i

nr

ELEMENTS OF QUATERNIONS.

ELEMENTS
OF

QUATERNIONS.
BY THE LATE

SIR

WILLIAM ROWAN HAMILTON, LLD.,

7

M.E.I. A.; D.C.L. CANTAB.;

FELLOW OF THE AMERICAN SOCIETY OF ARTS AND SCIENCES;
OF THE SOCIETY OF ARTS FOR SCOTLAND OF THE ROYAL ASTRONOMICAL SOCIETY OF LONDON
AND OF THE ROYAL NORTHERN SOCIETY OF ANTIQUARIES AT COPENHAGEN
CORRESPONDING MEMBER OF THE INSTITUTE OF FRANCE
HONORARY OR CORRESPONDING MEMBER OF THE IMPERIAL OR ROYAL ACADEMIES OF ST. PETERSBURG,
BERLIN, AND TURIN OF THE ROYAL SOCIETIES OF EDINBURGH AND DUBLIN
OF THE NATIONAL ACADEMY OF THE UNITED STATES
OF THE CAMBRIDGE PHILOSOPHICAL SOCIETY THE NEW YORK HISTORICAL SOCIETY
THE SOCIETY OF NATURAL SCIENCES AT LAUSANNE THE PHILOSOPHICAL SOCIETY OF VENICE
AND OF OTHER SCIENTIFIC SOCIETIES IN BRITISH AND FOREIGN COUNTRIES
;

ANDREWS PROFESSOR OF ASTRONOMY IN THE UNIVERSITY OF DUBLIN

AND ROYAL ASTRONOMER OF IRELAND.

SEOO3ST3D

EIDITIO3ST.

EDITED BY

CHARLES JASPER JOLY.

M.A.,

FELLOW OF TRINITY COLLEGE, DUBLIN

PROFESSOR OF ASTRONOMY IN THE UNIVERSITY OF DUBLIN,
AND ROYAL ASTONOMER OF IRELAND.
:

ANDREWS

VOLUME

I.

LONGMANS, GKEEN, AND

39,

PATERNOSTER ROW, LONDON.

NEW YORK, AND BOMBAY.
1899.

CO.,

V,

PRINTED AT THE

PON9ONBY

<Sk

WELDRICK,

ADVERTISEMENT TO THE FIRST EDITION,

IN

no instructions were

late father s Will

my

Writings, nor specially as to that of the

for his late fatal illness,

"

left

the publication of his

as to

ELEMENTS OF

would have been before now,

which, but

QUATEENIONS,"

in all their completeness, in

the hands of the Public.

My
to

H. Hamilton, who was named Executor, being too

brother, the Rev. A.

much engaged

in his clerical duties to undertake the publication, deputed this task

me.
It

was then

matter

First,

me

to consider

work

duty in this

and thirdly,

father,

found

as I

I could best fulfil

triple

to succeeding generations.

the

how

chiefly, to the dead; secondly, to the present public;

for

and

came

to the conclusion that

my

my

duty was

to publish

adding merely proof-sheets, partially corrected by

it,

and from which

removed

few typographical

errors,

my

late

and editing only in

the literal sense of giving forth.

Shortly before
subject of the

"

my

father

ELEMENTS."

death, I had several conversations with

In these he spoke of anticipated applications of

Quaternions to Electricity, and to

to develop,

bows

to be reserved for the

a good deal the nature of his

with

after dealing

questions in which the idea of Polarity

its

and mechanical
wrote out

all

as,

hands

of

another Ulysses.

own forthcoming Preface

more important

labour which the writing of the

And

all

applications which he never in his own lifetime expected

involved

him on the

topics,

is

to be able fully

He

also discussed

and I may intimate

that,

he intended to advert to the great

"ELEMENTS"

had

cost

him

labour both mental

besides a mass of subsidiary and unprinted calculations, he

the manuscript, and corrected the proof-sheets, without assistance.

here I must gratefully acknowledge the generous act of the Board of Trinity

College, Dublin, in relieving us of the remaining pecuniary liability,

ring the main expense, of the publication of this volume.

their intention to do so, gratifying as

it

Quaternions,"

defrayed

its

entire cost

me

was, surprised

bered that they had, after the publication of

my

father

and thus incur

The announcement
the

less,

when

former book,

"

of

remem

Lectures on

an extension of their liberality beyond what

781470

ADVERTISEMENT TO THE FIRST EDITION.

vi

was recorded by him at the end of his Preface to the Lectures," which doubtless
he would have acknowledged, had he lived to complete the Preface of the ELEMENTS."
"

"

He

intended also, I

know,

to express his sense of the care

bestowed upon the

typographical correctness of this volume by Mr. M. H. Gill of the University Press,

and upon the delineation of the figures by the Engraver, Mr. Oldham.
I

annex the commencement

of a Preface, left in

which he might possibly have modified

best fulfilled
in the

my

part as trustee of the unpublished

hands of the

manuscript by

or rewritten.
"

my

father,

and

Believing that I have thus

ELEMENTS," I

now

place

them

scientific public.

WILLIAM EDWIN HAMILTON.

January

1st,

1866.

PREFACE TO THE FIEST EDITION.

[1.]

THE volume now

principles as the

"LECTURES,"^)

about ten years ago

work can

The

is

founded on the same

which were published on the same subject

but the plan adopted

is

and the present

entirely new,

in no sense be considered as a second edition of that former one.

Table of Contents,

various

submitted to the public

by

collecting

Chapters and Sections,

may

into one view the headings of

suffice

to

give,

readers

to

acquainted with the subject, a notion of the course pursued

but

the

already
it

seems

proper to offer here a few introductory remarks, especially as regards the

method

of exposition,

which

it

has been thought convenient on this occasion

to adopt.
[2.]

The

present treatise

is

divided into Three Books, each designed to

develop one guiding conception or view, and to illustrate

it

but -not excessive number of examples or applications.

The

by a

sufficient

First

Book

relates to the Conception of a Vector, considered as a directed right line, in

space of three dimensions.

The Second Book introduces a

First Conception

of a Quaternion, considered as the Quotient of two such Vectors.

And

the

Third Book treats of Products and Powers of Vectors, regarded as constituting

********
********

a Second Principal

* This
fragment,

[W.E.

Hamilton.]

Form

of the Conception of Quaternions in Geometry.

by the Author, was found

in one of his

manuscript books by the Editor.

PREFACE TO THE SECOND EDITION.

SIR WILLIAM

BOWAN HAMILTON

died on

the

2nd

of

September, 1865,

He intended to have
leaving his great work on Quaternions unfinished.
an Index, and an Appendix con
added some account of the operator*
taining notes on Anharmonic Coordinates, on the Bary centric Calculus, and
v>

on proofs of his geometrical theorems stated in Nichol s Cyclopaedia. At

the time of his death, with the exception of a fragment of the preface, and
a small portion of the tahle of contents, all the manuscript he had prepared
was in type. As he rarely commenced writing before his thoughts were

no outline of the additions contemplated.

printed by direction of the Board of Trinity College,

fully matured, he has left

In

this edition,

Dublin, the original text has been faithfully preserved, except in a few
I have added notes,
places where trifling errors have been corrected.
distinguished in every case by square brackets, wherever I thought they
were wanted. I have rendered the work more convenient by increasing the

number

of cross-references,

by including in the page-headings the numbers

and

of the articles (for the original references are generally given to articles

not to pages), by dividing the work into two volumes, and by the addition
of an index.
The table of contents has been amplified by a brief analysis
of each article, designed as far as possible to assist the reader in following

and in recapitulating the arguments in the text. Hamilton indicated a

minimum course of study, amounting to rather less than 200 pages (or parts
of pages)," suitable for a first perusal, and he intended to have prepared a
"

table containing references to this course.

Such a table

will be

found at the

end of the table of contents, but for the convenience of students of Physics,
and of those desirous of obtaining a working knowledge of Hamilton s
powerful engine of research, I have amplified it somewhat, duly noting,
however, the

minimum

course.

* In the second volume

HAMILTON

I hope to devote an appendix to this important subject.

ELEMENTS OF QUATKRNIONS.

PREFACE TO THE SECOND EDITION.

I infer from the fragment of the author s preface that he proposed to

sketch an outline of the method of exposition, of an elementary character

and adapted
chiefly I

to those readers to

whom

address the following remarks

the subject

is

new.

To

those readers

According to the plan of this work, whenever a new conception or

notation

is

introduced, a series of illustrative examples immediately follows.

but occasionally a long and difficult

Intricate investi
investigation occurs even in the early parts of the book.
are
are
from the
which
omitted
everywhere
merely illustrative,
gations,

Most

of these involve

no real

difficulty,

selected course.

The

First

Book

deals with Vectors,

In a word,

considered without

reference to

concerned with the application of

=
to the algebra of vectors.
The sign - is first intro
the signs -f, -, and
= b.
duced, and the sign + follows from the formula of relation (b
a) + a
angles or to rotations.

it

is

and 4 (pp. 7-11) are occupied with a series of propositions con

the
commutative and associative laws of the addition of vectors, and
cerning
the multiplication of vectors by scalars, or algebraical coefficients.
Proposi
Sections 3

tions such as these often appear to a student to be

nately

it is

contrary.

mere truisms, and unfortu

not easy to find elementary examples to convince

The addition of vector-arcs, he will find on p. 156,

him
is

of the

not

com

few passages
associative.! With
mutative, though
noted in the table of a selected course, there is nothing in chaps. II. and III.
essential to a good knowledge of the subject.
They contain, however, an
the exception of a

it is

account of an extremely elegant theory of anharmonic coordinates, indepen

dent of any non-pro jective property, and intricate and powerful investigations
of geometric nets

and

The Second Book,

and

as involving

of systems of bary centres.

Quaternions considered as quotients of vectors,

angular relations. It opens with a first conception of a
treats of

quaternion as a quotient of two vectors, and thus the division of vectors is

introduced before that of multiplication, just as in the First Book subtraction
If q = j3
a is the quotient of two vectors, )3 and a, it is
precedes addition.
natural to define the product q.a by the relation q.a = |3. It is soon found,
:

any vector y is selected in the plane of a and )3, that the product q.y is a
vector in the same
plane whose length bears to that of y the same ratio as
if

the length of
*

|3

to that of a,

and which makes the same angle with y that

j3

In fact the commutative law of addition

depends on a property of a parallelogram, and there
on the validity of Euclid s fifth postulate. It does not hold except for Euclidean space.

fore ultimately

PREFACE TO THE SECOND EDITION.

makes with

Thus, from the

a.

first

xi

conception of a quaternion as a quantity

expressing the relative length and direction of two given vectors, we have
come to consider a quaternion as an operator on a special set of vectors, viz.
those in its own plane.
Observe that, so far, we have not arrived at the

conception of the product of two vectors, nor of the product of a quaternion

have only reached the limited conception of the
and an arbitrary vector.

We

product q

and a vector y in

y of a quaternion q

interpretation

is

assigned to q

y, as yet the product

its

plane,

j q
.

and while an

unknown.

is

After reviewing a class of quaternions derived by fixed laws from a

given quaternion, a special class of quaternions, called versors or radial
The product of a pair of versors is found
quotients, is considered in detail.
147) to depend on the order in which they are multiplied, that

(p.

is

qq

is

not generally equal to q q, or the commutative law of algebraic multiplication

is not true for versors, nor a fortiori for quaternions.

The

multiplication of a special set of versors of a restricted kind occupies

section 10, chap. I.

and on

p.

160 the famous formula

?=f=ft
is

dicular planes.

(A)

shown that the multiplication

it

related right versors

associative.

is

than two versors, and

it

yk=

/, and k are right versors* in three mutually perpen

This section contains the first example of a product of more

deduced, in which

law,

is

is

necessary to determine

Warned by
if

of these specially

the failure of the commutative

the remaining laws of algebra are valid

In algebra, if we first form the product Ic and then multiply

we have the same result as if we multiplied c by the product ah, and this

in quaternions.

by

#,

associative

law

expressed in symbols by the equation a

is

be

= ab

c.

This

also true for quaternions, and it may be regarded as the chief feature
which distinguishes quaternions from other systems of vector analysis. For
example, Grrassmann s multiplication is sometimes associative, but sometimes

is

it is
is

not.

It

is

necessary to prove, moreover, that quaternion multiplication

+ c) = ab + ac. This is not true if b and c are vector

distributive, or that a (b

arcs,

even

when

is

early investigations

number as shown on p. 156. Some

led him to a uon-distributive system of
a

of

Hamilton

multiplication

in 1830.f

Next a quaternion
the product of its tensor
*

is

decomposed
and its versor

in
;

two ways:

(1) in section 11, into

into the
(2) in section 12,

sum

of its

right versor turns a vector in

t Preface to Lectures

its plane thi ough a right angle.

on Quaternions, paragraph [41]. Scheffler has reproduced

this system.

PREFACE TO THE SECOND EDITION.

xii

scalar

and

may

shown,

This right or vector part, it is ultimately

right or vector part.
be identified with a vector at present it is regarded as a right

its

By the first of these

quaternion, or a quotient of two perpendicular vectors.
the multiplication of any two quaternions is reduced to the
decompositions,
"

arithmetical operation of multiplying

their

tensors,

and the geometrical

versors"; and by the second the addition of

quaternions is reduced to the algebraical addition of their scalar parts, and
Thus it is proved (Arts. 206,
the geometrical addition of their vector parts.

operation of multiplying their

207) that the addition of the vector parts is reducible to the addition of
vectors, and, as the addition both of scalars and of vectors is commutative
associative, so likewise is the addition of quaternions.

and

The

multiplication of right quaternions, or of the vector parts of quater

and, as any quaternion is
nions, is proved in Art. 211 to be distributive
;

sum

the

of a scalar

and a vector

multiplication of quaternions
follows,

some

of

is

part,

it

proved that the general

is also

distributive.

long

series

which are not easy, including Hamilton

of

examples
well-known con

struction of the ellipsoid.

Section 14

is

entitled

Standard Quadrinomial

On the reduction of the general Quaternion to the

Form (q = w + ix + jy + kz) with a First Proof of

"

the Associative Principle of the Multiplication of

Quaternions."

This proof

depends on the general Distributive Property lately proved, and on the

Associative Property of the particular set of versors i, /, k (Art. 161)
but
in chap. III. various proofs are given which are independent of these pro
;

perties.

The

The laws

first

associative

is sufficient

for all practical purposes.

now

of combination of quaternions are

(and subtraction)
is

proof

and

is

associative

established.

Addition

and commutative multiplication (and division)

;

distributive, but not commutative.

Passing over the second and third chapters in this Second Book, which
are chiefly complementary to the development of the theory, we find in
chap.

L, Book III., three lines of argument traced out in justification

In

of the identification of the vector part of a quaternion with a vector.

fact a

imposed, or a simplification is introduced, and this

simplification is shown to be consistent with the results

restriction

restriction

or

already obtained.*

is

In

much

the same

way

as a couple

We

or an

angular

have thus a new point of

Compare the note to p. 175, in which Hamilton remarks:
agreement, or of connexion, between right quaternions and their index-vectors, tending to justify the
ultimate assumption (not yet
made), of equality between the former and the latter."
"

PREFACE TO THE SECOND EDITION.

velocity

xiii

sometimes represented by a right Hue, a right quaternion and a

is

vector of appropriate length, perpendicular to the plane of the quaternion,

are now represented by the same symbol.*

The scope of the remainder of this volume is, I think, sufficiently indi
cated in the table of contents.
The foregoing sketch of the development of
the calculus of Quaternions necessarily presents but a meagre view of the
nature of this work however, my object has been to carry out, as far as I
;

could, the intention of

its illustrious

author expressed in the fragment of

his preface.

CHAELES JASPER JOLY.

THE OBSERVATORY, DUNSINK,
December, 1898.
*

With but

slight change, much of Books I. and II. might have been extended to space of
In Book III. advantage is taken of the peculiar simplicity of space of those dimensions
in which but one direction is perpendicular to a given plane, and a legitimate reduction of the number

n- dimensions.

of symbols

is

consequently made.

TABLE OF CONTENTS.

BOOK

I.

ON VECTORS, CONSIDERED WITHOUT REFERENCE TO ANGLES,

OR TO ROTATIONS,

....

CHAPTEB*

..

Pages

3-103

I.

FTJNDAMENTAL PRINCIPLES RESPECTING VECTORS.

SECTION!

On

the Conception of a Vector

and on Equality

of Vectors,

3-4

Geometrical subtraction of points Introduction

[Art. 1, Definition of a vector
Introduction of sign =, p. 3.
Art. 2, Equality of vectors
of the sign -, p. 3.
c = B
Art. 3, The equation
A, considered as an equidifference of points, admits
;

i>

of inversion and alternation, p. 4.]

SECTION

2.

On

Differences and

Sums

of Vectors,

taken two by two,

5-7

Art. 5, Intro
subtraction of vectors, p. 5.
[Art. 4, Definition of the geometrical
duction of sign +, regarded as a converse to
by the formula (b - a) + a = b, J p. 5.
Art. 6, The sum of any two co-initial sides of a parallelogram ABDC is the co-initial
,

diagonal, or

AB + AC = AD = AC + AB,
+ # = /8 +

commutative operation, or o

SECTIONS.

On Sums

of

Art. 7, Addition of

p. 6.
o,

two Vectors

p. 6.]

Three or more Vectors,

...

[Art. 8, Definition of the sum of any number of vectors, p. 7.

dition of vectors is an associative and commutative operation, p. 7.
Art. 11, Projection of a closed polygon, p. 8.]
polygon, p. 8.

* This
Chapter

is

7-8

Art. 9, The ad
Art. 10, Closed

may be referred to, as 1. i. the next as I. n. the first Chapter of the Second
Book, as II. i. and similarly for the rest.
t This Section may be referred to, as I. i. 1 the next, as I. i. 2; the sixth Section of the second
Chapter of the Third Book, as III. n. G and so on. [Article 180 is referred to as (180), and the
;

third sub-article of (180) as (180 (3.)).]

- a is added to a
[J This is, in words, b

sum
[

is I.

See

In (180

(6)

(3.)

and
)

and their sum

is b,

but not a

is

added to

(7).]

it is

shown that the

addition of vector arcs

is

not commutative.]

a and their

TABLE OF CONTENTS.

xvi

Pages

SECTION

4.

On

Coefficients of Vectors,

9-11

= (1 + m) a. if
is a positive or negative integer, p. 9.
[Art. 12, la + ma
Art. 13, The multiplication of vectors by positive or negative integers is a doubly
distributive operation,* p. 9.
Art. 14, This is also true when the coefficients are in
commensurable, i/a xa - (y x) a y (xa) = (yx) a. = yxa X (ft a) = xft xa, p. 10.
or
Art. 15, /3 = xa and a have the same or opposite directions according as x
0,
Art. 16, Division of a vector by a parallel vector, p. 10.
Art. 17, Intro
p. 10.
;

<

>

duction of the term scalar, p. 11.]

This short First Chapter should be read with care by a beginner

any misconception
being fatal to progress in the Quaternions. The
Chapter contains explanations also of the connected, but not all equally important, words

of the

meaning of the word

"Vector"

vector,"
"op
"pro vector,"
"transvector," "actual and null vectors,"
and successive vectors," "origin and term of a vector," "equal and unequal
vectors," "addition and subtraction of vectors,"
"multiples and fractions of vectors,"
&c.
with the notation B - A, for the Vector (or directed right line) AB and a deduction

or phrases,

"re

posite

of the result, essential but not peculiar^ to quaternions, that (what is here called) the
vector-sum, of the two co-initial sides of a parallelogram, is the intermediate and co-initial

The term u

diagonal.

Scalar" is

also introduced, in connexion with coefficients of vectors.

CHAPTER

II.

APPLICATIONS TO POINTS AND LINES IN A GIVEN PLANE.

SECTION

1.

On

Linear Equations connecting two Co-initial Vectors,

[Art. 18, Origin and change of origin, p. 12.

Art. 20, Or aa + b& = 0, p. 12.]
xa, p. 12.

Art. 19, Equation of a line OA

12-13

SECTION

2.

On Linear Equations between

three Co-initial Vectors,

13-19

a vector coplanar with a and

Art. 22, aa 4 bfr
+ bp
,
p. 13.
+ cy = is the condition that o, A, B, and c, or a, and 7, should be coplanar. Vector
is the condition of collinearity of
equation of a plane, p. 14. Art. 23, a -t- + c =
[Art. 21, aa

is

and c, p. 14.
Art. 24, a
b
c = BC : CA
AB, when a + b + c - 0, p. 14.
Art. 25, Equation of a line. Anharmonic of four collinear points, p. 15.
Art. 26,
Involution. Homographic division of lines, p. 16.
Art. 27, Vectors to points of
*
first construction, A = OA
Art. 28, Ratios
AB, % p. 17.
BC, B = OB
CA, c = oc
A, B,

of segments

OBC

OCA

of

sides

the triangle ABC, made by A , B , c

Hence, a b c =
Art. 29, Origin at mean point of triangle, p. 18.
Art. 30,

of

OAB, p. 18.

Projection of triangle and

After reading these two

mean

point, p. 19.]

Sections of the second Chapter, and perhaps the three

first Articles (31-33, pages
20-22) of the following Section, a student to whom the
subject is new may find it convenient to pass at once, in his first perusal, to the third
first

= w/8 ma is only true if a + )8 = + a. See (180 (3.)).]

[* m (0
a)
t Compare the second Note to page 206.
[J OA BC denotes the point of intersection of the lines OA and BC, DE ABC the point of intersec
tion of the line DE with the plane ABC.]
*

TABLE OF CONTENTS.

xvii
Pages

Chapter of the present Book

of the

and to read only the two first Articles (62, 63, pages 44-45)
Section of that Chapter, respecting Vectors in Space, before proceeding to the

first

Second Book (pages 107, &c.), which treats of Quaternions as Quotients of Vectors.

SECTION

On

3.

Plane Geometrical Nets,

.......

of second construction,
, Properties of points
Art. 32, Equation of six segments
AB, p. 20.
Art. 33, Points of second construction, A = OA B C , B
Art. 34, Construction of plane net, p. 22.]
p. 21.
[Art. 31

c"

A"

=AB

....

On Anharmonic

4.

in one Plane,
[Art. 35,
co-ordinates,

B"

=CA

CA

"

= OB

transversal, p. 21.
= oc * A B ,
C A , c
"

Co-ordinates and Equations of Points and Lines

Anharmonics of groups and

p

EC

made by

"

SECTION

=BC

20-23

xa + yb +

zc

p. 24.

...

pencils,

p.

23.

Art. 36,

23-29

Anharmonic

Art. 37, Anharmonic equation and co-

Art. 38, Examples. Line at infinity, p. 27.

ordinates of a right line, p. 26.
Arts. 39, 40, Independent geometrical interpretation of the coordinates of a line,
Art. 41, Concurrent and parallel lines, p. 29.]

p. 27.

SECTION

On Plane

5.

[Art. 42, Points

29-32

Geometrical Nets, resumed,

and

lines of a net are rational, p. 29.

Art. 43, Converse.

In

approximation to irrational points and lines, p. 30. Art. 44, Anharmonics

of nets are rational. Nets are homographic figures, p. 30.
Art. 45, Construction of
a net from one of its quadrilaterals, p. 31.]
definite

SECTION

........

On Anharmonic Equations and Vector

6.

in a given Plane,

Expressions, for Curves

.

32-43

[Art. 46, Conic touching sides of triangle at points of first construction, p. 32.
Art. 47, An ellipse or circle when o is interior to triangle, p. 33.
Art. 48, Case of
o exterior to triangle, p. 33.
Art. 49, Geometrical criterion of species, p. 34.
Art. 50, Determination of asymptotes of hyperbola and axis of parabola, p. 34.

Anharmonic symbol for tangent, p. 35. Art. 52, Vector to centre of conic,
Art. 53, Circumscribed conic having double contact with the inscribed conic,
Art. 54, Vector expression for a cubic with a conjugate point, p. 37.
p. 36.
Art. 55, Anharmonic generation of general cubic, p. 37.
Art. 56, Tangential Equa

Art. 51,
p. 35.

Art. 57, Examples of conies, p. 39.

Art. 58, Equations which are
tions, p. 38.
neither rational nor integral, p. 40.
Art. 59, Tangential equation deduced from

x*

y*

0, p. 41.

Art. 60, General Method, p. 42.

Art. 61, Application to

cubic, p. 42.]

other results of this Chapter, a theorem is given in page 38, which seems to
new geometrical generation of (plane or spherical) curves of the third order. The
anharmonic co-ordinates and equations employed, for the plane and for space, were sug
gested to the writer by some of his own vector forms ; but their geometrical interpretations
are assigned.
The geometrical nets were first discussed by Professor Mobius, in his Barycentric Calculus, but they are treated in the
present work by an entirely new analysis

Among

offer a

and, at least for space, their theory has been thereby

which we next proceed.

HAMILTON

ELEMENTS OF QUATERNIONS.

much extended

in the

Chapter

to

TABLE OF CONTENTS.

xviii

CHAPTER

III.

APPLICATIONS OF VECTOBS TO SPACE.

Pages

SECTION

On Linear Equations between

1.

Vectors not Complanar,

44-50

= 0, if a,
Art. 63,
7, 5 are any four vectors, p. 44.
[Art. 62, aa + b& +cy + d8
a+b-\-c + d=Q, if ABC and D are coplanar, p. 45. Art. 64, Intersections of lines
in a plane, p. 45.
Art. 65, For any five vectors aa + bfi + cy 4 dS 4 e* = 0, and
,

bc + d+e =

Art. 66, Intersection of a line and a plane, p. 47.

Q, p. 47.
67, Ratios of coefficients in (65) expressed as ratios of volumes of pyramids,
Art. 68, Convention respecting signs of volumes, p. 48.
Art. 69, I. OA.BC
p. 48.
4 OB CA 4 00 . AB = ; II. OA BCD - OB . CDA 4 OC DAB - OD . ABC =
III. OA .
Art.

UCDE 4 OB CDEA 4 oc DEAB + OB ABCD =

.

where in

I.,

A, B,

c are

in II., A, B, c, D are any four coplanar points;

D, E are any five points of space, p. 49.]

collinear points;
A, B, c,

been recommended to the student to read the first two Articles of this
first perusal of the Volume
and then to pass to the Second Book.

It has already

Section, even in his

SECTION
r

On Quinary Symbols

2.

any three

and in III.,

for Points

and Planes in Space,

Indefinite expression for a vector p

[Art. 70,

xaa 4

ybft

xa + yb

zcy

zc

4
+
+ wd + ve
Art. 72, (Q = (Q)
tvdS

50-55

vee
,

if
Symbol for a point; Congruence, p. 51.
)
= t (Q } 4
Art. 73, Condition of collinearity
(Q)
(U], where (Z7) = (11 11 1), p. 51.
of three points, p. 52.
Art. 74, Condition of coplanarity of four points, p. 52.
Art. 75, Quinary equation and symbol of a plane, p. 53.
Art. 76, Intersection of a

Art. 71, Quinary

p. 50.

and a plane,

line

given points,

SECTION

3.

p.

p. 54.

54.
Art. 77, Points, lines, and planes rationally related to five
Art. 78, Syntypical points, p. 55.]

On Anharmonic

Co-ordinates in Space,

.....

55-60

Art. 80,
[Art. 79, Quaternary or anharmonic symbol of a point in space, p. 55.
Anharmonic symbol and equation of a plane, p. 56. Art. 81, Trace of a plane on a
Art. 82, Formulae of collineation and coplanarity, p. 56.
given plane, p. 56.
Art. 83, Ratios of co-ordinates expressed as anharmonics of pencils of planes, p. 57.
Art. 84, Anharmonics of groups of points and of pencils of planes, p. 58.
Art. 85,

Quotients of two homogeneous and linear functions of the coordinates of a point

Art. 86, Reciprocal
expressed as the anharmonic of a pencil of planes, p. 59.
theorem and example, p. 59. Art. 87, Centre and plane of homology of pyramid,
p. 60.]

SECTION

4.

On

Geometrical Nets in Space,

Art. 89, First construction,

88, Net derived from five points, p. 61.
Art. 90, Second construction,
Art. 91, Quinary symbols of con
p. 62.
structed lines and planes, p. 63.
Anharmonic
Art. 92. Eight types of points P2.
Art. 93, Verification that the enumeration of the points PZ is
properties, p. 63.
[Art.

p.

61.

Art. 94, Arrangement of the points on lines and in planes, p. 76.

complete, p. 73.
Art. 95, Extension of results of
(44) and (45) to nets in space, p. 79.]

61-81

TABLE OF CONTENTS.
SECTION

On Bary centres

5.

plex Means
[Art. 96,
(1), (2)

Simple mean

Systems

of Points

of coinitial vectors

Mean

Com

and on Simple and

point

Gauche

81-87

quadrilateral
p. 81.

Barycentres, p. 84.]

On Anharmonic

6.

.........

Lines and planes determined by mean points of partial systems (3)-(6),

Art 97, Complex means

SECTION

of

of Vectors,

xix
Pages

and Curves

Equations, and Vector Expressions, of Surfaces

in Space,

...

87-96

It tiled quadric (1)-(10)

[Art. 98, Local and
Art. 99, Vector expression for a curve p =
twisted curve (11), (12), p. 87.
Line and plane (1) Ellipse and ellipsoid (2) quadric cone
for a surface p = (p(t, u}
Cubic cone (5) Ruled quadric (6)-(10), p. 94.]
(3), (4)

tangential equations uf surfaces;

<f>(t)

SECTION

7.

On

96-103

Differentials of Vectors,

[Art. 100, Definition of the differential of a vector;

surface (8)
Tangents to surfaces (9), p. 96.]

Hodograph

(5;

.Developable

An

application of finite differences, to a question connected with barycentres, occurs in

The anharmonic generation of a ruled hyperboloid (or paraboloid) is employed to
illustrate anharmonic equations ; and (among other examples) certain cones, of the second

p.

84.

have their vector equations assigned. In the last Section, a definition of

and scalars) is proposed, which is afterwards extended to differen
tials of quaternions, and which is independent of developments and of infinitesimals, but
Vectors of Velocity and Acceleration are mentioned
involves the conception of limits.

and third

orders,

differentials (of vectors

and a hint of Hodographs

is

given.

BOOK

II.

ON QUATERNIONS, CONSIDERED AS QUOTIENTS OF VECTORS,

AND AS INVOLVING ANGULAR RELATIONS,
.

CHAPTER

107-249

I.

FUNDAMENTAL PRINCIPLES RESPECTING QUOTIENTS OF VECTORS.

little, if any, of this Chapter II. i., should be omitted, even in a first perusal ,
contains the most essential conceptions and notations of the Calculus of Quater
nions, at least so far as quotients of vectors are concerned, with numerous geometrical
Still there are a few investigations respecting circumscribed cones, imagi
illustrations.

Very

since

it

nary intersections, and ellipsoids, in the thirteenth Section, which a student may pass
and which will be indicated in the proper place in this Table.

over,

SECTION

by

1.

Introductory Remarks

First Principles adopted from Algebra,

107-110

[Arts. 101-2, Comparison between Books I. and II. ; General principles adopted
Art. 103, I. Division must correspond to converse act of

definition, p. 107.

multiplication, or

by

=$

q,

p. 108.

Art. 104, II.

equal vectors give unequal quotients, p. 109.

Art.

Unequal vectors divided

105, III. Quotients equal to

C 2

TABLE OF CONTENTS.

xx

Pages
the same are equal to one another,

*v

p.

Art.

109.

106, IV.

and

*v

-:-=-;
a

Definition of addition, subtraction and division of quotients with a

Art. 107, V.

denominator, p. 109.

SECTION

First Motive for

2.

common

=
pan

-, p. 109.]

naming the Quotient

two Vectors

of

a Quater

110-113

nion,

The

Art. 109,
quotient of two vectors is not generally a scalar, p. 110.
But depends on their relative direction as well as on their relative length, p. 110.
Art. Ill,
Art. 110, The relative direction may be specified by a rotation, p. 111.
[Art. 108.

by its amount and the direction of its positive axis, p. 111.

Hence a geometric quotient is a function of four numerical elements or a

rotation is defined

Art. 112,

quaternion, p. 112.]

SECTION

Additional Illustrations,

3.

113-114

[Art. 113, Illustration of a quaternion by means of a desk on a table, p. 113.

Art. 115, Meaning of these elements,
Art. 114, Four numerical elements, p. 113.
Art. 116, A change in one of these alters the quaternion, p. 114.]
p. 114.
It

shown, by consideration of an angle on a desk, or inclined plane, that the complex

one vector to another, in length and in direction, involves generally a system of

is

relation of

four numerical elements.

Many

other motives, leading to the adoption of the name,

for the subject of the present Calculus, from its fundamental

the number
themselves in the course of the
Four," are found to present

connexion

"Quaternion,"

with

"

SECTION

On

4.

Equality of Quaternions

and on the Plane

work.

of a

Quater

115-119

nion,
in one plane
[Art. 117, The quotients of corresponding sides of similar triangles
Art. 118, But are unequal (and
are equal when the similarity is direct, p. 115.
Art. 119, Coplanar and diplanar
conjugate) when the similarity is inverse, p. 115.
Art. 120, Two geometric quotients can be reduced to a common
quaternions, p. 115.
denominator, and therefore their sum, difference, product, and quotients are quater

Art 121, Case of equal,

nions, p. 116.

p. 117.

Art. 122,

And of diplanar quaternions

X

reduced to a
8 HI a, 0,
HI

xa.

common

or

- and ft

q,

(I)

8 \\\q,

-, F
p. 118.

\\\

Art, 125, If

p. 118.

Art. 123, If

denominator, p. 117.

rt

--=

-,

being a sign of coplanarity, p. 117.

-, then, inversely,

ay

Art. 126,

=
xa

- and xq
a

|,

|||

a, 0,

and

Art. 124, Also

and

alternately,

<>

if

qx

5.
On the Axis and Angle of a Quaternion
a Right Quotient, or Quaternion,

SECTION

x isascalar,

and on the Index

...
;

p. 119.]

of

119-122

Art. 129, And denoted

axis of a Quaternion is defined, p. 119.
Art.
Art. 130, The angle of a quaternion, L q
v, P- 120.
Art. 132, Right quaternion or quotient of
131, Axis and angle of a scalar, p. 120.
perpendicular vectors Examples of geometrical loci expressed by the symbols Ax.
is deter
find l_, p. 121.
Art. 133, Index of a right quaternion;
right quaternion

[Arts. 127-8,
q, p. 120.

The

by Ax.

>

mined uniquely by

its

Index, p. 122.]

<

TABLE OF CONTENTS.
SECTION

On

6.

nion

The

134,

p. 122.

Ax. Kg- =

5-,

and conversely,

=-=

Art. 139,

Norm

= L

Kq =

q, if qis

Art. 141,

And

is

a
Art.

KK=

Which

K(-

(1(7)2, p.

SECTION

.q

q"

K^ =

L q

>

is

zero

or

<

if

g)

(1)

L q

0, if

and

= -K?;

<

p. 125.

or

-,

>

Art. 144, Of

of a

K- = =-\ qKq = Nq
Kq

127.]

On

7.

a scalar; and conversely, p. 124.

the equation of a plane

2
14o,K =

122-131

Ax. q
q

:
q"

Art. 143, Opposite of a quaternion, p. 126.

+-=

Quater

Art. 138, L

of a

Ax. q

Art. 136,

q p. 123.

More generally q + Kg

conversely.

126.

(2), p.

is

a scalar, p. 125.

is

p. 125.

a right quotient.
right line

p. 124.

+ Kq

And

Art 142,

in algebra q

= -

Art. 137, Conjugate of quaternion, p. 123.

Ax.

Art. 140, q

= -

reciprocal of q

As

Art. 135,

~i P- 123.

.......

the Reciprocal, Conjugate, Opposite, and

and on Null Quaternions,

[Art.

xxi
Pages

Radial Quotients; and on the Square of a Quaternion,

131-134

Art. 147,
1 are limiting
[Art. 146, Definition of a Eadial (or Versor), p. 131.
cases of radials.
Right radial, p. 132. Art- 148, The square of a right radial is- 1.

= - N#

Generally q*

number

finite

Equation of

if

of values of

-, p. 132.

Art. 149,

\/ -

has, in this Calculus, an in

two classes

circle, p. 133.

geometrical Reals and geometrical Imaginaries.

Art. 150, Reciprocal, conjugate and opposite of a Right

Radial, p. 134.]

......

8.
On the Versor of a Quaternion, or of a Vector
General Formulae of Transformation,

SECTION

and on some

135-143

[Art. 151-2, Radials and Versors differ only in the point of view from which they
are regarded, p. 135.
Art. 153-4, Deduction of properties proved in Arts. 147-8 when

a versor

regarded as a factor, p. 135.

is

Art. 155,

Uo denotes a unit vector having

n

same direction

i_

136.

Art.

156,

Uq =

= UK0,

L q and Ax. 113-= Ax. q; and conversely, p. 137.

the

TJfl

And U^ = U- = ==a
Ua

Art. 157, TJq depends only on relative direction, and

p. 136.

by

as a, p.

denotes the versor of

is

q,

uniquely determined

Art 158,

KU# =

= U-

or
Art. 159, Uxq = + TJq or - Uq according as the scalar x
0,
= U, p. 140.
a quaternion or a vector, p. 139.
Art. 160, U 3 =
Art. 161, Transformations of Uq.
Geoinetrica proofs and illustrations, p. 140.]
p.

whether g

In the
quaternion

138.

>

<

UU

is

foregoing Sections it is shown, among other things, that the plane of a

generally an essential element of its constitution, so that diplanar quaternions
but that the square of every right radial (or right versor] is equal to negative

five
is

are unequal ;

unity, whatever its plane may be.

in this as in several other

systems

The Symbol V
;

admits then of a real interpretation,

real, it is in the present Cal

but tuhen thus treated as

culus too vague to be useful

on which account it is found convenient to retain the old
signification of that symbol, as denoting the (uninterpreted ) Imaginary of Algebra, or
what may here be called the scalar imaginary, in investigations respecting non-real inter
:

sections, or

non-real contacts, in geometry.

TABLE OF CONTENTS.

xxii

Pages

On

Vector-Arcs, and Vector-Angles, considered as Represen

and on the Multiplication and
tatives of Versors of Quaternions
Division of any one such Versor by another,
143-156

SECTION

9.

.....

important, on account of its constructions of multiplication and division

that the product of two diplanar versors, and therefore of two such quaternions,

This Section

which show
is

is

not independent of the order of the factors.

Art. 163,
BA = r\ DC and n AC = n BD if r\ AB
[Art. 162, Vector Arcs, p. 143.
T
Arts. 164-5, Conditions of equality, p. 144.
Art. 166, Great semi
CD, p. 143.
r>

r>

Art. 167, Representation of the product of two versors by a

168, The multiplication of versors is not commutative,

circular arcs, p. 145.

vector arc, p. 146.

Art.

p. 147.

169,

= Kq q

versors qq

Art.

Art. 170, For right

Unless the versors are coplanar, p. 148.

=-7-5

qq

Art. 171, If their planes are at right angles, q q

p. 148.

qq is a right versor in the plane at right angles to hoth, 149.

=
sentation of division of versors, p. 150.
Art. 173, q
q)

Art. 172, Repre

(q"

and conversely,

q"

only

if
q"

q;

\\\

Art. 174, Vector angles, p. 151.

Art. 175, Employed to
Art. 177,
Art. 176, Second construction, p. 152.
p. 151.

p. 150.

construct the product q q,

Sense of the rotation produced

by q q, p. 152. Art. 178, Illustration by vector

Art. 179, Division of versors.
angles of the inequality of q q and qq
p. 153.
Conical rotation, p. 154. Art. 180, Sense of rotation round poles of sides of spheri
cal triangle.
Arcual sum. Spherical sum, p. 155.]
,

On

SECTION 10.

[Art. 181, Versors

i,

= - 1 ;/* = -!;

I.

ik
i,

j,

II.

Art.

j,ip. 157.

j,

fact,

Symbols,

QUATERNION may be

157-163

Art. 182,

p. 157.

ji=-k-

kj

= -i;

The

associative property of multiplication proved for

Formula i 2 -j^ = 2 = ijk = - 1. (A), p. 159. Art. 184,
(A), p. 161.]

make himself familiar with

to

ijk,

and k variously expressed as quotients,

III.
1.
II. ij=k;jk=i; ki=j.

183,

of the

=-

Fundamental Formula,

in the

In

Laws

and Jt Fundamental
and III. derived from

The student ought

Three Eight Versors, in three Rectangular

of

System

Planes; and on the

i*=ji = k* =

ijk

these laws, which are all included

=-

l.

(A)

symbolically defined to be a Quadrinomial Expression of the

form,

=w +

ix

+jy

-f

(B)

kz,

which w, x, y, z are four scalars, or ordinary algebraic quantities, while i, j, k are

three new symbols, obeying the laws contained in the formula (A), and therefore not

in

subject to all the usual rules of algebra

ij

=+

k,

but

since

we

ji=-k;

have, for instance,

and

.....

1 1
On the Tensor of a Vector, or of a Quaternion
Product or Quotient of any two Quaternions,

SECTION

and on the

Art. 186, Acts of Tension and Version.

[Art. 185, Tensor of a vector, p. 163.
Examples on the plane and sphere, p. 164. Art. 187, Tensor of a quaternion.
Examples, p. 167. Art. 188, Decomposition of a quaternion into Tensor and Versor,
Art. 189, Distinct and partial acts of Tension and Version, p. 169.
p. 169.
Art. 190, Transformations of Tq, p. 170.
Art. 191, Tensors and Versors of products

and quotients,

p. 171.

= -

Art. 192,

?
p.

173.

Art.

their indices,

= !?

K<?

ff

K<7.

K#

Examples on

circles,

9.

193, Quotient of two right quaternions is equal to the quotient of

1
if
or q
Art. 194, And q q = lq
I?
q = lq
lq, p. 174.

=, P-

175.]

163-176

TABLE OF CONTENTS.

xxiii
Pages

On the Sum or Difference of any two Quaternions; and on

the Scalar (or Scalar Part) of a Quaternion,
176-192
= q + q and
[Art. 195, For any tivo Quaternions addition is commutative, q 4- q
K(q + q} = K? + Kq, p. 176. Art. 196, Introduction^ symbol S. S = (1 + K) = SK.
Examples on the plane sphere and cyclic cone, p. 177. Art. 197, The sum of the
Art. 198,
scalars of any number of quaternions is the scalar of the sum, p. 185.
Art. 199, Or square, p. 187.
Scalar of a product, quotient, p. 186.
Art. 200,
Tensor and norm of the sum of two quaternions. Transformations, p. 189.]

SECTION 12.

On the Right Part (or Vector Part) of a Quaternion; and on

192-242
the Distributive Property of the Multiplication of Quaternions,
of
a
vector
and
at
Determinate
to
decomposition
along
right angles
[Art. 201,

SECTION 13.

a given direction, p. 192.

Art. 202,

Art. 203,

quotient, p. 193.

And

S -

of a quaternion into a scalar and a right

o

a and

#"

=V -

a are projections of OB along

Right line and cylinder, p. 194. Art. 204, Properties of

Art. 205, V is a distributive
Cylinders, spheroids, and ellipsoids, p. 196.

and

at right angles to OA.

\q.

= IVq + IVq p. 205. Art. 207, The general

symbol, p. 204. Art. 206, IV (q + q)
Art. 208, Quotient
addition of quaternions is commutative and associative, p. 206.
and product of two right parts. Spherical trigonometry, p. 207. Art. 209, Collinear
,

The multiplication

of collinear quaternions is doubly

Art. 211, Multiplication of right parts, p. 218.
Art. 213, Chords Art. 214, secants and
Art. 212, In general 2?2 / = 2qq
p. 219.
Art. 216, Ellipsoid, circular
Art. 215, tangent-cones to a sphere, pp. 220, 223, 225.
Art. 218,
Art. 217, Hamilton s construction, p. 232.
sections, cyclic planes, p. 230.

quaternions, p. 210.
distributive.

Art. 210,

Trigonometry,

p.

211.
,

Geometrical consequences of the construction, p. 235. Art. 219, Semi-axes. Spherical

Art. 220, Transformations of the Quaternion equation of the ellipsoid,
conies, p. 238.
p. 240.

On

SECTION 14.

the Reduction of the General Quaternion to a Standard

with a First Proof of the Associative Principle of

Quadrinomial Form

Multiplication of Quaternions,
Arts.

213-220 (with

.......

their sub-articles), in pp. 220-242,

may be

omitted at

first

242-249

reading.

Art. 222,
221, Standard quadrinomial form of a quaternion, p. 242.
Expression for derived functions. Law of the Norms, p. 243. Art. 223, Proof of the
associative principle of Multiplication.
Examples and Interpretations, p. 245. Art.
[Art.

224, Sketch of further treatment of the subject, p. 249.]

CHAPTER

II.

ON COMPLANAR QUATERNIONS,

OR QUOTIENTS OF YECTORS IN ONE PLANE; AND

ON POWERS, ROOTS, AND LOGARITHMS OF QUATERNIONS.

The

first six

Sections of this Chapter (II.

ii.)

may

be passed over in a

first

perusal.

SECTION 1.
On Complanar Proportion of Vectors; Fourth Proportional to
Third
Three,
Proportional to Two, Mean Proportional, Square Root
General Reduction of a Quaternion in a given Plane, to a Standard
250-256
Binomial Form,
;

Ail. 226, Fourth

[Art. 225, Quaternions and vectors in a given plane, p. 250.
Art. 227, Continued proportion.
proportional to three coplanar vectors, p. 250.
Mean proportional, p. 251. Art. 228, Standard binomial form. Couples, p. 254.]

TABLE OF CONTENTS.

xxiv

Pages

SECTION

2.

On Continued

Powers and Boots

[Art.

of

Proportion of Four or more Vectors

Quaternions; and Roots of Unity,

whole

256-261

Powers and

229,

Illustration, p.

256.

roots of quaternions, p. 256.

Art. 230, Cube roots.
Art. 231, Principal cube root, p. 257
Art. 232,
y/TT has

Art. 233, Fractional powers.

three real quaternion values, p. 257.
Art. 234, Scalar fractional exponents, p. 260.]

General roots of

unity, p. 258.

SECTION

3.

On

the Amplitudes of Quaternions in a given Plane and on

Expressions for such Quaternions, and for their
;

Trigonometrical

262-268

Powers,
[Art. 235, Amplitude of a quaternion, p. 262.
of amplitudes.
Examples, p. 264. Art. 237,
238,

SECTION

And with
4.

Powers

coplanar quaternion exponents, p. 268.]

the Ponential and Logarithm of a Quaternion

and on

with Quaternions for their Exponents,

of Quaternions,

268-276

Art. 240, Exponential property

[Art. 239, Ponential of a quaternion P(#), p. 268.
= P?
if
Art. 241,
+ iy} = P(a?) UP(# + iy}
P- 270.
Ill
and
connexion with trigonometry, p. 271. Art. 242, Imponential, p. 274
~2iy

P(? +

On

Art. 236, Addition aud subtraction

Art.
scalar, p. 266.

Powers with

?")

P?">

<?

TPO

Art. 243, logarithm of a quaternion, p. 275.]

5.
On Finite (or Polynomial) Equations of Algehraic Form, in
and on the Existence of n Heal
volving Complanar Quaternions
th
277-288
Quaternion Roots, of any such Equation of the n Degree,

SECTION

& q n + q\q n l -f ... + q n = has n

[Art. 244-8, Statements of the theorem thatF n #
277-78. Art. 249, Transformation of the equation, p. 278.
Art. 251, Construction of ovals, p. 279.
Art. 250, Geometrical statement, p. 279.

real quaternion roots, pp.

Art.
Art. 252, Geometrical proof, p. 280.
Art. 253, Quadratic equation, p. 281.
Art. 255, Construction of triangle, given base,
254, Second geometrical proof p. 284.
product of sides, and difference of base angles, p. 287.]
,

SECTION

6.

On

Equation

the n* - n Imaginary (or Symbolical) Roots of a Quaternion

n th Degree, with Coefficients of the kind considered in

of the

the foregoing Section,

.........

288-292

[Art. 256, Quaternion 01 couple equation equivalent to a system of two

The general
288.
Art. 257, Imaginary quaternion solutions.
p.
scalar

equations,

4
quaternion equation has w roots, p. 290.]

the Reciprocal of a Vector, and on Harmonic Means

with Remarks on the Anharmonic Quaternion of a Group
Four Points, and on Conditions of Concircularity,

SECTION

7.

Vectors

On

of
of

Art. 259, Reciprocal of a sum or

[Art. 258, Reciprocal of a vector, p. 293.
difference.
Anharmonic quaternion function of a group of four points, p. 293.
Arts. 260-1, Circular

and harmonic groups, pp. 295, 298.]

In this last Section (II. ii. 7) the short first Article 258, and the following Art. 259,
as far as the formula VIII. in p. 294, should be read, as a preparation for the Third
Book, to which the Student may next roceed.

293-300

TABLE OF CONTENTS.

CHAPTER
ON DIPLANAR QUATERNIONS,

OR

ITT.
OF

QUOTIENTS

xxv

VECTORS IN SPACE

AND

ESPECIALLY ON THE ASSOCIATIVE PRINCIPLE OF MULTIPLICATION OF SUCH

QUATERNIONS.

may he

This Chapter

omitted, in a

first

perusal.

Pages

SECTION

On some Enunciations

1.

of Multiplication of

...

of the Associative Property, or Principle,

Diplanar Quaternions,
= rq, and t = ss
q = sr, s

301-307

Art. 263, System of

t if
,
p. 301.
[Art. 262, q q
Art. 264, Enunciations of the
planes of the six quaternions q, r, s, s , q, t, p. 302.
and Art. 265,
302
principle in the form of theorems concerning vector-arcs, p.
and Arts. 266-7,
hexagon inscribed in a sphere, pp. 305,
Vector-angles, p. 304
;

306, and Art, 268,

SKCTION

pencil of six rays in space, p. 306.]

On some

2.

Geometrical Proofs of the Associative Property of

which are independent of the Distribu

........

Multiplication of Quaternions,
tive Principle,

Art. 270, Proof of the theorems of Art. 264

[Art. 269, Nature of proofs, p. 308.
of
by means of cyclic-arc properties of a sphero-conic, p. 308, and Art. 271, Of that
Art. 265 by its focal properties, p. 310, and Art. 272, Of that of Arts. 266-7 by stereo-

graphic projection, p. 310.]

SECTION

On some

3.

Additional Formulae,

....

313-317

Art. 274, Transformations of

313.
[Art. 273, Norm and Tensor of a vector, p.
the equation of the ellipsoid
Squaw root of a quaternion and of zero Biquater;

nions, p. 313.]

BOOK

III.

ON QUATERNIONS, CONSIDERED AS PRODUCTS OR POWERS

OF YECTORS AND ON SOME APPLICATIONS OF QUATER
;

NIONS,

CHAPTER
ON THE

321 to the end.

I.

INTERPRETATION OF A PRODUCT OF VECTORS OR POWER OF A VECTOR,

AS A QUATERNION.

The
SECTION

first six

On

1.

Sections of this Chapter ought to be read, even in a

a First

as a Quaternion,
[Art. 275-7,
vectors /3o =

HAMILTON

.....
Method

of Interpreting a

Introductory, p. 321.
:

Ra,

Art. 278,

tirst

Product of

perusal of the work.

Two

Vectors

First definition of a product of

p. 322.]

ELEMENTS OF QUATERNIONS.

TABLE OF CONTENTS.

xxvi

Pages

SECTION

On some Consequences

2.

of the foregoing Interpretation,

322-328

= Ka0, p. 322. Art. 280, Multiplication of vectors is doubly distri

[Art. 279, 0a
= )8a + )8a p. 323. Art. 281 Products of parallel and perpendicular
butive.
(a + a )
,

vectors.

Trigonometrical expressions, p. 323.

Examples.

2
reciprocal of a vector a

= - To 2

Ra = - =

Square and

Art. 282,

Examples on spheres,

p. 326.]

where
a
interpretation treats the product j3 a, as equal to the quotient
Ea) is the previously defined Reciprocal (II. ii. 7) of the vector a, namely a second
which has an inverse length, and an opposite direction. Multiplication of Vectors is
1

cT 1 (or
vector,

thus proved to be

(like that of Quaternions) a Distributive, but not generally a Commutative

The Square of a Vector is shown to be always a Negative Scalar, namely the
Operation.
negative of the square of the tensor of that vector, or of the number which expresses its

and some geometrical applications of this fertile principle, to spheres, &c., are
of the Eight Part of a Product of Two Coinitial Vectors, OA, OB, is
proved to be a right line, perpendicular to the Plane of the Triangle OAB, and representing
by its length the Double Area of that triangle while the Rotation round this Index, from
length

The Index

given.

This right part, or vector part, Vaj8, of the

the factors are parallel (to one common line)
and the scalar part,

the Multiplier to the Multiplicand,

when

product vanishes,
Saj8,

when they

positive.

are rectangular.

.......

On

a Second Method of arriving at the same Interpretation,

of Vectors,
Product
Binary

SECTION
a

is,

3.

Connexion between Right Quaternion and its Index.

= Iv
and conversely. II. I (v
Iv.
III. Iv
Iv = v
v)

I.

[Art. 283,

IRv, p. 329. Art. 284, The formula Iv

with the definition of 278, p. 329.]

Iv

v v

#a,

is

v.

Iv

of

329-330

It , if

IV. EIv

substantially identical

4.
On the Symbolical Identification of a Eight Quaternion with its
own Index and on the Construction of a Product of Two Rectangular

SECTION

Lines,

by

Third Line, rectangular to both,

.....

331-334

How

far is the substitution of a right quaternion for its index permis

Art. 286, This substitution is consistent with the First Book, p. 331.
Art. 287-8, And with the Second, p. 332.
Art. 289, And is therefore adopted, p. 333.

[Art. 285,
sible? p. 331.

Art. 290, Product of

SECTION

5.

On some

two rectangular

Simplifications of Notation, or of Expression, resulting

and on the Conception of an Unit-Line as a

..........

from this Identification

Right Yersor,

lines a line at right angles to both, p. 333.]

[Art. 291, Suppression of the symbols I and

of the terms Right Part and Index-vector, p. 335.

Ax. = UV,

Art. 292, and

p. 334.
Art. 293, Conception of a unit-line

as a right versor, p. 335.]

In this second interpretation, which is found to agree in all its results with the first,
is better adapted to an extension of the theory, as in the following Sections, to
ternary products of vectors, a product, of two vectors is treated as the product of the two
It is shown that, on
quaternions, of which those vectors are the indices (II. i. 5).

but

right
the same plan, the

Sum

of a Scalar and a YectoT

is

a Quaternion.

334-337

TABLE OF CONTENTS.
SECTION

6.

On

xxvii
Pages

..........

the Interpretation of a Product of Three or more Vectors

as a Quaternion,

337-356

[Art. 294, Multiplication of vectors is a special case of multiplication of Quater

nions.
Examples on products of three vectors, p. 337. Art. 295, Standard trinomial

Cartesian expressions.
Product of any number of vectors, p. 344.
Anharmonic
the product of sides of polygons inscribed in a sphere.

form for a vector.

Art. 296,

On

functions, p. 347.]

This interpretation

is

affected

by the substitution, as in recent Sections, of Right

Quaternions for Vectors, without change of order of the factors.

Multiplication of Vectors,
A vector,
like that of Quaternions, is thus proved to be an Associative Operation,
generally, is reduced to the Standard Trinomial Form,
p

ix

+jy +

kz

(C)

which i,j, k are the peculiar symbols already considered (II. i. 10), but are regarded
now as denoting Three Rectangular Vector-units, while the three scalars x, y, z are simply
rectangular co-ordinates; from the known theory of which last, ilhistrations of results
are derived.
The Scalar of the Product of Three coinitial Vectors, OA, OB, oc, is found
to represent, with a sign depending on the direction of a rotation, the Volume of the
so that it vanishes when they are complanar.
Parallelepiped under these three lines
Constructions are given also for products of successive sides of triangles, and other closed
in

polygons, inscribed in circles, or in spheres; for example,-a characteristic property of the

circle is contained in the theorem, that the
product of the four successive sides of an
inscribed quadrilateral is a scalar
and an equally characteristic (but less obvious) property
of the sphere is included in this other theorem, that the product of the Jive successive sides
of an inscribed gauche pentagon is
equal to a tangential vector, drawn from the point at
:

which the pentagon

Some general Formulae of Transformation of Vector

begins (or ends}.
Expressions are given, with which a student ought to render himself very familiar, as
they are of continual occurrence in the practice of this Calculus
especially the four
formulae (pp. 337, 339)
;

- 870
- /3S 7 a + 780)8
= S)87/? + 0Syap + 780)8?
(380)87
pSa)87 = V07Sop + V7oSp + Vo0Sp7

V 7Vo =
.

VyjSa

08)87

08)87

(D)

(E)

(F)

in

which

(G)

are signs of the operations

7, p are any four vectors, while 8 and
of taking separately the scalar and vector parts of a
On the whole, this
quaternion.
Section (III. i. 6) must be considered to be
an
(as regards the present exposition)
and if it have been read with care, after a perusal of the portions
important one
a,

previously indicated, no difficulty will be experienced in passing to any subsequent

applications of Quaternions, in the present or any other work.

SECTION

7.

On

the Fourth Proportional to Three Diplanar Vectors,

356-379

[Art. 297, The Quaternion fourth proportional to three diplanar vectors fia~ y.
Areas of spherical triangles and polygons, p. 356. Art. 298, Modifications when the
sides of the triangle are
Art. 299, Exceptional case
greater than quadrants, p. 372.
l

of quadrantal
triangle.

Fourth proportional to three rectangular

vectors, p. 377.]

8.
On an Equivalent Interpretation of the Fourth Proportional
Three Diplanar Vectors, deduced from the Principles of the Second
379-393
Book, .

SECTION
to

...

[Art. 300,

By Book

II.

a)

7 =

eu, u being a fourth proportional to three

ft

given rectangular unit-lines, p. 379.

Art. 301,

Before adopting

-7
a

=7
ft

if

a.

d2

TABLE OF CONTENTS.

xxviii

Pages

07/3

1,

p. 382.

Art. 302,

Two

tests are applied,

and found

to be satisfied,

Art. 303, Consequently, adopting the formula of 301, if v is a right

l
Art. 304, and as a further consequence (# 0)7 =
lv = u, p. 383.
5 -|- eu, u being now the same for all systems of mutually rectangular lines. Spherical
Art. 305, Series of spherical parallelograms, p. 387.
parallelograms, p. 385.

p. 382.

quaternion,

Art. 306, Construction of the series, p. 390.]

On

Method

of interpreting a Product or Function of

and on the Consistency of the Results of the
Interpretation so obtained, with those which have been deduced from
394-396
the two preceding Methods of the present Book,

SECTION

9.

the Third

Yectors as a Quaternion

....

[Art. 307, Fourth unit u, p. 394.]

These three Sections may be passed over, in a first reading. They contain, however,
theorems respecting composition of successive rotations (pp. 360, 361, see also p. 368)
expressions for the semi-area of a spherical polygon, or for half the opening of an arbitrary
pyramid, as the angle of a quaternion product, with an extension, by limits, to the
;

semi-area of a spherical figure bounded by a closed curve, or to half the opening of an

a construction (pp. 390-392), for a series of spherical
arbitrary cone (pp. 368, 369)
parallelograms, so called from a partial analogy to parallelograms in a plane; a theorem
;

(p. 393), connecting a certain system of such (spherical) parallelograms with the foci of a
and the conception (pp. 384, 394) of a
spherical conic, inscribed in a certain quadrilateral
Fourth Unit in Space (u, or + 1), which is of a scalar rather than a vector character, as
;

admitting merely of change of sign, through reversal of an order of rotation, although it

l
(ij
k) to Three Rectangular

presents itself in this theory as the Fourth Proportional

Vector Units.

SECTION 10.

On

[Art. 308,

quaternion
2C

2B

2A

the Interpretation of a

may

power of a vector

is

Power

of a

Vector as a Quaternion, 396-420

a quaternion,

be regarded as a power of a vector.

p.

396.

Art. 309,

and a

Proof of the equation

n
7^ 0* a = -

Art. 310, which includes the whole doctrine of Spherical

1, p. 399.
Art. 311, And arcual addition of
Spherical sum of angles, p. 404.
Triangles.
Art. 312, Solution of the equation of 309, p. 408.
Art. 313, Ex
sides, p. 407.
tension to spherical polygons, p. 414.
Art. 314, Geometrical loci and, p. 417.
Art. 315, Transformations connected with the powers of vectors, p. 420.]

It

may

be well to read this section (III. i. 10), especially for the Exponential
it establishes, between Quaternions and Spherical Trigonometry, or

Connexions which

by a species of extension of Moivre s theorem, from the plane to

For example, there is given (in p. 417) an equation of six terms,
which holds good for every spherical pentagon, and is deduced in this way from an
extended exponential formula.
The calculations in the sub-articles to Art. 312 (pp. 409414) may however be passed over; and perhaps Art. 315, with its sub-articles (p. 420).
But Art. 314, and its sub-articles, pp. 417-419, should be read, on account of the
exponential forms which they contain, of equations of the circle, ellipse, logarithmic spirals
(circular and elliptic), helix, and screw surface.
rather Polygonometry,
space, or to the sphere.

........

S MOTION 11. On Powers and Logarithms

some Additional Formulae,
[Art. 316,

may

suffice to

Diplanar Quaternions; with

Powers, logarithms, and trigonometrical functions of quaternions.

Supplementary formula,
It

of

p. 421.]

read Art. 316, and

its first

eleven sub-articles, pp. 421-423.

In

this

421-429

TABLE OF CONTENTS.

xxix
Pages

the simplest root, q

Section, the adopted logarithm, \q, of a quaternion q, is

of the

transcendental equation,
3

and

its

expression

found to be,

is

\q

in

which

lTq

+ /_q.UVq

(H)

and IT are the signs of tensor and versor, while L q

and IT. Such logarithms are found
be between

the angle of #, supposed

to be often useful in this

is

usually to
Calculus, although they do not generally possess the elementary property, that the sum
of the logarithms of two quaternions is equal to the logarithm of their product : this

apparent paradox, or at least deviation from ordinary algebraic

rides, arising necessarily

from the corresponding property of quaternion multiplication, which has been already
be
unless q and
not =
seen to be not generally a commutative operation (q
q"q

q"

And

complanar}.

here, perhaps, a student

might consider his

q"

first perusal of this

work

as closed.*

CHAPTER
ON

II.

DIFFERENTIALS AND DEVELOPMENTS OF FUNCTIONS OF QUATERNIONS; AND

ON SOME APPLICATIONS OF QUATERNIONS TO GEOMETRICAL AND PHYSICAL
QUESTIONS.
been already

It has

SECTION

1.

On

said, that this

Chapter

may

be omitted in a

first

perusal of the work.

the Definition of Simultaneous Differentials,

430-432

Art. 318, The usual definitions of differential

[Art. 317, Introductory, p. 430.
Arts. 319, 320,
coefficients and of derived coefficients being inapplicable, p. 430.
Differentials of quaternions are defined, p. 431.
tials,

Art. 321, Simultaneous differen

p. 432.]

SECTION 2.

Elementary Illustrations
Geometry,

of the Definition,

from Algebra and

432-437

[Art. 322,

Illustration

from Algebra,

p. 432.

Art. 323,

And from

geometry,

p. 435.]

In the view here adopted (cornp.

I. iii. 7), differentials

are not necessarily, nor even

But it is shown at a later stage (Art. 401), that the principles of this
Calculus allow us, whenever any advantage may be thereby gained, to treat differentials
as infinitesimals; and so to abridge calculation, at least in many applications.

generally, small.

* If he should choose to
proceed to the Differential Calculus of Quaternions in the next
Chapter (III. ii.), and to the Geometrical and other Applications in the third Chapter
(III. iii.) of the present Book, it might be useful to read at this stage the last Section
(I. iii. 7) of the First Book, which treats of Differentials of Vectors (pp. 96-102); and
perhaps the omitted parts of, the Section II. i. 13, namely Articles 213-220, with their
sub-articles (pp. 220-242), which relate,
among other things, to a Construction of the

by the present Calculus. But the writer Mill now abstain from
making any further suggestions of this kind, after having indicated as above what
appeared to him a minimum course of study, amounting to rather less than 200 pages (or
parts of pages) of this Volume, which will be recapitulated for the convenience of the
Ellipsoid, suggested

student at the end of the present Table.

TABLE OF CONTENTS.

xxx

Pages

SECTION

3.

On some

general Consequences of the Definition,

2
l
[Art. 324, Differential of j and of q~ , p. 438.

438-451

Art. 325, Notation proposed,

Art. 327, Differential quotients

440
Art. 326, Distributive property, p. 441.
and differential coefficients, p. 443. Art. 328, Differential of a function of several
Art. 329, Partial differentials, p. 446.
Art 330, Elimination
quaternions, p. 445.
p.

Art. 331, Differentiation of functions of functions, p. 449.]

of a differential, p. 448.

Partial differentials and derivatives are introduced

and

differentials of functions

of functions.

SECTION

4.

Examples

of

....

Quaternion Differentiation,

451-464

Art. 333, Transcendental

[Art. 332, Differentiation of algebraic and of, p. 451.
Art. 334, Differentiation of K#,
functions of a quaternion, p. 453.
Vg, T#, and
Art. 335, Differentiation of the axis and angle of a quaternion, p. 457.
p. 454.
Art. 337, And of
Art. 336, Differentiation of scalar functions of vectors, p. 459.
S<?,

U<7,

vector functions of scalars.

One
in situ

Examples,

of the most important rules

461.]

p.

to differentiate the factors of a quaternion product,

is,

thus (by p. 446),

The formula

(p.

439),

qq

d?

Aq

(I)

= -q~ l dq. q-\

(J)

for the differential of the reciprocal of a quaternion (or vector),

and so are the equations (p. 456),

Tq
and

(p. 454),

ya

^d*

also very often useful

\Jq

a*

is

(L)

It
q being any quaternion, and a any constant vector-unit, while t is a variable scalar.
is important to remember
(comp. III. i. 11), that we have not in quaternions the usual

equation,

dl*

^;
2

unless q

and Aq be complanar / and therefore that we have not generally,

*, if p

we have, in this Calculus, the scarcely less simple

useful in questions respecting orbital motion,

be a variable vector; although

equation, which

if

is

a be a constant vector, and

if

the plane of a and p be given (or constant).

TABLE OF CONTENTS.
SECTION

On

5.

xxxi
Pagee

..........

Successive Differentials and Developments, of Functions

of Quaternions,

Second

465-484

p. 465.

Art. 339, Simplification

differentials,
[Art. 338, Examples.
when cfiq = 0, or dq = const., p. 466. Art. 340, Special case of Taylor s theorem,
Art. 341, On the limiting ratio of two functions which vanish together.
p. 467.
Art. 342,

Geometrical example, p. 469.

differentials

and

Taylor

series

differences, p. 479.

several quaternions.

extended to quaternions,

Art. 344, Successive

Art. 345, Successive differentials of functions of

Examples of quaternion development,

Art. 343,

p. 473.

p. 476.

Scalar and Vector integrals, p. 479.]

In this Section principles are established (pp. 469-473), respecting quaternion functions
which vanish together and a form of development (pp. 473-475) is assigned, analogous*
to Taylor s Series, and like it capable of being concisely expressed by the symbolical
= e d (p. 480). As an example of partial and successive differentiation,
equation, 1 + A
;

the expression (pp. 480-481),

= rWkj-*rt,

which may represent any

vector, is operated

definite integration (pp. 482, 483), to

on

deduce the

and an application is made, by means of

known area and volume of a sphere, or

of portions thereof
together with the theorem, that the vector sum of the directed elements
of a spheric segment is zero: each element of surface being represented by an inward
normal, proportional to the elementary area, and corresponding in hydrostatics to the
;

pressure of a fluid on that element.

6.
On the Differentiation of Implicit Functions of Quaternions
and on the General Inversion of a Linear Function, of a Vector or
a Quaternion
with some connected Investigations,
484-568

SECTION

346-347, The

solution of a linear quaternion equation, or the Inversion of

[Art.
a linear quaternion function, p. 484.
Is reducible to the inversion of a linear vector
Art. 348, Transformations of the formula of solution, p. 489.
function, p. 485.
Art. 349, Quaternion constants or invariants of

Self -con jugate parts, p. 491.

Art. 350, Deduction of a symbolic cubic equation satisfied by fy and its conjugate
Art. 351, Case of a binomial function.
Fixed lines and planes, p. 497.
p. 494.
<>.

<

Case of equal roots. Depressed equation, p. 499. Art. 353, Case of

unequal roots, real and imaginary, p. 508. Art. 354, Case in which no root is zero.
Real and rectangular system for self -conjugate functions, p. 516. Art. 355, New
Art. 352,

Art. 356, Theorem of successively derived

proof of existence of the system, p. 523.
Art. 357, Rectangular and cyclic transformations, p. 527.
Art. 358,
lines, p. 525.

Focal transformations,

Art. 359, Passage from cyclic to focal forms, p. 535.

p. 530.
Bifocal and mixed transformations, p. 545.
Art. 361, Reciprocity of
Art. 362, Scalar function, linear with respect to vectors, p. 550.
forms, p. 547.
Art. 363, Linear and vector functions derived by differentiation, p. 551.
Art. 364,
Art. 360,

Solution of linear quaternion equation, p. 555.

Art. 365, Symbolic and biquadratic

equation, p. 560.]

In this Section

it is

shown, among other things, that a Linear and Vector Symbol,

and Cubic Equation, of the form,
<j>,

of Operation on a Vector, p, satisfies (p. 494) a Symbolic

= m - m + m ty* - 3
- m + 2 = ^,
w^r
<

<p

whence

m"(f>

(N)

another symbol of linear operation, which

* At a later
stage
new proof but still in
,

(Art. 375), a
a.

it

is

shown how

new Enunciation of Taylor

form adapted

(N

<J>

to quaternions.

to

deduce otherwise

Theorem

is

given, with a

TABLE OF CONTENTS.

xxxii
from
<p,

as well as the three scalar constants, m,

cubic (pp. 517, 518),

M=

m+mc+

m"c~

c3

m" .

The connected

algebraical

0,

(0)

and it is proved* (pp. 519, 520) that if

is found
have important applications
=
is said to he
independently of A. and p, in which case the function
S\(pp
while, in the same case,
self-conjugate, then this last cuhic has three real roots, c\, c 2 c 3
to

S/>0A,

<j>

the vector equation,

=
is

satisfied

(P)

0,

by a system of Three Real and Rectangular Directions

pp. 527, 528, and the Section III. iii.

surfaces of the second order, represented
Sptyp

Cases are discussed;

2
Op"

C",

7),

hy

in

and general forms

namely (compare

those of the axes of a (hiconcyclic) system of

the scalar equation,

which C and C are constants.

(Q)

(called cyclic, rectangular, focal, bifocal, &c.,

and
from their chief geometrical uses) are assigned, for the vector and scalar functions
one useful pair of such (cyclic) forms being, with real and constant values of g, A, /*,
(J>p

$p = gp

And

-f

VA/>,U,

gp

SAp/xp.

(R)

shown (pp. 560, 561) that if fq be a linear and quaternion function of a

then the Symbol of Operation, /, satisfies a certain Symbolic and Biquadratic

finally it is

quaternion,

q,

Equation, analogous to the cubic equation in

S/)0p

<J>,

and capable of similar applications.

simplified proof, of some of the chief results for this important case of selfis given at a later stage, in the few first sub-articles to Art. 415.

conjugation,

TABLE OF CONTENTS.
TABLE OF PAGES FOR THE FIGURES IN VOLUME

xxxiii

I.

TABLE OF A SELECTED COURSE.

This Course is recommended to those desirous of o 1
taining a good working knowledge of Quaternions.
The passages which are not distinguished jy a dagger are suitable for a first
perusal.

HAMILTON

ELEMENTS OF QUATERNIONS.

BOOK

I.

ON VECTORS, CONSIDERED WITHOUT REFERENCE TO ANGLES

OR TO ROTATIONS,

HAMILTON

ELEMENTS OF QUATERNIONS.

CHAPTER

I.

FUNDAMENTAL PRINCIPLES RESPECTING VECTORS.

SECTION

1.

On the Conception of a Vector and on

;

ART.

1.

direction, is

and

its final

JEquality of Vectors.

right line AB, considered as having not

said to be a

point B

VECTOR.

only length, but also

said to be its origin ;
(or

two extreme points or, more fully, to be the

its own origin from its own term
and, in con
it is also denoted by the
B
A a notation

to construct) the difference of its

result of the subtraction of

is

A vector AB is conceived to be

said to be its term.

is

Its initial point

symbol
formity with this conception,
which will be found to be extensively useful, on account of the analogies which
:

it

and algebraical
A
the extreme points A and B are

serves to express between geometrical

operations.
distinct,

(or

an

When

the vector A B or B - A

effective)

vector

but when

said to be

is

or B -

results, is said to

A and A -

an actual

B-^
A_ B

(as

a limit) those two

points are conceived to coincide, the vector

which then

vector

"

be

AA

or

A-

~~kevector.

A,

Opposite vectors, such as

null.

B, are sometimes called vector

and

revector.

AB and BA,
Successive

- A and c vectors, such as AB and BC, or B

B, are occasionally said to be
vector and protector : the line AC, or c - A, which
is drawn from the origin A of the first to the

term

c of the second,

transvector.

consider

being then said to be the

At a later stage, we shall have to

vector-arcs

and

vector-angles

but

at

present, our only vectors are (as above) right lines.

2. Two vectors are said to be EQUAL to each other, or the equation
AB = CD, or B - A = - c, is said to hold good, when (and only when) the origin
L>

ELEMENTS OF QUATERNIONS.
and term

[I.

i.

1,2.

of the one can be brought to coincide respectively with the corre

sponding points of the other, by transports

It follows that all null vectors are equal,

common

symbol, such as that used for zero

we may

so that

(or by translations) without rotation.

and may therefore be denoted by a

A-A

write,

B-B

= &c. =

but that two actual vectors, AB and CD, are

not (in the present full sense) equal to each
unless they have not merely equal

other,

but also similar

lengths,

If then

directions.

they do not happen to be parts of one

common line, they must be opposite sides of a parallelogram, ABDC the two
lines AD, BC becoming thus the two diagonals of such a figure, and conse
;

quently

bisecting

Conversely,

if

each

some point

in

other,

E.

the two equations,

D - E= E -

A,

and

are satisfied, so that the

commedial, or have a

two

- E = E -

lines

B,

AD and BC

common middle point

E,

are

then

they be parts of one right line, the equaTwo radii, AB, AC, of any one
tion D-C = B-AIS satisfied.
can never be equal vectors ; because their directions differ.
even

if

3.

An

equation between vectors, considered as

admits of inversion and alternation

1)

A -

then

or in symbols,

- C = B -

and

an

equidifference of points,

if

A,

D - B = c -

A.

B
Fig.

Two

B,

circle (or sphere),

5.

Fig.

and EF, which are equal to the same

each other ; and these three equal vectors

vectors, CD

also equal to
three parallel edges of a prism.

6.

third vector, AB, are

are,

in general, the

FUNDAMENTAL PRINCIPLES VECTORS.

ARTS. 2-5.]

SECTION

4.

In order to be able

we next

to write, as in algebra,

with

is co-initial

- A) =

(B

when a

define, that

AC which

- A

(c

first

to the

first

vector AB

- A =

if

B,

c - A,

subtracted

is

or from a third vector

it,

second vector, the remainder

term B of the

2.

Sum* of Vectors taken two by two.

Differences and

On

from a second vector

which

that fourth vector BC, which

is

term

c of the second vector

subtracted from a transactor (Art. 1), the remainder

is

is

equal to that

drawn from the

so that if a vector be

the projector corre

is

It is evident that this geometrical subtraction of vectors answers to

sponding.
and that, by such a decomposition
a decomposition of sections (or of motions)
;

of a null vection into two

(B

opposite vections,

A)

= (A - A) -

we have
-

(B

the formula,

= A - B

A)

an actual vector AB be subtracted from a null vector AA, the remain

If then we agree to abridge, generally, an expression
a to the shorter form, - a, we may write briefly, - AB = BA
of the form
a being thus symbols of opposite vectors, while a and - (- a) are, for
a and

so that,

der

is

if

the revcctor BA.

the same reason, symbols of one common vector

algebra, the identity.

5.

Aiming

still

at

-a

which we

we may

write, as in

agreement with algebra, and adopting on that account

(b

a)

shall say as usual that b

while relatively to

so that

a.

the formula of relation between the two signs, +

in

+ a =

-a

is

and

-,

b,

added to

and that

their

sum

is b,

be jointly called summands, we shall have the

they may
two following consequences
I. If a rector, AB or B - A, be added to
it

term B (Art. 1)
II. If

Q,

its

own

origin A, the

sum

is

its

and

pro-vector

BC be added to a vector AB, the

sum

is

the transvector AC;

or in symbols,
I.

In

fact,

the

(B

first

A)

+ A = B

equation

is

and

II.

(c

B)

+ (B - A) = c -

A.

an immediate consequence of the general formula

ELEMENTS OF QUATERNIONS.

when combined with

which, as above, connects the signs + and -,

tion (Art. 1) of

a vector as a difference of two points

the concep

and the second

2, 3.

i.

[I.

is

a result

same formula, combined with the definition of the geometrical subtraction

from another, which was assigned in Art. 4, and according
which we have (as in algebra) for any three points A, B, c, the identity,

of the

of one such vector

to

(c

It

is

A)

(B

= c -

A)

B.

clear that this geometrical addition of successive vectors corresponds (comp.

Art. 4) to a composition of successive vections, or motions;

and that the sum

and

revector) is

two

of

a null

BA + AB =

opposite vectors (or of vector

line

0, or

(A

so that

B)

(B

A)

0.

It follows also that the sums of equal pairs of successive

more

vectors are equal ; or

if

B - A = B -

and

A,

Fi 8- 7

fully that

the two triangles, ABC and A B C

- B =

then c - A = c - A

B,

being in general the two opposite faces of a

prism (comp. Art. 3).

6. Again, in order to have, as in algebra,
- B

(c

we

+ (B - A) = c -

the

to

sive

third to the

BV be
first,

first

the

and

vectors, AB, BC,

if

sum obtained by adding the

that fourth vector, AC, which

drawn from the origin A of the

the second.

B,

but not succes

equal to the second,

is

- B = c -

two successive

shall define that if there be

third vector

A, if c

first to

It follows that the

sum

is

the term c of

of

any two

co-

is

the

initial sides, AB, AC, of

any parallelogram ABDC,

intermediate and co-initial diagonal AD or, in symbols,

Fig.

8.

(c

A) + (B
A)

= D -

A, if

D -

= B - A

we have then (by 3) c - A = D - B.

The sum of any two given vectors has thus a

because
7.

of their order ; or, in symbols, a

|3

]3

a.

value

which

is

independent

If equal vectors be added to

equal vectors, the sums are equal vectors, even if the summands be not given
as successive (comp. 5) ; and if a null vector be added to an actual vector, the

sum

is

that actual vector

or,

in symbols,

abridge generally (comp. 4) the expression

+ a =
+ a to +

a.
a,

If then

and

if

we
still

agree to
denote a

FUNDAMENTAL PRINCIPLES VECTORS.

AETS. 5-9.]

then +

vector,

we

a,

and

-f

(+ a), &c., are other

symbols for the same vector

and

have, as in algebra, the identities,

-(-a)= + a,

=-(+)= -a,

+ (-a)

SECTION

On Sums of three
The sum

8.

(+ a)

+ (-

a)

0,

&c.

3.

or more Vectors.

of three given vectors, a,

/3,

7,

is

next defined to be that

fourth vector,
S =7

which

and in
the

is

(|3

-f

a),

or briefly,

=7 +

j3

a,

by adding the third to the sum of the first and second

manner the sum of any number of vectors is formed by adding
the sum of all that precede it also, for any four vectors, a, /3, 7, S,

obtained

like

last to

sum 8

+ a) is denoted simply byS + 7+|3 + a, without parentheses,

of summands.
number
any
9. The sum of any number of successive vectors, AB, BC, CD, is thus the
line AD, which is drawn from the origin A of the
and because, when
first, to the term D of the last
there are three such vectors, we can draw (as in
the

and

so

+ (7 +

on

13

for

BD of the (plane or
gauche) quadrilateral ABCD, and may then at plea

fig.

9) the

two

diagonals AC,

sure regard AD, either as the

sum

of AB, BD, or as

we

are allowed to establish the

the sum of AC, CD,
following general formu la
of association, for the case of any three summand lines, a, /3, 7 :

by combining which with the formula of commutation

(Art. 7), namely, with

the equation,
a +

|3

a,

which had been previously established for the case of any two such sum
mands, it is easy to conclude that the Addition of Vectors is always both an
Associative

and a Commutative Operation. In other words, the sum of any

vectors has a value which is independent of their order, and of

number of given

mode of grouping them so that if the lengths and directions

mands be preserved, the length and direction of the sum will

the

of the
also

sum

remain

ELEMENTS OF QUATERNIONS.

i.

3, 4.

except that this last direction may be regarded as indeterminate,

the length of the sum-line happens to vanish, as in the case which we are

unchanged

when

[I.

about to consider.

When any

n summand-lines, AB, BO, CA, or AB, BC, CD, DA, &c.,

one
order, are the n successive sides of a Mangle ABC, or of
any
a quadrilateral ABOD, or of any other closed polygon, their sum is a null line, AA
10.

arranged in

and conversely, when the sum

to zero, they

may

sides

successive

be

made
a

of

given system of n vectors

thus equal
any
any order, by transports without rotation) the

of

(in

is

polygon (plane or

closed

if there be given any such poly

a
pentagon ABCDE, it is possible to
gon (p), suppose
construct another closed polygon (p ), such as A B C D E

Hence,

gauche).

with an arbitrary

number of
be equal

initial point

AB

sides,

EA

but with the same

which new

(as vectors) to the old sides AB,

For example,

if

we

Fig. 10.

sides shall

EA, taken in

any

arbitrary order.

draw./b-wr successive vectors, as follows,

A B =

CD,

B C = AB,

D =

DE =

EA,

line E A

and then complete the new pentagon by drawing the

of the second figure (P

BC,

this closing side

will be equal to the remaining side

DE

of the first

figure (P).
11. Since a closed figure

projected

ABC

is still

a closed one,

when

all its

points are

on any assumed plane, by any system of

parallel ordinates

jected figure A B C

(although the area of the pro

.

to vanish),

may happen

it

number of given
be zero, and if we project them all
vectors a, ]3, 7,
on any one plane by parallel lines drawn from their
extremities, the sum of ihe projected vectors a j3 7

follows that

the sum of any

if

will likewise be null

so that these latter vectors,

Fi S- 1L
be so placed as to become the
of a closed polygon, even if they be not already such.
is considered as such a polygon, namely, as a triangle

like the former, can

successive sides
fig.

11,

A"B"C"

evanescent area

and we have the equation,

A"B"

as well as

AB +

sV +

A -

B"C"

0,

C"A"

0,

and AB + BC + CA =

0.)

(In
icitli

FUNDAMENTAL PRINCIPLES VECTORS.

ARTS. 9-13.]

SECTION

On
The

12.
(+ 1) a

a,

in like

or (+ 2)

regarded as a

a,

4.

Coefficients of Vectors.

simple or single vector, a,

and

is

also

denoted by la, or by 1 a, or by
a, is denoted by 2a, or
.

manner, the double vector, a +

&c.

coefficient

the rule being, that for any algebraical integer, m,

by which the vector a is multiplied, we have always,

la +

ma =

(1

+ m) a;

being here interpreted as in algebra. Thus, Oa = 0, the

zero on the one side denoting a null coefficient, and the zero on the other side
the symbol 1 +

denoting a null vector

because

by the

la + Oa = (1 + 0) a = la =

rule,

a,

and

Oa = a - a =

0.

= Oa = 0, we have (- 1) a = - a
1) a
Again, because (1) a + (- 1) a = (1
= - a == (1 - 2) a = (- 1) a = - a,
a
+
since
in
like
manner,
(- 2) a
(1)
(la)
=
=
we infer that (- 2) a
a
a
and generally (- m) a = - (ma),
(2a)
;

whatever whole number

m may

be

so that

we may, without danger

of

- la,
confusion, omit the parentheses in these last symbols, and write simply,

2a,

- ma.

13. It follows that whatever tivo whole numbers (positive or negative, or

Fig. 12.

null)

by

may be
and

/3,

represented

by

we have always,
na

and (compare

ma =

(n

and

n,

and whatever two

vectors

may be

denoted

as in algebra, the formulae,

m)

a,

n (ma) = (nm) a = ntna,

fig. 12),

m ()3

a)

= mf3

ma

so that the multiplication of vectors by coefficients is a doubly distributive operation,

C
HAMILTON S ELEMENTS OF QUATERNIONS.

ELEMENTS OP QUATEKNTONS.

10

numbers

at least if the multipliers be whole

[I.

4.

i.

a restriction which, however, will

soon be removed.

ma =

14. If

a multiple of a

the coefficient

]3,
;

and conversely

being

whole, the vector )3 is said to

still

(at least if the integer

&ero) 9 the vector a is said to

multiple of a vector is

7 = na then 7
9

is

be a sub-multiple of ]3.
said to be & fraction of that vector

a fraction of

which

]3,

is

be different from

multiple of a sub-

thus, if

/3

said to be multiplied

by the

fractional coefficient

product of this multiplication.

It follows that if x

vectors,

x} a,

(xa)

(yx) a

where x and y are incommensurable

yxa>

which include those of Art. 13, and

results

and 7

said to be the

is

then

xa = (y

ya

]3; also j3
"^

and y be any two fractions,

numbers being included), and if a and ]3

(positive or negative or null, whole

be any two

= ma, and

denoted as follows, 7 =
40

is

be

may

a)

(]3

xa

a?]3

be extended to the case

considered as limits of fractional

coefficients,

ones.

For any

15.

actual vector

and

a,

for

coefficient

any

of

any

of the

foregoing kinds, the product xa, interpreted as above, represents always a

vector

which has the same

/3,

has the opposite direction

a and

[3

each of

write
one

be any two actual

if

[|

0,

<

becoming

null if x

with directions

vectors,

which two cases we

/3

direction as the multiplicand-line a, if

0.

>

Conversely,

either similar or opposite, in

shall say that they are parallel vectors,

line],

we can always

which shall satisfy the equation

and

find, or conceive as
|3

= xa

or, as

we

found, a

coefficient

shall also write

it, /3

bear to + 1 the
positive or negative number x, so found, will
ratio, as that which the length of the line /3 bears to the length of a.

and the

results,

Hence

it is

from the

natural to say that this

of the vector

division

coefficient

shall

we

0,
<

= ax

the quotient which

;

and

to write,

)3

-f

a,

or x =

or x =

a,

shall have, identically, as in algebra, at least if the divisor-line a

vector,

same

~
x =

an actual

is

by the parallel vector a

/3,

accordingly,

so that

if

a (because both are then parallel, in the usual sense of the word, to

common

16.

but

0,

and
(|3

the dividend-line

if

a)

= -a =
a

j3,

which will afterwards be extended, by

|3

and

be

be parallel thereto, the equations,

Xa

a =

definition, to

= x
a

the case of non-parallel

FUNDAMENTAL PRINCIPLES VECTORS.

ARTS. 13-17.]

We may

vectors.

that the vector a

number x

write also, under the same conditions, a =

we

division of

have these other

shall

The

and may say

the quotient of the division of the other vector

8
-

17.

0&

is

so that

11

x = (ax =) p,

by the

identities,

ax

and

positive or negative quotient, x

]3

a.

= -, which

is

thus obtained by the

one of two parallel vectors by another, including zero as a limit, may

SCALAR; because it can always be found, and in a certain

also be called a

sense constructed,

by the comparison of positions upon

one

common

scale (or axis)

or can be put under the form,

A
AC
=
B A
AB
c

where the

three points, A, B, c, are collinear (as in the figure

simply the REALS (or real

Algebra; but, in combination with
the not less real VECTORS above considered, they
form one of the main elements of the System, or Calculus,

scalars are,

annexed).

therefore,

quantities) of

work

Such

In

-Tiff.

to

13*

which the present

shown, at a later stage, that there is

an important sense in which we can conceive a scalar to be added to a
Scalar plus Vector"
vector; and that the sum so obtained, or the combination
relates.

fact it

will be

"

is

a QUATERNION.

c 2

ELEMENTS OF QUATERNIONS.

12

CHAPTER

[I.

ir.

1,2.

II.

APPLICATIONS TO POINTS AND LINES IN A GIVEN PLANE.

SECTION

On Linear Equations
18.

WHEN

o,

and future

connecting two Co-initial Vectors.

several vectors, OA, OB,

that point
as OA,
such
vector,

point

sections

that all the vectors

1.

are all

drawn from one common

; and each particular

said to be the Origin of the System

is

In the present
the
be
not expressed,
contrary
always suppose,
which we may have occasion to consider, are thus

said to be the vector of its own term, A.

is

we

shall

a, ]3,

drawn from one common

if

origin.

But

be desired to change that origin

if it

o,

we shall only have to subtract, from

without changing the term-points A,
one common vector w, namely, the old vector oo
each of their old vectors a,
will be the new
o
since
the
of the new origin
w, (3
remainders, a
.

vectors a

]3

of the old points A, B,

a = OA =

A- o =

19. If two vectors a,

and

j3,

(A

o)

For example, we

(o

o)

annexed), then (by 16, 17) their quotient

to

an

have

opposite, so that the three points, o, A, B, are

situated on one right line (as in the figure

such as x

shall

= OA - oo = a - w.

or OA, OB, be thus drawn from a given origin

their directions be either similar or

if

o>,

and conversely, the equation

j3

~~I?

some

is

positive or negative scalar,

xa, interpreted

with this reference

B
origin 9 expresses the condition of collinearity, of the points o, A,

= 0, x - 1, corresponding to the particular positions, o

particular values x
of the variable point B,

20.

The

whereof the

linear equation,

indefinite right line

aa +

where a and

OA

is

the

connecting the two vectors a and

it thus

more symmetric form, when we write

b are two scalars, of

o.

the

and

A,

locus.
/3,

acquires a

bfl

which however only the

ratio is important.

POINTS AND LINES IN A GIVEN PLANE.

AETS. 18-21.]

The
x=

two points A and

condition of coincidence, of the

1, is

now

answering above to

more symmetrically,

or,

B,

13

a + b = 0.

Accordingly,

when

a=-

the linear equation becomes

b,

b(f$

a)

or

0,

]3

a = 0,

we do not suppose

that both the coefficients vanish

and the equation
=
B
should coincide with the point A a
a, or OB
OA, requires that the point
j3
case which may also be conveniently expressed by the formula,
since

B = A
coincident points

being thus treated

(in notation at least) as equal.

In general,

the linear equation gives,

a

OA +

OB =

and therefore

0,

SECTION

two (actual and

form aa +

b(3

and

= BO

OA.

Co-initial Vectors.

co-initial) vectors, a, f3,

equation of the

two scalar coefficients a

2.

On Linear Equations between three

21. If

be not connected by any

with any

0,

whatever, their
directions can neither be similar nor opposite
to each

they therefore determine a

which the (now actual) vector,

other

plane AOB, in

represented by the sum aa. + bfl, is situated.

For if, for the sake of symmetry, we denote
this

sum by

third scalar,

the symbol - cj, where

and 7 = oc

is

so that the three co-initial vectors, a,

aa +

and

if

some

c is

some third

15

vector,

/3,

bfi

7, are connected

+ cy =

by the

linear equation,

we make
~~
OA _

OJ3

_
c

then the two auxiliary points, A and B

will be situated (by 19)

indefinite right lines, OA, OB, respectively

and we

oc = OA + OB
so that the figure

A OB

is

shall

on the two

have the equation,

(by 6) a parallelogram, and consequently plane.

ELEMENTS OF QUATERNIONS.

14

22. Conversely, if c be
ordinates,

CA and CB

the three scalars,

w?/

pom

to the lines

a, b, c,

in the plane AOB,

OA and

OB,

so as to satisfy the

[I.

ii.

2.

we can draw from

it

the

and can determine the

ratios of

two equations,

OA

OB

OA

OB

which we shall have the recent expressions for OA OB with the relation
=
oc OA + OB as before and shall thus be brought back to the linear equation
= 0, which equation may therefore be said to
+
the condition
aa +
after

express

cy
bj3
of complanarity of the four points,

And

o, A, B, c.

we

if

write

it

under the

form,

xa +

and consider the vectors a and

+ %y =

2/j3

0,

as given, but 7 as a variable vector, while

/3

then be the

x, y, z are variable scalars, the locus of the variable point c will

given plane, OAB.

is

23. It may happen that the point c is situated on the right line AB, which
here considered as a given one.
In that case (comp. Art. 17, fig. 13), the

AC

must be equal

quotient

to

some

scalar,

suppose

so that

we

shall

have an

equation of the form,

*V ~~ CL

or y = a

t,

by comparing which

f>

n*.

(j3

or (1 -

a),

form with the

last

we

equation of Art. 21,

t)

a +

tfi

y =

linear

see that the condition

of collinearity of the three points A, B, c, in the

given plane OAB, is expressed by the formula,
a + b +

0.

may also be thus written,

OB
OA
-a -b
=

This condition

*.

k*. f

and under

known
24.

or

this last

form

OA
it

OB

Fig. 16.

expresses a geometrical relation, which

is

otherwise

to exist.

When we

have thus the two equations,

aa +

so that the three

and may

fy3

+ cy =

0,

co-initial vectors

and
a,

]3,

a + b +

0,

7 terminate on one right

on that account be said to be termino-collinear,

if

we

line,

eliminate,

POINTS AND LINES IN A GIVEN PLANE.

ARTS. 22-25.]

successively

and separately, each

of the three scalars a,

b,

c,

15

we

are conducted

to these three other equations, expressing certain ratios of segments

b (/3

a)

c (y

a)

a (a

c (y

0,

7)

+ a

j3)

=
7)

(]3

(a

|3)

0,

or

Hence

#.AB

C.AC = C.BC + ff.BA = #.CA

follows this proportion, between coefficients

We

might

also

= BC

CA

and

+ #. CB.

segments,

AB.

have observed that the proposed equations give,

_

bfi

+ aa
_ cy

+ cy

aa +

bfi

whence
AC - 7 - a
b
=
~
-a
AB
a
+
p
25. If

we

still

treat a

and

j3

= -

-,

&c.

as given, but regard

y and oc as

variable, the

equation

will express that the variable point c is situated someivhere

right line AB, or that it has this line for its locus

AB into segments, of which the

variable quotient

7
then, in like manner,

we

shall

"

we agree

to

is,

let its vector be,

+ y$
x + y

have this other

ratio of segments,

/
*

C B

then,

indefinite

of a

AC

If,

on the

divides the finite line

H.

and

line,

_=

it

CB

Let c be another point on the same

while

employ, generally, for any group offour

collinear points,

the notation,
(ABCD) =

AB CD
AB AD
=
BC DA
BC DC
.

so that this symbol,

(ABCD),

may

be said

to denote the anhartnonic function, or anharmonic quotient, or

ELEMENTS OF QUATERNIONS.

16

simply the anharmonic of the group.

A, B, c,

we

[I. n.

2.

in the present

shall have,

case, the equation,

i

ACBC

AC

,.

AC

26.

When

becomes

(as is

= yx

-?c B

CB

xy

the anharmonic quotient becomes equal to negative unity the group

If then we have the two equations,
well known) harmonic.

_ xa + yfi
x + y

_xa-

yfi

x - y

the two points c and c are harmonically conjugate to each other, with respect
and when they vary together, in consequence
to the two given points, A and B
;

11

of the variation of the value of

they form

-,

indefinite right line AB, divisions in involution

a well-known sense), on the

(in

the double points (or foci) of this

involution, namely, the points of which each is

points A and B themselves. As a verification, ^
if

we denote by

point

its

own conjugate, being the

^~Q

--g

the vector of the middle

JJL

of the given interval AB, so that

f3

Fig. 17.

=
fj.

JUL

or

a,

ju

= \

(a

|3),

we

easily find that

MC _ MB

y-ju_^-#_/3-ju
y +

j3-ju
so that the rectangle

gate points,

square

of

generally,

c,

under the

we

interval

MB

distances MC,

from the centre

half the
if

- M
y

JC

MC

of the

MC"

two

variable but conju

of the involution, is equal to the constant

between the two double

A, B.

points,

More

write

_
x + y
I

where the anharmonic quotient

Ix

my

iix*
-,

is

any constant

scalar,

then in another

xy

phraseology, the points c and c will form, on the inde

finite line AB, two homographic divisions, of which A and B are still the double

known and modern*

points.

More generally

still, if

xa +

we

establish the
,

1/8

Ixa

x + y
-

being

still

the two given

constant, but
lines,

variable points c

and

not

Ix

variable, while a

AB and A B

two equations

+
+ my

= OA

|3

= OB

and 7 = oc

by the
move along one common line.

are then homographically divided,

now supposed

* See the Geometric

Superieure of

M.

to

Chasles, p. 107.

(Paris, 1852.)

two

POINTS AND LINES IN A GIVEN PLANE.

ARTS. 25-27.]

27.

"When

tion a + b

+ cy =
subsists, without the rela
three co-initial vectors
the
coefficients, then

the linear equation aa +

between

its

17

bfl

complanar, but they no longer terminate on one right line ;

|3,
their term-points, A, B, c being now the corners of a triangle.
r
In this more general case, we may propose to find the vectors
]3
7 of

7 are

still

the three points,

A = OA
that

is

B = OB

BC,

CA,

to say, of the points in

= OC AB

which the

lines

drawn from the

origin o to the three corners

the three respectively
intersect
of the triangle

The

opposite sides.

OAA

three colUneations

&c.,

give (by 19) three expressions of the forms,

a = #a,

|3

7 =

f/|3,

27,

where #,

y, z are three scalars, which it is required to determine by means of the

three other collineations, A BC, &c., with the help of relations derived from the
1
principle of Art 23.
Substituting therefore for a its value or a, in the given

linear equation,
linear equation

and equating
which

results,

to zero the

we

7,

/3,

of the coefficients of the

b/3

each in

+ cy =
its

from the original equation

turn,

find the values,

-a

y =

b
z

whence the sought vectors are expressed

F.
+

new

namely,

arW +
and eliminating similarly

sum

-t

in either of the

two following ways

, _
7 ~ a+

+a

or

+ aa

II.
b

In

fact

we

and by the
on the line
sions for a,

see,

by

+ a

one of these expressions for a

other expression for the

BC.

As another

same vector a

verification,

we may

that

is

on the

line

OA

that the same point A is

observe that the last expres
,

7 coincide with those which were found in Art. 24, for a, )3, 7
on
the particular supposition that the three
themselves,
points A, B, c were
/3

collinear.
HAMILTON

ELEMENTS OF QUATERNIONS,

ELEMENTS OP QUATERNIONS.

18

28.

We may

write the last equations for a

and we

/3

7 under the form,

see that they then give the

required ratios, as follows

BA _
r

whence we obtain

c~

CB _ a
7
B !"?

at once the

known
BA
AC

AC

C B

It

o.

is

equation of six segments,

CB

AC

BA

CB

coefficients

BB

from the same

easy also to infer,

the following proportion of

as the condition of concurrence of the three

right lines A A
point, such as

2.

next propose to determine the ratios of the segments of the

made by the points A B c
For this purpose, we

sides of the triangle ABC,

may

[I. n.

and

cc

in a

common

ratios of segments,

areas,

= OBC OCA OAB,

:

in which we must, in
general, attend to algebraic sigm ; a triangle being
conceived to pass (through zero) from positive to negative, or vice versa, as com
pared with any given triangle in its own plane, when (in the course of any
continuous change) its vertex crosses its base.
It may be observed that with
this convention

(which

generalformula) we

is,

in fact, a necessary one, for the establishment of

have, for

any three points, the equation

ABC + BAG =
exactly as

we had

(in

Art. 5) for any two points, the equation

AB + BA =

More

fully,

we

0,

0.

have, on this plan, the formulae,

ABC = - BAC = BCA = - CBA = CAB = - ACB

and any two complanar

triangles, ABC,

negative ratio, according as the

two

A B

bear

rotations,

to

which

each other a positive or a

may

be conceived to be

denoted by the same symbols ABC, A B C are similarly or oppositely

29. If A and B bisect respectively the sides BC and CA, then
,

a = b =

directed.

c,

whence the known theorem follows, that the three bisectors

of the sides of a triangle concur, in a point which is often called the centre of
and which
gravity, but which we prefer to call the mean point of the triangle,

and

c bisects

AB

POINTS AND LINES IN A GIVEN PLANE.

ARTS. 28-30.]

here the origin

is

j3

At

o.

the same time, the

first

19

expresions in Art. 27 for

7 become,
~
7 -

whence

this other

known theorem

results, that

bisectors trisect each

other.

30.

The

linear equation

between

a, |3,

7 reduces

itself,

in the case last

considered, to the form,

a +
the three vectors

|3

+ 7 =

a, |3, 7,

or OA + OB + oc =

0,

or OA, OB, oc, are therefore, in this case, adapted

(by Art. 10) to become the successive sides of a

triangle,

by

without

transports

accordingly,

if

we complete

rotation

and

(as in fig. 19) the

parallelogram AOBD, the triangle OAD will have

the property in question.
It follows (by 11)
that if we project the four points o, A, B, c, by

JTl.

any system of parallel ordinates, into four other

points, o /? A /} B /? c /5 on any assumed plane, the
sum of the three projected vectors, a,, j3,, 7,, or
&c., will be null; so that

OA,

we

shall

have the new

linear equation,

or,

OA+
and

in fact

it is

evident (see fig. 20) that the

will be the mean point

projected mean point o/

of the projected triangle,

have

A /}

B,,

cy

We

shall

also the equation,

where

a/

- a =

^enoe
or the ordinate of the

Fi

o^ -

OA =

(oy A

AA,)

(oo y + O,A) =

20>

AAX -

oo,;

oox = 1 (AAX + BB, + c^),

mean point of a

triangle is the

mean of the

ordinates of

three corners.

tht

ELEMENTS OF QUATERNIONS.

20

SECTION

a, b, c
c"

do

Resuming

more general

the

are supposed to be unequal,

,the lines

triangle

or

B C

may

3.

Nets.

case of Art. 27, in which the coefficients

we may next

inquire, in

A B meet respectively the sides

n.

3.

On Plane Geometrical
31.

[I.

seek to assign the vectors

a",

]3",

7"

what points

A",

B",

BC, CA, AB, of the

of the points of inter

section (comp. 27),

B"=C

The

first

expressions in Art. 27 for

|3

-CA,

give the equations,

\c"

(c

a) /3

bfi

0,

(a

b)

(c

7 + cy =

whence
b(3-cy _
b - c

(a

+
(a

b)

y
b)

(c

a) fi

a)

but (by 25) one member is the vector of a point on BC, and the other of a
and similarly
of
each therefore is a value for the vector
point on B C
a"

A",

for

]3"

and

7".

We may
b(3
-

therefore write,

cy

_cy -aa

c-a

_ aa

and by comparing these expressions with the second

in Art. 27,

we

see (by 26) that the points

harmonic conjugates (as they are indeed

bfi

a-b

.
!

set of values of

A",

known

B",

to be)

/3

are, respectively, the

of the points A B , c ,

c"

POINTS AND LINES IN A GIVEN PLANE.

AKTS. 31-33.]

B, c

with respect to the three pairs of points,

notation of Art. 25, we have the equations,
(BA

= (CB

CA")

= (AC

AB")

And

because the expressions for

equation between those three vectors,

a",

|3",

BC")

c,

= -

A,

21

so that, in the

1.

conduct to the following linear

7"

(6-c)a +(c-a)/3 +(a-ft) /

-<>,

with the relation

(6

between

the recent

_ a +
)

(a

B",

line

A"BV

ABC

sides of a triangle

and because we have

A"C

CB

known theorem, that

of the dotted lines in
one
by

the
represent any rectilinear transversal, cutting

may

/r

.,

arrive (by 23) at this other

BA^ ~ a -/3
,

b)

are collinear, as indicated

c"

A",

21.

fig.

The

32.

we

its coefficients,

the three points

c)

AC

a"

b
-, as before,

while -j- = -, and -7- =

we arrive at this other equation of six
BA
c
c B
a
segments, for any triangle cut by a right line (comp. 28),
,

CB
B

which again agrees with known

33. Eliminating

and 7

j3

AC

BA"

I77^

c7B

results.

and 7 between,

either set of expressions (27) for

j3

with the help of the given linear equation, we arrive at this other
equation, connecting the three vectors a, ]3 7
,

= - aa +

(c

a) |3

(a

b}

Treating this on the same plan as the given equation between

that

if (as

in

fig.

"

= OA

BC

"

new

= OB

"

cV,

aa
W"

>

A B

^0

~2b

///

c+a

~~2c + a + b

or

nf.f
r*

find

may be expressed in either

shorter, but the second is, for

first is

some purposes (comp. 34, 36), more convenient

=

= oc

points of intersection

of the two following ways, whereof the

"

we

we make,

21)

the vectors of these three

a, j3, 7,

2aa + bQ + cy
^
*

*.

,.,.,
/*

2bfi

+ cy + aa

*y

2cy

ELEMENTS OF QUATEBNIONS.

22

And

the three equations, of which the following

(b

c)

(2b + c + a)

"_

with the relations between their

fi

(2c

"+

as indicated

",

we have

(a

b)

and

respect to B
c",

3>

4.

"=

0,

which are evident on inspection,

coefficients

and

we may

(25, 31),

(B

we

(c

collineations,

(a

a]
A",

b)

A"B

"C

B"C"

A"

",

Also, because

(c

a)

"

are harmonically conjugate with

two other

similarly for the

compared with c

",

employed

two points

see (by 26) that the

and

the two expressions,

__
we

b)

three of the dotted lines in the figure.

by

one,

+ a +

show (by 23) that we have the three additional

C"A"

is

[i.

A and with A

pairs of points,

B",

",

so that, in a notation already

write,

"C

A")

(C

B"

B")

= (A We") = -

1.

with any four complanar points, o, A, B, c, of

which no three are collinear, we can (as in fig. 18), by what may be called a
First Construction, derive from them six lines, connecting them two by two,
34. If

and

begin, as above,

new

intersecting each other in three

new
new

lines,

and then by a

we may connect

by three

which will give, by their intersections with the former

lines, six

points,

A",

and with the given

tion,

points,

these

Second Construction (represented in

".

We

points,

fig.

proceed to connect these with each other,

might
by sixteen new

namely, the four dotted lines of

three should be

21),

drawn from each

of

lines, or lines of

a Third Construc

fig. 21, and twelve other lines, whereof

the four given points and these would
:

be found to determine eighty-four new points of intersection, of which some

may be seen, although they are not marked, in the figure.

But however far

as to

these processes of linear construction

form what has been

may

be continued, so

called* a plane geometrical net, the vectors of the

points thus determined have all one

common property

namely, that each can

be represented by an expression of the form,

ma

+ zcy
xa + yb + zc

where the

coefficients x, y, z

are

ylfi

some whole numbers.

In

fact

we

see (by 27,

31, 33) that such expressions can be assigned for the nine derived vectors,
*

By

Prof. A. F. MOBIUS, in page 274 of his Barycentric Calculus (der barycentrische Calcul,

Leipzig, 1827).

POINTS AND LINES IN A GIVEN PLANE.

ARTS. 33-35.]
a

",

which alone have been hitherto considered

and

23

it is

not

difficult

from the nature of the calculations employed, that a similar result

must hold good, for every vector subsequently deduced. But this and other

to perceive,

connected results will become more completely evident, and their geometrical
signification will be better understood, after a somewhat closer consideration
of anharmonic quotients, and. the introduction of a certain system of (inharmonic
co-ordinates, for points

and

one plane, to which we shall next proceed

lines in

reserving, for a subsequent Chapter,

any applications

of the

same theory

to

space.

SECTION

On Anharmonic
35. If

equations

4.

Co-ordinates and Equations of Points and

Lines in one Plane.

we compare the last equations of Art. 33 with

of Art. 31, we see that the harmonic group BA

the corresponding
on the side BC of
CA",

into another such group,

fig. 21, has been simply reflected
on the line B C by a harmonic pencil of four rays, all passing through

the triangle ABC in

B A

"C

A",

the point o

and similarly

for the other groups.

More

generally, let OA, OB,

ABCD, be any pencil, with the point o for certex ; and let
in
the new ray OD be cut, as in fig. 22,
by the three sides of the triangle ABC,
the three points A I} B l5 d
let also
oc, OD, or briefly o

zcy

VAl
so that (by 25)

we

shall

Ul

yo + zc

have the anharmonic quotients,

and

let

us seek to express the two other vectors of intersection,

a view to determining the anharmonic ratios of

the groups on the two other sides.
The given

equation (27),
aa +

+ cy =

b(3

0,

shows us at once that these two vectors

oc.

71

y) fy3

+ zaa

"

(-*)**

are,

/3i

and

7,,

//Q>

with

ELEMENTS OF QUATERNIONS.

24

[I. n.

4.

whence we derive (by 25) these two other anharmonics,

(CD AE,)
so that

we have the

relations,

+ (CA BA,) = (BC

But

in general, for

any four

+ (BA

ACI)

collinear points A, B,

CAI)

c, D,

1.

it is

not

difficult to

prove that

AB
BC

whence by the

AC
CD +

Bl)

= DA

definition (25) of the signification of the

following identity

is

symbol (ABCDJ, the

derived,

=
(ABCD) + (ACBD)

Comparing

CB

this, then,

1.

with the recently found relations, we have, for

the following anharmonic equations

(CAB

B!)

fig.

22,

= (CA BA

X)

from known principles) the anharmonic

of the group does not change, when we pass from one side of the triangle,

and we

see that (as

was

to be expected

considered as a transversal of the pencil, to another such side, or transversal.

may therefore speak (as usual) of such an anharmonic of a group, as being

We

at the

of the

same time the Anharmonic of a Pencil ; and, with attention to the order
rays, and to the definition (25), may denote the two last anharmonics

by the two following

reciprocal expressions

At

(o

CABD)

= ;

(o

BACD)

= -

with other resulting values, when the order of the rays

understood that
(O.CABD) = (cVsV),
if

the rays oc, OA, OB, OD be cut, in the points c\ A

is

changed

it

being

E\ D\ by any one right

line.

36.

The

expression (34),

xaa + yb$ + zcy

xa 4 yb + zc

may

represent the vector of any point p in the given plane, by a suitable choice

POINTS AND LINES IN A GIVEN PLANE.

ARTS. 35-36.]

of the coefficients x,

For

or simply of their ratios.

z,

i/,

25

since (by 22) the three

complanar vectors PA, PB, PC must be connected by some linear equation, of

the form
f
a
PA + b PB + c PC = 0,
.

or

which gives
aa + b

Q+

a + b +

we have only

c>

to write

and the proposed expression for p will be obtained. Hence it is easy to infer,
on principles already explained, that if we write (compare the annexed
fig. 23),

PI

we

shall have,

= PA

P 2 = PB

BC,

or p ly

points of intersection, p,, p 2 , p 3

_
=

P 3 = PC AB,

with the same coefficients xyz, the following expressions for

the vectors OP!, op 2 OP S

pl

CA,

ybfi

/o 2 ,

of these three
c

zcy + xaa

+ zcy
_ xaa

p3

+ ybfS

A
which give at once the following anharmonics of

Fig. 23.

pencils, or of groups,
(A

BOCP)

(B

COAP)

(c

AOBP)

= (BA

CPI)

=-

x
;

y
whereof we see that the product

is

unity.

determine the position of the point p,

given

and therefore

proportional

to

those pencils,

when

Any

appears that the three coefficients x, y, s, or any scalars

them, of which the quotients thus represent the anharmonics of

may

it

be conveniently called the

that pointy p, with respect to the given triangle

itself

may

be denoted by the Symbol,

(a?,

HAMILTON

two of these three pencils suffice to

the triangle ABC, and the origin o are

ELEMKNTS OF QUATERNIONS.

?/,

z}.

ANHARMONIC CO-ORDINATES
and

origin

of
while the point P

ELEMENTS OF QUATERNIONS.

26

With this

notation, the thirteen points of

A"

"

(1, 0, 0),

(0, 1, 1)

(0, 1,

(2, 1, 1),

37. If PI

if

(1, 0, 1),

1, 0, 1),

For,

"

(1, 2, 1),

(1, 1, 0)

(1,

(I, 1, 1)

1, 0)

(1, 1, 2).

in the given plane,

P2 =

(a?,

a),

y*>

tXi

with the two former points, or

p^.

c"

o =

(0, 0, 1),

4.

scalar coefficients, then the following third point,

P =

right line

(a?i, ?/i, *i),

and u be any two

is collinear

and P 2 be any two points

PI

and

"

21 come to be thus symbolized

fig.

(0, 1, 0),

= (-

B"

1),

[I. n.

(in

other words)

situated on the

is

we make

if

and
x^aa +
x^a +

x z aa +

x*a +

xaa +

xa +

these vectors of the three points PiP 2 p are connected

t

(x^a

.)

in which (comp. 23), the

pi

+ u

sum

p cannot be collinear with

(x z a

.)

pz

of the coefficients

P2

PI,

unless

its

(xa

by the
+

is zero.

.)

linear equation,

Conversely, the point

co-ordinates admit of being thus

It follows that if a variable point p be obliged

expressed in terms of theirs.
to move along a given right line PiP 2 or if it have such a line (in the given plane)
,

for its locus, its co-ordinates xyz

must

degree, with constant coefficients

which, in the

may

known notation

of determinants,

be thus written,

or,

satisfy a homogeneous equation of the first

more

x,

y,

<TI,

2/1,

z
1

fully,

or briefly

where

/,

m, n are three constant

lx

my

scalars,

-f

nz,

whereof the quotients determine the

which is thus the locus of the point p. It is natural to

the equation, which thus connects the co-ordinates of the point p, the Anhar-

position of the right line A,

call

monic Equation of the Line

A and we
;

shall find

it

convenient also to speak of

POINTS AND LINES IN A GIVEN PLANE.

AETS. 36-39.]

the

coefficients

of that

Line

line

may

also be denoted

A
38.

being the Anharmonic Co-ordinates

n, in that equation, as

m,

Z,

which

For example, the three

m,

[/,

27

by the Symbol,

].

sides BC, CA,

AB

of the given triangle

have

thus for their equations,

x =

and

y =

0,

0,

[0, 1,

lines B C

A",

y +

-# =

[-

1],
C",

manner, for their

18, have, in like

fig.

-x =

0,

AB

B",

[0, 0, 1].

[0, 1, 0],

three additional lines OA, OB, oc,

equations and symbols,

y - 2 =

The

for their symbols,

[1, 0, 0],

The

0,

x - y=

0,

1, 0, 1],

[1,

0,
1, 0].

of fig. 21, are

+ v - y =

0,

0,

H-

0,

or

A"B

B"C

"C
",

2/

",

s-3#
[-

line

A"B"C"

0,

of

what

of the

same

figure, are in like

manner

+ # - 3y =
[1,

x + y - 3z =

0,

3, 1],

[1, 1,

3]

Q,

is

we may remark
is

",

3, 1, 1],

x + y +
Finally,

"B

by the equations and symbols,

represented

and the

C"A

"A

[1,1,-!];

[I,-!,!],

[-1,1,1],
the lines

0,

that on the

or

[1,1, 1].

same plan, the equation and the symbol

often called the line at infinity, or of the locus of all the infinitely

distant points in the given plane, are respectively,

ax + by +

cz

0,

and

[a, b, c]

because the linear function, ax + by + cz, of the co-ordinates x, y, 2 of a point p

in the plane, is the denominator of the expression (34, 36) for the vector p of
that point

so that the point p is at

an

infinite distance

from the origin

when, and only when, this linear function vanishes.

39. These anharmonic co-ordinates of a line, although above interpreted

o,

(37)

with reference to the equation of that line, considered as connecting the co

ordinates of a variable point thereof, are capable of receiving an independent

For the three points L, M, N, in which the line A.

geometrical interpretation.
or [/, m, n~], or Ix + my + nz = 0, intersects the three sides BC, CA, AB of the
E

ELEMENTS OF QUATERNIONS.

28

given triangle ABC, or the three given lines # =

evidently (on the plan of 36) be thus denoted

[I. n.

^ =

0,

0, z

4, 5.

(38),

may

L =

But we had

(0, n,

it is

= (-

N =

O,/);

n,

(tn,

/,

0).

also (by 36),

A"=(0,

whence

- w)

1,-1);

B"-(-

1,0,1);

c"=(l,-l, 0);

easy to infer, on the principles of recent articles, that

!L -

"

L-

"

"

with the resulting relation,

(BA"CL)

(CB"AM)

(AC"BN)

1.

proved, with the help of the

known and general relation between segments (32), applied to any two trans
and LMN, of any triangle ABC. In fact, we have thus the two
versals,
40. Conversely, this last equation

is easily

A"B"C"

equations,
BA"

CB"

AC"

A"C

B"A

C"B

BL

CM

LC

MA NB

on dividing the former of which by the

article results.

We

might therefore in

AN _
~

latter,

this

the last formula of the last

way have been

led, without

consideration of a variable point P, to introduce three auxiliary scalars,

I,

any
m,

n,

defined as

having their quotients

-,

equal respectively, as in 39, to the

three auharmonics of groups,

(BA"CL),

and then

it

(CB"AM),

(AC"BN)

would have been evident that these three

scalars,

/,

m, n

(or

any

others proportional thereto), are sufficient to determine the position of the right
line A, or LMN, considered as a transversal of the
given triangle ABC so that
:

they might naturally have been called, on this account, as above, the anharmonic co-ordinates of that line. But although the anharmonic co-ordinates of
a point and of a line may thus be independently defined, yet the geometrical
utility of such definitions will be found to depend mainly on their combination :
or on the formula Ix +

my +

nz =

of 37,

which

may

at pleasure be considered

as expressing, either that the variable

point (x, y, z) is situated somewhere upon
the given right line [/, m,
or
else
that the variable line [I, m, n] passes, in
n~\
;

some

direction, through the given point (xs y, z).

PLANE GEOMETRICAL NETS.

ABTS. 39-42.]

and A 2 be any two right

41. If AI

A2

or (in

may

infinite distance),

where
I,

and u are

given plane,

same plane, which passes through

other words) which concurs with them (at a

then any third right line

section A!

lines in the

29

in the

the inter
finite or

be represented (comp. 37) by a symbol of the form,

Or, what comes to the same thing,

scalar coefficients.

m, n be the anharmonic co-ordinates of the line A, then (comp. again

the equation

37),

/.

must be

if

satisfied

because,

- n\m^) + &c.

(ftiift?

if

(X, Y, Z) be the supposed point common to the

three lines, the three equations

IX +
must

mY+ nZ =

Y+

n,Z =

l,X +

0,

Y + n Z = 0,
2

Conversely, this co- existence will be possible, and the three lines
point (which may be infinitely distant), if the recent

co-exist.

have a

will

l,X + m,

0,

common

For example, because

condition of concurrence be satisfied.

\_a,

b, c]

has been

seen (in 38) to be the symbol of the line at infinity (at least if we still retain
the same significations of the scalars a, b, c as in Articles 27, &c.), it follows
that

A =
are symbols of
all

m,

[/,

n],

two parallel

and
lines ;

[I

+ ua,

ub,

because they concur at

n +

uc~],

infinity.

In general,

problems respecting intersections of right lines, collineations of points, &c.,

in the given plane,

when

between

eliminations

not here delay

same as in the

treated

by

this

anharmonic method, conduct to easy

on which we need

linear equations (of the scalar kind),

the mechanism of such calculations being for the most part the

known method

of trilinear co-ordinates

although

(as

we have

seen) the geometrical interpretations are altogether different.

SECTION

On Plane Geometrical
42. If

we now

geometrical

nets,

resume, for a

and

tively,

which have

lines, in

Nets, resumed.

moment, the consideration of those plane

in Art. 34
and agree to call those

which were mentioned

points

5.

the given plane, rational points and rational

lines,

respec

their anharmonic co-ordinates equal (or proportional) to whole

ELEMENTS OP QUATERNIONS.

30

numbers

but to say that a point or line

is

irrational, or that

irrationally related to the

it is

when

given system of four initial points o, A, B,

anharmonic co-ordinates are not thus all equal (or proportional)

its

integers

it

we may assume

clear that ichatever four points

is

however far the construction of the net

we

5.

because then the (inharmonic quotients, which were discussed in the

last Section, are rational ;

lines

[I. n.

which result will

begin with such

wards conduct

as initial,

c,

to

and

be carried, the net-points and net-

may

In fact?
be rational, in the sense just now defined.
and the subsequent eliminations (41) can never after

all

to any, that are of the contrary

kind

the right line which

two rational points being always a rational line and the point of
two rational lines being necessarily a rational point. The
assertion made in Art. 34 is therefore fully justified.
connects

intersection of

43. Conversely, every rational point of the given plane, with respect to the

four assumed initial points OABC, is a point of the net which those four points
determine. To prove this, it is evidently sufficient to show that every rational

=
point A!

on any one side BC of the given triangle ABC, can be so

(0, y, z),

constructed.

as in fig. 22,

Making,

= OAi

B!

we have

and

CA,

Ci

= OAi

AB,

(by 35, 36) the expressions,

B =
I

from which

it is

vV>

0,

y -

*),

(*,

y, 0)

easy to infer (by 36, 37], that

C B!

BC =

(0, y, z

y},

BC =

CI

and thus we can reduce the linear construction

in which the two whole

numbers y and

may

(0,

z, z)

of the rational point (0, y,

z),

be supposed to be prime to each

1), which has

depend on that of the point (0, 1,

been
constructed as A
It follows that although no
already
irrational point Q of the plane can be a net-point, yet every
other, to

such point can be indefinitely approached to,

linear construction

which

can be included within a quadri

or even within a triangular interstice

so that it

lateral interstice PiP 2 p 3 P4,

PiP 2 P 3

interstice of the net can

Analogous remarks apply

cide

with

net-lines,

then writing

and

if

be

made

as small as

to irrational lines in the plane,

we may

desire.

which can never

coin

be indefinitely approximated to by such.

any three collinear points of the net, so that the formulae
p be
fourth net-point (x, y, z ) upon the same line,

but

44. If p, p l5 p 2 be

of 37 apply,

by continuing the

x^a

may always
any
+ y^b +

ZiC

Vi,

x^a

y-J)

zz c

vz ,

PLANE GEOMETRICAL NETS.

ARTS. 42-45.]

we

shall

have two expressions of the forms,

tv pi
}

tVi

in

31

which the

coefficients tut

+ uv^pz

,_ fvipi
t Vi

Ul\

+ u
+

v-ipi

IIV*

are rational, because the co-ordinates

ii

are such, whatever the constants abc

be.

may

We

xy^

&c.,

have therefore (by 25) the

following rational expression for the anharmonic of this net-group

and similarly for every other group of the same kind. Hence every group of
four collinear net-points, and consequently also every pencil of four concurrent
net-lines, has a rational value for its anharmonic function ;

which value depends

only on the processes of linear construction employed, in arriving at that group

or pencil, and is quite independent of the configuration or arrangement of the

four

initial points

because the three

expression which results.

It

initial constants, a, b,

was thus

had the nine derived points A

</"

c,

disappear from the

that, in fig. 21, the nine pencils,

for their vertices,

were

all

which

harmonic

whatever manner the four points o, A, B, c might be arranged. In

general, it may be said that plane geometrical nets are all homographic
figures;* and conversely, in any two such plane figures, corresponding points
pencils, in

be considered as either coinciding, or at least (by 43) as indefinitely

approaching to coincidence, with similarly constructed points of two plane nets:
that is, with points of which (in their respective systems) the anharmonic co

may

ordinates (36) are equal integers.

45.

Without entering here on any general theory

of transformation of

anharmonic co-ordinates, we may already see that if we select any four netpoints GI, AI, BI, GI, of which no three are collinear, every other point p of the
same net is rationally related (42) to these; because (by 44) the three new
anharmonics of pencils,

(AI

SiO^p) =

(comp. 36) the new co-ordinates x^ y^

Si

are

&c.,

rational:

and therefore

of the point p, as well its old co-or

dinates xyz, are equal or proportional to whole numbers.

It follows (by 43)

that every point p of the net can be linearly constructed,

if

be given (no three being collinear, as above)

whole net can be reconstructed,^ if any one of

other words, that the

Compare the Geometric Superieure of M. Chasles,

or, in
its

any four such points

quadrilaterals (such as the

p. 362.

t This theorem (45) of the possible reconstruction of a plane net, from anyone of its quadrilaterals,
and the theorem (43) respecting the possibility of indefinitely approaching by net-lines to the points
above called irrational (42), without ever reaching such points by any processes of linear construction

ELEMENTS OP QUATERNIONS.

32
interstice in fig. 24)

[I.

As an example, we may

be known.

5, 6.

suppose that the

four points OA B C in fig. 21 are given, and that it is required to recover from
them the three points ABC, which had previously been among the data of the

For

construction.

this purpose, it is

three auxiliary points

A"

",

the three other auxiliary points

formulae, A = B

points A, B,

c,

*
B"

c",

",

only necessary to determine first the

OA B C &c. and next
*

as the intersections

A",

B",

c",

as B C

B"

",

&c.

after

which the

&c., will enable us to return, as required, to the

as intersections of

known

right lines.

SECTION

6.

On A nil arm 011 Ic Equations, and Vector

Expressions, for Curves

in a given Plane.

46.

When,

in the expressions

34 or 36 for a variable

three variable scalars (or anharmonic co-ordinates)

a?, ?/,

vector p

= OP, the

z are connected

by any

given algebraic equation, such as

fp (x,

y, *)

0,

th
supposed to be rational and integral, and homogeneous of the p degree, then
the locus of the tewn P (Art. 1) of that vector is a plane curve of the_p** order;

because (comp. 37)

is

it

cut in

imaginary), by any given right

For example,

if

we

p points (distinct or coincident, and real or

= 0, in the given plane.
line, Ix + my + nz

write
t

where
the

t,

sum

u, v are three
is zero,

then,

new

by

aa + u*bft + v*cy

"W+M +

tfc

variable scalars, of which

we

shall suppose that

eliminating these between the four equations,

x = P,

we

y =

u*,

z
,

+ u + v =

0,

are conducted to the following equation of the second degree,

=fp =
so that here

p=

2,

x*+y*

and the locus

+ z*- 2yz - Zzx

- %xy

of P is a conic section.

In

fact, it is

the conic

which touches the sides of the given triangle ABC, at the points above called A ,
=
for if we seek its intersections with the side BC, by making x
B c
^38),
,

of the land here considered, have been taken, as regards their substance (although investigated by a
treatise of MOBIUS, which was referred to in a
totally different analysis), from that highly original

former note

(p. 22).

Compare the remarks

in the following Chapter,

upon

nets in space.

ANHARMOMC EQUATIONS OF PLANE

AETS. 45-48.]

CUBYES.

33

which shows that

obtain a quadratic with equal roots, namely, (y - zf =
and similarly
there is contact with this side at the point (0, 1, 1), or A (36)

we

for the

two other

sides.

which the three right

47. If the point o, in

lines

AA

BB

cc concur, be

of that triangle are

(as in fig. 18, &c.) interior to the triangle ABC, the sides

then

all

cut internally,

by the points A B
,

c of contact with the conic

in this case (by 28) the ratios of the constants a,

are all positive,

b, c

so that

and the

denominator of the recent expression (46) for p can

not vanish, for any real values of the variable
scalars

t,

render

infinite

u, v

and consequently no such values can

The

that vector p.

generally in this case, as in

ellipse

conic is therefore

an

25,

fig.

which becomes however the

inscribed

inscribed circle,

when
or 1

b~

l
:

c~

- a

- b

denoting here the lengths of the sides of the

triangle, and s being their semi-sum.
a, b, c

48.

But

Fig. 25.

the point of concourse o be exterior to the triangle of tangents

if

ABC, so that two of its sides are cut externally, then two of the three ratios of

segments (28) are negative

be treated as
b

will

<

>

0,

and therefore one of the three

but each of the two others as

c

0,

>

0,

be a point on the side BC

AA

lines AC, AB,

bola,

circle)

a + b

0,

fig.

ellipse

>

if

0,

b, c

we suppose
>

but the points B

itself,

prolonged, as in

then the conic A B C will be an

the case of a

<

Thus

0.

>

constants a,

may
that

0,

o will be on the

26; and

(including

or a parabola, or an hyper

according as the roots of the quadratic,

(a

c) t*

+ 2ctu +

(b

c)

u2 =

0,

obtained by equating the denominator (46) of

the vector p to zero, are either, 1st, imaginary ;
or Ilnd, real

and

or Illrd, real

Fig. 26.

and unequal: that

is,

according as

we have
be

+ ca

or (because the product abc

a~ l

HAMILTON

is

b~

ab

-f

>

0,

or =

0,

or

<

here negative), according as

-f

c~

ELEMENTS OF QUATERNIONS.

l
<

0,

or = 0,

or

>

0.

ELEMENTS OP QUATERNIONS.

34

For example,

known

if

ratios of

the conic be what

[I.

often called the exscribed

is

n.

circle,

6.

the

segments give the proportion,

or 1

b~ l

c~

= -

- b

and

-s + s-c +
More

49.

if

generally,

be

c,

(as in fig. 26)

on that side prolonged, such that

cjc

AC = CB

writing then the condition (48) of

we

<

ing as cy c

which

AB

AB

AC = a +

an

ellipse, parabola, or

the arrangement being

still,

under the form,

hyperbola, accord

in other respects, that

Or, to express the same thing more symme

fig. 26.
we complete the parallelogram CABD, then according as the point

represented in

is

trically, if

>

and

c,

ellipticity (or circularity)

see that the conic is

or = or

<

a point upon the side AB, or

parallel to the chord B C , then

cc, is

AB = - a

s-b<0.

beyond the chord B C

falls, 1st,

with respect to the point A

or Ilnd, on that

chord; or Illrd, within the triangle AB C

figure being retained, the curve
other
lities,

is

the general arrangement of the same

,
In that
elliptic, or parabolic, or hyperbolic.

is

arrangement or configuration, which answers to the system of inequa

b>

0,

<?

>

+ b

0,

c<0,

the point A

on the line A A prolonged through A

a (b +

c)

-f

be

<

(b

is still

upon the

side BC

and

then the inequality,

be

shows that the conic

itself,

but

<

0,

whereof it is easily seen that

is necessarily an hyperbola
touched by the side BC at A while the other branch is touched
in B and c by the sides CA and BA prolonged through A.
The curve is also
hyperbolic, if either a + b or a + c be negative, while b and c are positive as
one branch

is

before.

50.

conic

When

is

the quadratic (48) has

an hyperbola, then the

its

substituting those roots, or the values of

any scalars proportional thereto),

p and similarly we can find the
;

case

when

roots real

t,

in the numerator of the expression (46) for

direction of the axis of the parabola, for the

the roots are real but equal

And

species of the conic,

equation,

<?&

f=

(46),

the

may

so that the

be found, by

u v which correspond to them (or

for

we

or direction, in which a right line OP must be

conic at infinity.

and unequal,

directions of the asymptotes

same

shall thus obtain the directions,

drawn from

o, so as to

meet the

conditions as before, for distinguishing the

be otherwise obtained by combining the anharmonic

of that conic, with the corresponding equation ax- + by

may

(38) of the line at infinity

so as to inquire (on

known

principles of

modern geometry) whether that

or touches

or cuts

it,

TANGENTS POLAKS.

DIFFERENTIALS

AKTS. 48-52.]

be considered as

meets that curve in two imaginary points,

line

in points which (although infinitely distant) are here to

it,

real.

51. In general,

if /(a?, y, z)

be the anharmonio equation (46) of any

plane curve, considered as the locus of a variable point p

tial* of this

35

and

the differen

if

equation be thus denoted,

y, z)

d/>,

Xdx

Fdy + Zdz

then because, by the supposed homogeneity (46) of the function/, we have the
relation

Xx
we

shall

have also

Xx
x

if

0,

but analogous relation,

this other

+ Ty + Zz =

Ty + Zz =

y - y
f

0,

= dx

dy

dz

= (a/, y , z ) be any point upon the

(by the principles of Art. 37), if P
=
and regarded as the limit
(a?, y, z),
tangent to the curve, drawn at the point p
at
p
this
of
a
secant.
The
tangent
may therefore be thus
symbol
(37)
of

that

is

written,

or

[X, Y, Z},

V,

[D,/,

D,/]

where D x %, D Z are known characteristics of partial derivation.

52. For example, when /has the form assigned in 46, as answering
&c. whence the
2
we have nx
conic
y
,

f=

lately considered,

at

any point

of this curve

(x, y, z)

[x-y

may

(x

z),

be denoted by the symbol,

y-z-x,

-z,

z-x-y]

in which, as usual, the co-ordinates of the line

proportional to them.

Thus

to the

tangent

at the point

may

be replaced by any others

or (by 36) at

(0,

1,1), which

is

evidently (by the form of /) a point upon the curve, the tangent is the line
that is (by 38), the side BC of the given triangle, as
[- 2, 0, 0], or [1, 0, 0]
;

was otherwise found before

(46).

And

in general

recent symbol denotes the right line, which

polar of the point

line \_X

Y\ Z

~\

is (in

easy to see that the

a well known sense) the

it is

with respect to the same given conic or that the

because the equation
the polar of the point (# y , z}

(x, y, z),
is

Xx

+ Yy + Zz =

0,

* In the
theory of quaternions, as distinguished from (although including) that of vectors, it will
he found necessary to introduce a new definition of differentials, on account of the non-commutative
property of quaternion-multiplication : but, for the present, the usual significations of the signs d and

D are

sufficient.

ELEMENTS OF QUATERNIONS.

36

which for a

conic

may

be written as

x+

condition requisite, in order that a point

Y y + Z z = 0,

(#,

s)

?/,

t/ 9

z) is

is

which

(by known

6/3

of the curve* should belong

satisfies

principles) the pole of the line

the conditions a~

therefore, for the present conic, the point

which the
aa +

expresses (by 51) the

at infinity (38), is the point

is

6.

that the centre of the conic, which

it

n.

which passes through the point (# i/ s ). Conversely, the point

so
(in the same well-known sense) the pole of the line [X, Y, Z~\

to a tangent
(#,

[I.

vector

+ cy =

OK

is

TC

by the help

easily reduced,

(b

X=

c,

b~

a, a

=
+

c~
b)

of

of the linear equation,

the form,

(27), to

(be

ca

+ ab)

with the verification that the denominator vanishes, by 48, when the conic is a
In the more general case, when this denominator is different from
parabola.

can be shown that every chord of the curve, which is drawn through
which point would
the extremity K of the vector K, is bisected at that point K

zero, it

way be seen again to be the centre.

53. Instead of the inscribed conic (46), which has been the subject of

therefore in this

recent articles,

we may,

as another example, consider that exscribed (or cir

cumscribed] conic, which passes through the three corners A, B, c of the given
The anharof fig. 21.
triangle, and touches there the lines AA", BB",
cc"

monic

equation of this

new

conic

is

easily seen to be,

+ zx + xy =

l/z

the vector of a variable point P of the curve

may

therefore be expressed as

follows,
t~

aa + u~ b3 +
l

v~ c
l

v~ c

with the condition

+ u + v =

0,

as before.

The

vector of

its

centre K

is

found to be
~
a* +

2 a*a +

V3

p+ &_

2bc

- 2ca - 2ab

a parabola, or an hyperbola, according as the denominator

And because these
of this last expression is negative, or null, or positive.

and

it is

an

two recent

ellipse,

vectors,

/c,

that the three points o, K,

bear a scalar ratio to each other,

it

follows (by 19)

are collinear; or in other words, that the line of

* If the curve
/= were of a degree higher than, the second, then the two equations above written
would represent what are called the jfirst polar, and the last or the line-polar, of the point (# y , z ),
with respect to the given curve.
,

VECTOR OF A CUBIC CURVE.

ARTS. 52-55.]
centres K,

37

of the two conies here considered, passes through the point of con

course o of the three lines

AA

BB

cc

More

L be the pole of any

if

generally,

with respect to the inscribed conic (46), and

[/, m, n\ (37),
if if be the pole of the same line A with respect to the exscribed conic of the
present article, it can be shown that the vectors OL, OL or A, X of these two
given right line

poles are of the forms,

A = k

where k and k are

one right

scalars

the three points

may

= P =
v

elimination of

t,

+ u +

u, v

suffice to

the three

ncy),

L are therefore ranged on

Pa +

u*b

an order higher

+
t

v*cy

v*c

3
,

3
y = u

3
,

we

find here

by

the anharmonic equation,

the locus of the point p

or briefly a cubic curve.
co-ordinates is so

be taken

Paa + w 3 i]3 +

Making x -

0, as before.

(x

may

o, L,

of a vector-expression for a curve of

than the second, the following

with

+ mbfi +

(laa

line.

As an example

54.

A = k

+ mbj3 + ncy),

(la a

much

is

+ y +

z)

- 27xyz =

therefore, in this example, a curve of the third order,

The mechanism

(41) of calculations with anharmonic

the same as that of the

known

trilinear method, that it

ABC are
briefly here that the sides of the given triangle

remark

(real) tangents

of inflexion ;

the points of inflexion being those

which are marked as
in
A",

21

fig.

and the

B",

c"

origin of vectors

o being a conjugate point* If

a = b = c, in which case (by 29)
this origin o becomes (as in
fig.
19) the mean point of the triangle,
the chord of inflexion
is then the line at infinity,

Fig>

A"B"C"

form represented in

fig.

27

having

27

and the curve takes the

three infinite branches, inscribed within

the angles vertically opposite to those of the given triangle ABC, of which the
sides are the three asymptotes.
55. It

would be improper

to enter here into

any

details of discussion of

such cubic curves, for which the reader will naturally turn to other works. f
*

is one of the imaginary cube-roots of

Answering to the values t = 1, u = 0, v = 3 where
which values of t, w, v give x = y = z, and p = 0.
t Especially the excellent Treatise on Higher Plane Curves, by the Rev. George Salmon, F.T.C.D.,
,

unity

&c.

Dublin, 1852.

ELEMENTS OF QUATERNIONS.

38

But

[I. n.

f>.

be remarked, in passing, that because the general cubic may be

represented, on the present plan, by combining the general expression of
Art. 34 or 36 for the vector p, with the scalar equation
it

may

= 27 to/s,

where

= x + y +

k denoting an arbitrary constant, which becomes equal to unity, when the

it follows that if p = (x, y, z) and
origin is (as in 54) a conjugate point
f
f
f
P = (xy y z] be any two points of the curve, and if we make s = x + y + z , we
;

shall

have the relation,

f

xyzs
in

which

it is

J=

not

PBP

%s

= x,y,z s 3.
,

difficult to

.
(A"

,*
3

B")

or

zs

ys

-.

^j

sx

sz

sy

prove that

PCP

(B"

C")

PAP

.
(c"

A")

We

obtain there
the notation (35) of anharmonics of pencils being retained.
the
sides
fore thus the following Theorem :
If
of any given plane* triangle
"

ABC

be cut (as in fig. 21) by

two points p and P in

(A"

any given

PBP

B")

(B"

PCP

then these two points p, p are on one

collinear points,

A",

c"

B",

for

AB of the

rectilinear transversal,

be such as to satisfy the

its plane

.
C")

(c"

common

PAP

=
A")

56.
degree

Whatever the

three tangents at those points

of the function

any point P =

order of a plane curve

f in

(x, y, z} is

[/,

we saw

46,

if

any

1,

three real points of inflexion,

its
its

and

cubic curve, which has the three

triangle for
seems to offer a new geometrical generation for curves of
BC, CA,

A"B"C",

anharmonic relation

may

and has

";

the sides

a result which

the third order.

be, or

whatever

may

be the

in 51 that the tangent to the curve at

the right line

m, n],

if

D^/,

D/,

n = D,/;

expressions which, by the supposed homogeneity of /, give the relation

Ix + my + nz = 0, and therefore enable us to establish the system of the two

following differential equations,

Zdx +

If then,

mdy + ndz =

by elimination

xdl +

0,

of the ratios of x, y,

z,

y&m
we

+ zdn =

0.

arrive at a new homogeneous

equation of the form,

= p
* This

Theorem may be extended, with

of the third order.

(D,/,

V,

scarcely

D a /),

any modification, from plane

to spherical curves,

ANHAEMOMC PEOPEETY

AETS. 55-57.]

as one that

(although

we

tion),

true for all values of #, y, z which render the

is

it

OF CUBIC CUEVES.

may

require to be cleared of factors, introduced

39

f unction

/=

this elimina

by

have the equation

shall

F
that must be

as a condition

m, n) =

(/,

assumed by that right

positions which can be

0,

satisfied by the tangent

to the curve, in all the

And, by comparing the two

line.

differential equations.

dp

we

see that

x y
:

= D?F

we may

n)

x&l +

0,

ydm

+ z&n =

and the symbol p =

D W F,

(D/F,

be, as above, the point of contact p of the variable line

positions, with the curve which

from the tangential equation F =
line

A,

locus

to the local

of a point p

/=

equation

0,

write the proportion,

Dm F

m,

(/,

m,

,]

in

if (x, y, z)

any one

of

its

Hence we can pass (or return)

a curve considered as the envelope of a right
of the same curve considered (as in 46) as the

is its envelope.

0, of

0,

we

since, if

[I,

D W F, DF),

obtain,

by elimination

of the ratios of

/,

m,

n,

an equation of the form

=/(DZF, D W F, D W F),
be necessary, of foreign factors) as a consequence of the
homogeneous equation F = 0, we have only to substitute for these partial deriva
tives, DjF, &c., the anharmonic co-ordinates x, y, z, to which they are propor
(cleared,

if

it

And when

tional.

the functions

them

/ and

F are not only homogeneous

(as

we

but also rational and integral (which it is some

times convenient not to assume them as being), then, while the degree of the
shall always suppose

to be),

function/) or of the local equation, marks (as before) the order of the curve,
the degree of the other homogeneous function F, or of the tangential equation
F =

0, is easily

seen to denote, in this anharmonic method

(as,

from the analogy

and older methods, it might have been expected to do), the class of
the curve to which that equation belongs or the number of tangents (distinct
of other

or coincident,

and

and imaginary), which can be drawn

real

from an arbitrary point in

As an example

57.

equations,
/ = x -

where

its

I,

Art. 46,

z,

plane.

(comp. 52),

to that curve,

x,

we

if

n =

eliminate

- x -

y,

Ix

x,

y,

my +

between the

nz = 0,

m, n are the co-ordinates of the tangent to the inscribed conic of

are conducted to the following
tangential equation of that conic,

we

or curve of the second class,

(I,

m,

n)

mn

+ nl + Im =

ELEMENTS OF QUATERNIONS.

40

with the verification that the sides [I,

among

the lines which satisfy this equation.

co-ordinates of contact x, y,

x = U/F =

as follows

z,

with the verification that the side

as
/,

an
m,

from

we should be brought back

/,

as before

/ = yz + zx + xy =

f=

differentiation the tangential co-ordinates*

and

= D xf = y +

0, of 46.

we can

z,

n = x +

= z + xy

from which we could

now

In

like

of the exscribed conic (53),

y,

by elimination tha tangential equation, namely,

F (/, m, n) = / 2 + m z + n z - 2mn - 2nl - 2lm =

so obtain

tangential

and then, by eliminating

to the local equation,

local equation

by

6.

manner, from the

derive

expressions for the

it

[1, 0, 0] touches the conic, considered

envelope, in the point (0, 1, 1), or

n,

if this

Conversely,

derive

y = n +

+ n,

&c. (38), of the triangle ABC are

0, 0],

equation were given we might (by 56)

|_I.

in turn deduce the local equation.

And

(comp. 40), the

very simple formula

Ix

my

+ nz =

0,

which we have so often had occasion to employ, as connecting two sets of

anharmonic co-ordinates, may not only be considered (as in 37) as the local

which a point p moves, but

equation of a given right line A, along

tangential equation of a given point, round ichich a right

as

we suppose

right line

A"B"C",

x + y +

tion

the set

/,

or [1,

m,

n, or

1, 1],

turns

also as the
:

according

Thus, while the

21, was represented in 38 by the equa

the set x, y,

of fig.

line

z,

to be given.

0, the point o of the same figure, or the point

(1, 1, 1),

may

be represented by the analogous equation,

I

because the co-ordinates

must

I,

+ n =

m, n of every

line,

which passes through

satisfy this equation of the first degree, as

may

this

point o,

be seen exemplified, in

the same Art. 38, by the lines OA, OB, oc.

58. To give an instance or two of the use of forms, which, although
homogeneous, are yet not rational and integral (56)

equation of the inscribed conic (46) as follows

#J

yh

%%

we may

write the local

=
;

* This name of
tangential co-ordinates" appears to have been first introduced hy Dr. Booth in a
Tract published in 1840, to which the author of the present Elements cannot now more particularly
See the reference in Salmon s
but the system of Dr. Booth was entirely different from his own.
refer
Higher Plane Curves, note to page 16.
"

LOCAL AND TANGENTIAL EQUATIONS.

AETS. 57-59.]

and then (suppressing the common numerical factor

41

J), the partial derivatives

are
I

or*,

n =

y~%,

form of the tangential equation for

so that a

/-i

-i

+ n

-i

z~$

this conic

= o

is,

which evidently, when cleared of fractions, agrees with the first form of the
with the verification (48) that a~ + b~ + c~ = when the curve is a
l

last Article

parabola

that

is,

when it is touched (50) by the line

we may write the local equation

exscribed conic (53),

ar 1

whence

it is

allowed to write
/

jr

z~

at infinity (38).

For the

thus,

also,

= #- 2

n =

/~

s~

2
,

aud

IM*

+ wi =

a form of the tangential equation which, when cleared of radicals, agrees again
with 57. And it is evident that we could return, with equal ease, from these

tangential to these local equations.

59.

For the

with a conjugate point

cubic curve

(54), the local

equation

may

be thus written,*
xl

we may

therefore assume for

/

and a form

its

+ yl +

=
;

tangential co-ordinates the expressions,

art,

si

n =

y~%,

of its tangential equation is thus

s~l

found to

be,

ft + m-b + n~b = 0.

Conversely,
equation,

if this

tangential form were given,

we might

return to the local

by making
x =

H,

y =

wf

?rl,

which would give a + y$ + si = 0, as before. The tangential equation just now

found becomes, when it is cleared of radicals,
=
or,

when

it is

J-*

also cleared of fractions,

= F =
*

HAMILTON

+ m~ z + n- 2 - 2

wV +

Compare Salmon

ri*P

Higher Plane Curves, page 172 [Art. 216, new

ELEMENTS OF QUATERNIONS.

ed.].

ELEMENTS OF QUATERNIONS.

42

which the biquadratic form shows (by 56) that

of

fourth
points

that
A"

roots

as indeed

class,
A",

B",

1)

( ~56),

known

The

to be.

- n =

a curve of the

inflexional character (54) of the

0,

recognised by the circumstance,

in order to find the four tangents from

the resulting biquadratic,

4/w 3 has
,

so that the line [1, 0, 0], or the side BC, counts as three,

fore a tangent of inflexion

the fourth tangent from

which touches the cubic at the point (-8, 1,

60. In general, the two equations (56),

nv x f -

may

this cubic is

6.

this curve is here

when we make
(0, *1,

it is

upon

c"

[I. n.

to z

f=

and

is

there

being the line [1, 4, 4],

1).

nVyf

0,

A"

three equal

mD z f =

0,

be considered as expressing that the homogeneous equation,

/ (nx,
which

is

ny,

obtained by eliminating

from f(x,

y,

z)

0,

Ix

my) =

0,

with the help of the relation

and which we may denote by

(#,

Ix

y]

my + nz = 0,
0,

has two

y, if /, m, n be still the co-ordinates of a tangent to the curve/;

which
obviously corresponds to the coincidence of two intersections
equality
of that line with that curve.
Conversely, if we seek by the usual methods

equal roots x

an

the condition of equality of two roots x

(a, y}

tf

=/

y of the homogeneous equation of the

(nx, ny,

Ix

my),

by eliminating the ratio x y between the two derived homogeneous equations,

= -D y we shall in general be conducted to a result of the dimension
= Dx
:

(f),

2p(p

1) in

<j>,

/,

m,

n,

and

of the form,

wPlP-O

(I,

m, n)

this elimina
by the rejection of the foreign factor w^- ), introduced by
=
in
be
will
general of
0, which
tion* we shall obtain the tangential equation F
such being generally the known class (56) of the curve of
the degree p(p - 1)
which the order (46) is denoted by p with (of course) a similar mode of

and

so,

a tangential to a local equation.

passing, reciprocally, from
when
the function /has the cubic form assigned in 54,
61. As an example,
the condition for the existence of two equal
we are thus led to
investigate

roots in the cubic equation,

=
*

(x,

y)

{(n

Compare the method employed

-l)x+(min

Salmon

I)

y}*

2
+ 27w

(Ix

+ my),

new ed.] to
Higher Plane Curves, page 98 [Art. 91,

the imaginary conic, z* + y* + z*

find the equation of the reciprocal of a given curve, with respect to
= 0, and /(*?) = are equa
F
then
r
he
deduced
function
ahove,
(xyz)
if
the
from/as
In general,
tions of two reciprocal curves,

LOCAL AND TANGENTIAL EQUATIONS.

ARTS. 59-61.]

by eliminating x

43

y between two derived and quadratic equations ; and the

result presents itself, in the first instance, as of the twelfth dimension in the

tangential co-ordinates

/,

m, n ; but

this division is effected, it is

imply that the curve

known

to be.

it is

found to be

divisible ~by

n 6 and
,

reduced to the sixth degree, thus appearing to

of the sixth class, as in fact the general cubic

is

when

A farther reduction is however

account of the conjugate point o

is

which introduces (comp. 57)

(54),

well

possible in the present case,

on
the

quadratic factor,
(I

and when
reduced

this

factor also

to the biquadratic

n)*

the tangential equation is found to be

form* already assigned in 59 the algebraic division,
is

set aside,

performed, corresponding to the known geometric depression of a cubic curve

with a double point, from the sixth to the fourth class. But it is time to close
this Section on Plane Curves; and to
proceed, as in the next Chapter we
last

propose to do, to the consideration and comparison of vectors of points in space.

* If we
multiply that form F =
biquadratic equation in I in, namely,

(59)

by

2
,

and then change nz

to

my, we obtain a

Ix

=^

(I,

m}

(l-

m?

(Ix

+ my) z + Mm

(I

m}

(Ix

+ my) z +

PmW

we then eliminate I
= n^, = Vmfy, we are conducted to
between the two derived cubics,
the following equation of the twelfth degree,
= % 3 y 3 z 3f(x, y, z), where /has the same cubic form as
in 54.
"We
are therefore thus brought back (comp. 59) from the tangential to the local equation of
z z A
the cubic curve (54)
complicated, however, as we see, with the factor x y z , which corresponds to
the system of the three real tangents of inflexion to that curve, each
tangent being taken three times.
and

if

The reason why we have not here been obliged

theory (60)

and not by

we might have

expected to be,

is

to reject also the foreign factor, z 12 as

that

by the general
2
multiplied the biquadratic function r only by z
,

we

z4 .

ELEMENTS OF QUATERNIONS.

44

CHAPTER

[I. in.

SI.

III.

APPLICATIONS OF VECTORS TO SPACE.

SECTION

1.

On Linear Equations between Vectors

not Complanar.

When three given and actual vectors OA, OB, oc, or a, /3, y, are not
contained in any common plane, and when the three scalars a, b, c do not all
vanish, then (by 21, 22) the expression act + bj3 + cy cannot become equal to
62.

; it must therefore represent some actual vector (1), which we may, for the
sake of symmetry, denote by the symbol - d$ where the new (actual) vector 8,
or OD, is not contained in any one of the three given and distinct planes, BOC,

zero

some

COA, AOB, unless

vanishes

one, at least, of the three given coefficients a,

and where the

shall thus

have a

neiv scalar, d, is either greater or less

linear equation between

aa +

which

will give

- aa

~d~
i

where OA, OB, oc

bj3

~d~~
aa

/
,

or

bfl

cy

~d~

into which the point

is projected,

three given lines OA, OB, oc,

b, c,

We

vectors,

d$ =

OD = OA + OB

oc

,
;

-cy

are the vectors of the three points A

-f

or

-6/3
,

four

+ cy

than zero.

d
,

on the

by planes drawn

parallel to the three given planes, BOC, &c.

so that they are the three co-initial edges of

a parallelepiped, whereof the sum, OD or

S, is

the internal and co-initial diagonal (comp 6j.

Or we may project D on the three planes, by
lines

given

OA

DA",

DB",

DC"

parallel

Fig. 28.

to the three

and then shall have

bfi + cy
OB + OC
&c., and S = OD = OA +
lines,

OA"

= OB +

OB"

= oc +

oc".

VECTORS OF POINTS IN SPACE.

ARTS. 62-64.]

And

45

evident that this construction will apply to any fifth point D of space,

it is

the four points OABC be still supposed to be given, and not complanar : but
that some at least of the three ratios of the four scalars a, b, c, d (which last
if

used as a mark of

letter is not here

differentiation] will

a =

0, if

D be situated in

planes through
63.
cients

S.

o.

We may

must

vary with the position

For example, we shall have

the plane BOC; and similarly for the two other given

of the point D, or with the value of its vector

inquire (comp. 23), what relation between these scalar

exist, in order that the point

D may be situated

coeffi

in the fourth given

what is the condition of complanarity of the four points, A, B, c, D.

Since the three vectors DA, DB, DC are now supposed to be complanar, they must

plane ABC; or

(by 22) be connected by a linear equation, of the form

a (a - S)

()3

S)

8)

comparing which with the recent and more general form

the required condition

a + b +

This equation

(62),

we

see that

is,

+ d =

0.

be written (comp. again 23) as

- b
- c
OB
OA
oc
_ +
= 1
+
+
or
+ __ - 1
d
d
d
OA
OB
oc

may
- a

and, under this last form,

known

expresses a

it

geometrical property of & plane

ABCD, referred to three co-ordinate axes OA, OB, oc, which are

common

origin o,

and terminate upon the plane.

We

of complanarity (comp. 28), the following proportion of

a
or,

- d = DEC

DCA

DAB

have

where

fig.

18

may

ABC

serve for illustration,

if

we

drawn from any

also, in this case

coefficients

and

areas

more symmetrically, with attention to signs of areas,

a b c d = BCD - CDA DAB - ABC
:

conceive o in that figure to be

replaced by D.
64.

When we

have thus at once the two equations,

aa +

bfi

+ cy + d$ =

so that the four co-initial vectors, a,

0,
|3,

and
j,

a + b +

+ d=0,

terminate (as above) on one

common

and may therefore be said (comp. 24) to be termino-complanar, it is

evident that the two right lines, DA and BC, which connect two pairs of the

plane,

four complanar points, must intersect each other in some point A of the plane,
at a finite or infinite distance.
And there is no difficulty in perceiving, on

ELEMENTS OF QUATERNIONS.

46

[I.

the plan of 31, that the vectors of the three points, A

which thus

result, are the following

,
A = BC

for

for B

for C

a =

~r

= AB

i.

c of intersection,

b3 + cy
aa + d$
=
b + c
a + d

DB,

DA,

= CA

)3

+ a

DC,

7 =

+ d

aa + b3

cy + d

a + b

+ d

expressions which are independent of the position of the arbitrary origin o, and
which accordingly coincide with the corresponding expressions in 27, when
we place that origin in the point D, or make S = 0. Indeed, these last results

when

hold good (comp. 31), even

the four vectors,

a, /3, y,

or the five points

are all complanar.

For, although there then exist two linear
four
those
between
which may in general be written thus,
vectors,
equations
o, A, B, c, D,

aa + b

+ cy + d S =

fi

0,

without the relations, a + &c. = 0,

we form from

a"

a"

a +

+ &c. =

i"/3

0,

c"y

+ eT8 =

between the

0,

coefficients,

yet

if

these another linear equation, of the form,

(a"

and determine

+ ta) a +

by the

(b"

tb

)3

(c"

tc

y +

+ td

(d"

0,

condition,
t

make

a"+b"+<f

= _

a =

+d"

+ ta, &c., and the two equations written at

the commencement of the present article will then both be satisfied and will
conduct to the expressions assigned above, for the three vectors of intersection

we

shall only

have

to

a"

which

vectors

thus be found, without

may

As an Example,

bfi

being necessary to employ those

of in the foregoing Chapter.

two given equations be (comp. 27,

let the

aa +

its

which were treated

processes of scalar elimination,

+ cy =

0,

(2a

c)

"

- aa =

33),
;

be required to determine the vectors of the intersections of the

and AB, CA
three pairs of lines BC, AA
Forming the combi
CA, BA

and

let it

"

"

".

nation,
(2a

and determining

by the

c)

a"

- aa +

(aa

(a

+ op +

Cy)

condition,

(2a + b +

c)

- a

-f

c)

0,

U,

YECTOES OF POINTS IN SPACE.

AETS. 64-66.]

which gives

= -

1,

we have
bfi

for the three sought vectors the expressions,

2aa +

cy + 2aa

+ cj
+

47

b(3

2a +

+ 2a

= a, by 27. Accordingly, in fig. 21, the line AA intersects

and although the two other points of intersection here
BC in the point A
which
considered,
belong to what has been called (in 34) a Third Construction,
"

whereof the

first

marked

are not

in that figure, yet their anharmonic symbols (36), namely,

otherwise found by combining the

(2, 1,0), might have been
=
=
and by combining s = 0,
lines
the
two
2z
for
x
and
CA, BA
equations y
=
x 2?/ for the remaining pair of lines.
(2, 0, 1)

and

"

In the more general case, when the four given points A, B, c, D, are not
plane, let E be any fifth given point of space, not situated on

65.
in

any common

any one

pyramid ABCD, nor on any such face

of the four faces of the given

= E. Then the four co-initial vectors, EA, EB,

prolonged and let its vector OE
do not
whereof
EC, ED,
(by supposition) no three are complanar, and which
some
terminate upon one plane, must be (by 62) connected by
equation of
;

the form
a

EA +

EB

-f

EC + d ED =
.

which we shall denote by - e,

are all different from zero.
Hence, because EA = o e, &c., we may establish
the following linear equation between five co-initial vectors, a, |3, y, S, c, whereof
where tliefour

scalars, a, b,

c,

and

d,

their sum,

no four are termino-complanar (64),

aa +
with the

relation,

whereof no one

now
e

66.

Under

a + b +

is, if

+ cy + d$ +

+d+

= (aa +

fy3

+ cy + dty

these conditions,

we denote by

Si

intersects the plane ABC,

= DE

if
*

ee

=0, between the

Hence

separately vanishes.

D!

that

bfi

we

(a

five scalars a,

b, c, d, c,

also,

d), &c.

write

ABC,

and

C-D!

Si,

the vector of the point DI in which the right line DE

we

shall

have

_ aa + ft/3 + cy
a + b + c
In fact, these two expressions are equivalent, or represent one common vector,
in virtue of the given equations
but the first shows (by 63) that this vector
;

ELEMENTS OP QUATERNIONS.

48
Si

[I.

terminates on the plane ABC, and the second shows (by 25) that

DE

nates on the line

intersection

of this

its

line

must therefore

extremity D!
with that plane.

1.

termi

be, as required, the

have

"We

it

m.

therefore the

two

equations,
.1.

(a

II.

Si)

b ()3

81)

c (y

SO =

^(S-W + ^e-SO-O;

whence (by 28 and 24) follow the two proportions,

I

IT.

= D^C

DXCA

= EDi

DiD

DIAB

the arrangement of the points, in the annexed fig. 29, answering to the case
where all the four coefficients a, b, c, d are positive (or

have one common

ing

sign] ,

and when therefore the remain

has the opposite sign).

coefficient e is negative (or

For the three complanar triangles, in the first

proportion, we may substitute any three pyramidal
67.

which

upon those triangles as their bases,

and which have one common vertex, such as D or E and

volumes,

rest

Fis>

29 *

because the collineation DEDI gives DDiBC - EDjBc = DEBC, &c.,

this other proportion,
.

I".

= DEBC

DEC A

we may

write

DEAB.

Again, the same collineation gives

EDi

we have

therefore,

by II

.,

DD = EABC
X

DABC

the proportion,

II".

= EABC

DABC.

DEBC + DECA + DEAB + EABC = DABC,

But
and

a + b +

d=-e;

we may

therefore establish the folio wing fuller formula of proportion, between

coefficients and volumes :
III.

a:b

:c

d: -

= DEBC

DECA

DEAB

EABC

DABC

the ratios of all these five pyramids to each other being considered as positive,
for the particular arrangement of the points which is
represented in the recent
figure.

The formula III. may however be regarded as perfectly general, if we

to
agree
say that a pyramidal volume changes sign, or rather that it changes its
68.

VECTORS OF POINTS IN SPACE.

49

algebraical character, as positive or negative, in comparison

with a ^V ew pyramid,

AETS. 68-69.]

and with a given arrangement

of points, in passing through zero (comp. 28)

any continuous change, any one of its

in the course of

namely when,

With

crosses the corresponding base.

the proportion III.

symmetric, hut equally general form :

III
a b c d e = BCDE
.

the

sum

shall have, generally,

DEBC = BCDE,

be

therefore

may

we

this convention*

DABC = - ADBC = ABDC = - ABCD,

CDEA

DECA = CDEA

in

expressed

DEAB

of these five pyramids heing always equal to zero,

following more

the

EABC

vertices

ABCD

when

signs (as above)

are attended to.

69.

We saw

the two equations,

(in 24) that

aa +

fy3

+ cy =

a + b +

0,

0,

gave the proportion of segments,

a

= BC

whatever might be the position of the origin

CA

AB,

In

o.

like

manner we saw

(in

63)

that the two other equations,

aa +

+ cy + d$ =

&/3

a + b +

0,

+ d=

0,

gave the proportion of areas,

a

d = BCD - CDA

DAB - ABC
:

And we

have just deduced (in 68) a

corresponding proportion of volumes from the two analogous equations (65),
a + b + c + d + e = 0,
+ d$ + ee = 0,
+
aa +

where again the origin

bfl

is

arbitrary.

cy

with an equally arbitrary origin. If then we conceive these segments, areas,

and volumes to be replaced by the scalars to which they are thus proportional,

we may

establish the three general formula;

OA

I.

OA

II.

BC + OB

BCD - OB

CA + OC
.

OA BCDE + OB

III.

where in

A, B, c are

I.,

in II.,

A, B, c,

and

A, B, c,

in III.,

AB =

CDA + OC DAB - OD
.

ABC =

CDEA + OC DEAB + OD
.

three collinear points

any

EABC + OE

ABCD =

D are any four complanar points ;

D, E are any five points of space ;

* Among the consequences of this convention respecting signs of volumes, which has already been
adopted by some modern geometers, and which indeed is necessary (comp. 28) for the establishment of
general formula, one is that any two pyramids, ABCD, A B C D bear to each other a positive or a nega
tive ratio, according as the two rotations, BCD and B C D , supposed to be seen respectively from the
,

points

A and A

HAMILTON

have similar or opposite

directions, as

ELEMENTS OF QUATERNIONS.

right-handed or left-handed.
JJ

ELEMENTS OF QUATERNIONS.

50

while o

[I. ni.

1,2.

of the three formulae, an entirely arbitrary point. It

must,
be
however,
remembered, that the additions and subtractions are supposed to be
performed according to the rules of vectors, as stated in the First Chapter of the
is,

in each

Book

present

the segments,

or areas,

which the equations

or volumes,

indicate, being treated as coefficients of those vectors.

We might

still

abridge the notations, while retaining the meaning of these formulae,

the

symbol of the arbitrary origin o

BC

-f

CA + C AB =
.

0,

SECTION

On Quinary Symbols
The equations

omitting

and by thus writing,*

collinear points with corresponding formulae

four
any
complanar points, and for any five points of space.

any three

for

by

further

and III

.,

for

2.

for Points

65 being

II".

and Planes in Space.

supposed to hold good, the vector

p of any point P of space may, in indefinitely many ways, be expressed (comp.
36) under the form
xaa + ybfi + zcy + ivd + vet
70.

of Art.

still

xa + yb + zc +

which the

in

ve

ratios of the differences of the five coefficients, xyzwv, determine the

In

position of the point.

common

wd +

fact,

because the four points ABCD are not in any

plane, there necessarily exists (comp. 65) a determined linear relation

between the four

drawn

vectors

them from the point

to

P,

which

may

be

written thus,

xa

PA + y

giving the expression,

TT
11.

which the

in

x = tx +

where
.

t
.

and
ee =

and a +

PC +

w d PD =
.

0,

xy

depend upon, and conversely

writing, then,

new and
.

x aa

y = ty +

v,

v are two
0,

z c

ratios of the four scalars

determine, the position of P

aa +

O =

PB +

v,

arbitrary
(65J,

we

tz

v,

scalars,

= twf +

v,

and remembering that

are conducted to the form for p,

assigned above.
*

We should thus have

different interpretations.

some of the notations of the Barycentric

Calculus, but

employed here with

QUINARY SYMBOLS FOR POINTS IN SPACE.

ARTS. 69-72.]

When

71.

the vector p

the Quinary Symbol (x, y,

thus expressed, the point p may be denoted by

w, v) and we may write the equation,

is

z,

?=(x,y,

But we

the same kind, (x

between

this

- v

f
:

//,

may
v

w, v).

z,

be denoted by this other symbol, of

also

provided that the following proportion

),

of coefficients (70) holds good

differences

x - v

Under

see that the same point

1

we

condition,

51

= x -

write the

therefore

shall

v.

following formula of

congruence,
f

(x

/,

(x, y, z,

w, v),

although not identical in composition,

have yet the same geometrical signification, or denote one common point. And
to express that these tico quinary symbols,

we

shall reserve the symbolic equation,

f

(x

f
,

to express that the five coefficients,

(x, y, z,

f
.

w,

v),

of the one symbol, are separately

equal to the corresponding coefficients of the other, x = x,

72.

Writing

v.

also, generally,
(tx, ty, tz, tw, tv)
f

(x

x,

v)

(x

and abridging the particular symbol*

may

briefly denote the

=
,

t (x,

y, z, w, v),

(x,

v),

(1, 1, 1, 1, 1) to

&c.,

(U), while (Q), (Q

quinary symbols

(x,

..?), (x

),

we may thus

establish the congruence (71),

(00 -(Q),
in which

and u

(0, 0, 0, 0, 1)

(Q)-t(Q )+u(U);

if

are arbitrary coefficients.

and

(1, 1, 1, 1, 0),

For example,
(0, 0, 0, 1, 1)

(1, 1, 1, 0, 0)

each symbol of the first pair denoting (65) the given point E; and each
symbol of the second pair denoting (66) the derived point DI. When the
coefficients are so simple as in these last expressions, we may occasionally omit
the

commas, and thus write,

still

more

(00001) - (11110)
*

briefly,

(00011) ^ (11100).

This quinary symbol (U} denotes no determined point, since

indeterminate vector p

= -

but

it

it

corresponds (by 70, 71) to the

admits of useful combinations with other quinary symbols, as above.

ELEMENTS OF QUATEBNIONS.

52

73. If three vectors, p, p

and

termino-collinear (24)

x a +

by w,
with a W// sum of

tion

tmp +

We

2.

expressed each under the first form (70), be

denote their denominators, xa +
xa +
.

they must then (23) be connected by a

coefficients, which may be written thus

limiar equa

m")

m.

p",

we

if

[I.

mp

tf

tm +

m"p"

*"w"

0.

have, therefore, the two equations of condition,

t

(xaa +
t

where

(xa

t,

scalars a

t"

e,

vet)

(x a +

some new

by

x +

+ v

et)

+ v

e)

t"

t"

(x"a

two equations

may

(65)

v"e)

t"x"

tv

v +

and they cannot be

scalar,

briefly be expressed

e,

and the

line.

five

but these equations

t"v"

= -u,

Hence

satisfied otherwise.

, p",

in

which the three vectors

),

(Q"),

by the equation,

so that if any four scalars,

equation, then,

v"et)

terminate, and of which the quinary symbols are (Q), (Q

p"

one right

supposing that

the condition of cottinearity of the three points p, p

p, p

(x"aa

while the five vectors a

scalars,

are subject only to the

tx
is

(xfaa
f

new

of condition are satisfied

where u

ve]

are three

t, t

can be found, which satisfy this

u,

t",

but not in any other

case, those three points

For example, the three points

D, E, DI,

PP

P"

last

symbolic

are ranged on

which are denoted

(72)

by the quinary symbols, (00010), (00001), (11100), are collinear ; because the
sum of these three symbols is (U). And if we have the equation,
=
(Q")

where

t,

right line

on the

line

74.
(Q")

(Q) + f

then

three scalars,

(Q

(Q"j

For example, the symbol

(0,

+ u (U),
on the
a symbol for a point
t
denote
0, 0, t,
any point
) may

is

p",

DE.

By

(Q"

u are any

PP

reasonings precisely similar it may be proved, that if (Q) (Q )

he quinary symbols for any four points ppW" in any common

plane, so that the four vectors

pp

p"p"

are termino-complanar (64), then

an

equation, of the form

t

must hold good

(Q) +

(Q

(Q")

and conversely, that

+
if

"

(Q

")

--u(U),

the fourth symbol can be expressed

as follows,

=
(QT")

(Q) +

(Q ) +

t"

(Q")

+ u (U),

QUIKAHY SYMBOLS OF PLACES.

ABTS. 73-75.]

with any scalar values of

plane PP

P"

t,

c,

then the fourth point p

u,

t",

For example, the four

(01000),

(10000),
or A, B,

of the other three.

D! (66), are complanar

is

situated in the

points,

(11100),

(00100),

and the symbol

53
"

(t,

t"

0, 0)

may

repre

sent any point in the plane ABC.

When

thus complanar with three given points, P O PI, P 2 we

v
have therefore expressions of the following forms, for the five coefficients x,
of its quinary symbol, in terms of the fifteen given coefficients of their symbols,
75.

a point P

is

and

And

of four

hence,

new and

arbitrary scalars

to

to

by

elimination of these four scalars,

linear equation of the

I

which
posed

may
locus

v =

x =

(x

tiXi

+ u

tiV l

u,

we

t 2 vz

u.

are conducted to a

form
v)

m (y -

v)

+ n

(z

v)

+ r

(tv

v)

0,

be called the Quinary Equation of the Plane P

or of the sup
the
p
because it expresses a common property of all the
of
point
PiP2>

points of that locus

and because the

three ratios of the four

new

coefficients

m, n, r, determine the position of the plane in space. It is, however,

metrical, to write the quinary equation of a plane II as follows,
Ix

where the fifth

my +

I,

more sym

nz + rw + sv = 0,

connected with the others by the relation,

coefficient, s, is

+ n + r +

and then we may say that [I, m, n, r, s]

the Plane n, and may write the equation,
II

=
[

I,

m,

is

(comp. 37) the Quinary Symbol of

n, r, s].

For example, the coefficients of the symbol for a point P in the plane ABC
be thus expressed (cornp. 74)

may

x =

u,

y =

ti

u,

t2

+ u,

w=

u,

= u

between which the only relation, independent of the four arbitrary scalars t
u, is
w-v=
this therefore is the equation of the plane ABC, and the symbol of that
.

!,-!]; which may (comp. 72) be sometimes written more

without commas, as [000 1 1 J. It is evident that, in any such symbol,
the coefficients may all be multiplied by any common factor.

plane

is [0, 0, 0,

briefly,

ELEMENTS OF QUATERNIONS.

54

[I.

m.

2, 3.

The symbol

76.

of the plane p PiP 2 having been thus determined, we

may
next propose to find a symbol for the point, P, in which that plane is intersected
by a given line P 3 P 4 or to determine the coefficients x
v, or at least the ratios
:

(a, y, s,

Combining, for

of their differences (70), in the quinary

= P = p PiP 2

v)

u>,

symbol

of that point,

P 3P 4

this purpose, the expressions,

(which are included in the symbolical equation (73),

and express the

PP 3 P 4 ), with the equations (75),

collinearity
Ix

sv

0,

(which express the complanarity pp PiP 2 ),

(&"3

which determines the

For example,

if

if it
t

be

t^

= u

we

(lx

77.

a point

0,

sr4 )

and contains the solution

line

= 4 + u

ic

of the problem.

DE, then (comp. 73),

v =

a point in the plane ABC, then

also

+ u

(75),

and therefore

hence

*3

(00011) + u (11111),

(Q) - (00011)

or

symbol had accordingly been found

= ABC DE.

(72) to represent the intersection

When
P,

the five coefficients, xyzwv, of any given quinary symbol (Q) for
or those of any congruent symbol (71), are any whole numbers (posi

tive or negative, or zero),

we

related to the five given points,

the System,

which those

five coefficients,

shall say (comp. 42) that the point P is rationally

five points

or briefly, that

determine.

Imnrs, of the quinary

equal or proportional to integers,

of the

are conducted to the formula,

last

(66), D!

On

+ h

sv 3 )

P be a point on the

(Q)

which

ratio

x = y =

but
t3 -

same System; or that

the contrary,

when

thus already whole

we

it is

symbol

And

a Rational Point of

manner, when the

(75) of a plane II are either

shall say that the plane is a Rational Plane

rationally related to the

same

five points.

the quinary symbol of a point, or of a plane, has not

coefficients,

and cannot be transformed (comp. 72)

have them, we shall say that the point or plane

given points

it is

in like

or briefly, that

it is irrational.

is

so as to

irrationally related to the

right line

which connects two

ANHAKMONIC CO-ORDINATES IN

ARTS. 76-79.]

two rational

rational points, or is the intersection of

SPACE.

55

be called, on

may

planes,

the same plan, a Rational Line ; and lines which cannot in either of these
two ways be constructed, may be said by contrast to be Irrational Lines. It
is

evident from the nature of the eliminations employed (comp. again 42), that

& plane, which

rational plane

common

is

determined as containing three rational points,

and

in like

manner, that a point, which

intersection of three rational planes, is

which

also every point

is

is

necessarily a

determined as the

always a rational point

obtained by the intersection of a rational

rational plane ; or of two rational lines with each other

complanar)

is

line

as

is

with a

(when they happen to be

when

78. Finally,

two points, or two planes, differ only

by the arrangement

(or order) of the coefficients in their quinary symbols, those points or planes

may

be said to have one common type

or briefly to be syntypical.

For ex

E, are thus syntypical, as being represented

the
and the ten planes, obtained by
by
quinary symbols (10000),
(00001)
the
combinations
of
all
those
five
ternary
taking
points, have in like manner
one common type. Thus, the quinary symbol of the plane ABC has been seen

ample, the five given points, A,

[00011] and the analogous symbol [11000] represents the plane

Other examples will present themselves, in a shortly subsequent

(75) to be

CDE, &c.
Section,

on the subject of Nets

in Space.

But

seems proper to say here a

it

few words, respecting those An harmonic Co-ordinates, Equations, Symbols, and

Types, for Space, which are obtained from the theory and expressions of the
present Section,
coefficients,

by reducing

(as

we

are allowed

in each symbol or equation,

from five

SECTION

On Anharmonic
79.

When we

that the fifth

number of the

3.

Co-ordinates in Space.

adopt the second form (70) for

coefficient

to do) the

to four.

in the first

form

p, or

vanishes,

suppose

we get

(as

we may)

this other general

expression (comp. 34, 36), for the vector of a point in space:

ma

yb(3 + zcy

xa + yb + zc +

and

may

wd

then write the symbolic equation (comp. 36, 71),

p =

and

+ wd

call this last the

(a?,

y, z, w),

Quaternary Symbol of the Point p

although we shall

ELEMENTS OF QUATERNIONS.

56

[I.

m.

3.

soon see cause for calling it also the Anharmonic Symbol of that point. Mean
while we may remark, that the only congruent symbols (71), of this last form,
are those which differ merely

by the introduction

three ratios of the four coefficients,

mine

the position of the point

of a

common factor

the

w, being all required, in order to deter

whereof those four

coefficients

may accordingly
be said (comp. 36) to be the Anharmonic Co-ordinates in Space.
80. When we thus suppose that v = 0, in the quinary symbol of the point
we may

suppress the fifth term sv, in the quinary equation of a plane n,

p,

Ix

the fifth

sv

(75)

coefficient, s,

and therefore may suppress

also (as here unnecessary)

in the quinary symbol of that plane,

which

is

thus reduced

to the quaternary form,

II

This
of

last

may

m,

[I,

n, r].

also be said (37, 79), to be the

which the Anharmonic Equation

Ix

Anharmonic Symbol of the Plane,

is

my +

nz

+ rw =

the four coefficients, Imnr, which we shall call also (comp. again 37) the An
harmonic Co-ordinates of that Plane II, being not connected among themselves
+ s = 0)
since their three ratios (comp.
by any general relation (such as / -f
the position of the plane
in
to
determine
order
all
in
are
79)
general necessary,
.

in space.

81. If

we suppose

that the fourth

symbol of a point, that point P

is

in,

coefficient,

the plane

w, also vanishes, in the recent

ABC; and

may

then be sufficiently
And if we attend

represented (as in 36) by the Ternary Symbol (x, y, z).

only to the points in which an arbitrary plane n intersects the given plane ABC,

we may
In

this

and

manner, then,

we

as being for such points unnecessary.

are reconducted to the equation, Ix +

my +

nz = 0,

A= [/, m, n], for a right line (37) in the plane ABC, considered
on that plane, of an arbitrary plane n in space. If this plane

to the symbol,

here as the
II

its fourth coefficient, r,

suppress

trace,

be given by

its trace

82.

ABC,

if

A,

its

quinary symbol (75),

by simply

we thus

obtain the ternary symbol for

suppressing the two last coefficients, r

In the more general

we denote (comp.

case,

72)

lished formulae of collineation

its

when the

point p

is

quaternary symbol

and complanarity

and

s.

not confined to the plane

by

(Q), the lately estab

(73, 74) will still

provided that we now suppress the symbol (U), or suppose

zero.
Thus, the formula,

hold good :
be

its coefficient to

ANHARMONIC CO-ORDINATES IN

ARTS. 79-83.]

expresses that the point P

vanish, the equation

signifies that

is

in the plane

is

p p

SPACE.

and

V";

if

57

the coefficient

"

which then remains, namely,

thus complanar with the two given points P ,

; or, in other words, that it is on the right

arbitrary third point

and with an

p",

line

P"

whence

(comp. 76) problems of intersections of lines with planes can easily be resolved.

In

like

manner,

for a p/#H0

if

we denote

by

briefly

[JK] the

r
expresses that the plane
II

IT,

IT";

quaternary symbol

[I,

w,

n, r]

the formula

II,

and

if

^"

jr

passes through the intersection of the

II

m=

we suppose

An?e planes

0, so that

the formula thus found denotes that the plane II passes through the point of
with any third plane ; or (comp. 41),
intersection of the two planes, FT,
n",

that this plane

II

contains the line of intersection of

three planes, II, IT,

n",

may

be said to be

II";

in

Hence

collinear.

which case the

it

appears that

either of the two expressions,

I.

may

..if (Q

II.

t"

(Q"),

be used as a Symbol of a Right Line

that line

in

f
.

[#] +

Space

t"

[JB"~|,

according as

we

consider

either, 1st, as connecting two given, points, or Ilnd, as being the

The remarks

intersection of two given planes.

points, planes,

and

lines

(77)

on rational and irrational

require no modification here

and those on

types (78J

adapt themselves as easily to quaternary as to quinary symbols.

83. From the foregoing general formulae of collineation and complanarity,
it follows that the point P
in which the line AB intersects the plane CDP
,

through CD and any proposed point p = (xyzw) of space,

p = AB

CDP =

(a?yOO)

may

be denoted thus

for example, E = (1111), and c = AB CDE = (1100).

In general, if ABCDEF
be any six points of space, the four collinear planes (82), ABC, ABD, ABE, ABF,
are said to form a pencil through AB
and if this be cut by any rectilinear
;

transversal, in four points,

C\D\ E

of this group of points (25)

Planes

which

may

is

be thus denoted,
(AB

HAMILTON

F\ then (comp. 35) the anharmonic function

called also the Anharmonic of the Pencil of
,

ELEMENTS OF QUATERNIONS,

CDEF)

ELEMENTS OF QUATERNIONS.

58

Hence (comp. again 25, 35), by what has

we may establish the important formula

[I. in.

shown respecting

just been

3.

and P

:
/Y*

(CD

AEBP) = (AC BP

= j

symbol (xyzw) for a variable point P (79),

represents the anharmonic of a pencil of planes, of which the variable plane CDP
In like manner,
is one ; the three other planes of this pencil being given.
so that this ratio of coefficients, in the

(AD

BECP) =

and

-,

(BD

CEAP) =

anharmonics

so that (comp. 36) the product of these three last

we have

the same plan

(BO

/y?

<5*

is

On

unity.

also,

AEDP) = -,

(CA

BEDP) = -,

(AB

CEDP) =

so that the three ratios, of the three first coefficients xyz to the fourth coefficient
10,

ABP, whereof the point P

suffice to determine the three planes, BCP, CAP,

common

intersection,

by means

is

the

of the anharmonics of three pencils of planes, to

And thus we

which the three planes respectively belong.

see a motive (besides

that of analogy to expressions already used for points in a given plane), for
a point in
calling the four coefficients, xyziv, in the quaternary symbol (79) for
space, the

84.

Anharmonic Co-ordinates of

In general,

that Point.

there be any four collinear points, P O

if

P3

so

that

(comp. 82) their symbols are connected by two linear equations, such as the
following,
(ft)

(ft)

then the anharmonic of their group

= t

(ft)

(ft),

(ft) +

(ft),

expressed (comp. 25, 44) as follows

may be

tit

fowd-jjpi
by considering the pencil (CD POP^PS), and the transversal AB (83).
in like manner, if we have (comp. again 82) the two other symbolic

as appears

And

equations, connecting four collinear planes

[12,]

[ft] + u [ft],

the anharmonic of their pencil (83)

[ft]
is

n
f

3,
f

[ft] + u [ft],

expressed by the precisely similar

formula,
(IIoILnjI,)
as

may

Jine AB,

uf
, ;

be proved by supposing the pencil to be cut by the same transversal

ANHABMONIC CO-OKDINATES IN

ARTS. 83-86.]

85. It follows that

if

linear functions of x, y,

Do

11.3

(82),

and

/ (xyzw)
z,

u",

and

we determine

if

59

any two homogeneous and

(xyzw) be

/i

SPACE.

four

collinear

planes

by the four equations,

/=0,

/:=/,

/i

fi-kf,

0,

have the following value of the anharmonic

any
determined
thus
of
the
of
function,
planes
pencil

where k

scalar

is

we

shall

derive this Theorem, which

Hence we

present system of co-ordinates to space

is

important in the application of the

The Quotient of any two given homogeneous and linear Functions, of the
anharmonic Co-ordinates (79) of a variable Point p in space, may be expressed as
"

the

Anharmonic (noI^EUIIa) of a Pencil of Planes

whereof three are

given,

while the fourth passes through the variable point p, and through a given right
line A which is common to the three former planes"
86.

And

Theorem

in

like

proved this other but

manner may be

analogous

The Quotient of any two given homogeneous and linear Functions, of the
anharmonic Co-ordinates (80) of a variable Plane II, may be expressed as the
Anharmonic (PQP^PS) of a Group of Points ; whereof three are given and
"

collinear

right line

and the fourth

is

the intersection,

A, with the variable plane

II,

of their

common and

given

II."

two given functions of Imnr be F and FI, and if we

determine three points P PiP 2 by the equations (comp. 57) F = 0, FI = F, F! = 0,
and denote by P 3 the intersection of their common line A with II, we shall

More

fully, if the

have the quotient,

For example,

if

we suppose

that

B 2 = (0101),

A 2 = (1001),
A 2 = (1001),
A 2 = DA

SO that

we

BCE, &C.,

find that the three ratios of

I,

= (0101),
and (DA 2 AA

m, n to

r,

= (0011),
= (OOlT),
2

c2
c
2)

= -

1, &C.,

in the symbol

II

\_lmnr~],

may

expressed (comp. 39) under the form of anharmonics of groups, as follows

/

where

Q, R, s

= (DA , Z AQ)

,
= (DB
2 Bn)
,

-=

denote the intersections of the plane

II

be

(DC 2 cs)

with the three given

I 2

ELEMENTS OF QUATERNIONS.

60

[I.

m.

3, 4.

And thus we have a motive (comp. 83) besides that

of analogy to lines in a given plane (37), for
calling (as above) the four coeffi

right lines, DA, DB, DC.

cients

I,

n, r, in the

m,

quaternary symbol (80) for a plane

the Anharmonic

II,

Co-ordinates of that Plane in Space.

87. It

may

same plane
notations
so

A",

B",

marked before

shall

if

(in 31, &c.), so that

= (OlIO),

B"

(CA"BL)

six given points,

A",

=
;

B",

c",

which the equation and symbol are

x + y +
sea?

= (1010),

write (comp. 36)

c"

= (1100),

~=

The

we may now

have (comp. 39, 83) these three other auharmonics of groups, with

their product equal to unity

and the

L,

c"

A"

we

we denote by

M, N the points in which the

cut by the three given lines BC, CA, AB, and retain the
for those other points on the same three lines which were

be added, that

II is

-=
(AB"CM)

Z,

2,

2,

(BC"AN)

are all in one given plane [E], of

+ w = Q

[E]

= [1111].

groups of points, of which the anharmonic functions thus represent

the six ratios of the four anharmonic co-ordinates, Imnr, of a variable plane
are therefore situated on the six edges of the given pyramid, ABCD

II,

two points in

each group being corners of that pyramid, and the two others being the inter
sections of the edge with the two planes, [E] and II.
Finally, the plane [E] is
centre
(in a known modern sense) the plane of homolvgy* and the point E is the
of homology, of the given pyramid ABCD, and of an inscribed pyramid AiBiCiDi,
where AI = EA BCD, &o. ; so that D! retains its recent signification (66, 76),

and we may write the anharmonic symbols,

AX =(0111),

Bl

d=(1101),

=(1011),

And

D^lllO).

if we denote
by A iBWiD the harmonic conjugates to these
with
points,
respect to the lines EA, EB, EC, ED, so that

we have

last

the corresponding symbols,

A\ = (2111),

B J = (1211),

G\ = (1121),

v\ = (1112).

other relations of position exist, between these various points, lines,

and planes, of which some will come naturally to be noticed, in that theory

Many

of nets in space to

which in the following Section we

* See
Poncelet

shall proceed.

Traite des Propriety Projectives (Paris, 1822).

GEOMETRICAL NETS IN SPACE.

ARTS. 86-89.]

SECTION

4.

On Geometrical Nets
When we

88.

others

have

(as in 65) five

we can

are complanar,

61

in Space.

E, whereof no four
given points A
them by a right line, and the three
.

two of

connect

any
by a plane, and determine the point in which these

another

last intersect

deriving thus a system of ten lines AI, ten planes Hi,

and

one

ten points

p l9

from the given system of five points P O by what

may be called (comp. 34) a

next propose to determine all the new and
which connect the ten derived points PJ with
,

We may

First Construction.
distinct lines,

2,

and planes,

2,

the five given points P O and with each other

,

new and

P 2 arise

distinct points

and may then inquire what

as intersections of lines with

(at this stage)

planes, or of lines in one plane with each other

all such

new

lines, planes,

then

we might proceed

for ever

To

to a Third Construction of the

same kind, and

so

on

building up thus what has been called* a Geometrical Net in Space.

express this geometrical process

planes,

and

And

points being said (comp. again 34) to belong to a Second Construction,,

and

and by quinary

lines,

by quinary symbols

(71, 75, 82) of points,

types (78), so far at least as to the

end

of the

second construction, will be found to be an useful exercise in the application of

and therefore ultimately in that METHOD OF

the subject of the present Book.
And the quinary form
here be more convenient than the quaternary, because it will exhibit more

principles lately established

VECTORS, which
will

clearly the geometrical

given points,

is

and

will

dependence of the derived points and planes on the five

thereby enable us, through a principle of symmetry, to

reduce the number of distinct

types.
89. Of the five given
PO
points,

be (10000)

while

corresponding type

of the

may

these two represent

AB

c AiBidAaiJaCz

ten derived points

be taken as (00011)

ADE = (01100),

=DA

any symbol or equation

By

quinary type has

PI,

been seen

provided that

A!

(78) to

of first construction, the

in fact, considered as symbols,

the

A2

A, B, c to B, c, A,

the

The nine other points

points A and DI.
and we have now (comp. 83, 87, 86) the symbols,

A = BC

also, in

= EA

PI

are

BCD = (10001),

BCE = (10010);

of the present form,

we

at the

it is

permitted to change

same time write the

Mobius, in p. 291 of his already cited Bary centric Calculus.

third, first,

ELEMENTS OF QUATEKNIONS.

62

L1

in

and second co-efficients, in the places of the first, second, and third thus,
B = CA* BDE= (10100), &c. The symbol f^/yOOO) represents an arbitrary point
on the line AB and the symbol [00;?rs], with n + r+ s = 0, represents an
:

arbitrary plane through that line

AB

as a symbol also of the line

each therefore

and

itself,

may

be regarded (comp. 82)

as a type of the ten

same time

at the

while the symbol [00011], of the plane ABC (75),

as a type of the ten planes IIx.
Finally, the five pyramids,
lines

Ai

may

and the

ten triangles,

may

pyramids,

ABDE,

CADE,

BCDE,

ABCE,

such as ABC, whereof each

be called pyramids

R^

is

be taken (78)

ABCD,

a common face of two such

and Mangles

of the First

TI,

Con

struction.

90. Proceeding to a Second Construction (88),

we soon

find that the lines

A may be arranged in two distinct groups ; one group consisting of fifteen lines
such as the line* AA DI, whereof each connects two points p and passes
A
2

2 ,i,

l5

also through one point P O , being the intersection of two planes III

ADE

point, as here of ABC,

such as B C
PO,

while the other group consists of thirty

each connecting two

of the thirty edges oifive new pyramids

points PI, but not passing

and being one

/

C B

A 2 Ai,

which pyramids

B A C-A,

A c B 2 Bi,

lines

2,2,

through any point

z,

A 2 B 2 C 2D!,

namely,

AiBiCiDi

R may be said (comp. 87) to be inscribed homologues of the

R the centres of homology for these five pairs of pyramids
z

former pyramids

five

through that

l9

being the five given points

whereof the

last

A E and the planes of homology being five planes

has been already mentioned (87), but which belong
.

[A]
[E],
The planes II 2 of second construction,
properly to a third construction (88).
form in like manner two groups ; one consisting of fifteen planes n 2 ,i, such
as the plane of the five points, AB B 2 c 1 c 3 whereof each passes through one point
.

PO,

and through four points

PI,

and contains two

ABiCa, ACiB-j, besides containing four lines

group

is

composed

2,2,

2,2,

A^C!, namely, the twenty

three points Pi, and
contains
each
whereof

pyramids

R^

but does not pass through any point P O

It is

to express these geometrical conceptions^ of \h forty-five lines

planes II 2

and the

* ABic 2 AB 2 ci, DA
,

AI,

lines

other

of twenty planes fI 2 , 2 , such as

faces of the five recent

three lines

A ,i, as here the

B^, &c. while the

lines

as here

five planes

EA

of homology of pyramids,

AO, are other lines of this

2 ;

now

required

the thirty-five

[A] ... [E],

by

group.

out the existence and

Calculus, p. 284, &c.) has very clearly pointed
chief properties of the foregoing lines and planes ; but besides that his analysis is altogether different
from ours, he does not appear to have aimed at enumerating, or even at classifying, all the points of
what has been above called (88) the second construction, as we propose shortly to do.

f Mobius

(in his

Bary centric

ARTS. 89-92.]

GEOMETRICAL NETS IN SPACE.

quinary symbols and

types, before

63

proceeding to determine the points p 8 of

second construction.

An

AA DI (90) may be represented by

and an arbitrary plane through that line by this other
symbol, Owwwrr], where m and r are written (to save commas) instead of - m
and - r; hence these two symbols may also (comp. 82) denote the line AA D,
91.

arbitrary point on the right line

the symbol (tuuQO)

and may be used

itself,

The

2 ,i-

[01111], of the last form, represents that particular

symbol

particular

as types (78) to represent the group of lines

plane through the last-mentioned line, which contains also the line AB^ of
the same group and may serve as a type for the group of planes n 2jl
The
line B C and the group A 2 2 may be represented by (stuQQ) and
if
we
[tttm],
/ /
s = t + u, and s = - s
while
the
to
write
B
c
A
and
the
2
agree*
plane
group
n 2)2 may be denoted by [11112], Finally, the plane [E] has for its symbol
;

[11114] and the four other planes [A], &c., of homology of pyramids (90),
have this last for their common type.
;

The

92.
lines

points P 2 of second construction (88), are

,

A and

above,

it is

not

more numerous than the

yet with the help of types, as

to classify and to enumerate them.
It will be suffi

difficult

down

these types, which are found to be eight, and to offer

remarks respecting them in doing which we shall avail ourselves of the

cient here to write

some

that construction

II 2 of

planes

eight following typical points,

which are
A"

A"

all

whereof the two

situated in the plane of

= (01100)

= (02100)

"

= (21100)
=

(12100)

ABC

first

have already occurred, and

A IV

=
1

(21100)
(32100)

A V = (02100)

(23100)

the second and third of these having (10011) and (30011) for
congruent symbols
(71). It is easy to see that these eight types represent,
respectively, ten, thirty,

thirty, twenty, twenty, sixty, sixty,

eight groups,

which we

the points P 2

is

290.

the end of what

fhefive given points P O

p2

belonging to

so that the total

number

of

(88) to close the present inquiry, at

defined to be the Second Construction, the total

P
which
are thus derived by lines aud planes from
2
p,,
is found to be
exactly three hundred : while the joint
,

of the net-lines, Ai,

be one hundred, so
*_With

distinct points,

we have above

number o/the net points,

number

mark as P 2)1
then we consent

shall

If

and sixty

2,

arid of the net-planes,

l9

II 2 ,

has been seen to

far.

this convention, the

line AB,

[QQtus], their point-symbol being (teOOO).

and the group AI,

may

be denoted by ike plane- symbol

ELEMENTS OE QUATERNIONS.

64

To

(1.)

4.

(_I.m.

the type P 2 ,i belong the ten points^

A 2B

2,

with the quinary symbols,

A"=

(01100),

= (lOOlO),

A\=

(10001),

= (00011),

T>\

which are the harmonic conjugates of the ten points p ls namely, of

A 2 B 2 C2
ABC
,

with respect to the ten lines Ai, on which those points are situated; so that
we have ten harmonic equations, (BA CA") = - 1, &c., as already seen (31, 86,
Each point P 2)1 is the common intersection of a line A with three lines
87).
x

2,2

thus

valent,

by

we may

establish the four fo]\owiug formula? of concurrence (equi

89, to ten such formulae)

A"

= BC B

= EA DiA 2 B

II 2 ,i

is

CI

C BI

= DA DiAi

six

planes

I1 3 , 2

IL

for example,

c2

fi

c B2

= DE AiA 2 BiB 2

also situated in three planes

and in

B 2 C2

Each point P 2)1

group

B^

CiC 2

in three other planes, of the

a point

is

A"

common

to the

twelve planes,

ABC, BCD, BCE

B C

Each
Each

B 2 C2 A 2

BiCiAi,

AI,

line, Ai, or

plane, Hi

A2

2,

or I1 2

DB^BiC

ABiCoCiBo,

B c A2

contains one point P 25l

,

contains

2,

three

but no line

such points

B 2 C 2 Di.

BjCiDi,

EB B 2 C C2

Cj,

A 2 ,i

contains any.

and each plane

contains two, which are the intersections of opposite sides of a quadrilateral

that plane, whereof the diagonals intersect in a point P O

Il 2 ,i

in

for example, the

diagonals BjC 2 B 2 Ci of the quadrilateral BiBaCad, which is (by 90) in one of the
planes n 2 ,i, intersect* each other in the point A while the opposite sides CiB!,
B 2 c 2 intersect in
and the two other opposite sides, BjB,, C 2 C! have the point
,

A"

on

The

for their intersection.

ten lines of third construction

A VxC

,,

of ten pairs of triangles 7\,

which the
line

centres of

Vc",

whereof the

A"B 2

2,

ten points P 2)1 are also ranged, three by three,

A
C

3,

2,

namely, on the axes of homology,

.

AWi,

which are situated

homology are the ten points

in fig. 21, is the axis of

latter is

homology

W,

in the ten planes IL,

PJ

of the

and

of

for example, the dotted

two

triangles, ABC, A B C

inscribed in the former, with the point o in that figure

The same
D! in fig. 29), to represent their centre of homology.
ten points p 2)1 are also ranged six by six, and the ten last lines A 3 are ranged
(replaced

by

Compare the

first

Note to page 62.

GEOMETRICAL NETS IN SPACE.

ABT. 92.]

four by four, in

five planes

3,

namely

in the planes of

pyramids, Hi, jR 2 already mentioned (90)

(87) the six points

A
which

WVaB^c

VA

2,

and the four right

a,

2,

homo logy

of fire pairs of

for example, the plane [E] contains

65

VB
2

lines,
A"B"C"

2,

latter are the intersections of HLQ four faces,

DAG,

DCB,

DBA,

pyramid ABCD, with the corresponding

of the

DA

DABi,

of its inscribed homologue AiBidDi

planes,

A2 B

in different planes,

We

of homology.

may

C1 ,

faces,

DjBiA!,

AiBiCi,

and are contained,

B2c A

C^

A 2 B 2 c2

the three triangles, ABC, AiBid,

although

ABC,

for

and having the

besides, in the four other

A2 B 2 C 2

instance, being all homologous,

A // B

line

also say, that this line

//

for their
is

A"B"C"

common

axis

the common trace

A^CI and A B c on the plane ABC and in

is the common trace, on that
like manner, that the point
plane Hi, of two
c
and
B
also
the
lines A
of
common
trace
of the two lines
B:Ci
being
namely
(81) of two planes II 2

2,

of

namely

2,

A"

2, 2

B^C

and B

On

(2.)

2,

which belong

to the third construction.

the whole, these ten points, of second construction,

considered to be already well

known

to geometers, in

A"

.,

may

be

connexion with the

theory of transversal* lines and planes in space but it is important here to

observe, with what simplicity and clearness their geometrical relations are
:

example, the

employed. For
(82) of the four planes, ABC, A^CI, A 2 B 2 c 2 and [E],

by the quinary symbols and quinary

expressed (88),

collinearity

becomes evident from mere

[00011],

which represent

types

inspection of their four symbols,

[H12l],

[H112],

[11114],

(75) the four quinary equations,

w-v = Q, x+y+z-2w-v

0,

x+y+

- w - 2v =

0,

o?

+ y+

s+0-4?=0;

with this additional consequence, that the ternary symbol (81) of the common
so that this trace is (by 38)
trace, of the three latter on the former, is [111]
And if we briefly denote the quinary
of fig. 21, as above.
the line
:

A"B"C"

symbols of the four planes, taken in the same form and order as above, by
* The
marked

and harmonic relations between the ten points, which we have above
and which have been considered by Mb bius also, in connexion with his theory of
appear to have been first noticed by Carnot, in a Memoir upon transversals.

collinear, complanar,

as p 2)i

nets in space,

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

66

[ft] [-RJ [ft],

[-So]

we

= -

[ft]

see that they are connected

[A] + [A]

[ft]

by the two

2[ft] + [ft]

if

we denote

(ILn.n .EQ

- 2

is its

mean point

n we
3,

have

a result which can be very simply verified, for the case

pyramid, and E (comp. 29)

relations,

the planes themselves by n,, II 2 n 2

(comp. 84) the following value for the anharmonic of their pencil,

whence

4.

[Lin.

when ABCD

the plane

II 3 ,

a regular

is

or [E], becoming

in this case (comp. 38) the plane at infinity, while the three other planes, ABC,
AjBid, A 2 B 2 c 2 , are parallel; the second being intermediate between the other

two, but twice as near to the third as to the first.

We

(3.)

must be a

little

more

concise in our remarks

on the

seven other

types of points P 2 which indeed, if not so well known,* are perhaps also, on
the whole, not quite so interesting although it seems that some circumstances
of their arrangement in space may deserve to be noted here, especially as
,

affording an additional exercise (88), in the present system of symbols and

The type p z 2 represents, then, a group of thirty points, of which
types.
A"

in

fig.

line

"

2,2

no other

line,

among

those which have been hitherto considered.

to describe here all the lines, planes,

aiming

called the third construction,

as well as

we may already

and that the planes

the pyramids R 2 and the

numerous

to be

an example; each being the intersection of a line A 2)1 with a

is the point in which AA intersects B C
as A
but each belonging to

21, is

3,

and

and the

triangles

points,

see that they

lines

3,

of

But without

what we have

must be expected

of that construction,

of the second construction,

z,

above noticed, can only be regarded as specimens, which in a closer study of

it becomes necessary to mark more fully, on the
present plan, as
n 3 ,i,
K, i. Accordingly it is found that not only is each point P 2 2 one of

the subject,
.

the corners of a triangle

3fl

of third construction (as

"

is

of

"B

"C"

in

the sides of which new triangle are lines A 3 2 passing each through
fig. 21),
one point P 2 ,i and through two points P 2 2 (like the dotted line A"B
of
such
also
each
P
but
is
the
two
intersection of
new lines of
point 2)2
fig. 21)
,

"

"C

* It does not appear that any of these other types, or groups, of points P 2 , have hitherto heen
noticed, in connexion with the net in space, except the one which we have ranked as the fifth, p 2 5 ,
and which represents two points on each line Ai, as the type P 2 ,i has heen seen to represent one point
on each of those ten lines of first construction hut \h&t fifth group, which may he exemplified hy the
,

DK with the two planes AiB[Ci and A 2 B 2 c 2 has heen indicated by Mobius (in
already cited work), although with a different notation, and as the result of a different

intersections of the line

page 290 of his

analysis.

GEOMETRICAL NETS IN SPACE.

ART. 92.]

67

whereof each connects a point P O with a point P 2)1

For example, the point A is the common trace (on the plane ABC) of the two

third construction,

3,3,

"

new

lines,

DA

I,

EA

because,

if

we adopt

for this point

"

the second of

its

two congruent symbols, we have (comp. 73, 82) the expressions,

We may
article)

therefore establish the formula of concurrence (comp. the

first

sub-

"

= AA

DA

B C

EA

which represents a system of thirty such formulae.

that the point A
may be represented, not only
(4.) It has been remarked
also
the
but
the
by
congruent symbol, (10011)
quinary symbol (21100),
by
"

if

then we write,
AO

= (11100),

BO

= (11100),

= (11100),

points A O B O C O in the plane of ABC, must be considered to be

syntypical, in the quinary sense (78), with the three points A^W", or to
belong to the same group P 2 , 2 although they have (comp. 88) a different
these three

new

ternary type.

again

fig.

to see that, while the triangle

It is easy

21) an inscribed homologue

3 ,i

of the triangle

AB

"

"B

is

"C

which

(comp.
is itself

it namely of
(comp. sub-article 1) an inscribed homologue T^\ of a triangle
new
the
common
axis
of
their
with
for
triangle A O B O CO is
homology,
ABC,
A"B"C"

on the contrary an exscribed homologue T3 2 with the same axis A 3}1 of the
But from the syntypical relation existing as above for
1\.
,

same given triangle

between the points A and A O we may expect to find that these two
points P 2)2 admit of being similarly constructed, when the five points P O are
"

space

treated as entering symmetrically (or similarly), as geometrical elements, into

The point A must

namely, on AA but also on a

the constructions.

A 2)i

line

AiA 8

line

and on two

which

therefore be situated, not only on a

lines

latter lines are

3, 3,

soon,

2,2 ,

which

is

= AA

AiA 2

and may consider the three points AO B O

,

BiB 2 , CiC2

while the three

to be
;

cc".

We

establish the formula of concurrence (comp. the last sub- article)

AO

found

easily

each connecting a point P O with a point P 2)1

seen to be BB" and
may therefore

new

lines

BB"

cc"

c as the traces of the three lines AiA a ,

AA", BB", cc",

which coincide

in position

with the sides of the exscribed triangle A O B O C O are the traces A 3 3 of three
3 ,i, such as ABtC 2 B 2 Ci, which
planes
pass through the three given points A,B,C,
K 2
,

ELEMENTS OF QUATERNIONS.

68

A 2?1 whereon

but do not contain the lines

are situated.

Every

other plane

[I.

m.

4.

the six points P 2)2 in their plane HI

contains, in like manner, six points P 2 of

III

the present group every plane n 2 ,i contains eight of them and every plane
n 2;2 contains three; each line A 2 ,i passing through two such points, but each
;

A 2)2

only through one. But besides being (as above) the intersection of
two lines A 2 each point of this group p 2 2 is common to two planes Hi, four
planes II 2 ,i, and two planes II 2 2 while each of these thirt}^ points is also a
line

common corner

of two different triangles of third construction, of the lately

mentioned kinds T^\ and T3)2 situated respectively in the two planes of first
construction which contain the point itself. It may be added that each of the
,

two points P 2)2 on a

,

line

the harmonic conjugate of one of the two points

is

2 ,i,

with respect to the point P O and to the other point

here the two harmonic equations,

PI,

PI

on that

line

thus

we have

by which the
(5.)

positions of the

two points A

and A might be determined.

"

third group, p 2)3 , of second construction, consists (like the preceding

group) of thirty points, ranged two by two on the fifteen lines A 2 ,i, and six by
six on the ten planes II l? but so that each is common to two sucli planes; eacli
two planes

also situated in

is

in two planes

Il 2 ,i,

fI 2)2 ,

and on one

line

A 3 ,i,

which (by sub-art. 1) these two last planes intersect each other, and two of
the five planes n 3 ,i each plane IJ 2 ,i contains four such points, and each plane
fl 2 2 contains three of them
but no point of this group is on any line AI,

in

or

2,2.

The

p2

six points

,3

which are

in the

plane ABC, are represented (like

the corresponding points of the last group) by two ternary types, namely by
(211) and (311) ; and may be exemplified by the two following points, of

which these

last are the

ternary symbols

AA
A!

The
A!

IV

= AA

iA 2 Ai

= AA

IV
sub-group A

but the three points

which
of the second sub-group are the corners of a new triangle, T

three points of the

IV
.

first

are colliuear

3>3

to the triangle ABC,

is

and

to all the other triangles in its plane

IV

C I IV , being the traces of the three planes

homologous
which have been hitherto considered, as well as to the two triangles A^CJ and
A 8 B 2 Ca the line of the three former points being their common axis of homology
;

and the

sides of the

new

of pyramids, [A], [B], [c]

(comp:.90) of
line

triangle,

IV

homology
is the common
A IV B IV C IV or
A"B"C"

as (comp. sub-art. 2) the

trace of the two other planes of the

same

GEOMETRICAL NETS IN SPACE.

AET. 92.]

group

3 ,i,

namely

of [D]

truce of the line A\A.\

planes

2, 2

and

We may also say that

[E].

and because the

in which that point

69

is

the point A t IV

is

the

B C O C B O are the traces of the two

lines

we may

contained,

write the formula of

concurrence,
AI

(6.)

3 ,i,

and

is

= AA

IV

A^A

C BO

CO

be also remarked, that each of the two points p z 3 on any line

the harmonic conjugate of a point p a 2 with respect to the point P O
It

may

two points

to one of the

on that

line

also

the harmonic con

being
other point
jugate of this last point, with respect to the same point P O and the
p 2 2 thus, on the line AA DI, we have the four harmonic equations, which are
not however all independent, since two of them can be deduced from the two
PI

others, with the help of the

(AA

And

"A

two analogous equations of the fourth sub-article:

A IV ) = (AA A O A IV ) = (AA^A^) =

the three pairs of derived points

P!,

p2

2,

P2

(AD^ A^)
3,

= -

on any such

1.

line

will

2 ,i,

be found (comp. 26) to compose an involution, with the given point P O on the
the other double point of this
line for one of its two double points (or foci)
:

involution being a point P 3 of third construction ; namely, the point in which

the line A 2 ,i meets that one of the five planes of homology n 3 ,i, which corre

Thus, in the present

sponds (comp. 90) to the particular point P O as centre.
X
example, if we denote by A the point in which the line AA meets the plane

which (by 81, 91) the trace on ABC is the line [411], and therefore is
has
been stated) the side B^CI of the lately mentioned triangle jT3 3 so
(as

[A], of

that

A X = (122) = AA

we

shall

BC"

CB

"

have the three harmonic equations,

=
)

(AA

X
"A

A O ) = (A

X
point A is the common harmonic conjugate of the
A IV A! IV and
O
given point A, with respect to the three pairs of points, A UJ, A

which express that

this

new

"A

therefore that these three pairs

AX

for its

form

(as

has been said) an involution, with A and

two double points.

(7.) It will be found that

we have now exhausted

the types of points

of second construction, which are situated
upon lines A 2jl ; there being
only four such points on each such line. But there are still to be considered

two new groups of points P 2 on

all

A and three others on lines A 3

lines, we may observe that each of the two

lines

1?

Attending first to the former set of

new types, p 2 4 p 2)5 represents twenty points, situated two by two on the ten
,

ELEMENTS OF QUATERNIONS.

70

A and therefore
common to three such
n
and each point p

but not on any line

lines of first construction,

the ten planes

111,

[I.

six

each point however being

4.

by six in

planes: also

each point P 2 4 is common to three planes

2 2
2 5 is situated
in one such plane
while each of these last planes contains three points P 2)4 ,
but only one point p z 5
If we attend only to points in the plane ABC, we
,

can represent these two new groups by the two ternary types
(021) and (021),
which as symbols denote the two typical points,

Av = BC

we have

c AiA 2

DiAiBi

D]A 2 B 2

and

B!

may
IV

that A VI

also the concurrence,

A V = BC
It

A VI = BC C/B^a = BC C B O

be noted that A

which
is

is

last point is

AO

AB

D^"

".

the harmonic conjugate of c

on the same trace

C AO, of the

with respect to A O
plane c A A 2 and
and B O on the
x

V
harmonically conjugate to B! , with respect to c

where B/ denotes (by an analogy which

trace of the plane c BiB 2 ,

become more evident) the intersection of that

we have the two equations,
(A

cW) =

(B oBl

VA

will soon

trace with the line CA

VI
)

= -

so that

1.

AI contains thus two points P 2 of each of the two last new

groups, besides the point P 2 ,i, the point PI, and the two points P O which had
been previously considered it contains therefore eight points in all, if we still
(8.)

Each

line

And

abstain (88) from proceeding beyond the Second Construction.

to prove that these eight points can, in two distinct modes, be so
involution, with two of them
attend only to points on the
we
Thus,
them by ternary symbols, we may write,

form (comp. sub- art

sent

an

6)

points thereof.

c = (001),

^=

A- = (021),

(021),

A =
A,*

(Oil),

A"

two double

line BC,

and repre

= (Oil)

A^ =

(012),

easy

arranged as to

for the

if

B = (010),

it is

(Ol2)

and the resulting harmonic equations

I.

II.

will then suffice to

(BA
(A

CA")

= (BA V CA VI ) = (BA/CA^) = -

=
BA"C)

show

1st,

(A

W)

(AV A V)

1,

--

1,

that the two points P O on any

,

double points of an involution, in which the points PI, P 2 ,i

conjugates, while the tico other pairs are of the

Ilnd, that the two points PI and p 2jl on any such

,

line Ai,

are the

one pair of

form
common form, P 2)4 P 2 ,s
,

and

line Ai, are the double points

of

a second involution, obtained by pairing the two points of each of the three other

GEOMETRICAL NETS IN SPACE.

ART. 92.]

71

Also each of the two points P O on a line Ai, is the harmonic conjugate
of one of the two points P 2 5 on that line, with respect to the other point of
groups.

the same group, and to the point PI on the same line

VI VI
=
(BA A! A )

(cAV^r

= -

thus,

1.

remains to consider briefly three other groups of points P 2 each

group containing sixty points, which are situated, two by two, on the thirty
lines A 2 2 and six by six in the ten planes Hi.
Confining our attention to
It

(9.)

plane ABC, and denoting them by their ternary

on
the line B C , the three new typical points, of the
thus,
three remaining groups, P 2)6 p 7 P 2 8

those which are in the

symbols, we have

2,

A"

A- =(321);

=(121);

A= =(231);

be combined these three others, of the same three types, and

on the same line B C
with which

may

= (112)

AX

= (312)

A!

Considered as intersections of a line

or with planes

(in

which

A IX

-n^r/
5 (u

AA

A vn

TJ

CQ.BI

(CO A

&c.,

V
",

II 2 , 2 (or its trace

&o., of the

contains six points p 2

6,

same plane

in the

u
B ri
C

n,,

its
,

trace

CO ) in

tweke points p 2

7,

r*

r<

UiU-j

A IX

A/",

&c.

The

line

on the plane ABC), in the

BB"

(or their

AA

1,

common

trace

DIB")

in

A vm

and similarly for the other

Each plane 2 contains twelve
:

same three groups.

and eight points P 2)8
7

points P 2)6 , eight points P 2

(2l3).

for the three other points,

points, AI

be thus denoted

vi-o vii
BA iv-rj
BI BI

C/A IX ) = -

B C thus intersects one plane II 2 ,i (or

vn it intersects two
point A
planes II 2

and one other plane

may

-n

v-r>vi

l/i

with the harmonic equation,

and with analogous expressions

iv/i

with lines

2,2

IX

character alone they belong to the second

latter

construction), the three points

A,

,i

while every plane II 2 2

and nine points p 2)8 Each point p 2)6
;

contained in one plane IIi in three planes n 2 ,i and in tioo

planes n 3 2
Each point P 2 7 is in one plane IIi, in two planes II 2 ,i, and in four planes II 2 2 .
And each point p 2)8 is situated in one plane Hi, in two planes II 2 ,i, and in
is

three planes

2, 2 .

The

(10.)
points of the three last groups are situated only on lines A 2 2
but, on each such line, two points of each of those three groups are situated
,

ELEMENTS OF QUATERNIONS.

72

[I.

m.

4.

which, along with one point of each of the tioo former groups, P 2)1 and P 2
and with the two points p l5 whereby the line itself is determined, make up a
system of ten points upon that line. For example, the line B C contains,
,2>

besides the six points mentioned in the last sub-article, the four others:

B = (101)

Of these

A"

(Oil)

"

(211).

mentioned, namely the points P 2)1 and P 2 2

are the double points (comp. sub-art. 8) of a new involution,

ten points, the two last

upon the line A 2)2

in which the two points
,

pair, as is

And

= (110)

of each of the four other groups

compose a conjugate

expressed by the harmonic equations,

the analogous equations,

(B

A"C

=
")

(B

WA)

(B

A^ C

=-

111
)

1,

on any line A 2 2 are the double points of another

involution (comp. again sub-art. 8), whereof the two points P 2 b P 2 a on that
line form one conjugate pair, while each of the two points p 2 6 is paired

show that the two points

PI

with one of the points P 2)7 as its conjugate. In fact, the eight-rayed pencil
/
/
/
//
A VIII A VII A VIII A VI1 ) coincides in position with the pencil (A BCA
(A c B A

A A

VI

Ai Ai

VI
),

and may be

said to be & pencil in double involution; the third

A"

and

fourth, the fifth and sixth, and the seventh and eighth rays forming one invo
lution, whereof the first and second are the two double* rays ; while the first

and seventh, and the sixth and eighth rays compose

another involution, whereof the double rays are the third and fourth of the
and second, the

fifth

pencil.
(11.)
selves,

If

we proceeded

to connect systematically the points P 2

and with the points PI and P O we should

,

find

among them

many remarkable

lines

and

which have been incidentally

noticed above for example, we should have a group n 3)2 of twenty new planes,
exemplified by the two following,

planes of third construction (88), besides those

;

[Ej = [11103],

[D

= [11130],

which have the same common trace A 3 ,i, namely the line
ABC, as the two planes AiBid, A 2 B 2 c and the two planes [D],

A"B"C",

2,

2,2

and

of these

3 ,i,

on the plane

[E], of the groups

which have been considered in former sub-articles

new planes

3, 2

would be found

* Compare page 172

to contain one point

of the Geom. Superieure of

M.

Chasles.

p,,,

and each

three points

GEOMETRICAL NETS IN SPACE.

ARTS. 92-93.]

p 2 ,i,

six points

twenty new

P2

2,

and

three points p 2

might be proved also that these

R 3 which are the

It

3.

73

planes are the twenty faces of five new pyramids

cxscribed homologues of the five old

pyramids R

with the

five

given points
But it would lead us heyond
p for the corresponding centres of homology.
the proposed limits, to pursue this discussion further although a few additional
:

(89),

remarks
tion

in

may

be useful, as serving to establish the completeness of the enumera

above given, of the lines, planes, and points of second construction.

93. In general, if there be any n given points, whereof no four are situated

any common plane, the number Noi the

obtained from them, as intersections

which are immediately

derived points,

A n

of line with plane (each line being

drawn through two of the given points, and each plane through
or the number of points of the form AB CDE, is easily seen to be,
AT~ f(n \ - n(n-l)(n-2)(n-3)

/W=

so that

(n

three others),

4)

2.2.3~

N = 10, as before, when n =

But

5.

if

we were

to

apply this formula

to the case n = 15, we should find, for that case, the value,

JV = /(15) = 15

and thus

fifteen given

14

13

11 = 30030

and independent points of space would conduct, by what

be called a First Construction (comp. 88), to a system

Yet it has been lately stated (92), that
of more than thirty thousand points.
from the fifteen points above called P O PI, there can be derived, in this way,

might

(relatively to them)

only two hundred and ninety points P 2 as intersections of the form* A II

and therefore fewer than three hundred. That this reduction of the number of
.

derived points, at the

end of what has been called

for the net in space, arising

points

PO,

would be found

from the dependence

to be so considerable,

(88) the Second Construction

of the ten points PI

on the five

might not perhaps have been

and although the foregoing examination proves that all the eight
do
types (92)
really represent points P it may appear possible, at this stage, that
some other type of such points has been omitted,
study of the manner in which
the types of points result, from those of the lines and planes of which they are
anticipated

2,

the intersections,
that

way

would indeed decide

this question

that the eight types, or groups, p 2jl ,

P2

and
,

8,

it

was, in

fact, in

of points of second

construction for space, were investigated, and found to be sufficient

yet

it

* T he
A of complanar lines, when
definition (88) of the points P 2 admits, indeed, Intersections A
they are not already points PO or PI but all such intersections are also points of the form A n so
that no generality is lost, by confining ourselves to this last form, as in the present discussion we
propose to do.
.

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

74

may be

[I. in.

4.

useful (compare the last sub-art.) to verify, as below, the completeness

of the foregoing enumeration.

admit of 105 binary, and of 455 ternary combi

but these are far from determining so many distinct lines &ud planes.

The fifteen points P O

(1.)

nations

In

fact, those

25

lines Ai,

PI,

15 points are connected by 25

2,

which

combinations of points

collineations,

therefore count as 75,

lines

among

the 105 binary

and there remain only 30 combinations

by the

represented

of this sort,

A 2 2 Again, there are 25 ternary

combinations of points, which are represented (as above) by lines, and therefore
do not determine any plane. Also, in each of the ten planes HI, there are 29

which are constructed by the 30 other lines,

6) triangles Ti, T2 because each of those planes contains 7 points P O P I}

connected by 6 relations of collinearity. In like manner, each of the fifteen
2 ,i contains 8 (= 10
2) other triangles Tz because it contains 5 points
planes

(=

35 -

There remain therefore only 20

PO, PI, connected by two collineations.
25
290
120) ternary combinations of points to be accounted for
(= 455
The completeness of the
and these are represented by the 20 planes
2)2
;

enumeration of the

and

and planes of the second construction

lines

is

therefore

verified
only remains to verify that the 305 points, P O p b p 2 above
considered, represent all the intersections A II, of the 55 lines Ai, A 2 with the
;

it

45 planes Hi,
(2.)

2 ,i,

Each plane HI

2,2

contains three lines of each of the three groups, Ai,

contains two lines A 2 ,i, and four lines A 2)2 and

H^

each plane

each plane
2 2 contains three lines

contained in three planes

planes

2 ,i

two planes

and each

II 2

2 ),

tion, there are

and plane

it

240

H!

line

Hence

2,2.

each line

2 ,a

2 ,i

in

(or because each line Ai

is

two planes HI, and in two

H^\ and in

in one plane HI, in two planes

follows that, without going beyond the second construc

(=

30 + 30 + 30 + 30 + 60 + 60)

so that the

number

cases

of cases of intersection

is

of coincidence of line

reduced, hereby,

55 45 = 2475, to 2235 (= 2475 - 240).

(3.) Each point P O represents twelve intersections of the form AI

from

because

it is

common

to four lines AI,

and

H!

to six planes HI, each plane contain

ing two of those four lines, but being intersected

point PO

as the plane ABC, for example,

AD and AE.

Again, each point P O

which contains any

is

by the two others in that

intersected in A by the two lines,

is

common

of the four lines AI

to three planes II 2 ,i,

through that point

therefore a system of twelve other intersections, of the form AI

each point P O

is

common

to three lines

2 ,i,

each of which

is

it

2)1

no one of
represents

Again,

contained in two

of the six planes II i, but intersects the four others in that point P O

which

GEOMETRICAL NETS IN SPACE.

ART. 93.]

therefore counts as twelve intersections, of the form

of the points P O represents three intersections,

2)1

2 ,i

IIi.

2 ,i

Finally, each

and

it

represents

form A

II, within the limits of the present

of
the
each
Thus,
five given points is to be considered as represent
inquiry.
ing, or constructing, thirty-nine (= 12 + 12 + 12 + 3) intersections of line

no other intersection, of the

with plane

and there remain only 2040

intersection

derived points, P I? P 2
(4.)

For

2235 - 195) other cases of such

by the 300

(=

II, to be accounted/or (in the present verification)

.

this purpose, the nine columns,

headed as

I. to

IX.

in the follow

ing Table, contain the numbers of such intersections which belong respectively
to the nine forms,

Ar

HI,

AI

2 ,i,

AI -112,2;

A
A

2 ,i

2 )2

for each of the nine typical derived points,

.

P2

,8

Column X.

states

lllj

A 2)2

2 ,i

ii 2 ,i,

A ,i*n

11 2 )1|

2 ,o

2>2

2 ,2j

A IX of the nine groups

,

contains, for each point, the

thus tabulated in the

preceding columns

number

iii,

sum

P!,

p 2)1

of the nine numbers,

and expresses therefore the

entire

which any one such point represents. Column XI.

the number of the points for each type ; and column XII. contains the
of intersections,

product of the two last numbers, or the

number

of intersections

II

which

are represented

(or constructed) by the group.

Finally, the sum of the
numbers in each of the two last columns is written at its foot and because
;

the 300 derived points, of

first

represent the 2040 intersections

is

and second constructions, are thus found to

which were to be accounted for, the verification.

seen to be complete: and no new

type, of points P 2 remains
,

(5.)

TABLE OF INTERSECTIONS A

to

be discovered.

II.

ELEMENTS OF QUATERNIONS.

76
It

(6.)

IIi,
2

to be

remembered

we have

that

not admitted,

any points which can only be determined by

(88),

is

2,

2 ,i

considered in recent sub-articles, as

of the third.

we

counted, as

lines

lines

struction, although they

the lines

definition

which can only be found as intersections of

any
For example, we do not regard the traces AA", &c., of certain

two such planes.

by our

intersections of three planes

as belonging to the second construction: nor have

of that construction,

planes

4.

[I.m.

among

the lines of second con

would present themselves early in an enumeration of

And any point in the plane ABC, which can only be

the present stage) as the intersection of two such traces, is not

student might find it however to be not useless, as
regarded as a point P 2
an exercise, to investigate the expressions for such intersections and for that

determined

(at

reason

it

noted here, that the

may be

traces of planes IL,

(comp. 81) of the forty-four

on the plane ABC which are found to compose a system

2,

ternartj types

of only twenty -two distinct lines in that plane, whereof nine are lines AI,

the seven following (comp. 38)

tioo],

[Til],

[oil],

[in],

[oil],

which, as ternary symbols, represent the seven

AA

BC,

B C

2,

are

A"B"C",

[211],

[ail]

lines,

DIA",

AA",

CO

Again, on the same principle, and with reference to the same defini
that new point, say F, which may be denoted by either of the two con

(7.)

tion,

gruent quinary symbols (71),

F = (43210) - (01234),
arid which, as a quinary type (78), represents a

space (and of no more,

type, with all

its Jive

on account of

coefficients unequal, represents generally a

distinct points), is not regarded

is

by us as a point p a

although

group of 120

this

new

point

easily seen to be the intersection of three planes of second construction,

namely, of the three following, which

[oilii],
or

new group of sixty points of

whereas a quinary

this last congruence,

AA DIC^,

cc

that each plane,

point being
planes.
94.

D^Ao, EB B

2 ,i

common

c c2

all

belong to the group

[iloii],
.

It

2 ,i

[imo],

may, however, be remarked

new group

in passing,

every such
evident from what has been shown) to three such

contains twelve points P 3 of thig

(as is

From the foregoing discussion it appears that the five given points P O
the
three hundred derived points PJ, P 2 , are arranged in space, upon \kv fiftyand

GEOMETRICAL NETS IN SPACE.

ARTS. 93-94.]

five lines Ai,

and in the

2,

III,

305 points, forming on

of the

contains ^#A

forty-five planes

Each

as follows.

2,

what may be

it

Each

sub-article (8.) to 92) a double involution.

77

line

2 ,i

line

AI

called (see the

contains seven points,

one, namely the (/wen point, P O , has been seen (in the earlier sub- art.
to be a double point of another involution, to which the three derived pairs of

whereof
(6.))

on the same line belong.

And each line A 2 2 contains ten
on
it
a
new
while
involution
;
eight of these ten points, with
points, forming
a different order of succession, compose still another involution* (92, (10.)).
points,

p,,

P2

Again, each plane HI contain $ fifty-two points, namely three given points, four
Eacli plane
2 ,i con
points of first and 45 points of second construction.

tains forty-seven points, whereof one

42 are points P 2

of

which

last,

a given point, four are points

is

88 are situated on

the six lines

PI,

and

in the plane,

* These theorems
respecting the relations of involution, of given and derived points on lines of
and second constructions, for a net in space, are perhaps new although some of the harmonic
to whom, indeed, as has
relations, above mentioned, have been noticed under other forms by Mcibius
been stated, the conception of such a net is due. Thus, if we consider (compare the note to page 66)

first

the two intersections,

El

we

easily find that they

= DE

be denoted

may

EI

they

are, therefore,

by Art.

92, the

E2

AiBiCl,

= DB

A 2 B2 C2 ,

by the quinary symbols,

E2

(00012),

two points

P2,s

on the

(00021)
line

DE

and consequently, by the theorem

stated at the end of sub-art. (8.), the harmonic conjugate of each, taken with respect to the other and to
the point DI, must be one of the two points D, E on that line. Accordingly, we soon derive, by comparison

of the symbols of these five points, DEDiEiE 2 the

,

the same type aa the two last of that sub- art.

(DiDE 2 Ei)

(8.)

=-

two following harmonic

equations,

which belong

to

(DiEEiEa)

=-

but these two equations have been assigned (with notations slightly different) in the formerly cited
page 290 of the Barycentric Calculus. (Comp. again the recent note to page 66.) The geometrical
last equation may be illustrated,
by conceiving that ABCD is a regular pyramid, and
mean point ; for then (comp. 92, sub-art. (2.) ), DI is the mean point of the base ABC; DID
is the altitude of the
DiE 2 are, respectively, the quarter,
pyramid and the three segments DIE,
the third part, and the half of that altitude
they compose therefore (as the formula expresses) a
harmonic progression ; or DI and EI are conjugate points, with
But in order to
respect to E and E 2
exemplify the double involution of the same sub-art. (8.), it would be necessary to consider three other
points p 2 on the same line DE whereof one, above called D I, belongs to a known group P 2 ,i (92, (2.))
but the two others are of the group p 2 ,4, and do not seem to have been
previously noticed. As an
example of an involution on a line of third construction, it may be remarked that on each line of the
group As,3, or on each of the sides of any one of the ten triangles T3 2 in addition to one given point
and that the two first
PO, and one derived point p 2 ,i, there are two points r 2 ,2, and two points r 2 ,e

meaning of the

that E

is its

]>iEi,

points are the double points of an involution, to which the two last pairs belong
AOBCO of the exscribed triangle AoBoCo, or on the trace of the plane BCiA 2 AiC2,

thus, on the side

we have

the two

harmonic equations,
(BA O B"CO)

(BA

=-

T
"B"CI

1.

")

Again, on the trace A CO of the plane A ciC 2 (which latter trace is a line not passing through any one
v
of the given points), CQ and BI IT are the double
points of an involution, wherein A is conjugate to Ci
and A IX to B TI . But it would be tedious to
such
instances.
multiply

ELEMENTS OP QUATERNIONS.

78

[I.

m.

4.

but four are intersections of that plane n 2 ,i with four other lines of second
Finally, each plane II 2 2 passes through no given point, but
contain $ forty-three derived points, whereof 40 are points of second construction.
construction.

And because the planes

groups of points, P O ,PI,

of

lines,

P2>4,

Ai,A 2

1*2)5?

,i>

2 ,2,

of first construction alone contain specimens of all the ten

Pa,i,

given or derived, and of

P2,8,

all

the three groups

at the close of that second construction (since the types

Ai are not represented

by any points

or lines in

any plane

2 ,i,

nor

are the types P O Ai, A 2 ,i represented in a plane n 2 2 ), it has been thought

convenient to prepare the annexed diagram (fig. 30), which may serve to
,

illustrate,

by some

P 2 in a plane

PO, PI,

of the nine lines Ai,

View of

the

selected instances, the arrangement of the fifty-two points

n
A

i5

namely, in the plane ABC; as well as the arrangement

in that plane,

and the

traces

As of other planes upon

Arrangement of the Principal Points and Lines

in a

it.

Plane of

First Construction.

In

this figure, the triangle

ABC

is

supposed, for simplicity, to be the equi

and DU
a regular pyramid ABCD (comp. sub-art. (2.) to 92)
inscribed
The
mean
its
be
to
is
first
point (29).
again replaced by o,
supposed
and the axis of homology
triangle, A B C
therefore, bisects the three sides
lateral base of

A"B"C"

is

the

line at infinity (38)

the

number

1,

on the

line

cV

prolonged,

the
to
being designed to suggest that the point
type P 2 ,i, or belongs to the first group of points of second construction.
A",

which that

line tends, is of

GEOMETRICAL NETS IN SPACE.

AETS. 94-95.]

second inscribed triangle,

indicated
this

by the

bisecting point

but with an
AoBoCo, to

"B

"C

",

number 2 placed
A

accent, 2

"

which

fig.

21

may

be consulted,

middle of the side B C

at the

this corner also belongs (by a

The
3,

IV
which
point A
,

on the dotted

accent, lower

infinitely distant, is

line at the top

down, marks the position

A!

And

IX
.

only

syn typical relation in

now

is

Finally, the ten other numbers, unaccented or accented, 4, 4

V
VI
V
TI
A! , A
A^, A
8, 8 , denote the places of the ten points, A
,

is

to suggest that

space) to the group p a 8

indicated by the number

A IX

The same number 2,

belongs to the second group of p 2
is placed near the corner A O of the exscribed triangle

remind us that

number with an

for

79

while the same

of the point A! IT
,

",

5, 5

6,

7,

AX, A vm A/
,

1
",

the principal harmonic relations, and relations of involution,

above mentioned,

may

be verified by inspection of this Diagram.

However far the series of construction of the net in space may be

continued, we may now regard it as evident, at least on comparison with the
95.

analogous property (42) of the plane net, that every point, line, or plane, to
or
which such constructions can conduct, must necessarily be rational (77)
;

that

it

must be

rationally related to the

system of ike five given points

because

the (inharmonic co-ordinates (79, 80) of every net-point, and of every net-plane,
are equal or proportional to whole numbers. Conversely (comp. 43) every point,

which is thus rationally related to the system of points

a
ABODE,
point, line, or plane of the net, which those five points determine.
Hence (comp. again 43), every irrational point, line, or plane (77), is indeed
line,

or plane, in space,
is

incapable of being rigorously constructed, by any processes of the kind above

described but it admits of being indefinitely approximated to, by points, lines,
:

or planes of the net.

Every anharmonic

ratio,

or of & pencil of net-lines, or of net-planes,

which depends only on the

whether of a group of net-points,

has a rational value (comp. 44),

processes of linear construction employed, in the

generation of that group or pencil, and is entirely independent of the arrange

Also, all relations of
ment, or configuration, of the five given points in space.
collineation,

and

of complanarity are preserved, in the passage

,

from one net

to

so that it may be briefly

another, by a change of the given system of points
said (comp. again 44) that all geometrical nets in space are homographic figures.
:

Finally, any five points* of such a net, of which no four are in one plane, are

* These
general properties (95) of the space-net are in substance taken from Mb bius, although (as
has been remarked before) the analysis here employed appears to be new
as do also most of the
theorems above given, respecting the points of second construction (,92), at least after we pass beyond
:

the first group p 2 ,i of ten such points, which (as already stated) have been
long.

known comparatively

ELEMENTS OF QUATERNIONS.

80

[I.

m.

4, 5.

of the whole net, or for the linear

(comp. 45) for the determination
construction of all its points, including the five given ones.

sufficient

the five points A B C D and E be now supposed to

be required to derive the four points ABCD, by linear con

As an Example,

(1.)

be given

and

let

it

let

In other words, we are now required to

a
exscribe a pyramid ABCD to
given pyramid A^C^, so that it may be homo
An
logous thereto, with the point E for their given centre of homology.

from these new data.

structions,

obvious process is (comp. 45) to inscribe another homologous pyramid, A 3 B 3 c 3 D 3

and then to determine the intersections of
so as to have A 3 = P:AI B^D!, &c.
,

corresponding faces, such as AiBiCt and A 3 B 3 c 3

will be in the

common plane [E]

of

for thQBQfour lines of intersection

of the three pyramids,

homology

and

will be

drawn through the

construct one corner A

the traces on that plane of ike four sought planes, ABC, &o.,

four given points

of the exscribed

common

If

&c.

DI,

were only required to

find the point above called A IV as the

it

pyramid, we might

intersection of three planes, as follows,

A = AiB^i

A^E

and then should have

this other

formula of intersection,

A = EAi

Or the

A 3 B 3C3

point A might be determined

33iA

Iv
.

by the anharmonic

(EAAiA.,)

equation,

3,

which for a regular pyramid is easily verified.

from one net in space to another, let
(2.) As regards the general passage
=
P
v
the symbols PI = (x
5
5 ) denote any five given points, whereof
(xv^),
l

no four are complanar; and let a b c d e and u be six coefficients, of which

the five ratios are such as to satisfy the symbolical equation (comp. 71,72),
(P,)

(p,)

(p,)

or the five ordinary equations which

a xi +

x& -

t (p

it

4)

=-

e (P.)

(17)

includes, namely,
v\

e v&

Let P be any sixth point of space, of which the quinary symbol

satisfies

the

equation,
(p

= xa (?0 + yV

(P 2 )

(P 3 )

+ wtf (p 4 ) + ve (p 6 ) + u (U)

found that this last point P can be derived from the five
p
5
points P!
by precisely the same constructions, as those by which the
=
As an example, if
point P
(xyzwv) is derived from the five points ABODE.
then

it

will be
.

= x + y +

w -

3v,

then the point (xyzwv

is

derived from A

MEANS OF VECTOBS.

ARTS. 95-96.]

81

thus A itself may be

by the same constructions as (xyzwv} from ABODE
=
P
is
A
E which would
constructed from AI
the
from
as
E,
(30001)
point
conduct anew to the anharmonic equation of the last sub-article.
;

be briefly added here, that instead of anharmonic

may

It

(3.)

ratios, as

connected with a net in space, or indeed generally in relation to spatial

problems,

we

are permitted (comp. 68) to substitute products (or quotients) of

quotients of volumes of pyramids; as & specimen of which substitution, it may

be remarked, that the anharmonic relation, just referred to, admits of being

replaced by the following equation, involving one such quotient of pyramids,

but introducing no auxiliary point :

EA
In general,
in space,

if xi/zic

we may

be

AiA = SEBidDi

(as in 79, 83) the

AiBjCiDi.

anharmonic co-ordinates of a point P

write,

PBCD

EBCD

PCDA

ECDA

with other equations of the same type, on which we cannot here delay.

SECTION

On Barycentres
96.

of Systems of Points and on Simple and Complex

Means of Vectors.
:

In general, when the sum 2a

ai = OAi,
.

is

divided (16)

by

5.

of

any number
a m = OA m

of co-initial vectors,

their number, m, the resulting vector,

= OM = - 2a = - SOA,
m

is

said to be the Simple

this

mean

Mean

vector terminates,

of those

and

of

vectors

and the point M,

which the position (comp. 18)

to be independent of the position of the

common

origin o,

Point (comp. 29), of the system of the

we have the equation,

m points,

AI,

=
or that the

sum

of the system,

that these

Ilnd, that
HAMILTON

if

(a,

of the

ju)

vectors,

(a m

/i)

= S

drawn from

(a

the

/i)

Am

is
.

is

in

which

easily seen

said to be the

Mean

It is evident that

2MA

mean point M,

to

the points A

Aud

hence (cornp. 10, 11, 30), it follows 1st,

vectors are equal to the m successive sides of a closed polygon ;
the system and its mean point be projected, by any parallel
is

equal to zero.

ELEMENTS OF QUATERNIONS,

ELEMENTS OF QUATERNIONS.

82

m.

[I.

5.

on any assumed plane (or line), the projection M of the mean point u,
is the mean point of the projected system : and Illrd, that the ordinate MM
of
the mean pointy is the mean of all the other ordinates, AIA I,
A WI A WI
It follows,
ordinates,

also, that if

N be the mean point

of another system, BI,

combining the two former, considered as partial systems

point of the total system, AI

denote the

vectors,

ON and

have the equations,

= Sa,
nifi

= n

mean point,

numbers,
(1.)

and

and

(v

mean

last

and

if s

be the

points obtained by
while v and cr may

points

then we shall

= Sa + S/3 = m/m + nv,

so-

cr),

MS = w

SN

situated on the right line MN, which connects

s, is

M and N

which are

SN,

of the

two

nv = Sj3,

the two partial mean points,

two segments MS

Bn

os, of these

m (a - p)
so that the general

Bn

+ n =

mean

and

divides that line (internally), into

inversely proportional to

the two whole

n.

As an Example,

mean point ; or more

let

ABCD be a gauche

quadrilateral,

and

let

E be

its

fully, let

OE = 5 (OA + OB + 00 + OB),

= |

or

that

is

to say, let a

6/,

(a

+ y +

j3

S)

OA!

OA 2 =

Ql

a2

= i (j3 + 7 +
= \ (a + 8),

S),

write the vector formulce,

&

= i

(a

72

= J

(7

have seven different expressions for the mean

shall

following

/3

7),

S),

vector,

namely, the

= !(
=

And

we

/ = i(+/3),

OA = a =J(|3 + 7 ),..

we

Then, with notations

in the equations of Art. 65.

lately used, for certain derived points DI, &c., if

+ 3a,) =

(S

+ 380

-..*(/+

4( +aO

7.)-

these conduct to the seven equations between segments,

AE = 3EAi

A E = EA

2,

DE = 3EDi
C E

= EC 2

which prove (what is otherwise known) that the four right lines, here denoted
DD!, whereof each connects a corner of the pyramid ABCD with the
by AAu
.

mean

point of the opposite face, intersect and quadrisect each other, in one
r
common point, E and that the three common bisectors A A^, B BJ, c c 2 of pairs of
;

MEAN POINTS OF SYSTEMS.

ART. 96.]

such as BC and DA, intersect and

opposite edges,

mean

point

so that tlie/owr middle points, c

sides AB, &c., of the

which

plane,

bisects

each other, in the same

bisect

c2 ,

83

A 2 of the four successive

,

gauche quadrilateral ABCD, are situated in one common

also the common bisector, B B 2 , of the two diagonals, AC

and BD.

point E

this

of the derived
is

nine

is

point

D being four, the

example, the numher s of the points A
lines, which thus cross each other in their general mean
seen to be seven ; and the number of the derived planes through that

In

(2.)

number

namely, in the notation lately used for the net in space, four

lines Aj, three lines

A 2 ,i,

planes, the six former

six planes III,

and three planes

Of these

2 ,i.

(in the present connexion) be called

may

because each contains three

lines (as

nine

triple planes,

the plane ABE, for instance, contains the

through the mean point E and the three latter

be
to
be
said,
contrast,
may
by
non-triple planes, because each contains only
tivo lines
that
determined
on the foregoing principles.
through
point,

lines AAi, BBi, c c 2 ), all passing

In general,

(3.)

mean points

let

(f)

(s)

denote the number of the

a total system of

s of

given points,

lines,

which

is

through the general

thus, in all possible

; let/(s) denote the number of the triple

ways, decomposed
planes, obtained by grouping the given points into three such partial systems
let ^ (s) denote the number of non- triple planes, each determined by grouping

into partial systems

those

( )

point

points in two different

+
:

ways into two partial systems and let F (s)

represent the entire number of distinct planes through the

so that

0(4) =7,

Then

it is

easy to perceive

furnishes two new

lines,

/(4)-6,
that if we

*(4) =

3,

F(4)=9.

introduce a new point

c,

each old

line

MN

we group the new point with one or

(M) and (N) and that there is, besides,

according as

other of the two old partial systems,

one other neic line, namely cs: we have, therefore, the equation infinite differences,
;

<p

(s

+ 1) =

(s)

2<j>

+ 1

which, with the particular value above assigned for

(4), or even with the
=
simpler and more obvious value,
1, conducts to the general expression,
(2)
<j>

(s)

= 25

1.

(4.) Again,
(M) (N) (P) be any three partial systems, which jointly
make up the old or given total system (S) and if, by grouping a new point a
if

with each of these in turn, we form three new partial systems, (M ) (N )

then each old triple plane such as MNP, will furnish three new triple planes,

M NP,

MN

P,

MNP

ELEMENTS OF QTJATEENIO^S.

84

[I. in.

KL, will give one new triple plane, OKI, nor can any
have, therefore, this
triple plane be obtained in any other way.
while each old

line,

We

equation in differences

But we have
if

we

then

seen that

(s

write, for a

moment,

= 20

1)

(s)

=
/(3)

Also,

1,

=*(*)

2X

and

= 3s

2/(s)

Finally,
Sf(s)

because the

+ +

= }* ()

(.s)

drawn through the mean point

so that

and

(*

(s)

1)

^(* +
equation in

= 2

relation,

= (2- -

1)

2s - 3

1) (2

1)

and the wow-triple planes, each

three,

the binary combinations of the

all

+ 3

- 4F

l)

1.

the whole system.

s of

= 2 25

(s)

and

last

ft ()

must jointly represent

2^

1,

- 2s +

each treated as

triple planes,

treated as one,

we have the

clear that

it is

= 3

(s)

5 1

!.

X (3)=4:

tf>(3)=3,

therefore,

which

this other equation in finite differences,

X (i + 1)-8 X () +

(5.)

new
new

/W+*W
we have

5.

25

+ 2

S ~2

(*)

- 3s -

- 3 5-

= 3

s-1

lines*

Hence,

1
;

- 25

-4^(s) =3/(s);

finite differences

admits of an independent geometrical

interpretation.

For

(6.)

instance, these general expressions give,

0(5) = 15;
so that

it

we assume

/(5)=25;

(5)

= 30

(5)

= 55

a gauche pentagon, or a system of five points in space, A

and determine the mean point F

E,

of this system, there will in general be a set

of fifteen lines, of the kind above considered, all passing through this sixth

point F

and these

be arranged generally in fifty -five distinct planes,

whereof twenty-five will be what we have called triple, the thirty others being
:

will

of the non-triple kind.

97.

More

initial vectors,

if ai

generally,

and

if

a^

a m be

a,,,

be, as before, a

any system

of

system of

given

and

co-

given scalars (17), then that

ARTS. 96-97.]

new

BARYCENTKES OF SYSTEMS OF POINTS.

n =
n

is

--!a

#!

by the equation

or

which

co-initial vector j3, or OB,

2a

(a

deduced from these by the formula,

+ amam
=
+ am
Sa

)3)

0,

85

or OB

20BA =

or

0,

be said to be the Complex Mean of those m given vectors a, or OA, con

sidered as affected (or combined) with that system of given scalars, a, as
It may also be said that the derived
coefficients, or as multipliers (12, 14).

may

voint B, of
is

which (comp. 96) the

position is

independent of that of the origin

the Barycentre (or centre of gravity) of the given system of points A!

o,
.,

and theorems of intersections

considered as loaded with the given weights i
of lines and planes arise, from the comparison of these complex means, or bary.

centres, of partial

and

which are entirely analogous

simple means of vectors and of points.

to

total systems,

lately considered (96), for

(1.)

As an example,

those

in the case of Art. 24, the point c

of the system of the two points,

A and

B,

is the barycentre
with the weights a and b while,
;

under the conditions of 27, the origin o is the barycentre of the three points
A, B, c, with the three weights a, b, c; and if we use the formula for p,
assigned in 34 or 36, the same three given points A, B, c, when loaded with
xa, yb, zc as weights, have the point p in their plane for their barycentre.

Again, with the equations of 65, E is the barycentre of the system of the four
given points, A, B, c, D, with the weights a, b, c, d and if the expression of
79 for the vector OP be adopted, then xa, yb, zc, ivd are equal (or proportional)
;

D must be loaded, in
weights with which the same four points A
In all these cases,
order that the point p of space may be their barycentre.
the weights are thus proportional (by 69) to certain segments, or areas, or volumes,
of kinds which have been already considered
and what we have called the
to the

anharmonic co-ordinates of a variable point p, in a plane (36), or in space (79),

may be said, on the same plan, to be quotients of quotients of weights.

The circumstance

that the position of a barycentre (97), like that of a

independent of the position of the assumed origin of
vectors, might induce us (comp. 69) to suppress the symbol o of that arbitrary
and foreign point ; and therefore to write* simply, under the lately supposed
(2.)

simple mean point

(96), is

conditions,

B =

S#A

or

= S0A,

if

2.

We should thus have some of the principal notations of the Sary centric Calculus
mainly with a reference to vectors. Compare the note to page 50.
*

hut used

ELEMENTS OF QUATERNIONS.

86
It

[I.

m.

5, 6.

easy to prove (comp. 96), by principles already established, that the

ordinate of the barycentre of any given system of points is the complex mean (in
is

the sense above defined, and with the same system of weights), of the ordinates
of the points of that system, with reference to any given plane : and that the
projection of the barycentre,

on any such plane,

is

the barycentre of the projected

system.
"Without

(3.)

any reference

barycentric notation B

may

7-

or to

to ordinates,

any foreign

be interpreted, by means of

the

origin,

o\xr fundamental

convention (Art. 1) respecting the geometrical signification of the symbol B - A,

considered as denoting the vector from A to B together with the rules for
multiplying such vectors by scalar* (14, 17), and for taking the sums (6, 7, 8, 9)
of those (generally new) vectors, which are (15) the products of such multipli
:

For we have only

cations.

to write the formula as follows,

S0
in order to perceive that

it

may

(A

B)

0,

be considered as signifying, that the system

of the vectors from the barycentre B, to the system of the given points AI,

when
i,

az

such a manner that these last vectors, a { BAi, # 2 .BA 2

.

successive sides of a closed polygon,

(4.)

system
becomes (generally) a new system of vectors with a null sum : in

multiplied respectively by
,

A2

the scalars (or coefficients) of the given

Thus

if

we meet

interpret

by transports without

can be

made

(10) the

rotation.

the formula,

B = i

we may indeed

it

(A!

+ A 2 ),

an abridged form of the equation,

as

OB = i (oA! + OA 2 )

o be any arbitrary point, and if o be the point which

then B is the point which bisects
completes (comp. 6) the parallelogram AiOA 2 o
the diagonal oo , and therefore also the given line A^, which is here the other

which implies that

if

diagonal.

But we may

also

regard the formula as a mere symbolical transfor

mation of the equation,

(A 2

which ( by the
vectors, from B

B)

(AJ

B)

Book) expresses that the two

and A 2 have a null sum ; or that

earliest principles of the present

to the

two given points

AI

they are equal in length, but opposite in direction

bisecting

A^,

as before.

which can only be, by B

ANHARMONIC EQUATIONS OF SURFACES.

ARTS. 97-98.]

Again, the formula,

(5.)

B!

(AI

87

+ A 2 + A 3 ), may be interpreted as an

abridgment of the equation,

OBi

= \

+ OA 3 + OA 3 ),

(OAj

which expresses that the point B trisects the diagonal oo of the parallelepiped
But the same
(comp. 62), which has OAI, OA 2 OA S for three co-initial edges.
,

formula

may

also be considered to express, in full consistency

interpretation, that the

AI,

A 2 , A 3 vanishes
,

the Mangle AiA 2 A 3

sum of

which

is

the three vectors,

from B

the characteristic property (30) of the mean point of

the legitimacy of such

And similarly in more complex cases

with the foregoing

to the three points

transformations being here regarded as a consequence of the original interpre

tation (1) of the symbol B - A, and of the rules for operations on vectors, so far
as they

have been hitherto established.

SECTION

On Anharmonic

6.

and Vector Expressions of Surfaces

and Curves in Space.

liquations,

When, in the expression 79 for the vector p of a variable point p of

the
four variable scalars, or anharmonic co-ordinates, xyzw, are connected
space,
(comp. 46) by a given algebraic equation,
98.

fp (#,

y, 2,

w)

0,

or briefly

/=

0,

th
supposed to be rational and integral, and homogeneous of the p dimension,
then the point P has for its locus a surface ofthep th order, whereof /=
may

be said (comp. 56) to be the local equation. For

the co-ordinates x
wy expressions of the forms,
.

X = txQ +
to indicate (82) that P

algebraic equation in

modern mode

if

we

substitute instead of

is
:

W=

collmear with
is of

of speaking),

(distinct or coincident,

USCi

and

the

tw + UWi

two given points P O

th

degree;

the surface

PI,

the resulting

so that (according to a received

may

be said to be cut in

real or imaginary*), by

points

any arbitrary right line p p lt

* It is to be
observed, that no interpretation is here proposed, for imaginary intersections of this
The language of
kind, such as those of a sphere with a right line, which, is ivholly external thereto.
modern geometry requires that such imaginary intersections should be spoken of, and even that they

should be enumerated :
tly as the language of algebra requires that we should count what are called
the imaginary roots of an
But it would be an error to confound geometrical imaginaries, of
equation.
this sort, with those
square roots of negatives, for which it will soon be seen that the Calculus of
exa<

Quaternions supplies, from the outset, a definite and real interpretation.

ELEMENTS OP QUATERNIONS.

88

And

mamner, when the four anharmonic co-ordinates Imnr

in like

plane

[I. in.

by an algebraical equation

(80) are connected

FJ

m,

(/,

n, r)

of the form,

where F denotes a rational and integral function, supposed

th

class,

with F =
I

tlQ

has for

its

r - tr Q

...

to

be homogenous

(comp. 56) a
because if we make

envelope

for its tangential equation

tili,

of a rariabfe

or briefly r = 0,

0,

of the q th dimension, then this plane II

surface of the q

6.

to express (comp. 82) that the variable plane II passes through a given right
line

IIi,

we

are conducted to an algebraical equation of the q th degree,

gives q (real or imaginary) values for the ratio

(real or

which

u, and thereby assigns q

drawn through any such

imaginary*) tangent planes to the surface,

but
given
arbitrary right line.
may add (comp. 51, 56), that if the
functions /and F be only homogeneous (without necessarily being rational and

We

integral),

then

V,

[Da/,

D/]

D,/,

the anharmonic symbol (80) of the tangent plane to the surf ace

is

point (xyzw)

its

0, at

the

D m F, D n F, D r F)

(DjF,
is

/=

and that

symbol for the point of contact of the plane [Imnr], with

vx
enveloped surface, F =
D/,
being characteristics of partial deri

in like manner, a

vation.
(1.)

As an example,

the surface of the second order, which passes through

the nine points called lately

A,

has for

B,

gives,

by

= x% - yw

^;

is

E,

A,,

differentiation,
x

so that

D,

its local equation,

=
which

C2 ,

C,

[s,

- w,

m=

Vyf =

r =

D W /= - y

#,

/]

a symbol for the tangent plane, at the point

(x, y, z,

w).

* As
regards the tot interpreted character of such, imaginary contacts in geometry, the preceding
note to the present Article, respecting imaginary intersections, may be consulted.

ANHAEMOMC EQUATIONS OP

ABT. 98.]

In

(2.)

boloid,

also

fact,

the surface here considered

is

on which the gauche quadrilateral ABCD

through the point

And

E.

if

Q = (^00),

p = (xyzw),
then QS and RT

we

SURFACES.

the ruled (or hyperbolic] hyper-

is superscribed,

and which passes

write

R = (0#sO),

annexed figure

89

T =

(OOssw),

namely, the
lines drawn through P to intersect the two pairs, AB,
CD,

and

(see the

31),

BC, DA, of opposite sides of that quadrilateral

two generating

ABCD, are the

through that point

lines,

or generatrices,

so that their plane, QKST, is the

tangent plane to the surface, at the point P.

If,

we denote that tangent plane by the symbol

we have the equations of condition,
=

Ix

my = my + nz = nz + rw = rw +

Ix

then,

[lmnr~\,

whence follows the proportion,

/

or,

because xz = yw,
I

r = ar 1

y~

- w

z~

- w~

l
:

l
;

y,

as before.

At

(3.)

the same time

we

see that

(AC BQ)

(DC 2 cs)

so that the variable generatrix QS divides (as is

AB and DC homo graphically*

if it

Conversely,
divides

known)

were proposed

for those

two given

the two fixed generatrices

AD, BC, and c c 2 being three of

to

its positions.

find the locus of the right line QS,

homographically (comp. 26) two given right

AB and DC

lines,

and AD,

lines in space,

which thus

we might take

BC, c c 2 (with the recent

of the letters) for three given positions of the variable line

meanings
and then should

have, for the two variable but corresponding (or homologous) points Q, s
selves,

and

for

any

them

arbitrary point P collinear with them, anharmonic symbols

of the forms,

Q =
because,

by

(s,

82,

u, 0, 0),

we

(0, 0, u, s),

P =

(st, tu,

uv, vs)

should have, between these three symbols, a relation of

the form
()-<()

if

then

write p =

we

(x, y, z,

HAMILTON

w),

we have

Compare 298

ELEMENTS OF QUATERNIONS.

+()
the anharmonic equation xz = yw, as

of the Geometric Superieure.

ELEMENTS OE QUATERNIONS.

90
before

whether of the

so that the locus,

line QS,

[I.

or of the point p,

m.

6.

is (as is

known) a ruled surface of the second order.

As

(4.)

R =

we

known

regards the

to observe that if

we

double generation of that surface,

may

it

suffice

write, in like manner,

T =

(OfrO),

(P)

(*00*>),

= u

(R)

(T),

have again the expression,

shall

p =

(st, tu,

xz - yw,

giving

uv, vs),

so that the same hyperboloid is also the locus of that other line RT,
which divides the other pair of opposite sides BC, AD of the same gauche quadri
r
lateral ABCD homographically
BA, CD, and A A 3 being three of its positions
as before

and the

lines

given point

each other in the

AO, c c 2 being still supposed to intersect

E.

The symbol

(5.)

art. 2)

of the form,

an arbitrary point on the variable

-f u
(x, 0, 0, w), or (ux, ty,
(0, y, z, 0)

symbol of an arbitrary point on the given line c c 2 is (If,

two symbols represent one common point (comp. fig. 31),
p = RT-C

when we suppose

line

of

if

u =

y,

Hence the known theorem

z,

CS

=
=

(y, y, z,

1,

u =

tz,

f
,

RT

uw)
u ).

(by subwhile the

is
;

And

these

),
^*

?y

= -

iv

results, that a variable generatrix, RT, of one system,

intersects three fixed lines, BC, AD, c c 2 ,

which are

generatrices of the other system,.

Conversely, by the same comparison of symbols, for points on the two lines RT
and c c z we should be conducted to the equation xz = yw, as the condition for
,

their intersection

and thus should obtain

this other

known theorem,

locus of a right line, which intersects three given right lines in space, is

that the

generally

an hyperboloid with those three lines for generatrices.

similar analysis
shows that QS intersects A r A 2 in a point (comp. again fig. 31) which may be
,

thus denoted

p"=QS-A

A 2 = (xyyx).

As another example of the treatment of surfaces by their anharmonio

and local equations, we may remark that the recent symbols for P and
(6.)

p",

combined with those of sub-art. (2.) for p, Q, R, s, T with the symbols of 83,
86 for c A c 2 A E and with the
=
equation xz yw, give the expressions
;

(P)

2,

(Q)

(s)

(R)

(T)

(P

= y

(c

(c 2 )

(R)

y
+ -

(T)

JO

(E)

(c

(c,)

= (A

(A,)

(P")

= y

(A

+x

(A,)

(Q)

+?

ANHARMONIC EQUATIONS OF SURFACES.

AET.98.]

whence

it

follows (84) that the two points p

lateral ABCD, divide the four

anharmonic

ratios

generating lines

and the

sides of the quadri

p
and
E in the following
through
P",

(c

(A

so that (as again is

EA 2

= - =
= (AA S
(BA CR)
DT)

EC 2 p

(QP"SP)

p")

= (RP TP) = - =
00

known) the

The

tangential equation of the present surface is easily found,

be the following

may

(cc 2 Ds)

variable generatrices, as well as fas fixed ones,

expressions in sub-art. (1.) for the co-ordinates

which

(BC AQ)

of the hjperboloid, are all divided homographicalty

(7.)

91

by the

Imnr of the tangent plane, to

= F =

In

- mr

be interpreted as expressing, that this hyperboloid

which touches the nine planes,

is

the surface of

the second class,

[1000], [0100], [0010], [0001], [1100], [0110], [0011], [1001], [1111];

or with the literal symbols lately

CD A,

BCD,

Or we may

DAB,

CDC

ABC,

ABC

DAA.",

87),

BCA

2,

same tangential equation F =

interpret the

again 86, 87, where

employed (comp. 86,

Q, L,

are

now

replaced

by

2,

and

[E].

as expressing (comp.

T, R, Q),

that the surface

is

the envelope of a plane QRST, which satisfies either of the two connected conditions
of

homography :
(BC AQ)

(CA BR)

= -- =
n

%=

2 Ds)
(cc
^

= (DA S AT)

a double generation of the hyperboloid thus showing itself in a new way. And
as regards the passage (or return), from the tangential to the local equation
(comp. 56),

we have

x = D/F = n

whence

xz -

(8.)

More

in the present

y = D OT F = r

yw =

example the formulae

0,

= DnF =

to

= Dr F

as before.

generally,

when

the surface

also of the second class, so that the

is

of the second order,

two functions / and

F,

and therefore

when presented
N2

ELEMENTS ON QUATERNIONS.

92

m.

[I.

6.

under rational and integral forms, are both homogeneous of the second dimen
w by the formulae,
r from x
sion, then whether we derive I

or

a?

w from

by the

X =

converse formulae,

^ = D OT F,

DjF,

= D n F,

W=

D,-F,

^IQ point P = (xyzw) is, relatively to that surface, what is usually called (comp.
=
52) the pole of the plane II
[//;] ; and conversely, the plane II is the polar
of the point P

may

other,

order

is

wherever in space the point p and plane

And

be situated.

known

thus related

to

each

because the centre of a surface of the second

be (comp. again 52) the pole of (what

to

II,

the

is called)

plam

at

infinity; while (comp. 38) the equation and the symbol si this last plane are,

respectively,

ax + by +

cz

the four constants abed have

if

dio

and

0,

[a, b,

c, d~\,

the same significations as in 65, 70, 79,

still

&c., with reference to the system of the five given points ABODE
that we may denote this centre by the
symbol,

K=
where F O denotes,

(D a F

DjF

D CF

DrfF

follows

still

a scalar

abridgment, the function F (abed), and d

for

it

is

constant.

In the recent example, we have F O = ac - bd

symbol for the centre of the hyperboloid becomes thus,
(9.)

KAccordingly

if

we assume

p =

where

s, t,

u, v are
s

if also

we

= bt +

any four
f

shall then

ac

p//= (*
x"=

and

b).

scalars,

= cu +

p = (sY, -

and P

new

is

u = dv +

ds,

uv

- v

s ),

point, while

at,

f
v = as + bu

yVW)

2est - cw

therefore,

w =

bd,

ast

btu

+ cuv + dvs

have the symbolic relations,

(*)+

write, for abridgment,

e

we

d, a,

(comp. sub-arts. (3.), (4.),)

uv, vs),

(st, tu,

cv,

(c,

and the anharmonic

be that

^(p)-M--(Oi

00 -*(*),
new

point, of

which the co-ordinates

rr

y = 2e tu + dw
ax"+

z"

by"+

cz"

= 2e uv - aw
dw"=

0.

w"

are,

= 2efvs + bw

ANHARMONIC EQUATIONS OF SURFACES.

ART. 98.]

93

which passes through the

be the harmonic conjugate of that fixed point, with
fixed point K, and if
= - 1, then this conjugate
variable
to
that
chord, so that (PKP
respect

That

is to

say, if PP be any chord of the hyperboloid,

p"

P")

point
point

or in other words, the fixed

on the infinitely distant plane [abed ]
bisects all the chords pp which pass through it, and is therefore (as above

P"

is

asserted) the centre of the surface.

With

the same meanings (65, 79) of the constants a, b, c, d, the

mean point (96) of the quadrilateral ABCD, or of the system of its corners, may
(10.)

be denoted by the symbol,

if

M=

(a-\ b~\ c-\

d->)

then this mean point be on the surface, so that

- bd =

ac

the centre

0,

on the plane [a, b, c, d~\

or in other words, it is infinitely
distant: so that the surface becomes, in this case, a ruled (or hyperbolic)
In general (comp. sub-art. (8.)), if F O = 0, the surface of the
paraboloid.

second order

is

is

a paraboloid of some kind, because

its centre is

then at

infinity,

in virtue of the equation

=
(aD a + bn b + CD C + dv d ) FO
or because (comp. 50,
58) the plane [_abcd^\ at infinity
planes, as satisfying its tangential equation, F = 0.
(11.)

It

is

is

then one of

evident that a curve in space

two anharmonic and local equations

section of two surfaces.
And then

may be represented by a
because it may be regarded as

its

its order,

tangent

system of
the inter

or the number of points (real or

imaginary*), in which it is cut by an arbitrary plane, is obviously the product

of the orders of those two surfaces
or the product of the degrees of their two
;
local equations,
(12.)

supposed to be rational and integral.

may also be considered as the edge of

curve of double curvature

regression (or arete de rebroussement) of a developable surface, namely of the

locus of the
tangents to the curve ; and this surface may be supposed to be cir

cumscribed at once
(98),

to

two given surfaces, which are envelopes of variable planes

and are represented, as such, by

a curve of double curvature

may

their tangential equations.

itself

In

this

view,

be represented by a system of two

and if the class of such a curve be

anharmonic and tangential

equations;
defined to be the number
of its osculating planes, which pass through an arbitrary
point of space, then this class
*

is

the product of the

Compare the notes

classes of

to pages 87, 88.

the two curved sur-

ELEMENTS OF QUATERNIONS.

94
faces just

now mentioned

6.

same thing) it is the product

by which the curve is (on this

comes

or (what

m.

[I.

to the

of the dimensions of the two tangential equations,

But we cannot enter further into
plan) symbolized.

these

details

the

mechanism of calculation respecting which would indeed be found to be the

known method

same, as that employed in the

(comp. 41) of quadriplanar co

ordinates.

anharmonic co-ordinates, we may consider any other system

xn which enter into the expression of a variable
#1,

99. Instead of

of n variable scalars,

And

then,

scalar,

scalar

p; for example, into an expression of the form (comp. 96, 97),

vector,

t,

those n scalars x be

if

we may regard

and

may

functions of one independent and variable

all

p as being

this vector

if

and

the n scalars x
u,

a function of that single

write,

L..p-f
But

itself

(Q.

be functions of two independent and scalar variables,

then p becomes a function of those two scalars, and we may write
.

accordingly,
II.

In the

p =

(t,

u).

term p (comp. 1) of the variable vector p has generally for

a curve in space, which may be plane or of double curvature, or may

1st case, the

its locus

even become a right line, according to the form of the vector-function $ and p
may be said to be the vector of this line, or curve. In the Ilnd case, p is the
;

vector of a surface, plane or curved,

according to the

manner

in

which

this vector

form of
<p

p depends on the two independent

(t,

u)

or to the

scalars that enter

into its expression.

As examples

(1.)

(comp. 25, 63), the expressions,

a +

T
L

"

a+ t3 + Uy
IL -P =
i + t + u

p= T7f

signify, 1st, that p is the vector of a variable point P on the right line

that

it

is the vector

Ilnd, that p
locus of
(2.)

is

itself*

AB

considered as the locus of a point

the vector of the plane ABC, considered in like

manner

or

and

as the

an arbitrary point p thereon.

The

equations,
I.

with

of that line

x-

y*

p = xa +

= 1 for the

signify 1st, that p

is

II.

yfi,

1st,

and

the vector of an

x?

ellipse,

p = xa +
if

/j3

sy,

1 for the Ilnd,

and Ilnd, that

it is

the vector of

VECTORS OF SURFACES AND CURVES IN SPACE.

ARTS. 98-99.]

an

common

with the origin o for their

ellipsoid,

and with OA,

centre,

95

OB, or OA,

OB, oc, for conjugate semi-diameters.

(3.)

The equation (comp.

expresses that p

t*a

+ U zf3 +

(t

y,

tt)

the vector of a cone of the second order, with o for

is

which

(or centre),

46),

is

touched by the three planes OBC, OCA, OAB

its vertex

the section of

made by

the plane ABC, being an ellipse (comp, fig. 25), which is

and the middle points A B c of the sides of that
inscribed in the triangle ABC
this cone,

triangle, being the points of contact of those sides with that conic.
(4.)

The equation (comp.

=

t~

a + u~

53),

fi

+ v~ l y,

with

u + v =

0,

is the vector of another cone of the second order, with o still for
but with OA, OB, oc for three of its sides (or rays). The section by the

expresses that p
vertex,

plane ABO

is

new

circumscribed to the triangle ABC,

ellipse,

and having

its

tangents at the corners of that triangle respectively parallel to the opposite sides
thereof.
(5.)

The equation (comp.

54),

= Pa + w 3 |3 +
p
signifies that

is

with

v*y,

+u + v =

0,

the vector of a cone of the third order, of which the vertex

is

the origin
its section (comp. fig. 27) by the plane ABC being a cubic
curve, whereof the sides of the triangle ABC are at once the asymptotes, and the
still

three (real) tangents of inflexion; while the mean point (say o ) of that triangle
is a conjugate point of the curve ; and therefore the right line oo
from the
,

vertex o to that
(6.)

mean

may

point,

The equation (comp.

be said to be a conjugate ray of the cone.

98, sub-art. (3.)

),

+ uvcy + vsd$
+ tub + uvc + vsd

staa + tubfi
sta
s

in which

- and

- are
v

constant scalars, and

two variable
a, j3,

scalars,

while

a,

c,

b,

7, S are four constant vectors,

variable vector, expresses that p

is

d are

still

but p

is

four

still

the vector of a ruled (or single-sheeted)

on which the gauche quadrilateral ABCD is superscribed, and which

passes through the given point E, whereof the vector e is assigned in 65.
hyperboloid,

(7.)

If

we make

(comp. 98, sub-art. (9.) ),

-tu
bp + tifv cy v
,_ s faa
s tfa

where

bt

cv,

= cu +

ifu b

ds,

+ uv

f
u = dv +

at,

d$_

d
v

as

bit,

ELEMENTS OF QUATERNIONS.

96

[I. ni.

6, 7.

then p = OP is the vector of another point p on the same hyperboloid; and

because it is found that the sum of these two last vectors is constant,
f

follows that K

it

the vector of a fixed point K, which bisects every chord PP

is

that passes through

it

or in other words (comp. 52), that this point

is

the

centre of the surface.

(8.)

The

three vectors,

are termino-collinear (24)

on a ruled hyperboloid,

~~o~^

K,

if

the

2~~

then a gauche quadrilateral ABCD be superscribed

common

bisector

of the two diagonals, AC, BD, passes

through the centre K.

(9.)

When

ac

bd, or

when we have the

sta +
st

or simply,

sta

-H

tuft

+ uvy +

is

still

is infinitely distant,
),

superscribed.

And

with

=u=

%,

In general,

it is

AB and DC

l,

mean point M of that

D because, when
the four given points A
.

of the

line,

is

last vector -expression for p,

which

two

divides similarly the

same gauche quadrilateral ABCD

or the other

and AD.

On
The equation

(1.) ),

8).

easy to prove, from the

SECTION

in which p = OP

centre (comp. 52, and 98,

but upon which the quadrilateral ABCD is

the locus of a right

is

pair of opposite sides, BC

100.

the variable vector p takes the value (comp. 96, sub-art.

that this paraboloid

opposite sides

+ u =

which the

p-j(a-i./3 + r +
(10.)

this surface passes through the

quadrilateral, or of the system of

s

vs$,

then the vector of a ruled paraboloid, of

sub-art. (10.)

equation,

tuft + uvy + vs$

+ tu + uv + vs

7.

Differentials of Vectors.

(99, I.),

p =

(0>

generally the vector of a point p of a curve in space,

gives evidently, for the vector oa of another point Q of the

expression of the form

P +

Ap =

tf>

(t

Atf)

PQ

same curve, an

DIFFERENTIALS OF VECTORS.

AST. 100.]

chord PQ, regarded as being

represented (4) by the finite difference,
so that the

PQ -

Suppose now
the n

th

A/o

A0

n th part of

(t

A)

scalar,

a vector, comes thus to be

(t).

that the other finite difference,

part of a new

A/o, or PQ,

(t)

itself

97

A,

is

u ; and that the chord

manner (comp. fig. 32), the

so that we may
new
a
vector, v n or PR
in like

is

write,

and wA/o = n

u,

we

if

Then,

variable (the
t 0efor

/o

PQ =

treat the two scalars,

/orm

<r

= PR.

and

u, as constant,

and the _pow

A and A/), and the multiple

but the number n as

being given), the
Q and R, the two

origin o,

p will be fixed: but the &0o points

vector

differences

mn/

and the

of the vector-function 0,

And

together.

if

this

number

be

n&p, or

<r

w , will

(in general)

made

indefinitely increased, or

to tewd

then e#cA of the two differences A, Ap

n~ l u) so that the
such being the common limit, of n~ u, and of
(t)
(t +
But
variable point Q of the curve will tend to coincide with the fixed point p.
will in general tend to zero;

fo infinity,

<f>

although the chord PQ will thus be

indefinitely shortened, its

n th multiple, PR or

or n ,

depending on the supposed continuity of

will tend (generally) to a finite limit*

to a certain definite vector, PT, or a^, or (say) r,

the function

(t); namely,
which vector PT will evidently be

(in general) tangential to the curve

other words, the variable point R will tend

We

shall thus

r = PT =

Km. PR =

that curve at p.

have a
<T

to

or, in

a fixed position T, on the tangent to

limiting equation, of the

= lim.

wA0

(t),

form

if

W-co

and u being, as above, two given and (generally) finite scalars. And if we
call the second of these two given scalars the differential of the

then agree to
first,

and

to denote it

by the symbol

dt,

we

shall define that the vector-limit, T or

the (corresponding) differential of the vector p, and shall denote it by the

so as to have, under the supposed conditions,
corresponding symbol, dp
rr^, is

u =

cU,

and

= dp.

Or, eliminating the two symbols u and r, and wotf necessarily supposing that P
is a j90m
o/ # curve, we may express our Definition^ of the Differential of a
*

HAMILTON

Compare Newton
S

Principia.

ELEMENTS OF QUATERNIONS.

Compare the Note

to

page 35.

ELEMENTS OE QUATERNIONS.

98

Vector p, considered as a Inunction $ of a Scalar

[I.

m.

7.

by the following General

t,

Formula :
= lim. n
in

which

and dp

is,

and dt are two arbitrary and independent scalars, both generally finite;
in general, a new and finite vector, depending on those TWO scalars,

according to a law expressed by the formula, and derived from that given law,
scalar, t.
whereby the old or former vector, p or (tf) depends upon the
sw</fe

(1.)

As an example,
p =

Then, making A^ =
number,

-,

let

the given vector-function have the form,

= It

<f)(l)

where w

a,
is

where a
<wy

<7M

0w0w

scalar,

+ JT-

and

is

a variable whole

we have

= au
and

is

finally,

writing d and

In general,

for

d/o

t
(

<r

u and

where a

a given or constant vector,

and /(^) denotes a scalar function of the scalar variable, t. Then because a is
a common factor within the brackets
of the recent general formula (100)
(2.)

let

()

a/(^),

is still

for dp,

we may

write,

provided that we

dp = d0

now

(t)

= d

af(t]

- ad/(*)

define that the differential

new scalar function of two independent

precisely similar formula

scalars,

of a scalar function, f(t)

and

dt,

is

determined by the

lim.

which can easily be proved to

agree, in all its consequences,

with the usual rules

for differentiating functions of one variable.

(3.)

For example,

if

we write d =

where h

nh,

is

th
namely, the n part of the given and (generally) finite
thus have the equation,

+
d/(0
-- ,./(*
urn.
dc
in

which the

first

member

is

h=o

A)-

a new variable scalar,

differential, dt,

we

shall

-/(Q.
,

here considered as the actual quotient of two finite

DIFFERENTIALS OF VECTORS.

ART. 100.]
:

scalars, df(t)

99

and not merely as a differential coefficient. We may, however,

from the expression of which the differential dt

dt,

as usual, consider this quotient,

and may denote

disappears, as a derived function of the former variable, t;

such,

by

it,

as

either of the two usual symbols,

f(t) and n tf(t).

In

(4.)
<[>(t),

like

manner we may

write, for the derivative of a vector-function*

the formula :

these two last forms denoting that actual and finite vector, p or $

obtained, or derived,

dp or

by

d(f)(t),

this last scalar

by

h,

we

which

is

dividing (comp. 16) the not less actual (or finite) vector,

And

the finite scalar, dt.

by

(t),

shall thus

Dtp

= D,0

(t)

th
again we denote the n part of

if

have the equally general formula

r
= hm.

+
^--

h)

<f>(t

h=o

ft

with the equations

dp

= Dtp dt = pdt
.

d<

()

D$

(tf)

dt =
<j>

(t)

dt,

the vector-function, p or 0, were a scalar function, f.

d = 1, gives thus dp = p so that the derived
(5.) The particular value,
vector p is (with our definitions) a particular but important case of the diffe
In applications to mechanics, if t denote the time, and if the
rential of a vector.

exactly as

if

term P of the variable vector p be considered as a moving point, this derived

f
vector p may be called the Vector of Velocity : because its length represents
the amount, and
off vectors

its direction is

And if, by setting

one origin, to represent thus the

the direction of the velocity.

ov = p (comp. again

fig.

32)

from

moving in space according to any supposed law, expressed

=
of which the corre
we construct a new curve vw
the
(t)
equation p
by

velocities of

a point

<

f
be written as p =

(Q, then this new curve has been

sponding equation may
old
PQ
defined to be the HODOGRAPH,! as the
may be called the orbit of the
.

motion, or of the moving point.

*

In the theory of Differentials of Functions of Quaternions, a definition of the differential

(q)
will be proposed, which is expressed by an equation of precisely the same form as those above assigned,
for d/(0, and for d0() ; but it will be found that, for quaternions, the quotient d(f>(y}: dq is not
that it cannot properly be called a derived function,
generally independent of dq ; and
d<

consequently
of the quaternion q alone.
(Compare again the Note to page 35.) [See 327.]
t The subject of the Sodograph will be resumed at a subsequent stage of this work. In fact, it
almost requires the assistance of Quaternions, to connect it, in what appears to be the best mode, with

such as

</>

Newton

(<?),

Law

of Gravitation.

[Compare 419.]

ELEMENTS OP QUATERNIONS.

100

(6.)

We may

differentiate

we

7.

a vector-function twice (or oftener), and so obtain

For example,

its successive differentials.

[I. in.

if

we

differentiate the derived vector

obtain a result of the form,

d/o

by an obvious extension
ddt or d z t, of the scalar

of notation
t is

d>

where

where

=p"dtf,

zero,

p"=

and

vp=v
t

if we

as usual, denotes (dt)

*p,

suppose that the second differential,

then the second

= ddp = d

differential

9-

p dt = dp dt =
,

and where

p".

of the vector p

is,

2
;

important to observe that,

with the definitions adopted, d p is as finite a vector as dp, or as p itself. In
applications to motion, if t denote the time,
may be said to be the Vector of
dtf

it is

p"

Acceleration.

We

(7.)
may also say that, in mechanics, the finite differential dp, of the
Vector of Position p, represents, in length and in direction, the right line
(suppose PT in fig. 32) which would have been described, by a freely moving point

p, in the finite interval

of this time

of time dt, immediately following the time

all foreign forces

had ceased

to

t,

^/at the end

act*

In geometry,

if p
^(t) be the equation of a curve of double curvature,
as
the
regarded
edge of regression (comp. 98, (12.)) of a developable surface, ihen
the equation of that surface itself, considered as the locus
of the tangents to the

(8.)

curve,

may

be thus written (comp. 99, II.)

P=
if it

utf(t)

<t>(t)+

or simply, p =

(j>(t)

+
d<j>(t),

be remembered that

u, or dt, may be any arbitrary scalar.

If
other
curved
(9.)
any
surface (comp. again 99, II.) be represented by an
=
of
the form, p $(x,y}, where
now denotes a vector -function of two
equation
independent and scalar variables, x and y, we may then differentiate this equation,
or this expression for p, with
respect to either variable separately, and so obtain

what may be

called two partial (but

finite)

partial derivatives,

differentials, d xp, d p, and two

v xp, v yp whereof the former are connected with the latter,

and with the two arbitrary (but

finite)
d xp = vxp dx
.

And

tj

scalars, dx, dy,

;

d yp =

T>

yp

by the

dy.

these two differentials

denote two tangential

relations,

(or derivatives) of the vector p of the surface

or
at least two vectors parallel to two tangents to
vectors,

* As

is

well illustrated

by Atwood

machine.

DIFFERENTIALS OE VECTORS.

ART. 100.]

that surface at the point P

so that their plane

is (or is

101

parallel to) the tangent

plane at that point.

The mechanism

of all such differentiations of vector-functions is, at the

same as in the usual processes of the Differential
the
present stage, precisely
most
the
because
Calculus ;
general form of such a vector-function, which has
(10.)

been considered in the present Book, is that of a sum of products (comp. 99)
of the form xa, where a is a constant vector, and x is a variable scalar : so that

we have only

to operate on these scalar coefficients

.,

by the usual rules

of the

calculus, the vectors a. .being treated as constant factors (comp. sub-art. 2).

But when we

shall

come

to consider quotients or products of vectors, or generally

those new functions of vectors which can only be expressed (in our system)

Quaternions, then

some few new

although deduced from the same

which have been established in the present
(11.)

function

As an example
word

(the

variables, let

"

by

of differentiation become necessary,

(or nearly the same) definitions, as those

rules

section.

of partial differentiation (comp. sub-art. 9) of a vector

vector

"

being here used as an

adjective)

of

two scalar

us take the equation

in which p (comp. 99, (3.)) is the vector of a certain cone of the second order ; or,
more precisely, the vector of one sheet of such a cone, if x and y be supposed
to

be real

and

scalars.

Here, the two partial derivatives of p are the following

therefore,

2p
so that the three vectors, p,

v xp, v y p,

if

drawn

(18)

from one common

origin, are

contained (22) in one common plane ; which implies that the tangent plane to
the surface, at any point p, passes through the origin o and thereby verifies
the conical character of the locus of th&t point P, in which the variable vector p,
:

or OP, terminates.
(12.) If, in the

whence

it

same example, we make

= l, y = -

follows that the middle point, say o

the points of the conical locus

99,

o?

and the recent

sub-art. 9)

1,

we have

of the right line AB, is

and that (comp. again the

the right lines OA and OB are
;

the values,

one of

sub-art. 3 to Art.
parallel to

two of

the tangents to the surface at that

point; so thut the cone in question

is

ELEMENTS OF QUATERNIONS.

102

[I.

m.

7.

And in like manner it

touched by the plane AOB, along the side (or ray) oo
may be proved, that the same cone is touched by the two other planes BOG and
.

and B

COA, at the middle points A

and CA

of the two other lines BC

therefore along the two other sides (or rays),

OA and OB

and

which again agrees

with former results.

variables,

Series,

it

(f>t

d2

and d,

may

generally be developed, by an extension of Taylor

%at* (as in sub-art. 1),

= ad

2
,

d3
<^

0, &o.

<}>(t

rigorously,

0,

d3 ^ =

where a

&c.

0,
is

(comp. sub-art.

a constant

vector,

6).

we have

Thus,

if

d0= atdt,

and

+ dO - a(t +

dt)*

+ atdt + J adt z

without any supposition that dt is small.

thus suppose A = d, and develop the finite difference, A$(rf)
-$(t), the first term of the development so obtained, or the term of

When we

(14.)

= $(t +

two scalar

under the form,

2
=
being supposed that d ^

sum of

It will be found that a vector function of the

(13.)

dt)

first dimension relatively to d^,

is

hence (by a theorem, which holds good for

vector-functions, as well as for scalar functions) the first differential dtyt of the

function

but we do not choose to define that this Differential is

because the Formula (100), which we prefer,

that first term:

postulate the possibility, nor even suppose the conception, of

Many

recent remarks will perhaps appear

more

clear,

means)
does not

(or

any such development.

shall come to

when we

connect them, at a later stage, with that theory of Quaternions, to

which we

next proceed.

[Compare generally III. ii. Two elementary illustrations of Hamilton s

method are given in 2 of the Chapter cited. It may be of interest to refer
to Art. xxvni. of J. Clerk Maxwell s
Another
Matter and Motion."
"

mode

of obtaining the

diagram

"

of velocities of a system at a given instant is

th
to take a small interval of time,
say the n part of the unit of time, so that
the middle of this interval
Take the
corresponds to the given instant.

diagram of displacements corresponding to this interval and magnify all its

dimensions n times. The result will be a diagram of the mean velocities of
the system during the interval. If we now
suppose the number n to increase
without limit the interval will diminish without limit, and the mean velocities

DIFFERENTIALS OF VECTORS.

AKT. 100.]

approximate to the actual velocities at the given instant.

n becomes infinite the diagram will represent accurately the
will

given

The

instant."

sub-art. (5).

In a

unit of time

letter to

Hamilton says
would suit me almost

n.

It

when

Finally,

velocities at the

of course not necessarily small

De Morgan,

vol. in., p. 629),

value of

is

103

compare

dated April 26th, 1852 (Graves

Life,

I lay no stress on the infinitely great

as well to define

"

though I think the other form a little clearer. But the important thing is
that I avoid
2nd, development in series
1st, commutation of factors
;

3rd, smallness of

differentials."]

BOOK

II.

ON QUATERNIONS, CONSIDEEED AS QUOTIENTS OF VECTORS, AND AS

INVOLVING ANGULAE RELATIONS,

HAMILTON

ELEMENTS OF QUATERNIONS.

CHAPTER

I.

FUNDAMENTAL PRINCIPLES RESPECTING QUOTIENTS OF

VECTORS.

SECTION
Introductory

Remarks

1.

First Principles adopted

from Algebra.

The only angular relations, considered in the foregoing Book, have

been those of parallelism between rectors (Art. 2, &o.)
and the only quotients,
ART. 101.

hitherto employed, have been of the three following kinds:

I.

II.

Scalar quotients of scalar-s, such as the arithmetical fraction

in Art. 14

Vector quotients, of rectors divided by scalar s, as - = a in Art. 16

III. Scalar quotients of vectors, with directions either similar or

opposite, as

= x in the

QUOTIENTS

But we now propose to

last cited Article.

(or geometric Fractions, as

- = - =

OA

ct

q,

with

we
|3

treat of other geometric

shall also call them), such as

not

\\

a (conrp. 15)

each of which the Divisor (or denominator), a or OA, and the Dividend
(or
numerator), |3 or OB, shall not only both be Vectors, but shall also be inclined

for

to each other at

an ANGLE,

distinct (in general)

horn zero, and from two*

right

angles.

102. In introducing this new conception,^ a General Quotient of

Vectors, with
Angular Relations in a given plane, or in space, it will obviously be necessary
to employ some properties of circles and spheres, which were not wanted for
* More

generally speaking, fron? every even multiple of a right angle.

ELEMENTS OP QUATERNIONS.

108

[II.

i.l.

the purpose of the former Book.

But, on the other hand, it will be possible
and useful to suppose a much less degree of acquaintance with many important
theories* of modern geometry, than that of which the possession was assumed,

Indeed

in several of the foregoing sections.

amount
found

it is

hoped that a very moderate

and trigonometrical preparation will be

render the present Book, as well as the early parts of the

of geometrical, algebraical,

sufficient to

preceding one, fully and easily intelligible to any attentive reader.

103. It may be proper to premise a few general principles respecting
quotients of vectors, which are indeed suggested\>j algebra, but are here adopted

by

And

definition.

(whatever

we here

its full

it

1st,

evident that the supposed operation of division

is

geometrical import

may

afterwards be found to be), by which

conceive ourselves to pass from a given divisor-line

dividend-line

j3,

to

what we have

a,

and from a given

called (provisionally) their geometric quotient,

must) be conceived to correspond to some converse act (as yet

not fall// known) of geometrical multiplication in which new act the former
quotient, q, becomes a FACTOR, and operates on the line a so as to produce (or
q,

may

(or rather

generate] the line

/3.

/3

even

if

We

= q

a,

shall therefore write, as in algebra,

or simply,

the two lines a and

other, as in fig. 33.

j3,

And

or

]3

qa,

although we

shall,

the following

for

a = q

OA and OB, be supposed to be inclined to each

and natural notation (comp. 16)

two following formulsB

qa
a

\/3
a =
pa=/3;
J a

the present, abstain

=q;

from writing

also

such

f ormulaef

as

j3a
-

where

|3

this very simple

will then allow us to treat as identities the

|3
-

when

a, |3 still

denote two

vectors,

/3,

and

-q a

q,

q denotes their geometrical quotient

* Such as
homology, homography, involution, and generally whatever depends on anharmonic ratio
although all that is needful to be known respecting such ratio, for the applications subsequently made,
may be learned, without reference to any other treatise, from the definitions incidentally given, in

Art. 25,

c.

theories, in

It was, perhaps, nut

of these modern geometrical
strictly necessary to introduce any
of the present work ; but it was thought that it might interest one class, at least,

any part

of students, to see how they could be combined with that fundamental conception of the
the First Book was designed to develop.

t It will be seen, however, at a later stage, that these

required, in the development of the Quarternion System.

VECTOK, which

two formulae are permitted, and even

FIRST PRINCIPLES ADOPTED FROM ALGEBRA.

ARTS. 102-107.]

because

we have not

even begun to consider the multiplication of one

yet

by another or the division of a quotient by a line.

104. As a Ilnd general principle, suggested

109
vector

it

down, that

= 2,
^
a
a
or in words,

by

algebra,

we

shall next lay

if

and under a

and

o =o,

then

/3

]3

slightly varied form, that unequal vectors, divided by

The importance

equal vectors, give unequal quotients.

of this very natural

and

obvious assumption will soon be seen in its applications.

105. As a Illrd principle, which indeed may be considered to pervade the
whole of mathematical language, and without adopting which we could not
usefully speak, in any case, of EQUALITY as existing between any two geome

we

trical quotients,
to the

same third*

shall next

assume that two such

quotient, without being at the

quotients can never be equal

same time equal

to

each other

or in symbols, that
f

if q

106. In the

IVth

q,

and

<?"

q,

then

=
q"

a preparation for operations on geometrical

quotients, we shall say that any two such quotients, or fractions (101), which
have a common divisor-line, or (in more familiar words) a common denominator,
place, as

among themselves, by adding, subtracting, or

dividing their numerators the common denominator being retained, in each
of the two former of these three cases.
In symbols, we thus define (comp. 14)

are added, subtracted, or divided,

:

that, for

any three (actual)

vectors, a,

|3,

y,

and

aa
aiming

still

at

agreement with algebra.

Yth principle, designed

107. Finally, as a

(like the foregoing) to assimilate,

so far as can be done, the present Calculus to Algebra, in its operations

geometrical quotients,

we

shall define that the following

on

formula holds good

q
\<3

4
>

It is scarcely
necessary to add,

identity q

q,

that

if

q,

then q

what

is
;

indeed included in this Illrd principle, in virtue of the

that we shall never admit that any two

or in words,

geometrical quotients, q and q are equal to each other in one order, without at the same time admitting
that they are equal, in the opposite order also.
,

ELEMENTS OF QUATERNIONS.

110

or that if two geometrical fractions , q

and

[II.

i.

1, 2.

be so related, that the denominator,

(here written towards the left-hand) is equal to the

/3,
numerator of the multiplicand q, then the product, q
q or qq, is that third
of the multiplier q

fraction, whereof the numerator is the numerator j of the multiplier, and the
denominator is the denominator a of the multiplicand: all such denominators,

or divisor-lines, being

still

supposed (16) to be actual (and not

SECTION
First Motive for

null] vectors.

2.

naming the Quotient of two Vectors a Quaternion.

we may

grounds for the application of the name,

QUATERNION, to such a Quotient of two Vectors as has been spoken of in recent
108. Already

In the

articles.

called (17) a

For

kind.

SCALAR

let

(actual) vector

(actual)

vector,

place, such a quotient cannot generally be

first

of the (so- called)

see

or in other words,

reals of algebra,

it

what we have

cannot generally be equal to any

whether of the

positive or of the negative

x denote any such (actual*) scalar, and let a denote any

then we have seen (15) that the product xa denotes another

say

|3

which

either similar or opposite in direction to a,

is

according as the scalar coefficient, or factor, x,

is positive or negative ; in
such
as j3, which is inclined to a,
any vector,
at any actual angle, whether acute, or right, or obtuse
or in other words
=
=
the
under
the conditions
or
is
oca
equation /3
(comp. 2),
impossible,
j3,
]3,
here supposed.
But we have agreed (16, 103) to write, as in algebra,

neither case, then, can it represent

Qftt

= x

we must,

therefore (by the Ilnd principle of the foregoing section,

stated in Art. 104), abstain

from writing

also

3=
a

x,

under the same conditions:

still
denoting a scalar. Whatever else a quotient of two inclined vectors
be found to be, it is thus, at least, a NON-SCALAR.

109.

Now,

in forming the conception of the scalar

two parallel^ vectors (17),

we took

In passing from a

as the quotient of

into account not only relative length, or ratio

of the usual kind, but also relative direction,

opposition.

itself,

may

to xa,

we

under the form of

similarity or

altered generally (15) the length of

one that is different from

1), we mean here
for product (15) a null vector; it is therefore
unnecessary to prove that the quotient of two actual vectors cannot be a null scalar, or zero.
f It is to be remembered that we have proposed (15) to extend the use of this term parallel, to
zero.

By an actual scalar, as by an actual vector (comp.

An actual vector, multiplied by a null scalar, has

the case of two vectors which are

(in the usual sense of the word) parallel to one common
when they happen to be parts of one and the same right line.

line,

even

QUOTIENT OF TWO VECTOKS A QUATERNION.

ARTS. 107-111.]

the line

a,

in the ratio of

+ x

to 1

of that line, according as the scalar

Ill

and we preserved or reversed the direction

And
coefficient x was positive or negative.

manner, in proceeding to form, more definitely than we have yet done,

the conception of the non-scalar quotient (108), q = j3 a = OB OA, of two in
in like

which for simplicity

clined vectors,

have

still

to take account

both of the

be supposed (18) to be

may

relative length,

and of the

co-initial,

we

relative direction,

compared. But while the former element of the complex

relation here considered, between these two lines or vectors, is still represented
of the

two

lines

by a simple RATIO (of the kind commonly considered in geometry), or by a

number* expressing that ratio the latter clement of the same complex relation
is now represented by an ANGLE, AOB
and not simply (as it was before) by
;

an algebraical

sign,

+ or

-.

110. Again, in estimating this angle, for the purpose of distinguishing one
quotient of vectors from another, we must consider not only its magnitude (or

PLANE: since otherwise, in violation of the principle

we should have OB OA = OB OA, if OB and OB were two

quantity], but also its

stated in Art. 104,

a cone of revolution, with OA for its axis in which

would necessarily be unequal vectors. For a similar reason,
we must attend also to the contrast between two opposite angles, of equal
magnitudes, and in one common plane. In short, for the purpose of knowing
distinct rays or sides of

case (by 2) they

fully the relative direction of two co-initial lines OA, OB in xpace,

not only how many degrees, or other parts of some angular

but also (comp. fig, 33)
unit, the angle AOB contains
the direction of the rotation from OA to OB
including a
of
the
in
which
the rotation is perplane,
knowledge

we ought

to

know

formed

and of the hand (as right or left, when viewed

from a known side of the plane), towards which the rotation
;

111. Or,

and

if

we agree

to select

to call all rotations positive

* This
number, which,

we

is directed.

some one fixed hand (suppose the

when they

rightf hand),

are directed towards this selected

shall

presently call the tensor of the quotient, may he whole or frac

even incommensurable with unity but it
may always be equated, in calculation, to a positive
scalar: although it might
perhaps more properly be said to be a signless number, as being derived solely
from comparison of lengths, without
any reference to directions.
t If riyht-handed rotation be thus considered as positive, then the
positive axis of the rotation AOB
in fig. 33, must be conceived to be directed,
downward, or below the plane of the paper. [Compare the
295 (2), and Art. 23 of Clerk Maxwell s
Hamilton compared the
Electricity and Magnetism.
Note^to
positive axis to a handle or turnscrew used in screwing a
It is now
right-handed screw into a nut.
tional, or

usual to regard the positive axis as drawn in the direction of the

translation of a right-handed screw
moving in a fixed nut, or Hamilton s left-handed rotation is now caUed
right-handed.]

ELEMENTS OF QUATERNIONS.

112

hand, but

when they

rotations negative

all

[II.

2, 3.

i.

are directed towards the other hand,

then, for any given angle AOB, supposed for simplicity to be less than two right
a rotation in a given plane from OA to
angles, and considered as representing
to that plane AOB as
oc
of one perpendicular
being the positive
OB, we

may

speak

axis of that rotation

and

same plane as
the rotation round the positive axis being
And then the rotation AOB may be considered to

of the opposite perpendicular oc to the

being the negative axis thereof

and

itself positive,

be entirely known,
right rotation

vice versa.
if

we know,

and Ilnd, the

1st, its quantity, or

the ratio which

direction of its positive axis4 oc

it

bears to a

but not without

a knowledge of these two things, or of some data equivalent to them.

But
whether we consider the direction of an Axis, or the aspect of a PLANE, we find
(as

an

indeed

well known) that the determination of such a direction, or of such

is

depends on TWO polar co-ordinates* or other angular elements.

112. It appears, then, from the foregoing discussion, that/or the complete

aspect,

determination, of

what we have

called the geometrical

QUOTIENT of two

co-initial

System of Four Elements, admitting each separately of numerical

Vectors, a

expression,

is

determine the

generally required.
relative length of

Of these four elements, one serves (109) to

the two lines compared and the other three
;

are in general necessary, in order to determine fully their relative direction.

Again, of these three latter elements, one represents the mutual

elongation, of the

tween them

two

lines

while the

perpendicular to their

tivo

inclination, or

or the magnitude (or quantity) of the angle be

others serve to

common

determine the direction of the

plane, round which a

axis,

through that
as
the
selected
positive one (or
previously
rotation

angle is to be performed, in a sense

towards a fixed and previously selected hand], for the purpose of passing
the simplest way, and
tion of the divisor-line,

to

(in

two

lines) from the direc

the direction of the dividend-line. And no more than

therefore in the plane of the

the
four numerical elements are necessary, for our present purpose because
are
altered
two
their
relative length of two lines is not changed, when
lengths
:

proportionally, nor is their relative direction changed,

they form

is

merely turned about,

in

its

own plane.

when

On

the angle which

account, then, of this

two lines, which is

complex relation (109) between
compounded of a relation of lengths, and of a relation of directions, and to which
we have given (by an extension from the theory of scalars) the name of a
essential connexion of that

* The

actual (or at least the frequent) use of such co-ordinates is foreign to the spirit of the
hut the mention of them here seems likely to assist a student, by suggesting an
rendered him familiar.
appeal to results, with which his previous reading can scarcely fail to have

present System

ADDITIONAL ILLUSTEATIONS.

AHTS. 111-114.]

FOUR

geometrical quotient, with a System of

already a motive* for saying, that

a

"

113

we have

numerical Elements,

the Quotient of tico Vectors

is

generally

Quaternion"

SECTION

3.

Additional Illustrations.

Some

113.

Quaternion,
34.

additional light

by the annexed

In that

be thrown, on this

first

conception of a

figure

the letters

figure,

CDEFG are designed

may

indicate

to

corners of a prismatic desk, resting

upon a horizontal
angle HCD (supposed

The

table.

to be one of

thirty degrees) represents a (left-

handed)

rotation,

the

whereby

horizontal ledge CD of the desk

is

conceived to be elongated (or re

moved) from a given horizontal

CH, which may be imagined

Fig. 34.

line

to be an edge of the table.

The angle GCF
the
contain
here
to
sloprf of the desk, or
forty degrees) represents
(supposed
On the face CDEF of. the desk are
the amount of its inclination to the table.

drawn two

similar

and

similarly turned triangles,

AOB and A O B

which are

be halves of two equilateral triangles in such a manner that

each rotation, AOB or A O B is one of sixty degrees, and is directed towards one

supposed to

common hand (namely the

attended

to,

the side O A
114.

the side OB
,

in the other

Under

hand

right

is to
;

in the figure)

while

if lengths

alone be

the side OA, in one triangle, as the side O B

is

to

or as the number two to one.

these conditions of construction,

we

consider the two quotients,

or the two geometric fractions,

.

OB OA and OB
:

as being equal to each other

having the same

,
:

because

OA

OB

,
,

or

O B

and -7-7,

we regard

and the same

relative length,

OA

o A

the two

lines,

OA and OB, as

relative direction, as

the

tico

other

* Several other reasons for thus

speaking will offer themselves, in the course of the present work,
angles, HCD and GCF, may thus he considered to correspond to longitude of node, and

t These two

inclination of orbit, of a planet or

HAMILTON

comet

in astronomy.

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

114
lines,

oV

And we

and oV.

as a Quaternion

because

its

[II.

i.

3, 4.

consider and speak of each Quotient, or Fraction,

complete construction (or determination] depends,

and

for all that is essential to its conception,

requisite to distinguish

others, on a system of four numerical elements (comp. 112)

which

it

from

are, in this

Example, ihefour numbers,

2,

115.

the

2, expresses that

double of the length

is

and

30,

40.

elements (to recapitulate

namely the number

1st,

or O B

Of these four

60,

what has been above supposed),

the length of the dividend-line, OB

of the divisor-line,

OA or O A

The Ilnd

numerical element, namely 60, expresses here that the angle AOB or A O B is
one of sixty degrees while the corresponding rotation, from OA to OB, or
,

from O A to O B

by

towards a knoicn hand (in this case the right hand, as seen
a person looking at iheface CDEF of the desk), which hand is the same for
,

is

two equal angles. The Illrd element, namely 30, expresses that
the horizontal ledge CD of the desk makes an angle of thirty degrees with a
known horizontal line CH, being removed from it, by that angular quantity,
both of these

in a

known

direction (which in this case happens to be towards the

from above).
that the desk has an
as seen

116.

Now

an

tion of the slope

Finally, the

IV th

hand,

elevation si forty degrees as before.

any one of these Four Elements, such as an altera

or aspect of the desk would make (in the view here taken) an
alteration in

essential change in the

Quaternion,

which

the two lines

(as the figure is in part

compared

no such change

left

element, namely 40, expresses here

is

although

conceived to take place,

is

(in

when

the same view) the Quotient of

designed to suggest)

the triangle AOB

is

merely turned

own plane, without being turned oxer (comp. fig. 36) or when
the sides of that
triangle are lengthened or shortened proportionally, so as to
about, in its

preserve the ratio (in the old sense of that word), of any one to any other of
those sides.
"We
may then briefly say, in this mode of illustrating the notion
of a QUATERNION* in
geometry, by reference to an angle on a desk, that the

Four Elements which

it

involves are the following

Ratio,

although the two

Angle,

Ledge,

and

Slope

latter elements are in fact themselves angles also,

but are not

immediately obtained as such, from the simple comparison of the two

which the Quaternion is the Quotient.
As

to the

"

Quaternio,"

lines,

of

mere word, Quaternion, it signifies primarily (as is well known), like its Latin original,
noun rerpaKrvs, a Set of Four: but it is obviously used here, and else

or the Greek

where in the present work,

in a technical sense,

AETS. 114-119.]

CONDITIONS OF EQUALITY OF QUATERNIONS.

SECTION

On Equality of Quaternions

115

4.

and on the Plane of a Quaternion.

an immediate consequence of the foregoing conception of a

two quaternions, or two quotients of vectors,
that
Quaternion,
supposed for
117. It

is

simplicity to be all co-initial (18), are regarded as being

EQUAL

to each other,

or that the EQUATION,

3
-= -,
by us considered and
two triangles, AOB and

is

and

turned,

in one

OD

OB

- =
oc
OA

or

defined to hold good,

COD, are

common

similar

when

the

and similarly

plane, as represented in the

annexed fig. 35: the RELATIVE LENGTH (109), and the

Fig. 35.
RELATIVE DIRECTION (110), of the two lines, OA, OB,
being then in all respects the same as the relative length and the relative
direction of the two other

Under

118.

OD.

lines, oc,

same conditions, we

the

shall write the

following formula of

direct similitude,

A AOB

COD

GC

reserving this other formula,

AAOBGC AOB
which we shall

call

or

A OB OCA OB

a formula of inverse similitude, to denote that

the two triangles, AOB and AOB

or

A OB and A OB

although

otherwise similar (and even, in this case, equal* on account of

their having a common side, OA or OA ), are oppositely turned

(comp.
mirror

fig.

36), as if one

were the

Fig. 36.

reflexion of the other in a

or as if the one triangle were derived (or

generated) from the other,
a
rotation
its
two
of
plane through
by
right angles.
may therefore write,
;

We

OB

OA

When

119.

.,
,

if

AOB

all

COD.

drawn from one common

them may be

origin o,

called the Plane of the Quaternion

amount of area, but with opposite algebraic signs (28). The two
OA, although not equal (110), will soon be denned to be conjugate quater
OA, and OB
Under the same conditions, we shall write also the formula,

quotients
nions.

oc

the vectors are thus

the plane AOB of any two of

* That

OD

is

OB

to say, equal in absolute

:

A AOB

oc

COD.

ELEMENTS OF QUATERNIONS.

116

the Quotient), OB

(or of

OA

and of course

[II.

i.

4.

also the plane of the inverse (or

And any tivo quabe

to be Complanar*
said
ternions,
o), may
but any two quaternions
Quaternions, or complanar quotients, or fractions
have
which
different planes (intersecting therefore in a right line
(or quotients),
OA OB.

reciprocal) quaternion (or of the inverse quotient),

which have a common plane (through

through the origin),

120.

Any

reduced to a

said,

by

contrast, to be Diplanar.

common denominator without change

Let

them, as follows.

and

be

may

two quaternions, considered as geometric fractions (101), can be

OA

and

of the valued of either of

be the two given fractions, or quaternions

oc

they be complanar (119), let OE be any line in their common plane but
they be diplanar (see again 119), then let OE be any assumed part of the
line of intersection of the two
planes so that, in each case, the line OE is
if

if

always conceive two other

lines,

We

can then
plane COD.
OF, OG, to be determined so as to satisfy the

and

situated at once in the plane AOB,

also in the

two conditions of direct similitude

(118),

A
and therefore

EOF

OC

AOB,

also the two equations

EOG

GC

COD

between quotients (117, 118),

OF

OB

OG

OD

OE

OA

OE

oc

and thus the required reduction is effected, OE being the common denominator
sought, while OF, OG are the new or reduced numerators. It may be added
that

if

have

be a

new

point in the plane AOB, such that

AOB,

we

shall

also,

OE

and

A HOE a

therefore,

OB

OF

by 106, 107,

OD

OB

OGOF

OD OB

OC

OA

OE

OC OA

OG
OF

OD OB
OC

whatever two geometric quotients (complanar or diplanar)

OA

may

OG
OH
be represented

by OB OA and OD oc.
:

* It

however, convenient to extend the use of this word, complanar, so as to include the case
which have equal
by angles in parallel planes. Indeed, as all vectors
of two such vectors,
lengths, and similar directions, are equal (2), so the quaternion, which is a quotient
in position,
ought not to be considered as undergoing any change, when either vector is merely changed
by a transport without rotation.
t That is to say, the new or transformed quaternions will be respectively equal to the old or
is,

of quaternions represented

given ones.

PLANE OF A QUATERNION.

ARTS. 119-123.]

117

now

the two triangles AOB, COD are not only complanar but directly
so
similar (118), so that A AOB oc COD, we shall evidently have A EOF a EOG
that we may write OF = OG (or F = G, by 20), the two new lines OF, oo (or the
121. If

two new points

The general

F, G) in this case coinciding.

construction (120),

common

denominator, gives therefore here only one

new triangle, KOF, and one new quotient, OF OF,, to which in this case each
(comp. 105) of the two given equal and complanar quotients, OB OA and

for the reduction to a

OD

oc, is equal.

But if these two latter symbols (or the fractional forms corresponding)
denote two diplanar* quotients, then the two new numerator-lines, OF and OG,
have different directions, as being situated in two different planes, drawn through
122.

new denominator-line OE, without having either the direction of that

the

itself,

even

line

they are therefore (by 2) unequal

whence it follows
to be equally long

or the direction opposite thereto

they should happen

(by 104) that the two new quotients, and therefore also (by 105) that the two

vectors,

if

old or given quotients, are unequal, as a consequence of their diplanar ity.

results, then,

from

this analysis, that diplanar quotients of vectors,

fore that Diplanar Quaternions (119), are always unequal; a

It

and there

new and compara

tively
process thus confirming the conclusion, to which we had
arrived by general considerations, and in (what might be called) a popular
way before, and which we had sought to illustrate (comp. fig. 34) by the con
technical

sideration of angles on a desk

namely, that a Quaternion, considered as the

quotient of two mutually inclined lines in space, involves generally a Plane, as

an

essential part

(comp. 110) of

its

constitution,

and

as necessary to the

com

pleteness of its conception.

123.

We propose

to use the

mark

as a Sign of Complanarity, whether of lines or of quotients

thus we shall

write the formula,

y
to express that the three vectors, a,
co-initial (18), are situated in one

And

plane, but

whatever.

by

/3,

plane

ft
y,

supposed to be

(or to

be made)

and the analogous formula,

for a scalar, considered as a quotient (17), has no determined

since it may be represented
as complanar with every geometric quotient
the quotient of two similarly or oppositely directed lines, in any proposed plane

therefore non-scalar (108)

must be considered

(or constructed)

III

ELEMENTS OF QUATEENIONS.

118

to express that the two quaternions, denoted here

that the four vectors,

just

a, /3, 7, S,

and

And

found (122) that diplanar quotients are unequal, we

of quaternions includes two complanarities of vectors

may

by

are complanar (119).

and

0,0,

HI

the equation of quotients, - =

OC
OA
in one

common

124.

With

is in (or

in

also

and therefore

we have

manner that we

-=;

if

0,0,

We shall

plane.

to express that the vector

HI

being impossible, unless

from o be

because

4, 5.

see that one equation

in such a

write,

i.

[II.

all the

four

lines

employ the notation

<?,

parallel to) the plane of the quaternion q.

the same notation for complanarity,

*ll

we may

write generally,

,/3;

two vectors, and x being any scalar

because, if a = OA and
)3 being any
= OB as before, then (by 15, 17) sea = OA where A is some point on the
/3
indefinite right line through the points o and A so that the plane AOB contains

a and

the

line

OA

For a

similar reason,

we have

generally the following formula

of complanarity of quotients,

fc
whatever two scalars x and
125. It

is

if

whence

it

evident (comp.

A AOB oc

is

COD,

be

may

//

111

a and

fig.

35) that

then

A BOA a

/3 still

denoting any two

and

DOC,

AOC

BOD

vectors.

easy to infer that for quaternions, as well as for ordinary or

algebraic quotients,
if

it

- =

-, then, inversely,

^N

ay

~ = 2 and alternately, - =

pc

73

being permitted now to establish the converse of the last formula of 118, or

to say that
..

OB

if

Under

the same condition,

also this other equation,

OD

oc

OA

,,

then

AOB

COD.

by combining inversion with

=

alternation,

we have

AXIS AND ANGLE OP A QUATERNION.

ARTS. 123-128.]

119

126. If the sides, OA, OB, of a triangle AOB, or those sides either

or

A"B",

way

AB
fig. 37) by any parallel,
to the base AB, we have evidently the rela

prolonged, be cut

(as in

tions of direct similarity (118),

A OB

oc

AOB,

A"OB"

whence (comp. Art. 13 and

we may

fig.

oc

12)

AOB

follows that

it

__

write, for quaternions as in algebra, the

general equation, or identity,

?;
xa

where x
see,

is

again any

Fig. 37.

scalar,

that for any quaternion

and
q,

a, ]3

are any two vectors.

and any

scalar x,

we have

It is easy also to

the product (comp.

107),

= ff/3

=
a.

jo

af)

= J3^ =

x~ a

/3

x~^a

so that, in the multiplication of a quaternion by a scalar

(as in the multipli
cation of a vector by a scalar, 15), the order of the factors is
indifferent.

SECTION

On
127.

5.

the Axis and Angle of a Quaternion ; and on the Index

of a Right Quotient, or Quaternion.

From what

has been already said (111, 112), we are

naturally led
more fully that the positive axis, of any quaternion

to define that the Axis, or

(or

geometric quotient] OB

that quaternion

and

is

OA, is a right line perpendicular to the plane AOB of

such that the rotation round this axis, from the divisor:

line OA, to the dividend-line

OB, is positive

or (as

we

shall henceforth assume)

directed towards the right-hand* like the motion of the hands of a watch.
128. To render still more definite this
conception of the axis of a qua
we
ternion,
may add, 1st, that the rotation, here spoken of, is supposed (112)
to be the simplest
possible, and therefore to be in the plane of the two lines

(or

of the
quaternion),

plane

being also generally less than a semi-revolution in that

that
the
axis shall be usually supposed to be a line ox drawn
Ilnd,

This

is,

of course, merely
conventional, and the reader may (if he pleases) substitute the left[The axis is supposed to be drawn outwards from the face of the watch. See

hand throughout.
Note, page 111.]

ELEMENTS OF QUATERNIONS.

120

[IT.

i.

5.

assumed origin o and Illrd, that the length of this line shall be
supposed to be given, or faced, and to be equal to some assumed unit of length
so that the term x, of this axis ox, is situated (by its construction) on a given

from

the

spheric surface described about the origin o as centre,

call

which surface we

may

the surface of the UNIT-SPHERE.

129. In this manner, for every given non-scalar quotient (108), or for

every given quaternion q which does not reduce itself (or degenerate) to a
positive or negative number, the axis will be an entirely definite vector,

mere

which

be called an UNIT- VECTOR, on account of

may

which we shall denote*

for the present,

then the usual sign of perpendicularity,

vectors

a, ]3,

the formula

130.

The ANGLE

Ax

by
we may now

J_,

length,

and

Employing
any two

q.

write, for

Ax.2j_/3

Ax.2_La;
a

assumed

its

the symbol

(a

of a quaternion, such as OB

the angle AOB letween the two

Ax.^JLJ^.
a

orbriefly,

of

lines,

OA, shall simply be, with us,

which the quaternion

is

the quotient

angle being supposed here to be one of the usual kind (such as are con
sidered by Euclid)
and therefore being acute, or right, or obtuse (but not of
this

any

from

class distinct

We

when

these),

the quaternion is a non-scalar (108).

by the symbol, L q and thus shall

shall denote this angle of a quaternion q,

have, generally, the two inequalities f following

L q

where

TT

131.

is

L q

>

<TT\

used as a symbol for two right angles.

When

the general quaternion,

axis (like the planety

angle takes, at the

q,

degenerates into a scalar, x, then the

becomes entirely indeterminate

same time,

in its direction

and the

either zero or two right angles for its value,

according as the scalar is positive or negative.

such scalar by x, we have

Denoting then,

as above,

any

Ax

x - an indeterminate unit-vector
0.
if x
L x = TT, if x

L x = 0,
* At a
the

less

suffice.

Ax q, of a quaternion q, by
but for the present, the notation in the text may

later stage, reasons will be assigned for denoting this axis,

arbitrary (or

<

>

more systematic) symbol,

13 Vq

[See 291.]

t In some investigations respecting complanar quaternions, and powers or roots of quaternions, it

is convenient to consider
but these may then
negative angles, and angles greater than two right angles
be called AMPLITUDES and the word Angle," like the word
thus be restricted, at
Ratio," may
:

"

"

least for the present, to its ordinary geometrical sense.

J Compare the Note to page 117.

the quaternion reduces itself to zero

The
;

angle,

unless

[See 235.]

as well as the axis,

we happen
n

dividend-line tends to become null, in the transition from -

to

know

A
to -.

becomes indeterminate, when

which the

a law, according to

CASE OF A RIGHT QUOTIENT, OR QUATERNION.

AKTS. 128-132.]

132.
angle

Of

non-scalar quaternions, the

as in the

is right,

121

most important are those of which the

B
and when we have

annexed figure 38

thus,

OB
OA

and OB

or L q

_L OA,

the quaternion q may then be said to be a

or sometimes, a Right Quaternion.
If then a

(1.)

points, but P

is

and OA

(3.)

(or fixed)

2
is

the plane through

o,

perpendicular

to

it is

for axis,

that the angles AOB

common

RIGHT QUOTIENT;*

equivalent to the formula of perpendicularity p _L a (129).

=
generally, if (3 OB, B being any third given point, the equation,

for

expresses that the foczs of p

vertex,

expresses that the locus of this point p

More

^,

a variable point, the equation

(2.)

TT

= OA and p = OP, where o and A are two given

the line OA

is

one sheet of a cone of revolution, with o for

and passing through

and AOP are equal

the point B

because

it

implies

in amount, but not necessarily in one

plane.

The equation (comp.

128, 129),

or
expresses that the locus of the variable point P is the given plane AOB
rather the indefinite half-plane, which contains all the points P that are at
once complanar with the three given points o, A, B, and are also at the same side
;

of the indefinite right line OA, as the point B.

(4.)

The system

of the two equations,

expresses that the point P

is

on the

situated, either

line prolonged through B, but not through o

case be said to be the indefinite half-line, or ray,

* Reasons

finite right line OB, or

so that the locus of p

which

sets

may

on that
in this

out from o in the

will afterwards be assigned, for equating such a quotient, or

quaternion, to a Vector

namely to the line which will presently (133) be called the Index of the Eight Quotient.
HAMILTON S ELEMENTS OF QUATERNIONS,

[See 290.]

ELEMENTS OF QUATERNIONS.

122

OB or

direction of the vector

understood to be a

]3

and we

may

write p =

xft,

5, 6.

i.

[II.

(x

>

being

scalar], instead of the equations assigned above.

This other system of two equations,

(5.)

expresses that the locus of P is the opposite ray from o

is situated on the prolongation of the revector BO
;

px

or that P
(1)

or that p =

xfi,

(Comp.

fig.

33,

or that

<

= x$

x>0,

if j3

=OB =-/3.

bis.)

Other notations, for representing these and other geometric loci, will
be found to be supplied, in great abundance, by the Calculus of Quaternions
(6.)

seemed proper to point out these, at the present stage, as serving

already to show that even the two symbols of the present section, Ax. and L,

but

it

when considered

as Characteristics of Operation

on

quotients of vectors, enable

us to express, very simply and concisely, several useful geometrical conceptions.

133. If a third line, 01, be drawn in the direction of the axis ox of such a
right quotient (and therefore perpendicular,
given rectangular lines, OA, OB)

and

by 127, 129,

if the length of this

to each of the two

new

line 01 bear to

the length of that axis ox (and therefore also, by 128, to the assumed unit of
the
length) the same ratio, which the length of the dividend- line, OB, bears to
length of the divisor-line,

the

INDEX of the Right

OA

then the line

Quotient.

And

01,

it is

thus determined,

evident, from

is

said to be

this definition of

such an Index, combined with our general definition (117, 118) of Equality
between Quaternions, that two right quotients are equal or unequal to each other,
according as their two index- lines (or indices) are equal or unequal vectors.

SECTION

On
134.
as q

=
,

6.

the Reciprocal, Conjugate, Opposite, and Norm of a

Ctuaternion ; and on Null Quaternions.

The EECIPROCAL
is

(or the Inverse,

comp. 119) of a quaternion, such

that other quaternion.

formed by interchanging the divisor-line and the dividend-line and in

thus passing from any non-scalar quaternion to its reciprocal, it is evident that

which

is

RECIPROCAL OF A QUATERNION.

ARTS. 132-137.]

123

the angle (as lately defined in 130) remains unchanged, but that the axis (127,
so that we may write generally,
128) is reversed in direction
:

a_=

p
135.
unity

The product

and each

we have, by

of

a_

/3

p~

a*

/3

^7

fo#0

always equal to

reciprocal quaternions is

wm tfy

equal to the quotient of

is

divided ly the other

positive

because

106, 107,
a 3
-.
a a

i.i
i
.

a/3a
~

ann
ana
I

77,

I
i.

any new or peculiar notation, to ex

a quaternion and its reciprocal
between
relation
mutual
the
existing
press
the
other
the
if
one
be
denoted
may (in the present System,
since,
symbol q,
by
It is therefore unnecessary to introduce

as in Algebra) be denoted

by the connected symbol,*

thus the two general formulae (comp. 134)

1

L - = L q
136.

Without

1
Ax .

q,

or -.

We

have

= -

Ax

q.

on the general theory of multiplication and

beyond what has been done in Art. 120, it may be

yet entering

divisions of quaternions,

here remarked that

if

each other, so that q

(as in algebra),

any two quaternions q and q be (as in 134) reciprocal to

= 1
be any third quaternion, then
q
(by 135), and if

we have

q"

the general formula,

22-0

.<//!.
-

>

if (by 120) we reduce q and

to a common denominator a, and denote
the new numerators by |3 and 7, we shall have (by the definitions in 106, 107),

because

137.
(as in

q"

When

fig.

two complanar triangles AOB, AOB

good, then the two unequal quotients,^

with a common

36) inversely similar (118), so that the formula

The symbol

and

AOB

are said to be

oc

side OA, are

AOB holds

CONJUGATE

for the reciprocal of a quaternion y, is also permitted in the present Calculus

it, until its legitimacy shall have been established, in connexion with a
general theory of powers of Quaternions.
[See 234.]
f Compare the Note to page 115.

but

we

q~

defer the use of

11 2

ELEMENTS OF QUATERNIONS.

124

QUATERNIONS
which

is

and

if

the first of them be

thus the conjugate of that

equal thereto,

still

or of

first,

denoted by the new symbol,

is

[II.

denoted by

any

Kg

q,

6.

then the second,

other quaternion

which

in which the letter

i.

is

K may

be said to be the Characteristic of Conjugation. Thus, with the construction

above supposed (comp. again fig. 36), we may write,

OB
- =
#;

OB

- =
OA

OA
138.

From

T;r

K?

OB
K OA
-r-r

this definition of conjugate quaternions, it follows, 1st, that if

hold good, then the line OB may be called (118) the

OA
reflexion of the line OB (and conversely, the latter line the reflexion of the
former], with respect to the line OA Ilnd, that, under the same condition, the
line OA (prolonged if necessary) bisects perpendicularly the line BB in some
the equation

OA

point A

(as represented in fig. 36);

nions (like

any two

but opposite axes

and Illrd, that any

reciprocal quaternions,

we may

so that
Z.

K#

tico

conjugate quater

comp. 134, 135) have equal angles,

write, generally,

= -

Ax K#

= Lq;

Ax

and therefore* (by 135),

zK?
139.

The

= zi;

Ax.K?=Ax.-.

reciprocal of a scalar, x, is

-, or
simply another scalar, &

or 1 ,

having

the same algebraic sign, and in all other respects related to x as in

But the conjugate K#, of a scalar x, considered as a limit of a quater
algebra.
nion, is equal to that scalar

itself",

but opposite angles, AOB and AOB

two right angles.

x, if

seen

by supposing the

tivo

equal

in fig. 36, to tend together to zero or to

We may therefore
K#

may be

as

write, generally,

x be any scalar

and conversely,!
q

= a

scalar, if

= q

K^

because then (by 104) we must have OB = OB BB = ; and therefore each of

the two (now coincident) points B, B must be situated somewhere on the
,

indefinite right

Hue OA.

* It will soon be seen that

these two last equations (138) express, that the conjugate and the
any proposed quaternion j, have always equal versors, although they have in general

reciprocal, of

unequal

tensors.

t Somewhat
that this last

is

[See 204, xiv.]

[See 157.]

and
be seen that the equation Kq
;
q may also be written as Vq =
another mode of expressing that the quaternion, q, degenerates (131) into a scalar.
later it will

CONJUGATE AND NULL QUATERNIONS.

ARTS. 137-142.]

140. In general,

sum (comp.

in the

by the construction represented

6) of the two numerators (or dividend-lines,

125

same

OB and OB

figure, the

),

of the two

and ~Kq (137), is equal to

whence (by 106), the sum of those two conjugate

conjugate fractions (or quotients, or quaternions), q

the double of the line OA

quaternions themselves

is,
-cK<?

sum

this

negative

is

if

+ q - q + i7
Kq =

therefore always scalar, being positive

if

the angle L q be acute, but

that angle be obtuse.

when the angle AOB is right, the interval

and the two lately
between the origin o and the line BB vanishes
mentioned numerators, OB, OB become two opposite vectors, of which the sum
141. In the intermediate case,

OA

is

null (5).

Now,

in general,

it is

natural, and~will be found useful, or rather

necessary (for consistency with former definitions), to admit that a null vector,

NULL QUATERNION

divided by an actual vector, gives always a

and

to denote this null quotient

by the usual symbol

as the quotient

In

for Zero.

fact,

we

have (by 106) the equation,

- a
_ a
_ a
a

a
-

"

JL

JL

\J

the zero in the numerator of the &/fc-hand fraction representing here a null
line (or a null vector, 1,
2); but the zero on the right-hnnd. side of the equation

denoting a null quotient (or quaternion). And thus we are entitled to infer
that the sum,
+ q, or q + Kq, of a right-angled quaternion, or right quotient
K<?

(132),

and

142.

of its conjugate, is

"We

always equal to

zero.

have, therefore, the three following formulae, whereof the second

exhibits a continuity in the transition

I.

..q +

from the first

Kq>0,

to the third

if
^?<^;

II.

III.

..?+K? = 0,

if

..q +

if

Kq<0,

^=|;
Lq>^.

And

because a quaternion, or geometric quotient, with an actual and finite

(as here OA), cannot become equal to zero unless its dividend- line

divisor-line

vanishes, because

by (104) the equation

*CL

= - requires the equation

a

]3

0,

ELEMENTS OP QUATERNIONS.

126
if

a be any actual and finite vector,

4-

Jq

we may

[II.

the lines OB, OB

becoming

sum

infer, conversely, that the

cannot vanish, without the line OA also vanishing

that

is,

6.

i.

without

and therefore the quaternion

opposite vectors,

We

are therefore entitled to establish the

becoming a right quotient (132).
three following converse formulae (which indeed result from the three former)

T.

IF.

Kq>0,

if

q +

Kq=Q,

then

Lq^-\
A

if

0,

then

Lq

IIF.

.if q +

~Kq<

then

Lq<\

>

/o

When

143.

two opposite vectors (1), as

common

(and actual) vector, a, we

obtained are OPPOSITE QUATERNIONS
q,

or of

any quotient

j3

may

a,

-j3

/3

are both divided

/3,

the two

- /3

= n -q =- q

one

thus

any quaternion

be denoted as follows (comp. 4)

by

quotients,

so that the opposite of

0-j3
- = _

and -

shall say that

while the quaternion q itself m^j, on the same plan, be denoted (comp. 7) by
+ q, or + q. The sum of any two opposite quaternions is zero,
the symbol

and

their quotient

(comp. again

is

negative unity

we may

so that

write, as in algebra

7),

(-q) + q = (+q) + (-q) = 0;

=-

(-q) :q

q=

(- 1) q

because, by 106 and 141,

^Vl.a^.o
a

The

-J3

g =1j = _ 1)&,
a

reciprocals of opposite quaternions are themselves opposite

or in symbols

(comp. 126),
1

1
= - -, because

- a

/3

/3

/3

Opposite quaternions have opposite axes, and supplementary angles (comp. fig.
33, fos); so that we may establish (comp. 132,, (5.)) the two following

general formulae,
L (144.

We may

also

q)

now

TT

- L
q

Ax.

(- ?)

=-

Ax

0.

write, in full consistency with the recent formulae

II. and IF. of 142, the equation,

II"...E--

if

OPPOSITE QUATERNIONS.

AHTS. 142-145.]

127

and conversely* (comp. 138),

..if

IF".

In words, the conjugate of a

is

quaternion (132),

must be a
(1) If

we

shall

is,

of a right-angled (or right)

not null) be opposite to

is

its

own

conjugate.

right quotient.

then

know

(as in 132, (1.)

On

right quotient, or

one which

we meet

the equation,

that p J_ a

that the locus of the point P

(2.)

then

the right quotient opposite thereto; and conversely, if an

actual quaternion (that

it

= -

and therefore
is

(if

and p =

a = OA,

the plane through

o,

OP, as before),

perpendicular

to

the line

OA

).

the other hand, the equation,

and therefore (by 131)

expresses (by 139) that the quotient p a is a scalar
that its angle L (p a) is either
or TT ; so that in this case, the locus of p is the
:

through the two points o and A.

the opposite of the opposite, or the reciprocal of the reciprocal, so

indefinite right line

145.

As

also the conjugate of the

conjugate, of

any quaternion,

that quaternion itself;

is

or in symbols,

so that,

by

abstracting

from

the subject

we may

of the operation,

write briefly,

It is easy also to prove, that the

conjugates of opposite quaternions are them
selves opposite quaternions ; and that the
conjugates of reciprocals are reciprocal
or in symbols, that
1.

&.(-q) = -Kq, or

aud

..K-

II.

= 1:K^,

or

* It will be seen at a later

stage, that the equation ~Kq
to this other equation,

quaternion q vanishes

Sq
:

and

that,

under

or that this quaternion

K?.K=-

q,

or q

-f

K# =

0,

may be

transformed

this last form, it expresses that the scalar

is

a right quotient (132).

[See 196, n.]

part of the

ELEMENTS OF QUATERNIONS.

128

The equation K(-q) =

(1.)

-Kq

is

this last equation

lines OB,

OB

common

scalar

any

in fig. 36, are multiplied

;

parallel to the line

(2.)

by any
by

To prove that conjugates

are reciprocal, or that

or that they are both cut

BB

6.

any

(comp. 126) may

scalar;
be proved, by simply conceiving that the two /

and

i.

included (comp. 143) iu this more

is

= xKq, where x
general formula, ~K(xq)

[II.

"Kq

K-

of reciprocals

we may

conceive that, as in the annexed figure 36,

bis,

Fig. 36,

while

we have

bis.

still

the relation

of inverse similitude,

AOB

as in the former figure 36, a

itself,

OC

new

AOB (118, 137),

point c

determined, either on the line OA

so as to satisfy either of the two

is

or on that line prolonged through A,

following connected conditions of direct similitude

A BOG OCAOB

B OC

oc

AOB

or simply, as a relation between the four points o, A, B,

A
For then we

BOG

c,

the formula,

AOB.

have the transformations,

shall

=11
K -1 = K OA
OB
q

OB

OB

OA -

OB

OG

oc

-,

^r-.

Kg

The two quotients OB OA, and OB oc, that is to say, the quaternion q
the conjugate of its reciprocal, or* the reciprocal of its conjugate, have
and
itself
the same angle, and the same axis we may therefore write, generally,
:

(3.)

z.

(4.)

Since OA

K - = L q;

OB and OA

a pair of conjugate quotients,

metric fractions,

^ and

Ax

K - = Ax

g.

OB have thus been proved (by sub-art. 2) to be

we can now infer this theorem, that any two geo
:

which have a common numerator

a,

are

conjugate

^>,

* It will be seen afterwards, that the common value of these two

equal quaternions,

K- and

q
X.q t
or in words, that it is
;
may be represented by either of the two new symbols, TJV/ Tq, or q
[See
equal to the versor divided by the tensor ; and also to the quaternion itself divided by the norm.
:

N<?

190, (3).]

GEOMETRICAL EXAMPLES.

ART. 145.]

quaternions, if the denominator

129

of the second be the reflexion of the denomina

j3

tor ]3 of the first, with respect to

that

common numerator

(comp. 138,

I.)

had only been previously assumed, as a definition (137), that such

conjugation exists, under the same geometrical condition, between the two other

whereas

it

8
fractions,

(or inverse)

be

and

f
r

the three vectors

being supposed to

a, ]3, j3

all co-initial (18).

(5.)

Conversely,

we

if

meet, in any investigation, the formula

OA

we

shall

the

line

know
OA

that the point B

OB =

is

(OA

OB),

the reflexion of the point B, with respect to

or that this line, OA, prolonged

necessary in either of two

if

opposite directions, bisects at right angles the line BB

either of the
(6.)

two figures 36 (comp. 138,

Under

some point A

in

as in

II.).

the recent conditions of construction,

elementary principles of geometry, that the

circle,

three points A, B, c, is touched at B, by the right line

it

follows from the most

which passes through the

and that this line is,

OB

a mean proportional between the lines OA, oc. Let then OD be such
a geometric mean, and let it be set off from o in the common direction of the
in length,

two

mentioned

last

the point D falls between A and c

lines, so that

vectors oc, OD be denoted

by the symbols 7,

we

also let the

shall then

have ex

pressions of the forms,

=
where a

is

some

positive scalar, a

>

(comp. sub-art. 2) with this scalar a,

OB

-=

OC

_
K

OA
OB

we

aa,

or

oc
-

OB

a?a,

and the vector

/3

of B will be connected

and with the vector

OB
= rr
OA

a* a
,

or -5-

= T^ ]3

by the formula

a,

a, this last formula expresses

suppose that 7
a
AOB
and
it expresses nothing more
BOC
the inverse similitude of triangles, A
or in other words, it is satisfied by the vector ]3 of every point B, which gives
that inverse similitude.
But for this purpose it is only requisite that the
(7.)

Conversely,

if

still

a geometric mean between the lengths of OA,

oc or that the two lines, OB, OD (sub-art. 6), should be equally long or finally,
that B should be situated somewhere on the surface of a sphere, which is described
length of

OB should be

(as above)

so as to pass through the point

its centre.

HAMILTON

ELEMENTS OF QUATERNIONS.

(in fig.

36, bis),

and

to

have the origin o for

ELEMENTS OF QUATERNIONS.

130

(8.)

in

we meet an equation

If then

which a = OA, p = OP, and a

is

[II.

6, 7.

i.

of the form,

a scalar, as before,

we

shall

know

that the

locus of the point p is a spheric surface, with its centre at the point o, and with
and also that if we determine a point c by the
the vector aa for a radius
;

= az a, this spheric locus of p is a common orthogonal to all the circles

equation oc
APC, which can be described, so as to pass through the two fixed points, A and c:
because evert/ radius OP of the sphere is a tangent, at the variable point p, to the
circle

APC, exactly as OB

is

ABC in the recent

to

figure.

fig. 36, bis, the similar triangles show (by elementary

(9.)
of BC is to that of AB in the sub-duplicate ratio
principles) that the length
or as the scalar a to 1.
If
or in the simple ratio of OD to OA
of oc to OA

In the same

we meet,
know that
then

in

any

research, the recent equation in p (sub-art. 8),

length of (p

cfa]

= a x length of (p -

a)

we

shall

while the recent interpretation of the same equation gives this other relation of
the same kind

length of p

= a

x length of a.

a subsequent stage [200 (3) ], it will be shown that the Calculus

of Quaternions supplies Rules of Transformation, by which we can pass from
any one to any other of these last equations respecting p, without (at the time)

At

(10.)

constructing

any Figure, or (immediately) appealing to Geometry: but

thought useful to point out, already, how much geometrical meaning*

tained in so simple a formula, as that of the last sub-art. 8.

The product

(11.)

NoRM,t and

is

of two conjugate quaternions

denoted thus

is

said to be their

it

is

was
con

common

qKq = Ng.
*

two equations of sub-art. 9, a sort of

APOLLONIUS of Perga, which
has been preserved through a citation made by his early commentator, Eutocius, and may be thus
enunciated Given any two points (as here A and c) in a plane, and any ratio of inequality (as here
that of 1 to a), it is possible to construct a circle in the plane (as here the circle BDB ), such that the
(lengths of the) two right lines (as here AB and CB, or AP and CP), which are inflected from the two
student of ancient geometry

may

recognise, in the

translation, into the language of vectors, of a celebrated local theorem of

common point (as B or p) of the circumference, shall be to each other in the given
(Auo SoOevTW ffrjimeicav, K. r. \.
Page 11 of Halley s Edition of Apollonius, Oxford, MDCCX.)
f This name, NOUM, and the corresponding characteristic, N, are here adopted, as suggestions
from the Theory of Numbers ; but, in the present work, they will not be often wanted, although it
may occasionally be convenient to employ them. For we shall soon introduce [in 187] the conception,
given points to any

ratio.

BADIAL QUOTIENTS, EIGHT RADIALS,

ARTS. 145, 146.]

N^ and

norm

131

It follows that

NKg

positive scalar

namely, the square of the quotient of the lengths of the two lines

that the

of which (as vectors) the quaternion itself

have,

by

sub-art. 6,

NT

the quotient (112).

In fact
the definition of a norm, the transformations

and by

OB

- =

OA

As

limit,

__ OB

N-OA

we may say that

is

we

OB

the

OB

oc

oc OB

OB OA

OA

N0

of a quaternion is generally a

oc

/OD

OA

OA

\length of aj

norm of a null quaternion

is

zero

or in symbols,

0.

(12.)

With

this notation, the equation of the sphericlocus (sub-art. 8),

has the point o for

the shorter form

its

and the vector aa

centre,

which

for one of its radii, assumes

aa

SECTION

On Radial

Quotients

and on

7.

tlie

Square of a Quaternion.

146. It was early seen (comp. Art. 2, and fig. 4) that any two radii, AB,
AC, of any one circle, or sphere, are necessarily unequal vectors ; because their
directions differ.
direction (110),
co-initial (18),

On

the other hand,

we may suppose

when we

are attending only to relative

that all the vectors compared are not
merely

but are also equally long

for the unit, they are all radii, OA, OB,

so that if their

of

common

length be taken

what we

have called the Unit-Sphere (128), described round

the origin as centre and may all be said to be Unit;

And

then the quaternion, which is

the quotient of any one such vector divided by any
Vectors (129).

other, or generally the quotient of any two equally long

vectors,

may

be called a Radial Quotient

(Compare the annexed figure

and the

or sometimes simply a KADIAL.

39.)

characteristic, of the Tensor, T^, of a quaternion, which is of greater geometrical utility than
it will he proved that this norm is simply the square,

the Norm, but of which

*K*-Nf-(T*).
Compare the Note

to sub-art. 3.

S 2

ELEMENTS OF QUATERNIONS.

132

The two

147.

[II.

and Negative Unity,

Unit- Scalar8, namely, Positive

7.

i.

may

be

considered as limiting cases of radial quotients, corresponding to the two extreme

and IT, of the angle AOB, or L q (131). In the inter
values,

when AOB

mediate case,

is

a right angle, or L q = -, as in
<i

or quaternion, may be called

fig. 40, the resulting quotient,
or simply, a EIGHT
Radial
a
Quotient
Right
(comp. 132)
;

The

HADIAL.

consideration of such

found to be of

right radiah will be o

importance, in the whole theory

greai>

and

Flg>

40

practice of Quaternions.

148.
is

The most important

that the Square of every Right Radial is equal to Negative

that we write generally, as in algebra,
understood
being

the following

Unity

and

it

general property of the quotients last mentioned

call this

product of two equal quaternions the SQUARE of each of them.

bis.

Fig. 41,

For

if,

as in

OB for the

41,

fig.

we

with o for

then the two right quotients, OB

bisecting radius t

are equal (comp. 117)

ABA

describe a semicircle

and therefore

their

common square

centre,

OA,

is

and with

and OA

OB,

(comp. 107) the

product,
J

where OA and OB
lines.

More

may

OB

Y=~

OA/

OA

OB

represent any

tivo

OA

OA

equally long, but mutually rectangular

generally, the Square of ever y Right Quotient (132)

Negative Scalar

namely,

to

the negative of the square

of the

represents the ratio of the lengths* of the two rectangular lines

equal

to

number, which

compared or
number which denotes (comp. 133) the length of
Index of that Bight Quotient
as appears from fig. 41, bis, in which OB is

to zero minus the square of the

the

is

Hence, by 145

(11.),

=-

if

L q

-.

SQUARE ROOTS OE NEGATIVE OTITY.

ABTS. 147-149.]

133

only an ordinate, and not (as before) a radius, of the semicircle ABA
have thus,
OB V .
/OB\ 8 OA
_
= _ (length of _
if OB J- OA.
OA
\length O/OA/
\OA/

for

we

"L

Thus

149.

every Eight Radial

Symbol

</-

in the present System, one of the Square

is,

and may therefore be said to be one of the Values of

which celebrated symbol has thus a certain degree of

Roots of Negative Unity

the

in this theory, on account

vagueness, or at least of indetermination, of meaning
For
of which we shall not often employ it.
although it thus admits of a

and geometrically real Interpretation, as denoting what has been

above called a Right Radial Quotient, yet the Plane of that Quotient is arbitrary;
and therefore the symbol itself must be considered to have (in the present
= or in other words the
values

perfectly clear

system) indefinitely many

has (in the Calculus of Quaternions) indefinitely

Equation,

Geometrical Reals

besides

which the same equation

other roots, of a purely symbolical character,

may

be conceived to possess, and which may be

Conversely, if q be any real quaternion,

Geometrical Imaginaries.-f

which

satisfies

42,

=-l,

<?

must

it

BOG,

we

/OB\*

oc

OB

oc

\OA/

OB OA

OA

that

and

the equation

we suppose

all

any

called

fig.

1,

(f

many Roots* which are

AOB

be a right radial

shall

for

if,

as in

have

cannot become equal to negative unity,

= - OA, or = OA in fig. 41 that is, by

this square of q

except by oc being
the lins OB being at right angles to the line OA, and being
at the same time equally long, as in fig. 40.
;

(1.)

If then

we meet

the equation,

Fig. 42.

where a = OA, and p = OP,

* It will be
subsequently
is unity, so that

we

as before,

shown

shall

[in 222], that if

a?,

know

y, z

that the locus of the point

be any three

scalars,

of which the sum of

the squares

and

if

i,

j,

k be any three right radials,

in three

mutually rectangular planes

ix

+jy

4-

then the expression,

kz,

denotes another right radial, which satisfies (as such, and by symbolical laws to be assigned) the
1
and is therefore one of the geometrically real values of the symbol V - 1.
equation q~ =
t Such imaginaries will be found to offer themselves, in the treatment by Quaternions (or rather
by what will be called Siquaternions}, of ideal intersections, and of ideal contacts, {^geometry [see214]
but we confine our attention, for the
to geometrical reals alone.
Compare the Notes to pages
;

present,

87 and 88,

ELEMENTS OF QUATERNIONS.

134

is

the circumference of a

with o for

circle,

has the same length as the line OA

its centre,

while the plane of the circle

7, 8.

i.

and with a radius which

In other words, the locus of p is a great

and the given point
of which the centre is the origin
to that given line.

is

perpendicular

circle,

A,

on a sphere

on the same

one of the poles of that circle.

is

spheric surface,

[II.

In general, the equation q 2 = - a*, where a is any (real) scalar, requires

that the quaternion q (if real) should be some right quotient (132)
the
(2.)

number

a denoting the

length of the index (133), of that right quotient or

quaternion (comp. Art. 148, and fig. 41, bis).

entirely arbitrary and therefore the equation

But the plane

of q is

still

- - a\

<f

like the equation q

= -

1,

which

the present system) indefinitely

(3.) Hence the equation,

which we

it

includes,

many

must be considered

to

have

(in

geometrically real roots.

suppose that
expresses that the locus of the point P is a
(new) circular circumference, with the line OA for its axis* and with a radius of
which the length = a x the length of OA.
150. It may be added that the index (133), and the axis (128), of a right
in

may

a>0,

radial (147), are the

and

its opposite

and that

same",

its reciprocal

(143), are all equal to each other.

of a given quaternion q be equal

to

right radial] because its square,

itself, divided by

its

opposite

(134), its conjugate (137),

Conversely, if the reciprocal

then q is a

the opposite of that quaternion,

q*, is

then equal (comp. 136) to the quaternion

and therefore (by 143)

to negative unity.

But

the conjugate of every radial quotient is equal to the reciprocal of that quotient;
because if, in fig. 36 [p. 115], we conceive that the three lines OA, OB, OB
are equally long, or

if,

in fig. 39,

we prolong

the arc BA,

by an equal

arc

AB we
,

have the equation,

And

OA

OA

OB

if

qKq =

conversely, f
if

then the quaternion q

*

OB

is

Kq

-,

or

1,

a radial quotient.

It being understood, that the axis of a circle is a right line perpendicular to the plane of that

and passing through its centre.

t Hence, in the notation of norms (Ho, (11.)), if N? =
norm of a radial quotient, is always equal to positive unity.
circle,

1,

then q

is

a radial

and conversely, the

RADIAL QUOTIENTS CONSIDERED AS VERSORS.

AKTS. 149-153.]

SECTION

On

135

8.

the Tersor of a ffcuaternion, or of a Vector; and on

some General Formulae of Transformation.

a quaternion = |3: a is thus a radial quotient (146), or when

the lengths of the two lines a and ]3 are equal, the effect of this quaternion q,
considered as a FACTOR (103), in the equation qa = |3, is simply the turning of
151.

When

<7

plane of q (119), and towards the hand determined

by the direction of the positive axis Ax q (129), through the angle denoted by
Lq (130) so as to bring that line a (or a revolving line which had coincided
the multiplicand- line

a, in the

namely, into that of the product-Hue /3. And

therewith) into a new direction
with reference to this conceived operation of turning , we shall now say that
:

every Radial Quotient

152.

is

a VERSOR.

Versor has thus,

a plane

in general,

an

and an angle

axis,

equal the only

namely, those of the Radial (146) to which it corresponds, or
difference between them being a difference in the points of mew* from which
is

they are respectively regarded namely, the radial as the quotient, q, in the
formula, q = j3 a and the versor as the (equal) factor, q, in the converse
;

formula,

j3

= q a
.

where

it

is still

supposed that the two vectors, a and

j3,

are equally long.

153.

by

its

A versor, like a radial

(147), cannot degenerate into a scalar, except

angle acquiring one or other of the

first case, it

becomes

positive unity

and

two

and

limit-values,

in the second case, it

TT.

In the

becomes negative

each of these two unit-scalars (147) being here regarded as a factor (or
coefficient, coinp. 12), which operates on a line, to preserve or to reverse its

unity

direction.

In

this view,

we may say

that -

Right Versor ( or versor with an angle = ^

inverts the line

on which

it operates,

is
)

1 is

an Inversor

a Semi-inversor

or turns

it

:*

and that
because

it

every
half-

through half of two right angles

* In a
slightly metaphysical

mode of expression it may be said, that the radial quotient is the

an analysis, wherein two radii of one sphere (or circle) are compared, as regards their relative
direction
and that the equal versor is the instrument of a corresponding synthesis, wherein one ratlins
is conceived to be generated,
by a certain rotation, from the other.
result of

semi-in versor," will not be often used but the introduction of it here, in passing,
t This word,
seems adapted to throw light on the view taken, in the present work, of the symbol^- 1, when
regarded as denoting a certain important class (149) of Reals in Geometry. There are uses of that
symbol, to denote Geometrical Imaginaries (comp. again Art. 149, and the Notes to pages 87 and 88),
1
considered as connected with ideal intersections, and with ideal contacts but with such uses of V
"

we

have, at present, nothing

to do.

ELEMENTS OF QUATERNIONS.

136

(comp.

For the same

fig. 41).

(like every right radial, 149,

that

it

reason,

we

roots of negative unity

from which indeed we have just seen, in 152,

from quotient), as being one of the square

observe that the

effect

own

hand, through a right angle.

If then q be such a versor,

have also (comp.

= - a

fig. 41), qfi

we have

q,

plane), is to turn that

so that, if a be

- a
qa =

line,

and

any

if

towards a given

qa =

/3,

we

shall

line in the plane of a

natural to write, under the same condition,

- <f

On

value of y/ - 1
illustrate.

1.

the equation,
q

as in 149.

of a right versor, considered

as operating on a line (in its

whence

8.

are led to consider every

right versor

or as one of the values of the symbol

we may

154. In fact

it is

i.

only as factor differs

differs

right versor

[II.

For

1,

the other hand, no versor, which

or can satisfy the equation (fa =
it is

is

not right-angled, can be a

a, as fig.

42

may

serve to

included in the meaning of this last equation, as applied

2^q, or through the double of

to the theory of versors, that a rotation through

the angle of q itself, is equivalent to

an

inversion of direction

and therefore

to

a rotation through two right angles.

155. In general, if a be any vector, and if a be used as a temporary*

symbol for the number expressing its length so that a is here a positive scalar,
;

which bears

to positive unity, or to the scalar

1,

the same ratio as that which

the length of the line a bears to the assumed unit of length (comp. 128) then
the quotient a a denotes generally (comp. 16) a new vector, which has the
same direction as the proposed vector o, but has its length equal to that assumed
;

unit

so that it is (comp. 146) the Unit- Vector in the direction of a.

denote this unit-vector

by the symbol,

Ua
that

is,

more

fully, if a be, as

incommensurable, but positive)

some

Ua

if

-,

and

We

shall

so shall write, generally,

a = length of a

above supposed, the number (commensurable or

which represents that length, with reference to

selected standard.

156. Suppose now that q = j3 a is (as at first) a general quaternion, or the

Such
quotient of any two vectors, a and ]3, whether equal or unequal in length.
:

a Quaternion will not (generally) be a Versor (or at least not simply such),

We

*
shall soon propose [in 185] a general notation for representing the lengths of vectors, accord
ing to which the symbol Ta will denote what has been above called a but are unwilling to introduce
more than one new characteristic of operation, such as K, or T, or U, &c., at one time.
;

ARTS. 153-157.]

VERSOR OF A QUATERNION, OR OF A VECTOR.

137

according to the definition lately given because its effect, when operating as
a factor (103) on a, will not in general be simply to turn that line (151) but
will (generally) alter the length* as well as the direction.
But if we reduce
;

the two proposed vectors, a and j3, to the two unit-vectors Ua and U/3 (155),
and form the quotient of these, we shall then have taken account of relative
direction alone

defined (151).

and the

result will therefore

We propose

be a versor in the sense lately

,

to call the quotient, or the versor, thus obtained,

the versor-element, or briefly, the VERSOR, of the Quaternion q

and

shall find

convenient to employ the same~f Characteristic, U, to denote the operation of

talcing the versor of a quaternion, as that employed above to denote the opera

it

tion (155) of reducing a vector to the unit of length, without

On

direction.

any change of its

and the

this plan, the symbol ~Uq will denote the versor of q

foregoing definitions will enable us to establish the General

- ==

Ua

Formula

which the tiro unit-vectors, Ua and U/3, may be called, by analogy, and for
other reasons which will afterwards appear, the versor s$ of the vectors, a and /3.
157. In thus passing from a given quaternion, q, to its versor, Ug, we have
in

only changed

(in general)

the lengths of the two lines compared, namely,

by

reducing each to the assumed unit of length (155, 156), without making any
change in their directions. Hence the plane (119), the axis (127, 128), and
the angle (130), of the quaternion, remain unaltered in this
so that

passage

we may

establish the

two following general formulae

= Ax
L Uq = L q
Ax
.

U<?

q.

By what we shall soon call an act of tension, which will lead us to the consideration of the tensor
of a quaternion.
t For the moment, this double use of the characteristic U, to assist in denoting both the unitvector Ua derived from a given line o, and also the versor U^ derived from a
quaternion q, may he
but as permitted, because the difference of the
regarded as established here by arbitrary definition
;

symbols, as here a and q, which serve for the present to denote vectors and quaternions, considered as
the subjects of these two operations U, will prevent such double use of that characteristic from
giving
rise to any confusion.
But we shall further find that several important analogies are by anticipation

Thus it will be found (comp.

expressed, or at least suggested, when the proposed notationis employed.
the Note to page 121), that every vector a may usefully be equated to that right quotient, of which it
is (133) the index; and that then the unit-vector Ua
may be, on the same plan, equated to that right

We

radial (147), which is (in the sense lately defined) the versor of that right quotient.
shall also find
ourselves led to regard every unit-rector as the axis of a quadrantal (or right] rotation, in a plane
which will supply another inducement, to speak of every such vector as a
perpendicular to that axis
;

On

the whole, it appeai-s that there will be no inconvenience, but rather a prospective
advantage, in our already reading the symbol Ua as "versor of a.
just as we may read the analogous
versor.

";

symbol
J

U<?,

as

"versor

of

q."

[Compare 286 and 290.]

Compare the Note immediately preceding.

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

138

More

generally

we may
L q = L

[II.

8.

i.

write,
q,

Ax / = Ax

and

q, if

U/ =

"Uq

the versor of a quaternion depending solely on, but conversely being sufficient to
determine, the relative direction (156) of the two lines, of which (as vectors) the
or the axis and angle of the rotation,
quaternion itself is the quotient (112)
so
in the plane of those two lines, from the divisor to the dividend (128)
;

that

any two quaternions, which have equal rersors, must

and equal

(or coincident)

also

have equal angles,

expressed by the last written formula.

axes, as is

Conversely, from this dependence of the versor IJq on relative direction* alone,
it follows that any two quaternions, of which the angles and the axes are
equal, have also equal versor s

Uy =

U^,

if

or in symbols, that

L q = L

q,

and

Ax

Ax

q.

example, we saw (in 138) that the conjugate and the reciprocal of any
quaternion have thus their angles and their axes the same it follows, there

JtTor

fore, that the versor of the conjugate is

cal

so that

we

always equal to the versor of

the recipro

are permitted to establish the following general formula, t

158. Again, because

/3

it

follows that the versor of the reciprocal of

time, the reciprocal of the versor

so that

11-=^;
q
IT?
Hence, by the recent result (157),

-;

or

any quaternion
we may write,

is,

at the

same

U^.ui = l.
q

we have
or,

also, generally,

* The

unit-vector Ua, which we have recently proposed (156) to call the versor of the vector a.
depends in like manner on the direction of that vector alone which exclusive reference, in each of these
two cases, to DIRECTION, may serve as an additional motive for employing, as we have lately done,
one common name, VEKSOK, and one common characteristic, U, to assist in describing or denoting both
all danger of
the Unit- Vector Ua itself, and the Quotient of two such Unit- Vectors, Uq = U/3 Ua
confusion being sufficiently guarded against (comp. the Note to Art. 156), by the difference of the two
,

symbols, a and q, employed to denote the vector and the quaternion, which are respectively the subjects
while those two operations agree in this essential point, that each serves to
;
of the two operations
eliminate the quantitative element, of absolute or relative length.

f Compare the Note to Art. 138.

ARTS. 157-159.]

EQUAL AND RECIPROCAL VERSORS, REVERSORS.

139

Also, because the versor IJq

is always a radial quotient (151, 152), it is (by

its
own
the
reciprocal and therefore, at the same time (comp.
conjugate of
150)
its
own
the
reciprocal of
conjugate-, so that the product of two conjugate
145),
,

versors, or

what we have

equal to positive unity

called (145, (11.)

common NORM,

their

is

always

or in symbols (comp. 150),

NU?

= TJ? KIT? =
.

1.

For the same reason, the conjugate of the versor of any quaternion is equal to
the reciprocal of that versor, or (by what has just been seen) to the versor of the
reciprocal of that quaternion; and therefore also (by 157), to the versor of the
conjugate

so that

the formula

we may

write generally, as a

of recent results,

summary

each of these four symbols denoting a new versor, which has the same plane,
angle, as the old or given versor ~Uq, but has an opposite axis, or

and the same

an opposite direction of rotation so that, with respect to that given Versor,

may naturally be called a KEVERSOR.
:

it

As

159.

regards the versor itself, whether of a vector or of a quaternion,

the definition (155) of TJa gives,

U#a

= + Ua,

or

= -

Ua

according as

>

or

<

because (by 15) the scalar coefficient x preserves, in the first case, but reverses,
in the second case, the direction of the vector a whence also,
by the definition
;

(156) of

"Uq,

we have generally (comp. 126, 143),

"Uxq

= +

"Uq,

or

= - Uy,

according as

>

or

0.

<

The

versor of a scalar, regarded as the limit of a quaternion

(131, 139), is equal
to positive or negative unity (comp. 147,
as
the scalar itself is
153), according

positive or negative

TJ# = +
the plane

and

or in symbols,
1,

or

= -

axis of each of these

1,

according as

two unit

(153), being (as we have already

quaternion (141) must be regarded

>

or

<

scalars (147), considered as versors

seen) indeterminate.

The

as ivholly arbitrary, unless

versor of a null

we happen

know

to

a laiv* according to which the quaternion tends to zero, before actually

in which latter case, the plane, the axis, and the angle of
reaching that limit
;

Compare the Note

to Art. 131.

ELEMENTS OF QUATERNIONS.

140

the versor*

The

UO may

all

become determined,

[II.

i.

8.

deduced from that law.

as limits

vcrsor of a right quotient (132), or of a right-angled quaternion (141), is

always a right radial (147), or a right versor (153)

and therefore

is,

as such,

one of the square roots of negative unity (149), or one of the values of the symbol
1

^/

in like

while (by 150) the axis and the index of such a versor coincide

manner

its reciprocal, its conjugate,

and

its opposite

and

are all equal to each

other.

160. It

is

evident that

if

a proposed quaternion q be already a versor (151),

in the sense of being a radial (146), the operation of taking

produces no change
unit-vector, it

length

that

and

that, if

remains the same vector, when

is,

axis of every

its

versor (156)

a given vector a be already an

it

divided (155)

is

by

own

its

For example, we have

by the number one.

in this case,

assumed (128, 129), that the

may

manner

in like

quaternion

is

an

unit-vector",

we

therefore write, generally, in the notation of 155, the equation,

U (Ax

.
</)

Ax

A second operation U leaves thus the result of

whether the

q.

the first operation

unchanged,
or
a
line,
quaternion we

subject of such successive operations be a

have therefore the two following general formulae, differing only in the
symbols of that subject

whence, by abstracting (comp. 145) from the subject of the operation, we

write, briefly

and symbolically,

UU = U.

IP =

161. Hence, with the help of 145, 158, 159,

we

easily deduce the follow

ing (among other) transformations of the versor of a quaternion

--..I

U?

We may also

r
KUq

= TJxq,

if

TJKy

may

>

= - Us?,

if

<

0.

write, generally,

* When the zero in this

symbol, UO, is considered as denoting a null vector (2), the symbol itself
denotes generally, by the foregoing principles, an indeterminate unit-vector although the direction of
this unit-vector may, in certain questions, become determined, as a limit resulting from a law.
;

GENERAL TRANSFORMATIONS OF A VERSOR.

AKTS. 159-161.]

the parentheses being here unnecessary, because (as will soon be

seen) the symbol U*/

141

more

denotes one common versor, whether we interpret

fully
it

as

denoting the square of the versor, or as the versor of the square, of q. The
present Calculus will be found to abound in General Transformations of this

which

sort;

all (or

nearly

but which, notwithstanding (or rather, perhaps,

simple geometrical conceptions

on account

extreme simplicity of their

this

of)

elements of a

the foregoing, depend ultimately on very

all), like

new kind

of Symbolical

are often useful, as

origin.,

Language in Geometry

and generally, as

instruments of expression, in all those mathematical or physical researches to

however, by no
means necessary that a student of the subject, at the present stage, should
make himself familiar with all the recent transformations of Uq although it

which the Calculus of Quaternions can be applied.

It

is,

be well that he should satisfy himself of their correctness, in doing which

the following remarks will perhaps be found to assist.

may

(1.) To give a geometrical illustration, which

the recent equation,

we may employ

36, bis [p. 128]

fig.

As

(2.)

Kq

points,

(2.),

we have

fig.
,

OB~

regards the equation,

new

by 145,

OA _ OB _ /OB
~
~

OB
_
~
o7

the three lines OA, OB, oc, of

in three

in which,

also serve as a proof, of

may

U(<f)

= (U^) 2 we have only

,

42, are cut (as in

by an

unit-circle (or

fig.

42,

to conceive that

bis)

by a circle

with a radius equal to the unit of length), which is described

about their common origin o as centre, and in their common
plane

new

for then

if

these three lines be called

a, /3, 7,

the three

oc are (by 155) the three unit-vectors de

noted by the symbols, TJa, Uj3, U7 and we have the trans
formations (comp. 148, 149),
lines

OA

OB

(3.)

As

regards other recent transformations (161), although

we have seen

(135) that it is not necessary to invent any new or peculiar symbol, to represent
the reciprocal of a quaternion, yet if, for the sake of present convenience, and
as a merely temporary notation, we write

B,-i,
employing thus, for a moment, the

letter

as a characteristic of reciprocation,

ELEMENTS OF QUATERNIONS.

142

or of the operation of taking the reciprocal,

equations (comp. 145, 158)

we

[II.

shall then

i.

8,9.

have the symbolical

EK = KB
BU = UK = KU = UK;
we have also (by 160), IP = U whence
easily follows that
U = BUB = BKU = BUK = KUB = KBU = KUK
= UBK = UKB = UKUB = UKBU =
(UK) = &o.
B

but

it

The equation

(4.)

U=Ij

or simply,

expresses that the locus of the point p

132,

which

(4.)),

direction

because

drawn from o

is

it is

On

the other

or

p =

or

-u,
a

>

0.

or

expresses (comp. 132, (5.)) that the locus of p

is

the opposite ray from o

the indefinite prolongation of the revector BO

it is

0/3,

hand the equation,

TTe =

that

line, or ray (comp.

not
in the opposite
but
O/OB,

equivalent to

U^ = 1
(5.)

the indefinite right

is

in the direction

because

it

may

or

be trans

formed to

U
(6.)

42

If

a, |3,

= - 1

7 denote

or

|j

TT

|j

(as in sub-art. 2)

(or of fig. 42, bis), so that

(by 149)

or

p =

#j3,

<

0.

the three lines OA, OB, oc of

we have

the equation

=
a

),

fig.

then

aj

this other equation,

expresses generally that the locus of p is the system of the two last loci ; or
that it is the whole indefinite right line, both ways prolonged, through the two

points o and B (comp. 144,

(7.)

41,

bis),

But

(2.)).

happen that the line 7, or oc, like OA in fig. 41 (or in fig.

has the direction opposite to that of a, or of OA, so that the last
if it

equation takes the particular form,

RE PRESENT ATIYE AND YECTOB, ARCS.

AETS. 161-163.]

then

U -a must be

(by 154) a right versor

with a plane containing

In

this case, therefore,

(1.))

the plane through

of these

o,

OA; and the recent equation

by a real* vector p, may be put under either

perpendicular

to the line

two

7T

^-=H;
&
a

pJ-a.

SECTION

On

reciprocally, every right versor,

a, will be (by 153) a value satisfying the equation.

the locus of the point p is (as in 132, (1.), or in 144,

supposed to be satisfied
earlier but equivalent forms

itself, if

and

143

9.

Vector-Arcs, and Vector-Angles, considered as Representatives

and on the Multiplication and
of Versors of Quaternions
Hi vision of any one such Versor by another.
;

162. Since every unit-vector OA (129),

drawn from the

origin o, terminates

some point A on the surface of what we have called the unit-sphere (128),
that term A (1) may be considered as a Representative Point, of which the
position on that surface determines, and may be said to represent, the direction
in

of the

And

line

OA in space

or of that line multiplied (12, 17)

then the Quaternion which

and which

vectors,

Versor (151),

is

be said to have the arc of a great

may

sphere, which connects

We may

the terms of the

also call this arc a

direction (comp. Art. 1),

in

fig.

39

[p.

what may be

131]

called

by any

positive scalar.

the quotient (112) of any two such unita

in one view
Radial (146), and in another view a
is

and

two

VECTOR ARC, on account

such as

is

circle,

AB,

upon the unit

vectors, for its Representative Arc.

of its

having a

definite

by a curved arrow
own opposite, or with

indicated (for example)

as being thus contrasted with its

by analogy the Revector Arc BA (comp. again

1)

this

latter arc representing, on the present plan, at once the reciprocal (134), and
the conjugate (137), of the former versor ; because it represents the corre
sponding Reversor (158).

163. This

mode

of representation, of versors of quaternions

would obviously be very imperfect, unless

equals.

We shall therefore define, as

arc, AB, upon the unit sphere,

*

Compare

149,

(2.)

is

also the second

it is

equals

by

vector arcs,

were to be represented by

otherwise natural to do, that a vector

equal to every other vector arc CD which can be

Note

to the

same Article

and the Notes to pages 87 and 88.

ELEMENTS OF QUATERNIONS.

144

derived from

it,

(.II.

i.

own great
In fact, the two

(or conceiving) it to slide* in its

by simply causing

reversal of direction.
circle, without any change of length, or

and plane triangles AOB, COD, which have the origin o for their
common vector, and rest upon the chords of these two arcs as bases, are thus
complanar, similar, and similarly turned so that (by 117, 118) we may here

isosceles

write,

A AOB

OC

COD,

OB

OD

OA

OC

the condition of the equality of the quotients (that

is,

here, of the versors),

We

shall
represented by the two arcs, being thus satisfied.
sometimes denote this sort of equality of two vector arcs, AB

and

by the formula,

CD,

n AB = n CD

and then

it is

clear (comp. 125,

and the

earlier Art. 3) that

we shall also have, by what may be called

alternation, these two other formulae of arcual
n BA = n DC

(Compare the annexed figure 35,

inversion

Fig. 35,

bit.

equality,

AC

and

BD.

bis.)

164. Conversely, unequal versors ought to be represented (on the present

plan)

by unequal

vector arcs

and accordingly, we purpose

any two

to regard

such arcs, as being, for the present purpose, unequal (comp. 2), even

when they

agree in quantity, or contain the same number of degrees, provided that they
differ

in direction

which

may happen in
may be opposite

either of

two principal ways,

as

arcs of one great circle

For, 1st, they
as, for
BA
and
the
arc
so
and
revector
a
arc
vector
AB,
may
corresponding
example,
OA, and the corre
represent (162) a versor, OB

follows.

sponding
the

tivo

reversor,

OA

AB and BC

in

fig.

43

represent two radial

planes

Or, Ilnd,

OB, respectively.

may belong

arcs

to different great circles, like

in

which

quotients

latter case,

(146)

in

they A

different

or (comp. 119) two diplanar versors, OB

and oc OB
:

but

it

OA,

has been shown generally (122),

that diplanar quaternions are always unequal

consider therefore, here again the arcs,

been said) unequal

AB and

we
BC, themselves, to

be

(as

has

vectors.

* Some aid to the

conception may here be derived from the inspection of fig. 34 [p. 113] in which
two equal angles are supposed to be traced on the surface of one common desk. Or the four lines OA,
or to be cut by a circle with o for
OB, oc, OD, of fig. 35, may now be conceived to be equally long
centre as in the modification of that figure, which is given in Article 163, a little lower down.
;

ARCITAL EQUATIONS, CO-ARCUALITY.

ABTS. 163-166.]

we may be

165. In this manner, then,

145

led (cornp. 122) to regard the

conception of a plane, or of the position of a great

on the unit sphere, as

circle

entering, essentially, in general,* into the conception of a vector-arc, considered

as the representative of a versor (162).
versors,

we may

see that

if,

in

But even without expressly referring to

we suppose that B is the middle point
that in a recent notation (163) we may

43,

fig.

of an arc A A of a great circle, so

establish the arcual equation,

n AB = n BA

we ought then (comp. 105) not

to write also,

n AB = n BC
because the two co-initial

arcs,

BA and

which terminate

BC,

differently,

must be

considered (comp. 2) to be, as vector-arcs, unequal.

On the other hand, if we
should refuse to admit (as in 163) that any two complanar arcs, if equally long,

and similarly (not

are equal vectors,

existing under

AB and CD in the recent

oppositely) directed, like

we could not

usefully speak of equality

We

any circumstances.

between

fig.

35, Us,

vector-arcs as

are then thus led again to include,

generally, the conception of a plane, or of one great circle as distinguished from

And hence an
another, as an element in the conception of a Vector- Arc.
equation
relations

between two such arcs must in general be conceived to include tivo

of co-arcuality. For example, the equation n AB = n CD, of Art. 163,

includes generally, as & part of its signification, the assertion (comp. 123) that
the four points A, B, c, D belong to one common great circle of the
unit-sphere;
or that each of the two points, c and D, is co-arcual with the two other
points,

A and

B.

166. There
arcs

may

is, however, a remarkable case of exception, in which two vector

be said to be equal, although situated in different planes namely,
:

when they

are both great semicircles.

great semicircle,
or

it

AA considered
,

In

fact,

upon

the present plan, every

as a vector arc, represents

represents negative unity (OA

OA = - a

a = -

1),

an inversor (153)

considered as one limit

but we have seen (159) that such a versor has in general an inde
terminate plane.
Accordingly, whereas the initial and final points, or (comp. 1)

of a versor

A and the term

the origin

B, of

a vector arc AB, are in general sufficient to

determine the plane of that arc, considered as the shortest or the most direct
path (comp. 112, 128) from the one point to the other on the sphere; in the
particular case when one of the two given points is diametrically opposite to
*

We

say, in general
considered as vector arcs,

HAMILTON

for it will soon be seen that there is a sense in

may

be said

to

ELEMENTS OF QUATERNIONS.

which

all great semicircles,

be equal to each other.

ELEMENTS OF QUATERNIONS.

146

[II,i.9.

the other, as A to A, the direction of this path becomes, on the contrary, inde
terminate.
If then we only attend to the effect produced, in the way of change

of position of a point, by a conceived vection (or motion] upon the sphere, we are
permitted to say that all great semicircles are equal vector arcs each serving
;

from one position to the

the factor
1, of which it is here the

simply, in the present view, to transport a point

and thereby

opposite-,

to reverse (like

representative] the direction of the radius which

is

drawn

to that point of the

unit sphere.
(1.)

The

in which

n AA = n BB

equation,

it

here supposed that A

is

is

opposite to A, and B to B, satisfies

because the four points

evidently the general conditions of co-arcuality (165)
ABA B are all on one great circle. It is evident that the same arcual equation
;

admits

163) of inversion and alternation

(as in

nAA = n B

We

(2.)

no

effect

may

so that

and n AB = n A B

B,

also say (comp. 2) that all null arcs are equal, as

on the position of a point upon the sphere; and thus

generally,

producing

may

write

n AA = n BB = 0,

with the alternate equation, or identity, r\ AB = n AB.

a representative, on the present plan,
(3.) Every such null vector arc AA is
of the other unit scalar, namely positive unity, considered as another limit of a
versor (153)

and

given, according to

given point
167.

A, to

The

law be
again indeterminate (159), unless some
which the arcual vection may be conceived to begin, from a

its

plane

is

an indefinitely near point B upon the sphere.

principal use of Vector Arcs, in the present theory,

in representing,

and

(so to speak)

in constructing,

by means

of a

is

to assist

Spherical

Triangle, the Multiplication and Division of any two Diplanar Versors (comp.
In fact, any two such versors of quaternions (156), considered as
119, 164).
radial quotients (152), can easily be reduced (by the general process of Art.

120) to the forms,

q

where

]3

a = OB

A, B, c are corners of

OA,

j3

= oc

OB,

such a triangle on the unit sphere

107), the former quotient multiplied

q

If then (on the plan of Art. 1)

= 7

by
:

a = oc

any two

and then (by

the latter will give for product

:

OA.

AB and BC in

successive arcs, as

be called (in relation to each other) vector

and projector

fig.

43,

while that third arc

MULTIPLICATION OF VERSOBS.

ARTS. 166-168.]

which

AC,

drawn from the

is

point of the

initial

first to

147

the final point of the

second, shall be called (on the same plan) the transactor we may now say
that in the multiplication of any one versor (of a quaternion) by any other, if
the multiplicand* q be represented (162) by a vector-arc AB, and if the multiplier
:

manner represented by a provector-arc BC, which mode of repre

always possible, by what has been already shown, then the product
q, is represented, at the same time, by the transvector-arc AC corre

q be in like

sentation
q

q,

is

or q

sponding.
168. One of the most remarkable consequences of this construction of the
multiplication of versors is the following: that the value of the product of two
diplanar versors (164) depends upon the order of the factors ; or that q q and qq
are unequal, unless q be complanar (119) with q.
For let AA and cc be any
two arcs of great circles, in different planes , bisecting each other in the point B,
as fig. 43 is
designed to suggest ; so that we have the two arcual equations
(163),

^ AB = n BA

and

BC = n C B

r\

then one or other of the two following alternatives will hold

good. Either,
1st, the two mutually bisecting arcs will both be semicircles, in which case the
two new arcs, AC and C A will indeed both
belong to one great circle,

namely

to that of

which B

A and

in this case,

(by 166,

is

(1.)

&pole, but will have opposite directions therein; because,

c will be diametrically opposite to A

and

c,

and therefore

the equation
r\

AC =

r>

but not the equation

n AC = n C A
will be satisfied.

each other in

Or, Ilnd, the arcs

B, will not both

AA and

cc

which are supposed to

be semicircles, even

if

one of

such; and in this case, the arcs AC, C A will belong to two distinct great
so that they will

bisect

them happen

to

be

circles,

be diplanar, and therefore unequal, when considered as

(Compare the 1st and Ilnd cases of Art. 164.) In each case, therefore,
AC and C A are unequal vector arcs but the
former has been seen (167) to

vectors.

represent the product q q;

and the

latter represents, in like

manner, the other

product, qq of the same two versors taken in the opposite order, because
the new transvector arc, when C B
(= BC) is treated as the new vector arc,
,

BA

(= AB) as the

new projector

arc,

as is indicated

it is

and

by the curved arrows in

Here, as in 107, and elsewhere, we write the symbol of the multiplier towards the left-hand, and
that of the multiplicand towards the
right.

ELEMENTS OF QUATERNIONS.

148

The

43.

fig.

qq and qq, are therefore

two products,

[II.

9.

i.

themselves unequal, as

above asserted, under the supposed condition of diplanarity.

On

169.

and qq
fig.

is

43,

made

round

to turn

its

and then the two new

the two factors, q and q

Thus we may conceive

are equal, as in algebra.

vanishes

when

the other hand,

are complanar

easy to prove, in several different ways, that their products, q q

versors, it is

middle point

B, until

that the arc cc

in

the spherical angle CBA

transvector-arcs, AC and C A

will evidently

become not only complanar but equal, in the sense of Art. 163, as being still
equally long, and being now similarly directed.
Or, in fig. 35, bis, of the last
cited Article, we may conceive a point E, bisecting the arc BC, and therefore
which is commedial therewith (comp. Art. 2, and the second
3
of
that
and then, if we represent the one versor q by either
figure
Article)
of the two equal arcs, AE, ED, we
may at the same time represent the other
versor q by either of the two other equal arcs, EC, BE ; so that the one product,
q q, will be represented by the arc AC, and the other product, qq , by the equal
also the arc AD,

arc BD.

Or, without reference to vector arcs,

we may suppose

that the two

factors are,

q =
OA, OB, oc being
fig.

35,

bis)

any

j3

a = OB

OA,

three complanar

for thus

we have only

= y a = oc OA,

and equally long right

lines (see again

to determine a fourth line, 8 or OD, of the

same length, and in the same plane, which shall satisfy the equation
=
/3:a (117), and therefore also (by 125) the alternate equation,
=
8 /3
y a and it will then immediately follow* (by 107) that
>:y

?-? =

We

may

75-

-=-8 = -S

7
Z=
y a

j3

pa

?Y-

therefore infer, for any two versors of quaternions, q

following reciprocal relations

if

the two

I-.-^-W
II.

and

= qq
q q

if

/Ilk (123);

then

q \\\q(l6S)

(as regards their places in the product] being thus at

once a consequence and &
proof of complanarity

convertibility of factors

f
170. In the 1st case of Art.
168, the factors q and q are both right versors
and because we have seen that then their two products, q q and qq are
(153)
;

It is evident
that, in this last process of reasoning,
of the four lines
compared; so that we

lengths

q q
q

qq

if

are versors.

jjj

q (123),

we make no use of the supposed equality of

might prove, in exactly the same way, that
without assuming that these two complanar factor-s, or quaternions, q and

ARTS. 168-171.]

MULTIPLICATION OF RIGHT VECTORS.

versors represented

by equally long but

circle, as in

the 1st case of 164,

are at once reciprocal (134),

are related as versor
I.

if

and

and
.

it

and

oppositely directed arcs of one great

follows (comp. 162) that these two products

conjugate (137), to each other; or that they

We may

reversor (158).

qq

Kq

149

and

q,

q be any two right versors

such versors, in two opposite orders,

II.

therefore write, generally,

qtf

-7-,

because the multiplication of any two

may always be represented or constructed

by a figure such as that lately numbered 43, in which the bisecting arcs AA
and cc are semicircles. The Hud formula may also be thus written (comp.
135, 154):
III.

and under

if

= -

and q z = -

1,

1,

then q q

qq = + 1

form it evidently agrees with ordinary algebra, because

under
the supposed conditions,
expresses that,

but

this

will be

it

found that

this last equation is not

it

cm identity in the general

theory of quaternions,
171. If the two bisecting semicircles cross each other at right angles, the

conjugate products are represented by two quadrants, oppositely turned, of one

great

circle.

It follows that if two right versors, in tivo mutually rectangular

planes, be multiplied together in two opposite orders, the two resulting products
ivill

be two opposite right versors, in a third plane, rectangular to the two former

or in symbols, that
if (f

1,

= -

1,

Ax

and

this case, therefore,

q q J_

Ax

(<?

and

q be

this result is

which

exists,

q,

to be

?)*

= - q\

q _L

Ax

= - ??

q q _L

Ax

q,

q.

in algebra a paradox,

namely

<f,

any two right versors, in tivo rectangular planes but we see that
not more paradoxical, in appearance, than the equation
;

under the same conditions.

in the last analysis,

it

<?

Ax

fa = -

find

Ax

we have what would be

the equation,

if

and

W~ (?--!,

then

In

=-

may

simply this

??

>

And when we come

to

examine what,

be said to be the meaning of this last equation,

:

we

that any two quadrantal or right rotations, in planes

ELEMENTS OF QUATERNIONS.

150
perpendicular

to

[II.

9.

i.

each other, compound themselves into a third right rotation, as

their resultant, in a plane perpendicular to each of

resultant rotation has

them

one or other of two opposite

and that

directions,

this third or

according to the

order in which the two component rotations are taken, so that one shall be
successive to the other.

172. "We propose to return, in the next section, to the consideration of

such a System of Right Versors as that which we have here briefly touched
upon but desire at present to remark (comp. 167) that a spherical triangle
ABC may serve to construct, by means of representative arcs (162), not only the
:

multiplication,

but also the

radial quotients)

by the

any one of two diplanar versors (or

In fact, we have only to conceive (comp.

division, of

other.

fig. 43) that the vector arc AB represents a given divisor, say q, or j3
a, and
that the transvector arc AC (167) represents a given dividend, suppose
or
:

<?",

y
same
:

for then the provector arc BC

plan, the quotient of these two versors,

versor

lately called q

we have

since

two

quaternions, as in algebra, the

173. It

must

is

namely

identities

however to be observed

we have found
f

identity, qq

= q

q, ceases

..

(q

(168), which

?(/ :?)=/

43,

q),

if q,

or qq

q
is

q"

for

by 106, 107, 120,

an equation as q

f
.

(q

q)

we

=
<?";

by 169, we may now

establish

be

if

f HI? (123);

then f\\\q.

still

been seen to be unequal thereto:

are always equally
arcs, AC and C A

as a vector arc has

true that these two last

and therefore subtend equal angles

that

we may

it is

at the centre o of the unit sphere

so

write, generally, for any two versors (or indeed for any two

quaternions}* q and

(106), or the

j3

represented by the arcs AB, BC,

not represented by AC, but by the different
,

long,

although

fact

qtf :q)=gr,

ff

product q

II. ..if

AC, the

or

that, for reasons already assigned,

In

be true.

to

I.

fig.

q,

on the

(168) that, for such versors, the ordinary algebraic

the two converse formulae

Accordingly, in

q":

generally,

not employ, for diplanar versors, such

because

arc C

(comp. again 167) will represent,

q",

the formula,

It Mall soon be seen

[see 191] that several of the formulae of the present section, respecting the
multiplication and division of versors, considered as radial quotients (151), require little or no modi
fication, in the passage to the corresponding operations on quaternions, considered as general quotients

of vectors (112).

EEPEESENTATIYE AND YECTOR ANGLES.

ARTS. 171-175.]

151

174. Another mode of Representation of Versors, or rather two such

modes, although intimately connected with each other,

may

new

be briefly noticed

here.

We may consider the

1st.

angle AOB, at the centre o of the unit-sphere,

when

conceived to have not only a definite quantity, but also a determined plane (110),
and a given direction therein (as indicated by one of the curved arrows in

by the arrow in fig. 33 [p. Ill]), as being what may be called

by analogy a Vector-Angle and may say that it represents, or that it is the
unitRepresentative Angle of, the Versor OB OA, where OA, OB are radii of the
fig.

39

[p. 131], or

sphere.

the
replace this rectilinear angle AOB at the centre, by
AC
at
what
be
called
the
Positive
S,
may
equal Spherical Angle
Pole of the representative arc AB so that C A and C B are quad

Ilnd. Or

we may

rants

and the

rotation, at this pole c

from the

first

of these

two

from a point outside the sphere),

which has been selected (111, 127) for the

quadrants to the second (as seen

has the direction

annexed figure 44

positive one, as indicated in the

we may

and then

consider this spherical angle as a new Angular Repre

sentative of the

same versor

q,

or OB

OA, as before.

now

that after employing a first spherical triangle ABC, to

construct (as in 167) the multiplication of any one given versor q, by any other
175. Conceive

given versor /, we form a second or polar triangle, of which the corners A B c

shall be respectively (in the sense just stated) the posi
,

of the three successive sides, BC, CA, AB, of the

tive poles

former triangle and that then we pass to a third tri

as part of the same lime B
with the
angle A
;

B"C

B"

the point diametrically opposite

shall be the negative pole of the arc

second, by taking for

to B

so that

B"

B"

CA, or the positive pole of

what was

the transvector-arc AC: also let

c"

lately called (167)

be, in like

the point opposite to c on the unit sphere.

may not only write (coinp. 129),

Ax

= oc

Ax

f
.

= OA

but shall also have the equations,

Lq = CA
Lq =
V,
B"C

B",

Ax

qq =

OB",

L qq

these three spherical angles,

namely the two

external vertical angle at

of the

B",

manner,

Then we

new

base-angles at c

or third triangle

and A

AW,

and the

will therefore

ELEMENTS OP QUATERNIONS.

152

represent, respectively,
plier^ q

on the plan of 174,

[II.

9.

i.

II., the multiplicand, q, the multi

and the product, q q. (Compare the annexed figure 45.)

Without expressly referring to the former triangle ABC, we can

176.

connect this last construction of multiplication of versors (175) with the general

formula (107), as follows.

Let a and j3 be now conceived to be two unit-tangents* to the sphere at c

two

perpendicular respectively to the

C

and drawn towards the same

arcs C

B"

and
?-

sides of those arcs

A and B respectively and let two

other unit-tangents, equal to these, and denoted by
the same letters, be drawn (as in the annexed figure
as the points

45,

at the points

bis)

same

to the

there

B"

and A

arcs C

so as to be

normal

and

to fall

and

B"

C A

towards the same sides of them as before.

two other unit-tangents, equal to each

each denoted by y, be drawn at the two

and A

Let
other,

also

and

last points

be both perpendicular to the arc

Then (comp.
A
and to fall towards the same side of it as the point c
174, II.) the two quotients, /3 a and 7 /3, will be equal to the two versors q
and q which were lately represented (in fig. 45) by the two base angles, at c

B"

so as to

B",

and A
is

the product, q
of the spherical triangle A B"C
,
therefore (by 107) equal to the third quotient, y a ;
;

q,

represented, as before,

which

by the

external vertical angle

C"B"A

of these

two

versors,

and consequently it is
of the same triangle,

evidently equal in quantity to the angle of this third quotient,

has the same axis OB", and the same direction of rotation, as the arrows in
45,

is

bis,

may

assist to

and
fig.

show.

177. In each of the two last figures, the internal vertical angle at

B"

is

thus

L q q, of the angle of the product and it is im

equal to the Supplement, IT
ff
to
that
the
observe
portant
corresponding rotation at the vertex v ,from the
;

side

B"A

to the side

the point c

is

B"C

or (as

positive] a result

we may
which

briefly express

is

reasoning of the foregoing Article.t

when

the multiplication of

of which the
*

By

tangential

two

is

unit-sphere,

from

the point

easily seen to be a general one,

to

by the

We may
is

then infer, generally, that

constructed by a spherical triangle,

base angles represent (as in the

an unit tangent
to the

any two versors

it)

two

last Articles) the factors,

here meant simply an unit line (or unit vector, 129) so drawn as to be
and to have its origin, or its initial point (1), on the surface of that

sphere, and not (as we have usually supposed) at the centre thereof.
and to look towards the arc A C , it would
t If a person he supposed to stand on the sphere at
appear to him to have a right-handed direction, which is the one here adopted as positive (127).
B",

AETS. 175-178.]

DEPENDENCE OE PRODUCT ON ORDER OP FACTORS.

153

while the external vertical angle represents the product, then the rotation round
the axis

of that product q

(OB")

axis (oc

of the multiplicand

cand, to the

q, is positive

to the

follows that the rotation

it

fully, since the only rotations hitherto considered

we may say

to the

then the rotation round

Ax

178.

on a plane perpendicular

of the multiplier q

plane ones (as in 128, &c.),

whence

same thing more

Ax

the axis (OA

from

Ax q of the multiplier, from the axis Ax q of the multipli

axis Ax q q of the product, is also positive.
Or, to express the

round the axis

of

q,

that

if

the two latter axes be projected

former, so as still to
the projection of

,from

have been

have a common origin

Ax

o,

q to the projection

will be directed (with our conventions) toivards the right hand.

q,

We

have therefore thus a new mode of geometrically exhibiting the

inequality of the two products, q q

taken as factors in two

Ax

and qq

For

different orders.

Ax

= OP,

of two diplanar versors (168),

when

this purpose, let

Ax

= OQ,

q q

= OR

and prolong to some point s the arc PR of a great circle on the unit sphere.
Then, for the spherical triangle PQR, by principles lately established, we shall
have (comp. 175) the following values of the two internal base angles at p
and

Q,

and

of the external vertical angle at

RPQ = L
and the rotation

at Q,

PQR = L

SRQ = L q q

from the side QP to the side QR will be right-handed.

an arcual perpendicular, RT, from the vertex R on

the base PQ, and prolong this perpendicular to R in such a
Let

fall

manner

as to

have
^ RT = A TR

We

shall then have a

prolong PR to some point s
new triangle PQR which will be a sort of reflexion (comp.

also

138) of the old one with respect to their

and
qq

For the

banded, as

it

rotation at P

ought

QPR = L
so that the

qq

common

to be

from PQ

and we

R QP = L q

to

shall

new external and spherical

PR

base PQ

new product,

this neiv triangle will serve to construct the

will be right-

have the equations,

QR

angle,

= Lqq

QR

OR =

HAMILTON

And

f
.

will represent the

as the old angle SRQ represented the old versor, q q, obtained

order of the factors.

Ax

qq

new

t ersor,

from a different

although, no doubt, these two angles, at u and R

ELEMENTS OF QUATERNIONS,

ELEMENTS OF QTJATEENIONS.

154

we may

are always equal in quantity, so that

[II.

9.

i.

establish (comp. 173) the general

formula,

L q q = L qq

>

yet as vector angles (174), and therefore as representatives of versors, they must
be considered to be unequal: because they have different planes, namely, the
or the two planes
tangent planes to the sphere at the two vertices R and R
;

which are drawn through

respectively parallel to these,

the centre o.

by means of Re
Thus to
(162).

179. Division of Versors (comp. 172) can be constructed

presentative Angles (174), as well as
divide

by

q"

q,

entirely similar to that last

representative arcs

Ax

OP =

to find a third point

fig.

to

46, by the two conditions,

OR =

q,

we have only

for multiplication,

employed

determine the two points p and R, in

and then

by

or rather to represent such division geometrically, on a plan

Ax

.
q",

Q by the two angular equations,

RPQ = L

QRP =

q,

TT

- L

q",

the rotation round p from PR towards PQ being positive; after which

we

shall

have,
Ax.fa":g)=oo;

Instead of conceiving, in

(1.)

nects the vertices of the

fig.

L(q":q)

46, that the dotted line

triangles, with PQ for their

two

RTR which con

,

common

base (178),

an arc of a great circle, perpendicularly bisected by that base, we may

imagine it to be an arc of a small circle, described with the point P for its

is

And

positive pole (comp. 174, II.).

then we

the versor

173) from

metrically performed

Ax

q,

q",

by

or q

q, to the

may

say that the passage (comp.

unequal versor q

a Conical Rotation of the

Axis

:
(q"

Ax

q),

q",

or qq

round

is

geo

the axis

=
through an angle 2/_q, without any (quantitative] change of the angle L

so that

we

q"

have, as before, the general formula (comp. again 173),

employ the construction of multiplication and

division by representative arcs, which fig. 43 [p. 144] was designed to illus
is determined in that figure by the
trate, and conceive that a new point
= n cV, we may then say that in the passage from the versor
condition * A
(2.)

Or

if

we

prefer to

c"

C"

q",

which

C",

is

represented by AC, to the versor q

the representative arc of

is
q"

made

:
(q"

q),

represented by

cV or

by

to move, without change of length, so as

CONICAL ROTATION OF AXIS OF VERSORS.

ABTS. 178-180.]

AB of

to preserve a constant inclination* to the representative arc

initial point describes the double

155

while

q,

from A to A

of that arc AB, in passing

its

be seen, by these few examples, that if, even independently

and U, new combinations of
of some new characteristics of operation, such as
It

(3.)

may

occur in the present Calculus, which are not

g(q":q),
wanted in algebra, they admit for the most part of geometrical interpretations,

old symbols, such as

and interesting kind

of an easy

and

in fact represent conceptions,

not well be dispensed with, and which

so

much

simplicity

and

it is

which can

useful to be able to express, with

(Compare the remarks in Art. 161

conciseness.

and

the sub-articles to 132, 145.)

180. In connexion with the construction indicated by the two figures 45,
c
it may be here remarked, that if ABC be any spherical triangle, and if A , B
,

be (as in 175) the positive poles of

rotation (comp. 177, 179)

The

&c., is positive.
is to

conceive that

we make

it

round

easiest

from B

to c

sides, BC, CA,

AB, then the

from

or that round B

c to

way,

on the

to the diametrically opposite triangle

which will not change the poles A

any

sphere^

Assuming then that these poles

we

are thus the near ones to the corresponding corners of the given triangle,
arrive without

perhaps, of seeing the truth of this assertion

the rotation round A from B to c be not already positive,

if

by passing

such,

A.

three successive

its

difficulty at the conclusion stated above

which has been

employed in our construction of multiplication (and division") of versors,

by means of Representative Angles (175, 176) and which may be otherwise

virtually

by the consideration of the unit-tangents of

justified (as before),

(1.)

round the
177)

bis.

45,

fig.

Let then

and

first,

a, ]3, y be any three given unit vectors, such that the rotation
from the second to the third, is positive (in the sense of Art.

let a

/3

y be three other unit

from these by the

vectors, derived

equations,

a =

Ax

(y

j3),

j3

then the rotation round a, from

have the converse formulae,

(2.)

being

Ax

(a

j3

to

y),

y =

Ax

(/3

will be positive also

a)

and we

If the rotation round a from

still

shall

to y were given to be negative, a, /3 y

]3
deduced from those three vectors by the same three equations as

before, then the signs of a,

* In a

manner analogous

/3,

to the

y would

all

require to be changed, in the three

motion of the equator on the

ecliptic,

by luni-solar

precession, in

astronomy.

ELEMENTS OF QUATERNIONS.

156
last (or reciprocal)
still

be

formulae

[II.

but the rotation round a, from

9, 10.

i.

j3

to

would

positive.

(3.)

Before closing this section,

it

may

be briefly noticed, that

it is

some

times convenient, from motives of analogy (comp. Art. 5), to speak of the
Transvector-Arc (167), which has been seen to represent a product of two
versors, as being the ARCUAL SUM of the two successive vector-arcs, which
represent (on the same plan) the factors Provector being still said to be added
but the Order of such Addition of Diplanar Arcs being not now
to Vector
indifferent (168), as the corresponding order had been early found (in 7) to
be, when the vectors to be added were right lines.
[Thus in fig. 43, n EC
;

= n cV. But n BA = r\ AB and n

= n BO,
+ n AB = n AC and n BA + n
= n cV. If a and
/3 are any two vector arcs, and if
consequently n AB + n BC
</B

</B

is
o equal to xa
is any scalar, x (a
Compare 14, and notice
j3)
#]3.
that the property there proved depends on the possibility of constructing

similar plane triangles of different sizes.]

(4.)

We may also speak occasionally, by

an extension of the same analogy,

SPHERICAL

of a spherical triangle, as being the

of the External Vertical Angle

of the two Base Angles of that triangle, taken in a suitable order of sum
mation (comp. fig. 46) the Angle which represents (174) the Multiplier being

SUM

then said to be added

(as a sort of Angular Provector) to that other Vectorwhich

represents the Multiplicand whilst what is here called the sum
Angle
of these two angles (and is, with respect to them, a species of Transvector;

Angle) represents, as has been proved, the Product.

of angular transaction becomes perhaps a little more
(5.) This conception
clear, when (on the plan of 174, I.) we assume the centre o as the common
vertex of three angles AOB, BOC, AOC, situated generally in three different planes.

For then we may conceive a

successive angular motions,

revolving radius to be either carried

from OA

to OB,

and thence

to oc

by two

or to be trans

ported immediately, by one sucli motion, from the first to the third position.
as regards the construction indicated by fig. 45, bis, in which
(6.) Finally,
tangents instead of radii were employed,
here, that

it

may

be well to remark distinctly

AW, in that figure, may be any given spherical triangle, for

the rotation round

and q

B"

from A

to c is positive (177)

which

and that then, if the two

which the internal angles at
;

be defined to be the two versors, of

c and A are (in the sense of 174, II.) the representatives, the reasonings of

factors q

Art. 176 will prove, without necessarily referring, even in thought, to any other
is (in the same sense) the
triangle (such as ABC), that the external angle at
B"

representative of

the product, qq, as before.

SYSTEM OF THREE EIGHT VERSORS.

ARTS. 180-182.]

SECTION

157

10.

Yersors, in Three Rectangular

on the I^aws of the Symbols, i, j, k.

On a System of Three Right

Planes

and

181. Suppose that

OK

GJ,

OT,

round the

rectangular unit-lines, the rotation

third being positive

and

let 01

oj

any three given and co-initial but

first from the second to the

are

OK be the

three unit- vectors respectively

opposite to these, so that

= -

01

GJ =

01,

Let the three new symbols

,/,

OK = -

oj,

OK.

k denote a system (oomp. 172) of three right

versors, in three mutually rectangular planes, with the

three given lines for their respective axes

Ax

Ax ./ = oj,

01,

Ax

so that

k = OK,

and
i

47

as figure

have

= OK

these

versors

j = 01 OK,

oj,

k = oj

We

serve to illustrate.

may

other expressions

the

for

01,

shall then

same three

= oj

j = OK
k = 01

= oj

OK = OK

oj

01

01

OK = OK

oj = oj

01

= 01

while the three respectively opposite versors

= oj

j = OK
- k =

01

may

OK

01

oj

be thus expressed

OK = OK

oj = oj

OK = OK

01

01

OK = OK

01

oj = oj

01

oj

01

oj

= 01

OK

= oj

01

And from

the comparison of these different expressions several important

symbolical consequences follow, which it will be worth while to enunciate
separately here, although some of them are virtually included in the results
of former sections.

182. In the first place, since

2-

we deduce (comp.
symbols

(oj

OK)

(OK

oj)

= oj

oj, &c.,

148) the following equal values for the squares of the

I.

1;

=-l;

new

ELEMENTS OP QUATERNIONS.

158

might indeed have been

as

i.

[II.

10.

from the circumstance

at once inferred (154),

that the three radial quotients, (146), denoted here

by

i,j, k, are all right

versors (181).

In the second

place, since
i

,j = (oj

OK

(OK

= oj

01)

01, &c.,

we have

the following values for the products of the same three symbols, or
versors, when taken two by two, and in a certain order of succession (comp.
168, 171)

II.

But

ij

= k

Jk =

ki=j.

in the third place (comp. again 171), since

j .iwe have

(01

OK)

(OK

oj)

= 01

oj, &c.,

and contrasted formulae, for the binary products of the

these other

same three right versors, when taken as factors with an opposite order
III.

ji

=- k

kj

= -

ik

- - j.

Hence, while the square of each of the three right versors, denoted by these
three new symbols, ijk, is equal (154) to negative unity, the
product of any

tivo

of

them

is

equal either to the third

itself,

or to the opposite (171) of that third versor, according as the

j\

multiplier precedes or follows the multiplicand, in the cyclical

succession,
i>j>

which the annexed figure 47,

(1.)

To

...

k, i,j,

and

bis,

may

of representative angles (174),

one of the four quadrantal

arcs, JK,

KJ

four spherical right angles, JIK, KIJ

common

pole

i,

The

IK

K K J, in fig.
K IJ, which
,

as representing the versor

with the introduction of the point

being at the back of the figure.
(2.)

bis.

give some help towards remembering.

connect such multiplications of i, j, k with the theory of repre

sentative arcs (162),

their

Fig. 47,

squaring of

i,

opposite to

or the equation

i*

= -

47, or

any one

which

1,

of the

those arcs subtend at

and similarly

i,

we may regard any

is

for

j and

k,

to be conceived as

comes thus to be geome

by the doubling (comp. Arts. 148, 154, and figs. 41, 42) of
an arc, or of an angle. Thus, we may conceive the quadrant KJ to be added
to the equal arc JK, their sum
the great semicircle Jj , which (by 166)

trically constructed

represents

may

being
an inversor (153), or negative unity considered as & factor.

Or we

add the riyht angle KIJ to the equal angle JIK, and so obtain a rotation

LAWS OF THE SYMBOLS,

ARTS. 182-183.]

J,

I,

K.

159

through two right angles at the pole i, or at the centre o which rotation is
equivalent (comp. 154, 174) to an inversion of direction, or to a passage from
;

the radius OJ, to the opposite radius QJ

or the equation ij = k, may in like manner
(3.) The multiplication of/ by i,
be arcuatty constructed, by the addition of K J, as a projector-arc (167), to IK
.

as a vector-arc (162), giving

u, which

arc, or arctial-sum (180, (3.)),

Or

is

a representative of

the same multiplication

constructed, with the help of the spherical triangle

/,

and

J represent respectively the multiplier,

the rotation round

from

J to

K being

positive

(180, (4.)), or the external vertical angle at

same product,
(4.)

The

manner be

i,

as the transactor-

and the

(comp. 175, 176), represents the

contrasted multiplication of

by

/, or of

into*

constructed, or geometrically represented, either

new

by

is negative,

so that the angle at

may

i,

in like

the addition of

which new process

or with the aid of the new
vector,

gives Ji (instead of u) as the new transvector

triangle UK (comp. figs. 46, 47), in which the rotation round
vertex

multiplicand,

while their spherical sum

k, as before.

the arc KI, as a new provector y to the arc JK as a

new

be angularly

may

in which the base-

UK;

angles at

k,

represents

now the

from

J to the

multiplicand,

and the resulting angle at the new pole K represents the new and opposite
- k.
product, ji =
183. Since we have thus ji = - ij (as we had q q = - qq in 171), we see
that the laws of combination of the new symbols,

same as the corresponding laws in algebra

i,

j, k, are not in all respects the

since the Commutative Property of

Multiplication, or the convertibility (169) of the places of the factors without

change of value of the product, does not here hold good: which arises (168)
from the circumstance, that the factors to be combined are here diplanar
It is therefore important to observe, that there is a respect in
which the laws of i, j, k agree with usual and algebraic laws namely, in the

versors (181).

Associative Property of Multiplication

or in the property that the

new symbols

always obey the associative formula (comp. 9),

t

whichever of them

may

multiplicand

is said to

K\ =

IK

A,

be substituted for

which equality of values we

*

i,

for

K,

and

for

in virtue of

omit the point, in any such symbol of a

may

be multiplied ly the multiplier

while, on the other hand, a multiplier

said to be multiplied into the multiplicand

a distinction of this sort between the two factors being
necessary, as we have seen, for quaternions, although it is not needed for algebra.

is

ELEMENTS OE QUATERNIONS.

160

and write

ternary product (whether of equal or of unequal factors),

as IK\.

jk =

= -

i*

= -

ijk

We may,

therefore,

1.

=/

# =

= -

ijk

shall occasionally refer, as to

"

basis for the v/hole

sufficient symbolical

shown that

x, y, z

Formula

ix

+ jy +

and therefore

ijk,

:*

shall

to be a

because

it

Form,

kz,

while

scalars,

j,

i,

k are the same

[See 221.]

= -

direct proof of the equation, ijk

definitions of the

and which we

A,"

Calculus of Quaternions

compose a system of four

three right rersors as above.

= w+

(A)

every quaternion can be reduced to the Quadrinomial

where w,

Formula

find to contain (virtually) all the laws of the symbols

(1.)

simply

= k.k = k z = - 1;

establish the following important

by 182,
i*

which we

ij

or briefly,

will be

it

10.

In particular we have thus,

i

to

i.

[II.

symbols in Art. 181

In

1,

be derived from the

may

we have only

fact,

to

remember

that those definitions were seen to give,

i

and
four

= oj

to observe that,
lines

may

OK,

01

by the general formula

be denoted by
?

j = OK

oj

of multiplication (107), whatever

we have always,

a, |3, 7, S,

L?i

7
)3

k = 01

*
= ? z = = l
a
a
j3 a
y

|3

or briefly, as in algebra,
S
~~

7/3_S
~

~~

~~

7*

the point being thus omitted without danger of confusion

ijk

= o/ oj = :

1,

so that

as before.

* This formula
(A) was accordingly made the basis of that Calculus in the first communication on
by the present writer, to the Royal Irish Academy in 1843 and the letters i, j, k,
continued to be, for some time, the only peculiar symbols of the Calculus in question. But it was
the subject,

and U, &c.),
gradually found to be useful to incorporate witli these a few other notations (such as
It \vas also thought to be instructive to establish the
for representing Operations on Quaternions.
principles of that Calculus, on a more geometrical (or less exclusively symbolical) foundation than at
Lectures on Quaternions
first ; which was accordingly afterwards done, in the volume entitled
:

again attempted in the present work, although with many differences in the
adopted plan of exposition, and in the applications brought forward, or suppressed.
(Dublin, 1853)

and

is

LAWS OE THE SYMBOLS

ARTS. 183, 184.]

Similarly,

we have

these two other ternary products

jki = (OK
kij

On

(2.)

(oj

OK

(OK

OK) = OK

01)

oj)

161

OK = -

01

= -

01

1.

the other hand,

kji

and

oj) (oj

(01

oj

01) (01

K.

I, J,

(oj

01) (01

OK) (OK

= oj oj = +
:

in like manner,
ikj

= +

and

1,

=+

jik

1.

(3.) The equations in 182 give also these other ternary products, in which
the law of association offactors is still obeyed
:

-/ =

ij

i.ji = i.-k =

ik

ij

i.jj

ik

=i.-l

ii.j,

=j

= -i = kj =

ki

iij

ij.i,

iji

ij.j,

with others deducible from these, by mere

ijj

cyclical

-J-,

= +j\
= -

i;

permutation of the

letters,

on the plan illustrated by fig. 47, bis.

(4.) In general, if the Associative Law of Combination exist for any three
symbols whatever of a given class, and for a given mode of combination, as for
addition of lines in Art. 9, or for multiplication of ijk in the present Article,

the same law exists for any four (or more) symbols of the same class, and
combinations of the same kind. For example, if each of the four letters i, K,

denote some one of the three symbols

one), we have the formula,

A,

fj.

(5.)

manner

IcA^t

\JLL

IK

\fJL

IK

i,j,

k (but not necessarily the same

fj.

- IK\

fJL

lK\fJL.

Hence, any multiple (or complex) product of the symbols ijk, in any
repeated, but taken in one given order, may be interpreted, with one

by any mode of association, or of reduction to partial factors,

can
be
which
performed without commutation, or change of place of the given
factors.
For example, the symbol ijkkji may be interpreted in either of the

definite result,

two following
ij

(among

.kk.ji =

ij

other)

-ji =

ways

i.-f

The formula (A) of 183

To show that it includes

184.
182.

the last cited Article,

HAMILTON

we may

ELEMENTS OF QUATERNIONS.

ii

= -

ijk

kji

= -

= -

1.

includes obviously the three equations (I.) of

also the six other equations, (II.),

observe that

it

(III.)>

of

gives, with the help of the

Y

ELEMENTS OF QUATERNIONS.

162

associative principle of multiplication

the absence of the point in the

by

ij

= -

ij

kk = -

ijk

symbol

& = +

7i-

jk =

= j jk =j k = -k;
= ij.j = if = ~ *
kj
.

And

then

it is

= -

jki = kij

kji

= Pj =

ki

= - ky = -ji 2 = +/.

we have

1,

=+

ijk

ik

easy to prove, without any reference

laws of the symbols be admitted, that

memory

ijk},

ji

10, 11.

i.

be suggested to the

may

(which

[II.

ij

to

-/

geometry,

if

the foregoing

also,

=jik =

ikj

=+

1,

and geometrically shown in recent sub-articles. It may be added

that the mere inspection of the formula (A) is sufficient to show that the three*
as otherwise

square roots of negative unity, denoted in

ordinary rules of algebra

it

by

i9

k>

cannot be subject

to all the

because that formula gives, at sight,

the non-commutative character (183), of the multiplication of such roots

among

themselves, being thus put in evidence.

[Conversely

if

three symbols

jk + kj

and

if

ki

0,

and k

/,

e,

ik

satisfy the equations

0,

ij

+ ji =

the associative property hold good,

/*

j=

= -

ij

ji

= -

ij

=j

i*.

is therefore commutative in multiplication with ? j, and k and with pro

ducts formed from them, and cannot be distinguished from a scalar.
Assum
ing therefore that the squares of the symbols are scalars, and that the symbols

have been multiplied by suitable numerical

2

are equal,

Again

.jk =

iQ =

and

The product Q

-i.kj =

is

ijk

z
=j =

kij

= -

k z = P.
kji

=jki = -jik =

jk = jki

likewise commutative with

able from a scalar.

Also

Q =2

ijk

coefficients so that their squares

kji

= -

j,
3
,

suppose,

Qi.

and
so if

k,

and is indistinguish
= - 1, Q =
1.

* It is evident that
1
k are also, on the same principles, values of the symbol V
i,
j,
More generally (comp. a Note to
because they also are right versors (153) or because ( q} 2 = q2
2
2 =
2
1, it will be proved,
page 133), if x, y, z be any three scalars which satisfy the condition # -f 7/ + z
;

at a later stage, that

TENSOR OF A VECTOR.

AETS. 184-185.]

Mr. Oliver Heaviside takes

z
=
consequently i .j=j but i ij
.

P=+
ik

but

163

i=jk,j=ki and

= - /, and in

his

k =

ij,

and

system the associative

property does not hold, or the product of three symbols has no definite mean
ing (see Art. 25 of a paper On the Forces in the Electro -magnetic Field,"
"

Trans. Koy. Soc. A., 1892).

Grassmann supposes
tive,

from

0,

and

his progressive multiplication is associa

but his regressive multiplication is not.

Q is taken as a scalar differing
zero. Hamilton (p. 61 of the preface to the "Lectures on Quaternions")

refers to the octaves of Messrs. J. T.

Graves and Arthur Cay ley as not obey

On the 8-square
ing the associative principle. See Prof. Cay ley s paper
When the associative principle
Imaginaries," Am. Jour, of Math., 1881.
does not hold, a distinct operation of grouping must be combined with multi
"

plication to render a product definite.]

SECTION

On

11.

the Tensor of a Vector, or of a (fcuaternion ; and on the

Product or duotieiit of any two Q,uaternions.

185.

Having now

sufficiently availed ourselves, in the

two

last sections,

of the conceptions (alluded to, so early as in the First Article of these

Ele

ments) of a rector-arc (162), and of a vector-angle (174) in illustration* of the

laws of multiplication and division of versors of quaternions we propose to
;

return to that use of the word,

the

first

eight sections of this

VECTOR, with which alone the First Book, and

First Chapter of the Second Book, have been

mean again, exclusively, by that

a
Directed
Line
in
And because we have already
vector,"
Right
(as
1).
considered and expressed the Direction of any such line, by introducing the
concerned: and shall therefore henceforth

word

"

conception and notation (155) of the Unit-Vector, Ua, which has the same
direction with the line a, and which we have
proposed (156) to call the Versor
that
we now propose to consider and express the Length of the
of
Vector, a
;

same

line a,

by introducing the new name TENSOR, and the new symbol^ Ta

* One of the chief uses of such

vectors, in connexion with those laws, has been to illustrate the
non-commutative property (168) of multiplication of versors, hy exhibiting a corresponding
property of
what has been called, by analogy to the earlier operation of the same kind on linear vectors
(5), the
addition of arcs and angles on a sphere.
Compare 180, (3.), (4.).

f Compare the Note to Art. 155.

ELEMENTS OF QTJATEBNIONS.

164

which

symbol we

latter

define it

[II.

shall read, as the Tensor of the Vector a

to be, or to denote, the

Number (comp. again

11.

i.

and

shall

155) which represents the

Length of that line a, by expressing the Ratio which that length bears
assumed standard, or Unit (128).

some

to

To

connect more closely these two conceptions, of the versor and the
we may remember that when we employed (in 155) the
letter a as a temporary symbol for the number which thus expresses the
186.

tensor of a vector,

Ua

length of the line a, we had the equation,

definition of the unit-vector denoted by Ua.

We

also these

two other forms

construction of

fig.

42,

vector, a,

(comp. 161,

OA

a,

as one

form

of the

might therefore have written

Ua,

and

= OB

(2.)),

For example, with the

Ua.

we may

OB

on each other,

of the scalar, a,

called (156) the versor,

bis

a = OA

a = a

Ua,

dependence of the

and on what has been

of equation (comp. 15, 16),

a = a
to express the

= a

write the three equations,

= oc oc
:

be thus the three positive scalars, which denote the lengths of the three
and these three scalars may then be considered as factors,
OA, OB, oc

if a, b, c

lines,

by which the three unit-vectors Ua, U/3, Uy, or OA OB ,

the
cited
(in
figure), are to be respectively multiplied (15), in order to
change them into the three other vectors a, |3, y, or OA, OB, oc, by altering

or as

coefficients (12),

oc

without any change in their directions. But such an exclusive

the Length (or on the extension] of a line, may be said to be an
on
Operation,
Act of Tension ;* as an operation on direction alone may be called (cornp. 151)
their lengths,

an

We

act of version.

have then thus a motive for the introduction of the

name, Tensor, as applied to the positive number which (as above) represents the
And when the notation Ta (instead of a) is employed for
length of a line.

such a

tensor,

we

see that

we may

equations (compare again 15, 16)

Ua
For example,

= a
if

Ta

Ta =

write generally, for any vector

the

Ua

a =

therefore, generally, whatever vector

Compare the

may

first

Ta Ua = Ua
.

Ta.

= a (160), then Ta = 1 and

be denoted by a, we have always,

a be an unit-vector, so that

a,

Note

Ua

in

page 137.

GEOMETEICAL EXAMPLES.

ARTS. 185, 186.]

For the same

reason, ivhatever quaternion

may

165

be denoted by

<?,

we have always

(comp. again 160) the equation,

T (Ax .?)

!.

Hence the equation

(1.)

where p =

OP, expresses that the locus of the variable point P is the surface of
the unit sphere (128).
= Ta expresses that the locus of p is the spheric
(2.) The equation Tp

surface with o for centre, which passes through the point A.

On

(3.)

A,

we have

the other hand, for the sphere through

which expresses that the lengths of the two

More

(4.)

o,

which has

its centre at

the equation,

lines, AP, AO, are equal.

generally, the equation,

(p

expresses that the locus of p

is

a)

(J3

a),

the spheric surface through B, which has

its

centre at A.

The equation

(5.)

of the Apollonian* Locus, 145,

under either of the two following forms

(p

from each of which we

stage,

(6.)

The

because

[and 200

a)

Tp = aTa

(p + a)
is

expresses that

from the two points A and A

if
.

(p

the plane through o, perpendicular to the line

OA = - OA, then the point P is equally distant
It represents therefore the same locus as the

^=
a

&

of 132, (1.);

or as the equation,

^ +

K^a =

0,

of 144, (1.);

or as
"^

geometry

a),

equation,

to

(3.), (4.)]).

equation,

it

be written

shall find ourselves able to pass to the other, at a later

expresses that the locus of p

;

GO

may

by general Rules of Transformation, without appealing

(comp. 145, (10.)

OA

= aT

a*a)

(8.), (9.),

Compare the

1, of

first

161,

Note

to

(7.)

page 130.

ELEMENTS OF QUATEENIOKS.

166

or as the simple geometrical formula,

found

by General Rules of
formulae into any other of them

five

possible,

which expresses that the

that this quaternion

In

(7.)

like

is

line

A and

in fact

this Calculus, to transform

it

i.

11.

will be

any one of these

or into this sixth form,

and therefore

a right quotient (132).

manner, the equation

because

And

J_ a (129).

scalar part* of the quaternion - is zero,

the plane which perpendicularly bisects the

expresses that p is equally distant from the two points

expresses that the locus of P

AB

[II.

it

is

B.

(8.)

The

limit) with a,

tensor,

we

Ta, being generally a positive scalar, but vanishing (as a

have,

T#a =

according as

#Ta,

>

or

<

thus, in particular,

T(-a) = Ta;
(9.)

and

TOa = TO =

0.

That

= T/3 + Ta,

a)

if

Uj3 = Ua,

but not otherwise (a and /3 being any two actual vectors), will be seen, at a
later stage, to be a symbolical consequence from the rules of the present
Calculus but in the mean time it may be geometrically proved, by conceiving
= oc - OA = AC
that while a = OA, as usual, we make
+ a = oc, and therefore
;

j3

/3

(4)

for thus

we

shall see that while, in general, the three points o, A, c are

corners of a triangle,

and therefore the

of the lengths of the

two other sides OA and

on the contrary, equal

triangle vanishes,

and

AC, or a

and

length of the side oc

to the latter

AC, the

is less

than the sum

former length becomes,

sum, in the particular case

when the

by the point A falling on the finite line oc in which case, OA

Ua = U/3
j3, have one common direction, as the equation
;

implies.
(10.) If

(Ua

a and

]3

be any actual vectors, and

if

their versors be unequal

not = U/3), then

T
an inequality which
*

Compare the Note

(j3

a)

results at once

to

page 127

<

T/3 +

Ta

from the consideration of the recent

and the following Section of the present Chapter.

TENSOR OF A QUATERNION.

ARTS. 186-187.]
triangle

by

OAC

but which

be found)

(as it will

rules of the calculus of quaternions.

(11.) If

T/3

>

or

<

= - Ua, then
but

U/3

Ta

a)

(T/3

>

(|3

may

167

be symbolically proved,

also

[See 210 (15.)]

- Ta),
+ a) =
according as
(T/3

- Ta),

= - Ua.
U/3 not

if

187. The quotient, U/3 Ua, of the versors of the two vectors, a and /3, has
been called (156) the Versor of the Quotient, or quaternion, q = |3 a and has
On the same plan, we propose
been denoted, as such, by the symbol, Ug.
:

now

to call the quotient, T/3

Tensor* of the Quaternion

And then,

symbol, Tq.
istic of

q,

as

Ta, of the tensors of the same two vectors, the

or

]3

and

to denote

it

called the letter

a,

we have

the operation of taking the versor, so

by the corresponding
(in

156) the character

we may now speak

of

a Vector,

and

may

a,

or of a Quaternion,

(]3

a)

= T/3

We

q.

Ta, as

say that as the versor

the quaternion q
(1.)

Hence

is

as the

shall thus have, generally,

we had

(]3

Uq depended

a)

U/3

Ua

(156)

relative length-^ (109), of

solely on, but conversely

sufficient to determine, the relative direction

(157), so the tensor

and determines the

whether of

Characteristic of the (corresponding) Operation of taking the Tensor,

was

Tq depends on

the two vectors, a and

]3,

of

which

the quotient (112).

the equation

T-

expresses that the locus of P

is

1, like

Tp = Ta,

to

which

it is

equivalent,

the sphere with o for centre, which passes

through the point A.

(2.)

The

equation (comp. 186,

(6.)),

p
expresses that the locus of P

is

- a

the plane through

o,

perpendicular to the

line OA.
*

Compare the Note

It has

to Art. 109, in page 111

and the first Note in page 137.
been shown, in Art. 112, and in the Additional Illustrations of the third section of the
present Chapter (113-116), that Relative Length, as well as relative direction, enters as an. essential
element into the very Conception of a Quaternion.
Accordingly, in Art. 117, an agreement of relative
lengths (as well a an agreement of relative directions) was made one of the conditions of equality,
between any two quaternions, considered as quotients of vectors so that we
may now say, that the
;

tensors (as well as the

versors} of equal quaternions are equal.
Compare the first Note to page 138, as
regards what was there called the quantitative element, of absolute or relative length, which was

eliminated from a, or from q,

;
by means of the characteristic
whereas, the new characteristic, T,
of the present section, serves on the
contrary to retain that element alone, and to eliminate what may
be called by contrast the qualitative
element, of absolute or relative direction.

ELEMENTS OF QUATEKNTONS.

168

Other examples of the same sort

(3.)

[II.

easily be

may

i.

11.

derived from the

186, by introducing the notation (187) for the tensor of a

quotient, or quaternion, as additional to that for the tensor of a vector (185).

sub-articles to

(4.)

The

(5.)

()3

a)

=, or

>,

according as Tj3

1,

<

tensor of a right quotient (132) is

>,

=, or

<

Ta.

always equal to the tensor of

its

index (133).

The

(6.)

generally,

tensor of a radial (146) is

by

always

positive unity

thus we have,

156,

TUg-1;
and in

particular,

by 181,
T?:

Txq =

(7.)

thus, in particular,

T/=T&

1.

xTg, according as x

or

>

<

= Tq, or the tensors of opposite quaternions are

(- q)

equal.

T# =

(8.)

thus, the tensor of a scalar

according as x

= Ta,

we may

establish the symbolical equation,

HP 2

TT

(comp. 145, 160),

Because the tensor of a quaternion is generally a positive scalar,

and its versor
is its own conjugate (139)
its angle is zero (131)

such a tensor
is

<

from, the subject of the operation

by abstracting

(159)

or

TT^T?;

so that,

(10.)

>

that scalar taken positively.

is

TTa

Hence,

(9.)

x,

positive unity

KTq
T (1

(11.)

or in symbols,

= T?
:

q)

L Tq =

T (a

0)

UT?

= Ta T]3 =
:

or in words, the tensor of the reciprocal of a quaternion

1.

T?
is

equal to the reciprocal

of the tensor.
(12.)

Again, since the two

long, the definition (137) of a

lines,

OB and OB

in

fig.

36

[p.

115], are equally

conjugate gives

or in words, the tensors of conjugate quaternions are equal.

(13.)

It

have equal

is

scarcely necessary to remark, that

tensors,

and equal

=q,

any two quaternions which

versors, are themselves equal

if

lY=Tfc

and

U/=W

or in symbols, that

TENSOK OF A QUATERNION.

ARTS. 187-189.]]

188. Since

we

169

have, generally,
Tj3

Uj3

TJ/3

Tj3

186

*
>

TJ

we may

>

two following general formulae of decomposition of a

tic o factors, of the tensor and versor kinds

establish the

quaternion into

to the formulae

which are exactly analogous

decomposition of a vector, into factors of the

I

a =

Ta Ua
.

IF.

for the

(186)

same two kinds

.

Ua

corresponding

namely,

Ta.

To illustrate this last decomposition of a quaternion, q, or OB OA, into factors,

we may conceive that A A and BB are two concentric and circular, but oppo
:

sitely

directed arcs, which terminate respectively

on the two
of those

lines

two

OB and OA, or rather on the longer

and on the shorter of them

lines itself,

prolonged, as in the annexed figure 48 so that OA

has the length of OA, but the direction of OB, while
;

on the contrary, has the length of OB, but the

and that therefore we may write,
direction of OA
OB

by what has been defined respecting

185, 186),

OA = Ta

U/3

versors

and

tensors of vectors (155, 156,

OB = T|3 Ua.
.

Then, by the definitions in 156, 187, of the versor and tensor of a quaternion,

U (OB

Uq

Tq

=T

(OB

OA)

= OA

OA = OB

OB

OA)

= OB

OA = OB

OA

whence, by the general formula of multiplication of quotients (107),

I.

= OB OA = (OB OA

II.

= OB OA =

(OA

OA)

= Tq

OA)

= ~Uq

U?

and
:

(OB

OB

(OB

Tq,

as above.

189. In words,

ing,

if

we wish

to pass

from the vector a

OA to the line OB, we are at liberty either,

from OA to OA and then to end by stretching, from OA

from the

line

serve to illustrate; or, Ilnd, to begin

HAMILTON

ELEMENTS OF QUATERNIONS,

by

stretching,

to the vector

)3,

or

1st, to begin by turn

to OB, as fig.

from OA

to

OB

48

may

and end
Z

ELEMENTS OP QUATERNIONS.

170

by turning, from OB

to OB.

The

[II.

11.

act of multiplication of a line a

by a quater
length and direction

considered as & factor (103), which affects both

(109), may thus be decomposed into two distinct and partial

acts, of

which we have called Version and Tension

acts

nion

i.

q,

and these two

the kinds

may be per
formed, at pleasure, in either of two orders of succession. And although, if we
attended merely to lengths, we might be led to say that the tensor of a quater
;

nion was a signless number* expressive of a geometrical ratio of magnitudes,

when the recent construction (fig. 48) is adopted, we see, by either of the

yet

two resulting expressions (188) for Tq, that there is a propriety in treating
this tensor as & positive scalar, as we have lately done, and propose systemati
cally to do.

190. Since

TKq

= Tq, by 187,

expressions

(12.),

U?, by 158, we may

conjugate, the two connected

UK?

and

any quaternion and

write, generally, for

its

I.

= Tq.TJq;

II.

.Kq = Tq:Uq

whence, by multiplication and division,

..q.Kq=

III.

(TqY

IV.

Kq

= (Uq)\

formula had occurred before; and we saw (161) that in it the

2
In like manner (comp.
parentheses might be omitted, because (U^) =
161, (2.) ), we have also
= Tq\
This

last

TJ(<f).

or in words, the tensor of the square of a

parentheses being again omitted
as appears (among
is
the
to
quaternion
square of the tensor
always equal
other ways) from inspection of fig. 42, bis [p. 141], in which the lengths of
;

OA, OB, oc

form a geometrical progression

OB
"

OA

At

the same time,

nions,

we

rn

OA

whence

T.OO
T.OA

/T.OB =
T.OA;

OA

qKq of two conjugate quater

common Norm, and denoted by

see again that the product

which has been called (145,

(11.)

their

the symbol Nq, represents geometrically the square of the quotient of the lengths
of the two lines, of which (when considered as vectors) the quaternion q is
itself

the quotient (112).

Y.
*

We may

2
qKq = T? = N?

Compare the Note,

therefore write generally,!

in page 111, to Art. 109.

VI.

T? -

=
y"

N<?

/ (q~Kq).

t Compare the second Note in page 130.

PRODUCT OR QUOTIENT OF TWO QUATERNIONS.

ABTS. 189-191.]
(1.)

We

have

by

also,

the tensor of a quaternion

YIL

171

the following other general transformations for

II.,
:

Tq = Kg. U?;

VIII.

Tq = U? K?;
.

which the geometrical significations might easily be exhibited by a dia

gram, but of which the validity is sufficiently proved by what precedes.
of

(2.)

(3.)

Also (comp. 158),

The

may now

We may

Tf

of that reciprocal,

(4.)

and the conjugate*

reciprocal of a quaternion,

be thus expressed

N?

_ ~

U? Tq

Tq

id

_L

J-

Nq

Tq

E?

Tq

tfy

_i

Kq

Tq

also write, generally,

IX.

Kg

= Tq

= Nq:q.

KUq

191. In general, let any two quaternions, q

and

be considered as multi

plicand and multiplier, and let them be reduced (by 120) to the forms j3 a
and j |3 then the tensor and versor of that third quaternion, y a, which is
:

(by 107) their product q

I.

II.

where

Tq>q

UY?

T/^ and

= T(y
=

U( 7

"Uq

q,

a)

a)

may

be thus expressed

Ty Ta = (T7

U7 U

= (U
7

T|3)

U/3)

(T/3

(U/3

Ta) = Tq

Tq

Ua) =

q are written, for simplicity, instead of T(q. q)

Hence, in any such multiplication, the tensor of the product

is

andllf^.

q).

the product of the

tensor
and the versor of the product is the product of the versors the order of
the factors being generally retained for the latter
(comp. 168, &c.), although it
may be varied for the former, on account of the scalar character of a tensor.
;

In

like manner, for the division of

any one quaternion /, by any other
have the analogous formulae

<?,

we

III.

T(q

-.

q)

=Tq

:Tq;

or in words, the tensor of the quotient of

* Compare

IV.

II (? :,?)=

Uq

any two quaternions

is

\Jq

equal to the

Art. 145, and the Note to page 128.

Z 2

ELEMENTS OF QUATERNIONS.

172
quotient of the tensors

and

[11.

11

i.

similarly, the versor of the quotient is equal to the

And

because multiplication and division of tensors are

performed according to the rules of algebra, or rather of arithmetic (a tensor
being always, by what precedes, a positive number], we see that the difficulty
quotient of the versors.

(whatever it may be) of the general multiplication and division of quaternions is

thus reduced to that of the corresponding operations on versors for which
:

latter

operations geometrical constructions have been assigned, in the ninth

section of the present Chapter.

(1.)

in

The two

two different

orders,

qq and

any two quaternions taken as factors

are equal or unequal, according as those two factors are

products,

qq

of

because such equality (169), or inequality (168), has

been already proved to exist, for the case* when each tensor is unity but we

complanar or diplanar

have always (comp. 178),

Tq
(2.)

Lq = Lq =

If

= Tqq

then qq =

|,

L q q = L qq

and

Kq q

(170)

so that the products of

two

right quotients, or right quaternions (132), taken in opposite orders, are always

conjugate quaternions.

(3.)

L q = L q = -,

If

Ax

and

q J_

Ax

then

q,

qq

= - q q,

4i

Ax

L qq = L q q = ^,
4i
so that the product of

tivo

q q _L

Ax

Ax

q,

when

the order of the factors

In general,
round Ax q, from
(4.)

(5.)

tensor,

Under

if

and

Ax

q to

is

reversed

angle, as q

Ax

q (171)

is

to its

is

a third

own oppo

changed
= k = -ji
ij
(182).

we had

be any two diplanar quaternions, the rotation

Ax
,

and
as

q q, is positive (177).

the same condition, q(q :q)

and same

q q _L

right quaternions, in two rectangular planes,

right quaternion, in a plane rectangular to both;

site,

but with a

is

a quaternion with the same

different axis

and

this

new

axis,

Ax q(q q), may be derived (179, (1.)) from the old axis, Ax q by a conical
rotation (in the positive direction) round Ax q, through an angle = 2 L q.
.

(6.)

The product

or quotient of

two complanar quaternions

is,

in general,

but if they be both scalar, or both

this
then
or
product
right,
quotient degenerates (131) into a scalar.
a third quaternion complanar with both

* Compare the Notes

to pages 148, 150.

PRODUCT OR QUOTIENT OF TWO QUATERNIONS.

ARTS. 191, 192.]

173

and q be complanar or diplanar, we have always as in

algebra (comp. 106, 107, 136) the two identical equations:
"Whether

(7.)

...(
Also,

(8.)

<?).?

by 190, V., and 191,

VII.

or in words, the

192. Let q 1

q q

VI. ..(q

<?

we have

I.,

.<f)

= q.

:q

formula

this other general

norm of the product is equal to the product of the norms.

=
a, and q
y ]3, as before then
/3
:

= 1

a)

= a

(a

so that the reciprocal of the product of

J3)

(1

q)

any two quaternions

duct of the reciprocals, taken in an inverted order

I.

y)

(/3

is

(1

q)

equal to the pro

or briefly,

Hq q = E? E/,
.

be again used (as in 161, (3.) ) as a (temporary) characteristic of recipro

cation.
And because we have then (by the same sub-article) the symbolical

if

equation,
q

KU =

UE,

or in words, the conjugate of the versor of

any quaternion

equal (158) to the versor of the reciprocal of that quaternion

is

versor of a product

is

equal (191) to the product of the versors

KUq q

UE? ?

UE? UE?
.

we

while the
see that

KUq KUq.
.

But

Kq
by 191

we

formula

= Tq. KUq, by 190, IX.

and Tqq = Tq. Tq = Tq Tq

.

arrive then thus at the following other important

and general

II.

..Kq

Kq.Kq

any two quaternions

an inverted order.

or in words, the conjugate of the product of

product of the conjugates, taken

(still)

in

is

equal to the

be illustrated, for versors (Tq = Tq = 1),

by the consideration of a spherical triangle ABC (comp. fig. 43 [p. 144] ) in
which the sides AB and BC (comp. 167) may represent q and q , the arc AC then
(1.)

These two

results, I., II.,

may

For then the new multiplier E? = Kq (158) is represented

whence the new
(162) by BA, and the new multiplicand E? =K? by CB
=
~Kq K? is represented by the inverse arc CA, and is there
product, Hq E?
representing q

q.

fore at once the reciprocal

E?

?,

and the conjugate

and

q be right quaternions, then

and the recent formula II. becomes, l&qq = qq
(2.) If q

"Kq

Kq
,

q,

= -

of the old product q q.

q,

Kq

as in 170.

= - q (by 144)

ELEMENTS OF QUATERNIONS.

174

In general, that formula

(3.)

denote any three vectors.

then
that, as in the annexed
Suppose

a, j3,

(4.)

may

be thus written

..Z^=K^.K;

III.

where

II. (of 192)

[Il.i.gll

may

fig.

49,

we have

the two follow

ing relations of inverse similitude of triangles (118),

AOB

BOG,

BOE

DOB

Gc

and therefore (by 137) the two equations,

pa op

75

we

shall have,

by

IF

73

III.,

= K-,
^
a
o

ADOCGC AOE;

or

so that this third formula of inverse similitude is a consequence

from the other

two.
(5.)

If then (comp. 145,

(6.)

any two

and

circles,

whether in one plane or in

from any point o, on the common

space, touch one
tangent BO, two secants OAC, OED be drawn, to these two circles; the four points
of section, A, c, D, E, will be on one common circle for such concircularity is an
another at a point B

if

easy consequence (through equal angles, &c.), from the last inverse similitude.
(6.) The same conclusion (respecting concircularity, &o.) may be otherwise
geometrically drawn, from the equality of the two rectangles, AOC and DOE,
each being equal to the square of the tangent OB which may serve as an
instructive verification of the recent formula III., and as an example of the

and

which calculations with quaternions conduct.

be
noticed
that
the construction would in general give three
may
but that if the two triangles
circles, although only one is drawn in the figure
ABC and DBE be situated in different planes, then these three circles, and of

consistency of the results, to

(7.) It

course the five points ABCDE, are situated on one

193.

An

common

sphere.

important application of the foregoing general theory of

Multi

and Division,

is the case of Right Quaternions

(132), taken in con
nexion with their Index- Vectors, or Indices (133).
Considering division first, and employing the general formula of 106, let

plication

|3

and 7 be each

-L a

right quotients, q =
planarities, /3

|||

]3,

|3

7,

and

a,

and 7

and 7 be the respective indices of the two

=
q
7 a. We shall thus have the two com-

let /3

and
|||

|3,

7 (comp. 123), because the four

lines

/3,

7,

TWO EIGHT QUATERNIONS.

CASE OP

ABTS. 192-194.]

175

7 are all perpendicular to a and within their common plane it is easy to

see, from definitions already given, that these four lines form & proportion of
vectors, in the same sense in which a, ]3, y, S did so, in the fourth section of
j3

the present Chapter

In

so that

we have (by

fact,

= T/3

Tj3

we may

write the equation of quotients,

133, 185, 187) the following relations of length,

T 7 =T 7 :T,

Ta,

and

while the relation of directions, expressed

U (/
is

easily established

L (i

We

by means
:

(IB

/3)

/3),

(1.)

tivo

/3

=
)

T( 7

/3)

by the formula,

U7

or

U7

U/3 =

U(3,

of the equations,

Ax. (/: 7 ) = Ax

0)

U.

that

Theorem (com p. again 133)

arrive, then, at this general

Quotient of any

i,j,

= L

y)

U (7

J3

T(/

.-.

Right Quaternions

is

equal

to the

For example (comp. 150, 159, 181), the

Quotient of their

u the

Indices"*

indices of the right versors

k are the axes of those three versors, namely, the lines

OK

01, QJ,

and we

have the equal quotients,

j
(2.)

In

like

= 01

oj = k = oj

manner, the indices of i

(3.)

- j = oj

01

i,

01,

- k are

-j,

= k = 01

&c.

In general the quotient of any two right

oj

01

oj

OK

and

&c.

versors is equal to the quotient

of their axes; as the theory of representative arcs, and- of their poles,

may

easily

serve to illustrate.

As

regards the multiplication of two right quaternions, in connexion

with their indices, it may here suffice to observe that, by 106 and 107, the
194.

product 7

a = (7

]3)

(]3

is

equal (comp. 136) to the quotient, (j

|3)

(a

j3);

We

*
have thus a new point of agreement, or of connexion, between right quaternions, and their
index-vectors, tending to justify the ultimate assumption (not yet made), of equality between the former
and the latter [see 290]. In fact, we shall soon prove that the index of the sum (or difference), of any

and shall find

right quotients (132), is equal to the sum (or difference) of their indices [see 206]
convenient subsequently to interpret the product fra of any tivo vectors, as being the quaternionafter
product (194) of the tivo right quaternions, of which those two lines are the indices (13)
which, the above-mentioned assumption of equality will appear natural, and be found to be useful.
(Compare the Notos to pages 121, 137). [In 198 the notation Iq is proposed as an abridgment of

two

it

"

Index of

?."]

ELEMENTS OP QUATERNIONS.

176

whence

it is

is

to

equal

[II.

i.

11, 12.

the product, q q, of any two Right Quaternions,

easy to infer that
the Quotient of the Index of the Multiplier, q divided by the Index of
"

the Reciprocal of the Multiplicand,

q"

It follows that the plane, whether of the product or of the quotient of

two

and therefore

also

right quaternions, coincides with ike plane of their indices

with the plane of their axes

because

we have,

generally,

by principles already

established, the transformation,

if

L q = o, then Index of

SECTION

q=Tq Ax
.

q.

12.

On the 8um

or Difference of any two Ctuaternions ; and on the

Scalar (or Scalar Part) of a Quaternion.

The Addition

195.

of

q considered as a geometrical
other
any
given quaternion q, considered also
as a fraction, can always be accomplished by the first general formula of

any given quaternion

quotient or fraction (101), to

when

Art. 106,*

these two fractions have a

common denominator

and

if

they

be not already given as having such, they can always be reduced so as to have
And because the addition of any two lines
one, by the process of Art. 120.

was early seen to be a commutative operation (7, 9), so that we have always
=
7 + j3 |3 + y, it follows (by 106) that the addition of any two quaternions
is likewise a commutative operation, or in symbols, that
I.

SUM

so that the

at least

two"f

+ q = q + q

Quaternions has a Value, which

is

independent of

and which (by what precedes) must be considered to be given, or

It
known, or definite, when the two surnmand quaternions are given.

their Order

is

of any

easy also to see that the conjugate of any such sum

is

equal to the sum- of the

conjugates, or in symbols, that

II.

The important formula

+
(<?

q)

EY

4-

Kg.

becomes geometrically evident,

(1.)
when it is presented under the following form. Let OBDC be any parallelo
gram, and let OA be any right line, drawn from one corner of it, but not
* [This formula
f

is

a definition.]

found [in 207] that this result admits of being extended to the case of three (or more}
but, for the moment, we content ourselves with two.
[As an example of non-commu

It will be

quaternions

last written

tative addition contrast Art. 180 (3.)]

SUM OF TWO QUATERNIONS.

ARTS. 194-196.]

177

Let the three other corners, B, c, D, be reflected (in

generally in its plane.
the sense of 145, (5.) ) with respect to that line OA, into three new points,
B

or let the three lines OB, oc, OD be reflected (in the sense of 138)

with respect to the same line OA

DD

which thus

bisects at right angles the three

Then each of
,
fig. 36 [p. 115].
joining lines, BB cc ,
the lines OB, oc, OD, and therefore also the whole plane figure OBDC, may be
,

as

it

does BB in

considered to have simply revolved round the line OA as an

rotation

axis, by a conical
and consequently the new figure OB D C like
one OBDC, must be a parallelogram. Thus (comp. 106, 137), we have

through two right angles;

that old

OD = oc + OB

Simple

(2.)

$=/+

and the recent formula

]3

(/

a)

(|3

a)

II. is justified.

as this last reasoning

draw any new diagram

to illustrate

is,

it,

and unnecessary

the reader

invited to the great simplicity of expression,

as it appears to be to

attention

with which

may be
many

once more

important

geometrical conceptions, respecting space of three dimensions, are stated in the

present Calculus and are thereby kept ready for future application, and for
easy combination with other results of the same kind.
Compare the remarks
:

already

made

in 132, (6.)

145, (10.)

161

179,

(3.)

192,

(6.)

of the shortly following sub-articles to 196, respecting

properties of

and some
an

oblique

cone with circular base.

196.

One

summands, q

most important cases of addition, is that of two conjugate

Kq of which it has been seen (in 140) that the sum is

of the

and

We

propose now to denote the half of this

thus writing generally,

always a

scalar.

symbol,

Bq

I.

or defining the
II.

new symbol
.

Sq = I

For reasons which

(q

S<?

q +

Kq

Kq

+ q = 2Sq

sum by

the

by the formula,

+ Kq)

or briefly, IF.

S = J

(1

+ K).

more fully appear, we shall also call this new

and
quantity, Sq, the scalar part, or simply the SCALAR, of the Quaternion, q
shall therefore call the letter S, thus used, the Characteristic
the
of
Operation
will soon

of taking the Scalar of a quaternion.

(Comp. 132, (6.) 137 156 187.)
It follows that not only equal quaternions, but also
conjugate quaternions, have
;

equal scalars

or in symbols,

III.

S? = Sq,

if
<?

=?;

and

IV.

SK?

= Sq

or briefly,

IV.
HAMILTON

ELEMENTS OF QUATERNIONS.

SK

S.
2

ELEMENTS OF QUATEENIONS.

178

i.

[II.

12.

because we have seen that ~Kq = + q, if q be a scalar (139), but that

=
K0 - q, if q be a right quotient (144), we find that the scalar of a scalar (con
sidered as a degenerate quaternion, 131} is equal to that scalar itself, but that the

And

We may therefore

scalar of a right quaternion is zero.

V.

Sa?

x be a scalar

x, if

and

VII.

Again, because OA in
plied thereby,

we may

is

|.

multiplied by

when OB

x,

is

multi

S#0 = #S0,

x be any scalar

if

in particular (by 188),

IX.
Also because

Zq =

if

0,

115]

[p.

S 2 = SS = S

write, generally,

VIII.

and therefore

36

fig.

write (comp. 160)

VI. .. SS0 = S0,

S0 -

..

now

Sq = 8 (Tq

SKq = S0, by IV.,

while

= Tq

11 q)

SU0.

U -, by 158, we have the general

KIJ0 =

equation,

XQTT
ou
.

1
dTT
= o
u

tt

-V
QTT ft = QTT
A....OU
ioU^5

or

et

!?

JO

whence, by IX.,

XL .S0-T0.SU-;
and therefore

"VTT
A.11.

C!
.

J$q

Q
o

T~i

JLq

Tq T - =
.

"NT

S^

-&.11

T^.SU
a

C!
.

Jo

AT
IN

I.

Q
o

combined with the recent definition

results of 142,

3;

1,

"VTT

=JN0.b
q

The

(V.), since

by 190,

also,

XI

or

73.

or II., enable us to

extend the recent formula VII., by writing,

XIII.

S0

>,

=, or

<

0,

according as L q

^,

according as S0

<>

=, or

>

and conversely,

XIV.
En fact,
fig.

36,

if

we

z.

we compare

<,

=, or

>

that definition

=, or

<

0.

with the formula of 140, and with

see at once that because, in that figure,

S (OB

we may

I.

>,

OA)

= OA

OA,

write, generally,

XV.

S0 = T0

cos

z0

or

XVI.

SU0 = cos/0;

SCALAR OF A QUATERNION.

ART. 196.]

179

equations which will be found of great importance, as serving to connect

quaternions with Trigonometry and which show that
;

XVII. ..Ltf=Lq,
the angle L q being

and

?r;

whence

still

taken

S U? = SU0,

if

not to

(as in 130), so as

fall

outside the limits

also,

XVIII. ..LJ=Lq,

T/ =

and

S/=S?,

if

Ty,

the angle of a quaternion being thus given, when the scalar and the tensor of that
quaternion are given, or known. Finally because, in the same figure 36

(comp. 15, 103), the

line,

OA = (OA

OA)

OA = OA

S (OB

OA),

be said to be the projection of OB on OA, since A is the foot of the perpen

dicular let fall from the point B upon this latter line OA, we may establish this

may

other general formula

XIX.

aS

= S -

= projection of (3 on a

a result which will be found to be of great utility, in investigations respecting

geometrical loci, and which may be also written thus
:

XX.

Projection of

on a =

Ua

Tj3

SU^;

with other transformations deducible from principles stated above.

It is
scalar
to
remark
on
of
the
account
character of Sq,
that,
scarcely necessary

we

have, generally,

by 159, and 187,

XXI.
while, for the

US?

the expressions,

(8.),

XXII.

TS? =

Sg;

same reason, we have always, by 139, the equation (comp.

XXIII.

KS?

= Sq

or

XXIII

KS

= S

IV.),

and, by 131,

XXIV.

S<?

0,

or

TT,

unless L q = -

tU

in

which

Sq = 0, by VII., and therefore L Sq is indeterminate:*

the same time indeterminate, by 159, but
vanishing, by

last case

US*? becoming at

TS<?

186, 187.
* Compare

the Note in page 120, to Art. 131.

2

ELEMENTS OF QUATERNIONS.

180

(1.)

The

S^=

equation,

[II.i.12.

0,

now seen to be equivalent to the formula, p _1_ a and therefore to denote

the same plane locus for p, as that which is represented by any one of the four

is

other equations of 186,

(6.)

by the equation,

or

T 1?
p-a
The

(2.)

1, of

187,

(2.).

equation,

or

expresses that BP _L OA

on OA

or that the points B

or that the locus of P

The

(3.)

(4.)

and OA

The

B,

perpendicular

projection

to the line

OA.

equation,

expresses (comp. 132, (2.)) that P

for vertex,

and p have the same

the plane through

is

for axis,

is

on one

sheet of a cone of revolution,

and passing through

other sheet of the same cone

is

with o

the point B.

represented by this other equation,

and

both sheets jointly

by the equation,

\
(5.)

The

8^

expresses that the locus of P

because

OA

it

itself ,
(6.)

expresses (comp.

is

or

1,

SU^
a

= T-,
p

the plane through A, perpendicular

XIX.)

or that the angle GAP

On

equation,

is

to the line

that the projection of OP on OA

right

or that

S-

is

OA;

the line

0.

the other hand the equation,

=
,

or

expresses that the projection of OB on OP is OP itself; or that the angle OPB is

right; or that the locus of P is that spheric surface which lias the line OB for

a diameter,

GEOMETRICAL EXAMPLES.

ART. 196.]
(7.)

Hence

the system of the

tico

S? =
a

represents the

the plane
(8.)

(5.), witli

And

equations,

S^-l,

l,

in which the sphere

circle,

new

with OB for a diameter,

on it from o.

(fj.),

OA for the perpendicular

therefore this

181

is

cut

by

let fall

equation,

obtained by multiplying the two last, represents the Cyclic* Cone (or cone of
the second order, but not generally of revolution], which rests on this last circle

and has the point o for its vertex.

In fact, the equation
evidently satisfied, when the two equations (7.) are so and therefore
every point of the circular circumference, denoted by those two equations, must
as its base,

(7.)

(8.) is

be a point of

the locus,

represented by the equation

tion remains unchanged, at least essentially,

any scalar

the locus

(8.) is,

But the

(8.).

when p

is

changed

latter

equa

to xp, x being

therefore, some conical surface, with its vertex at

and consequently it can be none other than that particular cone

(both ways prolonged), which rests (as above) on the given circular base (7.).

the origin, o

(9.)

The system

of the two equations,

s.s-i,
p
(in writing the

first

section

that

namely

s-i,
y

which the point may be omitted), represents a conic

section, in which the cone (8.) is cut by the new plane,

of

which has oc for the perpendicular

let

fall

upon

it,

from the origin of

vectors o.
every plane ellipse (or other come section] in space, of

(10.) Conversely,

which the plane does not pass through the origin, may be represented by a
because the cone which rests on
system of two equations, of this last /0m (9.)
;

any such
be a

and has

its vertex

at any given point o, is

known

to

cyclic cone.

(11.)

cone (8.)
*

conic as its base,

The
is

cut

curve (or rather the pair of curves), in which an oblique but cyclic

by a

concentric sphere (that is to say, a cone resting

Historically speaking, the oblique cone with circular base

may

deserve to be

on a circular

named

the Apollonian

from Apollonius of Perga, in whose great work on Conies (KUVIK&V), already referred to in a Note
to page 130, the properties of such a cone
although
appear to have been first treated systematically
the cone of revolution had been studied
by Euclid. But the designation cyclic cone" is shorter; and
it seerus more
natural, in geometry, to speak of the above-mentioned oblique cone thus, for the purpose
Cone,

"

of marking its connexion with the circle, than to call it, as is now usually -done, a cone of the second
order, or of the secottd degree
although these phrases also have their advantages.
:

ELEMENTS OF QUATERNIONS.

182

base by a sphere which has

[II.

12.

i.

centre at the vertex of that cone), has come, in

And any such conic may, on
times, to be called a Spherical Conic.

modern

its

the foregoing plan, be represented by the system of the two equations,

the length of the radius of the sphere being here, for simplicity, supposed to
be the unit of length. But, by writing Tp = a, where a may denote any
constant and positive scalar, we can at once remove this last restriction, if it be

thought useful or convenient

The equation

(12.)

(comp. 191, VII.)

do

to

may

(8.)

be written, by XII. or

XII

.,

under the form

8
J
or

so.

r Nr

briefly,

p
if

and

so that a

which

is

= Ta.U/3,

the cone

(8.)

on the sphere

(6.),

j3

and OB

aT^

Tj3

of Art. 188,

not only by the plane

is cut,

Ua

and
(5.)

fig.

48.

in the circle

but also by the (generally) new plane, S

which

in the (generally) neiv circle, in

sphere,

and

are here the lines OA

]3

Hence

(13.)
(7.),

=/3T|

this

or in the circle which

is

new plane

1,

cuts the (generally) new

represented by the system of the two

equations,

In the particular case when /3 a (15.), so that the quotient /3 a is a

which must be positive and greater than unity, in order that the plane

(14.)
scalar,

may

(5.)

cone

(8.)

(really] cut the

may
j3

and the
(15.)

||

be

real,

a*a,

sphere

we may
a>\,

circle (13.) coincides

In the same

circular base (that

is,

case,

(6.),

and therefore that the

circle (7.)

and the

write

(j8

with the

the cone

a)

= a2

a =

a,

/3

j3

circle (7.).
is

one of revolution

every point P of

its

of the circumference thereof) being at one constant distance

GEOMETRICAL EXAMPLES.

ART. 196.]
the vertex o,

from

equations

(7.) give,

(12.),

Conversely,

and 186,
if

:s2

(16.)

For, in the case supposed, the

by XII.,

=a
N a = S^:S-=l:Sa

(Compare 145,

= aTa.

at a distance

namely

183

To =

Or

aTa.

(5.).)

the cone be one of revolution, the equations (7.) must

conduct to a result of the form,

a*

=s
N^a = S^:Sa

S^^

or (comp. (2.)),

S-,
p

=
;

by our having ]3 = a, as
2
since otherwise we should have, by XIV., p J_ j3 - a a, and all the
in (14.)
points of the base would be situated in one plane passing through the vertex o,
which (for any actual cone) would be absurd.
which can only be by the line

]3

a* a

vanishing, or

(17.)
j3

]3,

Supposing, then, that we have not

as in (14.), nor even a

cone of revolution (or what

is

\\

a, /3

||

|3,

we

]3

||

a,

and therefore not a =

a,

see that the cone (8.) is not a

often called a right cone)

but that

it is,

on the

And we see
contrary, an oblique (or scalene) cone, although still a cyclic one.
that such a cone is cut in tico distinct series* of circular sections, by planes
mutually non-parallel) planes, (5.) and (13.)
or to two new planes, drawn through the vertex o, which have been called f the

parallel to the

two

distinct (and

two Cyclic Planes of the cone, namely, the two following

while the two

tico

from the vertex, OA and OB, which are perpendicular

lines

planes respectively,
(18.)

Of these two

of the sphere

when

is

may
lines,

which

(6.),

that cone

to these

be said to be the two Cyclic Normals.

a and )3, the second has been seen to be a diameter
be said to be circumscribed to the cone

may

considered as having the circle

(7.) for its base

(8.),

the second

plane (17.) is therefore the tangent plane at the vertex of the cone, to that
circumscribed sphere (6.).

cyclic
first

(19.)

The sphere

(13.)

in like

may

manner be

said to be circumscribed to

* These two series of

sub-contrary (or antiparallel] but circular sections of a cyclic cone, appear to
have been first discovered by Apollonius see the Fifth Proposition of his First Book, in which he
says, Ka\ela-8u Se rj roiavTi] ro^rj virevavria. (page 22 of Halley s Edition).
:

By M.

Chasles.

ELEMENTS OF QUATERNIONS.

184

the cone,

if

the latter be considered as resting on the

terminated by
or
is the line OB

new base

that circle as its

first cyclic

)3

normal

which

(17.)

lias

by

new

of this new sphere

(12.) the direction of the line a, or

(18.)

the first

12.

i.

circle (13.), or as

and the diameter

so that (comp.

[II,

cyclic

plane

of the

the

is

tangent plane at the vertex, to the second circumscribed sphere (13.).

(20.)

its

scalar co-efficient, is represented

by

V&

or

plane

circle,

of

fl

where

is

the equation,

a
S j3 =

|1

P
therefore cuts the cone in a

touches the first cyclic

diameter from the vertex = b

S--=l,
it

which

other sphere through the vertex,

and which therefore has

plane,

some

Any

77;

which (by

(12.)

the equation of the

is

S-y
so that the perpendicular

from

or

the vertex is b a

quently this plane of section of sphere

||

and cone

(comp.

|3
is

(5.)

parallel

to the

and conse
second cyclic

plane (17.).

In

(21.)

like

manner any

sphere, such as

1,

where

b is

any

scalar,

which touches
circle,

and

is

the second cyclic plane at the vertex, intersects the cone (8.) in a

which the plane has for equation,

of

therefore parallel

(22.)

written

The equation

to the first cyclic

plane.

of the cone (by IX., X.,

XVI.) may

also be thus

STj.STj
a

T^;
p

cos L

or,

P
a

cos

z-g

T;
p

the product of the cosines of the inclinations, of any

expresses, therefore, that
variable side (p) of an oblique cyclic cone, to two fixed lines (a and ]3), namely to
it

the two

cyclic

normals

inclinations, of the

namely

(17.), is constant

same variable

or that the product of the sines of the

side (or ray, p) of the cone, to two fixed planes,

to the two cyclic planes, is thus a constant quantity.

SCALAR OF A SUM OR DIFFERENCE.

ARTS. 196-197.]
(23.)

The

two great

the two cyclic planes,

circles,

185

in which the concentric sphere

Tp

= 1

cut by

is

have been called the two Cyclic Arcs* of the Spherical

Conic (11.), in which that sphere

is

by the

cut

cone.

It follows (by (22.))

that the product of the sines of the (arcual) perpendiculars, let fall from any point
p of a given spherical conic, on its tivo cyclic arcs, is constant.

These properties of cyclic cones, and of spherical conies, are not put
forward as new but they are of importance enough, and have been here
deduced with sufficient facility, to show that we are already in possession of a
(24.)

own

Calculus, with its

Rules ^ of Transformation, whereby one enunciation of a

geometrical theorem, or problem, or construction, can be translated into

several others, of which some may be clearer, or simpler, or more elegant

than the one

first

7 +

]3,

proposed.

7 be any three co-initial vectors, OA, &c., and

that OBDC is a parallelogram (6) then, if we write

197. Let
so

]3,

a =

q,

and suppose that B

points B, c, D on the
(OB =)
also

OD = 8

j3

But

let

]3

a = q,

and

D are the

we

line OA,

= aS?,

q"

= q + q
(106),

feet of perpendiculars let fall

shall have,

a =

= aSq

by 196, XIX.,
S =

OB = CD, and therefore OB = C D

= aS

aSq"

from the

the expressions,
(?

+?).

the similar projections of equal

lines

being equal; hence (comp. 11) the sum of the projections of the lines
must be equal to the projection of the sum, or in symbols,

OD = oc + OB

V= 7

|3

a = (7

a)

(/3

sum

or in words, the scalar of the

a).

Hence, generally, for any two quaternions, q and q, we have the formula
I... 8

]3,

(/+$)- 8^+80;
is

equal to the sum of the scalars.

It is easy

to extend this result to the case of

respective scalars; thus, if

{?"

(q

q)

q"

any three (or more) quaternions, with their

be a third arbitrary quaternion, we may write

=
}

S?"

+ S

(q

q)

=
80"

+ (S? + Sq)

where, on account of the scalar character of the summands, the last paren
may be omitted.
may therefore write, generally,

We

theses

II.

where

is

S2q = SS?,

or briefly,

used as a sign of Summation

*

HAMILTON

By M.

Chasles.

ELEMENTS OF QUATERNIONS.

and

SS
may

28

say that the Operation of

t Comp. 145,

(10.),

&c.
2

ELEMENTS OF QUATERNIONS.

186

taking the Scalar of a Quaternion

is

i.

12.

a Distributive Operation (comp. 13).

As

pi.

any one quaternion from any other, there is no

by the method of Art. 120, to the second general

to the general Subtraction of

difficulty in reducing

formula of 106

it,

nor in proving that the Scalar of

the Difference* is

always

In symbols,

equal to the Difference of the Scalars.

or briefly,

IY.

when A

SA

SA? = AS0,

AS

used as the characteristic of the operation of taking a

subtracting one quaternion, or one scalar, from another.
(1.)

is

It has not yet been proved (comp. 195) that the Addition of any

number of Quaternions,
(comp.

by

difference,

But we

9).

q,

is

q",

an

associative

and a commutative

sum

see, already, that the scalar of the

quaternions has a value, which

independent of their order,

is

operation

any such set of

and of the mode

of

of grouping them.
(2.)

If the

summauds be

all right

quaternions (132), the scalar of each

separately vanishes, by 196, VII. wherefore the scalar of their sum vanishes
also, and that sum is consequently itself, by 196, XIV., a right quaternion
;

a result which

it is

easy to verify.

In

fact, if j3 _L

a and 7 JL

a,

then 7 +

]3

then perpendicular to the plane of /3 and 7 hence, by 106,

the sum of any two right quaternions is a right quaternion, and therefore also the
J_ a,

because a

is

sum of any number of such quaternions.

(3.) Whatever two quaternions q and
algebra, the two identities (comp. 191, (7.)
V. ..(/-?) +

*-/;

may

be,

we have

always, as iu

VI. ..(q

+q)-q

= q

198. Without yet entering on the general theory of scalars of products or

quotients of quaternions,

we may

observe here that because, by 196,

scalar of a quaternion depends only on the tensor

of the

axis,

we

191).

It

may

the angle,

and

is

having thus always equal

XV.,

the

independent

are at liberty to write generally (comp. 173, 178,

the two products, qq and q

have been seen

and

and 191,

scalars, although they

have unequal axes, for the general case of diplanarity (168,
also be noticed that, in virtue of what was shown in 193,
q,

to

Examples have already occurred

in 196,

(2.), (5.), (16.).

SCALAR OF A PRODUCT, QUOTIENT, OR SQUARE.

ARTS, 197-199.]

187

respecting the quotient, and in 194 respecting the product, of any two right
quaternions (132), in connexion with their indices (133), we may now estab
lish, for

such quaternions, the formulse

any

III.

IV.

(q

= S (Iq

g)

Iq)

$q q = S(q .q) = S (lq

where the new symbol Tq

is

I -} = ~

f
:

(q

Tq

(Ax

cos L

Ax

cos L

q)

(Ax

q)

Ax

q)

used, as a temporary abridgment, to denote the

supposed here (as above) to be a right one. With

the same supposition, we have therefore also these other and shorter
Index of the quaternion
formulae

q,

V.

VI.

STJ

(q

q)

= q

SUq

= + cos L
(Ax
cos L

(Ax

Ax

Ax

which may, by 126, XVI., be interpreted as expressing

condition of r ect angularity of q and q

q)

q)

that,

under the same

In words,

VII.

VIII.

is

(q

q)

Ax

q)

q).

There

the Axes.

Angle of

proving these results otherwise,

;

Right Quaternions is equal to the

but the Angle of the Product, of two such quaternions,

equal to the Supplement of the

Art. 193

Ax

of the Quotient of two

the

Angle
Angle of their Axes

= L (Ax q
L q q = TT - L (Ax q
L

by
them by the consideration

nor in illustrating

is

no

constructions such as that

difficulty in

employed in

of isosceles quad-

rantal triangles,

upon the surface of a sphere.

199. Another important case of the scalar of a product

scalar of the square of a

quaternion.
[p.

133],

(q

we
,

L(q*)

by the adopted

= 2Lq,

case,

referring to Art. 149, and to fig. 42

2
(T?) , as in 190, and
)

also,

and

Ax

Ax.

however, by 196, XVI., we

III.

when

/ q is 0,

IV.

SU

(?)

(q

of

= -

may

= cos L

Ax

and

(?)

SU (f)

a formula which holds even

(f)

definitions of L q (130),

II...L(f) = Z(v-Lq\
In each

On

the case of the

we have always T(? 2 =

= (U?) 3 as in
161, we have
I.

but,

see that while

is

if

q,

Ax

Ax.

*-<!<%

q (127, 128),
if

?,

Lq>-

write,

= cos 2 L q

or -, or

= 2 (SU?) 2 -

TT,

and which

gives,

1.

2B

ELEMENTS OF QUATERNIONS.

188

Hence, generally, the scalar of

forms

[II.

may be put under either of

(f

12.

i.

the two following

V.

(</

where we see that

= T? 2

it

cos2z

VI.

q;

would not be safe to omit

(?

this last

(S,?)

T<f

the parentheses,

without some

2
without first deciding
made, and to write simply
be
shall
to
scalar of the square,
understood
the
symbol
signify

convention previously

whether

= 2

S<?

or the square of the scalar of q

The

these two things being generally unequal.

them, however, occurring rather oftener than the former, it appears

convenient to fix on it as that which is to be understood by S*/ 2 while the other
latter of

occasionally be written with a point thus, S

conventions respecting notation* we may write

may

q*

and then, with these

VII.

But the

(S?)

f=

is

any quaternion

(<?

).

easily seen to be the

we have generally (comp.

so that

VIII.

square of the conjugate of

conjugate of the square

formula

2
S? =

the

190, II.)

IX.

Kq

(q

= (Kq) z =

iy Ug
:

2
.

(1.)
quaternion, like a positive scalar, may be said to have in general
two opposite square roots because the squares of opposite quaternions are always
;

But of these two roots the principal (or simpler) one, and
equal (comp. (3.)).
that which we shall denote by the symbol
or yq, and shall call by
</q

eminence

the

We

obtuse.

Square Root of

q,

is

that which has

its

and not

angle acute,

shall therefore write, generally,

X.

L \/q = J L q

Ax ^q
.

with the reservation that, when L q = 0, or = TT, this

becomes (by 131, 149) an indeterminate unit-line.

Ax

common

axis of q

and \/q

Hence,

(2.)

XL..S-/0>0,

if

Lq<Tr\

while this scalar of the square root of a quaternion may, by VI., be thus trans

formed

XII.

a formula which holds good, even at the limit L q ~

IT.

2
2
*
while d (#2 ) is some
As, in the Differential Calculus, it is usual to write d# instead of (da;)
But as d% denotes a second differential, so it seems safest not to denote the
times written as d x*.
2
\vhich properly signifies SSg, or S^, as in 196, VI. ; the second
square of Sq by the symbol S
;

<7,

scalar (like the second tensor, 187, (9.), or the second versor, 160) being equal to t\\& first.
and the employment of the notation S 2
calculator will of course use his own discretion
;

as cos2 #

is

often written for (cos#) 2 ,

may

sometimes cause a saving of space.

Still

for
<?

every

(j)

TENSOR AND NORM OF A SUM.

ARTS. 199-200.]

The

(3.)

189

principle* (1.) that, in quaternions, as in algebra, the equation,

XIII.

= q\

(- 7)

an identity, may be illustrated by conceiving that, in fig. 42, a point B is

determined by the equation OB = BO for then we shall have (comp. fig. 33,

is

bis [p. 122]),

X2

OA
(os

oc

2
<7

OA

A AOB

because

B OC.

oc

200. Another useful connexion between scalars and tensors (or norms) of

In any plane triangle AOB, we havef

quaternions may be derived as follows.

the relation,
2
2 2 (T OA)
(T AB) = (T OA)
.

in

which the symbols

OA, &c.

therefore

but

we

if

by (T

OA)

I.

II.

(T OB)
.

cos

AOB+ (T

os)

OA, &o., denote (by 185, 186) the lengths of the sides

write q = OB

still
2

OA,

we have

= AB

OA

dividing

the formula becomes (by 196, &c.),

T (q-

1)

= 1 -

2T<?

SU? + T?

=
T<f

2S? +

or

But

sign,

and

we may

therefore change

write,

since

T(1 +

it is

2
<?

+ 2S? + T2 2

IV.

N(l + 0)

easy to prove (by 106, 107) that

whatever two quaternions q and q

VI.

we

1.

here a perfectly general quaternion

is

III.

And

.N(?-1) = N?-2S? +

may

be, while

.S^.N^ = S./K^=S

easily infer this other general formula,

VII.
which gives,

if

x be any

VIII.
*

N (q

q)

N/+

2S qKq + TSq
.

scalar,
.

(q

+ x) =

N? + 2xSq + x\

Compare the Note to page 162.

By the Second Book of Euclid,

or

hy plane trigonometry.

its

ELEMENTS OF QUATERNIONS.

190

[II.

i.

12.

We are

passages

now prepared to effect, by rules* of transformation, some other

from one mode of expression to another, of the kind which has been

alluded

to,

(1.)

and partly exemplified, in former

Take, for example,

sub-articles.

the formula,

- a

p
or the equivalent formula,

T(P +

a)

of 186, (6.);

TG>-),

which has been seen, on geometrical grounds, to represent a certain locus,

namely the plane through o, perpendicular to the line OA and therefore the
;

same

locus as that

which

is

represented by

S^=0,of
To pass now from

now

cited,

by

II.

(1.),

by

and

q,

by

calculation,

we have

to observe that the first or second

becomes then,

N (q + 1) = N (q-1)

or

Tte+l)=Tfe-l);
or finally,

196,

the former equations to the latter,

only to denote the quotient p

form, as just

the equation

and IV.,
&? =

0,

which gives the third form of equation, as required.

(2.)

- =
Conversely, from S

0,

we can

by the same general

return,

for-

cx

mulse II. and IV., to the equation

to

T
\a

+
\

or to

N
-

-- 1

a)

=
)

(p

^+

or to

a),

or

by
=

T
p-a

I.

and

1, as

III.

above;

and generally,
S? =

gives

T(?-1)=T(2 +

1),

or

= l;

while the latter equations, in turn, involve, as has been seen, the former.
(3.)

Again,

if

we

take the Apollonian Locus, 145,

the first of the two forms 186,

T
*

Compare

(5.) of its equation,

(p

145, (10)

= aT

(p

(8.), (9.),

namely,

a),

and several subsequent

sub-articles.

and employ

GEOMETRICAL EXAMPLES.

ART. 200.]

where a

is

a given positive scalar different from unity,

T(q-a*) = aT(q-l),
or

by VIIL,

N?

of

2a*8q +

a*

(Nq

we may

2Sg +

1)

N (q -

write
1)

it

as

- 2a
Sq, transposing, and dividing by c?
2
2
= fl
= Na
= Ta
or,
or,
N/o

last is the second

form 186,

Tp

<7

N<7

calculation alone,

N (q - a?)

or as

or, after suppressing

which

191

and

1,

thus deduced from the

is

first, by
without any immediate appeal to geometry or the construction
(5.),

any diagram.
(4.)

we take the

if

Conversely

N^

equation,

2
,

of 145, (12.),

which was there seen to represent the same locus, considered as a spheric
= 2
surface, with o for centre, and aa for one of its radii, and write it as
,

N<?

we can then by

N (q -

2
fl

or finally,
GO

N (q 2

a)

TP

(5.)

= aTa, or

It

is

P =
a

1),

or

= aT ( P -

this first form of that sub-article

from

form

calculation return to the

),

(q

as in 186, (5.)

(q

1),

being thus deduced from the second, namely

a.

from being the intention of the foregoing remarks,

far

to dis

courage attention to the geometrical interpretation of the various forms of expres

and general rules of transformation, which thus offer themselves in working

with quaternions on the contrary, one main object of the present Chapter
has been to establish a firm geometrical basis, for all such forms and rules.

sion,

But when such a foundation has once been laid, it is, as we see, not necessary
we should continually recur to the examination of it, in building up the

that

superstructure.

That each

of the two forms, in 186, (5.), involves the other

be proved, as above, by calculation

meaning of this result

to the

and in seeking

The

interesting to inquire

to interpret

it,

(q-

(q

be expressed under the form of an

IX.
which q

may

(q

N<?

= a\

identity, as follows

- N?) =

be any quaternion.

1), if

N?

(q

what

may

is

the

we should be led anew

result (4.) of calculation, that

in

it is

theorem of the Apollonian Locus.

(6.)

may

but

1)

ELEMENTS OP QUATERNIONS.

192
(7.)

Or,

191, VII., because

by

q (q

we may

write

If

(q

1)

1,

follows from the latter

as in algebra,

q,
:

..N(?-N<?)=N

then S - =

(f-q).

and conversely, the former equation

because each

may

be put under the form (comp.

~Nq=2$q.

196, XII.),
(9.)

1)

12, 13.

Jt

will soon be seen that

it

= qz -

as this other identity

it

X.
(8.)

pi.

Hence,

if

(p

a)

= Ta, then S

and

1,

reciprocally.

In

fact

(comp. 196, (6.) ), each of these two equations expresses that the locus of P is
or which has
the sphere which passes through o, and has its centre at A
;

OB = 2a for a diameter.
(10.)

By

changing
if

(11.)

Hence

Tq =
if

we

q to q + 1 in (8),

1,

then

Tp = Ta, then S

--

find that

0,

- =

p + a

0,

and

reciprocally.

and

reciprocally; because (by

106)

p + a
(12.)

Each

p + a

Ip

-i\

\a

(P

\a

two equations (11.) expresses that the locus of P is

which has its centre at o and their proved agreement

of these

the sphere through A,

a recognition, by quaternions, of the elementary geometrical theorem, that
the angle in a semicircle is a right angle.
;

is

SECTION

On

13.

the Right Part (or Vector Part) of a Quaternion ; and on the

IHstriVutive Property of the Multiplication of Quaternions.
201.

A given

vector

OB can always be decomposed, in one but in only one

way, into two component vectors, of which it is the sum (6)

OB in fig. 50, is para/lei (15) to another given Bvector OA, while the other, as
in the same figure, is
one, as

and

of

which
~~ _._,B

OB"

perpendicular to that given line OA namely, by letting

fall the perpendicular BB on OA, and drawing OB" = B B,
;

so that OB

a and
it

is

shall be a rectangle.
In other words, if
be
two
and
co-initial vectors,
)3
any
given, actual,
always possible to deduce from them, in one definite way, two other
BB"

EIGHT PART OF A QUATERNION.

and

co-initial vectors, (3

which

which need not however loth be actual

(1)

and

shall satisfy (comp. 6, 15, 129) the conditions,

when

ft vanishing,

(what we

may

functions,

it is

and that
is

j3",

193

call)

and

being null, when

]3"

]3

determinate vector-functions of a and

||

but both being

And

]3.

of these

two

evident that ft is the orthographic projection of ft on the line a ;

is the corresponding projection of
ft on the plane through o, which

j3"

perpendicular to

202.

|3

Hence

a.

it is

easy to infer, that there

always one, but only one way,

is

of decomposing a given quaternion^

q

= OB

OA =

ft

a,

summands (195), of which one shall be, as in 196, a scalar,

while the other shall be a right quotient (132).
Of these two parts, the former
has been already called (196) the scalar part, or simply the Scalar of the
into two parts or

Quaternion, and has been denoted by the symbol S^

to the recent figure 50,
I.

so that, with reference

we have

Sq = S (OB

OA)

= OB

OA

or,

And we now

propose to call the latter part the

quaternion, and to denote it by the new symbol

(]3

a)

= ft

a.

EIGHT PART*

same

of the

v?;
same

writing thus, in connexion with the

II.

V? =

V (OB

=
OA)

OB"

figure,

OA

V (/3

or,

a)

a.

ft"

The System of Notations, peculiar to the present Calculus, will thus have been
completed and we shall have the following general Formula of Decomposition
;

of a Quaternion into two

kinds

Summands (comp.

188), of the Scalar

and Right

III.
or, briefly

= 80 + V? = V? + S?,

= S +

and symbolically,
IV.

(1.)

In connexion with the same

V
because, by construction, B B =

(OB

V=Y+

fig.

OA)

50,

= B B

S.

we may
:

write also,

OA,

OB".

* This
Right Part, V#, will come to be also called the Vector Part, or simply the VECTOR, of the
Quaternion; because it will be found possible and useful to identify such part with its own IndexVector (133). Compare the Notes to pages 121, 137, 175 [and Art. 286].

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

194

In

(2.)

manner, for

like

V
Under

(3.)

In general,

V.

More

(5.)

VI.

:a)

Sq =

if

0,

and

0,

OA.

a)

(/3"

and

0,

S/ =

if
<?,

Also

(6.)

equation,

0.

evident that

it is

= AB
OA)

we have the

and

V<?

reciprocally.

generally,
f

(OB

[p. 115],

13.

i.

the recent conditions,

V(/3
(4.)

36

fig.

[II.

VII.
VIII.

or

V?

and

S<?,

=
Q3

V</

Lq =

if

0,

if

0,

with the converse.

V</

or

0,
|3

||

TT

the right part of a scalar being zero.

On

(7.)

the other hand,

K...VJ-J,
a ngr^ quaternion being
203.

We had
L

(196,

XIX.)

a formula which

OB = S (OB

OA)

OB"

V (OB

perpendicular

articles

to

(comp. 186,

this perpendicular

we must
we were

Tj3

the characteristic

the factor,

(6.),

/x

III.

or

OA,

may now
|3

- S

new symbol Vg,

OA)

OA,

or

be written thus,
a,

a
j3

on a

and we have

the analogous formula,

=
j3"

V^.a,

on the plane (through o) which is drawn so as to

a; and which has been considered in several former sub-

to express the projection of

\>Q

OB on OA, or of the vector

evidently, by the definition of the

^=

zY

to express the projection of

II.

if

|3

and 196,
=

TV

(1.)).

It follows (by 186, &c.) that

Ta = perpendicular

distance of E

from OA

being here considered with reference to its length alone, as

of the tensor implies.
It is to be observed that because

in the recent formula II. for the projection

j3",

is

not a scalar,

write that factor as a multiplier, and not as a multiplicand

at liberty, in

although

consequence of a general convention (15), respecting the

GEOMETRICAL EXAMPLES.

ARTS. 202-203.]
multiplication of vectors

and

195

denote the other projection

scalars, to

|3

under the

form,
T. ..j3
(1.)

The

(2.)

(196,

XIX.).

equation,

expresses that the locus of P

The

=aS^

is

the indefinite right

line

OA.

equation,
or

expresses that the locus of p

is

the indefinite right line

drawn through the point B, parallel to the line OA.
is

BB",

in fig. 50,

which

The equation

(3.)

S
a

or

0,

S
a

8,ofl96,
a

(2.),

seen to express that the locus of p is the plane through B, perpendi

if then we combine it with the recent equation (2.), we
cular to the line OA

has

"been

shall express that the point P is situated at the intersection of the

with the point B.

(4.) Accordingly, whether we take the two first or the two
recent forms (2.), (3.), namely,

mentioned

loci

or that

it

we can

case

^? =
a

0,

whence p

inferring, through 202, VI., that

by

last

last of these

or

0,

infer this position of the point p

through 202, V., that

two

coincides

j3

ct

in the

0,

= a

case

first

by 142
so that

by

inferring,

and in the second

we have in each

(comp. 104), or as a consequence from each system, the equality p =

OP = OB or finally (comp. 20) the coincidence) p = B.

case

/3,

or

(5.)

The equation

expresses that the locus of the point P

is

the cylindric surface of revolution, which

for it expresses, by
passes through the point B, and has the line OA for its axis
that
the
P
and
of
this
distances
latter line, are equal.
III.,
B,/r0w
perpendicular
;

2 c 2

ELEMENTS OF QUATERNIONS.

196
(6.)

The system

[II.

ay

a
expresses that the locus of P

the (generally)

is

elliptic section

of the cylinder

made by the plane through o, which is perpendicular to the line

(7.) If we employ an analogous decomposition of p, by supposing

(5.),

= p + p
.

the three rectilinear or plane

thus briefly written

//

oc.

that

"i

ii

JL a,

||

loci, (1.), (2.),

(3.),

may

have their equations

P "=0;

p"=0";

=p:

while the combination of the two last of these gives p =

(8.)

13.

two equations,

of the

i.

The equation

/3,

as in (4.).

of the cylindric locus, (5.), takes at the

same time the

form
V-T/3";

which

of the point p, on the

equation expresses that the projection
plane through o perpendicular to OA, falls somewhere on the circumference of
a circle, with o for centre, and OB" for radius and this circle may accordingly
last

p"

be considered as the base of the right cylinder, in the sub-article last cited.
204. From the mere circumstance that ~Vq is always a right quotient (132),
whence UV<? is a right versor (153), of which the plane (119), and the axis
(127), coincide with those of

Thus we have

by

generally,
I.

III.

q,

several general consequences easily follow.

principles already established, the relations:

V?

=
&

II.

KVq = - V?,

IV. ..SV? =

0,

or

Ax.V? =

KV = - V (144J

or

SV = 0(196,

VII.)

V... (UV?) = - 1(153, 159);

2

and

therefore,

because,

VI.

(V?)

= - (TV?) 2 = -

NV<?,*

by the general decomposition (188) of a quaternion

have

VII.

We have

V? = TV? UV?.
.

also (comp. 196, VI.),

VIII.

VS? =
*

0,

or

VS =

Compare the Note

to

(202, VII.)

page 132.

into factors,

we

PROPERTIES OF RIGHT (OR VECTOR) PART.

ARTS. 203-204.]

..YY? = Y?,

IX.
X.

and

or

YK = -Y,

or

.YK?=-Y<?,

YV = Y(202,IX);

197

because conjugate quaternions have opposite rig Jit parts, by the definitions in
For the same reason,
137, 202, and by the construction of fig. 36 [p. 115].

we have

this other general formula,

XL .Kq^Sq-Vq,

K=S-Y;

or

we had

but

=
S<?

or

V<?,

= 8 + Y, by 202,

IV.

III.,

hence not only, by addition,

q

but

also,

by

Kq = 2Sq,

or

K = 28,

as in 196,

I.,

subtraction,

XII.

-K? = 2Y?,

1-K = 2V;

or

Y, of the Operation of taking the Right Part of a

137 156; 187; 196), may be defined by either
Quaternion (comp. 132, (6)
of the two following symbolical equations

whence the

Characteristic,

XIII.

Y = 1 - S (202,

whereof the former connects

the characteristic

K;

Y = i (1 - K)

that of

K on S and Y expressed by
K by 196, IF. Again, the line

S on

is

if

50, be multiplied (15) by any scalar

fig.

BB

evidently multiplied

XY.

with the characteristic S, and the latter with

it

and

OB, in

and

XIY.

while the dependence of

the recent formula XI.

is

IY.)

by the same

Y#g = xVq,

by 188, 191,
XYI.
Y? = T? YD?,

coefficient, the perpendicular

hence, generally,

if

# be any scalar

therefore,

But the

and

XYII.

TY? = T? T VUj.
.

consideration of the right-angled triangle, OB

B, in

the same figure,

shows that

XYIII.
because,

and

we

by 202,

II.,

TY? = T?.sin^,

we have

TY? = T
T
.

OA)

(OB":

OB"

=
OB

T
.

OB":

sin

arrive then thus at the following general

quaternions with trigonometry

anew

XIX,

AOB

OA,

and useful formula, connecting

ELEMENTS OF QUATERNIONS.

198

[II.

13.

i.

by combining which with the formula,

SU<?

we

= cos z? (196, XYL),

arrive at the general relation

XX.
which

may

also (by

..

2
2
(SU?) + (TVU?)

=1;

XVII., and by 196, IX.) be written thus

XXL

..

and might have been immediately deduced, without

sines

and

cosines,

from the

right-angled triangle, by the property of the square of the hypotenuse, under

the form,
t
+(T.B B) - (T.OB)*.
/

(T. 000*

The same important relation may be expressed

example, we may write,

XXII.
where

it is

N? = T?

= S? 2

in various other

-V?

ways

for

2
,

assumed, as an abridgment of notation (comp. 199, VII., VIII.),

that

V? =
2

XXIII.

(V?)

XXIV.

but that

V f=V
And

the import of this last symbol remaining to be examined.

the definition of a norm, and by the properties of
and
S<?

XXV.
we may

NS? = S?

2
,

XXVI.

but

),

(<?

because,

by

V<?,

NV? = -

V<f,

write also,

XXVII.
a result which

is

N? = N

(Sq +

V?) =

NS? + NV?

indeed included in the formula 200, VIII., since that equa

tion gives, generally,

XXVIII. ..N (0 + *) -N0 + N*,

x being, as usual, any scalar.

we

It

may

if

Lq = 4il\

be added that because (by 106, 143)

have, as in algebra, the identity,

XXIX... -(/ +

<?)

= -q -q,

..

the opposite of the sum of any two quaternions being thus equal to the
the opposites,

we may

(by XI.) establish this other general formula

XXX.

-K? =

sum of

V<?-S?;

the opposite of the conjugate of any quaternion q having thus the same right
part as that quaternion, but an opposite scalar part.

GEOMETRICAL EXAMPLES.

ART. 204.]

From

(1.)

the last formula

q =

if

it

may

be inferred, that

Y/ = +

then

K<?,

199

but

Y<?,

=S<?

S<?

and therefore that

T/ = Tq,

Ax

and

= A.X

L q =

but

q,

IT

- L
q

might have been deduced from 133 and 143, without

the introduction of the characteristics S and Y.
which two

last relations

The

(2.)

equation,
\2

V)=(v],
a
the equation of 203,

like

or(byXXVL),

expresses that the locus of p

(5.),

or cylinder of revolution, with OA for

cylinder,

NV^-NVS

its

axis,

is

the right

which passes through

the point B.
(3.)

The system

of the

like the corresponding

two equations,

system in 203,
but

same right cylinder

section of the

(6.),
if it

represents generally an

happened that 7

[|

elliptic

the section

a,

then becomes circular.

The system

(4.)

of the

two equations,

<*-!,

with

>-!,

*<1,

the cylinder of revolution, with OA for axis,

represents the circle* in which
and with (1 - x?)^Ta for radius, is perpendicularly cut by a plane at a

distance

being

(5.)

to

xTa from

the vector of the centre of this circular section

xa.

1,

While the

scalar x increases (algebraically)

the connected scalar

decreases from 1 to

- # 2 at
)
\/(l

first

the radius of the circle

maximum =

from -

(4.) at

and thence

1 to 0,

increases from

to 1,

and then

the same time enlarging

Ta, and then again diminishing to zero while the

position of the centre of the circle varies continuously, in one constant direc

from zero

tion,

to a

from & first

By

the word

in like manner, the

surfaces,

limit-point

"

circle,"

words

and not volumes.

if

OA

in these pages,
"

"sphere,"

is

a, to

the point A, as a second

limit.

and
usually meant a circumference, and not an area
cone," &c., are usually here employed to denote
;

"

cylinder,"

ELEMENTS OP QTJATEENIONS,

200

The

(6.)

such

locus of all

therefore with o for centre

sented

by the equation

T/o

circles is

= Ta of 186,

p+a

but which

now

AA

13.

i.

for a diameter,

and

namely, the sphere which has already been repre

8^-^ =

by

the sphere, with

[II.

0, of

(2.)

or

by

1, of

187,

(1.)

or

200, (11.);

new

presents itself under the

form,

obtained by eliminating x between the two recent equations (4.).

from the last form to the second, and
(7.) It is easy, however, to return

thence to the
or

first,

or to the third,

by

rules of calculation already established,

In

relations between the symbols used.

be written, by XXII., under the form,

by the general

tion (6.)

may

N=l;
a
and therefore

also

TP

whence

=
a

= Ta, by 187, and S

1,

fact,

by 190, VI.

0,

the last equa

by 200,

(11.).

Conversely, the sphere through A, with o for centre, might already

have been seen, by the first definition and property of a norm, stated in
(8.)

145, (11.), to admit (comp. 145, (12.)) of being represented

N- = 1

and

therefore,

by XXII., under the recent form

write x to denote the variable scalar

(6.)

by the equation
;

in

which

-, as in the first of the

tions (4.),

we

if

we

two equa-

recover the second of those equations: and thus might be led

to consider, as in (6.), the sphere in question as the locus of a variable circle,

which

is (as

above) the intersection of a variable cylinder, with a variable plane

perpendicular to
(9.)

its axis.

The same

sphere

may also, by XXVII.,

have

its

equation written thus,

or

a
(10.)

If, in

each variable plane represented by the first equation (4.), we

circle, or that of the variable cylinder, to be multi

conceive the radius of the

plied

by any

constant and positive scalar

axis of the cylinder remaining unchanged,

of

circles,

represented by

this

new system
aa

a,

the centre of the circle and the

we

shall pass thus to a

of equations,

new system

GEOMETRICAL EXAMPLES.

ART. 204.]

The

(11.)
tion

locus of these

new

new

the centre of this

circles will evidently

201

be a Spheroid of Revolu

and the

surface being the centre o,

same surface being the diameter AA

axis of the

of the sphere lately considered

which

therefore either inscribed or circumscribed to the spheroid, according

is

sphere
as the constant a

>

or

<

because the radii of the

case greater, but in the second case

less,

new

circles are in the first

than the radii of the old

circles

or

because the radius of the equator of the spheroid = a Ta, while the radius of
the sphere = Ta.

The equations

(12.)

respectively the sphere

of the two co-axal cylinders of revolution,

and spheroid

which

envelope

(or are circumscribed thereto) are:

iY~i;
aa

or

or

The system

(13.)

of the two equations,

V]
=a?-l,
PJ
represents (comp. (3.)) a variable

with

Snot

||

a,

the scalar x be

ellipse, if

still

treated as a

variable.

The result of the

namely this new equation,

elimination of x between the two last equations,

(14.)

or

NS

+
a

NV-l,

by

XXV.,XXVL;

/3

or

N fe

J0

1,

by XXVII.

represents the locus of all such

representation, through

equal axes]

or finally,

ellipses (13.),

T (V + V

and

will be

quaternions, of the general

and the

by 190, VI.,

found to be an adequate

ELLIPSOID (with

that celebrated surface being here referred to

origin o of vectors to its points

1,

its

three mi-

centre, as the

six scalar (or algebraic) constants,

which enter into the usual algebraic equation (by co-ordinates) of such a central
ellipsoid, being here virtually included in the tivo independent vectors, a and ]3,
which

may

be called

its

two Vector-Constants*

* It will be
found, however, that other pairs of vector-constants, for the central

ellipsoid ,

o occasionally be used with advantage.

HAMILTON

ELKMENTS OF QUATKRNIONS.

2D

may

ELEMENTS OP QUATERNIONS.

202

The equation (comp.

(15.)

-l,

or

NV^-1,

13.

or

TV^-1,

of this ellipse of contact is represented

and touching

to the ellipsoid,

which answers to the value x =

ellipse

i.

(12.)),

represents a cylinder of revolution, circumscribed

along the

[II.

0,

in (13.)

it

so that the plane

by the equation,

the normal to this plane being thus (comp. 196, (17.)) the vector a, or OA;
while the axis of the lately mentioned enveloping cylinder is )3, or OB.
(16.)

of

Postponing any further discussion of the recent quaternion equation

may be noted here that we have generally, by XXII.,

the ellipsoid (14.), it

the two following useful transformations for the squares, of the scalar
of the right part

Y<?,

xxxi.
(17.)

of

2
s? =

In referring

upon occasion,
2

(S)

than S 2 =

ay

2
,

v?

2
;

&c.

and

xxxn.

v^ =

s?

ay.

and to the connected formula XXII.,

be somewhat safer to write,

. (T) 2 + (V) 2
+

briefly to these,

may

it

any quaternion

S<?,

2
2
2 (V) = (S)
(T)

2
,

(T)

(S)

(V)

2
because these last forms of notation, S , &c., have

been otherwise interpreted already, in analogy to the known Functional Nota

tion, or Notation of the Calculus of Functions, or of Operations (comp. 187, (9.))
and 204, IX.).
196, VI.

(18.)

In pursuance of the same analogy, any

scalar

may

be denoted by the

general symbol,

V- O;
because scalars are the only quaternions of which the right parts vanish.
(19.)

In

like

manner, a right quaternion, generally,

may

be denoted by the

svmbol,
S- O;

and

may

since this includes (comp. 204, I.) the right part of ant/ quaternion,
establish this general symbolic transformation of a Quaternion
:

we

QUATERNION EQUATION OF THE ELLIPSOID.

ART. 204.]
(20.)

With

this

form

we should have

of notation,

203

generally, at least for

real* quaternions, the inequalities,

2

(Y- 0)

Oj^O;

(S-

>0;

so that a (geometrically real] Quaternion is generally of the

form

Square-root of a Positive, plus Square-root of a Negative.

The equations 196, XVI., and 204, XIX.,

(21.)

quaternions and trigonometry, the formula

XXXIII.

tan L q =

give, as a

new link between

TVU? SU?
:

TV?

8q.

not be entirely in accordance with the theory of that Func

tional (or Operational) Notation to which allusion has lately been made, but it
will be found to be convenient in practice, to write this last result under one
(22.) It

may

or other of the abridged forms

XXXIV.

.tan L q =

which have the advantage of saving

when

XXXIV

or

-.q]

... tan L q

the repetition of the

= (TV

S) q;

symbol of the quaternion,

that symbol happens to be a complex expression, and not, as here, a single

letter, q.

(23.)

The transformation

generally,

by

II., for

XXXV.
so that

we may

194, for the index of a right quotient, gives

any quaternion

IVq = TV? Ax

by

II.,

And

because

we may

Ax

write also,

(1

Art. 149

t Compare the Note

%

At a

XXXVI.

IUV?

Ax

"Vq)

iiiv = AX.

= - Ax

.
V<?,

=by 135, and therefore

Ax

q,

by XXXV.,

XXXV.
* Compare

the formulae

establish generally the symbolical J equation,

xxxvr.
(24.)

q,

and the Notes

(1

V?) =

Ax

TV?.

to pages 87, 135.

to Art. 199.

later stage [286] it will be

found possible (conip. the Note to page 175, &c.), to write,

generally,

and then (comp. the Note in page 120 to Art. 129) the recent equations, XXXVI.,
take these shorter forms [291]

XXXVI

Ax

= UV^

Ax = UV.
.

.,

will

ELEMENTS OF QUATERNIONS.

204

[II.

13.

i.

205. If any parallelogram. OBDC (comp. 197) be projected on the plane

through o, which is perpendicular to OA, the projected figure

OB"D"C"

a parallelogram

is still

(comp. 11)

so that

or

oD"=oc"+OB"(6),

and

by 106,

therefore,

//
a = (7

ty/

a)

S"=y"+6";

few
+ Q3
/

\
:

a).

Hence, by 120, 202, for any two quaternions, q and

<?

we have the general

formula,

with which

easy to connect this other,

it is

..V

II.

Hence

also, for

and similarly

any
+

-?) =

(<?

three quaternions,

(?

q)

=
j

V (q

any greater number

for

generally (comp. 197,

III.

q,

of

-</.

<?

q, q

+q) =

+ (Vy + V?)

V?"

summands

we may

so that

write

II.),

V2? = 2V?,

III

or briefly

VS = SV

while the formula II. (comp. 197, IY.) may, in like manner, be thus written,

IY.

VA? = AY?,

or

IV.

YA = AY

mode of grouping them, in III., being

remain unaltered, although both those restrictions will

the order of the terms added, and the

as yet

supposed
soon be removed.

to

"We conclude then, that the characteristic Y, of the

operation
of taking the right part (202, 204) of a quaternion, like the characteristic S of

and the

taking the scalar (196, 197),

characteristic

of taking the conjugate

a Distributive Symbol, or represents a distributive operation

z., N,
U, T, of the operations of taking
respectively the axis (128, 129), the angle (130), the norm (145, (ll.))> the
versor (156), and the tensor (187), are -not thus distributive symbols (comp.
is

(137, 195*},

whereas the characteristics, Ax.,

186, (10.), and 200, YII.)

operating on
(1.)

its

We may

Indeed,

summands

it

but

or do not operate

upon a whole

parts (or summands).

now recover the symbolical equation

form (comp. 196, VI.

VIII.

(S

by

= 1 (145), under the

and 204, IV., VIII., IX., XI.)

2 SV - VS + V 2 = S + V = 1.
V) = S

202, IV.

(or sum),

has only been proved as yet (comp. 195,

this result will soon be extended [207].

(1.)

),

that

K2q = SK?,

for the case of two

RIGHT PART OP A SUM OR DIFFERENCE.

ARTS. 205-206.]

In

(2.)

manner we can recover each

like

VII.

=1-2V+ V =1-V =

-S

(1

or thus (comp. 196, II

terms of

- 28 + S 2 =

from

S, as in 196,

VI.

- S =
V, as in 201, IX.

and 204, XIV.), from the expressions

S and

for

V in

VIII.

S2 =

IX.

J (1 +

= i

(1

2
K) = 1(1 +
2

K)

K
2K + K

2K

i(l

i(l +

K) = S

= J(l - K) = V.

Similarly,

(3.)

and

.,

S 2 = (1-V) 2

S2

of the expressions for

the other, under the forms (comp. again 202, IY.)

VI.

205

X.

XI.

206.

As

SV

VS

= J

J (1
(1

+ K)

K)

(1

(1

K
i
-K
(I

K) =

+ K) =

(1

2
)

0, as in

204, IV.

0, as in

204, VIII.*

regards the addition (or subtraction) of such right parts, V^, ~Vq
any two right quaternions (132), we may connect it with the
,

or generally of

addition (or subtraction) of their indices (133), as follows.

any parallelogram (197, 205 j, but

pendicular to

its

Ta =
Let OB D

plane

l,

so that

^0:a)

= ^(
7 :a)=^(8:a) =

|,

8 = j +

(3.

be another parallelogram in the same plane, obtained by a positive

rotation of the former, through a right angle,

Then

Let OBDC be again

OA be now an unit- vector (129) per

let

the three right quotients,

/3

a,

a,

round OA as an axis

and

a,

may

so that

represent any two

and their sum, (f + q, which is always (by 197, (2.) )

q, q
a right quaternion
and the indices of these three right quotients are
three
the
lines )3 y S
so that we may write, under the
(comp. 133, 193)

right quaternions,
itself

foregoing conditions of construction,

form symbolical equations analogous to those in 161 (3.)

He may compare the equations obtained from the distri
butive symbols S, V and K, with those obtained from T, U and R, and may notice the pairs of
symbols commutative in order of operation, &c. It is well to combine the symbols as in a multipli
[It

from the

may be

six

cation table.]

instructive to the student to

symbols S, V,

K and T,

U,

B,.

ELEMENTS OF QUATERNIONS.

206

[II.

13.

i.

But

this third index is (by the second parallelogram) the sum of the two
= 7 + /3 we may therefore write,
former indices, or in symbols,
;

..I

I.

or in words the Index

Sum

because

Y<?,

sum

their

V#

no

=!<?

if

I?,

Sum* of any

^=/<?

(205, I.)

=|;

two Right Quaternions

and

Y (q

so that the index of the right part of the

is

equal

to the
9

IV+

are always right quotients (202, 204),

sum of

quaternions is the
is

?)

Hence, generally, for any two quaternions, q and q

the formula,
II.

of the

of their Indices.

we have

(?

the indices of the right parts.

In

q) is

sum

like

of

always
any two

manner, there

proving that

difficulty in

III.

.!(<?

-<?)

=!<?

if

I?,

L<f=L

^;

and generally, that

the Index of the Difference of any two right quotients, or of the right parts of
any two quaternions, being thus equal to the Difference of the Indices*

We

may

then reduce the addition or subtraction of any two such quotients, or parts,
a right quaternion being always
given, or known.

to the addition or subtraction of their indices

(by 133) determined, when its index is

207.
see, then, that as the MULTIPLICATION of any two Quaternions was

We

191) reduced to (1st) the arithmetical operation of multiplying their tensors,

(in

and (llnd) the geometrical operation of multiplying their versors, which latter
was constructed by a certain composition of rotations, and was represented (in
either of two distinct but connected ways, 167, 175) by sides or angles of a
spherical triangle

so the

ADDITION of any two Quaternions may be reduced

(by 197, I., and 206, II.) to, 1st, the algebraical addition of their scalar parts,
considered as two positive or negative numbers (16.) ; and, llnd, the geome
trical addition of the indices of their right parts,

this latter Addition of Lines

Parallelogram

* Compare
t
it

(6.).f

In

like

considered as certain vectors

according to the

(1.)

Rule of

the

being performed
manner, as the general Division of Quaternions

the Note to page 175.

It does not fall within the plan of these

ought to be distinctly stated that

Notes to allude often to the history of the subject but

what may be called Geometrical Addition
;

this celebrated Rule, for

of right lines, considered as analogous to composition of motions (or of forces), had occurred to several
writers before the invention of the quaternions
although the method adopted, in the present and in
A (1.) for the line AB,
a former work, of deducing that rule, by algebraical analogies, from the symbol B
:

may

The reader may compare the Notes to the Preface to the

possibly not have been anticipated.
s Volume of Lectures on Quaternions (Dublin, 1853).

author

GENEBAL ADDITION OF QUATERNIONS.

ARTS. 206-208.]

was seen

(in 191) to

207

admit of being reduced to an arithmetical

and a geometrical

division of versors, so

division

we may now

of

and

(by 197, III.,

206, IV.) reduce, generally, the Subtraction of Quaternions to (1st) an alge
braical subtraction of scalars, and (Ilnd) a geometrical subtraction of vectors
tensors,

beiug again constructed by a parallelogram, or even by a

And because the sum of any given
plane triangle (comp. Art. 4, and fig. 2).
set of vectors was early seen to have a value (9.), which is independent of their
this last operation

and of the mode of grouping them, we may now infer that the Sum of
number
any
of given Quaternions has, in like manner, a Value (comp. 197, (I.)),
which is independent of the Order, and of the Groining of the Summands or in
order,

other words, that the general Addition of Quaternions

is

a Commutative* and an

Associative Operation.
(1.)

The formula,

VS?

= SV?, of 205,

III.,

now

seen to hold good, for any number of quaternions, independently of the

arrangement of the terms in each of the two sums, and of the manner in

is

which they
(2.)

We

may

be associated.

can infer

anew

K (q

that

+q)= K/ +

Kq, as in 195,

II.,

under the form of the equation or identity,

S tf+q)(3.)

More

generally,

K2g

it

V (/ + q)
may

(S/

V/) +

(80

Vtf).

be proved, in the same way, that

= 2K?,

or

briefly,

whatever the number of the summands

208.

may

K2

= 2K,

be.

As

regards the quotient or product of the right parts, Vg and V/, of

let t and if denote the tensors of those two
parts, and
x denote the angle of their indices, or of their axes, or the mutual inclination

any two quaternions,

let

of the axes, or of the planes, f of the

that (by 204,

XVIII.)
t

and

TV?

two quaternions q and q

themselves, so

= Tq

sin

x = L (IV/

Lq,
:

TV/ = T/

IV?) = L (Ax

*
Compare the Note to page 176.
t Two planes, of course, make with each other,

sin L

/ Ax
:

<?

q).

in general, two unequal and

supplementary angles
here suppose that these are
mutually distinguished, by taking account of the aspect of each
as
from
the opposite aspect: which is most
plane,
distinguished
easily done (111), by considering the
axes as above.

but

we

ELEMENTS OF QUATERNIONS.

208

Then, by 193, 194, and by 204,

V?

V?

I.

II.

.V/.V?

and therefore (comp.

III.

V.

We

V?

XXXV

.,

(TV/ TV?) (Ax

:

IV/:I^-=- (TV/. TV?). (Ax./:

198], with the

Y?)

= -

costf

1ft

Ax.

q)

Ax.

q)

>,

temporary abridgments proposed above,

V?) = tt cos x\

z(V/.V?)

IV? = +

S (Vq

VII.

= IV?

XXXV.,

13.

i.

[II.

IV.

VI.

VIII.

= ^;

SU (Vq
SU (Vq

= + cos x

:Vq)

cos^;
.Vq) =

z(V/.Vy) =

7r-a?.

have also generally (comp. 204, XVIII., XIX.),

IX.

XI.

and in

TV (Vq
TV (Vq

V?) =

t t~

Vq) = ft

sin

sin x

X.

XII.

TVU (Vq
TVU (V?

V?) =

sin

V?) =

sin

particular,

XIII.

V (Vq

Vq) =

if

because (comp. 191,

XIV.

and

0,

V (Vq

Vq) =

q \\\q (123);

and 204, VI.) the quotient or product

(6.),

0,

of the right

two complanar quaternions (supposed here to be both non-scalar (108),

so that t and t are each
0) degenerates (131) into a scalar, which may be
parts of

>

thus expressed

XV.

.Vq :Vq = +

XVI.

and

ft-\

Vq Vq =
.

ft,

if

x=

t t,

if

x =

but

XVII.
the

first

V?

= -

Vq

t-\

and XVIII.

diplanarity (119),
their axes,

or in which those planes intersect,

round
mulae

it

from

I., II.,

XIX.

Ax

q to

Ax

and the second

case being that of coincident,

In the more general case of

which is perpendicular to both

Vq Vq = +

and therefore common

and which

case that of opposite axes.

we denote by

if

TT

S the unit-line

to their tivo planes,

so directed that the rotation

is

q is positive (comp. 127, 128), the recent for

give easily,
.

Ax

(Vq :Vq)= + $

XX.

Ax

(Vq

- S
Vq) =

and therefore (by IX., XI., and by 204, XXXV.), the indices of the right parts,
of the quotient and product of the right parts of any two diplanar quaternions,

may

be expressed as follows

XXI.
XXII.

IV

(Vq :Vq)= + $.

IV

(Vq

ft-* sin

- 3 ft
Vq) =
.

sin #.

QUOTIENT OR PRODUCT OF RIGHT PARTS.

ART. 208.]

209

Let ABC be any triangle upon the unit-sphere (128), of which the
spherical angles and the corners may be denoted by the same letters A, B, c,
(1.)

while the sides shall as usual be denoted by

a, b y c

and

let it

be supposed

that the rotation (comp. 177) round A from c to B, and therefore that round B

from A

to

c,

43

&c., IB positive, as in fig.

Then

[p. 144].

writing, as

we have

often done,
q

we

)3

and

a,

where

]3,

= OA,

&c.,

easily obtain the following expressions for the three scalars

for the vector 8

whence

In

fact

and

if

= sin

we have

= sin a

x =

- B

TT

8 = -

and

here,

Also

are as just stated.

if

c be (as in 175) the positive

the points A, B, c the negative poles (comp. 180, (2.)) of the

,

#,

]3.

poles of the three successive sides BC, CA, AB, of the given triangle,

AB

if ,

(2.)

new

and therefore
arcs B C

cV,

then

Ax
but x and

= oc

and the

are the angle

the quaternion which

is

Ax

f
.

= OA

axis of the quotient of these

two axes, or of

by the arc cV; therefore x is, as

and 8 is directed to the point upon

represented (162)

above stated, the supplement of the angle B,

the sphere, which is diametrically opposite to the point

B.

V. VII. VIII. IX. XI., for any triangle ABC on the

Hence, by
=
unit-sphere, with a
OA, &c., we have the formulse
III.

(3.)

XXIII.

XXIV.

svrV-sina
cosecccoss;
A P

XXVII.
XXVIII.

TV V

TV

V-

P
.

\
HAMILTON

ELEMENTS OF QUATERNIONS,

Va

= + sin a

sin c cos B

= + sin a cosec

c sin

sin a sin c sin B.

/
2

ELEMENTS OP QUATERNIONS.

210
(4.)

XX. XXI. XXII.,

by XIX.

Also,

[II.

13.

i.

the rotation round B from A

if

to c be still positive,

XXIX.

..Ax.v:Y
XXXI.

XXXII.

IV

XXX.

= -

IV (V I V^)
\
/
P

-3

= +

]3

sin a cosec c sin B

/3

sin a sin c sin B.

on the other hand, the rotation round B from A to c were negative,

then writing for a moment ai = - a, |3i = - ]3, 71 = - 7, we should have a new
(5.)

and
be

If,

A^CI, in which the rotation round

opposite triangle,

positive,

but the angle at

treating (as usual)

have

BI

all

= B, as well

BI equal in

as

= c, and

the four formulae of

new
)3,

75

Y-

P
(4.)

magnitude

= )

= + 8

and therefore,
sin

j3i

(6.)

in 204,

A! to Ci

for example,

cosec c : sin

sin a cosec c sin B

would

so that

by
we should

by

BI,

would therefore

still subsist,

direction of rotation in the given triangle,

in the second

provided that, for this

we were

to change the sign of

member of each.

Abridging, generally IV#

XXXIV

.,

to

(TV

S) q,

Sq to (IY S) q, as TVg S^ was abridged,

we have by (5.), and by XXIV., XXXII.,
:

any three unit-vectors

this other general formula, for

still

=a

a^

IV

from

the angles of a spherical triangle as positive,

Y~ V
or

B!

to that at B

a,

|3,

as terminating at the corners of a spherical triangle ABC

XXXIII.

(IV

S)

j3

tan B

7, considered

the upper or the lower sign being taken, according as the rotation round B
from A to c, or that round |3 from a to 7, which might perhaps be denoted by
the symbol afiy, and which in quantity
positive or negative.

When

is

equal to the spherical angle

B,

is

any three quaternions q, q q , considered as all

the
o
(119), contain any common line, those three may
origin
passing through
then be said to be Cottinear* Quaternions and because the axis of each is then
209.

the planes of

* Quaternions
Compare the

first

of which the planes are parallel

Note to page 116.

to

any common

line

may

also be said to be cottinear.

COLLINEAR QUATERNIONS.

ARTS. 208-210.]

211

perpendicular to that line, it follows that the Axes of Collinear Quaternions are
Complanar while conversely, the complanarity of the axes insures the collinearity of the quaternions, because the perpendicular to the plane of the axes is a
:

common

line

to the planes of the quaternions.

collinear but the converse propo

hold good, collinear quaternions being not necessarily corn-

Complanar quaternions are always

(1.)

sition does not

planar.
(2.) Collinear quaternions, considered as fractions (101), can always be
reduced to a common denominator (120)
and conversely, if three or more
can
be
so
as
to
quaternions
reduced,
appear under the form of fractions with
;

common denominator c, those quaternions must

e is then common to all their
planes.

be collinear: because the

line

two quaternions are collinear with any scalar

Any

(3.)

the plane of a, scalar

being indeterminate* (131).

Hence the

(4.)

scalar

and

right parts, Sq,

S/, Yg,

of

V<?

any two quater

nions, are always collinear with each other.

The

(5.)

conjugates of collinear quaternions are themselves collinear.

210. Let
line

common

other lines

j3,

q,

be any three collinear quaternions and let a denote^, a

Then we may determine (comp. 120)Vthree
such that
;

q"

to their planes.
y, S,

and thus may conclude that

because,

like

(as in algebra),

by 106, 107,

Q/O

^Ct

In

manner, at

least

CtOttO

under the same condition of collinearity,f

it

may

be

proved that
Operating by the characteristic
192, II., and 195, II.,
III.

IV.

where (by 209,

(5.))

we
.

K upon these two equations, and attending to

find that

K0".

Kq".

(Kq +
(Kq -

=
K<?)

Kq) =

Ktf + K0".
Kq". Kq

Kg".

~Kq"

Kq
Kq

the three conjugates of arbitrary collinears, K?,

Kq

Kq",

Compare the Note to page 117.

+ It will soon be
eeeu, however, that this condition

is

unnecessary.

ELEMENTS OF QUATERNIONS.

212

may

We

represent any three collinear quaternions.

same degree
V.

[II.

i.

13.

have, therefore, with the

of generality as before,

VI.

.
+<?"<?;

<?"(/+<?)=?"<?

q"(q

-q}=q"q

be any four collinear quaternions, we

q
formula (again agreeing with algebra)

If, then, q, q

-<fq.

may

q",

establish the

Vii.

and similarly

for

(f +

2")

+ q) =

(<?

any greater number,

VIII.

S/ =

S/ 2?

Y+ Y+

<f"i

we may

so that

write briefly,

= 2tfq,

and

+ qz +
+ qm
S/ = q\ + q z + + q
+ tf nfe
q\q m + q tfi +
S/2 = /i?i +

positive whole numbers.

where

qi

plication of Collinear* Quaternions

(1.)

Hence, by 209,

tion, for the product of

IX.
(2.)

Hence

qq

is

tionf (comp. 126; 199, VII.

we have

13.), the

Multi

Sq +

V? +

.
S<?

of any quaternion,

S<f+2S<?.V?

this general

transforma

and 204, XXIII.)

=
.<f

n,

In words (comp.

III.,

any two quaternions

= S/. Sq + V?

a Doubly Distributive Operation.

and 202,

(4.),

also, for the square

X..

and w being any

i"i

V?

</.

V?.

we have

the transforma-

2
.

Separating the scalar and right parts of this last expression,

these other general formulae
(3.)

we

find

XL
whence

dividing by

also,

XIII.

2
.S.? = S? + V?
2

SU (?

2
T<?

= (SU?) +

V.<f

2S<?.V?;

we have
2

XII.

(VU<?)

XIV.

VU (f) = 2SU?

VU<?.

= Kq, in IX., and therefore S/ = &q, V/ = - V#, and

(4.) By supposing q
transposing the two conjugate and therefore complauar factors (comp. 191,
(1.)

),

we

obtain this general transformation for a norm, or for the square of a

V. 202, III. ; and 204, XI.)

tensor (comp. 190,

XV.

Tq* =

N? = qKq =

which had indeed presented

(8q

-f

V?) (8q

itself before (in

- V?) = S?2 - V^ 2

204, XXII.), but

is

now

obtained

* This distributive
property of multiplication will soon be found (compare the last Note) to extend
more general case, in which the quaternions are not collinear.

to the

t [By means

of the formulae of 204

be effected on a square or product.]

many different

transformations involving K, S, V, and

T may

DISTRIBUTIVE MULTIPLICATION OF COLLINEARS.

ART. 210.]

new way, and

in a

any employment of sines, or

ivithout

well-known theorem respecting the square of

2

Eliminating

(5.)

even of the

the hypotenuse.

by XV., from XI., and dividing by

V<?

cosines, or

213

T<f,

we

find

that

XYI.

.
<f

= 2S? 2 -

T<f

XVII.

= 2 (SU?) 2 -

SU (f)

agreeing with 199, VI. and IV., but obtained here without any use of the
known formula for the cosine of the double of an angle.
(6.)

Taking the

and right parts

scalar

these other general expressions

XVIII. ..S/0- Sq

XIX.

We

.
S<?

+ S (V/.V?)

obtain

V?V = W- S? + V? S/ + V OY. V?)

which we

in the latter of

V?, S/.
formula

we

of the expression IX.,

or
(by 126) transpose the two factors V?
also (by 206, 207) write, instead of XIX., this other

may

may

S<7,

XIX

I Vq q =

IV? S? + IV? S/ + IV (Vq V?)

.

= Kq and transpose (comp.

f
the
two
(4.) )
complanar (because conjugate) factors, q + q and
(q + q), we
obtain the following general expression for the norm of a sum
(7.)

If

we

suppose, in VII., that

q"

= Kq, q

"

(q

q)K (q

q)

+ qKq + q

Kq

Kq

4-

qKq

or briefly,

XX.

(<?

g)

Nq

+ 2S

qStf +

N,

as in 200,

VII.

because
q

Kq = K
(8.)

Kq + x,

q Kq

by 192,

II.,

and

(1

+ K)

qKq = 2S

q Kq,

by 196, IF.

By changing q to x in XX., or by forming the product

where x is any scalar, we find that
XXI.

N (q + x) = N? + 2zSq + x\

as in 200,

VIII.

of q

+ x and

whence, in particular,

XXr.
(9.)

for

N (q - 1) = N? - 2Sq +

Changing q

any two

to

/3

vectors, a and

XXII.

(|3

1, as in

200, II.

a,

and multiplying by the square

j3,

the formula,

a)

= T/3 2 2T/3

Ta

SU

2
,

of Ta,

we

get,

ELEMENTS OP QUATERNIONS.

214
in

which Ta 2 denotes* (Ta) 2

[II.

i.

13.

because (by 190, and by 196, IX.),

and

To

S- Ta
a

(10.) In any plane triangle, ABC, with sides of which the lengths are as usual
denoted by a, b, c, let the vertex c be taken as the origin o of vectors then
;

a = cA,

we

j3

j3-a =

cB,

Ta =

AB,

T/3

STj

T(/3-o)=c,

a,

from XXII., the fundamental formula of plane

metry, under the form
2
2
c = a* - 2ab cos c +
XXIII.
recover therefore,

Z>

trigono

important to observe that we have not here been arguing in a

because although, in Art. 200, we assumed, for the convenience of the
It

(11.)
circle

= cosc;

is

student, a previous knowledge of the last written formula, in order to arrive

at certain applications, yet in these recent deductions from the

more rapidly

distributive property

VIII. of multiplication of

(at least) collinear quaternions,

we have founded nothing on the results of that former Article and have
made no use of any properties of oblique-angled triangles, or even of right;

angled ones, since the theorem of the square

anew

virtually proved

in

nor

(4.)

is it

of the

hypotenuse has been

necessary to the argument, that any

known, beyond the mere defini

properties of trigonometric functions should be

tion of a cosine, as a certain projecting factor,

was derived, and which

justifies us in

The geometrical Examples,

from which the formula 196, XVI.

writing cos c in the last equation (10.).

may therefore be read

in the sub -articles to 200,

again, and their validity be seen anew, without any appeal

to

even plane trigono

metry being now supposed.

= T?2 + V? as in 204, XXXI. and we
(12.) The formula XV. gives S?
know that
as being generally the square of a right quaternion, is equal to
2

V<?

a negative scalar (comp. 204, VI.), so that

XXIV.
in each of

vanishes

V?

which two cases

V#

unless

0,

<

0,

by 202,

0,

(6.),

or

TT,

and therefore

its

square

hence,

XXV.

2
.

S?

<

T?

(SU^)

<

1,

in every other case.

*
are not yet at liberty to interpret the symbol Ta2 as denoting also T (o2 )
because we have
not yet assigned any meaning to the square of a vector, or generally to the product of two vectors.
In
the Third Book of these Elements [282
or product can be
(3.)] it will be shown, that such a square
and then it will be found (comp. 190), that
interpreted as being a quaternion

We

T
whatever vector a

may

be.

(a

(Ta)

= Ta 2

APPLICATIONS TO TRIGONOMETRY.

AET. 210.]

215

might therefore have been thus proved, without any use of the
transformation SUg = cos L q (196, XVI.), that (for any real quaternion q) we
It

(13.)

may

have the inequalities,

XXVI.

SU?

<

1,

and

SU?>-1,

+ Tq,

Sq<

Sq>-Tq,

unless it happen that L q = 0, or = TT

SU^ being = + 1, and $q = + Tq, in the
whereas S~Uq = - 1, and Sq = - Tq, in the second case.
first case
2
= N?, and Tq Tq = T qKq = T q Kq = N? T (q q), while
(14.) Since T?
;

S q Kq = S
.

XXVII.
if

we adopt

Kq

.
N<?

S (/

(Tq + T?)
2

T (4 + q)

XXIX.

by XXVI.,

gives,

qKq = 2N? (T - S)

(<?

q)

>

0,

case,

>

Tq - Sq = (T -

<?

T (tf + q\

unless

member

each

q,

wo a positive scalar

is

S)

of

/=

this

hence,

and

<rq,

last

>

becomes

inequality

Tq.

Writing q =

(15.)

)3

a, c[

a,

and multiplying by Ta, the formula

becomes,

XXX.
in

Tq + Tq

which excepted

XXIX.

= 2 (T - S)

the abridged notation,

and suppose that the quotient q

-(!+*)

XX.

the formula

q),

XXVIII.

in

which

latter case,

anew

therefore arrive

7 =

a/3,

but not in any other, we have

Uy

Ty

T/3>T(

7 +

unless

/3),

^>0;

Uj3 (155).

We

at the results of 186, (9.), (10.), but without its

been necessary to consider any

having
was done in those former sub-

triangle^ as

articles.

On the other
by XXVI.,

hand, with a corresponding abridgment of notation,

(16.)

have,

XXXI.
also,

by XX.,

Tq + S? = (T +

S) q

>

0,

unless

L q =

TT

&c.,

XXXII.

T (q

+ qY - (Tq - T?) 2 = 2 (T + S) qKq =

.T(q +

2N? (T + S) (q: q)
.

hence,

XXXIII.

where either sign

(17.)

And

XXXIV.

we

may

q)>

(Tq

T (y

Tq)

unless

j = -xq,

>

be taken.

hence, on the plan of

..

|3)

>

(Ty

(15.), for

any two

unless

T/3),

vectors

|3,

7,

Ur = - U/3,

ELEMENTS OP QUATEKNIONS.

216

whichever sign he adopted

XXXV.

r +

/3)

13.

i.

on the contrary,

but,

[II.

(T 7

U7

*/

T/3),

= -

U/3,

all
or
the upper or the lower sign being taken, according as T 7
Tj3
in
was
from
conside
which agrees with what
inferred,
186, (11.),
geometrical
:

<

>

In fact, if we make
combined with the definition of Ta.
=
oc
then
will
and
OBC
in
be
oc,
7
general a plane triangle, in

rations alone,
/3

= OB,
=
y

which the length of the side EC exceeds the difference of the lengths of the
two other sides but if it happen that the directions of the two lines OB, oc
;

words that the lines OB, oc have opposite directions, then

the difference of lengths of these two lines becomes equal to the length of the
coincide, or in other

line BC

(18.)

With

the representations of q and #

we have

sides of a spherical triangle ABC,

Sq = cos
the equation

8q = cos

c,

XVIII.

(19.)

To

interpret,

the same side of the arc CA,

and

if

IV

(V#

by two

(7

a)

= COS

the

fundamentalformula

+ sin a sin

cos B.

= V<?)

j3

still,

&c.,

IV/ =

is

be

now

supposed, as in 208, (5.), that

positive, so that B and B are situated at

let it

B be

Then writing a = OA
sin c

(1.),

with reference to the same spherical triangle, the con

XIX

IVq = y

),

as follows

= cos a cos

cos b

nected equation XIX., or

the rotation round B from c to A

that arc.

Sq q = S

by 208, XXIV.,

gives therefore,

XXXVI.

assigned in 208,

the values,

a,

of spherical trigonometry (comp. (10.)

as in 208, (2.), the positive pole of

we have

a sin a

IVq

= -

|3

sin b

sin a sin c sin B (comp. 208, (5.)

with the recent values (18), for Sq and S/; thus the formula
by transposition of the two terms last written

),

XIX

becomes,

XXXVII.

/3

B = a sin a cos

sin a sin c sin

13

sin b

+ y sin

cos

a.

Let p = OP be any unit- vector then, dividing each term of the last
equation by /o, and taking the scalar of each of the four quotients, we have,
(20.)

by 196, XVI.,

XXXVIII.

this
.

new equation

sin a sin c sin

B cos PB = sin a cos

+ sin

cos a cos PC

cos

PA +

sin b cos

PB

APPLICATIONS TO SPHERICAL TRIGONOMETRY.

ART. 210.]

where
are

still,

217

are as usual the sides of the spherical triangle ABC, and A B , c

as in 208, (2.), the positive poles of those sides
but p is an arbitrary
c

of the sphere.
Also cos PA cos PB , cos PC are
the
sines of the arcual perpendiculars let fall from that point upon
evidently
those sides
being positive when p is, relatively to them, in the same hemi
point,

upon the surface

spheres as the opposite corners of the triangle, but negative in the contrary
case; so that cos

AA

and are the

&c., are positive,

sines of the three altitudes

of the triangle.
(21.) If

we

place p at B, two of these perpendiculars vanish,

formula becomes, by 208, XXVIII.,

XXXIX.

sin b cos

BB =

sin a sin c sin B

and the

last

-.
P

such then

the quaternion expression for the product of the sine of the side
CA, multiplied by the sine of the perpendicular let fall upon that side, from
the opposite vertex B.

Placing p at A, dividing by sin a cos c, and then interchanging B and

this other fundamental formula of spherical trigonometry,

(22.)
c,

is

we get

XL.
and we
tion

see that this

XIX.,

or

is

XIX

.,

cos

AA =

sin c sin B

sin b sin c

included in the interpretation of the quaternion equa

XXXYI. was seen in (18.) to be the

as the formula

XVIII.

interpretation of the connected equation

(23.) By assigning other positions to p, other formulae of spherical trigo

nometry may be deduced, from the recent equation XXXVIII. Thus if we

suppose P to coincide with B

and observe that (by the supplementary*

triangle),

V=

TT

A,

A =

7T

B,

AB =

7T

C,

while
cos

we

BB =

sin a sin c

sin c sin A,

by XL.,

easily deduce the formula,

XLI.

sin a sin c sin

sin B sin c

which obviously agrees, at the plane

sin B

- cos c

limit,

A + B +

cos c sin

A-

cos a cos

sin c

with the elementary relation,

TT.

No previous knowledge of Spherical Trigonometry, properly so called, is here supposed the

supplementary relations of two polar triangles to each other forming rather a part, and a very elemen
;

tary one, of spherical geometry.

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OE QUATERNIONS.

218

Again, by placing p at A

(24.)

XLII.

sin b cos c

with the verification that, at the plane

sin c cos a cos B

c cos B,

But we cannot here delay on such

metry

limit,

a = b cos c +

appeared

13.

the general equation becomes,

sin a cos c

i.

[II.

deductions, or verifications although it

worth while to point out, that the whole of spherical trigono
thus be developed, from the fundamental equation of multiplica
:

to be

may

when

operated on by the two

characteristics S and Y, and the results interpreted as above.
tion of quaternions (107),

the last

may next

now

without being

it,

reciprocals (135),

and

necessarily collinear, are right

their sums (197, (2.)

L q = L q = L

and

=
q"

by 106,

= 1

qq/

|,
,.

shall then have,

q",

in which case their

will be right also.

),

therefore,

Lq,=

We

is

be proved, as follows, that the distributive formula I. of

which enter
Article holds good, when the three quaternions, q, q

211. It

into

that equation

L(q"+ft

Let then

TT

194, 206,

and the

distributive property in question is proved.

as in 210, it is easy hence to infer, that the
(1.) By taking conjugates,

other distributive formula, 210, V., holds

nions

any three right quater

for

good

or that

(2.)

For any

three quaternions,

OY
V?
(3.)

The quaternions

+ Y?
(V?"

q,

we have

V? =

V0".

+ V/) = V?

q"

being

therefore the two equations

V? + Y/. V?
Y?"

still

V? Y?
.

arbitrary,

we have

thus,

by

210, IX.,
(?"+

(S?"+

SO

S? +

(Vq"+

Y/) Sq + Y?
.

(S?"+

(S/. Sq +Vq

S?

(Y/WY) V?
.

$q +Vq Sq VY?
.

GENERAL DISTRIBUTIVE PROPERTY.

ABTS. 210-212.]

so that the formula 210,

is

210, V.,

also (by conjugates) the

formula

valid generally.

The General*

212.

and therefore

I.,

219

Multiplication of Quaternions

210) a Doubly Distributive Operation

formula (comp. 210, VIII.)

generally, the

however many the summands of each

set

is

we may

so that

therefore (comp. 13,

extend, to quaternions

may

be,

and whether they

be, or be

not, collinear (209), or right (211).

Hence, as an extension of 210, XX., we have now,

(1.)

II.

N2?

2N?

2SS0K/

where the second sign of summation refers to

of the quaternions q, q

III.

ST?

unless all the quaternions

in which case the

sum of

binary combinations

q,

any

XXIX., we have

the inequality,

T2?,

>

bear scalar and positive ratios to each other,

two members of

the tensors, of

all possible

as an extension of 210,

And,

(2.)

become equal

this inequality

so that the

set of quaternions, is greater than the tensor of the

sum, in every other case.

In general,

(3.)

as

IV.

The

(4.)

an extension of 210, XXVII.,

.

(ST?)

formulae, 210,

8
(TS?) =

22 (T -

S) qKtf.

XVIII., XIX., admit

easily of analogous ex

tensions.
(5.)

We have also

in which,

by 210, IX.,

VI.
because,

(comp. 168) the general equation,

f
qq + q q = 2 (Sq

S? +

V?

.
S<?

V?

S? + S ( V?

by 208, we have generally

VII.

VIII.

or

(Comp. 191,

(2.),

V (V/.V?)
V/g

= -

-V^,

V (V? W)
.

if

Lq =

LJ = ^.

and 204, X.)

Compare the Notes to pages 211 and 212. [On page (35) of the Preface to the Lectures on
Hamilton refers to an early speculation of his (1831) on the
multiplication of lines for
which the product of sums was not equal to the sum of products. When addition is not
Quaternions,"

commutative,

multiplication even

by a

scalar is not distributive.

See 180 (3.)]

2F2

ELEMENTS OF QUATERNIONS.

220

213. Besides the advantage which the Calculus of Quaternions gains, from
the general establishment (212) of the Distributive Principle, or Distributive

Property of Multiplication, by being, so

far, assimilated to

Algebra, in processes

which are of continual occurrence, this principle or property will be found to

be of great importance, in applications of that calculus to Geometry
and
;

especially in questions respecting the (real or ideal*) intersections of right lines

with spheres, or other surfaces of the second order, including contacts (real or
ideal), as limits of

The following examples may

such intersections.

serve to

how

give some notion,

the general distributive principle admits of being

such questions in some of which however the less general prin

applied to

ciple (210), respecting the multiplication of collinear quaternions (209),

be

And

sufficient.

first

from a given point upon

we

shall take the case of chords of a sphere,

its surface.

From

a point A, of a sphere with o for centre, let

(1.)
draw a chord AP, which shall be parallel to a

given line OB

more

or

would

drawn

to

fully,

assign

it

be required to

the

= OP, of the
vector, p
extremity of the chord so
as
a
drawn,
function of the two given vectors,
a = OA, and |3 = OB or rather of a and Uj3,
;

since

evident that the length of the line

it is

]3

cannot affect the result of the construction,

r,

which

fig.

51

may

Since AP

(2.)

serve to illustrate.
||

OB, or p

- a

\\

j3,

Fig. 51.

we may

begin by writing the expression,

a +

which
line

AP

may
;

(15),

be considered (comp. 23, 99) as a form of the equation, of the right

in which it remains to determine the scalar coefficient x, so as to

and

satisfy the equation of the sphere,

To
/

In

short,

we

(2.)).

are to seek to satisfy the equation,

T
by some

= Ta(186,

scalar x

which shall be

substitute this scalar

(a

0/3)

= Ta,

different
(in general)

in the expression p = a +

#/3,

from zero

required vector p.
Compare the Notes

and then

to

in order to determine the

to pages 87, 88, &c.

INTERSECTIONS OF EIGHT LINES AND SPHERES.

ART. 213.]
(3.)

For

221

an obvious process is, after dividing both sides by

employ the formula 210, XXI., which had indeed

this purpose,

Tj3, to square,

and

to

occurred before, as 200, VIII., but not then as a consequence of the distribu
In this manner we are conducted to a
tive property of multiplication.
quadratic equation, which admits of division

the problem

(I.)

by

a*,

and gives then,

being thus resolved, with the verification that

]3

may

replaced by U]3, in the resulting expression for p.

(4.) As a mere exercise of calculation, we may vary the last process

by

dividing the last equation

This

as before.

last

by

(2.)

be

(3.),

Ta, instead of T/3, and then going on

procedure gives

and therefore
x = - 2S ^
a
(5.)

N a^ = _ 2S

In general, by 196, II

by

(by 196,

- 2S = -

new

expression for p, in which

(3.), to treat
(6.)

It is

the vector

now

the expression

-R-

a
.

J3

Kjj./3j
it

is

not permitted generally, as

easy to see that the second equation of

(5.) for

it

was in

of the multiplicand.
j3 as the multiplier,* instead
(2.) is satisfied

for

p gives (by 186, 187, &o.),

T
as

as before.

finally,

P
a

.),

(3.),

and

XII

.,

hence,

=
/0

= Ta,

T~.T]3

was required.
(7.)

To

interpret the solution (3.), let c in fig. 51 be the middle point of

the chord AP, and let D be the foot of the perpendicular let
then the expression (3.) for p gives, by 196, XIX.,

CA = i (a - p) = /3S

and accordingly, OCAD

is

= OD

a parallelogram.
*

Compare the Note

to

page 159.

fall

from A on OB

ELEMENTS OF QUATEEOTOKS.

222
(8.)

To

j3

a
75,
|3

or

OPOB

we have only to observe (oomp. 138)

i.

13.

which gives

interpret the expression (5.),

-p
=

[II.

=K

OA
OB

..

,
OP = PO,

if

that the angle AOP

or the supplementary angle AOP externally,

by

is

bisected internally,

the indefinite right line OB

(see again, fig. 51).

Conversely, the geometrical considerations which have thus served in

and (8.) to interpret or to verify the two forms of solution (3.), (5.), might

(9.)

(7.)

have been employed to deduce those two forms,

if

we had not

obtain them, by rules of calculation, from the proposed

seen

conditions

how
of

to

the

(Comp. 145, (10.), &o.)

is evident, from the nature of that question, that a ought to be
deducible from ]3 and p, by exactly the same processes as those which have
question.

(10.) It

served us to deduce p from

and the form

(5.)

/3

and

Accordingly, the form

a.

(3.) of

p gives

gives

XJ.3.
And

form can be recovered from the second, we see that each

leads us back to the parallelism, p - a
(5 (2.J.
x
solution
for
shows
that
The
(11.)
(3.)
since the first

\\

0,

p =

a,

P = A,

if

(a

]3)

0,

or

if

]3 JL a.

And

the geometrical meaning of this result is obvious ; namely, that a right

line drawn at the extremity of a radius OA of a sphere, so as to be perpendi
cular to that radius, does not (in strictness) intersect the sphere, but touches it
its second point of meeting the surface coinciding, in this case, as a limit, with
:

the first.
(12.)

Hence we may

infer that the plane represented

by the equation,

or
a
is

the tangent plane (comp. 196, (5.)

to the sphere here considered, at the

point A.
(13.)

Since

substitute for

it

j3

may
-

a, if

be replaced by any vector parallel thereto, we may

j = oc be the vector of any given point c upon the chord

IMAGINARY INTERSECTIONS.

ARTS. 213, 214.]

AP, whether (as in

by

(3.)

and

fig.

51) the middle point, or not

223

we may

therefore write,

(5.),

- a

214. In the examples of the foregoing Article, there was no room for the
occurrence of imaginary roots of an equation, or for ideal intersections of line
and surface. To give now a case in which such imaginary intersections may
occur,

we

shall proceed to consider the question of

drawing a

in a given direction, from a

given external point

secant to a sphere,
the recent figure 51 still

serving us for illustration.

Suppose then that

(1.)

which

it

required to

is

given line

j3

We

as before.

is

the vector of any given point E, through

or secant EP O PI, parallel to the same

draw a chord

have now,

if

pQ

= OP O

Ta = T

s
#0

being a new scalar

and

by transformations* which
articles

with attention.

similarly, if p l

DPI,

will easily occur to

And

the points P O

any one who has read recent

PI

be together

will

real,

or

together imaginary, according as the quantity under the radical sign

positive
that is, according as we have one or other of the two following
is

or negative

inequalities,

(2.)

The equation (comp. 203,

^-T"
T

J3-

(5.) ),

or

J3

represents a cylinder of revolution, with OB for its axis, and with Ta for the
radius of its base.
If E be a point of this cylindric surface, the quantity
* It does not seem to
be necessary, at the present stage, to supply so
it has hitherto been thought useful to give

Articles, or sub-articles, as
time to time, be given.

many
;

references to former

but such

may

still,

from

ELEMENTS OF QUATEBMONS.

224

under the radical sign

become equal. In this

case,

i.

13.

and the two roots #, %\ of the quadratic

then, the line through E, which is parallel to OB,

vanishes

(1.)

[II.

evident geometrically, since the

cylinder envelopes the sphere (comp. 204, (12.) ), and the line is one of its
If E be internal to the cylinder, the intersections P O PJ are real;
generatrices.
touches the given sphere;

as is otherwise

but

E be external to the same surface, those intersections are

if

ideal,

or

imaginary.

In

(3.)

this last case, if

and

we make,
and

Q,

for abridgment,

being thus two given and real

where v/ -

we may

scalars,

write,

o/c? and ordinary imaginary symbol of

Algebra, and is wotf
with any sort of Geometrical Interpretation* We merely express

1 is the

invested here

thus the fact of calculation, that (with these meanings of the symbols a, ]3,
and t) the formula Ta = T (e + #)3), (1.), wfow treated by the rules of quater

nions, conducts to the quadratic equation,

(x

which has no real

root

s)

0,

the reason being that the right

the present case, wholly external

intersect it at all;

to

the sphere,

line

through E

and therefore

is,

in

does not really

although, for the sake of generalization of language,

we may

agree to say, as usual, that the line intersects the sphere in two imaginary
points.

We

must however agree, then, for consistency of symbolical expression,

to consider these two ideal points as having determinate but imaginary vectors,
namely, the two following
(4.)

in

which

it is easy to prove, 1st, that the real part e +

the vector t of the
sj3 is
E
the
let
line through E
fall
o
on
the
from the centre
foot
of
perpendicular
which is drawn (as above) parallel to OB and Ilnd, that the real tensor tTj3
of each
of the coefficient of */ - 1 in the
;

imaginary part

the length of a tangent E

foot,

E"

to the sphere,

expression, represents

drawn from that external

Compare again the Note

to

page 87, and Art. 149.

point, or

IMAGINARY INTERSECTIONS.

ARTS. 214, 215.]

(5.)

In

we

fact, if

E E =

write OE =
t

= -

s]3

= |3S

which proves the 1st assertion

And

imaginary.

-~

shall

have

= projection of OE on OB
whether the points P O

(4.),

PI

be real or

because

"

we

we

sj3,

225

have, for the case of imaginary intersections,

tT(3

and the Ilnd

(6.)

An

assertion

= v/(T

- Ta2 =
)

/2

E",

(4.) is justified.

expression of the form

(4.),

or of the following,

and y are two real vectors, while ^/ - 1 is the

of algebra, and not a symbol for a geometrically real right
may be said to be a BIVECTOR.
in which

)3

(scalar)

imaginary

versor (149, 153),

manner, an expression of the form (3.), or # = s + t^/ - 1,

where s and t are ^wo real scalars, but y/ - 1 is still the ordinary imaginary
of algebra, may be said by analogy to be a BISCALAR.
Imaginary roots of
(7.)

In

like

algebraic equations are thus, in general, biscalars.

(S.)

And

if

a bivector

such as

(6.)

- =

be divided by a

+ ^

a
in which q Q

may

and

-l

(real)
/

are two real quaternions, but v/ - 1

vector,

the quotient,

-l
is,

as before, imaginary,

be said to be a BIQU ATE UNION.*

215.

The same

distributive principle (212)

tions respecting circumscribed cones,

and the

maybe employed

in investiga

tangents (real or ideal), which can

be drawn to a given sphere from a given point.

(1.) Instead of conceiving that o, A, B are three given points, and that
limits of position of the point E are sought, as in 214,
(2.), which shall allow
the points of intersection P O PI to be real, we may suppose that o, A, E
(which
may be assumed to be collinear, without loss of generality, since a enters only
by its tensor) are now the data of the question and that limits of direction of
,

Compare the second Note

to page 133.

[This word

is

used in a different sense by

Clifford.]

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

226

the line OB are to be assigned, which shall permit the same reality

being

drawn

still

N-- N
the quadratic in x

may

Ta =

T(e +

that

is,

roots are real

by

#|3)

and squaring, we have

TE,

therefore be thus written,

and unequal, or real and equal, or imaginary, according as

according as

EOB

sin
(3.)

Pi

+ x

and its

EP

parallel to OB, as in 214, (1.).

Dividing the equation

(2.)

13.

i.

[II.

or = or

<

>

OA

If E be interior to the sphere, then Te

<

OE.

Ta, T(a

can never exceed unity (by 204, XIX., or by 210, XV., &c.)
fore, in this case, the first of the three recent alternatives,

>

e)
;

but

we

TVUy

have, there

and the two

roots of

the quadratic are necessarily real and unequal, whatever the direction of /3 may
be.
Accordingly it is evident, geometrically, that every indefinite right line,

drawn through an internal

and real points.
(4.)
first

point,

If the point E be superficial, so that TE = Ta,

alternative (2.)

therefore

TVU

(]3

when

(a

in

e)

two

(5.)

e)

the angle EOB

But

TYU

exceed

finally, if E
(j3

E)

it; so that

then the

1,

and
except at the limit for which /3 JL
in which case we have the second alternative.

1,

is

is
:

now =

is
]3),

right.

In

0, for every direction

of

|3

may

and the

likewise always real, but vanishes for the

short,

we have

here the same system of

chords and of tangents, from a point upon the surface, as in 213

difference being, that we now write E for A, or for a.

then

distinct

still exists,

One root of the quadratic in x

other root, namely x = - 28 (s
case

must cut the spheric surface

be an external point, so that

>

the only

To, and T(a

e)

<

1,

either fall short of this last tensor, or equal, or

any one

of the three alternatives (2.)

may come

to exist,

according to the varying direction of j3.

(6.) To illustrate geometrically the law of passage from one such alter
native to another, we may observe that the equation

or

sin

EOF =

OA

OE,

CIECTJMSCEIBED CONES.

ART. 215.]

227

thus external) a real cone of revolution, with its vertex

and according as the line OB lies inside this cone,
at the centre o of the sphere
represents (when E

is

or on

it,

or outside

the

it,

or the second

first

or the third of the three alternatives

is

to be

line

adopted

through

drawn

E,

(2.)

the

or in other words,

parallel (as before)

to OB, either cuts the sphere, or touches

or does not (really) meet

the annexed

it

at all.

it,

(Compare

52.)

fig.

an external point, the

cone of tangents which can be drawn from it
(7.) If E be

still

and the equation

to the sphere is real;

its

vertex at E,

may

one form of

to p

by simply changing p

therefore, or at

it is,

it is,

TVU P
(8.)

Fig. 52.

be obtained from that

of the recent cone (6.),

least

of

with

this enveloping or circumscribed cone,

In general,

if

~
rn

(7.)

-:-

or

OEP =

sin

OA

any quaternion, and x any

q be

YTJ
the recent equation

T-;

(q

x)

Y?

(q

+ x)

rn

YQ
-:e).

rp
JL

OE.

scalar,

therefore be thus written

may

a
,

p-e

or
if

P P

EP =

P be the foot of the perpendicular

first

quotient
(9.)

is

=
evidently

We may

OA

let fall

OE,

from p on OE

and in

fact the

sin OEP.

also write,

-;

or

=s- N
I_N= N -N=I

or

as another

form of the equation of the circumscribed cone.

(10.) If

then

we make

also

"

"a

to express that the point P is on the enveloped sphere , as well as

on the enveloping

2G

ELEMENTS OF QUATERNIONS.

228

[II.

we find the following equation of the plane of contact, or of what

the polar plane of the point E, with respect to the given sphere
cone,

is

i.

13.

called

S-N-Y=0;
/
while the fact that

it is

the exponent 2, or by
(11.)

The

its

se-N

or

a plane of contact*

= 0;

exhibited by the occurrence of

is

equation entering through

its square.

vector,

N a = OE

= S C =

that of the point E in which the polar plane (10.) of E cuts perpendicularly
the right line OE ; and we see that
is

Tf Te = Ta2
.

as

was

to

or

OE

OE = (T OA) 2
.

be expected from elementary theorems, of spherical or even of plane

geometry.
(12.)

formed

The equation

the polar plane of

(10.), of

be thus trans

Sl-N^ = 0;

or

we

may easily

Bl-(s.Ni-^N-,
\
P
/
P
it

E,

If then
continues therefore to hold good, when and p are interchanged.
take, as the vertex of a new enveloping cone, any point c external to the

and

situated on the polar plane

FF
of the former external point E,
the new plane of contact, or the polar plane DD
of the new point c, will pass
the
E
a
former
vertex
of reciprocity, or of con
relation
through
geometrical
sphere,

jugation, between the

which

it

(13.)

two points

and

which

indeed well known, but

by quaternions t anew.

may

8^ = N^,
P

also be thus written,

^i.Nfi-V^K^
a p
a J
a
a
* In

is

In general, each of the two connected equations,

S^ = N,
which

E,

appeared useful for our purpose to prove

fact a

modern geometer would

say, that

we have

= s.

K,
aa

here a case of two coincident planes of inter

merged into a single plane of contact.

t In fact, it will easily be seen that the investigations in recent sub-articles are put forward,
almost entirely, as exercises in the Language and Calculus of Quaternions, and not as offering any

section,

geometrical novelty of result.

POLAR PLANES, CONJUGATE POINTS.

ART. 215.]

220

be said to be a form of the Equation of Conjugation between any two

points P and P (not those so marked in fig. 52), of which the vectors satisfy

may

expresses that those two points are, in a well-known sense, con

= Ta.
jugate to each other, with respect to the given sphere, T/o
it

because

it

(14.) If one of the

two

points, as p

be given by

its

vector p

while the

and vector p are variable, the equation then represents & plane
namely, what is still called the polar plane of the given point, whether

other point P
locus

that point be external or internal, or on the surface of the sphere.

(15.)

Let

P be thus two conjugate points and let it be proposed to

Assum
in which the right line PP intersects the sphere.
s, s

P,

find the points

ing (comp. 25) that

= a =
xp + yp

os

x + y =

1,

and attending

by

To-

to the equation of conjugation (13.),

the
200, VII.,
following quadratic equation in y

= Ta,

we
:

have,

by 210, XX.,

or

x,

which gives

(16.)
real,

Hence

it is

evident that,

one of the two points

p,

if

the points of intersection s, s are to be

interior, and the other must be

must be

because, of the two norms here occurring, one must be

And because the two roots of the
greater and the other less than unity.
or
the
two
differ
values
of
quadratic,
y x,
only by their signs, it follows

exterior to the sphere

(by 26) that the right line PP is harmonically divided (as indeed it is well
known to be), at the two points s,.s at which it meets the sphere or that in
a notation already several times employed (25, 31, &c.), we have the harmonic
:

formula,

(pspV) =
(17.)

From

1.

a real but internal point p,

we can

drawn to the sphere but if so,

*
are
and must consider
ideal, or imaginary
tangents
tangents, as being

an imaginary

circle

by quaternions,

as

of contact-, of

still

speak of a cone of

we must say

on

which the real but wholly external plane

is,

by modern geometry, recognised

as being (comp. (14.)

the polar plane of the supposed internal point.

Compare again the Note

that those

as terminating

them

to

page 88,

arid others

formerly referred

to.

ELEMENTS OP QUATERNIONS.

230

Some

216.

readers

may

find

[II.

i.

13.

useful, or at least interesting, to see here

it

a few examples of the application of the General Distributive Principle (212)

of multiplication to the Ellipsoid, of which some forms of the Quaternion

Equation were lately assigned (in 204, (14.) ) especially as those forms have
been found to conduct* to a Geometrical Construction, previously unknown,
;

for that celebrated

and important Surface

In what follows,

tions.

it

or rather to several such construc

will be supposed that

any such reader has made

himself already sufficiently familiar with the chief formulae of the preceding
and therefore comparatively few referencesf will be given, at least
Articles
;

upon the present

(1.)

To

subject.

prove,

that the locus of the variable ellipse,

first,

*2

which locus

represented by the equation,

is

II...
the two constant vectors

each other at

204, (13.)

-1>

(se)-(v

a, j3

204, (14.)

!,

being supposed to be

and

real,

some acute or obtuse (but not riglitj) angle,

second order, in the sense that

by an arbitrary

cut

it is

to be inclined to

is

a surface of the

rectilinear transversal

more than two, let us assume two

=
=
and let us seek to deter
OL,
OM, as given
points L, M, or their vectors A
mine the points p (real or imaginary), in which the indefinite right line LM

and

in two (real or imaginary) points,

in no

JJL

intersects the locus II.

number

or rather the

of such intersections,

which

will

be sufficient for the present purpose.

(2.)

Making then

p =

-^

(25),

we

have, for y

the following quad

*,

ratic equation,

III. ..

/a*
yB-

a /A
*SM

without proceeding to resolve which,

* See the
Proceedings of the Royal Irish

we

TT
T7^\
yV 3 +sV

- f

see already,

Academy

by

its

(y

mere

degree, that

for the year 1846.

f Compare the Note to page 223.

J If

j3

which the

o,

the system

I.

real or ideal, in
represents (not an ellipse but) a pair of riyht lines,
the second equation of that system, is cut by a plane
denoted

cylinder of revolution,

parallel to its axis }

and represented by the

by

first

equation.

EQUATION OF ELLIPSOID, RESUMED.

AET. 216.]

number sought

the

is

two

and therefore that the locus

II.

231

is,

as

above stated,

a surface of the second order.

The equation

(3.)

remains unchanged, when - p

II.

the surface has therefore a centre, and this centre

(4.)

written

is substituted for p
the origin o of vectors.
;

It has been seen that the equation of the surface

may

also be thus

IV.
it

is at

.T(S

204, (14.)

V|)=I;

gives therefore, for the reciprocal of the radius vector from the centre, the

expression,

iTfsSe+vI

v..
and

this expression has a real value,

value

may

assigned to p
(5.)

which never vanishes,* whatever real

is, whatever direction may be

be assigned to the versor U/o, that

:

the surface

is

therefore closed,

&&& finite.

Introducing two new constant and auxiliary

vectors, determined

by

the two expressions,

VT
VI.

7 = ^-f
p+a
-

a
d

a,

2P
-

- a

a,

which give (by 125) these other expressions,

"*-jrr.-ft

>-j-.-ft

we have

^ + 7 --2

VII
..

and under these

jo

70

7 is said to be the harmonic mean between the two

former vectors, a and ]3
and in like manner, 8 is the harmonic mean
between a and - ]3 while 2a is the corresponding mean between 7,
and
is
between
and
S.
so,
2/3
y
conditions,
;

(6.)

Under

the same conditions, for any arbitrary vector

/o,

we have

in

the

transformations,

* It is to be
remembered that we have excluded in (1.) the case where
can be shown that the
equation II. represents an elliptic cylinder.

j8

which case

it

ELEMENTS OF QUATERNIONS.

232

the equation IY. of the surface

therefore be thus written

may

(]~1;

[II.

i.

13.

X ...T^ + K^ = 1;

orthus,

meaning of which new forms will soon be seen.

The system of the two planes through the origin, which

the geometrical
(7.)

perpendicular to the

tively

new

vectors

y and

represented by the

is

S,

are respec

equation,

or

...

combining which with the equation

XIII.

- s

II.

ft/;

we get

or,

XIV.

j;

TP

T/3.

These two diametral planes therefore cut the surface in two circular sections,
with T/3 for their common radius and the normals y and S, to the same two
;

planes, may be called (comp. 196, (17.)) the cyclic normals of the surface;
while the planes themselves may be called its cyclic planes.
(8.)

Conversely,

centric sphere

if

XIV.,

we seek the

intersection of the surface with the con

which the radius

of

is T/3,

we

are conducted to the

equation XII. of the system of the two cyclic planes, and therefore to the
so that every radius vector of the surface, Which is
circular sections (7.)

two
not

drawn

in one or other of these

two planes, has a length either greater or

than the radius T/3 of the sphere.

less

By all these marks, it is clear that the locus II., or 204, (14.), is
above
its centre being at the origin (3.), and its
(as
asserted) an Ellipsoid
mean semiaxis being = T/3 while U/3 has, by 204, (15.), the direction of
(9.)

the axis of a circumscribed cylinder of revolution, of which cylinder the radius is

T/3 and a is, by the last cited sub-article, perpendicular to the plane of the
;

ellipse

of contact.

(10.)

Those who are familiar with modern geometry, and who have

caught the notations of quaternions, will easily see that this ellipsoid, II. or
IY., is a deformation of what may be called the mean sphere XI V., and is
the infinitely distant point in the direction of /3 being a
centre ofhomology, and either of the two planes XI. or XII. being a plane of

homologous thereto

homology corresponding.
217. The recent form, X. or
of the quaternion equation of the ellip
.,
soid admits of being interpreted in such a way as to conduct (comp. 216) to

CIRCULAR SECTIONS, CYCLIC PLANES.

ARTS. 216, 217.]

which
a simple construction of that surface
calculation, and then illustrate by geometry.
;

Carrying on the

(1.)

Eoman

we

233

shall first investigate

by

numerals from the sub-articles to 216, and

observing that (by 190, &c.),

y
the equation

X.

>

takes the form,

1TV
XV...

T |(_
J/

1
1

or

if

we make
7m^ = ^

when

and

which we

two new constant

K are

Trn =

and

vectors,

shall suppose to be positive, but of

and

T;>

t is

new

constant scalar,

which the value

may

be chosen

at pleasure.

The comparison

(2.)

interchanged,

although they
to assume that
ellipsoid,

XVIII.

T7

for the supposition

T (]3 + a)
which

latter case

(3.)

We

X. and

of the forms

>

Ty
=

may

not at

first

seem

T (]3 - a),

it is

XVIIF.

and

.-.

was excluded in 216,

(by 186,

(6.)

(1.).

have thus,

XXII.

TLc =

triangle, such that

.

CB =

ELEMENTS OF QUATERNIONS.

CA = K

/,

AE =

let also
S

do so

to

p.

U;

&c.)

may

equation of

be
the

therefore allowed

Tt

>

T*

= TS would give, by VI.,

Let ABC be a plane

HAMILTON

shows that y and S

and therefore that

TS,

XIX.. .TL = US

(4.)

or that they enter symmetrically into the

)3 J_

ELEMENTS OF QUATERNIONS.

234

[II.

i.

13.

Then if a sphere, which we shall call the diacentric sphere, be described round
the point c as centre, with a radius = T/c, and
L\T

therefore so as to pass through the centre A

(here written instead of o) of the ellipsoid,

and

if

D be the point

in

which the

AE

line

meets this sphere again, we shall have, by

213,

(5.), (13.),

KK-

XXIII.

and therefore

xxur.

-i +

so that the equation

XXIV.

K-.p;

XVI. becomes
?=

AE

Fig. 53.

DB.

The

point B is external to the diacentric sphere (4.), by the assump

a real tangent (or rather cone of tangents) to this sphere can
and if we select the length of such a
therefore be drawn from that point
(5.)

tion (2.)

we make each
equal to unity, and denote by D the second
BD with the sphere, as in fig. 53, we shall have

tangent as the value (1.) of the scalar

member

of the formula

that

t,

is

to say, if

XXI.

intersection of the right line

(by Euclid III.) the elementary relation,

XXY.
whence follows

f=

"BD,

DB

AE =

7
AE = BD

where AK denotes the length of the

let

BD

BD

somewhat more familiar notation,

XXVII.

fig.

this Geometrical Equation of the Ellipsoid,

XXYI.
or in

line AE,

and similarly

for

BD

(6.) The following very simple Rule of Construction (comp. the recent
53) results therefore from our quaternion analysis
From a fixed point A, on the surface of a given sphere, draw any chord AD;
:

same spheric surface with the secant

and take a radius vector AE, equal in

be the second point of intersection of the

drawn from a fixed external* point B

* It is
merely to fix the conceptions, that the point B is here supposed to he external (5.) ; the
calculations and the construction would he almost the same, if we assumed B to he an internal point,

or

Ti<T/c,

T7

<T5.

CONSTRUCTION OF THE ELLIPSOID.

ARTS. 217, 218.]

length to the line

AD

chord

BD

the locus

of the point E

B/or a point of its

ivith

an

ellipsoid, with

A for

to,

the

its centre,

and

surface.

of a plane but variable quadrilateral

//,

length

will be

Or thus

(7.)

and

in direction either coincident with, or opposite

and

235

and

ABED

in position, the two diagonals AE,

if their intersection

whereof the side

BD

be always situated

AD of the

quadrilateral

is

of which one side AB?S given in

be equal to each other in length,

upon the surface of a given sphere,

a chord, then the opposite side

BE

is

chord of a given
218.

From

either of the

two foregoing statements, of the Rule of Con

which quaternions have conducted, many geome

trical consequences can easily be inferred, a few of which may be mentioned
the present Calculus being, of
here, with their proofs by calculation annexed

struction for the Ellipsoid to

course,

still

(1.)
is

employed.

That the corner

B, of

what may be

called the Generating Triangle ABC,

in fact a point of the generated surface, with the construction 217,

(6.),

may be proved, by conceiving the variable chord AD of the given diacentric

sphere to take the position AG where G is the second intersection of the line
;

AB with that spheric

(2.)

If

surface.

D be conceived

(instead of A),

approach to A (instead of G), and therefore D to G

the direction of AE (or of AD) then tends to become tangential
to

to the sphere at A, while the length of

to

become equal

to the length of

circular section, in a plane

BG

AE

(or of

BD

tends,

by the

construction,

the surface has therefore a diametral and

which touches the diacentric sphere at

A,

and with

a radius = BG.

Conceive a circular section of the sphere through A, made by a plane

perpendicular to EC if D move along this circle, D will move along a parallel
(3.)

of BD or that of AE, will again be equal to

through G,
such then is the radius of a second diametral and circular section of the

and the length

circle

BG

ellipsoid,

made by

the lately mentioned plane.

The

construction gives us thus two cyclic planes through A

the per
which planes, or the two cyclic normals (216, (7.) ) of the
ellipsoid, are seen to have the directions of the two sides, CA, CB, of the
(4.)

pendiculars to

generating triangle ABC

(1.).

ELEMENTS OE QUATERNIONS.

236
(5.)

13.

i.

[II.

Again, since the rectangle

BD = BD

BG = BD

BA

we have the

AE = double area

ABE

of triangle

sin BDE,

equation,

XXVIII.
the third

perpendicular distance of E from AB = BG

side, AB, of the generating triangle

which

lution of a cylinder,

(1 .), is

envelops the ellipsoid,

and

sin

BDE

therefore the cms of revo

of

which the radius has

the same length, BG, as the radius of each of the two diametral and circular
sections.
(6.)

For the points

and

of contact of ellipsoid

we have

cylinder,

the

geometrical relation,

XXIX.
the point D

AB

is

BDE = a right angle

XXIX

or

ADB = a right angle

therefore situated on a second spheric surface, which has the line

for a diameter,

and

plane passes through

intersects the diacentric sphere in a circle,

and

A,

cuts the enveloping cylinder in

an

whereof the

ellipse

of contact

(comp. 204, (15.), and 216, (9.) ), of that cylinder with the ellipsoid.
(7.) Let AC meet the diacentric sphere again in F, and let BF meet
in F
(6.)

it again
the
common
the
of
last-mentioned
circle
and
plane
fig. 53)
ellipse
can then be easily proved to cut perpendicularly the plane of the gene

(as in

rating triangle ABC in the line AF

of contact

and therefore

so that the line F B

is

normal

also (by conjugate diameters, &c.)

to this

plane

to

the ellipsoid,

to

which

at B.
(8.)

These geometrical consequences of the construction (217),

others might be added, can all be

shown

many

and confirmed

to be consistent with,

by, the quaternion analysis from which that construction itself was derived.
Thus, the two circular sections (2.), (3.) had presented themselves in 216, (7.)

and

their

two

cyclic

normals

(4.),

or the sides CA, CB of the triangle, being

(by 217, (4.)) the two vectors K, i, have (by 217, (1.) or (3.) ) the directions of
the two former vectors 7, S; which again agrees with 216, (7.).
(9.)

Again,

it

will be

pairs of constant vectors,

found that the assumed relations between the

a, )3

7, 8

and

L,

K,

any one

of

which pairs

determine the ellipsoid, conduct to the following expressions

investigation is left to the student, as an exercise)
to

XXXI.

../3

-7 =

0-7

-8 = fpT--rTJ
~
1

0~7

(l

K)

(<-)-

is sufficient

(of

BO

three

which the

CONSEQUENCES OF THE CONSTRUCTION.

ART. 218.]

237

G referring here to fig. 53, while aj3y retain their former

meanings (216), and are not interpreted as vectors of the points ABCD in that
Hence the recent geometrical inferences, that AB (or BG) is the axis
figure.

the letters

B,

an enveloping cylinder

of revolution of

(5.),

and that F B

normal

is

to the

plane of the ellipse of contact (7.), agree with the former conclusions (216,
such a normal.
(9.), or 204, (15. ), that ]3 is such an axis, and that a is
j

(10.)

It

easy to prove, generally, that

is

gg +

(g-l)(Kg+l) = Ng-1
N(g + l)
(q + l)(Kq + l)

g-1
q +1

Ng-1

q-1

N(g-l)

whence
-

b
whatever two vectors

and

XXXIII.

the recent expressions

t*

(9.)

XXXIY.

_TY-TV
~

last

now

thus written

form 204,

TV -TV

But we have

be.

may

= 0V - TV, by 217,

and

for a

a = +

+
(<

]3

K)

(14.), of the

(5.)

here,
;

become, therefore,

^-^
+

The

l
)

- -

|3

(i

K)

equation of the ellipsoid,

may

therefore be

+ K

which the sign of the right part may be changed. And thus we verify by
calculation the recent result (1.) of the construction, namely, that B is a point
in

of the surface

for

we

XXXVI.

see that the last equation

.

/0

= AB =

-ic =

j3

is satisfied,

when we suppose

:S^;

a value of p which evidently satisfies also the form 216, IV.

(11.) From the form 216, II., combined with the value XXXIV. of
is

a, it

easy to infer that the plane,

xxxvu.

1,

tt

or

xxxvir.

s -^- = s
I

which corresponds to the value x = 1 in 216, I., touches the

point B, of which the vector p has been thus determined (10.)
the surface, at that point, has therefore the direction of

of FB, or of F B

by

K,

+ K

ellipsoid at the
;

the normal

or of a, that

to
is,

so that the last geometrical inference (7.) is thus confirmed,

calculation with quaternions.

ELEMENTS OF QUATERNIONS.

238
219.

noted

few other consequences of the construction (217) may be here

especially as regards the geometrical determination of the three prin

cipal semiaxes of the

elliptic

13.

i.

[II.

and diametral

of spherical

and the major and minor semiaxes

ellipsoid,

of

any

together with the assigning of a certain system

of which the surface may be considered to be the locus.

conies,

section

denote the lengths of the greatest, the mean, and the least
semiaxes of the ellipsoid, respectively then if the side BC of the generating
(1.)

Let

<?,

I, c

triangle cut the diacentric sphere in the points H and H the former lying
(as in fig. 53) between the points B and c, we have the values,
,

XXXVTTI

u\.J-\-J\. V -LJLJ..

n - :HM
RT?
Cl

==

h
U

TRT
JuGr

- TVFT
BH

r
C

so that the lengths of the sides of the triangle

ABC

may

be thus expressed, in

terms of these semiaxes,

"\T"

"\T"

~\TT

"v~

and we may

__

rp

-,_,

rri

rn

If, in

(2.)

rp.

write,

~v T
AL.

AH

T^
1/c

a = T^
It +

b
z>

rp

p - nn
= It
JU.
f

the respective directions of the two supplementary chords AH,

of the sphere, or in the opposite directions,

we

set off lines AL, AN,

with

the lengths of BH , BH, the points L, N, thus obtained, will be respectively a

major and a minor summit of the surface. And if we erect, at the centre A of
that surface a perpendicular
the point M (which will be

AM to the plane of the triangle, with a length = BG,

common to the two circular sections, and will be

situated on the enveloping cylinder) will be a

mean summit

thereof.

(3.) Conceive that the sphere and ellipsoid are both cut by a plane through
on which the points B and c shall be supposed to be the projections oi: B
and c then c will be the centre of the circular section of the sphere and if

A,

the line B C cut this

new

circle in the points

D I?

D*, of

which D may be sup

X

posed to be the nearer to B the two supplementary chords AD I? AD 2 of the

circle have the directions of the major and minor semiaxes of the elliptic section
,

of

the

ellipsoid

while

the

lengths

semiaxes

those

of

BA BG BD I? and BA BG BD 2 or BD^ and BD^,

meet the sphere again in D and D 2
.

(4.)

If these
t

two semiaxes

of the section be called a,

the tangent from B to the

sphere,

XLI.

BDi =

respectively,

denote by

are,

the secants BD : and BD 2

if

2
:

a / = aca~
/

l
;

we have

Bi) 3

= f

and

and

c^

thus,
c.

= ace;
i

1
;
*

if

we

still

SEMIAXES, SPHERICAL COLICS.

ART. 219.]

wo denote by;?

the inclinations of the plane of the section to the

and CB are perpendicular, so that

but

if

two

cyclic planes of the ellipsoid, whereto CA

and/> 2

239

the projections of these two sides of the triangle are

(C

ALdl.

A = CA

}-7

VC B

= CB

sin/^ = J (a

p2 = \

sin

(a

sin

p lt

sin
c)

;; 2 ,

c)

we have
BD 2 2 - BDi 2 = B D 2 2 - B
r

XLIII.

D!

= 4B

/
.

C A

2
)

sin/?! sin jt? 2

whence follows the important formula,

XLIV.

c~z - a

known and

or in words, the

squares of the seniiaxes, of a

c-

a-

useful theorem, that

plane and diametral

"

the difference of the inverse

section of

an

ellipsoid, varies as

the product of the sines of the inclinations of the cutting plane, to the two planes of

circular section.

As

(5.)

verifications, if the plane be that of the

generating triangle ABC,

we have

but

if

PI or

the plane be perpendicular to either of the two sides, CA, CB, then either

p =

0,

and

c/

a,.

(6.) If the ellipsoid be cut

sphere

XI V.,

by any concentric

mean

sphere, distinct from the

so that

XLV.

AE =

T/o

=r

<

b,

where r

is

a given positive scalar

then

XLYI.

BD = f r

<

acb~ l ,

<

that

is,

BA

so that the locus of what may be called the guide-point

through which, by
the construction, the variable semidiameter AE of the ellipsoid (or one of its
prolongations) passes, and which is still at a constant distance from the given
external point B,,is now again a circle of the diacentric sphere, but one of
r>,

which the plane does not pass (as it did in 218, (3.) ) through the centre A of
the ellipsoid.
The point E has therefore here, for one locus, the cyclic cone

which has A

for vertex,

and

rests

on the last-mentioned

circle as its base

and

on the concentric sphere XLV., it must be on one or other of

the two spherical conies, in which (comp. 196, (11.)) the cone and sphere last
since

it is

also

mentioned

intersect.

ELEMENTS OF QUATERNIONS.

240

[II.

13.

i.

of an ellipsoid uith a concentric sphere is therefore,

with the value of the radius r,
generally, a system of two such conies, varying
two
the
circular sections, for the parti
the
as
a
and becoming,
limit,
system of
cular value r = b\ and the ellipsoid itself may be considered as the locus of

The

(7.)

all

intersection

such spherical conies, including those two

And we

(8.)

see,

by

(6.),

that the

of any one of the concentric cones,

from

ttco cyclic

which

the two cyclic planes of the ellipsoid

circles.

planes (comp. 196, (17.), &c.)

on any such conic, coincide with

resulting, with the greatest ease,

rest

all this

which quaternions had conducted.

the figure 53, which was designed to illustrate that

the construction (217) to

respect to

"With

(9.)

construction, the signification of the letters

explained.

But

a second minor

summit

what may be

of the ellipsoid,

we may

made by

of

here add,

AN = NA

of the surface, so that

point in which the chord AF

AGF, intersects

ABCDD EFF GHH LN has been already

as regards the other letters

what we may here

Ilnd, that

the plane of the greatest and

the side

Illrd, that the tangent,

AB of

VKV

is

is

call the diacentric circle

called the principal ellipse* or the section

least axes,

plane of the generating triangle ABC, so that the lengths of

equal

1st, that

that

is

NBLEN
by the

AK and BF

are

to this ellipse at this point, is parallel to

the triangle, or to the axis of revolution of the enveloping cylinder

that
(5.), being in fact one side (or generatrix) of that cylinder; IVth,
the
AB
semidiameters
of
and
therefore
are
the
thus
two
AK,
conjugate
ellipse,
tangent TBT at the point B of that ellipse, is parallel to the line AKF or per

218,

BFF

thus the normal (comp. 218,

to
the
same
and
therefore
also to the ellipsoid, at B
(7.), (11.))
elliptic section,
Vlth, that the least distance KK between the parallels AB, KV, being = the
pendicular to the line

Yth, this latter line

is

radius b of the cylinder, is equal in length to the line BG, and also to each of
the two semidiameters, AS, AS of the ellipse, which are radii of the two circular
sections of the ellipsoid, in

planes perpendicular to the plane of the figure

that
AS
touches the circle at A ; and YHIth, that the point s is on the
Yllth,
chord AI of that circle, which is drawn at right angles to the side BC of the
;

triangle.

220.

The reader

will easily conceive that the quaternion equation of the

admits
of
ellipsoid
being put under several other forms
among which,
however, it may here suffice to mention one, and to assign its geometrical
;

interpretation.
* In the
plane of what

is called,

by many modern geometers, the focal hyperbola of the

ellipsoid.

SEMIAXES, SPHERICAL CONICS.

ARTS. 219, 220.]

For any three

(1.)

vectors,

we have

K, p,

*,

241

the transformations,

N- + N-+2S-p

p p
K

-N--fN-N- + 28--T TK

p p

-T- + K-T-VN(-T- + K-Tp

Kj

whence follows

\p

this other general transformation

XLYIII.

\
.

pi

K.

TT

TT"

JL

p
(2.) If then we introduce
defined by the equations,

T~V
AijlA.
"VT

two new auxiliary and constant

= - TT
U/c

and /,

TT
K=U.K,
T^

nn

1*,

vectors,

which give
L.

T/ =

we may write

the equation

precisely similar form

in

and

XYI.

ditions,

LII.

locus

of the point

AC = - K,

The only

(4.)

If

ic

=
)

T/ - T/2 =
2

K) ,

(*

is,

2
,

217) of the ellipsoid under the following

(in

and a new generating

"

D"D

C B

triangle,

AB

lengths of the sides are the

=+

(instead of

and therefore

BDD

),

and

will be the

same

diacentric

where B

sM0, by

AB =

set off a line

is

the con

AE in the direction

or in the opposite direction, with a length equal to that of

struction

draw any secant B

AD",

wew ,#red external point, but the

of

T (t

Tic,

with a new centre c

sphere,

have simply taken the places of and K.

Eetaining then the centre A of the ellipsoid, construct a new

which
(3.)

T/c

Tt,

BD W

the

ellipsoid as before.

inference which

that there exists (as

we
is

shall here*

draw from

this

new con

known) a second enveloping cylinder of

room

shall allow, a few additional remarks may be made, on the relations of the constant
&c., to the ellipsoid, and on some other constructions of that surface, when, in the
following Book, its equation shall come to be put under the new form, T(ip+ pa) = 2 - i 2 .
[See

vectors

t,

/c,

404.]

HAMILTON

ELEMENTS OF QUATERNIONS.

2 I

ELEMENTS OF QUATERNIONS.

242
revolution,

and that

axis

its

the side AB of the

is

the radius of this second cylinder

mean semiaocis, b, of the ellipsoid

AL

BAB

in fig. 53, bisects the angle

two circumscribed cylinders

new

triangle

equal to that of the

is

and that the major

AB

first,

but that

namely

to the

semiaxis, a, or the line

the plane of the

new

ellipse

of contact being

to that

and being perpendicular to the new

of contact was (by 218,
+ K,
(11.) ) to

+ K,

(7.)

between the two axes of revolution of these

geometrically determined by a process exactly similar

218,

13, 14.

i.

[II.

vector,

employed in

as the old plane

SECTION

On

14.

the Reduction

of the General Quaternion to a Standard

Qua drill omial Form ; with a First Proof of the Associative

Principle of multiplication of Quaternions.

01,

221. Retaining the significations (181) of the three rectangular unit-lines

GJ, OK, as the axes, and therefore also the indices (159), of three given

right versors, i,j,

index OQ of

any

k,

we can

in three mutually rectangular planes,

express the

under the trinomial form

other right quaternion, such as V#,

(cornp. 62),
I.

TVq = OQ = x

01

+ y OJ +
.

OK

where xyz are some three scalar coefficients, namely, the three rectangular
co-ordinates of the extremity Q of the index, with respect to the three axes
Hence we may write also generally, by 206 and 126,
01, OJ, OK.
II.

V<?

xi

+ yj + zk =

ix +///

+ kz

this last form, ix + jy + kz, may be said to be a Standard Trinomial Form,

which every right quaternion, or the right part
of any proposed quater
nion q, can be (as above) reduced.
If then we denote by w the scalar part, Sq,

and

to

"Vq

of the

same general quaternion

General Reduction
(183)

of a

Quaternion

we
to

by 202, the following

a STANDARD UUADRINOMIAL FORM
shall have,

III.
in

q,

which the/owr

Quaternion.
in like

And

w+

ix

wxyz,

may

manner,

IY.

+jy +

kz

be said to be the Four Constituents of the

evident (comp. 202, (5.), and 133), that if we write

scalars,
it is

q = (Sq + ~Vq =)

= ?/ +

ix*

+jy +

kz

STANDARD QUADBINOMIAL POEM.

Ants. 220-222.]

where

243

denote the same three given right versors (181) as before, then the

ijk

equation

V. ..? =

?,

and /,

between these two quaternions, q

between the constituents

VI.
which

is

new

includes the four following scalar equations

w =

x =

w,

y =

x,

y,

justification (comp. 112, 116) of the propriety of naming, as

we

have done throughout the present Chapter, the General Quotient of two Vectors
(101) a

QUATERNION.

222.

When

Form

the Standard Quadrinomial

is

(221)

we have

adopted,

then not only

I.

as before, but also,

Sq =

and

+ jy +

ix

V<?

kz,

by 204, XI.,

II.

Kq

= (Sq - V#

=)

w-

- jy -

ix

kz.

And

because the distributive property of multiplication of quaternions

(212),
combined with the laics of the symbols ijk (182), or with the General and

Fundamental Formula

namely with the formula,

of this whole Calculus (183),

=./

=^ = -l,

(A)

gives the transformation,

III.

we

have,

by 204,

(ix

+ jy +

kz)

&c., the following

IV.

V.

VI.
VII.

NV?

X.
XI.

new

= (TV?) 2 = -

Vg

kz)

= Tq = Sq* -

V?

2/

N0

</(&

=
TJq
(w +

SU?

+
ix

x*

(ix

+jy +

XIL..TVU,=

+jy +

v*

),

f+z

z
;

s );
2
v/0* + y* +

= tf +

+ yz +

= w: v/(^2 +

VII? =

y*

..TV?=V(^ +
UV? = (ix +jy +

(x*

expressions

VIII. ..Tq =

IX.

= -

kz)

x*

kz)

\**

2
V to* + x +

f+

z
;

2
;

^(w* +

+y*
:

x*

z*)

-f

^ (w*

x*

y*

g2 )

2
)

x*

y*

2
# +

12

ELEMENTS OF QUATERNIONS.

244
(1.)

To prove

of the multiplication (comp. again 182)

V#
ix

ix

= -

V?

kz

= -

V<?

V? = V?
2

We

+ jy +

kz,

kz-,

+ kxy - jxz

+ iyz,
jzx
izy
1
- z\
=
tf
V?
y
kyx

y>

= -

as follows the steps

x*

V<?

jy

(2.)

we may arrange

the recent formula III.,

z*

-i-

have, therefore,

XIII.
a result to which

(ix

+jy +

kz)

we have already

x*

if

1,

y*

z*

when considered

as denoting a right radial (149),

which the plane or the axis is arbitrary.

= q q, then N?" = N/ N?,
(3.) If

w +

&c.,

tv"

q"

symbol

/ - 1,

or a right versor (153), of

by 191,

1,

alluded,* in connexion with the partial

indeterminateness of signification, in the present calculus, of the

14.

i.

[II.

(8)

but

if

= w +

&o.,

&c., then

= w w - (x x + y y + zz)
= (iv x + xw] + (yz - z y),
= (lo y + y w) + (zx - x z],
f
Iw" = (wz + z w] + (x y - y x)
,

x"

y"

z"

and conversely these four

summed up

be

XV.
we ought

in,

w +

w"

ix"

+jy"

kz"

= (w +

ix

+ jy + kz )(w +

under these conditions XIV.,

z

and may

the quaternion formula,

therefore,

XVI.

scalar equations are jointly equivalent to,

ff

x"*

y^

z
z"

(w/

to

+ xz +

ix

+jy +

kz)

have the equation,

1

y"

) (tv*

+ xz + y z +

which can in
truth

may

fact be verified by so easy an algebraical calculation, that its

be said to be obvious upon mere inspection, at least when the terms

in the four quadrinomial expressions

w"

z"

are arrangedf as above.

Compare the first Note to page 133 and that to page 162.
t From having somewhat otherwise arranged those terms, the author had some
;

little

trouble at

in verifying that the twenty-four double products, in the expansion of w" 2 -f &c., destroy each
other, leaving only the sixteen products of squares, or that XVI. follows from XIV., when he was led
He believes, however, that the algebraic
to anticipate that result through quaternions, in the year 1843.
first,

theorem XVI., as distinguished from the quaternion formula XV., with which
had been discovered by the celebrated EULER.

it is

here connected?

ASSOCIATIVE PRINCIPLE OF MULTIPLICATION.

AKTS. 222, 223.]

223.

The

we

principal use which

form (221)

make

shall here

245

of the standard quadri-

the general associative property of

prove by
now with great ease be done, the
which
can
multiplication of quaternions ;
distributive* property (212) of such multiplication having been already proved.
nornial

In

we

fact, if

is

to

it

write, as in 222, (3.),

q
I.

w +

ix

+jy +

ix

+ jy

+ kz

ix"

+jy"

w"

<f

without

now assuming

that the relation

between the three quaternions

/,

g,

q",

kz,
,

kz",

= q

q, or any other relation, exists

and inquire whether it be true that
q"

the associative formula,

II. ..

holds good,

whether

we

this last

are replaced, in

by the

see,

J"

? .(?

.
q"

q q,

we have only

distributive principle, that

to try

when

formula

is

any one

common

valid

the three quaternion factors q, q ,

order on both sides of the equation, and

q"

with or without repetition, by the three given right versors ijk but this has
We arrive then, thus, at the important
already been proved, in Art. 183.
;

conclusion, that the General Multiplication of Quaternions

an Associative

is

had been previously seen (212) to be a Distributive one

although we had also found (168, 183, 191) that such Multiplication is not
or that the tico products, q q and qq are generally
(in general) Commutative
Operation, as

it

unequal.

member

We may

therefore omit the point (as in 183),

of the equation II.

q.
by the symbol
=
=
=
that
v
so
/,
(1.)
V#,
V?"
V0
and
and
196, 204,
therefore, by 191, (2.),
quaternions,

Let

Kfl

Let

we

t>"

be called

are

v"

t>,

Sv v = ^(v v + vv

this last right quaternion

shall then

may

denote each

q"q

= vv,

and

Nv

),

and

v,,

let

= \(yv =

S0

,,

any three right

vv).

so that v v = sy

+ v,

have the equations,

2

WX = v\ - v/

4\

sf

whence, by addition,
2 Wt>y =

* At a later
stage [II. in.
Principle of Multiplication,

v"

(V

vv -

v")

2], a sketch will be

which

v v

v"

v-v

(v"v

vv")

given of at least one proof of this Associative

will not presuppose the Distributive Principle.

ELEMENTS OF QUATERNIONS.

246

and therefore generally,

III.

if v, v

^vv =

be

v"

a formula with which the student ought

account of

its

With

(2.)

v"$v

The formula
V.

make himself completely familiar, on

to

W*, =
=

t>St>V

first

member may

still

supposed that
f
v = L v = L

v"Svv

v"

tWt>

t/

IV

?"

Wt>

is

= Iv

S0V

- Iv

(by 208,

planes of q
positive

&v"v

evidently complanar with the two indices Iv and

to

third index

the

Iv"

and

q".

the recent notations, the vector,

VI.
is

-.

therefore (by (L)) complanar with the third quaternion

With

Sw

often for simplicity be dispensed

same time (by 208) perpendicular

at the

(4.)

III. gives (by 206),

hence this last vector, which

is

v"v

/_

therefore this other very useful formula,

where the point in the

with and in which it is

Iv

Sv"v

the recent notations,

IV.

(3.)

- v

v"

extensive utility.

V
we have

as above,

still right,

v&v

[ILi.14.

Ivf =

r?Vt>

IV (V/

V?),

XXII.) a line perpendicular to both Iv and It?

and q being also such that the rotation round
;

while

&

or

it

common

to the

from Iv to Iv

is

length }

Tfy,

or
Tf>,,

or

TV

TV

or

tfy

(V/

V?),

Ae unit of length the same ratio, as that which the parallelogram under

foflrs

the indices, Iv
(5.)

To

and Iv,

bears to the unit of area.

interpret (comp. IV.) the scalar expression,

VII.
(because St/

s,

0),

we may employ the formula

208, V.

which gives the

transformation,

VIII.
where

T#"

StfVo =

Tfl"

denotes the length of the

TV,

line

cos

Iv",

(TT

x)

and

T^

represents

area (positively taken) of the parallelogram under Iv and Iv

by

(4.)

while x

is

the

(by

ASSOCIATIVE PRINCIPLE OF MULTIPLICATION.

ART. 223.]

247

This angle will be obtuse,

I0y
208), the angle between the two indices TV",
and therefore the cosine of its supplement will be positive, and equal to the
.

sine of the inclination of the line

of TV and TV,

the plane

to

TV"

to Iv be negative, that

round TV from TV

is, if

the rotation

if

the rotation round Iv from

It?

but that cosine will be equal the negative of this sine, if

We have therefore the important
the direction of this rotation be reversed.
to

TV"

be positive

interpretation

IX.

Sv"v

= + volume of parallelepiped under

TV

TV,

TV"

the upper or the lower sign being taken, according as the rotation round TV,
from TV to TV", is positively or negatively directed.
(6.)

For example, we saw that the ternary products

scalar values,

and

ijk

kji

have

namely,
= -

ijk

kji

1,

4-

by 183,

(1.),

(2.)

and accordingly the parallelepiped of indices becomes, in this case, an unit-cube

while the rotation round the index of i, from that of./ to that of k, is positive
;

(181).

In general,

(7.)

X.

any three right quaternions

= - St? V0
Svv

for
.

v"

vv v

we have the

",

formula,

and when the three indices are complanar, so that the volume mentioned in IX.
so that we may
vanishes, then each of these two last scalars becomes zero
;

write, as a

new Formula
XI.

while,

of Complanarity

Sv"v

on the other hand,

this scalar

quaternions (or their indices) be

it

if

0,

still

TV"

HI

TV

TV (123)

cannot vanish in any other case, if the

supposed to be actual (1, 144) because
;

then represents an actual volume.

(8.)

Hence

we may

also

for any three quaternions

XII.
that

is,

(9.)

establish the following

Formula

of Collinearity,

(Vq"

V?

V?) =

0,

if

IV?"

Ill

TVq

have any common line.

by 209, if the planes of q, q
In general, if we employ the standard trinomial form, 221,
,

V? =

ix

+ jy +

the laws (182, 183) of the symbols

XIII.

Sv"v

x"

i,

(z

q"

kz,

IV?

ix

&c.,

/, k give the

- y
z) +

y"

v"

ix"

II.,

namely,

&c.,

transformation,

(Jz

* *)

z"

(y

x -

ELEMENTS OF QUATERNIONS.

248

and accordingly

known

this is the

initial edges, if the

be 000, xyz,

</* ,

OP

lines OP,

OP"

for three co-

rectangular co-ordinates* of the four corners,

l/

14.

i.

expression for the volume (with a suitable

which has the three

sign) of the parallelepiped,

[II.

o, P,

P",

z".

Again, as another important consequence of the general associative

of
property
multiplication, it may be here observed, that although products
of more than two quaternions have not generally equal scalars, for all possible
(10.)

permutations of the factors, since we have just seen a case X. in which such
a change of arrangement produces a change of sign in the result, yet cyclical
permutation

S or in symbols, that for any three

extended
to any greater number of such
easily

permitted, under the sign

is

quaternions (and the result

factors) the following

is

formula holds good

XIV.
In

fact, to

prove this equality,

XIV.
and

to

remember that the

remains unaltered (198,

In

(11.)

like

XV.

I.),

K?

=
Y<?

we have only
S

(j

scalar

when

manner, by 192,
.

K (/

Y
of

to write

= S

?)

(q

it

q"q

thus,

),

the product of any two quaternions

the order of those two factors

q q)

~Kq q

changed.

be inferred that

may

II., it

is

Kq"

Kq Kq
.

Kq",

with a corresponding result for any greater number of factors whence by

192, I., if Uq and U q denote the products of any one set of quaternions taken
;

in

two

opposite orders,

xvi.
(12.)

XVIII.

But

if

we may

Kuq

write,

K<?

XYII.

K#

v be right, as above, then

..Kn*=n

XIX...SlIt>

t>;

En?

= -

Sn

v,

f;

n B?.

by 144

XX.

hence,

..

upper or lower signs being taken, according as the number of the right factors
is even or odd
and under the same conditions,
;

XXI.
as

was

Sllt>

= ^(Uv

n 0)

XXII.

lately exemplified (1.), for the case

* This result

may

any usual doctrine of

may

ease, important co-ordinate expressions.

Vn? = J

(Uv

where the number

is tico.

manner in which quaternions, although not based on

yet be employed to deduce, or to recover, and often with great

serve as an example of the

co-ordinates,

. .

COMPLANAR QUATERNIONS.

ARTS. 223, 224.]

(13.)

For the

case

where that number

XXIII.

XXIY.

S0Vt>

= -

is three,

= 1

S?t>V

249

the four last formulae give,

- vv

(v v v

v")

WVt? =

which obviously agree with X. and IV.

224. For the case of Complanar Quaternions (119), the power of reducing
each (120) to the form of a fraction (101) which shall have, at pleasure, for its
results

denominator or for its numerator, any arbitrary line in the given plane, furnishes
some peculiar facilities for proving the commutative and associative properties
of Addition (207),
tion (212,

223)

and the

distributive

and

associative properties of Multiplica

while, for this case of multiplication of quaternions,

already seen (191,

we have

(1.)) that the commutative property also holds good, as

it

does in algebraic multiplication.

It may therefore be not irrevelant nor
useless to insert here a short Second Chapter on the subject of such comin treating briefly of which, while

assuming as proved the existence

have an opportunity to say something
of Powers and Roots and Logarithms
and of the connexion of Quaternions
with Plane Trigonometry, and with Algebraical Equations. After which, in

planars

of all the foregoing properties,

we

shall
;

the Third and last Chapter of this Second Book, we propose to resume, for a
short time, the consideration of Diplanar Quaternions
and especially to show
;

how

the Associative Principle of Multiplication can be established, for them,

without* employing the Distributive Principle.
*

HAMILTON

Compare the Note

ELEMENTS OF QUATERNIONS,

to

page 245.

ELEMENTS OF QUATERNIONS.

250

CHAPTER

[II. n.

1.

II.

ON COMPLANAR QUATERNIONS, OR QUOTIENTS OF VECTORS

IN ONE PLANE AND ON POWERS, ROOTS, AND LOGARITHMS
;

OF QUATERNIONS.
SECTION

1.

On Complanar Proportion of Vectors

Fourth Proportional to
Third
to
Proportional
Three,
Two, Mean Proportional,
Square Root; General Reduction of a Quaternion in a
Given Plane, to a Standard Binomial Form.
;

The Quaternions of the present Chapter shall

Complanar (119) their common plane being assumed to
225.

be supposed to be
coincide with that of

all

the given right versor * (181).

And all the lines, or vectors, such as a, |3, 7,
&c., or o
ai, a 2
&c., to be here employed, shall be conceived to be in that
,

we may

given plane of /; so that

write (by 123), for the purposes of this

Chapter, HLQ formula of complanarity

1 1ll?

Under

226.

of the

symbol
line

by

may

also be

we may

ft 7

(j3

a)

O.HK&C.

^Ills

we can always (by

103, 117) interpret

7, as denoting a line S in the given plane

denoted (125) by the symbol (7 a)

two apparently equivalent symbols,
:

/3,

(]3

any

which

but not* (comp. 103)

so
7) a, (7 )3) a
:

write,

and may say that

>

IIK

Ill

!?"

these conditions,

form

either of the

that

Ill

this line

or to the three lines

is
a,

the Fourth Proportional to the three lines

the two Means, )3 and 7, of any such
7, |3
;

* In fact the
symbols j8 . 7, 7 j8, or #7, 7/8, have not as yet received with us any interpretation
and even when they shall come to be interpreted as representing certain quaternions, it will be found
.

(coinp. 168) that the

two combinations, - 7 and

have generally different

significations.

COMPLANAR PROPORTION OF VECTORS.

ARTS. 225-227.]

251

Complanar Proportion of Four Vectors, admitting thus of being interchanged,

Under

as in algebra.

so that

we may

as in algebra) the two Extremes, a

(still

of four lines a,

we may

the same conditions

J3, y, o\

make

also

may

write also (by 125),

and

likewise change places

S,

of

any such proportion

among themselves while

if we at the same time
:

the means become the extremes,

change the extremes to means. More generally, if a, /3, j, o\ E ... be any

odd number of vectors in the given plane, we can always find another vector p
in that plane, which shall satisfy the equation,

in ......

!!=
op

or

nr ......

STT--I;
o p
p

and when such a formula holds good, for any one arrangement of the nume
it can
and of the denominator-lines p, |3, 8
rator-lines a, y,
easily be
,

proved to hold

and
For example, whatever four
may be denoted by j3y&, we have the transformations,

good

also for

other

any

arrangement

of the numerators,

any other arrangement of the denominators.

(complanar) vectors

the two numerators being thus interchanged.

Again,

LTL-Z-L-X bvIV
by
S

TV

Sj3~|3S~/3

so that the

227.

means

two denominators

An

also

may change

places.

interesting case of such proportion (22(5) is that in which the

coincide

so that only three distinct lines, such as a,

and that we have (comp. Art. 149, and

fig.

42

[p.

]3,

y, are involved

133]) an equation of the

form,
I.

but not* y =

]3j3

--T=f-j3,

nor a =

a,

|3j3

y.

aj3y

form a Continued Proportion

and

]3 is

the

Mean

or

-ft

In this case, it is said that the three lines

which a and y are now the Extremes,

of

this line |3 being also said to

be

Mean Proportional

Compare the Note to the foregoing Article.

is in the
t We say, a mean proportional hecause we shall shortly see that the opposite line,
same sense another mean although a rule will presently be given, for distinguishing hetween them,
and for selecting one as that which may be called, by eminence, the mean proportional.
;

2K2

ELEMENTS OF QUATEBNIONS.

252

[II.n.l.

while y is the Third Proportional to

between the two others, a and 7
the two lines a and )3 and a is, at the same time, the third proportional to
;

y and

]3.

Under

the same conditions,

TT
11.
so that this mean,
#s0/*, as first,

)3,

and

/a

between a and

to those

to>

we have

a
77

y, is also the

otffor lines.

mi\/
P

7 a
= TT

whence

it is

fo

=~

\7/

(-=->

fourth proportional (226)

have also (comp. again 149),

"We

P
~

natural to write,

IV.

^
a

and therefore by

ay

V.7.

(103),

=
=
But because we
although we are not here to write ]3
(70)*, nor ]3
(ay)*.
have always, as in algebra (comp. 199, (3.) ), the equation or identity, (- q) z - q z
,

we

are equally well entitled to write,

7/7
the symbol

<p

\7,

denoting thus, in general, either of two opposite quaternions,

whereof however one, namely that one of which the angle is acute, has been
already selected in 199, (1.), as that which shall be called by us the Square
Hoot of the quaternion q, and denoted by +/q. We may therefore establish
the formula,

7 form, as above, a continued proportion the upper signs being taken

when (as in fig. 42) the angle AOC, between the extreme lines a, 7, is bisected

if a,
]3,

by the

line OB, or

the opposite

again 199,

line,

(3.)

]3, itself \

or

j3,
:

but the lower

when

T7

:Tj3

when

that angle

is

bisected

by

bisects the vertically opposite angle (comp.

j3

but the proportion of

VIII.

signs,

tensors,

= Tj3:Ta,

and the resulting formulae

IX.

2
.

T/3

in each case holding good.

Ta

T7

And when we

Tj3

(Ta

T 7 ),

shall speak simply of the

Mean

CONTINUED PROPORTION, MEAN PROPORTIONAL.

AET.227.]

make any

Proportional betiveen two vectors, a and y, which

obtuse angle with each other,

we

253

acute, or right, or

always henceforth understand the

of
these
two
bisectors
the
bisector OB of that angle AOC itself,
former
namely,
and not that of the opposite angle thus taking upper signs, in the recent
shall

formula VII.
(1.)

At

the limit

when the angle AOC

U/3 = each of these two unit-lines

common

as each of the

direction

vanishes, so that

and the mean proportional

two given extremes.

= Ua, then
has the same

Uy
]3

This comes to our

agreeing to write,

X.
if

//I = +

a be any positive scalar.

(2.) At the other limit,

and generally,

1,

when AOC =

Uy

or

?r,

</(a?)

=+

a,

= - Ua, the length of the

mean proportional ]3 is still determined by IX., as the geometric mean (in the
usual sense) between the lengths of the two given extremes (comp. the two
figures 41 [p. 132]

plane in which

all

even with the supposed restriction (225) on the

) ; but,
the lines are situated, an ambiguity arises in this case, from

the doubt which of the two opposite perpendiculars at

taken as the direction of the mean

the second line OB,

is

in this

is

To remove this ambiguity, we

vector.

suppose that the rotation round the axis of

sidered in this Chapter are,

the line AOC,

o, to

which

(to

axis all the lines

to

shall

con

by 225, perpendicular), from the first line OA

case positive
which supposition is equivalent
;

be

to

to

writing, for present purposes,

XI.*

v/ -

= +

and

XI

2
.

>/(-

ia,

if

>

0.

And

thus the mean proportional between two vectors (in the given plane)
becomes, in all cases, determined at least if their order (as first and third) be
:

given.
(3.)

If the restriction (225) on the

we might

then,

common plane

on the recent plan (227),

of the lines,

fix definitely

the length, of the mean OB, in every case but one

were removed,

the direction, as well as

this excepted case

being that

two given extremes, OA, oc, have exactly opposite

that the angle (AOC = TT) between them has no one definite

in which, as in (2.), the

directions; so
bisector.

In

this

case,

position, but only a

the sought point B would have no one determined

namely the circumference of a circle, with o for

locus

* It is to be
carefully observed that this square root of negative unity is not, in any sense,
imaginary, nor even ambiguous, in its geometrical interpretation, but denotes a real and given right
versor (181).

ELEMENTS OF QUATERNIONS.

254

[II. n.

1.

and with a radius equal to the geometric mean between OA, oc, while
plane would be perpendicular to the given right line AOC.
(Comp. again

centre,
its

and the remarks in 148, 149, 153, 154, on the square of a

or versor, and on the partially indeterminate character of the

the figures 41
right radial,

square root of a negative scalar,

when

interpreted, on the plan of this Calculus,

as a real in geometry.)

The quotient

228.

seen (191,

(6.)

any two complanar and right quaternions has been

of

to be a scalar

since then

we here suppose

(225) that q

\\\

i,

we

are at liberty to write,

I.

S? = x

and consequently may

given plane (of

i)

V?

= y

V# =

yi

iy

establish the following Reduction of a Quaternion in the

Standard Binomial Form* (comp. 221)

to a

x and y being some two

scalars,

which

be called the two constituents (comp.

then an equation between two quaternions?

may

again 221) of this binomial. And

considered as binomials of this form, such as the equation,
III.

breaks up (by 202,

constituents,

(5.)

q,

into

III

or

x +

iy

=x +

iy,

two scalar equations between their respective

namely,

IY.

tf

x,

y =

y,

notwithstanding the geometrical reality of the right versor,

(1.)

On comparing

i.

the recent equations II., III., IV., with those

marked

VI, in 221, we see that, in thus passing from general to complanar

quaternions, we have merely suppressed the coefficients ofj and k, as being for

as III., V.,

our present purpose, null

and

(2.)

As

the

convenient to

and

and have then written x and

y,

instead of

x.

word

call the

"

binomial

form

II. a

"

has other meanings in algebra, it may be

and the two constituent scalars x
;

COUPLE

which the values serve to distinguish one such couple from another,
not
may
unnaturally be said to be the Co-ordinates of that Couple, for a reason
which it may be useful to state.
y, of

* It
shorter,

is

permitted,

and perhaps

by 227, XI., to write this expression as x -f y V any ambiguity of interpretation.

less liable to

but the form x + iy

is

STANDARD BINOMIAL FOBM, COUPLE.

AETS. 227, 228.]

Conceive, then, that the plane of

(3.)

and that

of
*",

positive rotation

Ax

round

50

255

192] coincides with that

in that figure, directed towards

fig.
i is,

[p.

may be reconciled with our general convention (127), by

imagining that this axis of i is directed from o towards the back of the figure
This being assumed, and perpendiculars BB BB"
or beloiv* it, if horizontal.
the left-hand] which

being

let fall (as in

the figure) on the indefinite line OA

which normal we

to that line at o,

may

OA

call

and

may

suppose

a length equal to that of OA, with a left-handed rotation AOA

V.
while

OA =

= OB

j3

and

V.

or briefly,

OA,

OB",

j3"

a =

and on a normal

itself,

it

have

to

so that

ia,

as in 201, and q =

|3

a,

202

as in

then, on whichever side of the indefinite right line OA the point B

may

be

comparison of the quaternion q with the binomial form II. will

situated, a

give the two equations,

VI.

= Bq) =

so that these two scalars,

of the point

x and

referred to the

B,

j3

V?

0"

ia)

j3"

are precisely the two rectangular co-ordinates

/,

two lines OA

and OA

And

kind.

axes^ of the ordinary (or Cartesian)

= % +

two rectangular unit-

as

since evert/ other quaternion,

same way

in the given plane, can be reduced to the form j

a, or oc is to OA,
in the
a point in that plane, which can be projected into c and
or
constituents,
(comp. 197, 205), we see that the two new scalars,

x and

are simply (for the

where c

iy
is

c"

same pair of

referred to the

same reason) the

co-ordinates of the

new point

c,

axes.

It is evident (from the principles of the foregoing Chapter), that if

(4.)

we
express as couples (2.) any two complanar quaternions, q and q
shall have the following general transformations for their sum, difference, and
we thus

product

VII.

VIII.
(5.)

$q =

x,

..tfq=(afx)
.

(y

x - y y) +

(x

Again, for any one such couple,

= iy, as above, but also,
and

g,

y]

(x

y+ y

x}.

we have (comp. 222) not only

V<?

IX.

Kg =

x - iy

X.
rr

N? =

x*

-VTTT
XIII.

Compare the second Note

y*

XL

-=

to page 111.

Tq

X ~

^ &X
.

ELEMENTS OE QUATERNIONS.

256

Hence, for the

(6.)

quotient of

x +

any two such

-f

iy

x"

iy"

law of the norms (191,

The

+ y2

= xx + y y,

x"

(7.)

iy

x"

expressed here (comp. 222,

(3.)

in

which xyx y

(x"~

may

+ y

(x*

be any four

ir.

1,2.

have,

iy"

= y x - x y.

N^ = Nr/

N#,

is

the well-known algebraic equation, or

identity,

XV.

or the formula,

(8.) ),

by

y"

we

couples,

x"

XIY.

[II.

if)

x - y

(x

yf +

(x

y + y xf

scalars.

SECTION

2.

of Four or more Vectors; Whole

Powers and Roots of Quaternions ; and Roots of Unity.

On Continued Proportion
The conception

229.

of continued proportion (227)

of three to that of

a theory

be formed of cubes and higher

may

may

easily be extended

four or more (complanar) vectors

from the case

ichole

and thus

powers of quaternions, with a

correspondingly extended theory of roots of quaternions, including roots of

scalars,

and

we suppose that
expressed by the formulae,
Thus

in particular of unity.

afiyS form a continued proportion,

o
- = y = p
75

if

cs

I.

II.

p=
- = -y
-3
a

TT

u
whence

(by an obvious extension of usual algebraic notation,)

quaternion 8 a is the cube, or the third power, of /3 a
:

quaternion
relation

is,

may

we may say
;

that the

and that the

;

latter

which

last

or

(-,
\a/

Iir.

|3

f-V

(comp. 227, IV., V.).

\dj

z =
Q, in which
important to observe that as the equation q
a sought and Q is a given quaternion, was found to be satisfied by two

230.

But

it is

opposite quaternions
less

naturally be denoted by writing,

if

/ /D
/

conversely, a cube-root (or third root) of the former

III. ..

q is

the four vectors

q,

of the

simple equation q*
Q be actual and real

Q
;

form

</

is satisfied

by

(comp. 227, VII.), so the slightly

three distinct

whereof each, divided by

gives for quotient a real quaternion, which

is

and

real quaternions,

either of the other two,

equal to one of the cube-roots of

CUBE-ROOTS OF A QUATERNION, AND OF UNITY.

ARTS. 228-232.]
positive unity.

and

In

we conceive (comp. the annexed

fact, if

fig.

257

54) that

)3

two other but equally long vectors in the given plane, obtained
by two successive and positive rotations,

are
)3"

from

]3

each through the third part of a circumference, so that

IV>

"

jsr

IV.

or

W W
7/

and therefore

r_f

we

shall

*
have

(^
PAP

we are equally entitled, at this stage, to write, instead of III. or

IIF.,
these other equations :
so that

or

231.

(real

and

actual) quaternion

Q may

thus be said to have three

which however only one can have an

while none can have an
angle equal to sixty
degress, unless the proposed quaternion Q degenerates into a negative scalar.
distinct cube-roots

(real, actual,

angle

and)
than sixty degrees

less

In every other

of

case, one

of the three cube-roots of Q, or one of the three values

symbol Q4, may be considered as simpler than either of the other two,
because it has a smaller angle
and if we, for the present,
(comp. 199, (1.) )
of the

we shall call the Principal Cube-Root of the quaternion

Q, we shall thus be enabled to establish the formula of

denote this one, which

Q,

by the symbol

inequality,

VTTT
V 111.
232.
one of

its

At

the limit,

cube-roots

is

L/ sy/

/}
(^

^^
<

f\
(^

.
<

TT.

degenerates, as above, into a negative scalar,

itself a negative scalar, and has therefore its angle = TT ;
its

two roots of which the angles are

equal
S

L,

when Q

while each of the two other roots has

HAMILTON

:f
it

ELEMENTS OF QUATERNIONS.

angle = ^. In this case, among these

to each other, and are less than that
2

ELEMENTS OP QUATEKNTONS.

258
of the third,

we

IX.

and therefore

shall consider as simpler,

which answers (comp. 227, (2.)

and so shall be led to write,

..y-l

to a positive rotation

X.

and

[II. n.

2.

as principal, the one

through sixty degrees

Ll/ -

using thus the positive sign for the radical A/3, by which i
- 1 with the connected formula,
expression IX. for 2-J/

multiplied in the

is

although

it

might at first have seemed more natural

and to write thus,

to adopt as principal the

scalar value,

which

latter is in fact one value of the

(1.)

We have, however,
XL yi = 1
.

(2.)

The

symbol, (- l)i

on the present plan, as in arithmetic,

XI

and

y (a

if

>

0.

equations,

XII.

...-i

and

can be verified in calculation, by actual cubing, exactly as in algebra the only

difference being, as regards the conception of the subject, that although i
;

satisfies

the equation

=-

it is

regarded here as altogether real

namely,

as a real right versor* (181).

is no difficulty in conceiving how the same general principles
be extended (comp. 229) to a continued proportion of n + 1 complanar

233. There

may

vectors,
I.

a, di,

a2

a whole number greater than three

nexion therewith, the equations,

when n

is

II...

-)

n
;

III...

-=
a

\A /

a,
;

nor in interpreting, in con

/

\_i_

-p

IV...

\a /

oi-(\

a.

This conception differs fundamentally from one which had occurred to several able writers,
- 1 were inter
before the invention of the quaternions and according to which the symbols 1 and V
in
a
a
as
and
right
lines,
given plane. In
mutually rectangular
representing pair of equally long
preted
Quaternions, no line is represented by the number, ONE, except as regards its length ; the reason being,
;

mainly, that we require, in the present Calculus, to be able to deal with

no one right line is common to all such.

all possible

planes

and that

ARTS. 232, 233.]

FRACTIONAL POWERS, ROOTS OF UNITY.

Denoting, for the moment, what we shall

Q by

ternion

the symbol l/Q,

we

if

LQ<K

this last condition,

y- 1)

th

root of a

qua

have, on this plan (comp. 231, VIII.),

-,

VI.

call the principal n

259

VII.

^;

namely that there

V (/-

1)

>

0;

shall be a positive (scalar) coefficient

y of

x + iy (228), for the quaternion JJ/- 1, thus

i, in the binomial (or couple) form
th
root of negative
serving to complete the determination of that principal n
unity
a

>

or of

0, in

any

each of the two

quaternion,

be changed to - a, if
th
root of a
as to the general n

other negative scalar, since - 1

we may

write,

And

last formulae.

on the same
Viii.

principles,

Q* =

may

VQ

th

representing the general n root of positive unity, has n

different values, depending on the division of the circumference of a circle into
n equal parts, in the way lately illustrated, for the case n = 3, by figure 54;
and only differing from ordinary algebra by the reality here attributed to i.

where the factor

In

n th roots of unity is with us a real versor namely the

quotient of two radii of a circle, which make with each other an angle, equal to
the n th part of some whole number of circumferences.
fact, each of these

(1 .)

We

propose, however, to interpret the particular symbol

th
denoting the principal value of the n root of

whence

it

will follow that

when

H
,

as always

thus writing,

this root is expressed

under the form of a

two constituents x and y shall both be positive, and the

x
shall
have a smaller value than for any other couple x + iy (with
quotient y
constituents thus positive), of which the n th power equals *.

couple (228), the

:

(2.)

For example, although the equation

f
is satisfied

(3.)

by the two

And

values,

(1

(x

i)

+
:

</%,

*,

we

shall write definitely,

although the equation,

(x

+ iyY =

ELEMENTS OF QUATERNIONS.

260
is satisfied

by the

and

three distinct

real couples,

[II. n.

^/3)

(i

2,

and -

i,

we

2.

shall

adopt only the one value,

(4.)

In general, we

shall thus

VTT
XII.

have the expression,

7T

= cos

2n

sm

7T
;

2n

which we shall occasionally abridge to the following

XTT

xV__L X

and

it

w
- m
(Jlo
cj

C-

thus interpreted, denotes a rersor, which turns any line on

th
operates, through an angle equal to the n part of a right angle, in

this root,

which

the positive direction of rotation, round the given axis of i.

234. If
and n be any two positive ^uhole number*, and q

any quaternion,
the definition contained in the formula 233, II., of the whole power, q n enables
,

us to write, as in algebra, the two equations

I.

and we propose

m
.

q q

m+n

II. .. (q

n
)

m = mn
q

former to the case of null and negative whole

to extend the

exponents, writing therefore,

III.

<f

and

.. q

m~ n = m
q

1
.

q"

shall also extend the

formula

= - =

II.,

reciprocal* (134) of

rn

n
.

q.

by writing,
1

VI.

(q

m = n
q ,

be positive or negative so that this last symbol, if m and n be

whole numbers, whereof n may be supposed to be positive, has as many

whether
still

1
:

in particular,

V.

We

IV.

distinct values as there are units in the

when

reduced

to its least

terms

denominator of

among which

consider as the principal one, that which

is

values of q n

the

m th

Compare the

first

Note

to

we

shall naturally

th
power of the principal n

root (233) of q.
*

fractional exponent,

its

page 123.

FRACTIONAL POWERS, ROOTS OF UNITY.

AETS. 233, 234.]

(1.)

For example, the symbol

cube-root of q

it

q*

261

denotes, on this plan, the square of any

has therefore three distinct values, namely, the three values

same quaternion q but among these we

of the cube-root of the square of the

regard as principal, the square of the principal cube-root (231) of that proposed
quaternion.
(2.)

Again, the symbol

on the same plan, as denoting

q* is interpreted,

the square of any fourth root of q

but because (!*)*= 1* =

1, this

only two distinct values, namely those of the square root

and
exponent
being thus reduced to its least terms
;

g%,

square has

the fractional

among

the square of the principal fourth root, which square

these the
at the

principal value

is

same time, the

principal square root (199, (1.), or 227) of the quaternion q.

(3.)

of q

The symbol

q-*

is,

denotes, as in algebra, the reciprocal of a square-root

while

(4.)

q~* denotes the reciprocal of the square, &c.

If the exponent t, in a symbol of the form q*, be

still

a scalar, but a

surd (or incommensurable), we may consider this surd exponent, t, as a limit,

towards which a variable fraction tends and the symbol itself may then be
:

interpreted as the corresponding limit of & fractional power of a quaternion,

which has however

be of

little

this surd

(in this case) indefinitely

many

values,

and can therefore

or no me, until a selection shall have been made, of one value of

power as principal, according to a law which will be best understood

by the introduction

of the conception of the amplitude of a quaternion, to

which in the next section we shall proceed.

(5.)

Meanwhile (comp. 233,

symbol

i*

through

(4.)),

right angles,

Ax

according as this scalar exponent,

positive or negative

t,

and thus may

TTT r
-

VII.

i,

in the positive or negative direction,

whether rational or
establish the formula,

= cos

tir

+
<

or briefly (comp. 233,

XII

.),

VIII.

,t

definitely interpret the

line in the given plane,

which turns any

as denoting a versor,

round

we may already

ois~

tir

sin -^

irrational, is itself

ELEMENTS OF QUATERNIONS.

262

SECTION

On

[II.n.3.

3.

the Amplitudes of Quaternions in a given Plane

and on
for
such
and
Trigonometric Expressions
for
Quaternions,
;

their Powers.
235.

plane of

Using the binomial or couple form (228)

i
(225), if we introduce two new and real

for a quaternion in the

scalars, r

and

2,

whereof

the former shall be supposed to be positive, and which are connected with the
two former scalars x and y by the equations,

!...# =

we

shall then evidently

cos

y = r sin

s,

which

last expression

.
TJ<?

may
.

0,

iy)

(5.)

-r;

U (x + iy) = cos z +

sin z

be conveniently abridged (comp. 233,

234, VIII.) to the following

IV.

>

have the formulae (comp. 228,

II...Tq = T(x +
III.

z,

.,

and

XII

U<7

cis 2

so that

V.

=r

cis z.

And

the arcual or angular quantity, 2, may be called the Amplitude* of the

quaternion q\ this name being here preferred by us to "Angle" because we

have already appropriated the

latter

name, and the corresponding symbol L

q,

to denote (130) an angle of the Euclidean kind, or at least one not exceeding,

and

whereas the amplitude, z, considered as

obliged only to satisfy the equations L, may have any real and scalar value.

in either direction, the limits

We shall denote this

*

Compare the Note

TT

amplitude, at least for the present, by the symbol, f

am

q,

to Art. 130.

f The symbol V was spoken of, in 202, as completing the system of notations peculiar to the present
Calculus; and in fact, besides the three letters, i,j, k, of which the laws are expressed by the funda
mental formula (A) of Art. 183, and which were originally (namely in the year 1843, and in the two
following years) the only peculiar symbols of quaternions (see Note to page 160), that Calculus does not
habitually employ, with peculiar significations, any more than the jive characteristics of operation, K,

U, V, for conjugate, scalar, tensor, versor, and vector (or right part] : although perhaps the mark
for norm, which in the present work has been adopted from the Theory of Numbers, will gradually
come more into use than it has yet done, in connexion with quaternions also. As to the marks, /.,
Ax., I, R, and now am. (or am,,), for angle, axis, index, reciprocal, and amplitude, they are to be
S, T,

considered as chiefly available for the present exposition of the system, and as not often wanted, nor
employed, in the subsequent practice thereof; and the same remark applies to the recent abridgment
for cos -f i sin
to some notations in the present section for powers and roots, serving to express
the conception of one n th root, &c., as distinguished from another; and to the characteristic P, of
what we shall call in the next section the ponential of a quaternion, though not requiring that
cis,

notation afterwards.

No

apology need be made for employing the purely geometrical signs,

and complanarity although the

duced by the present writer, who has found it frequently useful.

for perpendicularity, parallelism,

last of

them was perhaps

_L,
first

||,

|||,

intro

AMPLITUDE OF A QUATERNION.

ART. 235.]
or simply,

am

and thus

q;

and

between amplitude

263

have the following formula, of connexion

shall

angle,

VI.

=)

(z

am

= 2mr

Lq_\

Ax

the upper or the lower sign "being taken, according as Ax .q =

being any whole number, positive or negative or null.

and n

i;

We may then

also (for

any quaternion q
VII.

(1.)

=
TJ#

Writing q =

ivhole

am q

cis

VIII.

the amplitude
its

quantity, expressing (with

from the line a to the line

diminished by any

i)

|||

a,

j3

the general transformations following

.

am

q,

= Tq
or

am

cis

Q3

write

am q.
:

a), is

thus a scalar

Ax

proper sign] the amount of rotation, round

i,

and admitting, in general, of being increased or

number of circumferences, or of entire revolutions, when
j3

only the directions of the two lines, a and /3, in the given plane of /, are given.
i itself, shall be con
(2.) But the particular quaternion, or right versor,
sidered as having definitely, for its amplitude, one right angle so that we shall
;

establish the particular formula,

IX.

am

.i

/.i=-^

When, for any other given quaternion q, the generally arbitrary

n
in VI. receives any one determined value, the corresponding value of
integer
the amplitude may be denoted by either of the two following temporary
(3.)

symbols,* which

we here

so that (with the

same rule

treat as equivalent to each other,

arn n .#,

formula than VI., the equation

X.
and may say that

or

may

we may

write, as a

more

definite

am w

.q = L n q =

this last quantity is the

the zero-value, amo^,

L n q\

of signs as before)

2mr

Lq;

n th value of the amplitude of q

be called the principal amplitude

while

(or the principal

value of the amplitude).

(4.)

may

With

these notations,

and with the convention, am

(- 1)

write,

XI.

XII.

XIII.

am q = Loq = Lq;
am n a = am n 1 = L n 1 = 2mr,

if

>

and
if

a be

still

am n

(- a)

= am

(- 1)

U (- 1) = (2w +

!)TT,

a positive scalar.
* Compare

the recent Note, respecting the notations employed.

= +

TT,

we

ELEMENTS OF QUATERNIONS.

264

From

236.

[II. n.

3.

the foregoing definition of amplitude, and from the formerly

established connexion of multiplication of versors with composition of rotations

(207),

obvious that (within the given plane, and with abstraction

it is

and

tensors] multiplication

braical) addition

so that, if the

interpreted in the general (or indefinite) sense of the equation 235, VI.,

may

write
I.

we

am (q

am q

q)

am q

II.

am (q

f
:

q)

am q - am q

that, in each formula, one of the values of the first

implying hereby
is

of

answer respectively to (alge

symbol am q be

division of quaternions

and subtraction of amplitudes

made

member

among the values of the second member; but not here specifying which
With the same generality of signification, it follows evidently that,

value.

any number of (complanar) quaternions, and

of any one quaternion, we have the analogous f ormulse
for a product of

a whole power

for

III.

am

n<?

= S am

IV.

am

cp

-p am
.

where the exponent p may be any positive or negative integer, or sero.

formula
(1.) It was proved, in 191, II., that for any two quaternions, the
TJq q U/ ~Uq holds good a result which, by the associative principle of
multiplication (223), is easily extended to any number of quaternion factors
.

(complanar or diplanar), with an analogous result for tensors

may

so that

we

write, generally,

v.

un? = nu?

vi.

..

Tn? =

Confining ourselves to the first of these two equations, and combining

with III., and with 235, VII., we arrive at the important formula:
(2.)

it

VII.

II cis

am q

(=

HTJq =

IJlI^

cis

am Uq)

cis

2 am

whence in particular (comp. IV.),

VIII.

(cis

am

q)

cis

(p

am

exponent p be still any whole number.

In these last f ormulse, the amplitudes am

#),

at least if the
(3.)

<?,

sent any angular quantities,

IX.

z,

&c.

n cis z = cis 2s

we may
X.

including thus, under abridged forms, some

respecting cosines

and

sines of

am

/, &c.,

therefore write
.

(cis

zf =

repre

thus,

ci

known and

sums and multiples of

may

them

arcs.

useful theorems,

POWEBS WITH SCALAR EXPONENTS.

ABT. 236.]

For example,

(4.)

number

the

if

IX

cis z

cis z

cis (z

form

of factors of the

have thus,
z)

265

2
.

(cis s)

be two, we

cis z

cis

2z

whence
cos
sin

(z

(z

z)

z)

cos 2z

= S (cis z
=r

cis z)

V (cis *
2

(cos s)

= cos z cos z -

cis g)

= cos z

(sin z)

sin z sin z

sin s

sin 2z

+ sin

z cos *

= 2 cos & sin 2

with similar results for more factors than two.

Without expressly introducing the conception, or at least the notation

we may derive the recent formula IX. and X., from the conside

(5.)

of amplitude,

ration of the power

That power of i, with a scalar exponent, t,

a symbol satisfying an equation which

(234), as follows.

i*

has been interpreted in 234,

may be written thus

(5.),

as

XI.

$ =

cis z,

if

Jfor

or geometrically as a versor, which turns a line through

may

be any

either the

scalar.

We

same or any

XII.
must hold good,

see then at once,

t*.

or

And

as in algebra.

this interpretation,

XIII.

that

where
if

(**)*

if

be

because the number of the factors

easily seen to be arbitrary in this last formula,

XIV.

right angles,

other scalar, the formula,

tf

from

we may

i*

is

write also,

***,

But the two last formulae may be changed by

IX. and X., which are therefore thus again obtained
although the later forms, namely XIII. and XIV., are perhaps somewhat
simpler
having indeed the appearance of being mere algebraical identities,
be any whole* number.

p
XL,

if

to the equations

although we see that their geometrical interpretations, as given above, are

important.
(6.)

symbol

In connexion with the same interpretation XI. of the same useful

i*,

it

may

be noticed here that

XV. ..K.^

=
r<;

* It will soon be
seen that there is a sense, although one not
quite so definite, in which this
formula holds good, even when the
exponent p is fractional, or surd namely, that the second member
is then one of the values of the first.
;

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATEENIONS.

266

[II.

3.

and that therefore,

XYI.
XVII.

** = i
co8^=S.
z

sin

^= r

Y.

+ r);

(i

= i r

t*

r).

(t*

iG

Hence, by raising the double of each member of XYI. to any

positive whole power p halving, and substituting z for Jfor, we get the
(7.)

equation,

XVIII.

2*- l (cos

)*

=J

(*

COS pz

r^ = i

(***

+ r*) + \p

+ p COS (p - 2)

(*&-*)

(HO*)

COS (p

.0

&c

4) 2

4-

&C.,

with the usual rule for halving the coefficient of cos Os, if p be an even
integer and with analogous processes for obtaining the known expansions of
;

2P

"

(sin z)

known

p
,

for

any

results of the

237. If p be
I.

still

positive

whole value, even or odd, of p

may

p
(r cis z)

be thus written
II.

and many other

same kind.
a whole number,

r p cispz

which (comp. 190, 161) the two

in

we have thus
= (Tq)*

cis

the transformation,

(p.&mq)

and versor kinds,

factors, of the tensor

=
(qY = (Tq)P

Tf

HI.

(?)

= (Uq)P = -UqP

and any value (235) of the amplitude am.g may be taken, since all will
And if, for L, we
conduct to one common value of this whole power qp
.

substitute this slightly different formula (comp. 235, (3.)

IV.

p
(q ) n

= TqP

cis

(p.*m*q), with

p = ~, n

),

>

0,

n being whole numbers whereof the first is supposed to be prime to the

second, so that the exponent p is here & fraction in its least terms, with & positive
9

Tqp

interpreted as a positive scalar (of which

the positive or negative logarithm, in any given system, is equal to p x the
logarithm of Tg), then the expression in the second member admits of n
which are precisely the
distinct values, answering to different values of n

denominator n, while the factor

is

values (comp. 234) of the fractional power qp on principles already established

the principal value of that power corresponding to the value n = 0.
n 9 we may say that the symbol (qp )
(1.) For any value of the integer
y

n>

th
p
by the formula IV., represents the n value of the power q
values, however, recurring periodically, when p is, as above, & fraction.

defined

such

POWERS WITH SCALAR EXPONENTS.

ARTS. 236, 237.]

(2.)

Abridging
V.

lp n

(l

cis

2pmr,

symbol

shall soon

remove

j,

.. liw

(4.)

fraction,

Thus, making successively p =

Principal value of IP

VII. ..!* =
VIII.

p be any
=

cis

l*o

Denoting in

we

(- l) n ,

= +

11

cis rnr,

particular,

p =

l*i

we have

-g-,

= -

t
l2 =

1,

--

&c.

,
3

2^~~

2"^

have, on the same plan

= +

1*2

li

the n th value of (- l) p

manner

like

and in

1.

1,

l*i=

l,

by 235, XII.,

thus, generally,

1P

VI.

TTTTT

if

we

a restriction which however

(3.)

we have

to l p n)

}n

267

by the abridged

(comp. 235, XIII.), for any

fractional* value of p,

IX.

X.

= dsp(2n +

p
(- l) n

(- 1)*

and

VT
XI.

iu =

/
.

(- l)*o

cis

= +

TT

whence (comp. 232),

(- l)*x

i,

-l +

1)

= -

i\i
(- l)*i
f

cis

~=i

iu =

(~ l)*a

1,

=+

(- l)*a

i,

-1

&c.

i,

\/3

&0.,

-g

these three values of (- 1)* recurring periodically.

(5.)

The formula IV.

XII.

so that the n th value of

gives, generally,
.

(f) n

p
q is

cis

(jP)

by V., the transformation,

2pmr =
!"

(q*>)

equal to the principal value of that

power of
and
of the same power of positive unity
<?,

multiplied by the corresponding value

th
it may be remarked, that if the base a be any positive scalar, the principal p
;

power,
(6.)

p
(a )o, is

simply,

The n

th

by our

value of the

definitions, the arithmetical value of of.

th

power of any

negative scalar,

a, is

in like

th

manner equal to the arithmetical p power of the positive opposite, + a, multi

or in
plied by the corresponding value of the same power of negative unity
;

symbols,

XIII.

- a* n = -

l p

p
(fl

)o

= (^

(7.) The formula IV., with its consequences V. VI. IX. XII. XIII., may
be extended so as to include, as a limit, the case when the exponent p being still
scalar, becomes incommensurable, or surd; and although the number of values

of the

power

becomes thus unlimited (comp. 234,

*

As

before, this restriction

is

(4.) ),

yet we can

only a temporary one,

2

still

ELEMENTS OF QUATERNIONS.

268
consider one of

them

[II. n.

as the principal value of this (now) surd power

3, 4.

namely

the value,

XIV.
which answers
nion

p
(q )o

Oy

cis (p

am

to the principal amplitude (235, (3.)

q},

of the proposed quater

q.

We

238.

may

therefore consider the symbol,

p

q
in

which the

being

any quaternion, while the exponent, p, is any scalar, as

fully interpreted", but no interpretation has been as yet assigned

now

to this other

base, q, is

symbol

of the

same kind,

?
in

which both the base

and the exponent q

q,

are supposed to be (generally)

quaternions, although for the purposes of this

To

Chapter complanar (225).*

this, in a way which shall be completely consistent with the foregoing con
ventions and conclusions, or rather which shall include and reproduce them, for

do

the case where the

new

quaternion exponent, q

degenerates (131) into a scalar,

one main object of the following section: which however will also
contain a theory of logarithms of quaternions, and of the connexion of both
will be

logarithms

and powers with the properties

shall

the ponential of a quaternion, and to consider which

call

we
we next

function, which

of a certain

proceed.

SECTION

4.

On the Ponential and Logarithm

of a Quaternion ; and on Powers

of Quaternions, with Quaternions for their Exponents.

we

239. If

consider the polynomial function,

I.

in

which q

is

P (q,

m) =

is

any quaternion, and

+ q + q2 +
1

any

qm ,

positive whole

number, while

it is

(for conciseness) that

supposed

TT

then

it is

not

_
Sm ~

/_

1.2.3..m \~ r (m

difficult to

+ I)/

prove that however great, but

[For the general case see 316.]

finite

and

given, the

PONENTIAL OF A QUATEKNION.

ARTS. 237-239.]

be, a finite

Tq may

tensor

..T(P(q,m +

III.

number

can be assigned, for which the inequality

a

if

n)-~P(q,m))<a,

0,

>

shall be satisfied, however large the (positive whole)

however small the (positive) scalar

if

words,

IY.
a

finite

we
.

number n may

provided that this last

is given.

be,

and

In other

write (comp. 228),

q

a,

269

= x +

iy,

value of the

number

Xm + iYm

m) =

(q 9

can always be assigned, such that the follow

ing inequality,

V.

shall hold good,

and

>

series,

(X...

XJ + (Y,IM - Ym

)*

<

<?,

however large the number

0) the scalar a

may

n,

and however small (but given

It follows evidently that each of the two scalar

be.

or succession of scalar functions,

VI.

..X

X^l + x,

l,

+ * +

Xm ,..

^^,..
4i

VII.

F, =

0,

Xn,

X, and the

and

converges ultimately to a fixed

or simply

y,

finite limit,

latter

F>,

function of the two scalars, x and y.

VIII.

we must
series

= y + xy,

or F,

consider this quaternion

Fw

whereof the one

and

of

may

which each

is

be called
a certain

Writing then

Q = X* + F, =

X+ iY,

(namely the

which the following

limit to

of quaternions,

IX.

(q, 0)

1,

P (q,

1)

=14

q,

converges ultimately) as being in like

(q,

2)

manner a

1+

|,

(q,m],

certain function,

which we

shall call the ponential function, or simply the Ponential of q, in consequence of

its

possessing certain exponential properties

and which may be denoted by

any one of the three symbols,

P (q, oo ),

We have therefore the

or

P(<?),

or simply

P^.

equation,

X ... Ponential ofq=Q = T*q = l +

qi

+ qz +

with the signification II. of the term q m

(1.) In connexion with the convergence of this ponential

+ q

series,

or with the

ELEMENTS OF QUATERNIONS.

270

inequality III., it may be remarked that

rm = Tq m we shall have, by 212, (2.),

we

if

[II. n.

write (comp. 235) r = Tq,

4.

and

XL

m + n)-P(<7,m))<P(r, m+w)-P(r,

T(P(?,

tive terms, rm+ i,

we

made

rm+n can be

have rp ^ \rp rp+z

^rp+l9 &c., so that a
can be assigned, such that r m
a and then,
shall

<

of the n posi

Now if we take a number^?

a.

<

m)

sum

it is sufficient then to prove that this last difference, or the

number

finite

<

>

>

2r -

>

>p

1,

2r-

<

XII.

ro

P(r,

the asserted inequality

In general,

(2.)

XIII.

if

AO +

+ n) -

P (r,

m)

<a(2~

2~

is therefore proved to exist.

an ascending series, with positive

A^ + A

2
2

+ &c.,

<?

AO

where

be convergent when q is changed to & positive

when q is a quaternion.

>

0, AI

2~ n )

<a\

0, &c.,

>

a fortiori converge,

scalar, it will

240. Let q and q be any two complanar quaternions, and let

sum, so that

I...

=/+?,

2"

q"\\\g

II.

.
q"

(o*

m=

d}

H~

/o

JL

<?",,

and with correspond

= q m q + g m _ii + q

m-tfr

= q\ = 1.
Hence, writing again r = Tq f rm = Tq m
=
manner / T/, &c., the two differences,
<?

IY.

and

be their

we have

where

III.

q"

\\\g;

then, as in algebra, with the signification 239, II. of qmy

ing significations of q m and

such as

coefficients,

P (/,
P (r

m)

+ r

(r,

2m)

m)

P (/,

(r

m)

tf

ogW,

and in

like

+ /, m),
.

P (r,

m),

can be expanded as sums of positive terms of the form r p .rp (one sum con
taining \m (m + 1), and the other containing m (m + 1) such terms)* but, by
;

239, III., the

sum

of these

given small positive scalar

Y.
*

[For the

On expansion
- %(m + 1)
III.

it is

total

number

(r

two positive differences can be made

#,

of terms in

(r

(r

+ r m)
,

+ /, m)
is 1

<

+2 +3+

(q

m}

P (q,

m}

-f

is

- P (q+g m)
t

<

(r

m)

(r,

if

a,

seen to be ^r p ,.rp where p -f p

4- 2) terms, all of which are positive : similarly for IV.
(
seen at once that

T (P

than any

since

+ /, 2m) -

of III. the series

less

m}

>

>

0,

(m + 1) = % (m + 1) (m + 2).
2
m, and there are (m + I)
From the expanded form of

-P

(r

).]

EXPONENTIAL PROPERTY.

ARTS. 239-241.]

provided that the number

when
~Pq is

the tensors,

r,

separately tends to 0, as
,

is

taken large enough

(1.)

VII.

where

oo

the

We

Law,

..P/ + ^=P.P=P.P

write (comp. 237,

PI =

is

scalar.

we

If

each difference, therefore,

a tendency which must exist a fortiori,

are replaced by the quaternions, q, q
The function

tends to

therefore subject to the Exponential

VI.

known

(5.)

VIII.

then

271

if

),

P# =

(*)

= arithmetical value off;

base of the natural system of logarithms, and x is any

simply e* to denote this principal (or

shall henceforth write

th
arithmetical) value of the x

of

power

and

e,

so shall have the simplified

equation,

viir.
(2.)

PZ =

*.

Already we have thus a motive for writing,

IX.
but this formula

which we

generally,

*;

..?<?

here to be considered merely as a definition of the sense in

is

interpret this exponential symbol,

&

as

namely

what we have

sum

called the ponential function, Pq, considered as the

lately

of the infinite but

converging series, 239, X. It will however be soon seen to be included in a

more general definition (comp. 238) of the symbol q q
.

(3.)

For any

X.
241.

we have by VIII.

scalar x,
.

the transformation

x - 1P# = natural logarithm of ponential of x.

The exponential law

(240) gives the following general decomposition

of a ponential into factors,

I.

in

P? = P

(x

iy)

which we have just seen thai the factor P#is a positive scalar. The other
P/, is easily proved to be a versor, and therefore to be the versor of Pq,

factor,

while

P#

is

the tensor of the

II.

since

and

IV.

same ponential

..P?.P(-?) = PO
.

(Kq)

m =

K (qm

=
)

1,

(say)

and

Kq

we have

because

III.

PK?

in general,

= KP?,

(comp. 199, IX.)

therefore, in particular (comp. 150, 158),

V.

Piy =

(- iy)

= KPiy,

or

VI.

NP*y =

1.

ELEMENTS OF QUATERNIONS.

272

We may

[II.ir.4.

therefore write (comp. 240, IX., X.),

VII.

TP? = PS? = P* =

IX.

UP? = PV? = P*y

this last transformation

X.
XI.

VIII.

**;

= e* =

cis

y (comp. 235, IV.)

being obtained from the two

.

SPwy = 1 i~

if

80- 1TP0;

+ &c. =

VPey = y - 77& o5 +

series,

cos y\

&c. = sin y.

Hence the ponential P? may be thus transformed

XII.
(1.)

nor to

we had not chosen

If

se/0c

(at first)

XIII.

(x

+ ///)=

to assume as

any one

c* cis y.

known the

series for cosine

..?/>=///

+ fyy,

/(-

y}

and

sine,

known one on which

unit of angle, such as that

we might then have proceeded

their validity depends,

we should

P? = P

as follows.

= +/y,

(- y)

= -

Writing

^,

have, by the exponential law (240),

XIV.

.f(y + jO = S(P*y .P*/) =///.//

XV.

/(y -

and then the functional

XVI.

2/ )

*y */;
.

/y .// + 0y

equation,

which

= cos -

x a right angle

results,

0/;

namely,

would show that

XVII.

/T/

whatever unit of angle

constant c

be adopted, provided that we determine the

may

by the condition,

XVIII.

= 1-5708,

least positive root

of the equation fy (= SPty) =

or nearly,

XVIII
* In

fact,

the value of the constant

polation between the

as the study of the series*

c

may

be obtained to

this

would show.

degree of accuracy, by simple inter

two approximate values of the function/,

/(1-5)

and of course there are

value can be found

artifices,

=+

0-070737,

/(1

6)

=-

0-029200

not necessary to be mentioned here,

by which

a far

more accurate

CONNEXION WITH TRIGONOMETRY.

ART. 241.]

(2.)

(TT

still

would thus

motive

rical constant, c

nume

arise for representing a right angle by this

or for so selecting the angular unit, as to have the equation

denoting two right angles),

XIX.

= 2c =

TT

giving nearly,

for thus

least positive root

XIX

we should reduce XVII.

As

(3.)

to the simpler form,

.

fy =

cos y.

to the function $y, since

XXI.
evident that

to be taken.

(fyY +

fyy

In

XXII.

iy)

l,

= +

$c

P<>/

y and it is easy to prove that the upper sign is

can be shown (without supposing any previous

sin
it

fact,

P (-

(wY

of cosines or sines) that

knowledge

- 1

of the equation fy

= 3-14159, as usual

TT

XX.

it is

273

is
</>c

or

1,

and therefore that

positive,

XXIII.

Pic =

whence

XXIY.

and

we

If then

XXVI.

replace c

$y =

The

(4.)

by

cos

series

^,

--

X. XI.

r Pty = SPt

(y

-c}=f(y- c),

= sin y

XXVII.

and

Pit/

cis y, as in

IX.

and sine might thus be deduced instead

and since we have the limiting value,

for cosine

known

XXVIII.

we have

of being assumed as

it

w=S

lim. y~ l sin
y=0

y = lim. y~ r
l

VP*>

1,

y=o

follows that the unit of angle, which thus gives Piy = cis y,

angle subtended at the centre by the arc equal

(or 2c) is to 1, as the circumference

is

to

radius

is (as

or that the

usual) the

number

ir

to the diameter of a circle.

If any other angular unit had been, for

any reason, chosen, then a right
would
of
course be represented by a different number, and not by 1*5708
angle
(5.)

nearly

but we should

still

XXIX.
though not the same
HAMILTON

have the transformation,

.

Py = cis

fy
(

\c

x a right

series as before, for cos

ELEMENTS OF QUATERNIONS.

angle

),

y and sin

y.
2

ELEMENTS OF QUATERNIONS.

274

The usual

242.

I.

if

2imr =

n be any whole number

P~V>

and

1,

II.

see,

(q

+ 2imr) = Pq,

follows, then, that the inverse ponential function,

it

what we may call the Imponential, of a given quaternion q, has

many values, which may all be represented by the formula,
III.

of

which each

satisfies

4.

by 241, XII., that

or

finitely

and

we

unit being retained,

[II. n.

P n-

inde

the equation,

IV.

while the one which corresponds to n = 0, may be called the Principal ImpoIt will be found that when the exponent p is any scalar, the definition
nential.
th
already given (237, IV., XII.) for the n value of the
to establish the formula,

V.

and we now propose

general case (238),

(q")

when the

exponent

VI.

the principal value of q q being

is

(<?

f
:

definition, to

thus writing generally,

(jT.-

j)

conceived to correspond to n =

still

the more

q the General Exponential Formula,

principal amplitude of q (comp. 235, (3.)

(1.)

by a new

a quaternion q

and
=

of q enables us

to extend this last formula,

for any two complanar quaternions, q

th

p power

0, or to

the

).

For example,
VII.

the ponential
therefore

now

P<?,

?
(

)o

= PfaPo-

e)

P!?,

because

which we agreed, in 240,

(2.),

Po^e =

Is

to denote simply

by

eq,

is

seen to be in fact, by our general definition, the principal value

of that power, or exponential.

(2.)

With
VIII.

the same notations,

.

e^ = cisy,

cos y

= |

(c*y

e-*),

sin

y =

- (&

tvi

?*)

these two last only differing from the usual imaginary expressions for cosine
and sine, by the geometrical reality* of the versor i.
*

Compare 232,

(2.),

and the Notes to pp. 253, 258.

LOGARITHM OF A QUATERNION.

ARTS. 242, 243.]

The

(3.)

defined

by

cosine

and

IX.

of a quaternion (in the given plane)

sine

the equations

may now

be

= i (& +

cos q

275

4
*)

X.

sin q

- (& -

e -*)

Alii

and we

may

write (comp. 241, IX.),

XI.

With

(4.)

cis

& = Py.

this interpretation of cis#, the exponential properties, 236,

X., continue to hold good

XII.

(<?*%

and we

(q lTq)

may

IX.,

write,
f

(iq

am,

q)

= (T0)

cis (q

am,

q)

a formula which evidently includes the corresponding one, 237, IV., for the
n th value of the p th power of q, when p is scalar.

The

(5.)

and VI., combined with 235, XII., give generally,

definitions III.

XIII.

ln

(!

this last equation including the

The same

(6.)

XV.
which

last

P 2imf
.

XIV.

(j*

= I,/, (^

)o

formula 237, XII.

definitions give,

.P -H- =

XVI.

&

equation agrees with a

.^)o = rF;

known

interpretation of the symbol,

V-i

considered as denoting in algebra a real quantity.

case where the exponent
(7.) The formula VI. may even be extended to the
f

is

a quaternion, which

planar with the base q

XYII.

not in the given plane of

thus

(*0.

is

we may

q
,

it

243.
plane),

As

it,

to the plan of this

after

it

as

though we

view

it is

see that there

symbol
would be no

what has been shown respecting complanars.

of a quaternion q (in the given
satisfies the equation,

any quaternion which

I.

aiid in this

Chapter to enter into any

of the interpretation of the exponential

regards the general logarithm q

we may regard

not corn-

write,

for this case of diplanar quaternions,

difficulty in treating

and therefore

would be foreign (225)

further details, on the subject
but

i,

..**

=?? =

?;

simply the Imponential

^,

of

which the n th value

2

is

ELEMENTS OF QUATERNIONS.

276

But the

expressed by the formula 242, III.

answers

(as above) to n

pL n

4j 5.

principal imponential, which

may be said to be the principal logarithm, or

the
Logarithm, of
quaternion ?, and may be denoted by the symbol,

simply

the

so that

we may

0,

i?;

write,

or

still

ITU? =

11

1?

am

1?

(T? U?) = IT? + 1U?,

and therefore,

0,

III.

We

more simply,
II.

because

= Po ? = IT? +
1

I.

1U? =

amo

fl

q.

have thus the two general equations,

IY.
in which IT?
(1.)

is still

SI? = IT?

The

(comp. 235,

this last denoting the

VIII.

The

(2.),

VI? = 1U?

n
.

th

or

?,

value

and

(4.)

may be

denoted by any one of the symbols,

log

or

?,

and then we

(log q) n

),

VII...l(-l)=tir.

ifir;

general logarithm of ?

log

(3.)

the scalar and natural logarithm of the positive scalar T?.

As examples

VI. ..lf =
(2.)

V.

1?

(log ?),
shall have,

+ 2W7T.

formula,

IX.

log

q q

= log ? + log

?,

if

\\\

?,

holds good, in the sense that every value of the first member is one of the
values of the second (comp. 236).
=
and one value of log q q = ? !?.
;
(4.) Principal value ofqi

(5.)

may

The

quotient of two general logarithms,

be said to be the general logarithm of

quaternion base, q

and we

independent integers, while

the quaternion, ?

to

the

complanar

see that its expression involves* two arbitrary

its

principal value

may

be defined to be

* As the
corresponding expression in algebra, according to Graves and

1?

Ohm.

1?.

and

EQUATIONS OF ALGEBRAIC FORM.

ARTS. 243-245.]

SECTION

On

277

5.

Finite* (or Polynomial) Equations of Algebraic Form, involving

Complanar Quaternions and on the Existence of n Real
;

Quaternion Roots, of any such Equation of the

We have

244.

is

Degree.

seen (233) that an equation of the form,

I.

where n

n ih

any given

f - Q = 0,

and

positive integer,

quaternion (144), has always n

anyf given, real, and actual

and unequal quaternion roots, q,
is

real, actual,

n
complanar with Q namely, the n distinct and real values of the symbol Q
This result is, however,
(223, VIII. ), determined on a plan lately laid down.
;

much more general Theorem,

Form namely, that if q q

included in a
Algebraic

z,

l9

respecting Quaternion Equations of

q n oe any n given, real,

and com

planar quaternions, then the equation,

II.

n
.

n~ l

q,q

has always n real quaternion roots, q

which roots

q.q

n
,

q",

q(

+ qn =

0,

\ and no more in the given plane

possible however that some, or all may become equal, in

of
certain
relations existing between the n given coefficients.
consequence
245. As another statement of the same Theorem, if we write,
of

it is

I.

n~ l

F n q ^ qn +

q,q

+ qn

we may say that evert/ such polynomial

a
n
equal
product of
real, complanar, and linear (or binomial)
of
the
form
or
that an equation of the form,
factors,
q
q
the coefficients

function, r n

qi

qn being as before,
to

is

<7,

II.

can be proved, in

rf - (f
q) (q

all cases to exist

q")

although we

(q

may

present methods, to assign expressions for the roots, q

coefficients q l ,
*

By

qn

S W),

not be able, with our

n

q(

\ in terms of the

saying finite equations, we merely intend to exclude here equations with infinitely many
P# = 1, which has been seen (242) to have infinitely many roots, represented by the

terms, such as

expression q = 2inir, where n may be any whole number.

t It is true that we have supposed Q ||| i (225) but nothing hinders us, in any other case, from
substituting for i the versor UVQ, and then proceeding as before.
;

ELEMENTS OF QUATERNIONS.

278

Or we may say

246.
nions

cf,

&c.,

HI i,

that there

which

[II. n.

always a certain system of n real quater

is

the system of equations, of

satisfies

5.

known

algebraic

form,
/

+.. +
+

in...

-&;

(0YY"

247.

Or because the

<?()=-<Zl

ff
\W +W
=
+
(/+<?

difference

T? n

- Fn q

+ ..- +

to\

&o.

is divisible

under the supposed conditions of complanarity (224),

by q

it is

that at least one real quaternion q always exists (whether

as in algebra,

sufficient to

we can

say
it

or

with the foregoing form (245, I.) of the polynomial function F.*
248. Or finally, because the theorem is evidently true for the case n =

J,

assign

not), which satisfies the equation,

IV.

.
?<?

0,

while the case 244, I., has been considered, and the case q n =
is satisfied by
the supposition q = 0, we may, without essential loss of generality, reduce
the enunciation to the following
:

Every

equation of the
I.

which q
plane, whereof
in

q",

formrf

..q(q- /)

(q

.
q")

(q

00-0) = Q,

and Q are any n real and given quaternions in the given

Q and q may be supposed actual (144), is satisfied by

at least

and complanar quaternion, q [see 253 (!.)]

the m - 1 last of the n - 1 given quaternions
Supposing
n ~^
- m first of them are actual, where m
vanish, but that the n
q(
may be
whole number, from 1 to n - 1, and
a new real, known,

at least one real, actual,

249.
.

<?

any

that

complanar, and actual quaternion q

II.

qo

which

introducing
the condition,

satisfies

m =
q

.
q".

n~ l
-r = F M _itf = q
+ q l tf-2 + q z n 3 + . . + tf n-i, which is of the form 245, I. If
q
is a second root of F,,^.]
then every equation of this form has a root, F,,.^" = 0, and
t The corresponding form, of the algebraical equation of the n th degree, was proposed hy Mourey,
in his very ingenious and original little work, entitled La vraie theorie des Qunntites Negatives, et des
Quantites pretendues Imaginaires (Paris, 1828).
Suggestions also, towards the geometrical proof of. the

[Thus

q-

<?

q"

theorem in the text have been taken from the same work in which, however, the curve here called
the inequality, here numbered as
(in 251) an oval is not perhaps defined with sufficient precision
It is to be observed that Mourey s book contains no hint of the
251, XII., being not employed.
present calculus, being confined, like the Double Algebra of Prof. De Morgan (London, 1849), and like
;

the earlier work of Mr.

Warren (Cambridge,

conception of the Quaternion involves, as

1828), to questions ivithin the plane: whereas the very

seen, a reference to Tridimensional SPACB.

we have

GEOMETRICAL EXISTENCE OF REAL ROOTS.

ARTS. 246-251.]

we may

write thus the recent equation

279

I.,

^-mj
and may (by 187, 159, 235) decompose
IV.

Tfq =

and

V.

TJfq

two following

into the

it

VI.

or

1,

am/? =

2pjr

some whole number (negatives and zero included)

250. To give a more geometrical form to the equation, let A be any given
and p, a, or OA, OB,
let it be supposed that a, ]3,
or assumed line
i, and
is a
and OP, os, are n - m + 2 other lines in the same planes, and that
in which

is

|||

</>p

known

function of p, such that

VII.

? A,

]3

p = ?A,

?"A,

a = ? A,

and
m

~ a

_ f - (P\
YP -jy

VIII

\pj

the theorem to be proved

points, o, A, B,

may

and

s,

may

is

T$p =

251. Whatever value

1||

f^\
\^

os y

OA

OB

then be said to be, that whatever system of real

and whatever positive whole member m,

same

always at least one real point p, in the

plane, which satisfies the two conditions

.

~
|3

in a given plane,

be assumed, or given, there

IX.

X.

am $p =

we may assume

2pir.

for the versor (or unit- vector)

Up, there always exists at least one value of the tensor Tp, which satisfies the
because the function T$p vanishes with Tp, and becomes
condition IX.
when
infinite
Tp = GO, having varied continuously (although perhaps with
;

Attending then only to the least value (if there

be more than one) of Tp, which thus renders
equal to unity, we can
conceive a real, unambiguous, and scalar function T//I, which shall have the
fluctuations) in the interval.

T</>p

two following properties

XI.

.
T<j>

And

in this

way

( i/,i)

= 1

XII.

p = $1,

or

shall call generally

perties

(XI-^L]

1>

<

1, if

>

0,

<

1.

an Oval,

XIV.

Up

i,

Tp =

^i,

and plane closed curve, which we

and which shall have the two following pro

be conceived to determine a

the equation, or system of equations,

XIII.

may

real, finite,

1st, every right line, or ray,

drawn from

the origin o, in

any arbitrary

ELEMENTS OE QUATERNIONS.

280

[II. n.

direction within the plane, meets the curve once, but once only

n-m

of the

other given points A, B,

is

5.

and Ilnd, no one

=
= 0.*
=
$j3

on the oval, because

<a

252. This being laid down, let us conceive a point p to perform one circuit
of the oval, moving in the positive direction relatively to the given interior
point o; so that, whatever the given direction of the line os may be, the

am

amplitude

(p

or),

four right angles, or

if

by

supposed to vary continuously,^ will have increased by

2?r,

in the course of this one positive circuit

and conse

of $p, will have

quently, the amplitude of the left-hand factor (p o-)
at
the
same
if
the
A
be also interior to
2mir.
time, by
Then,
increased,
point
:

the oval, so that the line OA must be prolonged to meet that curve, the ray AP
will have likewise made one positive revolution, and the amplitude of the
But if A be an exterior point, so
factor (p - a) a will have increased by Sir.
:

that the finite line OA intersects the curve in a point M,

and therefore never

again
prolonged, although the prolongation of the opposite line AO
must meet it once in some point N, then while the point p performs first what
meets

if

it

we may

call the positive half-circuit

positive half-circuit
initial

and

from N to

final direction,

the opposite direction-, in

from

to N,

again, the ray

namely

and afterwards the

AP has only

oscillated

other

about

its

that of the line AO, without ever attaining

this case, therefore,

the amplitude

am

(AP

OA), if

supposed to vary continuously, has only fluctuated in its value, and has
(upon the whole) undergone no change at all. And since precisely similar

still

remarks apply to the other given points,

am

(ftp,

increment = 2 (m +
points, A, B,

t) TT,

if

the amplitude,
has (by 236) received a total

B, &c., it follows that

of the product (VIII.) of all these factors,

be the number (perhaps zero) of given internal

m is (by 249) at least = 1. Thus, while P

while the number

* [A curve

traced out by a point moving so that the product of powers of its distances from fixed
equal to a constant parameter, consists of closed curves or ovals surrounding the fixed points
and enclosing all ovals corresponding to smaller parameters. If the parameter is small, each oval
certain
encloses but one fixed point, but as it increases, two ovals will combine into a curve with a
points

is

"

It is not generally true that a ray OP from one of the fixed points meets an
(254 (4.) ).
undulatory oval only once. In this case OP will oscillate in its motion as P traces out the oval. But
= in L POS 4- 2 (v - L PAO) = const., defines a set of curves diverging like half-lines or rays
am

undulation

"

</>/>

from the

fixed points, and

approximating to straight lines at great distances
perties of Conjugate Functions each of these curves which originates from

from them.

By the pro
cuts at right angles each
oval round
and does not meet it again. Near 0, am
is nearly equal to
L POS plus a con
stant.
From this it appears that IX. and X. can always be satisfied, and that as P traces out an
oval round
without oscillation, am
continually increases or diminishes without oscillation. The

</>

<f>p

ovals are lines of magnetic force, and the orthogonal curves are traces of equipoteutial surfaces for a
system of electric currents normal to the plane.]

t That

is,

so as not to receive

ferences (comp. 235, (1.)

).

any sudden increment, or decrement, of one

or

more whole circum

GEOMETRICAL ILLUSTRATIONS, QUADRATICS.

ARTS. 251-253.]

performs

times,

above) one positive

(as

and therefore

sequently the condition

condition, IX.,

oval

there

=
<t>p

or/<7

is

is

circuit, the

am

amplitude

<}>p

at least once, through a value of the

X. has been

at least

satisfied throughout,

has passed at least

form

once satisfied.

by the supposed

281

%pir

and con

But the

construction

other

of the

upon that curve, for which

therefore at least one real position P,

so that, for this position of that point, the equation 249, III.,

and therefore

The theorem

also the equation 248, I., is satisfied.

and consequently also, by 247, the theorem of 244, with

245 and 246, is therefore in this manner proved.

its

of Art. 248,

transformations

253. This conclusion

is so important, that it may be useful to illustrate

the general reasoning, by applying it to the case of a quadratic equation, of
the form,

OP
OS

AP
OA

Fig.55.

We

have now to prove fcomp. 250, VIII.) that a (real) point P exists, which
renders the fourth proportional (226) to the three lines OA, OP, AP equal
to a given line os, or AB, if this latter be drawn = os
;

or which satisfies the following condition of similarity of

triangles (118),

III. ..

which

AAOPOCPAB;

includes the equation of rectangles,

IV.

OP

AP = OA

T^
Fig. 55,

bis.

AB.

(Compare the annexed figures, 55, and 55,

nuous curve* is described as a locus (or as

bis.)

Conceive, then, that a conti

part of the locus) of P, by means of this

equality IV., with the additional condi
tion

when

necessary, that o shall be

within it; in such a manner that when
(as in fig. 56)

the general or total locus in several points, M,

* This curve

Fig. 56.

a right line from o meets

we

reject all

but the

of the fourth degree is the -well-known Cassinian ; but when it breaks up, as in
we here retain, as the oval of the proof, only the one round o, rejecting

into two separate ovals,

for the present that round A.

fig. 56,

HAMILTON

ELEMENTS OF QUATKUNIONS.

ELEMENTS OF QUATERNIONS.

282
point

which

is

nearest to o, as not belonging (comp. 251,

Then while p moves upon that oval,

from M to N, and from N to M again,

here considered.

[II. n.

5.

XII.) to the oval

in the positive direction

so that the ray OP per

forms one positive revolution, and the amplitude of the factor OP os increases
continuously by 2ir, the ray AP performs in like manner one positive revolu

relatively to o,

tion, or (on the

AP OA
:

whole) does not revolve at

2?r or

increases

by
In the one

to the oval.

increases

0, according as the point

by

case, therefore, the

4?r (as in fig. 55, bis)

by

(as in fig. 56)

one real position, p,

.

its initial

upon the
=

am

(j>p

value

oval,

2pir,

but

am

amplitude

in the other case,

by the

have been.

may

0/o
it

of the product

increases

by

2?r

through a value of

Hence, for

at least

we have

and therefore

VII.

throughout,

and

so that in each case, it passes at least once

the form 2pir, whatever

V.

and the amplitude of the factor

A is interior or exterior

all,

VI.

U^jo

= 1

1,

Tft>

by the equation

construction, or

of the

locus

IV.

the

=l

(II.) is therefore satisfied by at least one real vector

= 1 (I.) is satisfied by at least
and
the
quadratic equation fq
p
consequently
=
one real quaternion root, q
p \ (250, VII.). But the recent form I. has the

geometrical condition

<j>p

same generality

as the earlier form,

VIII.

where

qi

F2

<?

= q2 +

and q z are any two given,

q\q

+ qz =

(comp. 245),

and complanar quaternions

which satisfies

real, actual,

thus there is always a real quaternion q in the given plane,

the equation,

Vlir.
subtracting, therefore,

F2q =

(f*

+ qtf + ? 2 =

and dividing by q -

(comp. 247)

as in algebra (comp. 224),

obtain the following depressed or linear equation

IX.

g+

tf

qi

0,

or

IX

=
?"

we

q,

=-q -

<?i

(comp. 246).
ff

The quadratic VIII. has therefore a second real quaternion root, q related in this
manner to the first and because the quadratic function ~F z q (comp. again 245)
;

is

thus decomposable into two linear factors, or can be put under the form,

cannot vanish for any third real quaternion, q so that (comp. 244) the quad
real roots,
ratic
equation has no more than two such

it

RELATION S BETWEEN THE ROOTS.

ART. 253.]

The

(1.)

cubic equation

may

F 3 q = q +qiq + qtf + q3 = q(q- q }(q -

therefore be put under the form (comp. 248),

+ ?3 =

283

q")

has therefore one real root, say q\ by the general proof (252), which has been
above illustrated by the case of the quadratic equation subtracting therefore
= 0, and dividing by q-q\ we can depress the
(comp. 247) the equation F 3

it

</

cubic to a quadratic, which will have two

the cubic function

new

real roots,

and q*

and thus

be put under the form,

may

XI.

F 8 g = (q- q)

(q

(q

q")

q"),

which cannot vanish for any fourth real value of q; the cubic equation X.
and similarly

has therefore no more than three real quaternion roots (comp. 244)

for equations of higher degrees.

(2.)

vectors,

The existence of two real roots q of the quadratic I., or of two real
p and //, which satisfy the equation II., might have been geometrically
from the recently proved increase =

anticipated,

course of one circuit, for the case of

must be two

and p

4?r of

55, Us, in consequence of

fig.

in the

amplitude

<j>p,

which there

on the one oval of that figure, of which each

satisfies the condition of similarity III.
and for the case of fig. 56, from the
consideration that the second (or lighter) oval, which in this case exists, although
real positions, p

not employed above,

figure

is

related to o

is
;

related to

A exactly

as the first (or dark) oval of the

so that, to the real position p

on the

first,

there must

correspond another real position p , upon the second.

(3.) As regards the law of this correspondence, if the equation II. be put
under the form,

and

if

we now
XIII.

write
.

qa,

we may

write

XIV.

#1

=-

1,

q*

=-

<r

a,

comparison with the form VIII. and then the recent relation IX (or
246) between the two roots will take the form of the following relation
for

between

vectors,

= a;
..p + P

XV.

or

XV ..

r
op = /= a

-p = PA;

so that the point p completes (as in the cited figures) the parallelogram

and the

line

PP

this position of p

OPAP

we

226) the equation,

XVI.

by the middle point c of OA. Accordingly, with

have (comp. III.) the similarity, and (comp. II. and

bisected

is

A AOP

oc

P AB

XVII.

$p

(a

p)

$p

1.

202

ELEMENTS OF QUATERNIONS.

284

The

(4.)

other relation

[II. n.

5.

between the two roots of the quadratic VIII.,

namely (comp. 246),

XVIII...

and accordingly, the

OA, OP, and AP, or a,

The

(5.)

line

<r,

or os,

a fourth proportional to the three lines

is

and - p.

p,

actual solution,

complanar quaternions,

XIX...V=-*;
a

gives

<?"=?*>

<?

is

by

calculation, of the quadratic equation

performed exactly

XX...J

ike formula being,

/(ift-ft),

foi

in which, however, the square root

as in algebra

VIII. in

to be interpreted as a ra&J quaternion,

is

on

down.

principles already laid

(6.) Cubic and biquadratic equations, with quaternion coefficients of the

kind considered in 244, are in like manner resolved by the known formula of
algebra but we have now (as has been proved) three real (quaternion) roots
;

for the former,

254.

and/owr such

The following

is

real roots for the latter.

another

mode

of presenting the geometrical reason

ings of the foregoing Article, without expressly introducing the notation or

The equation $p =

conception of amplitude.

of

253 being written

as

follows,
I...<7

xp

-(i- a

or

we may thus regard the

point

p and

II.

Tcr

)>

vector

as a

<r

TX/O,

and

III.

known function of

vary together

although

UX

f>,

and A are fixed,

(and does) happen, that

may

it

TLr =

the vector p, or the

as & function of the point P; in the sense that, while o

s

to a former position without P having similarly returned.

Now

may

return

the essential

property of the oval (253) may be said to be this that it is the locus of the
points P nearest to o, for which the tensor TX/O has a given value, say b
namely
:

the given value of Tcr, or of os,

we

when

the point

s,

like o and A,

is given.

If then

conceive the point P to move, as before, along the oval, and the point s also
by the recent formula I., this latter

to move, according to the law expressed

point must move (by II.) on the circumference of a given circle (comp. again
and the theorem is, that in so
fig. 56), with the given origin o for centre;
moving, s will pass, at least once, through every position on that circle, while P

performs one

circuit of the oval.

And

this

may

be proved by observing that

(by III.) the angular motion of the radius os is equal to the sum of the angular
motions of the two rays, OP and AP but this latter sum amounts to eight right
;

angles for the case of fig. 55,

bis,

and

to four right angles for the case of fig.

56

CASSINIAN OYALS, LEMNISCATA.

ARTS. 253, 254.]

285

the radius os, and the point s, must therefore have revolved twice in the
case, and once in the second case, which proves the theorem in question.
(1.)

In the

first

of these

two

cases,

each of the three angular velocities

namely when A

is positive

first

an interior point,
throughout, and the mean angular
is

velocity of the radius os is double of that of each of the two rays OP, AP.

Bat

in

exterior, the mean angular velocity of the ray AP is

zero ; and we might for a moment doubt, whether the sometimes negative velocity
of that ray might not, for parts of the circuit, exceed the always positive velocity

when A

the second case,

of the ray OP,

and

is

so cause the radius os to

cannot he, however

for if

we

conceive P

move

to describe, like

other (or lighter) oval, in fig. 56, the point s

law
if

I.)

then

would again traverse the whole

it

could ever fluctuate in its motion,

through some given

by P

and

the coefficients, the quadratic equation

a circuit of the

dependent on

same circumference
it

it

by the

as before

would pass more than

twice

circle, during the successive

certain limiting values of
within
thus,
would have more than ttvo real roots a

series of real positions

description of the two ovals

(if still

of the

This

backwards, for a while.

on that

which has been proved to be impossible.*

(2.) While s thus describes a circle round

result

nected point B to describe

an equal

circle

round A

o,
;

we may

conceive the con

and in the

case at least of

easy to prove geometrically, from the constant equality (253, IY.)

AP and OA AB, that these two circles (with T U and T U
as diameters], and the two ovals (with MN and M N as axes), have two common
call the two given
tangents, parallel to the line OA, which connects what we may
fig.

56,

it is

of the rectangles OP

foci (or focal points) , o

and A: the new or

third circle,

thfa focal interval OA. as diameter, passing through the

the ovals, as the figure

may

which

is

described on

four points of contact on

serve to exhibit.

the same things by quaternions, we shall find it convenient

(3.)
to change the origin (18), for the sake of symmetry, to the central point c;

To prove

and thus

now CP by

to denote

and CA by a, writing also CA = Ta = a, and

each of the two equal circles by b. We shall
p,

representing still the radius of

then have, as the joint equation of the system of the two ovals, the following

V.

But because we have

VI.

.
<f

.T(<f-l)=2c,

if

and

= --

generally (by 199, 204, &o.) the transformations,

= 2S? 2 *

Tf = T?

+ 2 V? 2 = 2NS? -

[See the Note to 251, page 280.]

N? = N? - 2NV?,

ELEMENTS OF

286

the square of the equation Y.

the two following forms

may

QTJATERTSIOKS.

(by 210,

(8.)

[II. n.

5.

be written under either of

VII.

(N?

I)

4NV? = 4c

VIII.

(Nq +

I)

4NS<?

= 4c2

shows that the maximum value of T~Vq is c, at least if 2c 1,

and that this maximum corresponds to the
as happens for this case of fig. 56
= a results which, when interpreted, reproduce those of
value Tq = 1, or
whereof the

first

<

Tp

the preceding sub-article.

When

(4.)

2c

>

1,

it is

=
permitted to suppose Sg

and then we have only one continuous

if c

<

1,

though

0,

J, there exists a certain undulation in

>

=
NV<?

oval, as in the case of

N<?

fig.

55,

but becoming

a,

when

2
S<?

= 2c+

minimum for Sq =
= 1, and
a maximum when

(as before)

namely

1,

at the

T<?

two summits M,

N,

bis

but

the form of the curve

(not represented in that figure), TV*? being a

p JL

= 2c -

0,

or for

vanishing

where the oval meets the

axis.
(5.)

(as is

In the intermediate

known)

case,

written (cornp. 200, (8.)) under

2

IX...T(<?

when

2c = 1, the Cassinian curve IV. becomes

a lemniscata; of which the quaternion equation

-1) = 1;

or

any one
2

X...N<z

may, by V., be

of the following forms

XL

or

2S.<f ;

2
T? = 2SU

z
.

or finally,

XII.
which

last,

when

.T

written as
.

agrees evidently with

(6.)

known

C? = 2CA 2

COS 2ACP,

results.

This corresponds to the case when

XIII...

<r

~,

and

XIV...

o
j

=
j0

=+
^,

in253,XIL,

that quadratic equation having thus its roots equal; and in general, for all
degrees, cases of equal roots answer to some interesting peculiarities of form of

the ovals, on which

(7.) It

restriction

we cannot here

delay.

may, however, be remarked, in passing, that if we remove the

that the vector p, or CP, shall be in a given plane (225), drawn

through the line which connects the two foci, o and

A, the recent equation

V.

will then represent the surface (or surfaces)

generated by the revolution of the
oval (or ovals), or lemniscata, about that line OA as an axis.

CASSINTAN OYALS, LEMNISCATA.

AETS. 254, 255.]

we look back,
we shall see that

255. If

253, III.,
253,

but

IV.,

a moment,

for
it

on the formula of

similarity,

involves not merely an equality of rectangles,

an equality of

also

287

AOP

angles,

and

PAB

so

that

the

angle QAB represents (in the figures 55) a given difference of the base angles
AOP, PAO of the triangle GAP but to construct a triangle, by means of such
:

a given difference, combined with a given base, and a given rectangle of sides,
is a known problem of
elementary geometry. To solve it briefly, as an

AA

exercise, by quaternions, let the given base be the line

with o for
let

BAA

PAA

its

middle point, as in the annexed figure 57

represent the given difference of base angles,

AA P

and

let

AP A

angle of sides,

P.

OA AB be equal to the given rect

We shall then have the similarity
.

and equation,

I...AOA POCPAB;
whence

it

follows

II.

by the simplest

Fig. 57.

calculations, that

-a

or that p

mean

is

proportional (227) between a and

/3.

Draw,

therefore,

a line OP, which shall be in length a geometric mean between the two given
its extremity will be the
lines, OA, OB, and shall also bisect their angle AOB
;

AA P

required vertex, P, of the sought triangle

a result of the quaternion

analysis, which geometrical synthesis* easily confirms.

(1.) The equation III. is however satisfied also (comp. 227) by the
vector, OP = PO, or p =
p and because |3 = (p a) p, we have
:

IY.
so that

..

p + a

or

IT.

PB
.

OP

OB

OP

OA

OP

OA

ihefour following triangles are similar (the two

being equal)

V.

opposite

first of

them indeed

A OP

oc

AOP

POB OCAP B

geometry again would confirm.

(2.) The angles AP B, BPA, are therefore supplementary, their sum being
equal to the sum of the angles in the triangle OAP whence it follows that

as

* In
fact, the two triangles I. are similar, as required, because their angles at o and p are equal,
and the sides about them are proportional.

ELEMENTS OF QUATERNIONS.

288

the four points A, p, B, p

APBP
circle

AP

B,

are concircular

[II. n.

5, 6.

or in other words, the quadrilateral

:*

which (we may add) the centre c is on the

because
the angle AOB is double of the angle
57),

inscriptive in a circle, of

is

OAB

(see

again

fig.

by what has been already proved.

equations in quaternions

Quadratic

(3.)

may

also be

employed in the

decompose a
given vector into two others, which shall have a given geometrical mean, &c.
solution of

other geometrical problems

many

SECTION

On the

for example, to

6.

n2 - n

Imaginary (or Symbolical) Roots of a Quaternion

Equation of the n Degree, with coefficients of the kind
th

considered in the foregoing Section.

The polynomial

256.

on which

it

I.

Xn and Yn

may

depends,

and thus we may

or

Fn q

function

(245), like the quaternions

q l9

qn

always be reduced to the form of a couple (228)

establish the transformation (comp. 239),

Fn q = Fn (x + iy) = Xn + iYn

q,

Gn

and

n,

being two known,

On

(a?,

y)

+ iHn (x,

real, finite,

and

y),

scalar functions

and y which functions, relatively to them, are

each of the n th dimension, but which involve also, though only in the first
xn yn
And since the one
dimension, the 2n given and real scalars, # yi,

of the two sought scalars, x

Fn q

quaternion (or couple) equation,

is

0,

equivalent (by 228, IY.) to the

system of the two scalar equations,

II.

we

see (by

0,

Yn = 0,

or

III.

Gn (x, y}

0,

n (x, y}

0,

what has been stated in 244, and proved in 252) that such a
n th dimension, can always be satisfied by n

system, of two equations of the

and by

systems (or pairs] of real scalars,

IV. ../,y
*

sM.yM;

*",y";..

Geometrically, the construction gives at once the similarity,

A AOP cc
and

not more than n, such as,

if \ve

FOB,

whence

complete the parallelogram APA P

A OA P a

OP

B,

the

gives

L BPA

new

= OPA + PAO = POA

similarity,

L AP B = OA

-f

A PO

= AOP

thus the opposite angles BPA, AP B are supplementary, and the quadrilateral APBP is inscriptible. It
will be shown, in a shortly subsequent section [261, (6.)], that these four points, A, P, B, p form a
harmonic group upon their common circle.
,

IMAGINARY QUATERNION ROOTS.

ABTS.255, 256.]
it

although

may happen

become equal

(or

(1.) If x

that

tioo

289

or more of these systems shall coincide with

each other.

to)

and y be treated

(comp. 228, (3.)), the two

as co-ordinates

equations II. or III. represent a system of two curves, in the given plane

and

then the theorem is, that these two curves intersect each other (generally*} in n
real points, and in no more
although two or more of these n points may
:

happen

with each other.

to coincide

Let h denote, as a temporary abridgment, the old or ordinary imaginary,

1, of algebra, considered as an uninterpreted symbol, and as not equal to
Y/
real
any
versor, such as i (comp. 181, and 214, (3.) ), but as following the rules
(2.)

ofscalars, especially as regards the commutative property of multiplication (126)

so that

V.
(3.)

the

h9 + 1 =

and

0,

Let q denote

still

VI.

hi

\_q~]

imaginary algebraic quantity, or bi-scalar (214,

(7.)

just

VIII.

and

let

any

inform

= x+

iy,

IX.

but

biquaternion (214, (8.)), or (as

[/] +

i
[#"],

h not =

a real quaternion, or real couple, x + iy

now proposed

of h, let

meaning

VII.

but

ih,

i.

and with

denote the connected but

),

x + hy

\_q~\=

we may

x + hy

here call

be said to be complanar with

so that

-,

it)

BI-COUPLE, of

with the old notation

(123) of complanarity.
(4.)

nq =

Then, for the polynomial equation in real and complanar quaternions,

(244, 245), we may be led to substitute the following connected alge

braical equation, of the

X.

same degree,

[Fn q-] =

[q]

-f

n,

and

[grj

involving real scalars similarly

|>]-i

+ [0J =

2
which, after the reductions depending on the substitution V. of - 1 for h ,
receives the form,

where

Xn and Tn are the same real and scalar functions as in

(5.)

But we have seen

vanish together,

by

is

cause points of intersection, which are generally imaginary, to become

for instance the hyperbola, z z - y* = ,
real and distinct points, by the pair of right lines xy = 0, if the scalar a
or
;
2
2
and xy = 0, may be considered to
0, the two pairs of lines, # - / =
:

<

>

but for the case a

have/owr

coincident intersections at the

origin.

HAMILTON

to

may

but coincident with each other, and with former real roots

intersected in tivo

made

selecting any one of n real pairs IV. of scalar values, x and y\

* Cases of
equal roots
real,

I.

in II., that these two real functions can be

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

290

[II. n.

6.

then th Degree, has therefore n Real or

Form x + y */ - 1 and it has no more than n such

the General Algebraical Equation X., of

Imaginary Roots* of the

roots.

Elimination of y, between the two equations II. or III., conducts

(6.)

z
which equation has
generally to an algebraic equation in #, of the degree n
therefore n z algebraic roots (5.), real or imaginary namely, by what has been
;

lately proved, n real

y
y
same number
(W]

and

(comp. IV.) of

scalar roots

to correspond

[><)],

and which are

correspond,

0(1*1)

On

257.

would then have an

+ ^(n + i) con t r ary

?

the whole, then,

quaternions, of the n

th

to

it

by

n*

|>(-)] ;

(7.) ),

of

real root,

l)st

results that the equation

.
0",

or at least

the

form,

(252).

Fn q =
it

in complanar

admits of only

qto (244, &c.),

- n
bi-quaternions (214,

(8.) ),

by n(n

1) imaginary quaternion

or bi-couples (256,

(3.) ),

which

be thus denoted,
II.

and

(n

what has been proved

symbolically satisfied also (comp. 214, (3.)

may

.
&<")],

roots,

<?

or

degree, with real coefficients, while

I.

roots,

n real quaternion

is

XIII.

[>()] ;

by the supposed elimination, to imaginary or bi-scalar values of

and y^ n+l \ for example, could both be real, the quaternion

Fn q = 0,

equation
0(11+1)

\ with real and scalar values

and n(n - 1) other roots, with the
which may be thus denoted,
x^

either themselves imaginary (or bi-scalar, 214,

since if #(n+1

of corresponding values of y,

XII.

of,

of

III.

which the

first,

foC"*

)],

for example, has

[q^

the/om,

where / n+1 ),

n* l

x,(

n+1
yp* \ and y / /
l

are four realscalars, but h

is

the imaginary

of algebra (256, (2.)).

* This celebrated

Theorem of Algebra has long been known, and has been proved in other ways
seemed necessary, or at least useful, for the purpose of the present work, to prove it anew, in
or rather to establish the theorem (244, 252), to which in the present
connexion with Quaternions

but

it

Calculus

it

corresponds.

Compare the Note

to

page 278.

NEW SYMBOLICAL

ART. 256, 257.]

(1.)

BOOTS OF UNITY.

There must, for instance, be n(n -

the given plane of

(comp. 256,

1)

291

th
imaginary n roots of unity,

in

i\

besides the n real roots already deter

(3.) ),

and accordingly in the case n = 2, we have the four

and two imaginary
i, two real
following square-roots of I

mined

(233, 237)

|||

IV.
for,

by 256,

(2.),

- 1

1,

- hi

hi,

we have
V.

And

hiY =

= (-

=+

1) (- 1)

the two imaginary roots of the quadratic equation

1.

F q = 0,
z

which generally

be obtained by multiplying the square), may

root in the formula 253, XX. by hi
so that in the particular case, when that
radical vanishes, the four roots of the equation become real and equal zero
exist, at least as symbols (214, (3.)

having thus only

(2.)

Again,

itself for a square-root.

we

if

^T
VI.

write (comp. 237,

1*!

-1 +

plane

if

we

so that

whereof the

1*2

real cube-roots of positive unity, in the given

9 =

[ ?]

"

3
.

(as usual) the two ordinary (or algebraical)

then the nine cube-roots ofl

VIII.

write also,

and 6 Z are

roots of unity

q*

"

VII.

v/ 3

i
-

so that 1, q, (f are the three

and

(3.) ),

q, q*

9,

a raz/ scalar

first is

are the following

(\\\i)

2
;

Qq, Oq*

imaginary cube-

?,

0y

the two next are real couples, or quaternions

the two following are imaginary scalar s, or biscalars; and the four that
|||t
remain are imaginary couples, or bi-couples, or biquaternions.
;

(3.)

The

sixteen fourth roots of unity

IX.

+ A

(|||

are

*)

(1

A) (1

i)

the three ambiguous

signs in the last expression being all independent of
each other.

2P

ELEMENTS OF QUATERNIONS.

292

roots, of this sort, are

Imaginary

(4.)

sometimes

[II. n.

6, 7.

useful, or rather necessary,

in calculations respecting ideal intersections* and ideal contacts, in geometry :

although in what remains of the present Volume, we shall have little or no

occasion to

employ them.

We may,

(5.)

however, here observe, that when the

restriction (225)

on

the plane of the quaternion q is removed, the General Quaternion Equation of

th
the n
Degree admits, by the foregoing principles, no fewer than n* Roots,
real or imaginary ; because, when that general equation is reduced, by 221, to

the Standard Quadrinomial Form,

X.

Fn q =

+ iXn

+jYn

-f

kZn =

0,

breaks up (comp. 221, VI.) into a System of Four Scalar Equations, each
th
(generally) of the n dimension, in w, x,y,z; namely,

it

XI.
and

0,

Yn = 0,

0,

which therefore has

and similarly
nion

be eliminated between these four, the result

if x, y, z

scalar (or algebraical) equation of the degree n*,

constituent,

Zn =

is

(generally) a

relatively to the

n* (algebraical) values, real or

for the three other constituents, x, y,

z,

remaining
imaginary

of the sought quater

q.

It

(6.)

may

even happen, when no plane

is

given, that the

(or solutions) of a finite^ equation in quaternions shall

become

number of roots
infinite

as has

been seen to be the case for the equation q = - 1 (149, 154), even when we
confine ourselves to what we have considered as real roots. If imaginary roots
z

we may

be admitted,
h,

write, still

bi-scalar values,

the expression,

XII. ..(-!)* = * + hv
v

more generally, besides the two

and

if

being thus any

tivo

real

S*>

W=

Svv =

0,

and right quaternions,

Nt>

- N*/ = 1

in rectangular planes,

provided that the norm of the first exceeds that of the second by unity.
(7.)

And

in like

manner, besides the two real and

scalar values,

1,

we

have this general symbolical expression for a square root of positive unity,
with merely the difference of the norms reversed

XIII.

1*

+ hv
*

f
,

Sv = St/ =

Sw

0,

Comp. Art. 214, and the Notes there referred

t Compare the Note to page 277.

to.

EECIPROCAL OF A VECTOR,

AETS. 257-259.]

SECTION

293

7.

the Reciprocal of a Vector, and on Harmonic means of

with Remarks on the Anharmoiiic Quaternion
Vectors
of a Group of Four Points, and on Conditions of Con-

On

circularity.
258.

When
I.

two

and a

vectors, a

a = -

Ta Ta =

Ua

and therefore

Ta,

are so related that

II.

a = -

Uu

Ta,

or that

III.

we

and

1,

writing

V.

.a = Ba,

VII.

IX.

Ea

= Ba

VI.

and

every right quotient v,

-Ua

VIII.

Ta;

Ua

0,

the Reciprocal* of the other

is

Or

so that for every vector a,

denote this relation between them,

(at least for the present)

shall

Ua

two vectors

shall say that each of these

and

IV.

by

B a = EEa = a
2

and

One

Bit; =

IKt>

(comp. 161,

(3.),

and 204,

XXXV

.).

most important properties of such reciprocals is contained

in the following theorem
If any two vectors OA, OB, have OA OB for their reciprocals, then (comp.
fig. 58) the right line A B is parallel to the tangent OD, at the
259.

of the

origin o, to the circle

OAB

and the two

Or

are inversely similar (118).

I.

if

OA =

OB =

and

OA,

triangles,

OAB, OB A

in symbols,

OB,

then

OAB

OB A

Fig. 58.
(1.)

OA B

is

(2.)

Of

course,

under the same conditions, the tangent

at o to the circle

parallel to the line AB.

The angles BAO and OB A

(226) to AB, AO,

and

or

BOD being equal, the fourth proportional

and OB, has the direction of OD, or the

OB, or to BA, OA,

direction opposite to that of

AB

and

its

length is easily

proved

to

be the

*
Accordingly, under these conditions, we shall afterwards denote this reciprocal of a vector o hy
the symbol or 1 but we postpone the use of this notation, until we shall he prepared to connect it with
a general theory of products and powers of vectors. Compare 234, V., and the first Note to page 123.
And as regards the temporary use of the characteristic R, compare the second Note to page 262.
;

ELEMENTS OF QUATERNIONS.

294

same

reciprocal (or inverse) of the length of the

AB

line

[II. n.

7.

because the similar

triangles give,
II.
it

(OA

BA)

OB = (OB

AB

OB = 1

AB

being remembered that

III.

we may

OA OA = OB
.

OB = 1

therefore write,

IY.

(OA

BA)

whatever two vectors a and

Changing a and

(3.)

OB =

may

]3
j3

AB

R
The

-^-g

|3

B (Bj3 - Bo),

be.

to their reciprocals, the last formula becomes,

"

YL..E(/3-a)= Eq
(4.)

V.

or

inverse similarity

I.

VII.

or

B/3;

(OA :B

A ).OB

=B. AB.

gives also, generally, the relation,

K-^.
a

(5.)

Since, then,

IX.

K-^j3
a

by
l

195, II, or 207,

ri3
=T
Ki-

E]3

(2.),

v
X.

we have

Ea

EaEj3
.

E]3

-r
E(j3a)

-^775-

the lower signs agreeing with YI.

(6.)

In general, the

reciprocals of opposite vectors are themselves opposite

or in symbols,
(7.)

More

XL

x be any
(8.)

= - Ea.

Btfa = x~ l

Ea,

scalar.

we

find for any three vectors a,

j3,

to 7, dividing,

-B/3

Ba-Bj3

and taking con

7 (complanar or diplanar) the

-QA

BY
if

E (- a)

Taking lower signs in X., changing a

jugates,

formula

generally,

XII.
if

E7
Ba

VB(j3-y)

j3-a

BC

ABCO

a = OA,
(9.)

= OB, and
= oc, as usual.
]3
7
If then we extend, to any four points of space the notation
>

\TTTT
XIY.
interpreting each of these

AB CD
/
(ABCD) =
BC DA

two factor-quotients as a quaternion, and

their product (in /M S order)

is

(25.),

defining that

the anliarmonic quaternion function, or simply #ta

ANHARMONIC AND EVOLUTIONARY QUATERNIONS.

AETS. 259-260.]

295

Anharmonic, of the Group offour points A, B, c, D, or of the (plane or gauche)

Quadrilateral ABCD, we shall have the following general and useful formula of
transformation

where OA
(10.)

OB

oc are supposed to be reciprocals of OA, OB, oc.

this notation XIV., we have generally, and not merely for

With

collinear groups (35.), the relations

XVI.
(11.)

OA,

OD

shall

B C

(OCBA),

DA

(ABCD)
,

(ADCB)

1.

OD the reciprocals of

(OADC)

therefore,

XIX.
if

D be any five points, and OA

then have, by XV.,

XVIII.
and

XVII.

o, A, B, c,

we

= 1
(ABCD) + (ACBD)

Let

we agree

K (A B C D

to write generally, for

XX.
(12.)

the formula,*

six points,

any

/
N
(ABCDEF) =

If then the five points o

= (OADC) (OCBA) = - (OADCBA),

AB CD EF
BC DE FA
.

D be complanar

(225),

we

have,

by 226,

and by XIV.,

XXI.

K (A B C D

= (ABCD),

XXI

or

(A B C D

the anharmonic quaternion (ABCD) being thus changed to

four rays OA,

OD are changed

its

=
)

K (ABCD)

conjugate,

when

the

to their reciprocals.

260. Another very important consequence from the definition (258) of

reciprocals of vectors, or from the recent theorem (259), may be expressed as
follows

If any
(see

again

three coinitial rectors, OA, OB, oc, be chords of one

fig.

58)

their three coinitial reciprocals,

OA

OB

common
X

oc

circle,

then

are termino-

* There

is a convenience in calling, generally, this product of three quotients, (ABCDEF), the

evolutionary quaternion, or simply the Evolutionary, of the Group of Six Points, A . . F, or (if they be
not collinear) of the plane or gauche Hexagon ABCDEF because the equation,
:

(ABCA B C

1,

expresses either 1st, that the three pairs of points, AA BB , cc , form a collinear involution (26.) of a
well-known kind; or Ilnd, that those three pairs, or the three corresponding diagonals of the
,

hexagon, compose a complanar or a homoepheric Involution, of a new kind suggested by quaternions

(comp. 261, (11.)

).

ELEMENTS OF QUATERNIONS.

296
cottinear (24)

other words,

or, in

then the three points A

And

if

the four points

then

line,

7.

be concircular,

o, A, B, c

OA

OB

oc

thus terminate on

their three coinitial reciprocals, OA, OB, oc, are chords of

the tangent to which circle, at the origin,

c are situated on one right line.

conversely, if three coinitial vectors,

one right
one circle

[II.

parallel to the right

is

line; while the anharmonic function (259, (9.)), of the inscribed quadrilateral
OABC, reduces itself to a scalar quotient of segments of that line (which there

fore

own

is its

I.
if

by 139)

conjugate,

= B
(OABC)

namely,

B A =

GOA B C

OABC),

(o

the symbol GO be used here to denote the point at infinity on the right line
C
and if, in thus employing the notation (35) for the anharmonic of a

AB

we

plane pencil,

consider the null chord, oo, as having the direction* of the

tangent, OD.
(1.)

If p

= OP be the

variable vector of a point

quaternion equation of that circle

II.

Ep

may

Ej3 + x (Eo
E/3),

p upon the

be thus written

where

circle

OAB, the

III.

x = (OABP)

the coefficient x being thus a variable scalar (comp. 99, I.), which depends on
the variable position of the point P on the circumference.
(2.)

Or we may

write,

form of the equation of the same circle OAB with which may use
fully be contrasted the earlier form (comp. 25.), of the equation of the line AB,
as another

v
Or, dividing the second
jugates, we have for the circle,
(3.)

^TT

VI.

tp
.

=
+ Up

+ u

ta

member

...

while

u(3

~Tnr

of IY.

by the

VII.

first,

ta

uQ
+ -^ =

and taking con

u,

for the right line.

(4.)

Or we may

VIII
this latter
*

Y*

write,

E/

by

II.,

"

Ea -

or

Vlir

E/3

symbol, by 204, (18.), denoting any

Ka -

- Y-

B/3

scalar.

Compare the remarks in the Note to page 140, respecting the possible determinateness
symhol UO, when the zero denotes a line, which vanishes according to a law.

signification of the

of

AKHARMONIC OF A CIRCULAR GROUP.

AET. 260.]

Or

(5.)

more

still

IX.

briefly,

V (OABP)

while FOi,

their plane,

point in

by 259, XXI., we have the

X.

the four new points o^

(OABP)

and

still concircular,

PCi are the reciprocals

K (OABC)

= (OABC) = Y^

if

0.

p be any

of PO,

fifth

PC, then

1
;

are therefore generally concircular.

Ci

however, the point P be again placed on the

If,

(7.)

relation,

(O^jfixd)

IX

or

If the four points o, A, B, c be

(6.)

297

OABC, those four

circle

new

points are (by the present Article) collinear ; being the intersections of
In this case, therefore,
the pencil p OABC with a parallel to the tangent at P.
.

we have

the equation,

XI.

(P

OABC) = (oiA^Ci) = (OABC)

so that the constant anharmonic of the pencil (35)

thus seen to be equal to

is

what we have defined (259, (9.) ) to be the anharmonic of the group.

the anharmonic of a circular group is a scalar,
(8.) And because
(by 187,

(8.)

own

to its

therefore write, for

XII.

any

= +

that

is,

equal

we may

BC)

(AB

co),

the upper or the lower sign being

or does not fall, betiveen the points A

sides

taken, according as the point B falls,

c

= + (OA

(OABC)

= + a quotient of rectangles of opposite

and

inscribed quadrilateral OABC, the formula,

(OABC)

it is

negatively taken

tensor, either positively or

according as the quadrilateral OABC

is

an uncrossed or a crossed

one.
(9.)

Hence

it is

easy to infer that for any circular group

o, A, B, c,

we

have the equation,

XIII...TJ^

AB

TJ52;
CB

the upper sign being taken when the succession OABC is a direct one, that is,
when the quadrilateral OABC is uncrossed} and the lower sign, in the contrary
case, namely, when the succession is (what may be called) indirect, or when

while conversely this equation XIII. is sufficient

occurs, that the anharmonic (OABC) is a negative or a
and
therefore by (5.) that the group is circular (if not linear),
positive scalar,
as above.
the quadrilateral
to prove,

is

whenever

crossed

(10.) If A, B, c, D,

surface of one sphere),

HAMILTON

it

E be any five homospheric points (or points upon the

OF/
if o be any sixth point of space, while OA

and

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

298

are the reciprocals of OA,

[II. n.

OE, then the five new points A

7.

E are generally

but if o happens to be on the sphere ABODE,

homospheric (with each other)
E are complanar, their common plane being parallel to the tangent
then A
the
to
plane
given sphere at o with resulting anharmonic relations, on which
;

we cannot here

delay.

An interesting case

261.

of the foregoing theory

is

that

when

the generally

scalar anharmonic of a circular group becomes equal to negative unity

case (comp. 26), the group

circular
left to

is

in which

A few remarks upon such

said to be harmonic.

may here be briefly made the student being

up hints for himself, as what must be now to him an easy exercise

and harmonic groups

fill

of calculation.
(1.)

For such a group (comp. again fig. 58), we have thus the equation,
= - 1
A B = B C
I.
and therefore II.
(OABC)
.

or

III.

and under

= J (H

Ej3

R7

we shall say (comp. 216, (5.)

Mean between the two vectors, a and y.

this condition,

the Harmonic

(2.) Dividing, and taking conjugates (comp. 260,

thus obtain the equation,

TT7
IV.

..

)3
-

/3
*-

= 2;

V.

or

8=

that the Vector

(3.),

2a
-7=

y+a

and 216,

)3

(5.) ),

is

we

2y
y +a

or

VI.
e

|3

= - 7=^

a,

VII.

if

= i

(y

a)

thus denoting here the vector OE (fig. 58) of the middle point of the chord
may then say that the harmonic mean between any two lines is (as

AC.

We

in algebra) the fourth proportional to their semisum, and

(3.)

Geometrically,

VIII.

we have thus

A AOB a

EOC

to themselves.

the similar triangles,

VIII

A AOE oc

BOG

whence, either because the angles OBA and OCA, or because the angles OAC and
OBC are equal, we may infer (comp. 260, (5.) ) that, when the equation I. is
satisfied, the four points o, A, B, c, if not collinear, are concircular.
(4.)

We have
IX.

also the similarities,

.

A OEC a CEB,

and

IX

A OEA a

AEB

or the equations,

X...^- =^,
y

and

X ...^=^;
a c

CIRCULAR AND HARMONIC GROUP.

ARTS. 260, 261.]

in fact

we

have,

by VI. and VII.,

-

Hence the

(5.)

tt

mean proportional (227) between

line EC, in fig. 58, is the

the lines EO and EB

and the

299

or in words, the semisum (OE), the semidifference (EC),

excess (BE) of the

semisum over

harmonic mean (OB), form (as in

the

algebra) & continued proportion (227).

form thus a con
(6.) Conversely, if any three coinitial vectors, EO, EC, EB,

tinued proportion, and if we take EA = CE, then the four points OABC will
for example, the points APBP of
compose a circular and harmonic group
so as to form such a group.*
fig. 57 are arranged
;

(7.)

It is easy to prove that, for the inscribed quadrilateral

OABC of

fig.

58,

the rectangles under opposite sides are each equal to half of the rectangle under

the diagonals

which geometrical relation answers

monic equations (comp. 259,

XIII.
(8.)

(10.)

= + 2
(OBAC)

XIIF.

Hence, or in other ways,

to either of the

it

may

(OCAB)

= +

J.

be inferred that these diagonals,

which they belong

and that thus the

OB, AC, are conjugate chords of the circle to

that each passes through the pole of the other ,

(9.)

or

by

the same conditions,

we have

XIV.
AC

is

it is

DB

is

the

intersects

easy to prove, either by quaternions

the harmonic equations

(ABCO)

the harmonic

between BC and BA

(BCOA)

mean between AB and AO

In any such group, any two

quadrilateral), as for example o and

A and

BO

is

such a

mean

CB.

opposite points (or opposite corners of the

B,

may

be said to be harmonically conju

gate to each other, with respect to the two other points,

these two points

= (COAB) = - 1

and CA between co and

(10.)

when

in the sense
line

o.

Under

geometry, that

so that

which the chord AC prolonged

second tangent from the point D, in

the tangent at

two anhar-

A and

and we

see that

c are given, then to every third point o (whether

in a given plane, or in space) there always corresponds & fourth point B, which
is in this sense conjugate to that third
point this fourth point being always
:

complanar with the three points

*

Compare the Note

to 255, (2.).

In

A, c, o,

fig.

and being even

concircular with them,

58, the centre of the circle

OABC

is

concircular with the

three points o, B, B.

2Q

ELEMENTS OF QUATERNIONS.

300

[II.

7-in.

11.

1.

unless they happen to be collinear with each other in which extreme (or
limiting) case, the fourth point B is still determined, but is now collinear with
;

the others (as in 26, &c.).

When, after thus selecting two* points, A and c, or

or fixed, we determine (10.) the harmonic conjugates B, B

(11.)

given

treating
,

B",

them

as

with respect

then the three pairs of points,

any three assumed points, o, o
o B
o, B
may be said to form an Involution,^ either on the right
line AC, (in which case it will
only be one of an already well-known kind), or
to them, of
;

in
A,

o",

o",

B",

a plane through that line, or even generally in space and the two points
c may in all these cases be said to be the two Double Points (or Foci) of this
:

But

Involution.

Quaternions,
(12.)

is

We

the field thus opened, for geometrical investigation by

far too extensive to be more than mentioned here.

shall therefore only at present add, that the conception of the

harmonic mean between two vectors

such,

and need not be limited

extended to any number of

easily be

may

to the plane

since

harmonic mean of the n arbitrary vectors

.
<n,

we may define
a n when it
.

that

r\

is

satisfies

the
the

equation,

XV.

fl

= - (Eoi +

Ean

or

XVI.

wRu = 2Ka.

Finally, as regards the notation Ea, and the definition (258) of the
reciprocal of a vector, it may be observed that if we had chosen to define
(13.)

reciprocal vectors as

having similar (instead of

opposite) directions,

we should

but should
indeed have had the positive sign in the equation 258, VII.
have been obliged to write, instead of 258, IX., the much less simple formula,
;

RL>

= -

IRt>.

* There is a sense in which the geometrical process here spoken of can be applied, even when
the two fixed points, or foci, are imaginary.
Compare the Geometric Superietire of M. Chasles,
page 136.
f Compare the Note to 259,

(11.).

AnTs.261, 262.]

ENUNCIATIONS OF THE ASSOCIATIVE PRINCIPLE.

CHAPTER

301

III.

ON DIPLANAR QUATERNIONS, OR QUOTIENTS OF VECTORS IN

SPACE: AND ESPECIALLY ON THE ASSOCIATIVE PRINCIPLE
OF MULTIPLICATION OF SUCH QUATERNIONS.
SECTION

1.

some Enunciations of the Associative Property, or Principle,

Oil

of Multiplication of Diplanar (Quaternions.

we have

262. In the preceding chapter

as

had been proposed (224, 225),

given plane (that of

i)

confined ourselves almost entirely,

to the considerations of quaternions in a

alluding only, in some instances, to possible exten

But we must now return to consider, as in the

sions* of results so obtained.

First Chapter of this Second Book, the subject of General Quotients of Vectors

and

which has hitherto been

especially their Associative Multiplication (223),

only proved in connexion with the Distributive Principle (212), and with the
Laws of the Symbols, i,j, k (183). And first we shall give a few geometrical

which

enunciations of that associative principle,

distributive one,

and in which

multiplication of versors

it

shall be

independent of the

will be sufficient to consider (comp. 191) the

because the multiplication of tensors

is

evidently

an

associative operation, as corresponding simply to arithmetical multiplication,

or to the composition of ratios in geometry. f

throughout the present chapter, that g,

and distinct planes

versors, in three given

* As in
227, (3.); 242, (7.); 254,

r, s

are

We

shall therefore suppose,

some

three given but arbitrary

and our object

will be to

throw some

260,
257, (6.) and (7.)
(7.)
259, (8.), (9.), (10.), (11.)
and 261, (11.) and (12.).
t Or, more generally, for any three pairs of magnitudes, each pair separately being homogeneous.
J If the factors q, r, s were complanar, we could always (by 120) put them under the forms,
;

(10.);

and then should have (comp. 183,

(1.)

sr

the two equal ternary products,

q

-- =
= 5/8
fi

8
S 7
= s .rq;
-=--

y a
would be proved without any

so that in this case (comp.

224) the associative property

difficulty.

ELEMENTS OF QUATERNIONS.

302

by new enunciations in

additional light,

this section,

[II.

m.

1.

and by new demon

on the very important, although very simple, Associative

which may be written thus:

strations in the next,

Formula (223,

II.),

I.

or thus,

more

II.

and

sr

q=

rq

fully,
.

q q

if

t,

sr,

being here three new and derived

and

rq,

ss

versors, in three

new and derived

planes.

we may

263. Already

see that this Associative Theorem of Multiplication,

in all its forms, has an essential reference to a System of Six Planes,

namely

the planes of these six versors,

IV.

q, r, s, rq, sr, srq,

IT.

or

q, r, s, s

on the judicious selection and arrangement of which, the clearness and

elegance of every geometrical statement or proof of the theorem must very

much depend

while the versor character of the factors (in the only part of
the theorem for which proof is required) suggests a reference to a Sphere,

namely

to

what we have

called the unit-sphere (128).

And

the three following

arrangements of the six planes appear to be the most natural and simple that
can be considered namely, 1st, the arrangement in which the planes all pass
:

Ilnd, that in which they all touch

and Illrd, that in which they are the six faces of an inscribed
through the centre of the sphere

its

surface

We

solid.

proceed to consider successively these three arrangements.

264. When the first arrangement (263) is adopted, it is natural to employ
arcs of great circles, as representatives of the versors, on the plan of Art. 162.
Representing thus the factor q by the arc AB, and r by the successive arc BC,

we represent (167) their product rq, or s by AC or by any equal arc (165),

such as PE, in fig. 59, may be
supposed to be. Again, representing s by EF,
we shall have DF as the representative of the
,

ternary product

rq, or ss

order of association.

nary product,

sr

mine three new

q,

To

or

taken in one

t,

represent the other ter

or q

q,

we may
G,

points,

H,

*i,

first

deter

by arcual

equations (165), between GH, BC, and between

HI, EF, so that BC, EF intersect in H, as the
arcs representing s

and

had intersected in E

an arc GI which represents

sr or q

may

and then,

after thus finding

determine three other points K,

L,

M,

ASSOCIATIVE PRINCIPLE, SYSTEM OE SIX PLANES.

303

new

arcs,

ARTS. 262-264.]

by equations between KL, AB, and between LM, GI, so that

KL, LM, represent q and q and that AB, GI intersect in L

these two

shall

have an

arc,

theorem then

is,

sense of Art.

165

namely KM, which represents

that this last arc

KM

is

way we

for in this

And

q q as required.

the

equal to the former arc DF, in the full

or that when (as under the foregoing conditions of con

struction) the five arcual equations,

I.

r\

AB = n KL,

n BO = n GH,

exist 9 then this sixth equation

EF = n HI,

r\

of the same kind

II.

n DF = n

n AC =

DE,

r>

GI

= A LM,

is satisfied also,

KM

K and M, being both on the same great circle as the two pre
or D and M being on the great circle
determined
viously
points, D and F
DF
and KM, of that great circle, or the
two
F
K
and
the
and
arcs,
through
the two points,

two

dotted

DK,

arcs,

FM

in

the

figure,

being equally

and

long,

similarly

directed (165).
(1.)

Or, after determining the nine points

so as to satisfy the three

middle equations I., we might determine the three other points K, L, M, without
any other arcual equations, as intersections of the three pairs of arcs AB, DF
and then the theorem would be, that (if these three last points
AB, GI DF, GI
;

be suitably distinguished from their own opposites upon the sphere) the two
extreme equations L, and the equation II., are satisfied.
(2.)

The same geometrical theorem may

first, third,

tively

and fifth

also be thus enunciated

sides (KL, GH, ED) of a spherical hexagon

and arcually equal (165)

to the first, second,

and third

KLGHED

If the

be respec

sides (AB, EC, CA)

of

a spherical triangle ABC, then the second, fourth, and sixth sides (LG, HE, DK) of
the

same hexagon are equal

to the three successive sides

(MI, IF,

FM) of another

spherical triangle MIF.

be also

said, that if five successive sides (KL,

ED) of one
be
the
and
to
spherical hexagon
five successive
respectively
arcually equal
diagonals (AB, MI, BC, IF, CA) of another such hexagon (AMBICF), then the
(3.)

It

may

sixth side (DK) of the first is equal to the sixth diagonal (FM) of the second.
(4.)

Or,

if

we adopt

the conception mentioned in 180,

(3.), of

an arcual

sum, and denote such a sum by inserting + between the symbols of the two
summands, that of the added arc being written to the left-hand, we may state
the theorem, in connexion with the recent
III.

where B and F

may

n DF +

BA =

EF + n

fig.

BC,

59,
if

by the formula
n DA = A EC

denote any two points upon the sphere.

ELEMENTS OP QUATERNIONS.

304

We may

(5.)

IV.
(6.)

if

GH

^ ED + ^

a moment,

If, for

KL =

we agree
Y.

we may then
VI.
(7.)

D-E + H-G

still

referred to, the theorem

if

n AB =

HE +

n LG =

may
.

0.

B- A,

r"

little

more

lucidly thus

K-D + E-H + G-L = O.

then

be supposed to denote any three

)

above

be said to be, that

+ n

+ n =

n"

(n + n)

an Associative Operation.
any independent demonstration be given, of the truth

or in words, that Addition of Arcs on a Sphere

Conversely,

DK +

to be added according to the rule (180, (3.)

VII..

if

sub-art. (2.))

to write (comp. Art. 1),

L-K = O,

more simply,

which are

diplanar arcs,

(8.)

then

0,

express the recent statement IV. a

..if

Or

and compare

59,

fig.

+ n

l.

the same principle, although somewhat less

also express*

simply, as follows (see again

m.

[II.

is

any one of the foregoing statements, considered as expressing a theorem of

spherical geometry^ a new proof will thereby be furnished, of the associative
of

property of multiplication of quaternions.

265. In the second arrangement (263) of the

six planes, instead of repre

senting the three given versors, and their partial or total products, by arcs, it
natural to represent them (174, II.) by angles on the sphere. Conceive then
that the two versors, q and r, are repre
is

fig. 60, by the two spherical

EAB
and ABE and therefore (175)
angles,

sented, in

that their product, rq or

s , is represented
the
external
vertical
by
angle at E, of the
Let the second versor r be
triangle ABE.

by the angle FBC, and the

by BCF then the other

also represented

third versor

sented

will be repre

by the external angle

represent the first

binary product, sr or q

new

at F, of the

ternary product, t =

ss

rq,

triangle BCF.

we have only

Again, to
to take the

* Some of these formulae and

figures, in connexion with the associative principle, are taken,
though for the most part with modifications, from the author s Sixth Lecture on Quaternions, in
which that whole subject is very fully treated. Comp. the Note to page 160.
t Such a demonstration, namely a deduction of the equation II. from the five equations I., by

known

properties of spherical conies, will be briefly given in the ensuing section.

AETS. 264-266.]

THREE ARRANGEMENTS OF PLANES.

external angle at

of the triangle BCD,

if

305

D be a point determined by the

be equal to BCF, and DEC supple

two conditions, that the angle ECD shall

mentary to BE A. On the other hand, if we conceive a point D determined
by the conditions that D AF shall be equal to EAB, and AFD supplementary to
CFB, then the external angle at
second ternary product, q q = sr

AFD

of the triangle

will represent the

must
q, which (by the associative principle)
be equal to the first. Conceiving then that ED is prolonged to G, and FD to H,
their
the two spherical angles, GDC and AD H, must be equal in all respects
.

vertices

D and D

coinciding,

being not only equal

in

and the

rotations (174, 177)

Or, to express the

amount, but also similarly directed.

same thing otherwise, we may enunciate (262)

saying, that when the three angular equations,
I.

which they represent

ABE = FBC,

BCF = ECD,

the Associative Principle

DEC =

7T

by

- BEA,

are satisfied, then these three other equations,

II.

DAF = EAB,

For not only

are satisfied also.

FDA = CDE,
is this

AFD =

7T

CFB,

theorem of spherical geometry a con

sequence of the associative principle of multiplication of quaternions, but con

versely

any independent demonstration*

of the

theorem

is,

at the

same time,

a proof of the principle.

266. The third arrangement (263) of the six planes may be illustrated
by conceiving a gauche hexagon, AB CA BC to be inscribed in a sphere, in such
,

a manner that the intersection D of the three planes,

C AB , B CA , A BC , is on the surface ; and therefore that
the three small

circles,

denoted by these three

last

symbols, concur in one point D while the

second intersection of the two other small circles,

triliteral

AB

C,

CA

annexed
first

B,

may

be denoted by the letter D as in the

Let it be also for simplicity at
,

fig. 61.

supposed, that (as in the figure) the five circular

Fig. 61.

successions*
I.

are all direct

* Such as

we

spherical conies.

HAMILTON

AB

D,

AB CD

B CAD,

A BC

or that the five inscribed quadrilaterals, denoted

shall sketch, in the following section, with the help of the

to the foregoing Article.

D,

by these symbols

known

properties of the

Compare the Note

ELEMENTS OF QUATERNIONS.

ELEMENTS OP QUATERNIONS.

306

are all uncrossed ones.

I.,

three versors, q, r,

TT

260,

(9.)

is

it

^
U

B D

DC

>

AB
= + TT
AC

-.

= + TT
U DA
IT
BD

TT
r =

1.

allowed to introduce

each having two expressions, as follows

s,

II.

Then (by

[II. in.

CA

r-

-.

CB

TT BD

^ CD

A B

CA

although (by the cited sub- article) the last members of these three formulae
should receive the negative sign, if the first, third, and fourth of the successions
I.

were to become

We have thus

ones.

(by 191) the derived expressions,

- TT DA/ - TT A B
DC
BC

TTT

the corresponding quadrilaterals were crossed

indirect, or if

whereof, however, the two versors in the

signs,

the fifth succession

if

if

formula,

IV.
since,

by the

ss

rq

formula would

first

were indirect

the second succession were such.

TTT

and

I.

- TT CD - TT D A
CB
AB

BD
= TT
U
BC

-,

q q

associative principle, these

differ in their

and those in the second

Hence,

sr

two

DA
= TT
U
7
AC

last versors are to

be equal,

follows that, under the supposed conditions of construction, the four points,
B, c , A, D , compose a circular and direct succession ; or that the quadrilateral,

it

BC AD /JS plane, inscriptible,* and uncrossed.

267. It is easy, by suitable changes of sign, to adapt the recent reasoning
and thus to
to the case where some or all of the successions I. are indirect
;

fronV the associative principle, this theorem of spherical geometry if

AB CA BC be a spherical hexagon, such that the three small circles C AB B CA A BC

infer,

concur in one point D, then, 1st, the three other small

in another point,
the

number^pf

conversely,

and Ilnd, of the six

those which are indirect

is

circles, AB

C,

CA

B,

circular successions, 266, 1.,

always

any independent demonstration!

even-

BC A, concur

and BC AD

(including zero).

of this geometrical

new proof of the associative principle.

268. The same fertile principle of associative

And

theorem

will be a

multiplication

may

enunciated in other ways, without limiting the factors to be versors,

*

Of

AD are known to be homospheric (comp. 260, (10.) ), the

would follow from its being plane, if the latter were
here deduced from the equality of the two versors IV., on the plan of

course, since the four points BC

inscriptibility of the quadrilateral in a circle

otherwise proved
260,

but

it is

(9.).

An

be

and

in the following section.

elementary proof, by stenographic projection, will be proposed

AETS. 266-268.]

THIRD ARRANGEMENT, SPHERICAL HEXAGON.

307

without introducing the conception of a sphere. Thus we may say (comp.

264, (2.) ), that if o ABCDEF (comp. 35) be any pencil of six rays in space,
and o A B C any pencil of three rays, and if the three angles AOB, COD, EOF of
.

C OA , A OB of the
pencil be respectively equal to the angles B OC
another
then
of
three
o
can
be assigned, such
second,
pencil
rays,
that the three other angles BOC, DOE, FOA of the first pencil shall be equal to

the

first

A"B"C",

the angles

B"OC",

C"OA",

A"OB"

of the third

equality of angles (with one vertex)

and

being here understood (comp. 165) to include complanarity,

similarity of

direction of rotations.

Again (comp. 264,

(1.)

(4,)),

which the four vectors afly

are

any two

lines in space

= %p,

it

7
under

establish the following formula, in

(226), but

and ?

O
It,
...
for,

we may

form a complanar proportion

this last condition,

we have

]3

(comp. 125),

a y
a
7
Another
enunciation
of
the
associative
(2.)
principle is the following
o
y
O
P =
TTT
P =S
-111.
if then
-,
-;
a 7
o
7 a
if we determine
six
new
(120)
vectors, r\Qi, and /cA^u, so that

</

J.T-

for

$
- = -,
7

(0

)3
- = rj

IV.

r)

8
Z
whence - = -,

and

>

we

shall

have the transformations,

V
(3.)

true,

fjL

-^.L

-^.

l^-^L^-!!!-^
""

^ - %
=

VT

Conversely, the assertion that this last equation or proportion VI. is

are connected by the five proportions
JUL

whenever the twelve vectors a

a form of enunciation of the associative principle

IV.,
(comp. IV. and V.) to the equation,
is

---

atleastif

for it conducts

|(>e;

but, even with this last restriction, the three fact or- quotients in

represent any

VII.

three quaternions.
2

may

ELEMENTS OF QUATERNIONS.

308

SECTION

[II.

2.

2.

On some Geometrical Proofs

of the Associative Property of multi

plication of Ctuaternions, which are independent of the
Distributive* Principle.

269.

We propose,

in this section, to furnish three geometrical

Demonstra

tions of the Associative Principle, in connexion with the three figures (59-61)
which were employed in the last section for its Enunciation ; and with the
three arrangements of six planes,
first

which were described in Art. 263.

of these proofs will suppose the

knowledge

of a

The two

few properties of spherical

but the third will only employ the doctrine of stereographic

therefore
be of a more strictly elementary character. The
and
will
projection,
Principle itself is, however, of such great importance in this Calculus, that its
conies (196, (11.)

nature and

its

evidence can scarcely be put in too

many

different points of

view.

270.

The only

Article assume as

properties of a spherical conic, which we shall in this

known,f are the three following 1st, that through any
:

three given points on a given sphere, which are not on a great circle, a conic
can be described (consisting generally of two opposite ovals), which shall have
a given great circle for one of its two cyclic arcs
Ilnd, that if a transversal arc
;

cut both these arcs, and the conic, the intercepts (suitably measured) on this

the vertex of a spherical angle move

along the conic, while its legs pass always through two fixed points thereof, those
legs intercept a constant interval, upon each cyclic arc, separately taken.

transversal are equal

and Illrd, that

if

properties, we see that if, in fig. 59, we conceive a

conic
to
be
described, so as to pass through the three points B, F, H,
spherical

Admitting these three

Compare 224 and 262 and the Note to page 245.

t The reader may consult the Translation (Dublin, 1841, pp. 46, 50, 55) by the present Dean
Graves, of two Memoirs by M. Chasles, on Cones of the Second Degree, and
Spherical Conies.
[If a cone have one system of cyclic sections parallel
to APB, on inversion from o, the vertex of the cone, it is seen to have a
;

second system parallel to the tangent plane at o to the sphere through the
In the figure the circle OAB is the section of
vertex and the circle APB.
this sphere by the plane through two edges of the cone OA and OB, while
oc (parallel to AB) and oc (tangent to OAB) are the traces of the cyclic

BAO = ZAOC proves the Ilnd property. Again, if OA

/.C OB =
and OB are the traces on the cyclic plane parallel to APB by the planes POA
and POB respectively, Z.APB = /.A OB
OA being parallel to PA and OB to
But as P moves along the circle APB, the angle APB is constant, and thus the Illrd property
PB.

planes.

also proved.]

is

PROOFS BY SPHERICAL CONICS.

ABTS. 269, 270.]

309

have the great circle DAEC for one cyclic arc, the second and third
equations I. of 264 will prove that the arc GLIM is the other cyclic arc for this

and

to

conic

and

the

first

equation

I.

the arcual chord

if

proves next that the conic passes through K

prolonged, the two remaining
;

FK be drawn and

equations prove that it meets the cyclic arcs in D and M after which, the
equation II. of the same Art. 264 immediately results, at least with the
;

arrangement* adopted in the figure.

The

(1.)

1st property is easily seen to correspond to the possibility of

circumscribing a circle about a given plane triangle, namely that of which

the corners are the intersections of a plane parallel to the plane of the given

with the three radii drawn to the three given points upon the
but it may be worth while, as an exercise, to prove here the Ilnd

cyclic arc,

sphere

property by quaternions.
a
(2.) Take then the equation of

XII.) be written thus

I.

196,

cyclic cone,

(8.),

which

may

(by 196,

..S-S-^ = N-S;
a

]3

and

II.

let

j3

75

N 75*
p

p and p being thus two rays (or sides) of the cone, which may also be con
sidered to be the vectors of two points P and P of a spherical conic, by
supposing that their lengths are each unity. Let r and / be the vectors of
the two points T and T on the two cyclic arcs, in which the arcual chord PP

them

of the conic cuts

III.

so that

.S-=0,
a

The theorem may then be

V.

if

S^ = 0,

and

IV.

Tr = T/=1.

stated thus

= XT + X T,

that

VI.

then

or that this expression VI. satisfies II.,

if

p = X T + XT

the equations

I.

III.

IV. V. be

satisfied.

Now, by

III. V. VI.,

we have

.a

whence

it

follows that the

first

* Modifications of that
arrangement

adapt the reasoning.

members
may be

/3

]3

of I.

and

II. are equal,

conceived, to which however

it

and

it

only

would be easy to

ELEMENTS OF QUATERNIONS.

310

remains to prove that their second members are equal

if

[II. ni.

also, or that T/o

2.

= Tp,

T/=Tr.
Accordingly we have, by Y. and VI.,

VIII.

and the property

^2

in question

X -X

T
L
= g-iQ, by 200, (11.), and 204,
(19.)

T
r
.

is

proved.

To prove

the associative principle, with the help of fig. 60, three

other properties of a spherical conic shall be supposed known :* 1st, that for
every such curve two focal points exist, possessing several important relations

271.

to

it,

one of which

is,

that

conic can be constructed

tangents be

drawn

if

these two foci

Ilnd, that

to the conic,

and

if,

and one

tangent arc be given, the

from any point upon the sphere, two

also two arcs to the foci, then one focal arc

makes with one tangent the same angle as the other focal arc with the other
tangent and Illrd, that if a spherical quadrilateral be circumscribed to such
;

a conic (supposed here for simplicity to be a spherical ellipse, or the opposite

ellipse being neglected), opposite sides subtend supplementary angles, at either of
the two (interior) foci. Admitting these known properties, and supposing the

arrangement to be as in fig. 60, we may conceive a conic described, which

shall have E and F for its two focal points, and shall touch the arc BC
and
;

then the two

first

of the equations

I.,

in 265, will prove that

it

touches also

the arcs AB and CD, while the third of those equations proves that it touches
AD, so that ABCD is a circum scribed f quadrilateral: after which the three
equations II., of the same article, are consequences of the same properties of
the curve.J
272. Finally, to prove the same important Principle in a more com
pletely elementary way, by means of the arrangement represented in fig. 61,
or to prove the theorem of spherical geometry enunciated in Art. 267,

* The reader

may

again consult pages 46 and 50 of the Translation lately cited.

two by two.

In

we

strictness,

there are of course four foci, opposite

f The writer has elsewhere proposed the notation, EF

(. .)

ABCD, to denote the relation of the focal

points E, F to this circumscribed quadrilateral.

J [The two cyclic arcs and a point determine a spherical conic.

Referring to the Note on 270,
describe a sphere to touch one cyclic plane at the point o.
Then if OA is given, take the section APB
of the sphere by a plane parallel to the second cyclic plane, and the cone is determined.
Recipro

The Ilnd property is the reciprocal of the Ilnd of 270, and

cating this, the 1st property follows.
the Illrd is easily derived by reciprocating the Illrd of 270, remembering that for a point P on the
remaining arc of the circle APB, L AP B + L APB = TT.]

PROOF BY STEREOGRAPHIC PROJECTION.

AETS, 270-272.]

may assume

the point

311

as the pole of a stereographic projection, in

which the

three small circles through that point shall be repre

but the three others by circles,

And then (inter
changing accents) the theorem comes to be thus
sented by right

all

lines,

being in one

stated

common plane.*

J/A

c be any three points (comp. fig. 62)

on the sides BC, CA, AB of any plane triangle, or on

those sides prolonged, then, 1st, the three
I.

will meet in one point

AB

circles,

A BC

and Ilnd, an

B CA

number

even

(if

any) of the

six (linear

or circular) successions,
II.

AB

C,

BC A, CA

and

B,

IF.

AB

D,

A BC

D,

B CA D,

will be direct
an even number therefore also (if any) being indirect. But,
under this/orm,f the theorem can be proved by very elementary considerations,
and still without any employment of the distributive principle (224, 262).
is evident from the
(1.) The first part of the theorem, as thus stated,
;

Third Book of Euclid

but to prove both parts together,

may

it

be useful to

proceed as follows, admitting the conception (235) of amplitudes, or of angles

as representing rotations,

which

may

have any

values, positive or negative,

and

are to be added with attention to their signs.

(2.)

We may

thus write the three equations,

III.

AB

= mr,

to express the three collineations,

odd or

even,

AB

according as the point B

longation of that line

is direct

BC A = M

or indirect

C,

TT,

CA B =

w"zr,

&c. of fig. 62 ;* the integer, n, being

is

on the

finite line AC, or

or in other words, according as the

and similarly

for the

two other

on a pro

first succession II.

coefficients, n

and

n" .

[Invert figure 61 from the point D. The sphere becomes a plane, and the circles through D
right lines, the other circles remain circles.]
t The Associative Principle of Multiplication was stated nearly under this/orm, and was illus
trated by the same simple diagram, in paragraph XXII. of a communication by the present author,

and has been printed in the First and Second Editions of

Cyclopedia of the Physical Sciences (London and Glasgow, 1857 and 1860). The
same communication contained other illustrations and consequences of the same principle, which it
has not been thought necessary here to reproduce and others may be found in the Sixth of the

which was

entitled Letters on Quaternions,

the late Dr. Nichol

already cited Lectures on Quaternions (Dublin, 1853), from which (as already observed) some
of the formulae and figures of this Chapter have been taken.

author

| [AB C being the angle through which B A must be turned in the positive direction so as to coincide
with B c.l

ELEMENTS OF QUATERNIONS.

312

Again,

(3.)

if

OPQR be any four points in one plane, we

formula,

IV.

plane of the triangle ABC,

V.

and

AB D + DB

2, 3.

establish the

may

POQ + QOR = POR + 2mir,

with the same conception of addition of amplitudes

the

[II. in.

we may

write,

BC D + DC A = n

= mr,

then D be any point in

if

CA D + DA B =

tr,

n"ir

therefore,

VI.

(AB D + DC A) + (BC D +

DA

+ (CA D + DB

B)

C)

(n

+ n +

TT.
n")

Again, if any four points OPQR be not merely complanar but

we have the general formula,

(4.)

cular,

VII.

concir-

OPQ + QRO = pir 9

the integer p being odd or wen, according as the succession OPQTI is direct or
indirect ; if then we denote by D the second intersection of the first and second

whereof c

circles I.,

VIII.

p and p being

odd,

& first intersection,

is

AB D + DC A = prr,

we

when

the two

first

shall

have

BC D + DA B =

successions II

p ir,

are direct, but even in

the contrary case.

Hence, by VI., we have,

(5.)

IX.

CA D + DB C = p"w,

the third succession II

through the intersection
say,

p"

is

odd

QIC

even,

p+p +

p"

= n + n +

therefore always circular, or the third circle

is

X.

where

of the two first

and

it is direct

n"

I.

passes

or indirect, that

is

to

according as the number of even coefficients, among the

is itself even or odd ; or in other words,
according

five previously considered,

as the

number of

indirect successions,

among

the five previously considered,

is

even (including zero), or odd.

(6.)
is even,

In every case, therefore, the total number of successions of each kind

and both parts of the theorem are proved the importance of the
:

second part of

II

.)

(respecting the even partition, if any, of the six successions IT.

from
the necessity of proving that we have always, as in algebra,
arising

XI.
if q, r, s
(7.)

it

sr

= +

rq,

and

never

XII.

sr

=-

rq,

be any three actual quaternions.

The

associative principle of multiplication

the distributive principle,

by

which we cannot enter here.

may

also be proved, without

certain considerations of rotations of a system, on

ADDITIONAL FORMULA, NORM OF A VECTOR.

AETS. 272-274.]

SECTION

On some

313

3.

Additional Formulae.

273. Before concluding the Second Book, a few additional remarks

may

be made, as regards some of the notations and transformations which have

And first as to notation,
already occurred, or others analogous to them.

although we have reserved for the Third Book the interpretation of such
2
expressions as |3a, or a , yet we have agreed, in 210, (9.), to abridge the
2

2
and we now propose to abridge it
frequently occurring symbol (Ta) to Ta
still further to No, and to call this square of the tensor (or of the
length] of a
;

Norm

the

vector, a,

2
Tg = N#, and

shall therefore

as we had (in 190, &c.), the

Vector",
equation
the norm of the quaternion q (in 145, (11.)).
write generally, for any vector a, the formula,

of that

now

We

N^

called

I ... (Ta) 2 =

The equations (comp.

(1.)

II.

N/>

- 1

III.

V.

186,
.

Ta 2 = Na.

(1.) (2.) (3.) (4,)),

NP = N

IY.

Nfo -

a)

Na

..N(p-a)=N(j3-a),

represent, respectively, the unit-sphere; the sphere through A, with o for

centre; the sphere through

with the same centre A.
(2.)

The equations (comp.

YI.

N(p +

with A for centre; and the sphere through

o,

186,

= N(p -

a)

a)

(6.) (7.)

B,

),

VII.

Nfo -

]3)

= N(p -

),

represent, respectively, the plane through o, perpendicular to the line OA

and the plane which perpendicularly

274.

As

bisects

the line AB.

regards transformations, the few following

may

here be added,

which relate partly to the quaternion forms (204, 216, &c.) of the Equation
of the Ellipsoid.
(1.)

XVI.,

Changing

complanar, while 1
HAMILTON

Ec, by 259, VIII., in the equation 217,

and observing that the three vectors /o, Up, and EK are

K( K

of the ellipsoid,
:

T/o

p ) to E/o

TEp by

ELEMENTS OF QUATERNIONS.

258, that equation becomes,

when
2

divided
S

ELEMENTS OP QUATERNIONS.

314

by TE/o, and when the value 217,

employed

(5.)

for

[II. in.

3.

taken, and the notation 273

is

is

I...

of

which the

T (tp

member

first

and the second member

PK),

admit of being written* as

will soon be seen to

as K*

?.

connexion with the earlier forms (204, 216) of the equation of

the same surface, we introduce a new auxiliary vector, a or os, such that (comp.
(2.)

If, in

216, VIII.)

the equation may,

simple form

by 204,

be reduced to the following extremely

(14.),

III.

= T/3

T<r

which expresses that the locus of the neio auxiliary point s

while the line PS, or & called the mean sphere, 216, XIV.
;

is

what we have

p,

which

connects

any two corresponding points, p and s, on the ellipsoid and sphere, is seen to
to the fixed line j3; which is one element of the homology, mentioned

be parallel

in 216, (10.).
(3.)

It

is

IV.

if

p and

easy to prove that

and therefore

..Sf-S^SC,
a
6
o

be the vectors of two

o-

V.

6000

S~: S^ = S^:S^

new but corresponding

points,

p and

on

the ellipsoid and sphere ; whence it is easy to infer this other element of the
homology, that any two corresponding chords, PP and ss , of the two surfaces,
intersect each other

216,

(7.))

TTT

VI.

and

on the

cyclic

plane which has S for

normal (comp.

its cyclic

in fact, they intersect in the point T of which the vector

..T-

= xp

this point is

+ xp
-j-

X + X

xa + x a

X + X

if

a p
# = S^-,
C

and

a?

is,

op

=-S^;

on the plane just mentioned (comp. 216, XL), because

VII.

.S

Compare the Note

0.

to page 241.

HOMOLOGIES OF ELLIPSOID AND SPHEEE.

ART. 274.]

Quite similar results would have followed,

(4.)

vni.

..<r=-

if

315

we had assumed

P
which would have given again, as in

IX.

Ter

III.,

BS;
ay

X...8-

but with

T/3,

7
the

otfAer cyclic

been taken

plane, with

y instead of 8

for its normal,

might therefore have

(as asserted in 216, (10.) ), as another

plane of homology of ellipsoid

and sphere, with the same centre of homology as before namely, the point at
infinity on the line /3, or on the axis (204, (15.) ) of one of the two circum
:

scribed cylinders of revolution (comp. 220, (4.)

The same

(5.)

ellipsoid

).

in two other ways,

is,

homologous

to the

same

mean

sphere, with the same two cyclic planes as planes of homology, but with
a new centre of homology, which is the infinitely distant point on the axis of

the second circumscribed cylinder (or on the line AB

of the sub-article last

cited).

Although not specially connected with the

(6.)

general transformations

XXXIV.)

XL

ellipsoid, the following

be noted here (comp. 199, XII., and 204,

may

..

(7.)

XIII.
or the

The equations 204, XVI. and XXXV., give

.

UV?

UVU?

XIV.

UIV?

Ax

easily,

XV.

TIV? = TV?

..

more symbolical forms,

XIIF.

UVU = UV

XIV.

UIV =

Ax.

XT.

TIV

TV

and the identity 200, IX. becomes more evident, when we observe that

XVI.
(8.)

We have

*=1 =
q

The

which

= q(l-Kq).

(g"

(q

1 )(

Kg + 1

+ l)(Kq +

l)

and 218,

(10.)),

+
= Ng-l 2Vg
N? + 1 + 2Sq

formula,*

XVIII.
in

.q-Nq

also generally (comp. 200, (10.)

XVII
(9.)

U (rq + Kqr)

= U(Sr Sq +

and r may be any two

Vr V?)
.

^ (/Y)* q~\

quaternions, is not perhaps of

any great

* This formula was

given, but in like manner without proof, in page 587 of the author
Lectures on Quaternions. [It may be expressed in terms of p = (r2 g- 2 )i.
Use 210, XI. and XII.]
2 S 2

ELEMENTS OF QUATERNIONS.

316

importance in

[II. HI.

3.

but will be found to furnish a student with several useful

itself,

exercises in transformation.
(10.)
root,

x +

When

it

was

said, in 257, (1.), that zero

had only

itself for

a square-

the meaning was (comp. 225), that no binomial expression of the form

iy (228) could satisfy the equation,

XIX.

= q2 =

(x

it/)*

(x*

if)

2%,

any real or imaginary values of the two scalar coefficients x and y, different
from zero ;* for if bi-quaternions (214, (8.) ) be admitted, and if h again
denote, as in 256, (2.), the imaginary of algebra, then (comp. 257, (6.) and
= 0, the imaginary
(7.) ) we may write, generally, besides the real value, 0*

for

expression,

..04 =

XX.
v

and

v being thus

V,

So-Sff-Swf-Nff-Nfj-O;

if

any two real right quaternions, with equal norms

(or

with

equal tensors), in planes perpendicular to each other.

(11.)

For example, by 256,

and by the laws (183)

(2.)

of

ijk,

we have

the

transformations,

XXL
so

the biquaternion

that

symbol

=
..(* + hJY
i

-f +

+ hj

is

(#+/) =

+ AO =

one of the imaginary values of the

0*.

(12.)

In general, when bi-quaternions are admitted into calculation, not

only the square of one, but the product of two such factors may vanish, without
a circumstance which may throw some
either of them separately vanishing
:

light on the existence of those imaginary

which were treated of in 257.
(13.)

(or symbolical}

roots of equations,

For example, although the equation

XXII.
has no real roots except

1,

2
.

!=(?-!) (?+!)

and therefore cannot be

+ hv

0.

verified

of any other real scalar, or real quaternion, for

bi-quaternion-f v

if

we

by the

substitution

substitute for q the

q, yet
with the conditions 257, XIII., this equation

XXII.

is

verified.

# Compare

the Note to page 289.

f r a symbolical square-root of positive unity.
f This includes the expression
hi, of 257,
Bi- quaternion in
Other such roots are
hk.
hj, and
[It is probable that Hamilton used the word
order to distinguish clearly the V
1 of algebra from the geometrical reals i, j, and k of the new
and in a Paper read before the
Calculus. In his earlier writings i, j, and k are called imaginaries
(!.)>

Boyal Irish Academy on November 11, 1844, the scalar of a quaternion is called the
and the vector, the
imaginary part." See p. 3, vol. iii., of the Proc. K.I. A.]
"

"real

part,"

BI-QUATEENIONS.

AET.274.]

It will be found, however, that

(14.)

when

317
two imaginary but non-evanes

norm of each is zero provided that we

2
or
the formula Ng = Sq* (204, XXII.)

cent factors give thus a null product, the

agree to extend to bi-quaternions

Right Part

V<?

Norm

of a Biquaternion (like that of an ordinary or real

the
to
Square of the Scalar Part,, minus the Square of the
equal

to define that the

quaternion)

is

each of these

two parts being generally imaginary, and the former

being what we have called a Si-scalar.

With

(15.)

and

this definition, if q

q be

any two

real quaternions,

we may

h be, as above, the ordinary imaginary of algebra,

formula

and

if

establish the

XXIII.
or (comp. 200, VII.,

As

N (q+hq

= (Sq + AS/) 2 - (Vq + /AY) 2

and 210, XX.),

XXIY.
(16.)

N (q + hq

N? - Nq

27*8

qKq.

regards the norm of the sum of any two real quaternions, or real

vectors (273), the following transformations are occasionally useful (comp.

220, (2.)):

XXV,
XXVI.
in each of

which

it is

N (q + q) = N(0y. Uq + Tq. TJj

N + a) = N (T/3 Ua + Ta TJ/3)
)

(|3

permitted to change the norms to the tensors of which

they are the squares, or to write

for

N.

BOOK

III.

ON QUATERNIONS, CONSIDERED AS PRODUCTS OR POWERS OF VECTORS

AND ON SOME APPLICATIONS OF QUATERNIONS,

CHAPTER

I.

ON THE INTERPRETATION OF A PRODUCT OF VECTORS, OR

POWER OF A VECTOR, AS A QUATERNION.
SECTION

On a

First

1.

of interpreting a Product of Two Vectors

as a duaternion.

Method

ART. 275. In the First Book of these Elements we interpreted, 1st, the
Ilnd, the sum of two
difference of any two directed right lines in space (4.)
the
of
one
such
or more such lines (5-9)
Illrd,
product
line, multiplied by
;

or into a positive or negative number (15)

divided by such a number (16), or by

SCALAR

IVth, the quotient of such a

what we have

line,

called generally a

and Yth, the sum

of a system of such lines, each affected

(97)
with a scalar coefficient (99), as being in each case itself (generally) a Directed
(17)

Line* in Space, or what

we have

called a

VECTOR

(1).

276. In the Second Book, the fundamental principle or pervading con

ception has been, that the Quotient of two such Vectors is, generally, a

QUATERNION

(112, 116).

to be remembered, that we have

which usually relates to what may be
or the quotient of two lines in space making either

It is

however

included under this general conception,

an Oblique Quotient,
an acute or an obtuse angle with each other (130), the three following particular
cases
1st, the limiting case, when the angle becomes null, or when the two

called

lines are similarly directed, in

positive scalar

right angles, or

which case the quotient degenerates (131) into a

Ilnd, the other limiting case,

when the

when

lines are oppositely directed,

the angle

and when

is

equal to

tico

in consequence

the quotient again degenerates, but now into a negative scalar ; and Illrd, the
intermediate case, when the angle is right, or when the two lines are perpen
dicular (132), instead of being parallel (15),
* The Fourth
Proportional to any three complanar

being another

HAMILTON

line in the
S

and when therefore their quotient

lines

has also been since interpreted (226), as

same plane.

ELEMENTS OF QUATERNIONS,

zT

ELEMENTS OF QUATERNIONS.

1, 2.

i.

[III.

becomes what we have called (132) a Right Quotient, or a EIGHT QUATERNION

which has been seen to be a case not less important than the two former
:

ones.

But no

277.

going
other

POWER

which

PRODUCT of

of a Vector

I.

in

Interpretation has been assigned, in either of the

Books, for a

<yj3a,

denote

fore

or for the SQUARE, or

so that the Symbols,

.

|3o,

a, /3,

two or more Vectors

two

and

II.

vectors,

but

a3

1
,

denotes a scalar, remain as yet

and we are therefore free

to assign, at this stage, any

entirely uninterpreted
meanings to these new symbols, or new combinations of symbols, which shall not
contradict each other, and shall appear to be consistent with convenience and
;

And

analogy.

do so will be the chief object of this First Chapter of the

which is designed to be a much

to

Third (and last) Book of these Elements

shorter one than either of the foregoing.

As

278.

a vector

a.

is

plied into*

commencement

we

of such Interpretation

multiplied by another vector

shall here define, that

or that the latter vector

j3,

informer, or that the product

j3a is

is

multi

obtained, when the multiplier-

by the reciprocal Ea (258) of the multiplicand-line a ; as we had

proved (136) that one quaternion is multiplied into another, when it is divided by
the reciprocal thereof.
In symbols, we shall therefore write, as a first defi
nition, the formula
line /3 is divided

I...|3a

j3:Ea;

where

.Ea = -Ua Ta(258, VIL).

II..

And we

proceed to consider, in the following section, some of the general

consequences of this definition, or interpretation, of a Product of two Vectors,
as being equal to a certain Quotient, or Quaternion.

SECTION

On some Consequences
279.

The

of the foregoing Interpretation,

definition (278) gives the formula

I.

it

2.

gives therefore,

/3a

=-

and

by 259, VIII. the general

,

IL../3a = Kaj3;

or

similarly,

II

Compare the Notes

a/3

relation,

a/3

= K/3a.

to pages 147, 159.

t-

ARTS. 276-281.]

INTERPRETATION OF A PRODUCT OF TWO VECTORS.

The Products of two

jugate Quaternions
(168),

(generally) a

is

opposite orders, are therefore

Multiplication of Vectors, like that of

Non- Commutative

from

(i.) It follows

two

Vectors, taken in

and the

323

Con

Quaternions

Operation.

II. (by 196, cornp. 223, (1.)), that

III.

Sj3a

-f

Saj3

= i

()3a

a/3).

It follows also (by 204, comp. again 223, (1.)), that

(2.)

IV.

= i
Y/3a = Ya]3

(]3a

a/3)

we have the

280. Again, by the same general formula 259, VIII.,

trans

formations,
"

follows, then,

+ a )~

by taking conjugates

also,

Ea"

from the definition (278), that

IL..j3(aW)
whence

JL.

"

"E(a

it

+ K-^- -A
-K^L-K^
Ea
E/3
E/3
Ej3

___

III.
Multiplication of Vectors

is,

(a

(279),

+ a

j3

=|3a +

/3a

we have
=

a/3

this other general equation,

aft.

therefore, like that of Quaternions (212), a

Doubly

Distributive Operation.

As we have not yet assigned any signification for a ternary product

vectors, such as y]3a, we are not yet prepared to pronounce, whether the
281.

of

Associative Principle (223) of Multiplication of Quaternions does or does not

But we can already

extend to Vector-Multiplication.
consequences from the
attention
Vectors

derive several other

definition (278) of a binary product, )3a

among which,
be called to the Scalar character of a Product of two Parallel
and to the Right character of a Product of two Perpendicular Vectors,
;

may

or of two lines at right angles with each other.

(1.)

The

definition (278)

may

I.../3a
it

be thus written,

= -Tj3.Ta.TJ(/3:a);

gives, therefore,
II.

T]3a

= T]3 .Ta

III.

Uj3a = U(j3

a)

= Uj3

.Ua

the tensor and versor of the product of two vectors being thus equal (as for
quaternions, 191) to the product of the tensors, and to the product of the versors,
respectively.

2T2

ELEMENTS OF QUATEBMONS.

324

Writing for abridgment (comp. 208),

= Ax
b = T/3,
a = Ta,
IY.
7

(2.)

we have

(/3

i.

[III.

x = z

a),

(]3

2.

..,

thus,

Y.

VII.

Tj3a

= ba

SUjSa

VI.

= -

SUa/3

..

Sj3a

CQStf

Sa/3

= - ba cos*

VIII.

^ j3a

any two

so that (comp. 198) the angle of the product of

- x

TT

vectors

is

the supple

ment of the angle of the quotient.

(3.) We have next the transformations (comp. again 208),

IX.

XI.

XII.
so that

TV/3a = TVa/3 = la
IV/3a

= -

sin

yba

IUV/3a = Ax

sin

X.

XF

= -

)3a

XII

TVUjSa
IVa/3

= +

IUVa)3 =

TVUa/3 =
-yah sin

Ax

a/3

sin x

= + 7

Ae rotation round the axis of a product of two vectors, from the multiplier

to the multiplicand, is positive.

(4.)

It follows also,

product, j3a,

by IX., that the

tensor of the right part of such a

equal to the parallelogram under the factors or to the double of

whereof those two factors a, j3, or OA, OB, are two

is

the area of the triangle OAB,

coinitial sides

so that if

we denote here

this last-mentioned area

by the

symbol

A OAB,
we may

write the equation,

XIII.

and the
gram, of

TVjSa = parallelogram under

a, j3,

= 2 A OAB

index, IVjSa, is a right line perpendicular to the plane of this parallelo

which

line the length represents its area, in the sense that they bear

equal ratios to their respective units (of length and of area).

(5.)

Hence, by 279, IV.,

XIV.

T (|3a

For any two

= XV.

(6.)

Sj3a

a/3)

vectors, a,

Na

.S(/3

= 2 x parallelogram = 4 A OAB.
/3,

a)

XVI.

Na.V(/3
V/3a =

a)

or briefty,*

XVII.

..j3o

= -No.()3 :o),

with the signification (273) of Na, as denoting (Ta) 2

* All the
consequences of the interpretation (278), of the product 0a of two vectors, might be
deduced from this formula XVII.
which, however, it would not have been so natural to have
assumed for a definition of that symbol, as it was to assume the formula 278, I.
;

PRODUCTS OF PARALLELS AND PERPENDICULARS.

AET. 281.]
(7.)

two

If the

factor- lines be perpendicular to each other, so that x is a

becomes a

right angle, then the parallelogram

(4.)

becomes a right quaternion (132)

so that

XVIII.
(8.)

Under

XIX.
(9.)

S/3o

Sa/3

j3a

write,

and

j3 JL a,

if

0,

and the product

rectangle,

we may

reciprocally.

the same condition of perpendicularity,

z/Ba

On

325

= Za]3 =

XX.

the other hand,

= - yba;

I/3a

XXI.

the two factor-lines

if

\)Q

Ia/3

parallel, the rig lit part

which

of their product vanishes, or that product reduces itself to a scalar,

opposite directions

we may

for

XXII.

if

]3

is

two vectors multiplied have similar or

negative or positive according as the

establish the formula,

V/3a =

then

a,

||

Va/3

0,

and, under the same condition of parallelism,

XXIII.
the tipper or the
(10.)

j3

Sj3a

= Sa/3 = +

XXY.
(11.)

If

a, |3,

and

also write (by 279, (1.)

perpendicularity and. formula of parallelism

XXIY.

then

if /3 JL a,
if

]3

ba,

being taken, according as x =

/0t0er s/pw

We may

a/3

a,

||

j3a

then

7 be any three

j3a

= -

a/3,

a]3,

and reciprocally

considered as vectors of the corners

a, b, c, then because, by XYIL, /3a =

articles to 208 allow us to write,

XXVII.
XXVIII.

/3

a,

and

new

yj3

S (Y 7 /3 V)3a) =

sin a sin c cos B

IV (V7 j3

(IV

S)

V)3a)

positive scalars,

=-7

with the converse.

A, B, c of a spherical triangle, with sides equal to three

XXYI.

TT.

(2.)) the following formula of

= +

unit-lines,

0, or

(V 7 /3 V)3a) =
.

]3

the sub-

sin a sin c sin B

]3

j3,

tan B

upper or lower signs being taken, in the two last formulae, according as the
rotation round ]3 from a to
7 or that round B from A to c, is positive or
,

negative.
(12.)

The equation 274,

XXIX. ..T(^ + P K)

I.,

of the Ellipsoid,

IV -IV;

or

may now

XXX.

be written thus

T (tp + /OK)

= Ni

N*.

ELEMENTS OF QUATEKNTONS.

326

Under

282.

[Til.

i.

2.

the general head of a product of two parallel vectors, two

which furnish two first examples of Powers of Vectors

when
the
case
the two factors are equal, which gives this remark
namely, 1st,
able result, that the Square of a Vector is always equal to a negative Scalar;
and Ilnd, the case when the factors are fin the sense already defined, 258)
interesting cases occur,

which case

reciprocal to each other, in

their product

it

equal to Positive Unity

is

follows from the definition (278) that

so that each

may, in this case, be con

Power of that other

sidered as equal to unity divided by the other, or to the

which has Negative Unity for its Exponent.
(1.)

When

square of

/3

a,

the product /3a reduces itself to what

and may denote by a 2

a,

and thus we may

we may

call the

write, as a particular

but important case of 281, XXIII., the formula (comp. 273),

norm (273) of
which expresses

so that the square of any vector a is equal to the negative of the

that vector

or to the negative of the square of the number Ta,

same

(185) the length of the

(2.)

II.

(3.)

vector.

More immediately, the

.

2
a = aa = a

Hence (compare the

III.

definition (278) gives,

:

Ea = -

(Ta}

= - Na,

as before.

notations 161, 190, 199, 204),

S a2 = .

Na

IV.

2
a =

and
V. ..T.a 2 = T(a 2 )= + Na=(Ta) 2 = Ta 2

the omission of the parentheses, or of the point, in this last symbol of a tensor,*
for the square of a vector, as well as for the square of a quaternion (190), being

thus justified

and in

like

VI.

manner we may

a =
2

U (a

write,

= - 1 = (Ua) 2 =

Ua

2
;

the square of an unit-vector (129) being always equal to negative unity, and

parentheses (or points) being again omitted.

(4.)

The equation

VII.
it

z
.

= a2

gives

represents therefore,
through the point A.

Vir...N/>

by 186,

(2.),

= Na,

or

VII" .

TP = Ta

the sphere with o for centre, which passes

Compare the Note

to

page 214.

SQUARE OF A YECTOE.

ART. 282.]
(5.)

The more general

327

equation,

VIII.

(/3

a)

(comp.* 186,

(4.),

which passes through the point B.

represents the sphere with A for centre,
(6.)

For example, the equation,

IX.

represents the sphere with

(7.)

(p

a)

(p

of = a\

for centre,

The equations (comp.

X.

(p

186,

which passes through the origin

XL

(p

/3)

(p

The

o.

a)

to the line OA;

represent, respectively, the plane through o, perpendicular

and the plane which perpendicularly bisects the line AB.
(8.)

(3.),

(6.), (7.)),

a)

(comp. 186,

distributive principle of vector-multiplication (280),

and the formula

279, III., enable us to establish generally (comp. 210, (9.)) the formula,

XII.

a)

(]3

the recent equations IX. and X.

IX
(9.)

The

2
.

2
.

/o

therefore be thus transformed

may

= 2SaP

2
2Sj3a + a

/3

and

Sap =

0.

equations,

XIII.

2
p +

XIY.

2
p +

0,

represent the spheres with o for centre, which have a and 1 for their respec
2
tive radii; so that this very simple formula, p + 1 = 0, is (comp. 186, (1.))

a form of the Equation of the Unit-Sphere (128), and

importance in the present Calculus.
(10.)

The

equation,

XV.
may

as such, of great

is,

z
.

2Sap +

0,

be transformed to the following,

XVI.

N0>

a)

= -

(p

a)

- a2 =

Na

or

XVF.
it

(p

a)

=
</(<>

*)

=
*/(<>

+ Na)

represents therefore a (real or imaginary) sphere, with

last radical
radius.
(if real) for

with this

* Compare

also the sub-articles to 273.

for centre,

and

ELEMENTS OF QUATERNIONS.

328

This sphere

(11.)
if this

than -

Na

but

The

(12.)

it

therefore necessarily real,

is

scalar constant,

i.

[III.

2, 3.

be a positive scalar; or

if c

2
though negative, be (algebraically) greater than a

c,

becomes imaginary,

if

Na

or

0.

<

radical plane of the two spheres,

XVII.

p~

2Sap +

- 2Sa +
p

Q,

0,

has for equation,

it is

the given vectors a, a and the given scalars

one or both of the spheres themselves be imaginary.

therefore always real,

be such, even

if

if

c,

(13.) The equation 281, XXIX., or XXX., of the Central Ellipsoid (or of
the ellipsoid with its centre taken for the origin of vectors), may now be still
further simplified,* as follows
:

XIX.
The

(14.)

XX.
whence

it is

+ pK ) =

(ip

i\

definition (278) gives also,

.

afta = a

XX

or

Ea

a =

Ea

Ea =

natural to write, t

XXI.

= 1

Ea

a = a-

so far anticipate here the general theory of powers of vectors, above

alluded to (277), as to use this last symbol to denote the quotient, of unity

if

we

divided by the vector a;

equation,

XXII. ..a.
It follows,

(15.)

XXIII.
If

(16.)

symbol a~

have

as to

so

l
,

identically,

=o-1 .o=

or for every vector, the

1.

by 258, VII., that

1
a- = -

Ua Ta
:

we had adopted

then the formula

interpretation for the

symbol

and

the equation

XXIV.
/3a.

XXIV.
XXIII.

/3a

a".

as a definition + of the

might have been used,

But we proceed

j3

as a formula of

to consider

an entirely

of
different method, of arriving at the same (or an equivalent) Interpretation
to
as
this latter
or of a
Product of Vectors, considered
equal

symbol

Binary

a Quaternion.
*

Compare the Note

to

page 241.

J Compare

t Compare the Note to page 293.

the Note to page 324.

AETS. 282-284.]

SECOND INTERPRETATION OF A PRODUCT.

SECTION

329

3.

of arriving at the same Interpretation, of a

Binary Product of Vectors.

On a Second Method
283. It cannot

fail to

have been observed by any attentive reader of the

Second Book, how close and intimate a connexion* has been found to exist,
between a Eight Quaternion (132), and its Index, or Index- Vector (133).
Thus, if v and / denote (as in 223, (1)., &c., any two right quaternions, and
if Iv,

Iv denote, as usual, their

I.

II.

L/ =

(v

v)

which

may

if

Iv,

= Iv

III. ..Iv :Iv =

to

we have

indices,

alre

ady seen that

and conversely

v,

Iv (206)

:v (193)

EL? =

IR.V (258, IX.).

284. It could not therefore have appeared strange,

establish this new formula of the same kind,
I.

Iv

Iv = v .v = v

to interpret the product of

if

we had proposed

to

v,

as a definition (supposing that the recent definition

whereby

be added the more recent formula,

IV.

us),

(133)

278 had not occurred to

any two indices of right quaternions, as

being equal to the product of those two quaternions themselves. And then, to
interpret the product |3a, of any two given vectors, taken in a given order, we
should only have had to conceive (as we always may) that the two proposed
factors, a and )3, are the indices of two right quaternions, v and v , and to
multiply these latter, in the same order.

For thus we should have been

led to

establish the formula,

II.

or

we should have

|3a

v v,

this slightly

if

a = 10,

and

]3

= Iv

more symbolical equation,

in which the symbols,

I- a

are understood to denote the

a

and

/3

and

I^jS,

two right quaternions, whereof the two

lines

are the indices.

*

HAMILTON

Compare the Note

ELEMENTS OF QUATKRNIONS.

to

page 175.

aU

ELEMENTS OF QUATERNIONS.

330

To

(1.)

now

establish

[HI.

r.

3, 4.

the substantial identity of these two

interpretations,

278 and 284, of a Unary product of vectors /3u,

notwithstanding the difference
of form of the definitional equations by which they have been expressed, we
have only to observe that it has been found, as a theorem
that
(194),

IV...

vv

= Iv
:I(l :v)=Iv :IRv;

but the definition (258) of Ea gave us the lately cited

equation, ~RIv = IHv
we have therefore, by the recent formula II., the

equation,

V.
as in 278, I.

.L/.L>

a and

]3 still

VI.

or

L/:EL>;

denoting any two

j3

.a =

vectors.

j3

The

Ea,

two interpretations

therefore coincide, at least in their results, although they have been obtained
by different processes, or suggestions, and are expressed by two different formula.

The result 279, II., respecting conjugate products of vectors, corre

thus
to the result 191, (2.), or to the first formula of 223,
sponds
(1.).
(3.) The two formulae of 279, (1.) and (2.), respecting the scalar and right
(2.J

parts of the product

223,

article,

The

(4.)

j3a,

answer

to the

two other formulae

of the

same sub-

respecting the corresponding parts of v v.

doubly distributive property (280), of vector-multiplication,

(1.),

this plan seen to be included in the corresponding but

is

on

more general property

(212), of multiplication of quaternions.

changing IVq, IY/, t, f, and S, to a, |3, a, b, and y, in those

formulae of Art. 208 which are previous to its sub-articles, we should obtain,

By

(5.)

with the recent definition (or interpretation)

II.

of ]3a, several of the con

from the former

sequences lately given (in sub-arts, to 281), as resulting

definition, 278, I.

Thus, the equations,

VI., VII., VIII., IX., X., XI., XII., XXII., and

of 281, correspond

to,

and may (with our

last definition)

XXIII.

be deduced from,

the formulae,
V., VI.,

VHL, XL,

XII., XXII., XX., XIV., and XVI., XVIII.

(Some of the consequences from the

of 208.

sub-articles to

208 have been

already considered, in 281, (!!.))

(6.)

The

geometrical properties of the line IY/3a, deduced from the first

definition (278) of ]3a in

281,

that line, from

its

tion

by

j3

to a

(3.)

and

(4.),

(namely, Hue positive rotation round

perpendicularity to their plane

and the

representa

the same line of the parallelogram under those two factors, regard being

had

to units of length

223,

by means

(4.),

and

of area,)

might

also

have been deduced from

of the second definition (284), of the same product, (3a.

EIGHT QUATERNION EQUAL TO ITS INDEX.

ARTS. 284-286.]

SECTION

On

331

4.

the Symbolical Identification of a Right Quaternion with its

own Index and on the Construction of a Product of two
Rectangular Lines, by a Third Line, Rectangular to both.
:

285. It has been seen, then, that the recent formula 284, II. or III.,
may replace the formula 278, I., as a second definition of a product of two

same consequences, and therefore ultimately to

the same interpretation of such a product, as the first. Now, in the second
formula, we have interpreted that product, j3a, by changing the two factor-lines,
vectors,

a and

which conducts

]3,

to the

to the two right quaternions, v

are the indices

and

f
,

or I~ a a

and

which they

I^jS, of

and by then defining that the sought product j3a

is

equal to

two right quaternions. It becomes, therefore, im

this
at
to
portant
stage, how far such substitution, of I~ a for a, or
inquire,
of v for Iv, together with the converse substitution, is permitted in this
the product v

v,

of those

Calculus, consistently with principles already established.

For

it is

evident

such substitutions can be shown to be generally legitimate, or allow

\ve
shall thereby be enabled to enlarge greatly the existing field of
able,
that

if

interpretation

and

to treat, in all cases, Functions of Vectors, as being, at the

same time, Functions of Right Quaternions.

286. We have first, by 133 (compare 283,
I.

In the next

I- jS

TX

by 206 (comp. 283,

place,
or subtraction,

if

II.),

1.),

J3

the equality,
a.

we have the formula

of addition

with these more general results of the same kind (comp. 207 and 99),
III.

In the third

place,

I-

Sa = Sl- a

IV.

by 193 (comp. 283,

.. I-

III.),

we

have, for division, the

formula,

Y. ..I- /3:I- a = j3:a;

1

while the second definition (284) of multiplication of vectors, which has been
proved to be consistent with the first definition (278), has given us the

analogous equation,

VI.

I- /3

I- a

|3

a =

j3a.

2U2

ELEMENTS OF QUATERNIONS.

332
It

would seem, then, that we might

VII.
or

still

more

operation I

..T a
1

Quaternion, that the

VIII.

or

and symbolically,

briefly

is

= a;

4.

i.

at once proceed to define, for the

purpose

of interpreting any proposed Function of Vectors as a

following general Equation exists

[III.

if it

I* =

v,

if^ = ^;

be understood that the

subject of the

always a right quaternion,

IX.

I =

1.

But, before finally adopting this conclusion, there is a case (or rather a class
of cases), which it is necessary to examine, in order to be certain that no
contradiction to former results can ever be thereby caused.

287. The most general form of a vector function, or of a vector regarded as

a function of other vectors and of scalars, which was considered in the First

Book, was the form

comp. 275),

(99,

I.

and we have seen that

if

we change,

sponding right quaternion I

a,

= 2#a

in this

form, each vector a to the corre

and then take the index

of the

new

right

quaternion which results, we shall thus be conducted to precisely the same

or in symbols,
vector p, as that which had been otherwise obtained before
;

that
II.

But

2#a = IStfl^a (comp. 286, IV.).

another form of a vector-function has been considered in the Second

Book;

namely, the form,

in which a,

j3,

7, S,

are any odd number ofcomplanar vectors.

And

before

we accept, as general, the equation VII. or VIII. or IX. of 286, we must

cominquire whether we are at liberty to write, under the same conditions of
planarity,

and with the same

IV..
288.

To examine

signification of the vector

/o,

the equation,

.-!(.. .g|.j^. I-,}

this, let there

be at

first

only three given complanar

in which case there will always be (by 226) & fourth vector p,
vectors, y ||| a, ]3
in the same plane, which will represent or construct the function (y /3) a ;
;

namely, the fourth proportional to

/3,

7, a.

Taking then what we may

call

PRODUCT OF TWO RECTANGULAR LINES A LINE.

AUTS. 286-290.]

the Inverse Index-Functions, or operating on these four vectors

characteristic

may

be denoted by

v",

..J":v

VI.

or

= (p:a =

..*/"=(/

VTT

so easy to

)*;

~~~
_ P _ 7 _

I"V

.....
it is

:*

y:p=)v":ifi

Or, more symbolically,

which proves what was required.

number

p by the

v"

Y.

And

a, /3, y,

obtain four collinear and right quaternions (209), which

f
f
and we shall have the equation,
;
v, v

we

1
I"

333

I"

.
>

>.

extend this reasoning to the case of any greater odd

of given vectors in one plane, that

we may now

consider the recent

formula IY. as proved.

We

289.

shall therefore adopt, as general, the symbolical equations

VIII. IX. of 286

and

"VII.

shall thus be enabled, in a shortly subsequent section,

to interpret ternary (and other) products of vectors, as well v& powers

Functions of Vectors, as being generally Quaternions

and other

although they may, in

may become right quaternions

may, in virtue of the same principle, be
their own indices (133), and so be treated as

particular cases, degenerate (131) into scalars, or

(132)

iu which latter event they

represented by,
vectors.

and equated

In symbols, we

to,

shall write generally, for

any

of vectors

set

a, |3, y,

and any function f, the equation,

I.

q being
right, or

./(a,

some quaternion

j3,

7,

.)

=/(!-

<,,

I-J3, I-y, ..-)

while in the particular case

when

<?-

this quaternion is

when
=

we

shall write also,

II.

= S- 1

= TV,

and usually by preference

./(, ft

y,

.)

- I/XI-a,

(for that case), the

I-

/3,

r-7,

=
.)

formula,

p,

p being a vector.
290. For example, instead of saying (as in
281) that the Product of any
two Rectangular Vectors is a Right Quaternion, with certain
properties of its
Index, already pointed out (284, (6.) ), we may now say that such & product is
And hence will follow the important consequence, that
equal to that index.

ELEMENTS OF QUATERNIONS.

334

Product of any two Rectangular Lines

the

in

structed by) a Third Line, rectangular to both

Line,

is

Space

equal to (or

the Rotation round

the Multiplier-Line to the Multiplicand- Line,

from

[III.

may

this

4, 5.

i.

be con

Product-

being Positive

and

the Length of the Product being equal to the Product of the Lengths of the
Factors, or representing (with a suitable reference to units) the Area of the

And

generally we may now, for all purposes

calculation and expression, identify* a Right Quaternion with its own Index.
Rectangle under them.

SECTION

On some
from

291.
that

5.

Simplifications of Notation, or of Expression, resulting

this Identification; and on the conception of an Unitas a Right \ ersor.

An

immediate consequence of the symbolical equation 286, IX.,

we may now

is

suppress the Characteristic I, of the Index oj a Right Quater

formula into which

nion, in all the

it

has entered

and

so

may

simplify the

Thus, instead of writing,

Notation.

Ax
Ax

or

we may now

= ITJY?,

or

Ax. = IUY,

as in 204, (23.),

= UIY?,

or

Ax. = UIY,

as in 274, (7.),

write simply t,
I.

Ax

q =

The Characteristic Ax., of

(132, (6.)

of

),

may

UY?;

the

or

II.

Ax. =

UY.

Operation of talcing the Axis of a Quaternion

therefore henceforth be replaced

whenever we

may

think

fit

to

and Y, which
dispense with it, by this combination of two other characteristics,
are of greater and more general utility, and indeed cannot^ be dispensed with,
in the practice of the
present Calculus.
*

Compare the Notes to pages 121, 137, 175, 193, 203.

t Compare the first Note to page 120, and the third Note to page 203.
J Of course, any one who chooses may invent new symbols, to denote the same operations on
quaternions, as those which are denoted in these Elements, and in the elsewhere cited Lectures, hy the
letters U and V
but, under some form, such symbols must he used and it appears to have been
hitherto thought expedient, by other writers, not hastily to innovate on notations which have been
As
already employed in several published researches, and have been found to answer their purpose.
to the type used for these, and for the analogous characteristics K, S, T, that must evidently be a
:

mere affair of taste and convenience and in

some examination-papers by the author.
:

in

fact

they have

all

been printed as small

italic capitals,

CONCEPTION OF UNIT-LINE AS A EIGHT VEESOE.

AETS. 290-293.]

335

We are now enabled

also to dimmish, to some extent, the number

of
been
have
which
technical terms,
employed in the foregoing Book. Thus,
whereas we defined, in 202, that the right quaternion V# was the Eight Part
of the Quaternion q, or of the sum Sq + V?, we may now, by 290, identify

292.

that part with

its

part, or simply the

own

index-vector

TVq, and so

be led to

may

VECTOR,* of that Quaternion

q,

call it the vector

without henceforth speaking

although the plan of exposition, adopted in the Second Book,

do so for some time. And thus an enunciation, which
we
should
that
required
was put forward at an early stage of the present work, namely, at the end of
of the right part

the First Chapter of the First Book, or the assertion (17) that
Scalar plus Vector equals

"

becomes entirely

For we are in

intelligible,

manner

this

as being added to a Line,-f

and acquires a perfectly

led to conceive a

when

it

Number

definite signification.

(positive or negative)

added (according to rules already estab

is

which the

lished) to that right quotient (132), of

we

Quaternion"

line is

In symbols,

the Index.

are thus led to establish the formula,

I.

= a +

a,

when

whatever scalar, and whatever vector,

because either of

we

are

entitled

these
to

two parts,

say, that

both

II.

may

= a + I-1 a

be denoted by a and

or summands,
Scalars and

Lines, are included in the Conception of a

may

vanish

or

Vectors,

Quaternion, as

a.

And

separately,

Numbers and

now

enlarged or

modified.

293. Again, the same symbolical identification of Iv with v (286, VIII.)

leads to the forming of a new conception of an Unit-Line, or Unit- Vector (129),
as being also a Right Versor (153)

turn a

line,

rotation

or an Operator, of which the

effect is

to

in a plane perpendicular to itself, through a positive quadrant of

and thereby

to oblige the Operand-Line to take a

new

direction, at

And then
right angles to its old direction, but without any change of length.
the remarks (154) on the equatiou q* = - 1, where q was a right versor in the
former sense (which is still a permitted one) of its being a right radial quotient
*

Compare the Note

On

to page 193.

account of this possibility of conceiving a quaternion to be the sum of a number and a line,
it was at one time suggested
by the present author, that a Quaternion might also be called a Grammarithm, by a combination of the two Greek words 7/30^^ and apidpo s, which signify respectively a
Line and a Number.
t

ELEMENTS OF QUATERNIONS.

336

[III.

two equally long but mutually rectangular

(147), or the quotient of
immediately applicable to the interpretation of the equation,

where p

is still

an

vector p,

we have

lines,

become

+l = 0(282, XIY.);

unit-vector.

Thus (comp.

(1.)

or

p --l,

5, 6.

i.

fig.

41, p. 132), if a be

any

line perpendicular to such a

the equations,

L..

2
II...p a = pP = a = -a;

oa = /3;

being another line perpendicular to p, which is, at the same time, at right
third line a,
angles to a, and of the same length with it and from which a

j3

or -

a, opposite to the line a, but

still

equally long,

by (what we may here

the operation, denoted

formed by a

is

repetition of

the characteristic p

call)

or

having that unit-vector p for the operator, or instrument employed, as a sort of

handle, or axis* of rotation.

More generally (comp.

(2.)

or constructs, or

is

if

a, j3,

y be any three

if

and

equal to,

order of the factors (comp. 279) be observed: so that

tion (comp. 281,

XXI.),

III...aj3 = 7,

if

lines at right

the length of y be numerically equal to the

then (by what precedes) the line y represents,
/3,
tliQ product of the two other lines, at least if a certain

angles to each other, and

product of the lengths of a

290),

IV...j3_La, 7 J_a, y

J.

we may

and

]3,

V.

write the equa

Ta

T/3

= Ty,

provided that the rotation round a, from j3 to y, or that round y from a to

&c., has the direction taken as the positive one.

In

(3.)
line

more general

this

we may

)3 is

still

]3,

now by an operation

product-line y; but not

of

case,

a has operated on the multiplicand-line

(generally) multiplied

by

/3,

conceive that the multiplier-

so as to produce (or generate] the

of version alone, since the tensor

that of a, in order to form,

by V., the

tensor

of the product y.
(4.)

And

then to a)

we

if

)3,

fig.

repeat this

(comp. 189), or
plane of

(comp.
if

we

41,

bis,

compound

in

which a was

first

operation, of tension

multiply again by

a,

we

obtain a fourth

y, but with a direction opposite to that of

generally different

VI.

namely

.y = aa/3 = a j3=/3
*

/3,

in the

and with a

length

the line,
2

changed to )3, and

and version combined

=
-<rj3,

Compare the second Note

if

to page 137.

a = Ta.

line j3

TEENAEY PEODUCT OF VECTOES.

ARTS. 293, 294.]

The

(5.)

negative scalar,

or aa, is therefore equivalent, in its effect on |3, to the

,
2
or - (Ta) , or - Na, considered as a coefficient, or as a

operator a

-a

(scalar) multiplier (15)

whence the

equation,
,

may
as

now with

be again deduced, but

we

337

I.),

a new interpretation, which

however,

is,

see, completely consistent, in all its consequences, with the one

first

pro

posed (282).

SECTION

On

6.

the Interpretation of a Product of Three or

as a Quaternion.
294. There

now no

is

more

Vector*,

difficulty in interpreting a ternary product of vectors

some

(comp. 277, I.), or a product of more vectors than three, taken always in

namely, as the result (289, I.) of the substitution of the corre

sponding right quaternions in that product which result is generally what we
have lately called (276) an Oblique Quotient, or a Quaternion with either an

given order

acute or

an obtuse angle (130)

become

itself

but

Quaternions (223), in which indeed

Operation

three vectors, a,

(1.)

/3,

III.

Y7 |3a is

see that it is included, is

III.

[On account

an Asso

II.), for

any

and IV., are now replaced by the following

.V. 7 Y0a = aSj3 7 - j3S 7 a

.

V7 j3a

written, for simplicity, instead of

of the importance of these formulae,

it is

to,

*
;

V( 7 |3a),

or V. 7 j3a;

and

a student of this Calculus

worth while

to notice that, using the

principles of the present Book,

V 7 0a =
+
HAMILTON

(1

- K) 7 j8a = (? +
()8a +
J
0y) =
= 7 So0 - 0870 + 08/87.]
jSy)

(7)8

oj8)

- J (ya +

ay)

-4-

ELEMENTS OF QUATERNIONS.

= aS|3 7 /3Sya + 7 Saj3

with which, as with the earlier equations referred

will find it useful to render himself very familiar.
*

may

7, the Formula,

The formulae 223,

which

we now

we may write generally (comp. 223,

or that

II.

in

degenerate (131) into a scalar, or

a right quaternion (132), and so be constructed (289, II.) by a

It follows (comp. 281), that Multiplication of Vectors, like that of

neic vector.

ciative

may

aX

ELEMENTS OF QUATERNIONS.

338

(2.)

Another useful form

(3.)

of the equation II. is the following

The equations IX. X. XIV.

formula

V.

Sy/3a

6.

i.

..VVa3.

IV.

three vectors, the

[III.

223 enahle us now

of

to write, for

any

=-

Saj3 7

Say|3

=-

SjSya

=-

S/3ay

volume of parallelepiped under

6 x volume of pyramid OABC

Sya/3

a, /3, y,

upper or lower signs being taken, according as the rotation round a from
to

positive or negative

is

/3

or in other words, the scalar Sy/3a, of the ternary

product of vectors yj3a, being positive in the

first

but negative in the

case,

second.
(4.) The condition of complanarity of three vectors,
expressed by the equation (comp. 223, XI.)

a,

|3,

y, is therefore

VI.
(5.)

If a,

/3,

gives

SyjSa

VI

or

y be any

three vectors,

VIII.

SyS =

Sa|3 T

=
;

&o.

complanar or diplanar, the expression,

IX.

and

0,

Saj3S

represents therefore (comp. II. and IV.) a fourth vector

or in symbols,
dicular to y, but complanar with a and j3
it

8,

which

is

perpen

X.

SJ_ 7

XI...

and

(Compare the notations 123, 129.)

(6.) For any four vectors, we have by

and each
is"

XII.

XIII.

V (Va/3
V (Vaj3

at once

complanar with

a, j3,

any

XIV.
is

and IV. the transformations,

VyS) = SSajSy ySajSS

VyS)

= aS/3yS - jSSayS

found

to

and with

y, 8

or a Hue OE, which

e,

is

which
in the

OAB and OCD.

Comparing them, we

as a linear function of

which

a, /3.

of these three equivalent expressions represents & fifth vector

intersection of the two planes,

(7.)

II.

HI

see that

any

three given diplanar vectors, a,

pSajSy

= aS/3 7/o +

be one of extensive

may

arbitrary vector p

j3Sya/o

utility.

j3,

y,

ySa/3/o

by

be expressed
the formula

ANALYSIS OF A VECTOR.

AKT. 294.]

Another very useful formula,

(8.)

XY.

same kind,

of the

= V/3 7 Sap + Vya

pSafiy

339

is

the following

+ Va/3 Sjp
.

S/3/>

member of which, the points may be omitted.*

One
mode
of proving the correctness of this last formula XV., is
(9.)
operate on both members of it, by the three symbols, or characteristics of

in the second

to

operation,

XVI... S.
the

common

results

a,

S.y;

S.j3,

on both sides being respectively the three scalar products,

XVII. ..Sap.
where again the points

Saj3 7 ,

S/3/o

Syp

Sa/3 7 ,

Saj3 7

be omitted.

may

We

here employ the principle, that if the three vectors a, j3, 7 be

(10.)
actual and diplanar, then no actual vector \ can satisfy at once the three scalar
equations,

XYIII.
because

it cat t not be

SX = 0,

S|3A

0,

S 7 A = 0;

perpendicular at once to those three diplanar vectors.

any investigation with quaternions, we meet a system

form XYIII., we can at once infer that

(11.) If, then, in

of this

XIX.. .A =
while, conversely,
vectors, or Saj3 7
(12.)

Hence

if

be

0, as in

also,

0,

an actual

VI

if

XX...Sa/3?^0;

vector, then a,

j3,

7 must be complanar

under the same condition XX., the three scalar equa

tions,

gve

xxn...x= M

Operating (comp. (9.)) on the equation XV. by the symbol, or

S S, in which S is any new vector, we find a result which may
be written thus (with or without the points)
(13.)

characteristic,

XXIII.
where
*

a, /3,

S,

Sap

may

Sj3 7 S

S/Bjo

denote any five

S 7 Sa

4-

S 7jo Sga|3 - SSp S]3 7

.

vectors.

[Another method of proving XIV. is to assume p = xa + y/3 + zy. Operating by S .VjSy, S&yp
and similar expressions
may be found for y and z. To prove XV. assume p =
z Voj8, and
operate in turn by S a, S j8, and 8.7],
;

2X2

ELEMENTS OF QUATERNIONS.

340

In drawing

(14.)

when

holds good, even

must be

we assume

this last inference,

the three vectors,

a, |3,

We

(15.)

y be

ever so

little

has been proved, by

out of the plane of a and )3.

have therefore this new formula

which p

denote any fourth

may

whether

vector,

and

(9.)

fact

(12.),

if

0,

Plane of

XV.

which in

6.

i.

that the equation

y are complanar

true, as a limit, since the equation

to be valid, if

in

[III.

in,

Sa/3 7

or out

the common

of,

a, /3, y.

If p be perpendicular

(16.)

each term of the

change p

first

the last formula

to that plane,

is

evidently true,

member vanishing

to a vector 8 in the plane of a,

and if we
separately, by 281, (7.)
we
are
to
the
follow
conducted
)3, y,
;

ing equation, as an interpretation of the same formula XXIV., which expresses

a known theorem of plane trigonometry, including several others under it
:

XXV.

sin

BOG cos AOD + sin COA cos BOD + sin AOB cos COD =

for any four complanar

and

co-initial lines, OA, OB,

oc OD.
f

passing from OD to a line perpendicular thereto, but in their

plane, we have this other known* equation:

common

XXVI.

sin

BOG sin AOD + sin COA sin BOD + sin AOB sin COD =

which, like the former, admits of

here as offering

XXIV.

(18.)
this

0,

By

(17.)

formula

many

transformations, but

Operating on that formula by S

only mentioned

is

itself naturally to our notice, when we seek to interpret the

obtained as above by quaternions.

new equation

XXVII.

8,

and changing p

to

f,

we have

- SacS/ByS + SjSeSyaS + SyaSa/33,

if

Saj3y

which might indeed have been at once deduced from XXIII.

(19.)

when

The equation XIV.,

a, j3,

y are complanar

XXVIII.

aS/3y/>

as well as

XV., must hold good

hence

+ pSyap + ySafip =

Compare page 20

0,

if

of the Geometric Superieure of

Sa/3y
M.

0.

Chasles.

at the limit,

COMPLANAR VECTORS.

ABT. 294.]
(20.)

This

formula

last

is

evidently true, by

plane of the three other vectors

and

341

(4.), if

we suppose

if

p be in the

common

to be perpendicular to

it

that plane, so that

XXIX.
and

by 281,

therefore,

Y7 a

||

since

S|3 7/o

Vaft

||

(Sj3y

= S (Yj3y

p)

0.

= Y/3
7 P &c.,

p)

by p, and so obtain this other formula,

divide each term

XXXI.

Y/3 7

||

(9.),

XXX.
we may

+ 7 Ya/3 =

aY/3 7 + /3Y 7

if

0,

Saj3y

0.

(21.) In general, the vector (292) of this last expression vanishes

the expression is therefore equal to its own scalar, and we may write,

XXXII.
whatever three vectors
(22.)

For the

equally long,

may

be denoted by

II.

the three vectors are

the proportion,

XXXIII.

Yj3 7

XXXI.

XXXIY.
where OA, OB, oc are

a, |3,

we suppose that

case of complanarity, if

we have

and the formula

= 3Saj3 7
Y]3 7 + j3Y 7 a + 7 Yj3

by

Y7 a

Yaj3

sin BOG

sin COA

sin

AOB

becomes thus,
OA

OB

sin BOC +

COA + oc

sin

ant/ three radii of one circle,

sin

AOB =

and the equation

is

interpreted

as in Articles 10, 11, &c.

The equation XXIII. might have been deduced from XIY.,

of XY., by first operating with S 8, and then interchanging S and p.

(23.)

instead

(24.)

vector p

may

in general be considered (221) as depending*pn three

scalars (the co-ordinates of its term)

than three scalar equations

(25.)

any

As an example

three given

a, b, c

vector p has for

XXXYI.

it

it cannot then be determined by fewer

be eliminated between fewer than four.

of such determination of a vector, let a,

and diplanar

XXXY.
in which

nor can

vectors

Sap =

and
a,

let

SjS/o

b,

S 7f =
>

are supposed to denote three given scalars.

its

expression,
e-

(ffY/3 7

/3,

7 be again

the three given equations be,

;

Then

the sought

by XY.,

+ &Y 7 a +

cYa/3),

if

XXXYII.

= Sa)3 7

ELEMENTS OF QUATERNIONS.

342
(26.)

As another example,

XXXVIII.
same

then, with the

As an example

equations,

XL.
then,

Sap

(28.)

This

last

6.

Syap =

e~

(a a

+ b

e,

fi

we

Safip

XIV.,

have, by

c y).

of elimination of a vector, let there be the four scalar

vectors, a

XLI.

by XXIII., we have

but only the four

= a

Sfiyp

i.

the three equations be,

let

signification of the scalar

XXXIX.
(27.)

[III.

a,

Sj3/>

c,

this resulting equation, into which

.

and the four

8,

Sj3yS

equation

Syp =

b,

may

SySa +

scalars, a

p does not

- d.
=
SajSy

SSa/3

enter,

0.

therefore be considered as the condition of con

by the four scalar equations XL., in one

has not been expressly stated before, it follows

currence of the four planes, represented

common point

for,

although

it

evidently from the definition 278 of a binary product o/ tutors, jOombinedJwith

196, (5.), that every scalar equation of the linear form (comp. 282, XVIII.),

XLII.
in which a = OA,

and p =

vector of the foot

s,

SajO

a,

or

Spa

a,

OP, as usual, represents a plane locus of the point p

the

of the perpendicular on that plane from the origin, being

XLIII.
(29.) If

os =

o-

= aRa = aa

we conceive a pyramidal volume

(282,

(68) as

XXI.).

having an algebraical

(or

be capable of bearing either a positive or a negative

ratio to the volume of a given pyramid, with a given order of its points, we may

scalar) character, so as to

then omit the ambiguous

ternary product of vectors

sign, in the last expression (3.) for the scalar of a

:

and

so

may

write, generally,

OABC denoting such a

volume, the formula,

XLIV.
= a
oc

positive or

is

Sa|3y

= 6 OABC,
.

a negative scalar, according as the rotation round OA from OB to

negative or positive.

(30.)

More

generally, changing o to D,

thus the formula

XLV.
in which

it

and OA or a

to a

8,

&c.,

we have

may

be observed that the expression

is

changed

to its otvn opposite,

ELIMINATION OF A VECTOR.

ART. 294.]

or negative, or
or

when any

and

is

multiplied by

1,

when any two of

two of the four points, A, B,

c,

the

343

four

vectors, a, )3, 7, S,

D, change places with each other

former value, by a second such binary

therefore is restored to its

inter

change.
(31.)

Denoting then the

XLIY., XLY.,

XLYII.

by

we have

E,

first,

by

result (comp. 68)

&c.)

under the more symmetric form

BCDE + CDEA + DEAB + EABC + ABCD =

E may denote any jive points of space.

an analogous formula (69, III.) of the First Book, for any

A, B, c, D,

And

(32.)

7, 8

DABC = EABC - EBCD + ECDA - EDAB

and may then write the

= BECD =
(because^- EBCD

which

|3,

the equation,

XLYI.

in

we?0 cngraVi of a,

six

points OABCDE, namely the equation (comp. 65, 70),

XLYIII.

OA BCDE -f OB CDEA
.

-f

OC

DEAB + OD EABC
.

-f

OE

ABCD =

0,

in which the additions are performed according to the rules of vectors, the

volumes being treated as scalar

going principles and

results.

coefficients, is

In

from the fore

easily recovered

by XLYII.,

fact,

this last

formula

may

be

written as

XLIX.

ED EABC = EA EBCD + EB

or, substituting a, /3, 7, 8 for

L.

which

is

ECAD + EC EABD
.

EA, EB, EC, ED, as

only another form of XIY., and ought to be familiar to the

student.

The formula 69, II. may be deduced from XXXI. by observing

when the three vectors a, /3, 7 are complanar, we have the proportion,

(33.)

that,

LI.
if signs (or

Yj37

Y7a

Ya]3

Y ()3y + ya + a|3) = OBC

OCA

OAB

algebraic or scalar ratios) of areas be attended to (28, 63)

formula 69, L, for the case of three

written as follows

LII.

collinear points A, B, c,

ABC,

aQ3- 7

+ ]3(7-a) +

7(-j3)

the three coinitial vectors a,

]3,

7 be

=2

2Y(j3-a)( 7 -a) =0,

termino-collinear (24) .

and the

may now

=
if

be

ELEMENTS OF QUATERNIONS.

344

[III.

6.

i.

The case when four coinitial vectors a, j3, y, 8 are termino-complanar

(64), or when they terminate in four complanar points A, B, c, D, is expressed
by equating to zero the second or the third member of the formula XLV.*
(34.)

Finally, for ternary products

(35.)

formula

of vectors in general,

we have

the

LIIL

2
c.

/3V

+ (Sa/3y) 2 = (VajSy) 2 = (aS/3y - /3Sya + ySa/3) 2

2
= a2
(Sj3y) +

The

295.

j3

2
2
2 (Sya) + 7 (Sa/3)
2S/3ySyaSa/3.f

identity (290) of a right quaternion

with

important versors,

i,

identical with, their

j, k, as constructed by,

own

and even

its index,

now

ception (293) of an unit-line as a right versor, allow us

and the con

to treat the three

as (in our present view)

OK

axes; or with the three lines 01, oj,

of 181, con

sidered as being each a certain instrument, or operator, or agent in a right

which causes any line, in a plane perpendicular to itself,

),
turn in that plane, through & positive quadrant, without any change of its

rotation (293, (1.)

to

this conception, or construction, the

"With

length.

are

still

Laws of

the

Symbols

ijk

included in the Fundamental Formula of 183, namely,

,-/ = # = #fc .

-1;

(A)

we now, in conformity with the same conception, transfer the Standard

Trinomial Form (221) from Right Quaternions to Vectors, so as to write

and

if

generally an expression of the form,

I.

ix

+ jy +

where xyz and abc are

kz,

or

a = ia

+jb

+ kc, &c.,

(namely, rectangular co-ordinates), we can

recover many of the
foregoing results with ease and can, if we think fit,
connect them with co-ordinates.
scalars

(1.)

P =-

As

to the laws (182), included in the

1, &c.,

reversal
(2.)

may

which

The

Fundamental Formula A, the law

be interpreted on the plan of 293,

results

from two

representing the

(1.), as

successive quadrantal rotations.

contrasted laws, or formulae,

ttvo

ij

= +

k,

ji

= -

(182, II.

k,

and

III.)

may now

be interpreted as
expressing, that although a positive rotation through
a right angle, round the line i as an
axis, brings a revolving line from the position

j to the position

round

k,

the line j, as

* [And

or +

k, yet,

a new

axis,

on the contrary, a

the equation of the plane ABC

t [Since Kafiy

=-

yfta.]

positive quadrantal rotation

brings a neiv revolving

is

S/)V (0y

+ ya +

from a new

line

ctj8)

initial

STANDARD TRINOMIAL FORM FOR A VECTOR.

ARTS. 294, 295.]

position,

to a

i,

position, +
(3.)

that

we

new final position, denoted by -

345

or opposite* to the old final

k,

Jc.

= - 1 (183) may be interpreted by conceiving,

Finally, the law ijk
operate on a line a, whicli has at first the direction of + j, by the three

lines, A-,/,

in succession

i,

in the directions of

i,

which gives three newlout equally long lines, ]3, 7, 8,

+ k, - j, and so conducts at last to a line - a, which
;

has a direction opposite to the initial one.

of ijk, which are
(4.) The foregoing laws

all (as

has been said) included

Formula A, when combined with the recent expression

(184) in the

give (comp. 222,

(1.)

II.

for the square of that vector the value

2
.

(ix

|0

+ jt/ +

= -

Icz)

+ if + s 2 )

(x*

I.

for p,

this square of the line p is therefore equal to the negative of the square of its

length

(185), or to the negative of

Tp

the former resultf 282,

The

(5.)

or

(1.)

its

norm

Np

which agrees with

(273),

(2.).

condition of perpendicularity of the two lines p

and

a,

when they

by the two trinomials I. and I ., may be expressed

XVIII.) by the formula,
= Sap = III.
(ax + by +

are represented

cz)

(281,

which agrees with a well-known theorem of rectangular co-ordinates.

The

(6.)

condition of complanarity of three lines, p, p

f
,

represented by

the trinomial forms,

IV.
is

ix

+ jy +

kz,

ix

+ &c.,

ix"

+ &c.

p"

(by 294, VI.) expressed by the formula (comp. 223. XIII.),

V.

8p"p

*"

(sty

/*) +

,/

(x z

agreeing again with known results.

(7.) When the three lines p, p,

or OP, OP

recent expression for Sp

294,

p",

pp

gives,

by

(3.),

z x)

OP",

"

(y

x-

*fy)

are not in one plane, the

the volume of the parallelepiped

* Jn the
ceptions,

the west,

Lectures, the three rectangular unit-lines, i, j, Jc, were supposed (in order to fix the con
and with a reference to northern
latitudes) to be directed, respectively, towards the south,
and the zenith and then the contrast of the two
formulae, ij - + &, j i =
k, came to be
;

illustrated

by conceiving, that

round an axis

\ve at one

time turn a movcable

(or handle) directed to wards the south,

with

line,

which

is

at first directed westward,

right-handed (or screwing] motion, through

zriyht angle, which causes the line to take an upward position, as its final one and that at another
time we operate, in a
precisely similar manner, on a line directed at first southward, with an axis
directed to the west, which
obliges this new line to take finally a downward (instead of, as before, an
a.

upward]
t

direction.

Compare

HAMILTON

also 222,

IV.

ELEMENTS OF QUATKKNIONS.

ELEMENTS OF QUATERNIONS.

346

[III.

6.

i.

(comp. 223, (9.)) of which they are edges and this volume, thus expressed, is
a positive or a negative scalar, according as the rotation round p from p to p is
that is, according as it has the same direction as
itself positive or negative
;

that round + x from

+ y

to

z (or

round

from j

to k],

or the direction

opposite thereto.
(8.)

It

be noticed here (comp. 223, (13.)

may

and V.) we have

three vectors, then (by 294, III.

VI.

VII.

),

that

if

y he any

a, )3,

Saj3y

= - SyjSa = J

Va/3y

= + Vy|3a = \ (j3y + yj3).

(aj3y

7 /3a)

More generally (comp. 223, (12.) ), since a vector, considered as repre

a
right quaternion (290), is always (by 144) the opposite of its own
senting
conjugate, so that we have the important formula,*
(9.)

VIII.

we may write

Ko

= -

IX.

and therefore

o,

KOa

any number of vectors, the transformations,

for

X.
XT.

Sll a = J (Ila

SDa =

Vila = + Vll a = J

(Ila

Il a),

Il a),

upper or lower signs being taken, according as that number

being understood that

XII.
(10.)

The

Il a,

a =

7 /3a,

Ila = a/3y

if

is

even or odd

_L

Ax.

it

relations of rectangularity,

XIII.

Ax.

_L

Ax. /;

Ax. j

_L

Ax. k

which result at once from the definitions (181),

briefly, as follows

Ax. &

may now

be written more

jk;

XIV. ..tJ-y;
and similarly in other

cases,

where the

kLi;

axes, or the planes, of

any two right

quaternions are at right angles to each other.

But, with the notations of the Second Book, we might also have
written,
123, 181, such formulae of complanarity as the following, Ax./||| i,
(11.)

by

to express (comp. 225) that the axis of

might cause some confusion,

*

if

/ was a

we were now

line

and

it

formula to/|||

t.

in the plane of

to abridge that

manner, we interpret, on our present plan, the symbols Ua, Ta, No as equivalent to
TI^o, Nl^a, we are reconducted (compare the Notes to page 137) to the same signification of
those symbols as before (155, 185, 273) and it is evident that on the same plan we have now,
If, in like

UI-X

So =

0,

Vo =

o.

PRODUCT OF ANY NUMBER OF VECTORS.

ARTS. 295, 296.]

In general,
sign

|||,

it

seems convenient that we should not henceforth employ the

three
except as connecting either symbols of

complanar

linear (209),

347

lines,

considered

still

as

or else symbols of three right quaternions, considered as being colbecause their indices (or axes] are complanar or finally, any two
:

complanar quaternions (123).

will result,
(12.) On the other hand, no inconvenience

we now

if

insert

the sign of parallelism, between the symbols of two right quaternions which
for example, we may write, on our
are, in the former sense (123), complanar
;

present plan,

XV.

w\\j,

..**||,

*k\\k,

xyz be any three scalars.

296. There are a few particular but remarkable cases, of ternary and other
products of vectors, which it may be well to mention here, and of which some
if

be worth a student

may

while to remember

especially as regards the

products of successive sides of closed polygons, inscribed in

If A, B,

(1.)

c,

D be any four

articles to 260, that their

is

scalar

concircular points,

circles,

or in spheres.

we know, by

anharmonic function (ABCD), as

denned

the sub-

in 259, (9.),

being also positive or negative, according to a law of arrangement ot

those four points, which has been already stated.

But, by that definition, and by the scalar (though negative] character

(2.)

we have

of the square of a vector (282),

quadrilateral ABCD, the formula

I.

in

(ABCD)

which the

tive or of

= AB

coefficient

BC

260,

a positive scalar, namely the product of two nega

e* is

ABCD be a plane and

DA 3

>

0.

inscribed quadrilateral,

we

have, by

the formula,
III.

AB BC
.

CD

DA =

according as this quadrilateral in a

(4).

If then

(8.),

the continued product of the four sides

two positive squares, as follows

2
2
II.
6 = BC
DA 2 = BC 2
.

(3.)

CD DA =

generally, for any plane or gauche

The product

a crossed or an uncrossed one.

circle is

a]3y of any three complanar vectors is a vector, because

scalar part Safiy vanishes,

successive sides AB,

a positive or negative scalar,

BC,

by 294,

(3.)

and

(4.);

and

if

CD of a quadrilateral thus inscribed in a

product has either the direction of the fourth successive

side,

circle,

their

DA, or else the

opposite direction, or in symbols,

IV.

according as the quadrilateral

AB BC CD
.

ABCD

is

DA

>

or

<

its

the factors be three

0,

an uncrossed or a crossed one.

2

ELEMENTS OF QUATERNIONS.

348

(5.)

[III.

i.

6.

conceiving the fourth point D to approach, continuously and inde

point A, we find that the product of

By

finitely, to the first

the three successice sides of

by an equation

V.

AT being a

CA = AT;*

line (corap. fig. 63)

scribed circle, or (more fully)

ABC of that

circle,

initial direction

given

AB BC

triangle, ABC, is

any plane
form

of the

at

which touches the circum

which touches

the point A;

the segment

or represents the

of motion, along the circumference, from A through B

the length of this tangential product

line,

AT,

is

equal

to,

to

while

or represents, with the

usual reference to an unit of length, ihe product of the lengths of the three sides,

same inscribed triangle ABC.

(6.) Conversely, if this theorem respecting the product

of the

of the sides of

supposed to have been otherwise proved, and

since
then
it will give in like manner the equation,
remembered,

inscribed triangle be

VI.
if

D be any fourth point,

AC CD
.

theorem

it

be

DA = AU,

concircular with A, B,

c,

while AU

as in the

is,

annexed

new segment ACD, we can

figures 63, a tangent to the

recover easily the

if

an

respecting the product

(3.),

of the sides of an inscribed quadrilateral ;

and thence

can return to the corresponding theorem (260, (8.) ),

respecting the anharmonic function of any such figure

ABCD

for

we

shall thus

by V. and VI., the

have,

Fig. 63

bis.

equation,

VII.

AB

BC

CD

DA

AT

2
N.AC, or Ac

AU

CA

AC

always positive (282, (1.)), but

the dividend AT.AU is negative (281, (9.)) for the case of an uncrossed quadri
lateral (fig. 63), being on the contrary positive for the other case of a crossed
in which the divisor CA.AC or

one

(fig.

63,

is

bis).

circle through a given point A, which

line
OT = r, as represented in fig. 64, we
o
a
touches at a given origin
given
shall then have by (5.) an equation of the form,
(7.)

If P be

any point on the

VIII.

OA AP PO = X
.

[AT
Or

directly

CA

by Euclid u AB = u CB

or

OT,

AT
CA

=u

AB

"1

CB J

PRODUCTS OF SIDES OF INSCRIBED POLYGONS.

AET. 296.]
in

which x

some

is

which varies with the position of

scalar coefficient,

Making then OA =
shall have

and OP =

a,

IX.

a(p

349

p, as usual,

P.

we

= -XT,

a}p

or

- a = XT

1
p"

a p

2
,

or

Vrp- =
1

IX".

and any one

of these

may

of the equation of the

be considered as a form

circle,

determined by the

given conditions.
Geometrically, the last formula

(8.)

expresses, that the line p

IX".

- a

1
,

- Ea, or A P
=
= Ea = E OA, and
(see again fig. 64), if OA
OP = p = E OP, is parallel to the given tangent T at o which agrees with
fig. 58, and with Art. 260.
1

or

a"

E/>

If B be the point opposite to o

(9.)

or

]3,

as being J_

X.
in

upon the

l
r, so that rfi~ is a vector, is

or

circle,

then the diameter OB,

given by the formula,

.

|3

= -r

which the tangent

r admits, as it ought to do, of being multiplied by any

without the value of j3 being changed.
(10.) As another verification, the last formula gives,

scalar,

XL
(11.)

TVUref = OA
1

sin ACT.

If a quadrilateral OABC be not inscriptible in a circle, then whether

we can always

be plane or gauche,

and

OB = T/3 = Ta

circumscribe (as in

OBC, about the two triangles,

then, on the plan of

(6.),

fig.

65) two

circles,

formed by drawing the diagonal OB

we can draw two tangents

it

OAB

and

OT, ou, to the two segments OAB, OBC, so as to repre

sent the two ternary products,

OA AB
.

after

and

BO,

OB

BC

co

which we shall have the quaternary product,

XII.
where the

OA AB

divisor,

positive scalar,

os

2
,

BC

co = OT

or BO

ou OB 2

OB, or

but the dividend OT

also the quotient in the second

quaternion.

OB, is a

ou, and therefore

member, or the product in the first member,

is

ELEMENTS OF QUATERNIONS.

350
(12.)

The

two tangents

quaternion is perpendicular to the plane TOU of the

and therefore to the plane itself si the quadrilateral OABC, if that
axis of this

be a plane figure

but

if it

scribed sphere at the point o

round

it,

(13.)

from OT to ou,

The

6.

i.

[III.

be gauche, then the axis

being also in

normal

is

all cases such,

to the

circum

that the rotation

is positive.

angle of the

same quaternion

is

the supplement of the angle TOU

between the two tangents above mentioned it is therefore equal to the angle
U OT, if ou touch the new segment OCB, or proceed in a new and opposite direc
;

from o

tion

(see

the two arc8,

again

fig.

65)

it

may therefore

be said to be the angle between

OAB and OCB, along which a point should move, in order

on the two circumferences,

to

go from

o,

the opposite corner B of the quadrilateral OABC,

the
two
other
A
and c, respectively or the angle between the
corners,
through
to

arcs OCB, OAB.

(14.)

These

and

results, respecting the axis

successive sides, of

angle of the product of the four

any quadrilateral OABC, or ABCD, apply without any modifi

and
(9.) ) of the same quadrilateral

cation to the anharmonic quaternion (259,

the axis becomes indeter

although, for the case of a quadrilateral in a circle,

minate, because the quaternary product and the anharmonic function degene
rate together into scalars, or because the figure may then be conceived to be
inscribed in indefinitely

many

spheres t yet the angle

the same rule as in the general case

and uncrossed quadrilateral

one

(fig.

(fig.

63)

this angle

but =

may

still

being =

TT,

be determined by
for the inscribed

for the inscribed

0,

and

crossed

63, bis).

For the gauche quadrilateral OABC, which may always be conceived

we may say, by (13.), that the angle of
the quaternion product, L (OA
AB BC co), is equal to the angle of the lunule,
bounded (generally) by the two arcs of small circles OAB, OCB with the same
(15.)

to be inscribed in a determined sphere,

.

construction for the equal angle of the anharmonic,

L (OABC),

L (OA

or

AB BC
.

co).

evident that the general principle 223, (10.), of the permissibility

of cyclical permutation of quaternion factors under the sign S, must hold good
(16.)

It

for the case

still

is

when those quaternions

degenerate (294) into vectors

and

it is

more obvious, that every permutation of factors is allowed, under the

whence cyclical permutation is again allowed, under this other sign SU

sign T

and consequently

also (corup. 196,

XVI.) under the

sign

L.

PENTAGON IN A SPHERE.

ART. 296.]
(17.)

Hence

generally, for any four vectors,

XIII.

SajSyS

S/3 7 Sa

XY.
and in

XIV.

we have
.

the three equations,

particular, for the successive sides of

ABCD, we have

351

any plane

or gauche quadrilateral

the four equal angles,

XYI.

L (AB

BC

CD

DA)

= L

(BC

CD DA
.

= &o.

AB)

with the corresponding equality of the angles of the four anharmonies,

XVII.

L (ABCD) = L (BCDA) = L (CDAB) = L (DABC)

XVII. j,

or of those of the four reciprocal anharmonics (259,

XVII".

L (ADCB) = L (BADC) = L (CBAD) = L (DCBA).

(18.) Interpreting now, by (13.) and (15.), these last equations, we derive
from them the following theorem, for the plane, or for space
Let ABCD be any four points, connected
four circles, each passing through
:

"by

three of the points

then, not only

the angle at A, between

is

the arcs ABC, ADC, equal to the angle at

but also

it is

equal (comp.

fig.

c,

between CDA and CBA,

66) to the angle at B, between

the two other arcs BCD and BAD, and to the angle at D, between
the arcs DAB, DCB.

Again, let ABCDE be any pentagon, inscribed in a

and conceive that the two diagonals AC, AD are drawn.
shall then have three equations, of the forms,

(19.)

sphere

We

XVIII.

AB BC
.

CA = AT

AC CD DA = AU
.

where AT, AU, AV are three tangents

is

AB

BC

We

to the sphere at A, so that their

CD DE EA = (AT

AW

= a new

AU AV)
.

2
:

(AC

The product of

inscribed in a sphere,

pentagon begins and

product

vector, which touches the sphere at A.

the five successive sides of

is

AD 2 )

have therefore this Theorem, which includes several others under

"

the

AD DE EA = AV

But the equations XVIII. give

a fourth tangent at that point.

XIX.

Fi

it

any (generally gauche] pentagon

equal to a tangential vector, drawn from the point at which
ends."

ELEMENTS OF QUATERNIONS.

352

[III.

Let then p be a point on the sphere whicli passes through

through three given points A, B, c we shall have the equation,
(20.)

i.

o,

6.

and

XX.

= S (OA AB BC CP PO) = 8 Q3 - a) (7 - 0)
= rfSfiyp + fiSyap + y 2 Safto p SajSy.

vectors a,

co-initial

passes through their

XXI.

A\ B

point P,

equation

if

we

on the

point P,

7)

p)

see that the condition for the four

yfi

9
p =

then

37,

+ yfr + zf.

x<?

the variable point p

project (comp. 62)

three given chords OA, OB, oc,

into

points

through that
to
the
shall
have the
we
respectively parallel
planes BOG, COA, AOB,
three planes

by

XXII.
(23.)

(p

p thus terminating on one spheric surface, which

origin o, may be thus expressed

7,

|3,

common

p = xa +

If then

(22.)
v

Comparing with 294, XIV., we

(21.)

OP 2 = OA

OA"

+ OB OB + OC OC\
.

That the equation XX. does in fact represent a spheric locus for the
is evident from its mere/orw (comp. 282,
and that this sphere
(10.) )
;

four given points, o, A, B, c, may be proved by observing

that the equation is satisfied, when we change p to any one of the four
passes through the

vectors, 0, a,
(24.)

7.

/3,

Introducing an auxiliary

XXIII.
or

by the system

XXIV.

the equation

XX.

282,

is

IX .;

(25.)

/3

Vya + 7

Va/3,

a = SSo,

S&r =

j3
1

2
.

SS/3-

=SSr

SS/3,

= S

becomes simply,

XXV.

or

SSp,

S^

= 1

is

a diameter (comp.

(6.)).

The formula XXIII., which determines

way

XXVI.

XXVI

),

the point of the sphere opposite to o, and

and 196,

written in this oilier

or

= a 2 V/3 +
7

SSajSy

of the sphere

XXV.
so that

determined by the equation,

S,

of the three scalar equations (comp. 294, (25.)

XXIY.
or

OD or

vector,

this

diameter,

may

be

SSaj3 7

Va

(J3

OABC OD = .

where the symbol OABC, considered

a) (7

V (OA

|3)

AB

7;
BC

co)

as a coefficient, is interpreted as in 294,

FOBMS OF THE EQUATION OF A SPHERE.

AKT. 296.]

XLIY.

353

namely, as denoting the volume of the pyramid OABC, which

an inscribed one.
;

(26.)

This result of calculation, so far as

OA AB

axis of the quaternion

BC

is

here

regards the direction of the

co, agrees with, and may be used to confirm,

it

the theorem (12.), respecting the product of the successive sides of a gauche
quadrilateral, OABC
including the rule of rotation, which distinguishes that
;

axis

from
(27.)

its opposite.

The formula XXIII.

XXVII.

XX.

S<T

y-

/3-

V (/3-

and the equation

for the diameter 3

- a

of the sphere

XXVIII. ..0 = 8
which expresses (by 294,

(/3

Y Q3-

(y-

(y-

1
a"

+ a

1
a"

comp. 260,

(34.),

+ y-

be transformed to the following

may

a"

- a

also be thus written

may

(p

1
a"

(10.)), that

tbe/owr reciprocal

vectors,

XXIX.

OA = a =

a-

are termino-complanar (64)

being

parallel to the

let fall

on

this

OB = ft =

tangent plane

to the

=
</

OP = p =

which they

in

sphere at o

y-

all

p-

terminate,

because the perpendicular

8-1,

from the three scalar equations,

xxxi.
(28.)

oc =

is

XXX.
as appears

the plane A B C P

plane from o

J3

In general,

if

Sa s = s3 s = S s =

D be the

i.

foot of the perpendicular

from

o,

on the plane

ABC, then

XXXII.
because this expression

satisfies,

Sa/3y

and

V (|3y + ya + a/3)

may

be deduced from, the three equa

tions,

XXXIII.
As

Sa^ =

a verification, the formula shows that the length

or altitude, OD,

is

equal to the sextuple volume of the

the double area of the triangular base ABC.

T,

of this perpendicular,

pyramid OABC, divided by

(4.), and 294, (3.),

(Compare 281,

(33.).)

(29.) The equation XX., of the sphere OABC, might have been obtained by
the elimination of the vector S, between the four scalar
equations XXIV. and
XXV., on the plan of 294, (27.).
HAMILTON

ELEMENTS OF QUATER MOXS.

zZ

ELEMENTS OF QUATERNIONS.

354
(30.)

And

[III.

6.

another form of equation of the same sphere, answering to the

XXVIII., may be obtained by the analogous elimination of

development of

the same vector

(31.)

between the four other equations, XXIV. and XXV.

of any even number of complanar vectors is generally a

S,

The product

quaternion with an axis perpendicular

plane but the product of the

hexagon ABCDEF, or any other even-sided figure, inscribed in a
a scalar because by drawing diagonals AC, AD, AE from the first (or last)
to

their

successive sides of a
circle, is

point A of the polygon, we find as in (6.) that it differs only by a scalar coefficient,
or divisor, from the product of an even number of tangents, at the first point.

On

(32.)

the other hand, the product of any odd number of complanar

same plane and in particular (comp. (19.) ),

the product of the successive sides of a pentagon, or heptagon, &c., inscribed in a
circle, is equal to a tangential vector, drawn from the first point of that inscribed

vectors is always a line, in the

and odd-sided polygon because it

product of an odd number of such

differs

only by a scalar coefficient from the

tangents.

(33.) The product of any number of lines in space is generally a

and if they be the successive sides of a hexagon, or other
(289)

quaternion

polygon, inscribed in a sphere, the axis of this quaternion (comp.

normal

even-sided
(12.)

is

to that sphere, at the initial (or final) point of the

polygon.
of
successive
sides
of
a
the
the
heptagon, or other odd(34.)
product
sided polygon in a sphere, is equal (comp. (19.) ) to a vector, which touches the
sphere at the initial or final point ; because it bears a scalar ratio to the

But

product of an odd number of vectors, in the tangent plane at that point.*

called
(35.) The equation XX., or its transformation XXVIII., may be
the condition or equation of homosphericity (comp. 260, (10.)) of the five points
and the analogous equation for the five points ABODE, with
o, A, B, c, P
;

from any arbitrary origin o, may be written thus

= S (a XXXIV.
e) (e
a)
8) (8
j3) (/3
7) (7

vectors ajSySe

XXXV.

or thus

ac? + b$*

+ cy + d& +

ee ,f

*
p
[The inscription of polygons in a sphere is treated very fully in the Lectures." If pi, pi,
= pa ps, &c. denote the vector
are the vectors from the centre to the vertices, and if n = pz
pi, t-z
"

sides,

where

then by 213
q

when n

=
is

n t n _i

p2

(5.)
.

t2 tl

= -npiir
Hence

p\q

pa
n

l
1 =
l
and p + i = pi =
0Pi rS
12 n pi n~ u~
when n is even piVg V# pi or V# pi but
and
affords the conditions
Sg =
piSq + SpiV# =

=-

(-)

odd the quaternion equation

qpi

izpz if

(-)"

or

||

SpiV? = 0, or q is a vector at right angles to pi. See Lecture VI., Art. 336.]
t [On change of origin XXI. may be written in the form
(a-e) + i(j3-)

<!(

7-

+ rf(8_

6)

0,

a (a-

3
e)

+ b(B -

2
)

+ c(y-

e)

2
<*(5-

e)

= 0.

Introducing e defined by XXXIX., XXXV. and XXXVII. follow. Eliminating a, b, c, d, and e

to
five equations connecting the
squares of the mutual distances between the points, analogous
that here given, a determinant relation is at once
found.]

from

EQUATION OF HOMOSPHEEICITY.

ART. 296.]
six times the second

member

the second

XXXVI.
or

more

member

a = BODE,

found to be equal to

of this last formula being

of the one preceding

= CDEA,

355

it, if

= DEAB,

d = EABC,

- ABCD,

fully,

XXXVII.
so that,

6a = S ( y - 0)(8 -

by 294, XLYIII. and

j3)(e

j3)

= S ( 7 8e -

XLVIL, we

8ej3

/3

78), &o.

have also (comp. 65, 70) the

equation,

XXXYIII.

with the relation between the

(36.)

(as

The equation

XL.

..0 = a +

(27.) ),

/cm

of a vector

K,

282, (10.), namely

a - 2S*a + g =

0,

being the vector of the centre

may be written as

XLI.
g being some scalar constant
fifth point E is situated.

/3

XLII.
(38.)

The

*e,

+ d+e,

which,

XLIY.

its

(3.),

2S/ce

= 0;

which the equation

0,

by the

vector

Y (- j3) 0-7)

equation of the circle ABC,

XLIII.

+ 0=0,..

- 2S
=
Kp + g

and on

in general be written thus

scalar g, between five scalar

the five following,

of the sphere ABCD, of

*

be obtained as the result of

and of a

2S/c/3

(37.) By treating this fifth point, or

the condition or equation of
concircularity
the formula,

may

XXXY. may

or condition

rfS

above) the origin of vectors to be arbitrary.

an elimination (294,
equations of the

+ cy +

bfi

coefficients,

XXXIX.
which allows

= aa +

condition referred to, the

g,

as arbitrary,

we

of the four points A, B,

y -8)

and the

(8

recover
c,

or

-a).

equation of the sphere ABCD,

V(a-j3)

(/3- 7 )

(y-p)

(p

-a)

= S a
(
-/3) 03- y ) (7-8) (8-p) (p-a)

p being as usual the vector of a variable point p, on the one or the other locus.
(39.) The equations of the tangent to the circle ABC, and of the tangent
plane

to the

sphere ABCD, at the point A, are respectively,

XLY...O
and

XLYI.

V(a-j3)(/3- 7 )( 7 -a)(p-a),

= S

(a

0) (/3- 7 ) (7-

8)

(8-

a) (p

a).

2Z2

ELEMENTS OF QUATERNIONS.

356

we combine the two equations XLIII. and

XLIV. and XLVI., we find in each case the equation,

or

XLVII.
it

6, 7.

i.

Accordingly, whether

(40.)

XLY.,

[III.

(p

a)

being supposed that the three points

four points.
(41.) If

A, B, c,

giving

0,

p = A (20)

or

a,

and that the

A, B, c are not collinear,

are not corn-planar.

the centre of the sphere, ABCD be taken for the origin

XLVIIL

a =
2

/3

f = & = - r\

XLIX.

or

o, so

that

To = Tj3 = Ty = TS = r,

the positive scalar r denoting the radius, then after some reductions we obtain
the transformation,

L...V(a-0)0- 7 )(y-8)(8-a)-2SO-a) (7-0) (*-),

(42.)

Hence, generally,

have the equation (comp.

LI.
(43.)

if

be, as in (36.), the centre of the sphere,

XXVI
Y (AB BC

we

.),

We may therefore

CD

DA)

= 12KA ABCD.
.

enunciate this theorem

The vector part of the product of four

successive sides, of a gauche quadri

lateral inscribed in a sphere, is

equal to the diameter

the polygon, multiplied

by the sextuple volume of the

drawn

to the initial

pyramid, which

point of

its four points

determine"

In effecting the reductions (41.), the following general formula of

transformation have been employed, which may be useful on other occasions
(44.)

LIL
where a

may

aq + qa = 2 (aSq + S?a)

be any vector, and q

may

LII

z
aqa = a Kq + 2aS?a

be any quaternion.

SECTION

On

7.

the Fourth Proportional to Three IMplanar Vectors.

297. In general,

when any four

quaternions,

q,

q",

",

satisfy the

equation of quotients,

I...q

":q"

=q

or the equivalent
formula,

we

shall say that

they

form a Proportion

the Fourth Proportional to the

first,

and that the fourth, namely q

second, and third quaternions, namely
;

",

is

to

FOURTH PROPORTIONAL TO DIPLANAR VECTORS.

ARTS. 296-297.]

357

and
taken in this given order. This definition will include (by 288)
q, q
the one which was assigned in 226, for the fourth proportional to three com,

q"

l
=
plane, 8
fia~ y,

namely that fourth vector in the same

which has been already considered and it will enable us to

planar

vectors, a,

y,

J3,

interpret (comp.

289) the symbol

III.

when 7

a
.

j3a~ 7,

new

as denoting not indeed a Vector, in this

which

may

these three

not

a, /3,

|||

case, but at least

a Quaternion,

be called (on the present general plan) the Fourth Proportional

Diplanar Vectors,

to

Such fourth proportionals possess some

a, j3, 7.

interesting properties, especially with reference to their vector parts, which it

will be useful briefly to
consider, and to illustrate by showing their connexion

with spherical trigonometry, and generally with spherical geometry.

(1.) Let a, )3, 7 be (as in 208, (1), &c.) the vectors of the corners of a
triangle

ABC on the

whereof the sides are

unit-sphere,
I

IV..

= cos a =
87/3
= cos b =

=-

=-

Scry"

and

a, b, e;

let

us write,

8)87,

870,

n = cos o = S/3cf = - Saj3

*

where

it is

understood that

V.

f3

=-

1,

VI.

or

Ta =

Tj3

T7

it
being also at first supposed, for the sake of fixing the conceptions, that
each of these three cosines, /, m, n, is greater than zero, or that each side of

the triangle ABC

(2.)

is less

than a quadrant.

Then, introducing three new

vectors,

s,

defined by the equations,

f 7 = Vyef /3 = mfi + ny - la,

a = Vaj3 -1 y = ny + la - mf3,
a = la +
J

mfi

we

find that these three derived vectors have all one

because they have one common norm

VIII.

IX.

so that
(3.)

This

common

we

shall

and the

TS = Te = T? = r =
X.

r, is less
.

80)87

common

length, say r,

namely,

NS = Ne = N? =

length,

ny,

+ n 2 - 2lmn = r z

</(P

than unity
1

8j3cT 7

-2
tow).

for if

we

write,

e,

have the relation,

scalar e is different

from

zero, because the vectors a,

|3,

7 are diplanar.

ELEMENTS OF QUATERNIONS.

358

Dividing the three lines

(4.)

their versors (155, 156)

DEF,

and

three

XII.

new

OD =

The

ew

so obtain a

we change them

r,

to

triangle,

US = r-

OE =

=r

Ue = r^e

?.

opposite to D, E, F, in this new or derived triangle, are

67, by the corners A, B, c of the old or given triangle

sides

as in

bisected,

their fewgtfA,

unit-vectors,

OF =
(5.)

by

i.

which the corners are deter

^Ae unit-sphere, of

<w

mined by the

S,

[TIL

fig.

because we have the three equations,

XIII.

= 2la

+ 8 = 22/3

8 +

2wy

c (so that the arc

Denoting the //te of the 0z0 sides by
2r/, &c.), tlie equations XIII. show also, by IY. and IX., that
r

(6.)

EF =

</,

XI Y.

cos a

r cos #

cos b = r cos b

cos c

= r cos

the cosines of the half-sides of the new (or bisected) triangle, DEF, are therefore
to the cosines

proportional

The equations

(7.)

XY.
we have

21 = -

therefore,

of the sides of the old (or bisecting) triangle ABC.

()3y

may

and express

2m

+ y/3),

(1.),

= - ( 7 a + ay),

2n = -

(a/3

|3a)

by YIL, the three following equations between quaternions,

XYI...
which

by 279,

(IY.) give,

j32

T8 = 7

S/3,

also be thus written,

in a

new way

the relations of bisection

(5.).

(8.)

We have therefore

the equations between vectors,

(9.)

Hence

or because a,

or

by Y.,

also,

XYIir...

or
(10.)

tion

7 are

unit-vectors,

In general, whatever

XYII.

|3,

the length of the vector a

expresses that the line

e is

with respect to that vector a; because

(comp. 138)
it

may

may

be,

the

1
.

Sa-

=a

= Kea-

or

equa
,

be put (comp. 279) under the

form

XIX.

first

the reflexion of the line

XIX

. .

aa

Ka>.

EXPRESSIONS FOR CONICAL ROTATION.

ART. 297.]
(11.)

= ajar

1
,

Another mode

same

of arriving at the

to conceive

is

decomposed

one parallel and the other perpendicular to

xx.. .?=?
for then

we

shall have,

XXI.

= a? a- +

in such a

and

",

manner that

(10.), the transformations,

by 281,
1

a,

vectors,

r-La;

nio,

+?",

interpretation of the equation

two summand

into

359

a"a

= Z aa - g
1

?"

being thus preserved, but the perpendicular part being

the parallel part of

the operation a

or
)
(
by
from
g = a^cr to the form ca = a, that is, to the
return
Or
we
may
(12.)
and then this equation between quaternions will show,
first equation XVF.
as suggested in (7.), that whatever may be the length of a, we must have,

reversed,

XXII.

so that the rfwo lines

to that

from a

common

plane.

(13.)

to

XXIII.
(14.)

we

more

XXV.
(15.)

T,

Ax.*

ta

= Ax.

An

L a =L a

aZ,

S are equally long, and the rotation from

these

two rotations being

also write the equations

new

XXIV.
is,

=a

a,

&0.

XXIII

to a is

and

similarly directed,

XVII. XVII
?= a

7
.

a,

&o.

S=

jSer fa/3

XXIV

or

7
.

.,

SjSa

ajS

briefly,
.

8-

M-

and

XXV

7
.

^q

if

XXVI.

fia

expression of this form, namely one with such a symbol as

..<?(

for an operator, occurred before, in 179, (1.),

)<r

and in 191

(5.)

and was seen

to indicate a conical rotation of the axis of the operand quaternion (of

symbol

XVII

equation,

xxvii.

is

6w/

in one

under the forms,

in the second equation

Substituting this last expression for

derive this

that

c,

We may

to be conceived as being written within the

parentheses)

which the

round

the axis

through an angle = 2 L q, without any change of the angle, or of the tensor,

of that operand-, so that a vector must remain a vector, after
any operation of

of

q,

* It was remarked in
291, that this characteristic Ax. can be dispensed with, because
UV but there may still be a convenience in employing it occasionally.

being replaced by

it

admits of

ELEMENTS OE QUATERNIONS.

360
as being

tliis sort,

still

a right-angled quaternion (290)

[III.

i.

or (comp. 223,
(10.)

because

XXVIII.

1
l
Sqpq- = $q- q p = 8,0-0.

(16.) If then we conceive two opposite points, p and p, to be determined

on the unit-sphere, by the conditions of being respectively the positive poles of
the two opposite arcs, AB and BA, so that

XXIX.
we can

infer

OP = Ax.

from

of the angle

AOB

And

(17.)

XXIV.

round

conical rotation

1
jSa"

that the

the line

o be

= Ax.

and

q,

line

OP as an

OP = p o = Ax.

OD may
axis,

apt

= Ax.

q~

be derived from the line OE, by a

through an angle equal

the centre of the sphere)

(if
in like manner we can infer from
still

to the

double

XXI V

that the line OE

.,

admits of being derived from OD, by an equal but opposite conical rotation,
round the line OP as a new positive axis, through an angle equal to twice the

angle BOA.

To

(18.)

illustrate these

and other connected

annexed figure 68

results, the

which p represents, as above, the

of
the arc BA, and arcs are drawn
positive pole
from it to D, E, F, meeting the great circle
is

drawn

in

through A and B in the points

R, s, T.

other letters in the figure are not,

(The

for

the

moment, required, but their significations will

soon be explained.)

Ll

This being understood, we see, first,

that because the arcs EF and FD are bisected (5.
(19.)

,)

at

A and

B, the three arcual perpendiculars, ES,

FT, DR, let fall

on the great circle

and that
are equally long

from

through A and

B,

E, F, D,

therefore the point p

is

the interior pole of the small

DEF

circle

if

F be the

so that a conical rotation round this pole p, or

point diametrically opposite to F
axis OP, would in fact bring the point D, or the line OD, to the posi
tion E, or OE, which is one part of the theorem (17.).
:

round the

(20.)

Again, the quantity of this conical rotation, is evidently measured by

circle with p for pole
but the bisections above mentioned

the arc RS of the great

give (comp. 165) the two arcual equations,

XXX.
and the

n RB = n BT,

other part of the

n TA = n AS

same theorem

whence

XXXI.

(17.) is proved.

RS = 2

r\

BA,

CONICAL ROTATION.

ART. 297.]

The

(21.)

point F

may

be said to be the

point D, with respect to the point B,

thus

we may say
and

rotation,

round

of a small

reflexion,

on the sphere, of the

between them

bisects the interval

and

that two successive reflexions of an arbitrary point upon a

from D

sphere (as here

points (B

which

361

and then from F

to F,

to E), ivith respect to two given

A) of a given great circle, are jointly equivalent to one conical

the pole (p) of that great circle

circle,

or to the description of an r*rc

round that pole, or parallel to that great circle and that the
:

angular quantity (DPE) of this rotation is double of that represented by the arc
or is the double of the angle (BPA), which
(BA) connecting the two given points
the
same
at
that given arc subtends,
pole (p).

There

(22.)

of rotation

but

is,

as

it is

we

see,

no

difficulty in geometrically proving this theorem

remarkable how simply quaternions express

it

namely by

the formula,

XXXII.

/3>j3

= a/^ 1 p
.

1
.

jStf

denote any three vectors and which, as we see by the

points, involves essentially the associative principle of multiplication.
(23.) Instead of conceiving that the point D, or the line OD, has been
in

which

a, |3,

may

the position F, or OF, with respect to the

point B, or to the line OB, with a similar successive
reflexion from F to E, we may conceive that a point has

reflected into

moved

along a small semicircle, with B for pole, from

to

,.-

/
:

A for

of these two successive

VTT

DE

arc

of a third small

circle,

which

is

is

equivalent to a

~f;

-_..-\y

\l/_

and then along another small

and we see that the
pole, from F to E

and semicircular motions

/Sv

Fig. 69.

F, as indicated in fig. 69,

semicircle, with

result, or effect,

motion along an

parallel (as before) to the great circle

through B and A, and has a projection RS thereon, which

(still

as before) is

double of the given arc BA.

And

instead of thus conceiving two successive arcual motions of a point

D upon a sphere, or two successive conical rotations of a radius OD, considered as
(24.)

compounding themselves into one resultant motion of that point, or rotation of

that radius,

we may

conceive an analogous composition of two

rotations of a solid body (or rigid system],

which

is

fixed in space (and in the body}

round axes passing through a point o,

and so obtain a theorem respecting

such rotation, which easily suggests itself from

we may perhaps return.

(25.) But to draw some
&c.,

and from the recent

HAMILTON

successive

what precedes, and on which

additional consequences

from the equations VII.,

the
fig. 68, especially as regards the Construction of

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QTJATEBNIONS.

362

[III.

i.

7.

Fourth Proportional to three diplanar vectors, let us first remark, generally,

that when we have (as in 62) a linear equation, of the form
aa +

co-initial vectors a

connecting four

HUB fifth

bfi
.

it

0,

whereof no three are complanar, then

S,

vector.

= aa +
is

+ cy + d$ =

evidently complanar (22) with a,

is therefore part of the indefinite

bf3

- - cy -

d$,

/3,

and

with

also

y, 8

(comp. 294, (6.) )

of intersection of the plane AOB, COD, of

line

these two pairs of vectors.

(26.)

And

if

we

divide this fifth vector

by the two (generally unequal)

- d,

scalars,

a +

and

b,

the two (generally unequal) vectors,

(aa

bfi)

which are obtained as the

(a

and

b),

(cy

quotients of these

+ d$)

two

(c

+ d),

divisions, are (comp. 25, 64)

the vectors of two (generally distinct) points of intersection, of

namely the two following

lines

with planes,

AB OCD,
(27.)

When

the two

lines,

and

AB and

CD,

CD OAB.

happen

to intersect each other, the

two

and thus we recover, in a new way, the con

dition (63), for the complanarity of the four points o, A, B, c, or for the terminolast-mentioned points coincide

complanarity of the four vectors a,

which

y, S

/3,

namely the equation

be compared with 294, XLV., and L.

(28.) Eesuming now the recent equations VII., and introducing the

may

new

vector,

XXXIII.
which

A =

la

= \

mfi

(e

8),

gives,

XXXIV.
we

..

S7X =

0,

and

XXXV.

..

TX = v/(r - ;r) = r sin

2

two arcs BA, DE, prolonged, meet in a point L (comp. fig. 68),
=
for which OL
UX, and which is distant by a quadrant from c: a result which
be
confirmed
may
by elementary considerations, because (by a well-known
see that the

theorem respecting transversal arcs] the common bisector BA of the two

DE and EF, must meet the third side in a point L, for which
sin

DL =

sin EL.

sides,

CONSTRUCTION OF A FOURTH PROPORTIONAL.

ART. 297.]

To prove

(29.)

common

we

value,

and to assign their

by quaternions this last equality of sines,
have only to observe that by XXXIII.,

XXXVI.
TS\ = TcA =

in which,

363

VSA = V A =

sin c

TVSe =

and

r sin 2c

the sines in question are therefore (by 204, XIX.),

XXXVI
On

(30.)

in

XXXVIII.

which

we may

similar principles,

XXXVII.
.

TVUeA = Jr

TVUSA

V/3A

sin 2c

interpret the

sin c

r* sin c

two

= cos

vector-equations,

VaA = mV/3a,

= /Vj3,

TA TV/3a = r

sin c

= tan c tan
:

c,

an equivalent to the trigonometric equations,

XXXIX.
(31.)

we

if

Accordingly,

fig. 68), so that

bisects RS,

tan CD

cos BC

cos

AC

tan AB

sin

BL

sin

AL"

the perpendicular CQ on AB (see again

points M, N by the

let fall

and

we determine two new

if

arcual equations,

XL.
ND

the arcs MR,

MN and

CB are drawn,

<~>

LM =

n AB =

will be quadrants

construction (18.),

pole of

is

QR,

LNM

is

n LN =

r\

CD,

and because the angle

the pole of DR, and

the angle

we have

r\

right

DM

is

at

a quadrant;

is

right

whence D

is

by
the

conceiving then that the arcs CA and

and LMN], right-angled at

Q and N, which show, by elementary principles, that the three trigonometric
quotients in XXXIX. have in fact a common value, namely cos CQ, or cos L.
(32.)

To prove

three triangles [BCQ, ACQ,

this last result

auxiliary points M, N, Q, R,

by

we have

quaternions,

and without employing the

the transformations,

because

and

XLIII.
it

= - S/3- 7 X- = - 88X- = 1

being remembered that A

-L

V 7A = 7 A = - A7

7,

whence
2

7 A)

2
2
= A2
7 A

S7A

0.
3

A2

ELEMENTS OF QUATERNIONS.

364

At

(33.)

BA,

and

if

negative

is

see that

if

(as before)

the positive) pole of CQ, whereby

wo may

their significations determined,

XLIY.

p be

7.

i.

the positive pole of

be the negative and positive poles of DE, while i/ is the

K,

(as

we

the same time

[III.

OP = UV]3a

the letters in

fig.

68 have

write,

OK = - 7 UX;

OK = 7 UX;

all

oi/ =

-UX;

OL = + UX, as before.

while
(34.)

Writing

also,

XLV.

XLV.

so that

XL VI.

we have

= - yX,

OK = UK,

1
.

jScr

\ = yx,

or

OM =

and

\K~ =

and

X,

|3a"

TJ/z,

fj.\"

ju

fj.K~

ihis fourth proportional, to the three equally long but diplanar vectors, a,

therefore a versor, of which the representative arc (162)

sentative angle

KDM, or L DR, or EDP

is

(174)

versor, or quaternion, the expression

XLVII.
The

(35.)

jSa-

and we may write

DPE

= cos L DR + OD

for this

sin L DR.*

the"

sum

of the two base-

and because the two other

y, is

double of this representative angle is

angles of the isosceles triangle

)3,

KM, and the repre

is

sum

triangles,

EPF

what remains,
when we subtract the vertical angle F, of the triangle DEF, from the sum of the
or when we
supplements of the two base-angles D and B of that triangle
F PD, are also isosceles

(19.),

the lune FF shows that this

is

subtract the

We

sum of the

have therefore

Proportional

same triangle from four

three angles of the

right angles.

very simple expression for the Angle of the Fourth

this

XLVIII.
(36.) Or, if

L Pa~ y = L

DR =

we introduce the

7T

(D

+E+

F).

area, or the spherical excess, say S, of the

triangle DEF, writing thus

XLIX. ..S
we have

D+E

these other expressions

L.

(BeT

y=

\TT

- iS

LI.

OD = TJS =

because

+ F-7T,

pa^y =

r~ $,

by

sin

|S +

cos

}2

XII.

7 (jScr )- = 7 suppose, c is brought to a point c by a

where p is the opposite of P (XXIX.). Hence c and c are
the points of intersection of small circles whose poles are D and p and c is the reflexion of c to the
This shows that the angle of the quaternion fia.~ y is CDP.]
great circle PD.

[Since fior^y

conical rotation round

(jSa

OD

1
^)- =

a- ]

or round OP

ANGLE OF A FOURTH PROPORTIONAL.

ART. 297.]
(37.)

we

Having thus expressed

y,

fi<r

365

new appeal

require no

figure, in order to express this other fourth proportional, ya~

negative of its conjugate, or has

an

opposite

fl,

which

to the
is

the

hut an equal vector part

scalar,

the geometrical difference being merely

(comp. 204, (1.), and 295, (9.) )
this, that because the rotation round a from (B to y has been supposed to be
:

negative, the rotation

(38.)

We may
LIL

round a from y
= -

1
.

to

must

j3

on the contrary,

be,

positive.

thus write, at once,

.
7<r

/3

KjSeT y

= -

^+r

sin

iS

$ cos

and we have, for the angle of this new fourth proportional, to the same three
vectors a, |3, 7, of which the second and third have merely changed places with
each other, the formula
:

LIIL
(39.)

VII.

Lya

fi

But the common

we have

/=

..

+E+

F)

JTT

follows

cos

JS

when

by

(6.)

that

which the area

triangle DEF, of

sin

is

by

the rotation round a from

when
is

the sides 2a

S, are bisected

another spherical triangle, of which the sides* are

LV.

cos a

It follows also,

RDK, or MDN, or the arc

triangle

S.

|3

to

is

nega

above supposed.

(40.) It

(41.)

by XI.,

the upper sign being taken,

tive, as

(D

vector part of these two fourth proportionals is 8,

therefore,

LIV.

= RDL =

cos a = cos b

cos b

2b

by the

a, b, c,

= cos

2c

of a spherical

corners A, B, c of

then

cos c = cos |S.

from what has been recently shown, that the angle

MN

in

fig.

68, represents the semi-area of the bisected

DEF; whence, by the right-angled triangle LMN, we can

the sine of this semi-area

is

ini er that

equal to the sine of a side of the bisecting triangle

ABC, multiplied into the sine of the perpendicular, let fall upon that side from
the opposite comer of the latter
triangle because we have
;

LVI.
(42.)

sin is = sin

The same

MN = sin LM

conclusion can be

sin L

sin

AB

sin CQ.

drawn immediately, by quaternions,

from the expression,

LVIL

sin

in which one factor

iS = e = Saf^y = S (Vj3a 7
.

is

= TV/3a

SU (V/3a

the sine of AB, and the other factor

is

7)

the cosine of CP,

or the sine of CQ.

*
These sides abc, of the bisecting triangle ABC, have been hitherto supposed for
simplicity (1.) to
be each less than a quadrant, but it will be found that the formula LY. holds
good, without any such
restriction.

ELEMENTS OF QUATERNIONS.

366

Under

(43.)

same conditions,

the

LYIII.

we may

a=

(e

[III.

7.

i.

since

Q = jr

a)

&c.,

),

write also,

LIX.

is = SU

sin

and XIII.,
limn - - S
LX.

)(

8) (8

= S8

4/mn

in which, by IY.

Hence

(44.)

JS = r =

cos

(r

tan

this last form,

Te,

of the edges at

As

(45.)

LXIII.

=-J
(4/m/O

LXI Y.

we ought then

if

(46.)

(e+

J8 +

8e).

- rS

+ 8+

(E?

Se)

1 -

a general expression for the tangent of half the

the lengths TS,

be.

we have
2

(^

(8

g)

+ e) a = 2

= (SS

^)

(r

-r
(S

(;-

Se^) (r

S?8)

(r

SSe)

gives

+ S?8 + S&)

2
)

= - r2

(SSeZ)
;

Under

= r 6 - ra

and in

8
2
(S^) + (S8) +

(S8e)

fact this equality results

2Sc2S?8S8,

immediately from the

LIU.

the same condition, respecting the equal lengths of

8,

also the formula,

LXYI.

whence other
(47.) If

ciple

- S

r*

is between LIX. LXI.

8
.

general formula 294,

we have

(e

(4/ww)

2)

to find that

LXY.
2

may

a verification,

but the elimination of

we have

any triangular pyramid ODEF, whatever

spherical opening at o, of

e)(e

^ + 28 + 8e)

and under

(S

by LIY.,

also,

LXI.
TirTT
LXIl.

LXII.

-Y

(8 +

e) (a

verifications

may

(^

8)

= 28

(r

- Se -

S8 -

S&) = SlmnS

be derived.

denote the area* of the bisecting triangle ABC, the general prin
enables us to infer that

<r

T VT7TT

LXYII.

tan -

Sa3
1 - 8j3y

Sya 8

sin

cos a

+ cos

+ cos

* The reader will observe that the more usual

symbol 2, for this area of ABC,
to denote the area of the exscribed triangle DEF.

is

here employed

CONNEXION WITH SPHERICAL AREA.

AET. 297.]
if

the perpendicular CQ from c on AB, so that

p denote
e

(48.)

367

= sin

siup =

sin b sin c sin

A = &c. (comp. 210,

(21.)

).

But, by (IX.) and (XL),

LXVIII.

+l +

-a

4 cos

c\ z

cos-

cos^j;

hence the cosine and sine of the new semi- area are,

^ T^
LXIX.

cos s =

+ cos a +

cos b

+ cos

OS

OS

C S

/C

sin - sin = sin c

-TT-cr
LXX.

sm - =

0"

COS

2
(49.)

Returning

LXXI.
if

to the bisected triangle, DBF, the last

sin b sin

-7

cos c

- = sin
p

sm c

sec c

denote the perpendicular from F on the bisecting arc AB, or FT in

j2 = cos c sec c by LY. hence

but cos

Accordingly, in

FT =

sin
(50.)
its

The

68,

fig.

sin ES

tan

we

have, by spherical trigonometry,

sin L

= cos LN sin

MN

cosec

LM = tan MN

cot AB.

which thus represents in quantity the semiarea of DEF,

and may be considered as the representative arc

pole at the point D,

(162) of a certain new quaternion Q, or of its versor, of which the axis

radius OD, or US and this new quaternion
may be thus

expressed

LXXIII.
its

68;

JS = sinX tan c = sin FT tan AB.

= sin LE

arc MN,

fig.

LXXII.

has

sin a
JS = -

sin

formula gives,

is

the

tensor and versor being, respectively,

LXXIY.
(51.)

follows

An

.TQ=r = cosJS; LXXV.

important transformation of this

LXXVI.

UQ =

so that

LXXVII.

S =

TTQ = cos J2
last versor

may

be obtained as

ELEMENTS OF QUATERNIONS.

368

[HI.

7.

Im

these powers of quaternions, with exponents each = i, being interpreted as

1
square roots (199, (1.) ), or as equivalent to the symbols x
), &o.

/(-

(52.)

and

conjugate (or reciprocal) versor,

may

sentative arc,
c

The

be deduced from

UQ by

the corresponding quaternion

LXXVIII.
and we have

LXXIX.

cos

Q =

NM

which has

simply interchanging

for its repre

and

]3

7, or

is,

KQ =

JS - OD

the rotation round D, from E to F, being

sin

supposed to be negative.
the
other
be
point upon
sphere, and let OH = rj also let
any
(53.)
DFH
new
then the same reasoning
the
be the area of
spherical triangle,

Let

UQ"

still

shows that

LXXX.
if

cos

is + OD

the rotation round D from F to

two co-axal

cation of the

the analogous formula

LXXXI.
where S +

versors,

sin

is = (Sf

be negative

LXXYI.

and

)*

(&T
and

1
)

(n^

)*,

therefore,

LXXX., we

by

have by

multipli

LXXV.,

cos i (S +

S)

+ OD

sin \

(S +

S)=

(&T

)*

(-^ (&f )! (n^

denotes the area of the spherical quadrilateral, DEFH.

(54.) It is easy to extend this result to the area of any spherical polygon, or

to the spherical opening (44.) of any

extension of

it,

pyramid

and we may even conceive an

as a limit, to the area of any closed curve upon the sphere, con

sidered as decomposed into an indefinite number of indefinitely small triangles,

such as the point D, on the spheric surface, and with

of the curve, for their respective bases or to the
indefinitely small arcs EF, FH,
spherical opening of any cone, expressed thus as the Angle of a Quaternion,

with some common

vertex,

which

is

the limit* of the product of indefinitely

many

square-root of a quaternion, which differs indefinitely

factors, each equal to the

little

from

unity.

or rather, we may
is closely analogous to a definite integral, of the ordinary kind
a Definite Integral, but one of a new kind, which could not easily have been introduced
without Quaternions. In fact, if we did not employ the non-commutative property (168) of quaternion
so that they
multiplication, the Products here considered would evidently become each equal to unity
would furnish no expressions for spherical or other areas, and in short, it would be useless to speak
of them.
On the contrary, when that property or principle of multiplication is introduced, these
in spherical
expressions of product-form are found, as above, to have extremely useful significations
the
a
remarkable
and
it
will
and
theorem,
respecting
be
seen
that
geometry;
embody
they suggest
resultant of rotations of a system, round any number of successive axes, all passing through one fixed
*

This Limit

say that

it is

point, but in other respects succeeding each other with

any gradual or sudden changes.

AREA OF POLYGON OR CURYE OK SPHERE.

ART. 297.]

To

(55.)

assist the recollection of this result, it

(3.) for

(comp. 180,
"

The Arcual

the definition of an arcual sum)

may

369

be stated as follows

Sum

of the Halves of the Successive Sides of any Spherical

an arc of a Great Circle, u-hich has the Initial (or Final)

Polygon, is equal to
Point of the Polygon for its Pole, and represents the Semi-area of the Figure"
it being understood that this resultant arc is reversed in direction, when the
,

half-sides are (arcually)

added in an opposite order.

that in the
regards the order thus referred to, it may be observed
the
to
inLXXVL,
arcual addition, which corresponds
quaternion multiplication
(56.)

we

As

conceive a point to move,

half-side of the triangle

first,

from B to

DEF answers

F,

through half the arc DF

which

to the right-hand factor, or square-root

We then conceive the same point to move next from F to A, through

(2S- )*.
half the arc FE, which answers to the factor placed immediately to the left of
1

the former
arc

BA

through the resultant

any equal arc (163), such as ML
(3.) ),
we conceive a motion through half the arc ED, or

having thus moved, on the whole,

a transvector, 180,

(as

And

in fig. 68.

through any

finally,

so far,

or through

arc equal to that half, such as the arc

LN in the same

(or total

figure, to

extreme left-hand factor in the formula the final resultant

transvector arc), which answers to the product of the three square roots,

correspond to the

as arranged in the formula, being thus represented by the final arc MN, which
has the point D for its positive pole, and the half-area, ^S, for the angle (51.)
of the quaternion (or versor) product which it represents.
(57.)

Now

to be that

the direction of positive rotation on the sphere has been supposed

and therefore along the perimeter, in the order
D, from F to E

round

DFE, as seen* from any point of the surface loithin the triangle that is, in the
order in which the successive sides DF, FE, ED have been taken, before adding
:

compounding) their halves. And accordingly, in the conjugate (or reciprocal)

LXXIX., we took the opposite order, DEF, in proceeding as usual
from right-hand to left-hand factors, whereof the former are supposed to be
(or

formula

multiplied

fo/f

the

latter-,

while the result was, as

we saw

in (52.), a new

* In this and other cases of the

sort, the spectator is imagined to stand on the point of the sphere,
round which the rotation on the surface is conceived to be performed his body being outside the sphere.
And similarly when we say, for example, that the rotation round the line, or radius, o A, from the line
;

to the line oc, is negative

(or left-handed), as in the recent figures, we mean that such would appear
be the direction of that rotation, to a person standing thus with his/<?^ on A, and with his body in
the direction of o A prolonged or else standing on the centre (or origin) o, with his head at the point A
Compare 174, II. ; 177 and the second note to page 152.
t Compare the Notes to pages 147, 159.

OB
to

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

370

versor, in the expression for

own

to its

changed

which, the area

[III.

i.

7.

was simply

of the triangle

negative.

To

give an example of the reduction of the area to sero, we have

only to conceive that the three points D, E, F are co-arcual (165), or situated
(58.)

on one great

circle

by the laws*

or that the three lines

LXXXII.
thus cos

JS =

(59.)

(Sr

and 2 =

1,

)* (E?- )!

(2S-

this case,

the formula,

S&

if

1,

=
;

0.

in (53.) let the point

Again,
;

we have

of complanar quaternions,

For

are complanar.

S, c,

be co-arcual with D and

F, or let

then, because

Lxxxir.

(&r

>*

(us- )*

= (ss-

if

)*,

ss&i =

o,

the product of four factors LXXXI. reduces itself to the product of three
factors LXXVI.
the geometrical reason being evidently that in this case the
added area S vanishes so that the quadrilateral DEFH has only the same area
;

as the triangle DEF.

But this added area (53.) may even have a negative^ effect, as for
example when the new point H falls on the old side DE. Accordingly, if we
(60.)

write

LXXXIII.
and denote the product

Q, =

LXXXI.

(- )* (&,-

(,r ),

of four square-roots

by Q 2 we
,

shall

have

the transformation,

LXXXIY.
which shows (comp.

(15.)

Qt =

(gc- )*

Qi

(^ft

if

S8, =

that in this case the angle of the quaternary product

is that of the ternary product Q l9 or the half-area of the triangle EFH

DEF
DHF), although the axis of Q 2 is transferred from the position of the
(=
axis of Qi, by a rotation round the pole of the arc ED, which brings it from

OE to OD.
(61.)

From

this

example,

it

may

be considered to be sufficiently evident,

LXXXI. may be applied and extended, so as to represent

area of any closed figure on the sphere, with any assumed point
the
(comp. (54.))

how

the formula

Compare the Second Chapter of the Second Book.

f In some investigations respecting areas on a sphere, it may be convenient to distinguish (comp.
between the two symbols DEF and DFE, and to consider them as denoting two opposite
(28.), (63.)
But for the present, we are content to express this distinction, by
triangles, of which the sum is zero.
means of the two conjugate quaternion products (61.) and (52.).

AREA OF POLYGON OR CURVE ON SPHERE.

AET.297.]

D on

the surface as a sort of spherical origin

not situated on the perimeter, but

A new

(62.)
(50.),

quaternion

is

even when this auxiliary point

OB

axis

as the quaternion

but with a double angle, and with a tensor equal to unity,

by simply
effected by removing

and although

UQ;

the fractional exponents* in the

this

can easily be accomplished in other ways.

LXXIV., and by VII. IX. X., the transformations!
.

.
<2

UQ

(g 7 ]3)

-in fact, because S

e

LXXXVI.
To show

(ya/3)

-=

may

LXXYL,

yet

it

by LXXIII.

have,

= - (e-S)

yajSS
2

gya/3

=r -e
2

+ 2eS;

2
Qo = - (ya/S) = cos S + OD

sin 2.

geometrical signification, let us conceive that ABC and

have the same meanings in the new fig. 70, as in fig. 68 and that

(63.)

LMN

of

= r OD, by XII., the trigonometric values LIV. for r

enable us to write this last result under the form,

and
and

r~

be formed

squaring cannot be

formula

For example we

LXXXV.

is

either external or internal thereto.

with the same

squaring the versor

371

its

new

are three

points, determined by the three arcual equations (163),

LXXXVIL

n BC = n CBi,

n AC = ^ CAi,

which easily conduct to

this fourth equation of the

MN

= n

NM

A_

jt_

same kind,

LXXXVIF.
This
fig.

n LM! = n

B^.

thus (comp. 167, and

1
=
while
yjSr

rc L,MI represents

w0>

43) the product

a^"

jSy"

ya"

Fig. 70.

equal BA (31.), represents

whence the arc MMi, which has its pole at D, and is numerically equal to the
whole area S of DEF (because MN was seen to be equal (50.) to halfih&t area),
2
the
The formula
or or Q
the old arc ML, or

a/3"

product

represents

LXXXVI.

its

ya^jSy"

ajS"

(ya/3)

has therefore been interpreted, and

may

be said to have been

proved anew, by these simple geometrical considerations.

see, at the same time, how to interpret the symbol,
(64.)

We

LXXXVIII.
namely

as denoting a versor, of

which the

*
Compare the Note to (54.).
t The equation 870/2= 70)88 is not valid generally

W=

p(l>

Qo = I

but

$ ^
7 P

axis is directed

we have

here 8

to,

70)8

or from, the

and in general,

tfpll Vtf.
3

ELEMENTS OF QUATERNIONS.

372
corner

[III.

7.

i.

of a certain auxiliary spherical triangle DEF, whereof the sides respec

by the

tively opposite to D, E, F, are bisected (5.)

round a from

as the rotation

angle represents, or
triangle, at least

if

is

we

|3

to

numerically equal
still

given points A, B,

negative or positive

is

we have

suppose, as

the area

to,

and

c,

according

of

which the

of that auxiliary

hitherto for simplicity done

the given triangle ABC are each less than a quadrant.

(1.), that the sides of
298. The case when the sides of the given triangle are all greater, instead
of being all

than quadrants,

less,

considered;

briefly)

deserve next to be (although more

are all equal to quadrants, being

may

when they

the case

reserved for a short subsequent Article and other cases being easily referred
to these, by limits, or by passing from a given line to its opposite.
:

Supposing now that

(1.)

or that

we may

I.

II.

0,

<

>

-,

<

>

retain the recent equations

still

0,

0,

<

-,

>

IY. to XT.

XIII.

XV.

and

to

of 297; but we must change the sign of the radical, r,

XII. and XIV., and also the signs of the versors, US, Ue,

in the equations

desire that the sides of the auxiliary triangle, DEF,

be

XXVI.,

figures 67, 68)

a, b, c

are

now

tensor of S,

E,

IV.
(2.)

in XII., if

and therefore being

OD = cos a

still

US

= - r

= cos

rS

OE =

cos b

</,

supposed to be

Ue = - r

= - r cos

The equations IV. and VIII.

new

of

297

still

2r cos a cos

cos c

cos a

fz

= - r cos

we must

= - r~^
c

holding good,

+ cos

0,

>

equations,

OF = -

cos

itself

write,

V.

we

bisected (as in

by the corners of the given triangle, ABC, of which the sides

each greater than a quadrant. Thus, r being still the common

write now, under these new conditions I. or II., the

III.

may

still

cos c

we may now
1,

we adopt positive values (297), or negative values (298), for the

the value of r being still
n
of the sides of the bisecting triangle
/, m,
supposed to be positive.
that
(3.) It is not difficult to prove (comp. 297, LIV., LXIX.),

according as
cosines

VI.

r =

cos J2,

the recent formula V.

VII.

may

2 cos a cos

according as

>

0, &c.,

or

<

0,

&c.

therefore be written unambiguously as follows

f

cos c cos

and the formula 297, LV. continues

JS =

cos a

to hold good.

cos b

+ cos

- 1

CASE OF SIDES GREATER THAN QUADRANTS.

AKTS. 297, 298.]

373

In like manner, we

may write, without an ambiguous sign (comp.

the
297, LI.),
following expression for the fourth proportional fla~ y to three
unit-vectors a, |3, 7, the rotation round the first from the second to the third
(4.)

being negative

YIII.

f3a~

where the scalar part changes

j = sin

iS + OD

when the

sign,

JS

cos

rotation

is

reversed.

It is, however, to be observed, that although this formula YIII. holds

not
good,
only in the cases of the last article and of the present, but also in
that which has been reserved for the next, namely when I = 0, &c.
yet
(5.)

because, in the present case (298)

US

longer the (positive) axis

we have

the area

>

TT,

the radius OD
l

of the fourth proportional (5ar y

any longer, as in 297, L., the (positive) angle of that versor.

we have now, for this axis and angle, the expressions

nor

no
- JS

is

is JTT

On the contrary

IX.
(6.)

in

fig.

To

Ax

1
.

fin

- OD
? = DO =

illustrate these results

by

X.

L fia y = i (S l

TT).

we may remark

a construction,

that

if,

AB be supposed each greater than a

form from it a new figure, analogous to 68,

67, the bisecting arcs BC, CA,

quadrant, and if we proceed to

the perpendicular CQ will also exceed a quadrant, and the poles p and K will
fall between the
points c and Q; also M and R will fall on the arcs LQ and QI/

and although the arc KM, or the angle KDM, or I/DR, or EDP, may
be considered, as in 297, (34.), to represent the versor ficr y, yet the

prolonged

still

corresponding rotation round the point D

(7.)

And

is

now

of a negative character.

as regards the quantity of this rotation, or the

magnitude

of the

angle at D,

it is
again, as in fig. 68, a base-angle of
one of three isosceles
triangles, with P for their
common vertex; but we have now, as in fig. 71,

a new arrangement; in virtue of which this

angle is
to be found by
what
when
the
remains,
halving

sum

of the supplements of the

angles at D and E, in the triangle DBF, is
subtracted from the angle at r, instead of our
subtracting (as in 297, (35.))
the latter angle from the former sum ; it is therefore now, in
with

agreement

the recent expression X.,

XI.
(8.)

The negative

f3a~

= i
(D

4-

E +

of the conjugate of the

Xll.

1
.

<r

F)

TT.

formula YIII. gives,

3 = - sin -2 + OD cos
.

ELEMENTS OF QUATERNIONS.

374

and by taking the negative

to the following

[III.

of the square of this equation,

we

7.

i.

are conducted

--

XIII.

= -

= cos

(yaW

S + OD

sin

a result which had only been proved before (comp. 297, (62.), (64.) ) for the
S TT and in which it is still supposed that the rotation round a from

case
)3

to

<

is

(9.)

negative.

With

we have

the same direction of rotation,

also the conjugate or

reciprocal formula,

XIY.
(10.) If

= cos

(/Ba-^)

- OD sin S.
.

happened that only one side, as AB, of the given triangle ABC,
while each of the two others was less than a quadrant, or that

it

was greater,

we had

- = y

^
p

0,

>

>

but n

0,

<

and

if

we wished

to represent the fourth

proportional to a, j3, y by means of the foregoing constructions we should

= y
only have to introduce the point c opposite to c, or to change y to y
for thus the new triangle ABC would have each side greater than a quadrant,
:

and

so

would

fall

under the case of the present Article after employing the

we should only have to change the resulting versor to
;

construction for which,

its

negative.
(11.)

And

in like

manner,

if

we had

and

negative, but n positive,

might again substitute for c its opposite point

construction of Art. 297 and similarly in other

any

In general,

we

if

arbitrary (but positive)

initial vectors 8,

e,

of

cases.

(12.)

we

and so fall back on the

begin with the equations 297, XII., attributing

value to the common tensor, r, of the three co-

which the versor s, or the

unit-vectors

US,

&c., termi

25 ,
nate at the corners of a given or assumed triangle DBF, with sides =
sides
with
we may then suppose (comp. fig. 67) that another triangle ABC,
2<?

2</,

denoted by
this one,

by

and with their cosines denoted by I, m, 11, is derived from

and therefore by the equations
its sides

the condition of bisecting

LVIIL),

(comp. 297,

XY.

b, c,

oA = a =

U(

),

oB =

/3

U(

with the relations 297, IY. V. VI., as before

or

oc =

g),

<y

U(S

..

+?

2ra cos a

% + S = 2rj3 cos b

),

by these other equations

(comp. 297, XIII. XIY.),

XYI.

= 2ry cos

MODIFICATIONS OF THE CONSTRUCTION.

AET. 298.]

When

(13.)

simple construction

this

is

adopted,

we have

375
at once (comp.

297, LX.), by merely taking scalars of products of vectors, and icithout any
reference to areas (compare however 297, LXIX., and 298, VII.), the
equations,

XVII.

4 cos a cos b cos c = 4 cos b cos c cos a = 4 cos c cos a cos V

= - r 2 S (? + 8) (3 + e) = &o. = 1 + cos 2a + cos 26 + cos 2c

or

cos a

cos a

_cosb _

cos c

cos 6

cos c

f*

_ cos # + cos b + cos c

2 cos # cos & cos c

which can indeed be otherwise deduced, by the known formulse of spherical

trigonometry.
(14.)

We see,

then, that according as the

the half-sides, of a given or

or equal to unity, or

triangle,

squares of the cosines of

DEF,

is

greater than unity,

than unity, the sides of the inscribed and bisecting

less

angle, ABC, are together

assumed spherical

sum of the

less

tri

than quadrants, or together equal to quadrants, or

together greater than quadrants.

Conversely, if the sides of a given spherical triangle ABC be thus all

(15.)

or all greater than quadrants, a triangle DEF, but only one* such triangle,

less,

can be exscribed to

so as to

it,

have

its

sides bisected, as

above

process being to let fall a perpendicular, such as CQ in fig. 68,

&c.

and then

draw new

to

through

arcs,

the simplest

from

on AB,

c, &o., perpendicular to these

perpendiculars, and therefore coinciding in position with the sought sides

DE, &c., of DEF.
(16.)

The

trigonometrical results of recent sub-articles, especially as regards

the area-f of a spherical triangle, are probably all well known, as certainly
some of them are but they are here brought forward only in connexion with
;

quaternion formula?

present subject,

mentioned, wherein
is

important

XIX.

one of that

as

includes the
a, /3,

class,

which

is

not irrelevant to the

formula 294, LIIL, the following may be

three vectors, but the order of the factors

7 denote any

a
.

and

and

(a)3y)

* In the next
Article,

= 2
we

+ a 2 03 7

j3V

2
)

j3 (a

2
2
2 4 7 Saj3S/3 7
7 ) + 7 (aj3)

.:|:

shall consider a case of indeterminateness, or of the existence of indefi

namely, when the sides of ABC are all equal to quadrants.

be taken of referring to an interesting Note, to pages 96, 97 of Luby s
Trigonometry (Dublin, 1852) in which an elegant construction, connected with the area of a spherical
triangle, is acknowledged as having been mentioned to Dr. Luby, by a since deceased and lamented
A construction nearly the same, described in the
friend, the Rev. William Digby Sadleir, F.T.C.D.
nitely

many

exscribed triangles DEF

t This opportunity

may

sub-articles to 297, was suggested to the present writer

+
[Using the relation VatfySa/Sy = a~\Py$l3y +
easily follows

on squaring (V

+ S) a&y

by quaternions, several years ago.

2
2
7 Vaj8Soj3 + Vya (Sya + 2Saj3S^7),
or multiply XIX. by 2 and put a/8 = r, &y = p, and 0~ay =

this
rp. J

ELEMENTS OF QUATERNIONS.

376

And

(17.)

we suppose that a,
by o- the

as in 297, (1.), &c.,

if,

vectors, OA, OB, oc,

and denote,

ABC, the principle expressed

[m.

vv
XX.

a +

P
which admits

o-

+ a

stated

under

sm a

7 + a

a+/3

unit-

of several verifications.

We

(18.)

Qy + ~afi+y = cos

7.

area of the triangle

by the recent formula XIII. may be

+ 7

-~

7 are three

|3,

as in 297, (47.),

apparently different, but essentially equivalent form

this

x>

for instance, transform

may,

XXI

K(a +
+

+ 7)

)3)(j3

m
+m
+

(7

+ a tan ^

cos

- a tan -

cos

&
(T\

(cr

^
&

<j

- + a sm .

cos

- a sm -

<*

o"

+ a sin H

pr

LXYII.)

+2e

a)

+ n + ea
+ M - a

as follows (comp. 297,

it

cos

o-

+ a

sin a, as above.

^/

This seems to be a natural place for observing (comp. (16.) ), that if

a, j3, 7, S be 5^ /owr vectors, the lately cited equation 294, LIIL, and the
square of the equation 294, XV., with 8 written in it instead of p conduct
(19.)

and symmetric formula

easily to the following very general

XXII.

a 3

22 S 2

287080)88)38878

7 (SaS)
2

+ a 2 g (S/3 7 )

[Since

/jBU
"

/8

(j3

+ 7

7) bisects the angle

V"

(5)

~
This

is

/3

?+^
+

j8)

)3

between

^therefore

S (S 7 a)

j8

+ 72 S 2

(jB

7)

a direct transformation from 297,

(Sa)3) .t

and 7,

+
_--

/\l//3\i/7U
y

/+

a)\/3 +

LXXVI.

to

(a

0)

(-)

(-)

U (a

)3

7+
7

g +

a+/3*7+

XX.]

aa +
+ cy + d$ =
by
f [This may perhaps be more rapidly derived by operating on
from the four
determinant
of
a
form
the
d
in
and
and
and S5.,
eliminating a, b, c,
Set., S/3., 87.,
t>&

results of operation,]

CASE OF SIDES EQUAL TO QUADRANTS.

AHTS. 298, 299.]

then we take any spherical quadrilateral ABCD, and write

(20.) If

XXIII.
treating

= cos AD = - SUaS,

377

= cos BD = - SU/3S,

ri

= cos cu = SUyS,

7 as the unit- vectors of the points A, B, c, and /, m, n

we have the equation,

a, /3,

as the

cosines of the arcs BC, CA, AB, as in 297, (1.),

XXIV.

+ Pf* + m*m

2mnm

+ nV* +

n + 2nln

2/mV + 2mn ?+ 2nl m

+ 2lmn

which can be confirmed by elementary considerations,* but

merely as an interpretation of the quaternion formula XXII.

is

here given

In squaring the lately cited equation 294, XV., we have used the

(21.)

two following formulae of transformation (comp. 204, XXII., and 210,

XVIII.), in which a, (3, y may be any three vectors, and which are often found
to be useful

XXV.

2
.

299.

(Va]3)

The

ABC, are all

(Saj3)

- a2 2
/3

XXVI.

S (V/3y Vya) = 7 2 Saj3 - S/3ySya.

.

two cases, for which the three sides

less,

two foregoing

a, b, c,

of the given triangle

or all greater, than quadrants, having been considered in the

Articles, with a reduction, in 298, (10.)

other cases to these,

it

and

(11.), of certain

only remains to consider that third principal

which the

case, for

or to inquire
sides of that given triangle are all equal to quadrants
on our general principles, the Fourth Proportional to Three Rectangular
And we shall find, not only that this fourth proportional is not itself
Vectors.

what

is,

a Vector, but that

zero

it

does not even contain any vector part (292) different from

although, as being found to be equal to a Scalar,

(131, 276) in the general conception of a Quaternion.
:

In

(1.)

fact, if

suppose, in 297,

m = 0,

n=

IIL..S/3 7 = S ya =

Saj3

I.

or

we

0,

0,

= 0,

(1.),

it is still

included

that

or that

while

II.

IV.

a= b=

Ta =

T|3

-,

T7 =

1,

the formulae 297, VII. give,

V...S =

0,

0,

= 0;

but these are the vector parts of the three pairs of fourth proportionals to the
*

formula equivalent to this last equation of seventeen terms, connecting the six cosines of the
join, two by two, the corners of a spherical quadrilateral ABCD, is given at page 407 of

which

arcs

Carnot

Geometrie de Position (Paris, 1803).

HAMILTON

ELEMHNTS OF QUATERNIONS.

ELEMENTS OE QUATERNIONS.

378

three rectangular unit-lines, a,

evanescence of vector parts

)3,

7,

taken in

[III.

all possible

must evidently take

orders

7, 8.

i.

and the same

the three given lines be

place, if

only at right angles to each other, without being equally long.

Continuing, however, for simplicity, to suppose that they are unit

(2.)

lines, and that the rotation round a from j3 to j is negative, as before, we

= 1, in 297,
and that thus the six
see that we have now r = 0, and
(3.)
;

fourth proportionals reduce themselves to their scalar parts,

In

positive or negative unity.

ditions, the values

VI.

jSeT

manner we

this

(here) to

under the supposed con

find,

y/B^a

= ay-

ft

=+

VI

l
.

ya~

f3

aftr

For example (comp. 295) we have, by the laws

(3.)

namely

y = fty~ a =
l

(182) of

j, k,

i,

the

values,

VII.

In

tico

ij-

k =jk- l i =

ki-

j= +

VII

the two fourth proportionals,

fact,

the

l
.

ternary products,

ijk

and -

ij-

kji,

kj~ i

ik~

j =ji~ k = l

1.

k and kj~ l i, are respectively equal to

and therefore to +

laws included in the Fundamental Formula

and -

1,

by the

(183).

To connect this important result with the constructions of the two last
Articles, we may observe that when we seek, on the general plan of 298, (15.),
(4.)

to exscribe a spherical triangle, DBF, to a given tri-quadrantal (or tri-rectangular]

triangle, ABC, as for instance to the triangle

manner
old,

UK

that the sides of the new triangle shall be

the problem

is

found

to

admit of

indefinitely

in
(or JIK) of 181,

bisected

many

such a

by the corners of the

solutions.

Any point

may be assumed, in the interior of the given triangle ABC and then, if its
so that
reflexions D, E, F be taken, with respect to the three sides, a, b, c,
;

(comp.

fig.

72) the arcs PD, PE, PF are perpendicularly

bisected

by those three

FD, DE

will be bisected

quired

sides, the three other arcs EF,

by the

points A, B,

c,

as re

because the arcs AE, AF have each the same

length as AP, and the angles subtended at A by PE and

PF are together equal to two right angles, &c.
(5.)

The

positions of the auxiliary points, D, E, F,

are therefore, in the present case, indeterminate, or variable

the angles at those three points is constant,

because,

by the

and equal

but the sum of

to

four right angles",

on PD, PE, PF as bases, that sum of the
the sum of the angles subtended by the sides

six isosceles triangles

three angles D, E, F

is

equal to

of the given triangle ABC,

at the

assumed

interior point p.

The

spherical

OTHER VIEW OF A FOURTH PROPORTIONAL.

ARTS. 299, 300.]

DBF

excess of the triangle

is

therefore equal to two right angles,

379

and

its

area

S TT as may be otherwise seen from the same figure 72, and might have
been inferred from the formula 297, LY., or LYI.
=

(6.)

The

radius OD, in the formula 297,

for the fourth propor

XLYIL,

tional /Ba- y, becomes therefore, in the present ease, indeterminate ; but because
the angle L DR, or J(?r - S), in the same equation, vanishes, the formula

becomes simply

jScf

y =

VI.

as in the recent equations

1,

other examples, of the class here considered.

(7.) The conclusion, that the Fourth Proportional
Lines

is

a Scalar,

to

Three Rectangular

ways be deduced, from the principles

For example, with the recent suppositions, we may

in several other

may

Book.

of the present

and similarly in

write,

VIII.

VIIF.

1
.

jSeT
1

ya

= -

= +

7,

7J31

a/3-

j3,

= = +

= -

ay

a,

fly

y,

/3

= + a

the three fourth proportionals VI. are therefore equal, respectively, to - a 2 - 2 and
consequently to + 1 while the corresponding expressions VI
/3
2
2
are equal to + |3 + 7 + a2 , and therefore to - 1.

7",

Or (comp. (3.)
282, XXI.),
IX.
(8.)

in which the factor cf

we may

l
.

fla

is

y =

write generally the transformation (comp.

a-

2
.

flay,

cT

if

a2 ,

always a scalar, whatever vector a

may

the vector part of the ternary product flay vanishes, by 294, III.,
recent conditions ofrectangularity III. are satisfied.

be

while

when

the

Conversely, this ternary product flay, and this fourth proportional

y, can never reduce themselves to scalar s, unless the three vectors a, ft, 7

(9.)
l

fla~

(supposed to be

all actual (Art. 1)

are perpendicular each to each.

SECTION

On an

8.

equivalent Interpretation of the Fourth Proportional to

Vector*,, deduced from the Principles of the

Three Diplanar
Second Book.

300. In the foregoing section, we naturally employed the results of pre

ceding sections of the present Book, to assist ourselves in attaching a definite
signification to the

and

Fourth Proportional (297) to Three Diplauar Vectors

l

thus, in order to interpret the symbol fla~ y,

we

availed ourselves of the

interpretations previously obtained, in this Third Book, of a

as a

line,
3

and

of

ELEMENTS OF QUATERNIONS.

380

But

aj3y as quaternions.

a)3,

i.

8.

be interesting, and not uninstructive, to

may

it

[III.

inquire hoiv the equivalent symbol,

I.

(y

a)

or

y,

with y not

y,

\\\

a, /3,

might have been interpreted, on the principles of the Second Book, without at first
assuming as known, or even seeking to discover, any interpretation of the three

mentioned symbols,

lately

II.

It will be

where

297

while u

Proportional

and OP

is

the same vector, and

is
;

a-

e is

as a

employed

(/3

aj3y.

a/3,

found that the inquiry conducts

III.

to

a)

to

an expression of the form,

7 = 8 + eu

the same scalar, as in the recent sub-articles

temporary symbol, to denote a certain Fourth

Three Rectangular Unit Lines, namely, to the three lines OQ,

to

68

with reference to the construction represented

by that figure, we should be led, by the principles of the Second Book, to
oi/,

in

fig.

write the equation

IY.

And when we

;* so that,

(OB

OA)

oc = OD

cos

JS +

(oi/

OQ)

OP

sin

JS.

proceed to consider what signification should be attached, on

same Second Book, to that particular fourth proportional,

here the coefficient of sin ^S, and has been provisionally denoted by u,
find that although it may be regarded as being in one seme a Line, or at

the principles of the

which

we

is

least homogeneous with a line, yet it

rather analogous, in Geometry,

to the

must not be equated to any Vector

Scalar Unit of Algebra, so that it

being
may be
:

naturally and conveniently denoted by the usual symbol 1, or + 1, or be equated

to Positive Unity.
But when we thus write u = 1, the last term of the formula
III. or IV., of the present Article, becomes simply
this term (or part) of the result

e,

or sin

J2

and while

comes to be considered as a species

of

Geome

the complete Expression for the General Fourth Proportional to

Three Liplanar Vectors takes the Form of a Geometrical Quaternion and thus

trical Scalar,

the formula 297,

in

XLVIL,

for the present,

it,

here the recent

(/3

symbols

or 298, VIII.,
a)

7 for

is

f3a~ j, to

reproduced, at least if

we

substitute

avoid the necessity of interpreting

II.

u ln the abstract
published in the Proceedings (Eoyal Irish Academy, November llth, 1844),
and the sought fourth
South, West, Up were used at first, instead of the symbols i, j, k
Preface to
proportional to jik, which is here denoted by u, was called provisionally, Forward.
*

ihe words

"

Lectures, p. (54).]

SCALAR UNIT IN SPACE.

ART. 300.]

The

(1.)

construction of

c,

&c., as before,

V.
VI.

OB

OA = OR

OQ = cos c +

oc = OQ cos p + OP
.

for the multiplication of quaternions (as

VIII.

(OB

OA)

to hold

good

does for multiplication in algebra)

it

may

OQ)

OP

oc =

+ OP cos
it

snip.

u = (oi/

the quadrinomial expression

in which

OQ) sin c

for abridgment,

VII.

we have

which has been proved (212)

pretation, that distributive property

and writing

(oi/

as is natural, for the purposes of the sought inter

Admitting then,

(2.)

68 being retained, but no principles peculiar

write, with the same significa

fig.

Book being employed, we may

to the Third
tions of

381

oi/. sin c

sinp + u

cosp + OQ cos
.

sin c sin p

cosp

be observed that the sum of the squares of the four

coefficients

of the three rectangular unit-vectors, OQ, oi/, OP, and of their fourth proportional,
u, is

equal to unity.

But the

(3.)

coefficient of this fourth proportional,

as a species of fourth unit,

IX.

which

sin c

sin^ = sin

MN =

sin

JS =

we must therefore expect to find that the three other

when divided by cos JS, or by r, give quotients which

coefficients in

Q,

X.

or that a point

sin c

(4.)

cosp =

Accordingly

when we substitute D
XI.

upon the unit-sphere, from the

x can be assigned, for which

r cos L

oi/. sin c GOSJS

cos c cosp

= r cos QX

found that these three

it is

for

+ OQ cos
.

cosp + OP cos
.

three points

cos PX.

equations are satisfied,

have the transformation,

last

and therefore that we

cos c sinp

VIII.,

are the cosines of the

arcual distances of some point x

be regarded

may

is

c sin

/;

= OD cos JS =
.

S,

whence follow the equations IV. and III. and it only remains to study and
interpret the fourth unit, u, which enters as a factor into the remaining part
;

of the quadrinomial expression VIII., without

those of the Second Book

284, of

/3a,

&c.

employ iog any principles except

and therefore without using the Interpretations 278,

ELEMENTS OF QUATERNIONS.

382

301. In general, when two sets of three vectors,

connected by the relation,

it is

)3

i.

8.

are

/3

~
a

and a,

a, )3, 7,

or

[III.

natural to write this other equation,

M
m...-y=
7 /;
TTT

and

f"*

to say that these tivo fourth proportionals (297), to a,

are equal

each other

to

whatever the full

7,

]3,

and

to a

two

signification of each of these

last

moment to be not yet fully known, may be after

short, we may propose to make it a condition of the

symbols III., supposed for the

In

wards found to be.

sought Interpretation, on the principles of the Second Book, of the phrase,

Fourth Proportional

"

and

Book, and
of

two equivalent Symbols 300,

of either of the

III. shall folloic

to three Vectors,

from

I.

or II.

just as, at the

that the recent Equation

I.,

commencement

of that

Second

before concluding (112) that the general Geometric Quotient

any two

lines in space is

a Quaternion,

we made

a condition (103) of the

it

interpretation of such a quotient, that the equation

|3

(j3

a)

a =

j3

should be

satisfied.

302. There are however two

tion III.

tests

must be submitted, before

(comp. 287), to which the recent equa

adoption in order that we may

its final

with the previous interpretation (226) of a Fourth

to
Three
Complanar Vectors, as a Line in their common plane
Proportional
and Ilnd, with the general principle of all mathematical language (105), that
be sure of

its consistency, 1st,

things equal to the same thing, are to be considered as equal to each other.
And it is found, on trial, that both these tests are borne so that they form no
:

objection to our adopting the equation 301, III., as true by definition, whenever

the preceding equation II., or

(1.)

line S,

have by

It

may happen
226

as in

I., is satisfied.

that the

first

namely, when

member

a, |3,

of that equation III.

7 are complanar.

In

is

equal to

this case,

we

II. the equation,

TTT

1 V ...

7
so that a,

/3

their fourth

7 are

7 =

7 7
also

/3

or

TTT/

IV.

^
7
a
/3"

3
a

7;

complanar (among themselves), and the

proportional likewise

and the equation III.

line

is satisfied,

is

both

NEW

TESTS FOR

ARTS. 301-303.]

members being symbols

common

for one

the intersection of the two planes,

a/By

happen

INTERPRETATION.
line,

and a

which

fi

383
general situated in

is in

although those planes

may

to coincide, without disturbing the truth of the equation.

Again, for the more general case of diplanarity of a, )3, 7, we may

conceive that the equation* II. co-exists with this other of the same form,
(2.)

a y
if

a
If then that definition be consistent with

the definition 301 be adopted.

we ought

general principles of equality,

to find,

and VI., that

III.

by

third equation between two fourth proportionals holds good

VIII.

or that

VII...V=Cy";
a
a

4, =

this

C,

II. and V. are satisfied.

And accordingly, those two
the
of
the
Second Book, respecting
equations give, by
general principles
quaternions considered as quotients of vectors, the transformation,

when the equations

/3

T7

7 =

ay

303. It

-^77

/3

-,.

7
]3
-^ = -

7 =

-17

ay

/3"

^77-,

as required.

then permitted to interpret the equation 301, III., on the prin

Second Book, as being simply a t ran.ifor mat ion (as it is in

is

ciples of the

algebra) of the immediately preceding equation II., or

I.

and therefore

to

write, generally,

I...qy =
where

y,

7 are any two

this last condition.

if v

q (y

/) = q

and

q,

q are any

and

be any two right quaternions, we have

vectors,

Now,

II.

if

tivo

quaternions,

which satisfy

(by 193, comp. 283) the equation,

III.

10

L? =

= 00 M

or

IV.

v~ (Iv

Iv

=
)

-1

whence

V.

tr

1
.

Iv = v -\ Iv

It follows, then, that

principle which has just been enunciated.
l
a right Line (Iv) be multiplied
by the Reciprocal (v~ ) of the Right Quaternion

by the

of which

it is

the Index, the

Product

(v~ Iv) is

is

(1.).

(v),

always supposed to be the one which

it may have been in connexion

before, although perhaps

Compare 217,

if

independent of the Length, and of

* In this and other

cases of reference, the numeral cited
(with the same number) has last occurred

with a shortly preceding Article.

"

ELEMENTS OF QUATERNIONS.

384
the Direction, of the

Line thus operated on

or constant value

8.

i.

Pro

this

Value, for all possible Lines (a) in Space: which

common

duct has one

other words, that

or, in

";

[III.

common

be regarded as a kind of new Geometrical Unit, and

may

we have

lately denoted, in 300, III., and VII., by the

u
because, in the last cited formula, the line OP is the
temporary symbol
index of the right quotient OQ OL
Retaining, then, for the moment, this
is

equal to what

symbol,

we

u,

quaternion,

v,

line

have, for every

a in space, considered as the index of a right

the four equations:

VI.

tr

o = u

VII.

IX.
in

which

understood that a =

it is

a = vu

= u

v~

TV,

VIII.
a

= a

and the three

last are

here regarded as

deduced and interpreted as

being merely transformations of the first, which is

above.
And hence it is easy to infer, that for any given system of three

X.
where the

XL

we have

lines a, |3, 7,

angular

(]3

7 = xu,

a)

the general expression

scalar co-efficient, x, of the

x =

(T|3

This coefficient x

is

Ta)

T7

a JL

if

new

unit, u, is

_L 7,

J- a

determined by the equation,

XII.

according as

/3

|3,

UT =

Ax

other words, according as the rotation round the first line, from
the third, is on the contrary negative or positive (compare 294, (3)
(1.)

In

about,

from which

we may

by

line 7,

or in

).

which has

u, while the system of the three rectangular

conceived to be derived

it is

it is

the second to

illustration of the constancy of that fourth proportional

been, for the present, denoted

unit-lines

j3).

but

the first line a, is itself positive or negative

to

j3,

according as the rotation round the third

positively or negatively taken,

from

(a

therefore always equal, in magnitude (or absolute quantity),

to the fourth proportional to the lengths of the three given lines oj3y

the second line

rect

is

any manner turned

in

observe that the three equations, or proportions,

XIII.

/3

a = a

j3

= 7

/3,

conduct immediately to this fourth equation of the same kind,

XIV.
if

we admit

that this

new

a = 7

j3,

u = (7

quantity, or symbol,

* In
equations of this form, the parentheses
are here retained.

or*

may

u, is

]3)

to be operated on at

all,

be omitted, though for greater clearness they

FOURTH PROPORTIONAL RESUMED.

ABTS. 303, 304.]

385

or combined with other symbols, according to the general rules of vectors

and

quaternions.
It

(2.)

is,

then, permitted to change the three letters

permutation, to the three other letters,

by a

a, |3, 7,

cyclical

7, a (considered again as represent

|3,

without altering the value of fas fourth proportional, u\ or in

other words, it is allowed to make the system of the three rectangular lines

ing

unit-lines),

through the third part of four right angles, round the interior and co-

revolve,
initial

diagonal of the unit-cube, of which they are three co-initial edges.

(3.)

And

it is still

more evident, that no such change

we merely cause the system

place,
any angle, in its own plane, round the
merely substitute, for the factor

j3

third line as an axis

round an origin supposed to be

(4.)

as

And

a. 7

since thus

another factor equal thereto.

a,

combining these two last modes of rotation,

ever,

of value will take

of the two first lines to revolve, through

if

we can represent any

we shall
But by

rotation ichat-

fixed.

as regards the scalar ratio of

any one fourth proportional, such

kind here considered, such as )3 a 7, or u, it

any
is sufficient to
without
suggest that,
any real change in the former, we are
allowed to suppose it to be so prepared, that we shall have
/3

other, of the

to

XV..
x being some scalar

=;

/-*/;

=/3;

and representing the

coefficient,

ratio required.

304. In the more general case, when the three

given lines are not rect
angular, nor unit-lines, we may on similar principles determine their fourth
proportional, without referring to
real loss of generality,

perpendicular to each other

sary, for the quotient

]3

a,

ft

and 7 are

to each other

:a),

by 203,

IV.

let

..8"

I.

and

ELEMENTS OF QUATERNIONS.

II.

j3"

and

II.,

= y-.a,
a

HAMILTON

[p. 360], as follows.

that the planes of a,

comes merely

since this

parallel to a, but

so that,

68

we

j3

7"

/3

|3",

Having thus
7 = 7 +

are perpendicular to

shall

7"

Without any
/3 and a, 7 are

to substituting, if neces

another quotient equal thereto.

L..Ax.(j3:a)J-Ax.( 7
where

fig.

we may suppose

7",

it,

have the expressions,

V 2a

a.

and

ELEMENTS OF QUATERNIONS.

386

We may

[III.

then deduce, by the distributive principle (300,

8.

i.

the tranfor-

(2.) ),

mations,

7 +

/3

ft"

j +

7 = o + #W

0"
-

-f

where
VI.

2+
7

S = I3S

&

"S

The

= TS
+
a

/3"S

what we have

latter part, xu, is

and

VII.

$- 7
a

xu =

".

called (300) the (geometrically) scalar

part, of the sought fourth proportional ; while the former part

may (still)
be called its vector part and we see that this part is represented by a line,
:

which

at once in the two planes, of

is

j3,

7",

and of

7,

two planes

or in

|3";

may be generally constructed as follows, without now assuming that the

Let 7 be the projection of the
planes a/3 and ay are rectangular, as in I.

which

line
)3

7 on the plane of

a, |3,

and operate on this projection by the quotient

which is drawn through the line |3 a 7 so

a as a multiplier; the plane

obtained, at right angles to the plane

and the plane through

7,

And

locus for that line.

and the magnitude

VIII.

which

aj3, is

one locus for the sought line 8

perpendicular to the plane

is

TS

77

is

another

as regards the length of this line, or vector part S,

(or quantity) of the scalar part xu,

cos

== t

and

s,

IX.

it

x =

easy to prove that

is

sin

s,

where

X.
if c

= T/3

Ta T 7
.

XI.

and

denote the angle between the two given lines

to their

of the third given line

sin

a, j3,

sin c

smp,

and j9 the inclination

the sign of the scalar

plane
7
being positive or negative, according as the rotation round
:

coefficient, x,

from

j3

to

is

negative or positive.
(1.)

Comparing the recent construction with

condition

I. is satisfied,

the four unit-lines

tions of the four radii oc, OQ, on,

what

may

(2.)

It

68,

we

see that

US

when

the

take the direc

Uy, Ua, Uj3,

which terminate at the four corners of

be called a tri-rectangular quadrilateral CQRD on the sphere.

may

to halfihG area

be remarked that the area of this quadrilateral is exactly equal

S of the triangle DEF which may be inferred, either from
;

the circumstance that

by

oi>,

tig.

the angle

MDN

or

its

spherical excess (over four right angles)

is

constructed

from the triangles DBR and KAS being together equal

SPHERICAL PARALLELOGRAM.

ARTS. 304, 305.]

DESK

to the triangle ABF, so that the area of

is

is

387

and therefore that

2,

oE

CQRD

JS, as before.

The two

(3.)

QR

sides CQ,

of this quadrilateral,

obtuse angle at D, being still called

to c

still

being

called

denoted by d

let

p and

d=

cos

which

side CD

JTT),

cos^j cos c

is

opposite

p be

now

and

let s

or the area of the quadrilateral.

We

DR which

opposite to

is

QD be denoted by d and d

also let the diagonals OR,

and the

c,

the side

denote the spherical excess (CDR shall then have the relations,

which are remote from the

= cos p cos

cos d

tan;/= cos

f
;

XII.

tan

(cos s

of which

= cosp tan
=

cos^? sec^/

= cos c

sec c

some have virtually occurred before, and

tan j?
cos

d sec c?

all are easily

proved by

right-angled triangles, arcs being when necessary prolonged.

(4.) If we take now two points, A and B, on the side QR, which satisfy the
arcual equation (comp. 297, XL., and
68),
fig.

XIII.

and

DD

if

we then

it is

join AC,

and

let fall

on this new arc the perpendiculars BB

D of the side BD on the arc AC is
,

easy to prove that the projection B

equal to that arc, and that the angle

new

AB = nQRJ

DBB

is

right

so that

we have

the two

equations,

XIV.. .AB D = nAC;

and the new quadrilateral BB D D

Hence

is

XV.. DBB =
.

JTT

also tri-rectangular.

the point D

(5.)
may be derived from the three points ABC, by any two
of the four following conditions
1st, the equality XIII. of the arcs AB, QR ;
Ilnd, the corresponding equality XIV. of the arcs AC, B D
Illrd, the
:

tri-rectangular character of the quadrilateral

CQRD

IVth, the corresponding

character of BB D D.
(6.)

In other words,

this derived point

is

the common intersection of the

BB

erected at the four points

four perpendiculars,
,
four
D
B
, c,
R,
CQ, BB being still the perpendiculars from c and B, on AB and AC ;
and R and D being deduced from Q and B by equal arcs, as above.
305. These
consequences of the construction employed in 297, &c., are
to the

arcs AB, AC, CQ,

here mentioned
merely in connexion with that theory of fourth proportionals
to vectors, which
they have thus served to illustrate but they are perhaps
;

3D2

ELEMENTS OE QUATERNIONS.

388

numerous and interesting enough,

"

to justify us in

[III.

suggesting the name,

Spherical Parallelogram,"* for the quadrilateral CABD, or BACD, in

(or 67), p.

360; and in proposing to say that D

when

completes such a parallelogram,

given

upon the sphere,

as first, second,

fully observed, that the analogy

the general

case,

when

68

c, A, B,

or B, A,

c,

are

must however be care

It

third.

fig.

the Fourth Point, which

is

the three points

and

8.

i.

here thus far imperfect, that in

plane
the three given points are not co-arcual, but on the con
to the

is

trary are corners of a spherical triangle ABC, then

for the three first points of a

if

we

take

c,

D, B, or B, D, c,

new spherical parallelogram, of the kind here con

new fourth point, say AI, will not coincide with the old second point
A; although it will very nearly do so, if the sides of the triangle ABC be small:
the deviation AAi being in fact found to be small of the third order, if those
sidered, the

sides of the given triangle be supposed to be small of the first order

being always directed towards

on BC.

To

the foot of the perpendicular, let fall

and

from A

be still any two given

/3, 7
with
other
an
each
unit-vectors, OB, oc, making
angle equal to a, of which
the cosine is /; and let p or OP be any third vector.
Then, if we write,
(].)

investigate the law of this deviation, let

I.

the

new

pi

(p)

or derived vector

JNp

parallelogram.

We

0p or

and BQCQI

7 +

the two fourth proportionals, to

of the length of

\P

p, ]3, 7,

shall also

and

distributive

III.

0p

HI 15,

= Upi,

j3,

common

vector part of

multiplied by the square

lately called a spherical

have the transformation (compare 297,

7,

(2.) ),

will be such that

symbol of operation
.

to p, 7,

be what we have

and the

C-Q!

or OPI, will be the

pi,

will

OQ = Up,

13 J,

and
tf>

p=p,

if

/o|||

but

IV...0p = -/p,
(2.)

This being understood,

if

p|lAx.( 7

:/3).

let

J3y the same analogy, the quadrilateral CQKD, in fig. 68, may be called a Spherical Rectangle.
= 7 and
= #. So $ (yft + zy) = zfr + yy.]
t [In fact
4>

<>

SERIES OF SPHERICAL PARALLELOGRAMS.

ART. 305.]
so that

p, or OP

is

the projection of p on the plane of |3y

the part (or component) of p, which

we

shall

tivo

such

have an indefinite

(p,

p\

p"

pi

PI,

or

OP",

pi,

p +

(j>*p

p2

p.>,

is

Then

or rather

l*p" ;

&o.

ttvo

which each

geometrical progression, in

by the constant factor

(3.)

p"

being thus ranged, alternately,

pp
and
which
are let fall from the points p
Pip b
perpendiculars,
on the given plane BOC; and the intervals, PP 2 PiP 3 P 2 P4,
forming a
,

and

the two series of derived points, p 1? p 2 P 3 P 4

on the

vectors,

succeeding each other alternately, as follows

series,

and

perpendicular to that plane.

is

of derived

series

389

by the

or

1,

If then this angle be

is

equal to the

multiplied

it,

negative of the cosine of the given angle BOC.

still

supposed

from the intermediate value

also in general

one before

to

|TT,

be distinct from

we

shall

and

have the two

TT,

and

limiting

values,

VII.

p*n

p zn +i

f/i,

n =

if

oo

or in words, the derived points p 2 P 4

the other derived points,

of odd orders, tend to the other point P

limiting positions

p3

PI,

of even orders, tend to the point p

and
I,

as

these two limit points being the feet of the two (rectilinear)

from p and P on the plane BOC.

deviation pp 2 is small of the third order,

perpendiculars, let fall (as above)

(4.)

But even the first

the length

if

OP be considered as neither large nor small, and if the sides of

the spherical triangle BQC be small of the first order.
For we have by VI. the

Tp

of the line

following expressions for that deviation,

VIII.

pp 2 = p 3 - p = (P -

ff

1)

= - sin 8 *

where p a denotes the

smp a Tp
,

Up"

inclination of the line

p to the plane j3y or the arcual
The
perpendicular from the jpowrf Q on the side BC, or a, of the triangle.
statements lately made (305) are therefore
been
to
have
correct.
proved
(5.)

And

may

we now resume and extend

is

BA 2 c, &c., through an indefinite

fourth point of

and

the spherical construction, and con

deduced from BAiC, as AI was from BDC, or D from BAC while

be supposed to be deduced by the same rule from BDjC, and D 2 from

ceive that D!

A2

if

series

of spherical parallelograms, in which the

any one is treated as the second point of the next, while the first
remain constant we see that the points A I? A 2
are all

third points

situated on the arcual perpendicular let fall from A on

BC; and that in like
manner ihe points D 1} D 2
are all situated on that other arcual perpendicular,
,

ELEMENTS OF QUATERNIONS.

390

which
DW

is let fall

from D on

We

BC.

coincide precisely with the feet

theorem, which

see also that the ultimate positions,

of those tico perpendiculars

would perhaps be

it

[III.

difficult to prove,

8.

A X and

a remarkable

any other

by

i.

method

than that of the Quaternions, at least with calculations so simple as those

which have been employed above.
be remarked that the construction of

68 might have been

otherwise suggested (comp. 223, IV.), by the principles of the Second Book,
(6.)

if

It

may

we had sought

fig.

to assign the fourth proportional (297) to three right quater

for example, to three right versors, v, v ,

whereof the unit lines a, j3,
be
to
the
axes.
For
should
be
the
result
would be in general a
supposed
y
l
v
with
e
for
its
scalar
with
v~
and
for
the index of its right
S
quaternion
part,
nions

v"

v",

part

and

denoting the same

and the same

scalar,

vector, as in the sub-

articles to 297.

306. Quaternions

which

also be

may

shall complete (comp. 305, (5.)

a new construction,

to furnish

employed

the graphical determination of the two

series of derived points,

I.

when the

A2

D, AI, DI,

three points A, B, c are given

D2

&c.,

upon the unit-sphere

and thus

shall

of
visible (so to speak), with the help of a new
figure, the tendencies
those derived points to approach, alternately and indefinitely, to the feet, say
D and A , of the two arcual perpendiculars let fall from the two opposite corners,

render

D and

A, of the first spherical parallelogram, BACD,

which diagonal

(as

we have

seen)

is

common

on

its

given diagonal BC

to all the successive paral

lelograms.
(1.)

The given

triangle ABC being supposed for simplicity to have

its sides

abc less than quadrants, as in 297, so that their cosines linn are positive, let

c be the feet of the perpendiculars let fall

points A, B, c

also let

M and N

on these three sides from the

be two auxiliary points, determined by the

equations,
II.

so that the arcs

dicular

ND on

AN and BC

BM =

n MC,

AM =

bisect each other in M.

MN
Let

fall

from N a perpen

BC, so that

IIL..A B D = nA c;
and

let

B",

c"

be two other auxiliary points, on the sides

which satisfy these two other equations,

sides prolonged,

IY.

= n AC,

n c

c"

= ^ AB.

and

c,

or on those

CONSTRUCTION OF THE SERIES.

ARTS. 305, 306.]

Then

(2.)

last points,

the perpendiculars to these last

B"

and

sides,

391

CA and AB, erected at these

will intersect each other in the point D,

c",

and the foot of

(305) the spherical parallelogram BACD;

which completes

the perpendicular from

point D, on the third side BC of the given triangle, will coincide (comp.
305, (2.) ) with the foot D of the perpendicular on the same side from N ; so that
this

this last perpendicular

ND

is

one locus of the point D.

To

obtain another locus for that point, adapted to our present pur
(3.)
E
let
denote
now* that new point in which the two diagonals, AD and BC,
pose,
intersect each other; then because (comp. 297, (2.)

V.

we may
VI.

OD = u (w/3 + ny -

write (comp. 297, (25.), and (30.)

OE = u (m[3 + ny), whence VII.

the diagonal

AD thus dividing the

we have the

expression,

la),

),

sin

BE

sin

EC = n

m = cos BA

which the

arc BC into segments, of

cos

sines are

proportional to the cosines of the adjacent sides of the given triangle, or to the

BA and A

cosines of their projections

adjacent to the
points

A and

The

(4.)

on BC; so that the greater segment

and the middle point M

lesser side,

is

of BC (1.) lies between the

E.

intersection E is therefore a

known

point,

and the great

circle

through A and E is a second known locus for

D which point may therefore be found, as
;

the intersection of the arc AE prolonged,

with the perpendicular

because E

And

ND from N (1.).

leyond the middle point

of BC, with respect to the foot A of the
lies (3.)

perpendicular on BC from A, but (as

it is

easy to prove) not so far beyond M as the

point D or in other words falls between M
,

and D (when the arc BC

posed,
arc

less

is,

as above sup

than a quadrant), the prolonged

AE cuts ND

between

N and D

or in

other words, the perpendicular distance of

the

sought fourth point D,

diagonal BC of the parallelogram,

A,

the given

from

is less

from the same given diagonal.

*

It will be observed that M, N,

present letters c and

c"

Fig. 73.

than the distance of the given second point

(Compare the annexed

fig. 73.)

E have not here the same significations as in

correspond to u and E in that figure.

fig.

68

and that the

ELEMENTS OF QUATERNIONS.

392

III.

8.

i.

(5.) Proceeding next (305) to derive a new point AI from B, D, c, as D has

been derived from B, A, c, we see that we have only to determine a new*

by the equation,

auxiliary point F,

YIII.

and then
which

to

draw DF, and prolong

EM =

it till it

MF

meets A A in the required point

AI,

will thus complete the second parallelogram, BDCAi, with BC (as before) for

a given diagonal.

In

manner, to complete (comp. 305, (5.) ), the third parallelogram,

BA^D!, with the same given diagonal BC, we have only to draw the arc AjE,
and prolong it till it cuts ND in DI after which we should find the point A 2
(6.)

like

of & fourth successive parallelogram BDiCA 2

The

(7.)

by drawing D

and

F,

so

on for ever.

constant and indefinite tendency, of the derived points D, DI

the limit-point B
other limit-point

and
,

A2

of the other (or alternate} derived points AI,

becomes therefore evident from

final (or limiting] results of

which, we may

this

new

by

express

to

to the

construction

the

these two equations

(comp. again 305, (5.)),

IX.

.D. = D

A. = A

But

the small ness (305) of HIQ first deviation AA when the sides of the
given triangle ABC are small, becomes at the same time evident, by means of
(8.)

: ,

the same construction, with the help of the formula VII.

which shows that
the interval^ EM, or the equal interval MF (5), is small of the third order, when
;

the sides of the given triangle are supposed to be small of the first order
agreeing thus with the equation 305, YIII.

The theory

(9.)

of such spherical parallelograms admits of

applications, especially in connexion with spherical conies

some interesting

on which however

we cannot

enter here, beyond the mere enunciation of a Theorem, } of which

(comp. 271) the proof by quaternions is easy
:

* This new
point, and the intersection of the perpendiculars of the given triangle, are evidently
not the same in the ne\v figure 73, as the points denoted hy the same letters, F and P, in the former
are conceived to bear to each other the same relations
figure 68
although the four points A, B, c,
in the two figures, and indeed in
BACD being, in that figure also, what we have proposed
fig. 67 also
;

r>

to call a spherical parallelogram.

Compare the Note

to (3.).

t The formula VII. gives easily the relation,

VIP.

tan

EM =

tan

MA

a\
tan -

2
;

hence the interval EM

is

small of the third order, in the case

(8.)

here supposed

and generally,

if

7T

<

-, as in (1.), while b

and

c are

unequal, the formula shows that

or than D M, so that E falls between

M and B

as in

this interval

EM

is less

than MA

(4.).

J This Theorem was communicated to the Royal Irish Academy in June, 1845, as a consequence
of the principles of Quaternions.
See the Proceedings of that date (Vol. III., page 109).

SPHERICAL QUADRILATERALS.

ART. 306.]

any spherical quadrilateral, and

"7/KLMN be

also

we complete

any point on

the sphere;

if

the spherical parallelograms,

X.
and determine

393

the poles

LIMB,

KILA,

MING,

NIKD,

KM and LN of the quadrilateral :

of a spherical conic, inscribed in the derived

K and F of the diagonals

then these two poles are the foci*

quadrilateral ABCD, or touching

its

four

sides."

Hence, in a notation J elsewhere proposed, we

these conditions of construction, the formula
(10.)

shall have,

under

XI.
(11.)

EF

Before closing this

XI

or

ABCD;

(. .)

article

and

EF

BCDA; &c.

(. .)

section,

seems not irrelevant to

it

remark, that the projection 7 of the unit-vector j, on the plane of a and

is

given by the formula,

VTT

XII.

a sin a cos B + 8 sin b cos

-r-r-

,
7 =

sine

]3,

A
;

and that therefore the point p, in which (see again fig. 73) the three arcual
perpendiculars of the triangle ABC intersect, is on the vector,

XIII.
(12.)

It

may

= a tan A +

13

tan B + 7 tan

c.

be added, as regards the construction in 305,

(2.),

that the

right lines,

XIV.
however far their

series

nately, in two points s

PP I?

may

and

T,

P^,

be continued,
of

tiro

point P

or assumed.

is given,

namely the
E and F.

real,

points,

intersect the given

plane BOO, alter

are,

fixed points in the plane,

* In the
language of modern geometry, the conic in question

four

and which thus become

space

p3 p 4

which the vectors

-v TT

in

P2 p3,

may

when

the position of the

he said to touch eight given arcs

DA; and four imaginary, namely two from each of the focal

sides AB, uc, cu,

= \r l K, r = /*r A, s = vr p, and t = Ki~ l v

then srq = -vi~ K = Kt, rq = /a.K, sr =
andqt = \v. On reference to fig. 60, p. 304, there is no difficulty in seeing that a conic
having the given foci may he drawn to touch the four sides, produced when necessary.]
t [Take q

t>\,

ts=ien,

Compare the second Note

HAMILTON

to page 310.
ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

394

SECTION

[III.

9.

i.

9.

On a Third Method of

interpreting a Product or Function of

and on the Consistency of the
Vectors as a Quaternion
Results of the Interpretation so obtained, with those which
have heen deduced from the two preceding Methods of the
present Book.
;

307.

The

Conception of the Fourth Proportional

to three

Rectangular UnitUnit in Geometry, is eminently

and offers a Third Method of interpreting

Lines^ as being itself a species of Fourth

characteristic of the present Calculus

a Product of two Vectors as a Quaternion

Book

present

more

however found

is

to be

of being easily extended to products of three or

and admits

lines in space,

which

two former methods (278, 284) of the

consistent, in all its results, with the

and generally

to Functions of Vectors (289).

In

fact

we

have only to conceive* that each proposed vector, a, is divided by the new or
and that the quotient so obtained, which is
u, above alluded to
whereof the vector a is the
always (by 303, VIII.) the right quaternion I

fourth unit,

index, is substituted for that vector

the resulting quaternion being finally,

if

we think it convenient, multiplied into the same fourth unit. For in this way
we shall merely reproduce the process of 284, or 289, although now as a con
sequence of a
tation

different train

of thought, or of a distinct but Consistent Interpre

which thus conducts, by a new Method,

to the

same Rules of Calculation

as before.

The

(1.)

z
equation of the unit-sphere, p

(282,

XIV.), may thus be

conceived to be an abridgment of the following fuller equation

* It was in a somewhat
analogous

way

that Des Cartes showed, in his Geometria (Schooten

Edition, Amsterdam, 1659), that all products and powers of lines, considered relatively to their lengths
alone, and without any referenda to their directions, conld be interpreted as lines, by the suitable

introduction of a line taken for unity, however high the dimension of the product or power might be.
Thus (at page 3 of the cited work) the following remark occurs
:

"Ubi

notandum

concipiam, licet

But

it

est,

illas,

was much

quod per

vel

b-\

similesve, communiter,

non

nisi lineas

omnino simplices

ut nominibus in Algebra usitatis utar, Quadrata aut Cubos, &c. appellem."

more difficult to accomplish the corresponding multiplication of directed lines in

to all other lines,

line, which is symmetrically related
planes (comp. the Note to page 258). The Unit of Vector -Multiplication
cannot properly be itself & Vector, if the conception of the Symmetry of Space is to be retained, and
at least
duly combined with the other elements of the question. This difficulty however disappears,
when we come to consider that new Unit, of a scalar kind (300), which has been above
in

space

or

on account of the non-existence of any such

common

to

all possible

theory,

denoted by the temporary symbol u, and has been obtained, in the foregoing section, as a certain
Fourth Proportional to Three Rectangular Unit- Lines, such as the three co-initial edges, AB, AC, AP of

CONCEPTION OF THE FOURTH UNIT.

AKT. 307.]

the quotient p

395

u being considered as equal (by 303) to the right quaternion,

I^p, which must here be a right versor (154), because

its

square

is

negative

unity.
(2.)

The

equation of the ellipsoid,

T (ip + piO =?-*

(282,

XEL),

be supposed, in like manner, to be abridged from this other equation

may

and similarly in other

cases.

of the sphere and

(3.) We might also write these equations,
under these other, but connected forms

ellipsoid,

III

UH

IV...T(-p +

-cV-K--*
u
u
u
J

\tr

with interpretations which easily offer themselves, on the principles of the

foregoing section.
(4.)

adopt

however, to be distinctly understood, that ice do not propose to

Form of Notation, in the practice of the present Calculus: and that

It

this

we merely

is,

suggest

it,

in passing, as one

which

may

tional light on the Conception, introduced in this

serve to throw

some addi

Third Book, of a Product of

tu O Vectors as a Quaternion.

In general, the Notation of Products, which has been employed

throughout the greater part of the present Book and Chapter, appears to be
(5.)

what

\ve have called an Unit-Cube: for this fourth proportional, by the proposed conception of it,
undergoes no change, when the cube ABCD is in any manner moved, or turned; and therefore may be
considered to be symmetrically related to all directions of lines in space, or to all possible sections (or

translations] of a point, or body. In fact, we conceive its determination, and the distinction of it (as 4- u)
from the opposite unit of the same kind ( u], to depend only on the usual assumption of an unit of
length, combined with the selection of a hand (as, for example, the right hand), rotation towards which
hand shall be considered to be positive, and contrasted (as such) with rotation towards the other hand,
round the same arbitrary axis. Now in whatever manner the supposed cube may be thrown about in

space, the conceived rotation round the edge AB, /row AC to AD, will have the same character, as righthanded or left-handed, at the end as at the beginning of the motion. If then ihe fourth proportional
to these three edges, taken in this
order, be denoted by 4 u, or simply by 4- 1, at one stage of that
it
may (on the plan here considered) be denoted by the same symbol, at every other
while the opposite character of the (conceived) rotation, round the same edge AB, from AD to
AC,
- u, or to 1.
leads us to regard the fourth
proportional to AB, AD, AC as being on the contrary equal to
It is true that this conception of a new unit
for space, symmetrically related (as above) to all linear

arbitrary motion,
stage

directions therein,
may appear somewhat abstract
can of course confine their attention to the rules

and metaphysical but readers who think it such

of calculation, which have been above derived from it,
;

and from other connected consideiations and which have (it is

hoped) been stated and exemplified,
in this and in a former
volume, with sufficient clearness and fulness.
:

3E2

ELEMENTS OF QUATERNIONS.

396

much more
Quotients

convenient,

for actual use in

either such as has been just

[III.

i.

9, 10.

than any Notation of

suggested for the sake of illus

calculation,

now

was employed in the Second Book, in connexion with that

First Conception of a Quaternion (112), to which that Book mainly related, as
the Quotient of two Vectors (or of two directed lines in space).
The notations
tration, or such as

of the two

Books

however, intimately connected, and the former was

judged to be an useful preparation for the latter, even as regarded the
while the Characteristics of
quotient-forms of many of the expressions used
are,

Operation, such as
S,

V, T,

K, N,

TJ,

are employed according to exactly the same fates in both.

In short, a reader
of the Second Book has nothing to unlearn in the Third; although he may be

supposed to have become prepared for the use of somewhat shorter and more
convenient processes, than those before employed.

SECTION

On

the Interpretation of a

10.

Power of a Vector

as a

Quaternion.
308. The only symbols, of the kinds mentioned in 277, which we have
not yet interpreted, are the cube a 3 and the general power a*, of an arbitrary
vector base,, a, with an arbitrary scalar exponent, t
for we have already assigned
,

interpretations (282,
a~

(1.), (14.),

and 299,

(8.)) for the particular symbols a

And we

which are included in this last/orm.

cular interpretations
of the present
nion,

which

as follows

if

we now

1
,

shall preserve those parti

this

Power

a* is

generally a Quater

be decomposed into two factors, of the tensor

may

consistency with the principles

define, in full

and preceding Books, that

and

versor kinds,

L..o*-Ta*.TJa*;

IV

denoting the arithmetical value of the

th>

power of the positive number Ta,

which represents (as usual) the length of the base-line a and Ua* denoting a
versor, which causes any line p, perpendicular to that line a, to revolve round it as
;

an

axis,

tion,

through

right angles, or quadrants,

according as the scalar exponent,

(comp. 234, (5.)).

t,

and

is itself

in a positive or negative direc

& positive or negative number

POWER OF A YECTOR AS A QUATERNION.

AKTS. 307, 308.]

397

As

regards the omission of parentheses in the formula I., we may

observe that the recent definition, or interpretation, of the symbol a*, enables
(1.)

us to write (comp. 237, II. III.),

II.

The

T (a) =

= Ta

(Ta)

III.

U (a

= (Ua)* = Ua*.

and angle of the power a*, considered

generally determined by the two following formulae
(2.)

axis

as a quaternion, are

IY.

Ax

Ua

V.

and

a*

= 2mr

J tir

(positive or negative or null)

so chosen as to bring the angle within the usual limits,

being

and the

the signs accompanying each other,

integer, n

TT.

In general (comp. 235), we

(3.)

speak of the (positive or negative)

same power, with reference to the

may

product, J^TT, as being the amplitude of the

line a as

an

and may write accordingly,

axis of rotation;

VI. ..am.a =

We may

(4.)

VII.

Ua* =

write also (comp. 234, VII. VIII.),

cos -J +

Ua

sin

In

VIII.

or briefly,

4>

(5.)

ifcr.

Ua

= cas
6

4>

particular,

IX.

Ua

"

= cas

nir

IX

Ua 8w+1 = Ua

upper or lower signs being taken, according as the number n (supposed to be

whole) is even or odd. For example, we have thus the cubes,

X.
(6.)

The

Ua

Ka

a 3 = - aNa.

power a may be thus expressed (it

through ^t-rr round + a, is equi
through + \t-n round
a)
of the

to turn a line _L a

valent to turning that line

.

Ua

and norm

conjugate

being remembered that

XI.

=-

Ta*.

U-

(- a)

XII.

Na* = Ta

2*
;

=-

a.
parentheses being unnecessary, because (by 295, VIII.) Ka
The
the
same
are
and
of
scalar, vector,
reciprocal
power
(7.)
given by the
formulae
:

XIII.

XV.

S. a = Ta. COS
1

a*=

IV.

~
6

Ua-<

XIV.

V. a = Ta .Ua

sin

^
<i

<r*

Ka<

Na* (comp. 190,

(3.)).

ELEMENTS OF QUATERNIONS.

398
If

(8.)

we decompose any

tively parallel

XYI.
=

the

new

a*pa*

round

for,

the new quaternion q

Ax

q,

through

to

p",

a* = p +

p")

Ua

round

for,

191, (5.)),

a,

2
.

p" ,

revoke conically through an angular

an axis (comp. 297, (15.) ).

if q be any quaternion, and

formed from q by such a

is

10.

a* (p

the line a as

More generally (comp.

(9.)

vector obtained by causing

quantity expressed by

axis

we have

to o,

i.

and
which are respec
the general transformation:*

vector p into parts p

and perpendicular

[III.

if

conical rotation of

without any change of

its

angle L

q,

its

own

or of its

tensor Tq.

Treating

(10.)

as three

ijk

unit-lines

rectangular

(295),

the symbol, or

expression,

XVIII.
in

= rtffkj-k*,

XIX.

or

P =

rk^k^,

which

XX. .r^O,

may

represent

any

vector

= 0,

= l, ^0,

t^2,

the length or tensor of this line p being r its incli

and the angle through which the variable plane kp
;

nation^ to k being STT;

may be conceived to have revolved, from the initial position
initial direction towards the
position kj, being for.
(11.)

In accomplishing the transformation XYI., and

expression

employ

XVIII.

to the less

Id,

in passing

with an

from the

symmetric but equivalent expression XIX., we

the principle that

XXI.

s
kf =

=-

S- 1

which easily admits of extension, and

VII. or VIII.

may

K (kj-) =j k
s

be confirmed by such transforma

tions as

(12.) It is scarcely necessary to remark, that the definition or interpreta

tion I., of the power a* of any vector a, gives (as in
algebra) the exponential

property,

XXII...
whatever scalars

more than two

*

may

be denoted by

aV=a"

and

and similarly when there are

factors of this form.

Compare the shortly following

suh-article (11.).

f If we conceive (compare the first Note to page 345) that the two lines i and j are directed
respectively towards the south and west points of the horizon, while the third line k is directed
towards the zenith, then sir is the zenith-distance of p and tir is the azimuth of the same line,
measured from south to ivest, and thence (if necessary) through north and east, to south again.
;

EXPRESSIONS FOR YERSORS AS POWERS.

AETS. 308, 309.]

As

(13.)

ing a vector,

XVIIL,

verifications of the expression

we may observe
XXIII... P

that

it

399

considered as represent

gives,

XXIV.

and

-E>;

(14.) More generally, it will be found that

the eminently simple transformation

if

z
.

= - r\

u* be any

scalar,

we have

XXV.

u
.

= (rVj

kj- k-*)*

rWjWj- k-*.

the two last expressions denote generally two equal quaternions, because
= ru Ilnd, equal angles, each = L u ) and
(k
they have, 1st, equal tensors, each
k by one common system
Illrd, equal (or coincident) axes, each formed from

In

fact,

of

two

round

successive rotations,

one through

sir

round j, and the other through

k.

309.

Any

the form

a*,

quaternion,

by

q,

which

not simply a scalar, may be brought to

and of the exponent, t\ which

is

a suitable choice of the base, a,

and 2

moreover be supposed to fall between the limits

for this purpose we have only to write,
latter

tn

may

I.

2^1

II.

Ta =

III.

T0*;

Ua = Ax

since

7T

and thus

posed quaternion
the base

may
may

Quaternion, on a Scalar and a Vector

the general dependence of a

Element, presents

a, of

itself in
is

a new way (comp. 17, 207, 292).

Tq =

a versor,

1,

the equivalent poiver a

we have thus Ta = 1
,

is

an

unit-line.

"When

the pro

or in other words,

Conversely, every versor

be considered as a poiver of an unit-line, with a scalar exponent,

be supposed to be in general positive, and less than two so that
,

t,

which

we may

write generally,

IV. ..U? = a

with

V. ..a =

Ax.^ = T-

l,

and
VI.
although

and the

if

>

0,

this versor degenerate into 1 or

<

1,

the exponent

becomes

or 2,

an indeterminate or arbitrary direction. And from such

transformations of terrors new methods maybe deduced, for treating questions
of spherical trigonometry and
generally of spherical geometry.
*

base a has

The employment of this letter u, to denote what we called, in the two

preceding sections, a
fourth unit, &c., was stated to be a merely temporary one. In general, we shall henceforth simply
equate that scalar unit to the number one; and denote it (when necessary to be denoted at all] by the
usual symbol, 1, for that number.

ELEMENTS OF QUATERNIONS.

400

Conceive that

(1.)

in

P, Q, R,

46

fig.

and let
unit- vectors, a, /3, 7 as usual
determined by the three equations,
VII.

= 2A,

a?7T

10.

i.

153] are replaced by A, B, c, with

z be three scalars between
and 2,

[p.
y,

a?,

[III.

= SB,

yir

ZTT

= 2c

denote the angles of the spherical triangle. The three versors,

indicated by the three arrows in the upper part of the figure, conie then to be
thus denoted
=
VIII... = a*; q =

where

A, B, c

&;

s>

we have the

so that

diately follow

&

y*~

In X. we

OVJ3* -

3
.

j3V =

and multiplications, these two others imme

]3

X".

/3vy =

- 1

7 being again supposed to be negative.

to

a x = CaSA

and then the formula becomes,

above

- 1

write (by 308, VIII.),

may

XI.

rotation is as

X.

or

the rotation round a from

(2.)

a*

divisions

by easy

last,

f->;

equation,

IX.
from which

q q

y
]3

for

cjSsB

y*

= cysc

triangle, in

any spherical

which the order of

XII.

cysc

c]3sB

casA = - 1

or (comp. IX.),

XIII.

- cos

Taking the

(3.)

= - cos
ing that Sj3a

+ y sin c =

scalars
c,

Taking the

(4.)

XV.
which

is

on both sides of

= a

|3

sin

this last equation,

result

sin A)

and remember

derive one form of the fundamental

cos B

= cos

this other

A cos B +

j3

sin

c sin

A sin

formula

B.

B cos A + V/3a sin A sin B

easily seen to agree with 306, XII.,

The

A + a

sin B) (cos

namely, the equation,

we have

pared with the equation 210,

(5.)

+ cos A

cos c

vectors,

sin c

we thus immediately

equation of spherical trigonometry

XIV.

(cos B

and

may

also be usefully

com

XXXVII.

XV. may

be enunciated in the form of a Theorem, as

follows:
"

If there

be

amj spherical triangle ABC, and three

lines

centre o of the sphere, one towards the point A, with a length

be

sin

drawn from

cos B

the

another

EXPRESSIONS FOR VERSOR8 AS POWEKS.

ART. 309.]

toicards the points, with a length

plane AOB, and towards the

sin B

and

if,

sin

= sm B cos A; and

side

of

it

the third perpendicular to the

as the point c. icith a length

sin c sin

will be directed towards c,

from o

and

the

will have a length

c."

Dividing both members of the same equation XY. by p, and taking

we find that if P be any fourth point on the sphere, and Q the foot of

(6.)

scalars,

the perpendicular let fall

from

common hemisphere,

P are situated at one

of the

through A and B divides the spheric surface),

XYI.

sin c cos PC

common

two into which the great

we have

cyclically

= sin A cos B cos PA + sin B cos A cos PB +

changing abcAKC

side of
circle

then,
sin

A sin B

a formula which might have been derived from the equation 210,
first

PQ

this point on the arc AB, this perpendicular

being considered as positive when c and

that arc (or in one

by

with these three lines as edges, we construct a parallelepiped

intermediate diagonal

same

401

sin c sin PQ

XXXYIIL,

and then passing from the former

and from which many less general equa

to bcavcA.,

triangle to Us polar or supplementary

may be deduced, by assigning particular positions to P.

be the centre of the circum
(7.) For example, if we conceive the point P to

tions

scribed small circle ABC,

by

and denote by

R the

arcual radius of that circle,

the semisum of the three angles, so that 2s =

A+B+

denote, as in 297, (47.), the area* of the triangle ABC,

XYII.
the formula

XYI.

PA = PB = PC = R,

and

sin

PQ =

7r

if

o-

<r,

and

again

whence
sin

R sin (s

c),

gives easily,

XYIII.

2 cot

R sin - = sin A sin B sin c

<>

a relation between radius and area, which agrees with known results, and
from which we may, by 297, LXX., &c., deduce the known equation
:

VTV
XIX.

tan

T>

b
c
a
- 4A sin sin - sin ^
&
At
.

<&

in

which we have

still,

as in 297, (47.), &c.,

XX.
(8.)

In

like

equation 210,

(SajSy =) sin a sin b sin c

manner we might have supposed,

XXXYIIL,
*

HAMILTON

= &c.

in the corresponding general

that P was placed at the centre of the inscribed

Compare the Note

ELEMENTS OF QUATERNIONS*

to the cited sub -article.

3

ELEMENTS OF QUATERNIONS.

402

the sides

and that the arcual radius of that circle was

and thus should have with ease deduced
being

relation,

which

small

circle,

6-

is

a sort of polar reciprocal of

XXI.

the semisum of

known

this other

XVIIL,

2 tan r sin s =

r,

10.

i.

[III.

e.

But

these results are mentioned here, only to exemplify the fertility of the
formulaa, to which the present calculus conducts, and from which the theorem
in

(5.)

was early seen

We

(9.)

might

to be a consequence.

develop the ternary product in the equation XII.,

as

we

developed the Unary product XIII. compare scalar and vector parts and
New general theorems, or at least
operate on the latter, by the symbol S p
thus
of
which it may be sufficient in this
new general forms, would
arise,
;

place to have merely suggested the investigation.

(10.) As regards the order of rotation (I.) (2.),

it

is

clear,

from a mere

formula XV., that the rotation round j from |3 to a, or that

round c from B to A, mmt be positive, when that equation XV. holds good*, at
inspection of the

least if the angles A, B,

c,

of the triangle ABC, be (as usual) treated as positive:

because the rotation round the

281,

line

V/3a from

J3

to a

is

always positive (by

(3.)).

(11.) If, then, for

any

given spherical triangle, ABC, with angles

posed to be positive, the rotation

round

from B

to

still

A should happen

sup

to be (on

the contrary) negative, we should be obliged to modify the formula XV. which
could be done, for example, so as to restore its correctness, by interchanging a
;

with

j3,

and

at the

same time A with

B.

is, however, a sense in which the formula might be considered

(12.)
as still remaining true, without any change in the mode of writing it
namely,

There

if

we were

case last

supposed

tion X.,

we

(11.).

Accordingly,

x, y, z

are negative,
(13.)

if

get this other equation,

XXII.
where

denoting negative angles, for the

we take the reciprocal of the equa

to interpret the symbols, A, B, c as

On

are positive, as before,

if

the rotation round

the whole, then,

er*/3-V =

and therefore the new

a

from

to
]3

y be

exponents, -x,

(1.).

7 be any given system of three co-initial

can always assign a system of three scalars,

if a, /3,

and diplanar unit-lines, OA, OB, oc, we

x, y, z, which shall satisfy the exponential equation X., and shall have
of the form VII. to the spherical angles A, B, c; but these three scalars,

mined

-y,-z,

itself negative, as in

so as to fall between the limits

2,

will be

all positive,

relations
if

deter

or all negative,

EXPONENTIAL FORMULA FOR THE SPHERE.

AET. 309.]

according as the rotation round a from

]3

to

is negative,

as in

403

(1.),

or positive,

as in (11.).

As

(14.)

regards the limits just mentioned, or the inequalities,

XXIII.

..#<

2,

z<2;

y<2,

x>-2,

y>-2,

>

2,

they are introduced with a view to render the problem of finding the expo
nents xyz in the formula X. determinate
for since we have, by 308,
;

XXIV.
we might

a =
4

|3

= +

if

1,

Ta =

Tj3

Ty =

1,

otherwise add any multiple (positive or negative) of the number four,

exponent of any unit-line, and the value of the resulting

to the value of the

power would not be altered.

=
(15.) If we admitted exponents
satisfying the equation

X.

2,

we might render

indeterminate in another

way

the problem of

for

would then

it

be sufficient to suppose that any one of the three exponents was thus equal to
- 2, and that the two others were each =
or else that all three were
2, or

of the

form

(16.)

2.

When

formula X.,

if

it

was

lately said (13.), that the exponents, x, y,

limited as above,

tacitly excluded, for

would have one common

z,

in the

sign, the case was

which those exponents, or some of them, when multiplied

each by a quadrant, give angles not equal to those of the spherical triangle
but equal to the supplements of
ABC, whether positively or negatively taken
;

those angles, or to the negatives of those supplements.

(17.)

In

fact, it is

2
evident (because a =

or the reciprocal equation

of xyz, will

still

be

XXII.,
when we

= y2 = -

1), that the equation X.,

be satisfied by any one system of values

if it

satisfied,

]3

divide or multiply any

ttvo

of the three

exponential factors, by the squares of the two unit-vectors, of which those factors
are supposed to be powers
or in other words, if we subtract or add the number
in
each
two
two,
of
exponents.
:

(18.)

We

may,

XXV.

for example, derive

= -

-^y-

which, when the rotation

be interpreted as follows.

is

from XXII.

or

XXVI.

this other equation

.

a^/3^ = y

2 2
"

as supposed in (1.), so that xyz are positive,

may

Conceive a lime cc with points A and B on its two bounding semi

and with a negative rotation round A from B to c or, what comes to
the same thing, with a positive rotation round A from B to c
Then, on the
plan illustrated by figures 45 and 46, the supplements TT - A, ir B, of the
(19.)

circles,

ELEMENTS OF QUATERNIONS.

404

[III.

10.

i.

angles A and B in the triangle ABC, or the angles at the same points A and B in
the co-lunar triangle ABC will represent two vcrsors, a multiplier, and a multi
2 *
plicand, which are precisely those denoted, in XXVI., hy the two factors, a
,

and

/3

versor,

and the product of these two factors, taken in this order, is that third
which has its axis directed to c and is represented, on the same general

"

plan (177), by the external angle of the lune, at that point, c

which, in quantity,
- c.
is equal to the external
angle of the same lune at c, or to the angle n
This product is therefore equal to that power of the unit-line oc , or - y, which
2
= 2 - z we have therefore, by this construction, the
has its exponent = - (TT ;

<;)

7T

equation,

XXVII...
which (by 308,

The

310.

(6.)

a-*j3

-y=

2 2

(-T)

XXVI.

agrees with the recent formula

equation,
2

I... j

2n

fr

2A

a=

-1,

are the unit-vectors OA, OB,

which results from 309, (1.), and in which
ft, y
oc of any three points on the unit-sphere; while the three scalars A, B, c, in
,

the exponents of the three factors, represent generally the angular quantities
of rotation, round those three unit-lines, or radii, a, j3, y, from the plane AOC
to the plane A OB,

from BOA

to BOC,

and from COB

to COA,

and are

positive or

or
negative according as these rotations of planes are themselves positive
must be regarded as an important formula, in the applications of

negative
the present Calculus. It includes, for example, the whole doctrine of Spherical
not merely because it conducts, as we have seen (309, (3.) ), to one
Triangles
form of tlic fundamental scalar equation of spherical trigonometry, namely to the
:

equation,
II.

but also because

it

cos c

c sin

gives a vector equation (309,

the angles, or the rotations, A,

OA, OB, oc, for

+ cos A cos B = cos

any system

B, c,

with

A sin B

(4.)

),

which serves

to connect

the directions* of the radii, a,

of three diverging right lines

from one

ft,

7, or

origin.

It

* This
may be considered to bo another instance of that habitual reference to direction, as distin
mere quantity (or magnitude), although combined therewith, which pervades the present
guished from
hud aimed to
Calculus, and is eminently characteristic of it; whereas Lea Cartes, on the contrary,
all
reduce all problems of geometry to the determination of the lengths of right lines; although (as
who use his co-ordinates arc of course well aware) a certain reference to direction is even in his theory
inevitable, in connexion with the interpretation of negative roots (by

him

called inverse or false roots)

SPHERICAL SUM OF ANGLES.

ARTS. 309, 310.]

may,

therefore, bo

make

not improper to

405

here a few additional remarks,

respecting the nature, evidence, and extension of the recent formula I.

inverse expo(1.) Multiplying both members of the equation I., by the
2o

nential y

*",

we have

the transformation (comp. 309,

2s

III...(3

2A

2c

a* =

c)

2(?r

-7

(1.)

= 7

2A

(2.)

formula

Again, multiplying both members of

7r

into* a

I.

we

obtain this other

2c

2A

IV.

..y"p

A)

2(jr

=-a^=<T

ir
.

2A

2c

n
Multiplying this last equation IV. by a and the equation III.
we derive these other forms
(3.)

into

7%

2A

V.

a-

2B

2c

7-"^

2a

= - 1

VI.

so that cyclical permutation of the letters, a,

equation I.

J3

j3,

2A

2c

= - 1

a"

7"

7,

and

A, u, c,

in

allowed in the

was to be expected, from the nature of the theorem

as indeed

which that equation expresses.

(4.) From either V, or VI. we can deduce the formula
2A

by comparing which with

2n

2c

2(ir-B)

and IV., we

III.

see that cyclical permutation of

letters is permitted, in these equations also.

(5.)

Taking the

pare 309,

XXII.)

reciprocal (or conjugate) of the equation

this other equation

2A

VIII.

of equations.
of Des Cartes,

Thus

B)

2(_7T

2c

= - 1;

7~~

a"*/3~~

2(7T-_A)

IX.

or

2u

2(TT

"

L, we obtain (com

in the first sentence of Schooten

0)

ft

7""

-f

recently cited translation (1659) of the Geometry

facile ad hujusmodi terminos reduci
peasant, ut deinde ad illorum constructionoin, opus tan turn sit rectarum quarundam
loiigitudineru

wo

find

it

said

"

Omnia Geometric) Problemata

cognoscere."

The very difFerent view of geometry, to which the present writer has been
led, makes it the more
proper to express here the profound admiration with which ho regards the cited Treatise of Des Cartes
as
it
does the germs of so large a portion of all that has since been done in mathematical
containing
:

science, even as concerns imaginary roots of equations, considered as marks of

geometrical impossibility.
* For the distinction
between multiplying a q-uateriiion into and by a factor, see the Notes to
pages
(
lu
J.
147,

ELEMENTS OF QUATEBMONS.

406

in which cyclical permutation of letters

from III.) we can

The equation X. may

2(7r

XL

for

A)

2u
fS~"

2c

-y

7
.

also be thus written (comp. 309,

8(710)

B)

2(ir

)3~~~ = 7

2A

..a^

2(7r

= (-

And all the foregoing equations may be

(7.)

and

again allowed, and from which (or

is

at once derive the formula,

X
(6.)

10.

i.

[III.

XXVII.)

Q)

7)""""

interpreted (comp. 309, (19.)

),

same time proved, by a reference to that general construction (177)

the multiplication of vcrsors, which the figures 45 and 46 were designed to
at the

illustrate

exponent,

t,

through

if
is

mind

in

that a power a, of an unit-line

(by 308, 309) a versor, which has the

right angles,

The

(8.)

we bear

effect

a,

with a scalar

of turning a line J_ a,

round a as an axis of rotation.

principle expressed

by the equation

I.,

from which

all

the sub

sequent equations have been deduced, may be stated in the following manner,
we adopt the definition proposed in an earlier part of this work (180, (4.) ),

if

for the spherical

"For

a suitable Order,

In

(9.)

sum

of

any spherical

fact,

is

two angles on a spheric surface

the Spherical

triangle?,

equal to

when the

Two Right

of the three angles, if taken in

Angles."

negative,

if

we

the spherical sum so obtained

is

(by

rotation round A from B to c

add the angle B to the angle

spherically

Sum

A,

is

the definition referred to) equal to the external angle at c if then we add to
this sum, or supplement of
or total sum,
c, the angle c itself we get a final
;

which

is

exactly equal to

?r

addition of spherical angles at one vertex,

therefore in one plane,

being accomplished in the usual
spherical
to those

and

manner; but the

summation of angles with different vertices being performed according

new rules, which were deduced in the Ninth Section of Book II.,

and were connected (180, (5.) ) with the conception of angular

or
of the composition of angular motions, in different and successive
transvection,

Chapter

I.

planes.
(10.)

Without pretending

we may

it

to attach importance to the following notation,

which may serve to recall and

in passing, as one

just propose
represent the conception here referred

to.

Using a plus

in parentheses, as a

symbol or characteristic of such spherical addition of angles, the formula

be abridged as follows
:

XII.

c (+) B (+)

A =

TT

I.

may

ADDITION OF ARCS ON A SPHERE.

ART. 310, 311.]

407

the symbol of an added angle being written to the left of the symbol of the
because such addition corresponds
angle to which it is added (comp. 264, (4.) )
;

(as above) to a multiplication of versors,

and we have agreed

to write the symbol

of the multiplier to the left* of the symbol of the multiplicand, in every multi
plication of quaternions.

311. There is, however, another view of the important equation 310, L,
according to which it is connected rather with addition of arcs (180. (3.) ), than
with addition of angles (180, (4.) J and may be interpreted, and proved anew,
;

with the help of the supplementary or polar triangle, A B C , as follows.

c being still supposed to be negative,
(1.) The rotation round A from B to
let

a,

j3

to

j3

y
is

and let
c be (as in 175) the positive poles of the sides BC, CA, AB
from
round
a
rotation
the
because
be their unit-vectors.
Then,
y
positive (by 180, (2.)), and is in quantity the supplement of the
;

will be (by 281,

spherical angle A, the product

which a

is

products,
(2.)

the

y /3
and A the angle

axis,

(3.)

(2.),

a versor, of

with similar results for the two other

/3V.
If then we write (comp. 291),
c*Y>

I.

a = UV/3y,

=UVya,

/3

UVa/3,

supposing that
II.

we

shall

To =

T/3

have (comp. again 180,

T7

(2.)

and

1,

III.

Saj3y

>

),

and
V.

A = Ly

fi

B =

Lay

C =

Z/3V

whence (by 308 or 309) we have the following exponential expressions for these
three last products of unit-lines.
2A

VI.
(3.)

a"

2c

SB

Y=

|3-

|3

V = -f

Multiplying these three expressions, in an inverted order,

therefore, the

new product
2c

T7TT
\ JLL.

Compare the Note

f [Here

2B

2A

/Q^T
/Q
2/Q 2
y n u=^ja.ay.y^j = ytjct =
~TT~

TT"

and the equation 310,

*

|3

to

I. is

(3.

in this

way proved

anew.

page 147.

UVjS 7

= UVVyaVafl = U (-

aSafl?)

=-

aUSajS?.]

we

have,

ELEMENTS OF QUATERNIONS.

408
(4.)

And

we might have

because, instead of VI.,

VTTT

n*

__ V
3-

- -

ft*
P

see that the equation to be

proved

i.

10.

written,

a
.

0"

we

[III.

be reduced to the form of the

may

identity
/-j/

"

TX

1JS-

7
and may be

~~7

interpreted as expressing,

-- +

7^7

-I>

what

a point be
supposed to move first along the side B C of the polar triangle A B C from B
to c
then along the successive side C A , from c to A
and finally along the
remaining side A B from A to B it will thus have returned to the position
is

evident, that

if

from which
(5.)

In

it set out,

this

view,

or will on the whole have not changed place at

then,

we perform what we have

addition of arcs (instead of angles as in 310)

(264, (4.)

),

we may

and in a

all.

elsewhere called an

notation already used

express the result by the formula,

X.

AV + CV +

n B C =

r.

each of the two left-handed symbols denoting an arc, which is conceived to be

added (as a successive vector-arc, 180, (3.)), to the arc whose symbol immediately
follows

it,

or

is

written next

it,

but towards the right-hand,

The expressions VI. or VIII., for the exponential factors in 310, I.,
new way the necessity of attending to the order of those factors, in
that formula for if we should invert that order, without altering (as in 310,
VIII.) the exponents, we may now see that we should obtain this new product:
(6.)

show

in a

2A

2B

2c

which, on account of the diplatwrity of the lines a /3 y is not equal to

the properties of which may be
negative unity, but to a certain other versor
,

inferred from

we cannot

what was shown

in 297, (64.),

and

in 298, (8.), but

upon which

here delay.

312. In general (comp. 221), an equation, such as

!.../=?,
between two quaternions, includes a system offour* scalar equations, such as the

following

S/3?

* The
propriety, which such results as this establish, for the use of the name, QUATERNIONS, as
applied to this whole Calculus, on account of its essential connexion with the number FOUR, does not
require to be again insisted on.

AKTS. 311,312.]

where

409

be any three actual and diplanar vectors and conversely,

-y may
such vectors, then the four scalar equations II. repro
three
be
any
y
:

a, j3,

if a,
j3,

and are

duce,

A QUATERNION EQUATION INCLUDES FOUR.

But

sufficiently replaced by, the one quaternion equation I.

an equation between two

vectors is equivalent

equations, such as the three

last

XXII.

one vector equation

equations II.

only to a system of three scalar

for example, in 294, (12.), the
;

equivalent to the three scalar equations

is

XXI.,

under the immediately preceding condition of diplanarity XX. In like manner,

an equation between two versors of quaternions* such as the equation
IIL..TV=U<? ,

includes generally a system of three, but of not more than three, scalar equa
tions; because the versorlJq depends generally (comp. 157) on a system of three

namely the two which determine

scalar s,

determines

its

angle L q

Ax

axis

its

q,

and the one which

or because the versor equation III. requires to be

combined with the tensor equation,

IV.

Oy =

in order to reproduce the quaternion equation

310,

I., is

evidently of this versor-form III.,

If then

unit-lines.

we met

compare 187

Tg,

Now

I.

the recent equation,

7 be

still supposed to be
one of inform had occurred to

if a, |3,

that equation, or if

(13.),

without any knowledge of its geometrical signification, we might propose to

resolve it, with respect to the three scalar s A, B, c, treated as three unknoivn
us,

The few following remarks, on the problem thus proposed,

be not out of place, nor uuinstructive, here.
quantities.

(1.)

Writing

may

for abridgment,

V.

cot

A =

cot B

t,

u,

cot c

v,

and
YI.

= - cosec A cosec B cosec

c,

the equation to be resolved becomes (by 308, VII., or 309, XII.)

VII.

(v

+ y)

(u

)3) (t

a)

in which the tensors on both sides are already equal, because

VIII. ..s 2 =
* An

Up = Up,

a
t?

lw + l^+l.

= TJV0, between two versors of vectors (156), or between the

or
(29 L), is equivalent only to a system of two scalar equations; because the
direction of an axis, or of a vector,
depends on a system of two angular elements (111).
axes of

equation,

U\Y

two quaternions

HAMILTON

ELEMENTS

OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

410

[III.

i.

10.

- a, and
Multiplying the equation VII. by t + a, and into t
dividing
2
the result by t + 1, we have this new equation of the same/orm, but differing
(2.)

by

(comp. 310,

cyclical permutation

IX.

and

in like

(t

a) (v

(3.)

+ y) (u +

manner,

X. ..(U +

j3) (t

Taking the half difference

that (by 279, IV., and 294, II.)
-

f* fj3y

.JL\_ -L

ay/3)

a) (V

+ y) =

of the

(3.)

= V.

arrive at this

two

|3Vy

8.

and observing

last equations,

ySajS

S/3y,

J (/3y

we

|3)

new

equation, of vector form

XII.

t>V/3o

yj3)

V/3y,

+ Vj3y + ySa/3 - aSjSy

which is equivalent only to a system of two scalar equations, because

= 0, when
operated on by S ]3 (comp. 294, (9.) ).

it

gives

(4.)

It enables us, however, to determine the iwo scalars,

operate on

it

by S

XIII.

and

if

we

a,

and

for

if

we

we get (comp. 298, XXVI.),

= a 2 S/3y - Sj3a Say = S (V/3a .Vay)

ifSajSy

operate on the same equation XII. by S

XIV.
(5.)

SajSy

y Sa/3

Say Sy/3

y,

we

get in like manner,

= S
(Vay .Vyj3).

Processes quite similar give the analogous result,

XV.

and thus the problem

Sya

]3

is resolved,

for the three sought scalars,

angles A, B, c

wSa]3y

Sy/3 Sj3a

in the sense that expressions have been found

u, v, or for

t,

= S (Vy/3
.V/3a)

whence the fourth

scalar,

the cotangents V. of the three sought

s,

in the quaternion equation VII.,

can easily be deduced, as follows.

(6.)

Since (by 294,

(6.),

changing S

muting) we have, for any three

vectors a,

XVI.

/3Sa/3y

to a,
]3,

and afterwards

cyclically per

y, the general transformations,

V (V/3a .Vay),
V

=
=V(Vy/3.V]3a),
(Vy/3 .V]3a),

ySa/3y=V (Vay .Vy/3),

SOLUTION OF THE EXPONENTIAL EQUATION.

AET. 312.]

the expressions XIII.

XV. XIY.

XVII.

411

give,

(u

(3)

Sa]3 7

((;+ 7 )Sa/3 7

VT

Va 7 .V7 j3;

/3

.V|3a

whence, by VII.,

XVIII.
and thus the remaining

And

(7.)

expressions

(Sa|3 7 )
s, is

scalar,

the equation VIII.

XVII.

= (V 7 j3) 2 (Vj3a) 2 (Va 7

also entirely determined.

be verified, by observing that the

may

give,

XIX.

(+ 1) (S/37

2
)

2
= (V 7 2
/3) (Vj3a)

The equations XIII. XIV. XV. XVI.

new expressions

(8.)

these

2
)

give,

by elimination

of Sa/3 7 ,

XX.

a*-

= (V

S) ( Vj3a

^=

Va 7

fin-*

- (V 8) (V
7j3
:

we

I.,

find that the three scalars,

cotangents of the angles opposite to the sides a, b,

which the three given unit-lines

Vj3)

(V:S)(V 7 .V 7 j3);

by comparing which with the formula 281, XXVIII.,

the characteristic

c,

t,

after suppressing (291)

u, v,

are either 1st, the

of the spherical triangle in

7 terminate, or Ilnd, the

a, ]3,

negatives of those

being as usual supposed to be

rotation
round
as
the
a from /3 to 7 is negative
according

cotangents, the angles themselves of that triangle

positive (309, (10.)

or positive-, that

),

is

XVIII., according
second

member

(294, (3.)), according as Sa|3 7

as the fourth scalar,

of that

equation

s, is

XVIII.

>

or

<

0; or

finally,

by

negative or positive, because the

ahvays negative, as being the

is

product of three squares of vectors (282, 292).

(9.)

In the

1st case,

which

and VI., that we are permitted

formula 310,

I.,

usually denoted

equation

we

see then anew,

by V.

which are

TT

to

letters.

But we

those three angles,

one even,

see also, that

icithout

and two odd multiples

preserve a positive product of cosecants, because

VI, by

(1.),

exponential

by the same

any

that of 309,

as equal to the angles of the spherical triangle (8.),

even multiples of

or

is

to interpret the scalars A, B, c, in the

of

s is,

we may add any

disturbing
IT,

in

any

the

exponential

order, so as to

for this case, negative in

(8.).
3

ELEMENTS OF QUATERNIONS.

412

i.

[III.

we may,

10.

which

is

that of 309, (11.),

reasons, interpret the scalars A, B,

c,

in the formula 310, I., as equal to the

(10.)

In the Ilnd

case,

negatives of the angles of the

requires, because s

is

now

triangle

and

as thus having,

for similar

what YI. now

positive (8.), a negative product of cosecants, while

have the values required. But we may also add, as in (9.),

any multiples of IT, to the scalars thus found for the formula, provided that
the number of the odd multiples, so added, is itself even (0 or 2).
their cotangents

(11.)

The

conclusions of 309, or 310, respecting the interpretation of the

and might have been anticipated,

exponential formula, are therefore confirmed,

by the present new analysis

in conducting which

evident that

it is

we have

been dealing with real

scalars, and with real vectors, only.

If
this
last
restriction
were removed, and imaginary values admitted, in
(12.)

the solution of the quaternion equation VII.,

as in II., on that equation,

XXI.

we might have begun by

operating,

by the four characteristics,

S

S,

a,

and

]3,

8.7;

which would have given, with the significations 297, (1.), (3.), of /, m, n, and
relation between those four scalar data,

e,

and therefore with the following

XXII.

e*

- m* -

+ 2lmn,

n*

a system of four scalar equations, involving the four sought scalars, s, t, u, v

from which it might have been required to deduce the (real or imaginary)

values of those four scalars, by the ordinary processes of algebra.

(13.) The four scalar equations, so obtained, are the following

=
=

XXIII

nv - tuv +

+ mtu + ntv + uv -

et

- - eu

-<

mu +

It

Itu

+-

tv

+ nuv +

- 2ln

ev

tu

Itv

+ muv - n

eliminating uv and u between the three last of which, we find, with the help
of

XXII., the determinant,

1,

XX1Y.

..0 =

m,
n,

mt, ntv +

It

and analogous eliminations

XXY.
and

XXYI.

Uv +

t,

(f + 1)

e,

et

ev

tv

e (t

1) (ev

-n +

Im)

- 2ln

give,
.

(e

e (t*

uv - (m

+
-

1) (eu
nl) (n

-m + nl),

Im)

(1

(et

- /+
mn)

SOLUTION OF THE EXPONENTIAL EQUATION.

ART. 312.]
(14.)

Rejecting then the factor f

we

-t-

find, as

the problem (12.), the following system of values

XXVII.

XXVIIL

and

et

e*s

=-

mn
(1

m - nl

eu =

(1

the only real solution of

ev=n
- w2
)

(1

Ini

which correspond precisely to those otherwise found before, in

and might therefore serve to reproduce

413

(4.) (5.) (6.),

the interpretation of the exponential

formula (310).
(15.)

But on the purely

that the four equations

solutions,

algebraic side,

XXIII.

represented by the following formulse

(

XXIX.

found, by a similar analysis,

by a system of four imaginary

it is

are satisfied also

+ 1 =

=
;

( s

= tuv

It

mu - nv - e =

be sufficient to have mentioned in passing, since they do not

appear to have any such geometrical interest, as to deserve to be dwelt on here

which

it

may

though, as regards the consistency of the different processes

employed,
may
from the equation VII. to IX., after
certain preliminary multiplications, we divided by f + 1, as we were entitled
to do, when seeking only for real solutions, because t was supposed to be a
be remembered that in passing

it

(2.)

scalar.

This seems to be a natural occasion for remarking that the following

general transformation exists, whatever three vectors may be denoted by a, j3, j
(16.)

XXX.

S (Vj3y Vya Va/3) = - (Sj3y) 2

.

which proves in a new way (comp. 180), that

from Vya

to

Vaj3,

is

always positive; or

is

round

the rotation

the line V/3y,

directed in the same sense (281,

(3.)),

round Vaj3 from a to /3, &c.

In like manner we have generally,

as the rotation
(17.)

XXXI.
XXXII.

and

S (Va/3 .Vya V/3y) = +

(Sa/3y)

S (Vy|3 Vay V/3a) = +

(Sa/3 7 )

so that the rotation round

Vy/3

ment

from Vay

the three diplanar vectors

(18.) If

then

BC, CA, AB, of

A",

B",

c"

any spherical

a, /3,

^6>Vj3a

7 may

is

2
,

negative,

whatever arrange

have among themselves.

be the negative poles of the three successive

triangle,

the rotation round

A"

from

B"

to

sides,
c"

is

ELEMENTS OF QUATERNIONS.

414

which

negative:

is

A quantitative

assigned
be

for

i.

10.

entirely consistent with the opposite result (180), respecting

A B, c

the system of the three positive poles

(19.)

[III.

we may

any pyramid, and

interpretation of the equation

from

infer

OA

if normals

XXX. may also be easily

and 294,

(by 281, (4.),

(3.) ) that if OABC
OB oc to the three faces BOG, COA, AOB have

it

bearing the same

a
similar
to
reference
ratios to units, &o.), then (with
units) the volume of the
new pyramid, OA B C , mil be three quarters of the square of the volume of the old
their lengths numerically equal to the areas of those faces (as

pyramid, OABC.

But an allusion was made, in 310, to an extension of the exponential

has lately been under discussion and in fact, that formula
which
formula
admits of being easily extended, from triangles to polygons upon the sphere
313.

for

we may

write, generally,
2A /J _ 1

2A,,

I.
if

AiA 2

A,,-!

2A 2

^a^~.

an

2Aj

a^ a^

A n be any spherical polygon, and

= (-

1)",

the scalars

if

AI, A 2 ,

in the

exponents denote the positive or negative angles of that polygon, considered as

the rotations A n AiA 2 AiA 2 A 3
namely those from AIA W to AiA 2 &c. while n
,

is

any positive whole number*

2.

>

Let
(1.) One mode of proving this extended formula is the following.
oc = y be the unit- vector of an arbitrary point c on the spheric surface and
conceive that arcs of great circles are drawn from this point c to the n suc
cessive corners of the polygon.
shall thus have a system of n spherical
;

We

triangles,

two

and each angle

of the

polygon

will (generally) be

(positive or negative) partial angles, which

II.

III.

CAiA 3 = A

A w AiC =

CA 2 A 3

I,

AiAjjC

A"!,

may
=

2,

A"

decomposed into

be thus denoted

so that, with attention to signs of angles in the additions,

Also

IV.

V.

A!

=A +
I

A"I,

A2 = A

A" Z

&C.

let

A 2 CA! =

C],

A 3 CA 2 =

C 2 , &C.

and therefore
VI.
which reduces

GI

Ca

itself to 2?r in

angles, and with the point

c,t

= an even multiple of

TT,

the simple case of a polygon with no re-entrant

c in its interior.

* The formula admits of

interpretation, even for the case n

2.

EXTENSION TO SPHERICAL POLYGONS.

ARTS. 312, 313.]

Then, for the triangle CA^ Z of which the angles are d, A

(2.)

have,

!^j

VII.
like

I,

A"

2,

we

III., the equation,

by 310,

and in

415

az

^j.

= - 7
01"^

"

ff
;

manner, for the triangle CA 2 A 3 we have

,

2^3
~

2A^s

VIII.

=-7

ff

a2

"

^202
*

&c.

But, when we multiply VII. by VIII., we obtain, by IV., the product,

2A"

IX.
and

so proceeding,

a^ ai

we have

at last,

X.

ai~ a n

2(o 1

2A

2A 2

a2

310,

The theorem

(3.) ).

(3.)

when

itself to I.,

is

it is

of the form,

V cn~ = (_

which reduces

+ c2
*

by VI., a product

2A,,

2A"i

~=+

2A

2A 2

03"""

1)

^j

!l!i

multiplied by a

and

into a

*"

(comp.

therefore proved.

In words (comp. 310,

"

the spherical

(8.) ),

sum of

the successive angles

of any spherical polygon, if taken in a suitable order, is equal to a multiple of two

right angles, which is odd or even, according as the number of the sides (or corners)
the definition formerly given (180, (4.) ),
of the polygon is itself odd or even
7

Sum

of a Spherical

may

When

be briefly stated thus.

spherical surface, as in

ends of any one

the point c

And

of Angles, being of course retained.

side, is

(1.),

an arbitrary point c

is

the reasoning

taken on the

the spherical sum of the two partial angles, at the

the supplement of the angle which that side subtends, at

but the sum of

all

such subtended angles

is

either four right angles,

some whole multiple thereof therefore the sum of their supplements can
differ only by some such multiple from nw if n be the number of the
or

sides.
(4.)

Whatever that number may

product in the formula

the equations

whereof the former

If any

conically

we have

be, if

we denote by p n

for every vector p,

and

the exponential

for every quaternion

q,

XI.

"

I.,

..

may

line OP,

round any n

~
p n pp n = p

XII.

(by 308,

drawn from
radii, OA
;,

(8.)

pn qpn

~l

= q

be thus interpreted

the centre o of a sphere,

.

oA n

as

successive axes

be

made

to revolve

of rotation, through

ELEMENTS OP QUATERNIONS.

416

[III.

10.

i.

angles equal respectively to the doubles of the angles of the spherical polygon AJ

AW,

the line will be brought back to its initial position, by the composition of these

rotations"

Another way

(5.)

of proving the extended formula I., for

any spherical

analogous to that which was employed in 311 for the case of a

r
r
a
on
Let A A 2
A n be the
triangle
sphere, and may be stated as follows.
Aw A
and let a\, a 2
a n be the
positive poles of the arcs AiA 2 A 2 A 3
is

polygon,

} ,

Then

unit- vectors of those n poles.

the point AI

the positive pole of the

new arc A.\A. n and tlie angle AI of the polygon at that point is measured by
the supplement of that arc
with similar results for other corners of the
is

Thus we have the system

polygon.

of expressions (comp. 311, VI.)

2 A,,

2Aj

XIII.

= a a

ai

an n = a

n -i

so that the product of powers in I. is equal to the following product of n squares

th
of unit-lines, and therefore to the n power of negative unity.

Al V

"VTTT"
.

n n w _i

and thus the extended theorem

n _ia n _ 2

a 2a

f
x

^a n

\n

{- 1)

proved anew.
(6.) This latter process may be translated into another theorem of rotation,
on which it is possible that we may briefly return,* in the Second and last
is

Chapter of this Third Book, but upon which we cannot here delay.
(7.) It may be remarked however here (comp. 309, XII.), that the extended

XV.
(8.)

be thus written

may

exponential formula I.

For example,

Ca n S A n
if

Ca n _iS A_i

ABCD be any

AD

unit vectors of

its

Hence (comp.

XVII.
and

(cos c

therefore,

XVIII.

Ca 2 S A 2

CaiS AI

(- l)

which the

by

A,

D, so that

to AB, &o.,

while

a,

A represents the
|3,

7, S are the

corners, then

XVI.
(9.)

spherical quadrilateral, of

angles (suitably measured) are denoted

positive or negative rotation from

CSD

=
CySC C/3SB CaSA +

sin c) (cos

we may

D+

D-3

scalars

sin D)

on both

sin c sin

309, XIII.),

by taking
cos c cos

write also,
(cos

sides,

cos CD

1.

B+

j3

sin B) (cos

and changing

= - cos B cos A +

Compare 297,

(24.).

A+a

sin A)

signs,

sin B sin

cos

BA

ARTS. 313, 314.]

in fact, each

FORMULA FOR A SPHERICAL QUADRILATERAL.

417

XI Y.)

to the

member

of this last formula

is

equal (by 309,

cosine of the angle AEB, or CED, if the opposite sides AD, BC of the quadri
lateral intersect in E.

Let p = OP be the unit vector of any fifth point, p, upon the spheric
then operating by S p on XYIL, we obtain this other general

(10.)

surface

formula,

sin

sin c cos

cos B cos

AP +

cos CP

sin B cos

sin

A cos BP +

DP +

cos c cos

sin

A sin B sin AB sin PQ

sin c sin

sin

CD sin PR

in which the sines of the sides AB, CD are treated as always positive

but the

PQ and PR, on those two sides, are regarded as

according as the rotations round p, from A to B and from

of the perpendiculars

sines

positive or negative,

c to D, are negative or positive

to P, several other

but

less

and hence, by assigning particular

positions

general equations of spherical tetragonometry can

be derived.

For example, if we place P at the intersection, say F, of the opposite

AB, CD, the two last perpendiculars will vanish, and two of the six terms
and a similar reduction to
disappear, from the general formula XIX.

(11.)
sides

will

four terms will occur,

if

we make

the arbitrary point P the pole of a

side, or

of a diagonal.

The

314.

definition of the power a

which was assigned in 308, enables us

form some useful expressions, by quaternions, for circular, elliptic, and spiral
loci, in a given plane, or in space, a few of which may be mentioned here.

to

Let a be any given unit- vector OA, and

(1.)

perpendicular to

it

then,

I.

by the

f3

any other given

definition (308), if

OP = p =

/3,

Ta =

Sa]3

1,

the locus of the point p will be the circumference of a

and OB for radius, and in a plane perpendicular to OA.
(2.)

we

If

the product
if

we

take

retain the condition

Ta =

1,

we
=

line OB,

write,
0,

circle,

with o for

centre,

but not the condition Sa]3 = 0, then

and not merely a vector but

will be in general a quaternion,

a*]3

its

vector-part (292),
II.

III.

we can form

OP = p = Y. a fj3 =

j3

this

cos x

new vector-expression,

+ y

sill x,

where

and now the

oc for

its

locus of

is

2x =

and

IY.

ellipse,

with

tir,

a plane

major and minor semiaxes

= oc =
Yoj3

its centre at o,

and with OB and

while the angular quantity,

x, is

what

often called the excentric anomaly.

HAMILTON

ELEMENTS OF QUATERNIONS.

is

ELEMENTS OF QUATERNIONS.

418
If

(3.)

V,

we

OB =

/3

and P

so that B

under the same conditions

write,

= V/3a

a = a~ y,

and

YI.

we have

VII.

)3

rectangular

then,

is

the same conditions, the

VIII.

IX.

or

IX

which

last

Syfto,

X.

by

OA

II.,

aVpa,

jS

o,

the equation,
;

with OB and oc for two

elliptic locus (2.),

TVap

TVa/3

of the point p

itself,

(5.) ),

If

we

Sa/>

j3

Ty,

S a/o = SajSSjSp (comp. 298, XXVI.)

form we have, by
Saj3 cos x,

retain the condition Saj3

Sj3p

/3

and IV.,

II.

cos x.

but not now the condition

(1.),

Ta = 1, we may again write the equation I. for

now be a logarithmic spiral, with o for its pole,
to

Vpa

the plane,

as a confirmation of

(5.)

OP = p =

circle,

the section of the right cylinder (compare 203,

made by

radii.

Under

(4.)
is

(2.),

cos0 + 7 sin # = a

so that the locus of this projected point p

10.

i.

and p on a plane drawn through

are the projections (203) of B

at right angles to the unit-line OA,

[III.

but the

locus

of p will

in the plane perpendicular

because equal angular motions, of the turning

line OP,

correspond

now

to equal multiplications of the length of that line p.

(6.)

For example, when the

the revolving unit

line,

returns (comp. 309,

of

was made, the

of the locus,

is

XL

XXIV.)
length Tp

scalar exponent

Up

is

which

it

of the turning line p

multiplied by To.

by

4, so that

= TJaMJ/3

to the direction

increased

had before the increase

itself,

or of the radius vector

which constant and positive scalar

is

not

now

equal to unity.
(7.)

If

we

reject both

the conditions

Ta =

(1.),

and

l,

Sa]3

0,

an unit-line, nor
power a*, is now neither
as a factor,
perpendicular to j3, namely to the line on which that power operates,
we must again take vector parts, but we have now this new expression
so that the line a, or the base of the

XII.
in which

we have

OP = p = V.

a*/3

(/3

cos^ + 7 sinz)

written, for abridgment,

XIII.

..fl

= Ta,

<y=.V(Ua./3).

EXPRESSIONS FOR CERTAIN SPIRALS.

ART. 314.]

In

(8.)

be said to be

may

now an

case, the focus of

more complex

tliis

a plane curve, and

is still

logarithmic spiral

elliptic*

419

for if

we

suppress the

scalar factor, a*, we fall back on the/om II., and have again an ellipse as the
but when we take account of that factor, we find (comp. (2.) that equal
locus
)

increments of excentric anomaly (#), in the auxiliary ellipse so determined, cor

to equal multiplications of the length (Tp), of the vector of the new

respond
spiral.

We may

(9.)

also project B

and

P, as in (3.), into

plane through o perpendicular to OA, which plane

of the auxiliary vector

the equation of

tlie

we

and

it is

a*

../>

(ft cos #

easily proved that

Ta

spiral, of the

known and

on the

contains the extremity c

>

or

+ 7 sin x] = a

are brought back to the case (5.),

be a logarithmic

still

projected spiral becomes (with

XIY.
so that

then, since

points B and P

/3

<

Ua

/3

1),

and the projected curve

is

seen to

ordinary kind.

(10.) Several spirals of double curvature are easily represented, on the same
general plan, by merely introducing a vector-term proportional to t, combined

or not with a constant vector-term, in each of the expressions above given, for

For example, the equation,

the variable vector p.

XY.
while

c is

cta

a%

Ta =

with

l,

and

Saj3

any constant scalar different from zero, represents a

VIII.

0,

helix,

on the right

circular cylinder

And

we introduce a new and

right-hand term, and so write,
(11.)

if

XVI.

variable scalar, u, as & factor in the

eta

/3,

we

shall have an expression for a variable vector p, considered as depending on

two variable scalars (t and u), which thus becomes (99) the expression for a
vector of a surface

namely

of that important Screw Surface,

which

is

the locus

from the various points of a given helix, on the

axis of the cylinder of revolution, on which that helix, or spiral curve, is

of the perpendiculars, let fall

traced.

The

usual logarithmic spiral might perhaps bo called, hy contrast to this one, a circular loga
Compare the following sub-article (9.), respecting the projection of what is here called

rithmic spiral.

an

elliptic

logarithmic spiral.

3H2

ELEMENTS OF QUATERNIONS.

420

[III.

10, 11.

i.

315. Without at present pursuing farther the study of these loci by quater
1
nions, it may be remarked that the definition (308) of the power a , especially

when Ta =

1, combined with the laws (182) of t, j, /c, and with

the identification (295) of those three important right versors with their own
indices, enables us to establish the following among other transformations,

for the case

which

will be

(1.)

Let

found useful on several occasions.

a be
I.

III.

any

= S

r<

a^S.a
V.

(2.)

Let a and

VI.

+aS.a*-

(4.)

We

have

X.
xi.
(5.)

The

also,
*

by

the

eT

IV.

1
;

= S

= - S OM

+1

- aS O M

a*

let

be

still

S.a^S.a*

any scalar

then

a*"

1
.

/, k,

= *v.

k<

= s

a**

1
.

same principles and laws,

j V.

k*

= V.

A?v.y*

AV.

= - v.

**

= V.y

v.

AJ*

= - v.y.

may be

put under

may

XIII.

V. a*= aS

vector p,
expression 308, (10.), for an arbitrary

it

a^=l.

and
.

a*

a* = S

VII.

=/w

V./ = V. #

XII.

And

i,

v.

yv. ^ = - v. *

the following form

(6.)

aV.a ^a

Hence, by the laws of

ix.

be any scalar] then,

unit-vectors,

= S

a*

VIII.

(3.)

2
2
=
.(S.a0 +(S.a^)

let

II.

a*;

any two

be

and

unit-vector,

= rV. & 2S+1 + rA 2 *V.

be expanded as follows
.

(i

cos

tir

+j

sin

2S

.*

tir)

sin

sir

+ k cos STT)

Second Chapter of this Book [337],

on some of these last; expressions, in connexion with differentials and derivatives
(7.)

We

shall return, briefly, in the

of poivers of vectors

but, for the purposes of the present section, they

may

suffice.

[Since

/* =

8 -1

./Si

Sk*>

this

follows at once from

rk*

j->

A 1 --, remembering

that

ABTS. 315, 316.]

LOGARITHMS OF DIPLANAB QTJATEBNIONS.

SECTION

421

11.

of Biplanar Quaternions; with some

Additional Formulae.

On Powers and logarithms

We

Chapter with a short Supplementary

Section, in which the recent definition (308) of a power of a vector, with a
scalar exponent, shall be extended so as to include the general case, of a Power
316.

shall conclude the present

of a Quaternion, with a Quaternion Exponent, even

so

combined are diplanar

less general one

with the

and a connected

when

the two quaternions

definition shall

be given (consistent

same kind, which was assigned in the Second

of the

Chapter of the Second Book), for the Logarithm of a Quaternion in an arbitrary

Plane * together with a few additional Formulae, which could not be so con
:

veniently introduced before.

(1.)

"We

propose, then, to write, generally,

and

q being any quaternion,

(or

being the real and

of logarithms, of real

Napierian) system

and

known

base of the natural

positive scalars

so that (as

usual),
II.

+ i + -11 + &o. = 2-71828.

(Compare 240,
(2.)

for

We

what we

and

(1.)

any quaternion

shall call its principal logarithm, or

shall

4i

(2.).)

shall also write, for

III.

and thus

1?

= 1T? +

q,

the following expression

simply

its

Logarithm

Z?.UV<?;

have (comp. 243) the equation,

IV.

&=

q.

(3.) When q is any actual quaternion (144), which does not degenerate (131)
into a negative scalar, the formula III.
assigns a definite value for the logarithm,

\q

which

is

such (comp. again 243) that

V.

VII.

SI? = IT?

VI. .. VI? = L q .UV?;

UV1? = UV?

VIII.

TV1? = L

The quaternions considered, in the

Chapter referred to, were all supposed to be in the plane of
the right versor
But see the Second Note to
page 277.
.

ELEMENTS OP QUATERNIONS.

422

[III.

11.

i.

the scalar part of the logarithm being thus the (natural) logarithm of the tensor
and the vector part of the same logarithm Iq heing constructed by a line in the
;

direction of the axis

Ax

of

q,

which the length bears,

to the

assumed unit of

length, the same ratio as that which the angle L q bears, to the usual unit of

angle (oomp. 241, (2.), (4.)).

(4.)

If

it

were merely required to

IX.
in which q
to

is

supposed

any negative

X.
where n

is

..f-q,
and actual quaternion, which

to be a given

scalar (3.),
.

satisfy the equation,

we might do
= (log
q

q) n

this

not equal

by writing (compare again

243),

+ 2mrUVq,

Iq

any whole number, positive or negative or null

what we have

is

and

in this view,

quaternion q, is only what may

be considered as the simplest solution of the exponential equation IX., and may,
as such, be thus denoted
called the logarithm, \q, of the

XI. ..lq=
(5.)

The

excepted case (3.),

where q

(log 0)

a negative scalar, becomes on this

is

plan a case of indetermination, but not of impossibility

since

we have,

for

example, by the definition III,, the following expression for the logarithm of
negative unity,

XII.
which in

its form

(- 1)

iV - 1

agrees with old and well-known results, but

is

here inter

preted as signifying any unit-vector, of which the length bears to the unit of
length the ratio of TT to 1 (coinp. 243, VII.).

We propose

also to write, generally, for any tivo quaternions, q and q,

even if diplanar, the
following expression (comp. 243, (4.) ) for what may be
called the principal value of the power, or
simply the Power, in which the
(6.)

former quaternion q

is

the base, while the latter quaternion q

XIII.
and thus
tions I.
(7.)

tf

is

the exponent

this quaternion

power receives, in general, with the help of the defini

a
III.,
perfectly definite signification.
When the base, q, becomes a vector, p, its angle becomes a right angle

and

the definition III. gives therefore, for this case,

and

this is the quaternion

which

is

XV.

to be multiplied
f

&**.

by /,

in the expression,

POWEES AND FUNCTIONS OF QUATERNIONS.

AET. 316.]
(8.)

last

for the

When,

formula becomes

same

XVII.
we

I.,

* up

^ = TV.

../>=

the relation (Up)

= cos x +

Up

sin x,

U
P,

= -

2z =

if

reproduced, as one which

for

we may

XVIII.

(9.)

the

by the

(UpY =

tTr

XVII

or briefly,

c*

power a

17

is

cpsx,

in this

way

included in the more general definition XIII., of the

is

write,

** Up

mentioned

last

definition,

= cps - (comp. 234,

VIII.)
/c

XVI. and XVII.,

with the recent values

that the

t,

1 gives,

see that the former definition, 308, I., of the

power q q

scalar,

XYI.
and because, by

becomes a

vector-base, the exponent q

423

x and t^p.

of

In the present theory of diplanar quaternions, we cannot expect to find

sum of the logarithms of any two proposed factors, shall be generally

equal to the logarithm of the product

but for the simpler and earlier case of

may be considered to exist, with

complanar quaternions, that algebraic property

due modifications for multiplicity of value*
(10.)

The

definition III. enables us, however, to establish generally the

very simple formula (comp. 243, II. III.)

XIX.
in which (comp. (3.)

\q

(Tq .Uj) = ITq + 1U?

),

z.0;

(11.)

We

any quaternion

XXII.

UlUg = TJV?.

have also generally, by XIII., for any scalar exponent,

base, q,

XXIII.

t,

and

the power,
.

q*

&=

(Tq)<.

(cos

Lq +

UV?

sin

q)

or briefly,

XXIir.
in

q*

= iy. cus

which the parentheses abput Tq

XXIV.

(g

may

q,

if

UV?

be omitted, because

= (Tqy=Tq*
(comp. 237,

II.).

* In
243, (3), it might have been observed, that every value of each member of the formula IX.,
there given, is one of the values of the other member ; and a similar remark
applies to the formulae I.
and II. of 236.

ELEMENTS OF QUATERNIONS.

424

[III.

11.

i.

of a power are two rectangular vectors,

(12.) When the base and exponent
their
and
whatever
lengths may be, the product p\p is, by XIV., a
p
p, then,
a
vector
but t is always a versor,
;

XXV.
we have

Ta + Ua

or in words, the power p

T. P P =

For example (comp. 242,

XXVII.
and generally

to the base,

Thus whatever

XXVIII.

= ** = - k

*>

(15.)

c is

If, for

=0;

/= &- +

the tms of the power

scalar c

may

i cj

this power, then, is a versor (12.),

when

Pj

and the shortly following formula

(7.),*

a scalar, which in certain cases

tive or negative

S,

if

1,

rm s becomes

unless the direction of that

the case

the base be an unit-line, and the exponent a line of awy length,

if

but perpendicular

(14.)

any vector

a versor, under this condition of rectangularity.

is

XXVIII.),

itself to

sin Ta, if a be

therefore,

XXVI.

(13.)

cos

and

any whole and

(153), and the

may

e-i

a line perpendicular

indeterminate,

c *-

= cos

rms

its

to

both

by the power reducing

happen.

we may

be,

t cfli

is

is

write,

-k

sin

j-

generally the line

+ k

but in

grew number, this versor degenerates into posi

^%

real quaternion

any

XXIX...UV0 = 0,

axis
q,

becomes indeterminate

we

(131).

write again,

XXX... ^ = ^, and XXXI...0

and therefore

= -l,

the process of 239 will hold good, when we change i to v the series, denoted
in I. by ? is therefore always at last convergent,^ however great (but finite) the
tensor Tq may be
and in like manner the two following other series, derived
;

from

it,

which represent (comp. 242,

analogy to known expressions, the

(3.)

cosine

what we

and

shall call, generally,

by

sine of the quaternion q, are

always ultimately convergent:

XXXII.

XXXIIL

cos? - J
sing -

-")

J3 + L/8.4

~) - | - j-f^ +

- &0

T^O

&C>

* In the
a certain multiplicity
theory of complanar quaternions, it was found convenient to admit
a notation for the
of value for a power, when the exponent was not a whole number and therefore
section enable us
principal value of a power was employed, with which the conventions of the present
;

now

to dispense.

t In

fact, it

can be proved that this final convergence

is replaced
by a biquaternion (214, (8.) )

nary, or

when

any but

real quaternions.

it

even when the quaternion is imagi

we have no occasion here to consider

exists,
;

but

POWERS AND FUNCTIONS OF QUATERNIONS.

ART. 316.]

We

(16.)

shall also define that the secant, cosecant, tangent,

of a quaternion, supposed

XXXIV.

to be real, are the functions

still

sec? =

We

cos

c v!7

cos q

+ v

~v

sin q,

therefore, as in trigonometry (comp. 315, (1.)

and

XXXVII.
And

//

may

whatever quaternion q
(18.)

2u

cosec? =

q,

&c.

shall also have,

XXXVI.

^w?o

have the usual relations, sec q - 1

shall

(17.)

and cotangent

XXXV... tan,
and thus

-^

425

(cos

qY +

(sin q}*

".

cos q

- v

sin q

),

&* = e

1,

be.

the formula of trigonometry, for cosines

and

swzes of

swms of

or wore arcs, &c., will thus hold good for quaternions also, provided that the

quaternions to be combined are in any common plane

XXXVIII.
(19.)
(9.)

it

cos (q

= cos /

q)

cos q

for example,

sin q sin

q,

if

\\\

q.

This condition of complanarity is here a necessary one; because (comp.

is necessary for the establishment of the exponential relation between

sums and powers.

(20.)

Thus, we

may

indeed write,

XXXIX.

e^ =

but, in general, the developments of these

XL.

? +?

gg + terms

V = g?

e,

if

\\\q<,

two expressions give the

of third

difference,

and higher dimensions

<o

XLI.

and

(qq

- q

q)

V? V/),
.

an expression which does not vanish, when the quaternions q and q

are

diplanar.

few supplementary formulae, connected with the present Chapter,

be
here, as was mentioned at the commencement of this Article
appended
may
And
first
it may be remarked, as connected with the
(316).
theory of powers
(21.)

of vectors, that
HAMILTON

if a,
]3,

7 be any

ELEMENTS OF QUATERNIONS.

three unit-lines, OA, OB, oc,

and

if

o-

denote the
3 I

ELEMENTS OF QUATERNIONS.

426

XX. may

area of the spherical triangle ABC, then the formula 298,

written

i.

[III.

11.

be thus

p+7

a+

p 7+

the exponent being here a scalar.

The immediately preceding formula,

(22.)
vectors,

the relation

XLTII.

2
2
2
(Uaj3y) + (Uj3y) + (Uay) + (Uaj3) + 4TJay

for example,

(23.)

7 be

if a,
|3,

made equal

1-1-1-1

tion becomes,

gives for any three

XIX.,

298,

The following

is

= -

to

/,

j, k,

the

SU/3

member

first

= - 2

SUj3 7

of this

equa

2.

much more complex

identity, involving as

it

not only three arbitrary vectors a, |3, 7, but also four arbitrary scalars, a,
and r\ but it has some geometrical applications, and a student would find
for himself.

does
ft,

it

c,

To

to investigate a proof of it
are
the
three
vectors
a, |3, 7, and the three scalars a, ft, c,
abridge notation,
sums
are
formed
which
to
considered as each composing a cycle, with respect
2,

good exercise in transformations,

and products

II,

on a plan which

XLIV.

20V/3y = aVj3y +

be thus exemplified

may

+ cYa/3

ftV 7 a

This being understood, the formula to be proved

XLY.

+ (2aY/3 7

(Sa/3 7 )

2H (r

+ 8)87 +

+ S(r 2 + a 8 +

2
)

be]

2H (r +
2

{(V/37)

W.

Ha* =

the following

+ r 2 (2V/3 7 ) 2 - r(2a

is

(/3

formula gives, for any three

XLYI.

vectors a,
2

(Sa/3 7 )

/3,

For example, by making the four

7)

az)

2ftc(r

the sign of summation in the last line governing all that follows
(24.)

scalars a,

ft,

c,

it.

r each

0, this

7, the relation,

+ 2nS/?7 =

2H

+ 2

8
.

(Yj3 7 )

which agrees with the very useful equation 294, LIIL, because

XLYII.
(25.)

Let

a a (V)3y)

= a2

{(Sj3 7 )

]3y) =

(aS]3y)

7 be the vectors of three points A, B, c, which are

which the radius is r, and the equation is,

a, /j,

a given sphere, of

- Ua\

XLYIII.

2
2
=
p + r

(comp. 282, XIII.)

exterior to

ADDITIONAL FORMULAE, CONTACTS ON A SPHERE.

ART. 316.]

427

denote the lengths of the tangents to that sphere, which are

shall then have the rela
drawn from those three points respectively.

and

let a, b, c

We

tions

XLIX.
thus r 2 +
the

a?

= - a2

formula

of that

significations of the letters

the eight (real or imaginary) small

circles on a sphere.

We

= -

z
;

of the formula

and thus a theorem

XLY.

vanishes

therefore also equal to zero, for these

is

to be extremely useful, in the investigation

(26.)

= 72 +

and the second member

&c.,

member

first

a 2 + a2 =

obtained, which

is

of the

by quaternions

which touch a given

circles,

cannot enter upon that investigation here

found

system of

of three small

may remark

but

because the vector p of the foot P, of the perpendicular OP

on the right line AB, is given by the expression,
)3

set

is

that

the origin o

let fall

V/3

may be proved in various ways, the condition of contact of that right

with the sphere
VIII. is expressed by the equation,
as

line

AB

XL

LI.
or

TVfBa = rT

(a

/3)

LIL

or

2
.

by another easy transformation, with the help

This

(27.)

+ Sa/3) 2 =

LIII.

(r

(r

(r

(V/3a)
of

r z (a

j3)

XLIX.,
= aW.

equation evidently admits of decomposition into two factors,

last

representing two alternative conditions, namely,

LIV.
and

if

we

r2

+ Sa/3 - ab =

consider the tangents a

still

LV.

and

+ Sa|3 + ab =

b (25.) as positive, it is

in several different ways, that the first or the second factor

according as the point

P, at

not fall between the points A

of that line

In

(28.)

LVI.

AB

line

touches the sphere, does or does

or in other words, according as the length

equal to the sum, or to the difference, of those two tangents.

fact we have, for the first case,

T(/3

easy to prove,

to be selected,

is

a)

a,

in virtue of the relations

LVII.

which the

and B

is

T (]3 - a) =

(b

or

XLIX.
-

a),

or

(|3

a)

(b

a)

= - 2

+ Saj3

(r

ab),

but, for the second case,

=
(]3

a)

(b

a)

=-2

(r

+ Sa/3 + ab)

312

ELEMENTS OF QUATERNIONS.

428

and

it

be remarked, that

may

we might

[III.i.ll.

way have been

in this

led to find the

and thence the equation LIIL, or

system of the two conditions (27.)

its

trans

formations, LII. and LI.

(29.)

We may

XLVIIL,

conceive a cone of tangents from A, circumscribing the sphere

it along a small circle, of which ike plane, or the
polar

and touching

plane of the point A,

LYIII.
and in

like

is

Sap +

found to have for

easily

r*

its

equation,

(comp. 294, (28.), and 215, (10.)

manner the equation,

LIX.

..8/3,0

+ r2 =

0,

represents the polar plane of the point B, which plane cuts the sphere in a
second small circle and these tivo circles touch each other, when either of the two
:

conditions (27.)

is satisfied

LY.

internal for the case

(30.)

The

such contact being external for the case LIY., but

condition of contact (26.), of the line

and

sphere,

might have been

otherwise found, as the condition of equality of roots in the quadratic equation

(comp. 216, (2.)),

LX.

o = (xa + yfiY +

x +

yY r\

or

LXI.

(r

z
+ a 2 ) + 2xy(r* + 8a0) + y (r* +

the contact being thus considered here as

/3

) ;

case of coincidence of inter

sections.

which expresses that

(31.) The equation of conjugation (comp. 215, (13.)),
each of the two points A and B is in the polar plane of the other, is (with the
present notations),

LXII.

2
r + Sa)3 =

the equal but opposite roots of LXI., which then exist

if

the line cuts the

division of the secant line

sphere, answering here to the well-known harmonic

AB (comp. 215, (16.)), which thus connects two conjugate points.
(32.)

In

like

manner, from the quadratic equation 216, III., we get

this

analogous equation,

LXIII...S-S^-sfY^.Y^)
a
a
\
PJ
P
connecting the vectors A,
tively to the ellipsoid

ju

of

216, II.

l,

any two points L, M, which are conjugate rela

and if we place the point L, on the surface, the

ADDITIONAL FORMULA, CONTACTS OK A SPHERE.

AKT. 316.]

429

equation LXIII. will represent the tangent plane at that point L, considered as
the locus of the conjugate point M; whence it is easy to deduce the normal, at
any point of the ellipsoid. But all researches respecting normals to surfaces

can be better conducted, in connexion with the Differential Calculus of Qua

ternions, to which we shall next proceed.
(33.) It may however be added here, as regards Powers of Quaternions
with scalar exponents (11.), that the symbol q rq~ f represents a quaternion
formed from r, by a conical rotation of its axis round that of q y through an
= 2t L q and that both members of the
angle
equation,
;

LXIY.
are symbols of one

[Some
(?*

qt

This

t"

quaternion.

of the expressions q qfq

"

and

the definition XIII.,

2"is

5
<?

i<z

"

=
(<f."f

is

(qrq-J

must be taken in the interpretation

care

By

)*"

common

and

(0

1
t*"

)3"

^ =

^"^

0*"*

quite consistent with the rule that in an operating product the

factor to the right operates first

on ihe operand. If the expression \q had

been interpreted as equal to \q q instead of q 1 q, then indeed the equality
= i2.
would have held good, but the general rule would
(jsy :=
have been disobeyed.]
qf

"

"

ELEMENTS OE QUATERNIONS.

430

CHAPTER

[III. n.

1.

II.

ON DIFFERENTIALS AND DEVELOPMENTS OF FUNCTIONS OF

QUATERNIONS; AND ON SOME APPLICATIONS OF QUATER
NIONS, TO GEOMETRICAL AND PHYSICAL QUESTIONS.

SECTION

On

1.

the Definition of Simultaneous Differentials.

317. IN the foregoing Chapter of the present Book, and in several parts of
Book preceding it, we have taken occasion to exhibit, as we went along,

the

a considerable variety of Examples, of the Geometrical Application of Quater

but these have been given, chiefly as assisting to impress on the reader
the meanings of new notations, or of new combinations of symbols, when such
nions

presented themselves in turn to our notice. In this concluding Chapter, we

desire to offer a few additional examples, of the same geometrical kind, but
dealing,

more

surfaces

and

freely than before, with tangents

to give at least

to Physical Inquiries.

But

some
it

and normals

to curves

and

specimens, of the application of quaternions

seems necessary that we should

first

establish

here some Principles, and some Notations, respecting Differentials of Quater

nions,

and

318.

of their Functions, generally.

The

usual definitions, of differential

coefficients,

and

of derivedfunctions,

are found to be inapplicable

generally to the present Calculus, on account of
the (generally) non-commutative character of quaternion-multiplication (168,
It becomes, therefore,
191).
necessary to have recourse to a new Definition of
Differentials,

which yet ought to be so framed, as

have been seen to be

of Quaternions.

and

to

because scalar 8 (131), as well as

included, under the general Conception

nclude, the usual Rules of Differentiation

vectors (292),

to be consistent with,

DEFINITION OF DIFFERENTIALS.

ARTS. 317-320.]

319. In seeking for such a

new

431

definition, it is natural to

go back

to the

and to consider how the

principles of the whole subject of Differentials
to
have
dealt with the question,
be
Fluxions
of
Inventor
supposed
might
great
first

if

he had been deprived of that powerful resource of common

is

supplied by the commutative property of algebraic multiplication

familiar equation,

calculation,
;

or

which

by the

xy = yx,

considered as a general one, or as subsisting for every pair offactors, x and y\

while limits should

still

be allowed, but infinitesimal* be

still

excluded:

and

indeed the fluxions themselves should be regarded as generally finite* according

to what seems to have been the ultimate view of NEWTON.
320.

which a study of the Principia appears

contained in the following Definition, which we believe to be a

The answer

to suggest, is

to this question,

and which we propose

perfectly general one, as regards the older Calculus,

adopt for Quaternions

"Simultaneous Differentials (or

Corresponding Fluxions] are Limits of Equi

multiples^ of Simultaneous and Decreasing

And
then,

if

any

any system

tend to vanish together, according to

differences,

any

of

any system

of

law, or system of laws

equimultiples of those decreasing differences all tend together to

of finite limits, those Limits are said to be Simultaneous Differentials

of the related Variables of the System

the

Differences"

conversely, whenever any simultaneous

variables, all

to

letter d,

and are

denoted, as such,

by prefixing

Symbol of each such

as a characteristic of differentiation, to the

variable.

to the Second Lemma of the Second Book of the Principia (Third

and especially the following passage (page 244)
Neque enim spectatur in hoc Lemmate magnitudo momentorum, sed prima nascentium proportio.
Eodem recidit si loco momentorum usurpentur vel velocitates incrementorum ac decrementorum (quas
etiam motus, mutationes et fluxiones quantitatum nominare licet) vel finite qusevis quantitates velo-

Compare the remarks annexed

Edition, London, 1726)

"

citatibus hisce proportionales."

t As regards the notion of multiplying such differences, or generally any quantities -which all
diminish together, in order to render their ultimate relations more evident, it may he suggested by
various parts of the Principia of Sir Isaac Newton but especially by the First Section of the First
;

Lemma

under which such expressions as the following

d produci," .... "ideoque rectae
The direction, "ad puncta longinqua produci," is repeated in connexion
semper finitse Ab, Ad,
with the Eighth and Ninth Lemmas of the same Book and Section while under the former of those
two Lemmas we meet the expression, "triangula semper finita," applied to the magnified representa
tions of three triangles, which all diminish indefinitely together
and under the latter Lemma the words
occur, "manente longitudinae Ae," where Ae is a finite and constant line, obtained by a constantly
Book.

See for example the Seventh

occur:

"

intelligantur semper
.

A~B

et

(p. 31),

AD ad puncta longinqua b et

."

increasing multiplication of a constantly diminishing line

AE (page

33 of the edition cited).

ELEMENTS OP QUATERNIONS.

432

More

321.

fully

and symbolically,
I.

[III. n.

let

q, r, s,

denote any system of connected variables (quaternions or others)

II.

A^, Ar, As,

1,2.

and

let

denote, as usual, a system of their connected (or simultaneous} differences

such a manner that the sums,
III.
shall be a

ever they

q + Ag,

+ Ar,

+ A*,

in

new system of variables, satisfying the same laws of connexion, what

may be, as those which are satisfied by the old system I. Then, in

new system

returning gradually from the

gradually from the old

made

to the

to the old one, or in proceeding

the
simultaneous
new,
differences II. can all be

to zero, since it is evident that they may

while
the
if,
differences themselves are thus supposed to
decrease* indefinitely together, we multiply them all by some one common but

(in

general) to approach together

all vanish together.

But

increasing number, n, the system of their equimultiples,

IY.

may
when

tend to
this

become equal

happens, as in all

some determined system of finite limits. And

ordinary cases it may be made to do, by a suitable
to

adjustment of the increase of n to the decrease of A#, &c., the limits thus obtained
are said to be simultaneous differentials of the related variables,
denoted, as such,

q, r, s

and are

by the symbols,
Y.

dq,

dr,

SECTION

d-s,

2.

Elementary Illustrations of the Definition, from Algebra and

Geometry.
322.

To

leave no possible

doubt, or obscurity, on the import of the

foregoing Definition, we shall here apply it to determine the differential of

a square, in algebra, and that of a rectangle, in geometry in doing which we
;

shall show, that while for such cases the old rules are reproduced, the differen
tials

treated of need not be small

* A quaternion
may be said
when that tensor tends to zero.

to decrease,

and that
when

it

its tensor

would be a
decreases

and

vitiation,

and

not a

to decrease indefinitely,

ILLUSTRATION FROM ALGEBRA.

ARTS. 321, 322.]

433

any additional terms were introduced

correction, of the results, if

into their

expressions, for the purpose of rendering all the differentials equal to the cor

responding
in the

first

though some of them may be assumed

Example, one, and in the second Example, tivo.
differences

In Algebra, then,

(1.)

namely,

y = x\

Ay =

y +

(x

therefore, as usual,*

III.
or

so,

gives,
II.

and

be

us consider the equation,

let

I.

which

to

what comes

where n

to the

Ay =

same thing,

an arbitrary multiplier, which

be a positive whole number.
is

Conceive

(2.)

now

be supposed, for simplicity, to

may

A# and Ay, remaining always

that while the differences

connected with each other and with x by the equation III., decrease, and
tend together to zero, the number n increases, in the transformed
equation IV.,
and tends to infinity, in such a manner that the product, or multiple, n&x, tends
to

some finite

satisfy

limit a

which

may

A0 to

happen, for example, by our obliging

always the condition,

V.

after a previous selection of

(3.)

pression

We

A# =

some

shall then have,

by IV.,

a,

given

n-Ay

= 2xa + n~

and therefore a 2

is

nitely tends to zero,

and

and

b (since

its

nkx =

or

and finite

with this

increases indefinitely, the term n~ l

quantities, a

n~ l a,

for the equimultiple

VI.

But because

a,

value for

last condition V., the

wAy, of the other

l

a*

given
a~,

a.

+ n~ a\

and

difference,

if

following ex

Ay

= 2xa.

ft

while the

finite, (2.),

number n

in this expression VI. for wAy, indefi

limit is rigorously

Hence

null.

the two finite

be

supposed
finite), are two simultaneous limits,
to which, under the
supposed conditions, the two equimultiples, n&x and
*

is

to

2
write here, as is common, A# 3 to denote
x 2 would be written, on the same
(Ao;) ; while A
2
In like manner we shall write d# 2 , as usual, for (d#) 2 and shall
plan, for A (x }, or Ay.
-

"We

known

denote d(z 2 ) by d

HAMILTON

x-.

Compare the notations S? 2 S

ELEMENTS OF QUATERNIONS.

2
,

and V# 2

2
,

in 199

and 204.
3

ELEMENTS OF QUATERNIONS.

434

tend-* they are, therefore,

x and y

and we may

by the

And

it

d# =

elimination of

VIII.
(4.)

definition (320), simultaneous differentials of

write accordingly (321),

VII.
or, as usual, after

2.

ir.

[III.

= 2xa

a,

z
dy = d x = 2%dx.

dy =

a,

would wo improve, but

vitiate,

according to the adopted

defi

nition (320), this usual expression for the differential of the square of a variable x

in algebra, if

we were

for the difference

and

to

add to

For

x*.

it

?
the term d# in imitation of the formula III.
,

this

would come to supposing

A#, the fern w

finite value, a, of dx, or of

or w^da?

that, for a given

2

in the expres

VI. for n Ay, could fail to tend to zero, while the number, n, by which the
square of d# is divided, increases without limit, or feM^s (as above) to infinity.

sion

(5.)

As an

arithmetical example, let there be the given values,

IX. ..^ =
and

let it

y=

2,

;r

4,

d^ = 1000;

be required to compute, as a consequence of the

the simultaneous differential, dy.

arithmetical value of

definition (320),

We

the

have now the

following equimultiples of simultaneous differences,

X.

n Ax

= dx = 1000

wAy = 4000 + lOOOOOO^r

but the limit of the n th part of a million (or of way greater, but
number) is exactly zero, if w increase without limit the required

1
;

0w0>*

an

tYZ/we

of

dy

is,

therefore, rigorously, in this example,

XI.
(6.)

And we

see that these

XII.

dy = 4000.

two simultaneous
dx = 1000,

differentials,

dy = 4000,

are not, in this example, even approximately equal to the two simultaneous
differences,

XIII.

= d^ = 1000,

which answer to the value n =

Ay = 1002 2

2 2 = 1004000,

although, no doubt, from the very conception

* In this
a but
case, indeed, the multiple nAx has by V. a constant value, namely
convenient to extend the use of the word, limit, so as to include the case of constants
constant is its own limit.
generally, that a

it

is

found

or to say,

ILLUSTRATION FROM GEOMETRY.

AETS. 322, 323.]

embodied in the

of simultaneous differentials, as

they must

and

d#

as to be nearly equal, for large values of the

taneous and decreasing

definition (320),

of themselves taken,

admit of having such equisubmultiples

XIY.

435

n~ dy,

number

n, to

some system of simul

differences,

XV.

A#

and

Ay

and more and more nearly equal to such a system, even in the way of ratio, as
to vanish.
they all become smaller and smaller together, and tend together

For example, while the

(7.)

XYI.

n~ l

&x = 0-001,

and the same value

of

differentials

two simultaneous

and

number n

of the

dy = 0*004,

gives,

n = 1000000

if

by X., the equally rigorous values

differences, as follows,

XYII.

A0 =

and

0-001,

so that these values of the decreasing differences

to

themselves retain the constant

XII., their millionth parts are, respectively,

values

XY. may

be nearly equal to the two equisubmultiples,

simultaneous differentials, XII.

= 0-004001

Ay

XIY.

And it is evident that

already be considered
or

XYI.,

of the

be improved, by taking higher values of the number, n,

constant values XII., of d# and dy, being at all affected thereby.
= # 2 , and x =
(8.) It is, however, evident also, that after assuming y
in IX.,

we might have assumed any

instead of the value

1000

and should then have deduced a

but there would always

function, y,

and

exist, in this example, or for this

is

form

differentials,

of the

between

d# and dy,

XVIII.
which

different (but still

and not the formerly deduced value,

for this value of the variable, x, the rigorous relation

the two simultaneous

2, as

other finite value for the differential d#,

finite) value for the other differential, dy,

4000

two

approximation would
without the rigorous and
this

dy = 4d*,

obviously a case of the equation VIII., and can be proved by similar

reasonings.

323. Proceeding to the promised Example from Geometry (322), we shall

again see that differences and differentials are not in general to be confounded
with each other, and that the latter (like the former) need not be small. But

we

shall also see that the differentials

(like the difference*),

which enter into a

3

ELEMENTS OF QUATERNIONS.

436

[III. n.

2.

statement of relation, or into the enunciation of a proposition, respecting

quantities which vary together, according to
themselves

homogeneous among
homogeneous

?#A

differential, as

it

the variable to

being

any law or

which

it

correspond*,

and

of

should be

which

it

definition (320) enables us to construct the differential of a rectangle, as the

other (finite) rectangles, without

tiuo

of

and without even the thought

area,

of

the

is

It will also be seen that the

of line, or area of area.

line

laws, need not even be

sufficient that each separately

any reference

to

sum

units of length, or of

employing any numerical

calculation

whatever.
Let, then, as in the annexed figure 74, ABCD be any given rectangle,

(1.)

BE and DG be any arbitrary but given and

its sides, AB and AD.
Complete
the increased rectangle GAEF, or briefly AF, which

and

let

finite

increments of

will thus exceed the given rectangle AC, or CA,

the
or

sum

by

of the three partial rectangles, CE, CF, CG

by what we may call the gnomon* CBEFGDC.

diagonal CF take a point

i,

so that the line ci

Oil the

B H

may

be
Fig. 74.

any arbitrarily selected submultiple of that diagonal

and draw through T, as in the figure, lines HM, KL, parallel to the sides
AD, AB and therefore intercepting, on the sides AB, AD prolonged, equisub;

multiples BH,

DK

of the

two given increments, BE, DG,

of those

two given

sides.
(2.)

or that

Conceive

we

now

that, in this construction, the point

take a series of

new

on the given diagonal

approaches

to c,

CF, nearer and

points i,
nearer to the given point c, by taking the line ci successively a smaller and
Then the two new linear intervals, BH, DK, and
smaller part of that diagonal.

the neiv gnomon, CBHIKDC, or the

sum

of the three new partial rectangles, CH,

CK, will all indefinitely decrease,

and

will tend to vanish together

however, always a system of three simultaneous

two given
(3.)

sides,

But the

ci,

remaining,

differences (or increments), of the

AB, AD, and of the given area, or rectangle, AC.

given increments, BE and DG, of the two given sides, are always

(by the construction) equimultiples of the two first of the three new and decreasing
differences
they may, therefore, by the definition (320), be arbitrarily taken as
;

two simultaneous differentials of the

tivo sides,

AB and AD, provided

that

we

then

treat, as the corresponding or simultaneous differential of the rectangle AC, the

Book

The word, gnomon,

of Euclid.

is

here used \vith a slightly more extended signification, than in the Second

ILLUSTRATION FROM GEOMETRY.

ART. 323.]

new gnomon (2.), or of the decreasing

AI, whereof the first is given.

limit of the equimultiple of the

between the two

AI, or the

BE

sum

(2.) of

to BH, or of

For

difference

AC and

rectangles,

are then, first, to increase this new gnomon, or the difference of AC,

"We

(4.)

437

DG

to

the three partial rectangles, CH,

DK

and

by the

this last limit will,

ci,

CK, in the ratio of

secondly, to seek the limit of the area so increased.

definition (320),

sought differential of the rectangle AC

BE and DG, be assumed (as by (3.) they may)
to the

be exactly and rigorously equal

if the given

be the

to

and

finite increments,

differentials

of the

sides,

AB, AD.

Now when we

(5.)

we get

thus increase the

tico

new partial rectangles, CH and CK,

and CG which, as being

precisely the two old partial rectangles, CE

given and

constant,

must be considered

to be their

own limits*

increase, in the same ratio, the other new partial rectangle

ci,

But when we

we do

not recover

but obtain the new rectangle

the old partial rectangle CF, corresponding to it
the
or
which
is
not
constant, but diminishes indefinitely
equal rectangle CM,
CL,
;

as the point
this

new
(6.)

approaches to c

rectangle CL or CM,
//, then,

tials of the given

is

in such a

manner

the given increments, BE, DG, be

sides,

AB,

that the limit of the area, of

rigorously null.

assumed to be the

differ-

AD (an assumption which has been seen

to be

still

per milted), the differential of the given area, or rectangle, AC, is proved (not
assumed) to be, as a necessary consequence of the definition (320), exactly and
because such
rigorously equal to the sum of the two partial rectangles CE and CG
;

is

the limit
(7.)

(5.)

And

of the multiple of the new gnomon

if

any one were

to suppose that

value for the differential of a rectangle,

term, or part, so as to

make

it

by adding

(2.), in the construction.

he could improve this known

to it

the rectangle CF, as a

equal to the old or given

gnomon

(1.),

new

he would

commit a geometrical error, equivalent to that of

that
the
two similar rectangles ci and CF, bear to each other the
supposing
(the definition being granted)

simple ratio, instead of bearing (as they do) the

homologous

sides.

Compare the note

to

page 434.

duplicate ratio,

of

their

ELEMENTS OF QUATERNIONS.

438

SECTION

On some

[III. n,

3.

3.

general Consequences of the definition.

324. Let there be any proposed equation of the form,

and

let dq,

... be any assumed (but generally finite) and simultaneous

of the variables,
whether scalars, or vectors, or quaternions,
r,
.

differentials

<?,

on which

dr,

is

supposed to depend,

(or simultaneous) differential of

by the equation
their function,

tion 320, compare 321) to the following limit

II...

where n

is

expresses,

dQ = lim.

n {F(q +

n~ 1 dq,

Then

I.

Q,

is

the corresponding

equal (by the defini

+ n~ dr,...) - F(q,
l

r,

...)};

any whole number (or other positive* scalar) which, as the formula
conceived to become indefinitely greater and greater, and so to

is

And

tend to infinity.

in particular,

if,

volving only one variable

q,

IY.

Q,

as in

Q=

dQ - dfq = lim.

consider the function

so that

III...
then

we

n=

n {f(q + n

dq)

fq\

cc

a formula for the differential of a single explicit function of a single variable, which
agrees perfectly with those given, near the end of the First Book, for the
differentials of a vector, and of a scalar, considered each as a function (100) of

a single scalar variable,

but which

general definition (320), to the case

now

extended, as a consequence of the

when the connected

variables, q, Q,

and their

dQ, are quaternions: with an analogous application, of the

more general Formula of Differentiation II., to Functions of several Quater

differentials,
still

is

dq,

nions.

As an example

(1.)

its square, so

of the use of the formula IV., let the function of q be

that

V.

..

Q = fq =

q\

Then, by the formula,

YI.

* Except

may

dQ = dfq = lim.n

z
{(q+ n~ dq)
l

-q

z
}

=lim.

(q

.dq + dq.

q+n

in some rare cases of discontinuity, not at present under our consideration, this scalar n
as well be conceived to tend to negative infinity.

CONSEQUENCES OF THE DEFINITION.

AKT. 324.]

where dg 2

of
signifies* the square

VII.

dq
.

that

is,

= q dq + dq

(f

439

or without the pointsf between q and dq,

VII

= qdq + dq q

(f

an expression for the differential of the square of a quaternion, which does not
because q and dq are not generally
in general admit of any further reduction
:

When, however,

commutative, as factors in multiplication.

it

happens, as in

= dq q, by the two quaternions q and d^ being complanar,

algebra, that q dq
the expression VII. then evidently reproduces the usual form, 322, VIII., or
.

becomes,

(2.)

VUL

= 2?dft

As another example,

let

d?

if

q (123).

I]]

the function be the reciprocal^

IX...

Q=fq

q-i.

Then, because

X.

f(q +
=

(q

=of which,

when

multiplied by

n~ l dq)

-fq =

+ n-*dq)- 1
+

n~ l (q

(q+ n dq)

is

n~ l dq)~ l

the limit

n,

{q

+ ir dq)
1

(q

q~

}q~

dq
l

q~

dq

q~

we have

the following

expression for the differential of the reciprocal of a quaternion,

XI... d.q- = -q~ l .dq.ql

or without the points^ in the second

1
,

(as in

member, d^ being treated

VIF.)

as a whole symbol,

XI

d q
.

= - q

dq

q~

an expression which does not generally admit of being any farther reduced, but
becomes, as in the ordinary calculus,
XIL..d.<r

that
*

is,

=
-<r<l<7,

if

d?

for the case of complanarity, of the quaternion

Compare the note

HI ?,

and

its

differential^

to page 433.

f The point between d and q z , in the first member of VII., is indispensable, to distinguish the
But just as this latter square is denoted
differential of the square from the square of the differential.
2
the symbol, dq,
so the products, q dy and dq q, may be written as qdq and d^ q
briefly by
d<?

being thus treated as a whole one, or as if it were a single letter. Yet, for greater clearness of expres
sion, we shall retain the point between q and d#, in several (though not in all) of the subsequent
formulae, leaving it to the student to omit it, at his pleasure.
J Compare the note immediately preceding.
[See 329 (4.) for a result including XI.]

ELEMENTS OP QUATERNIONS.

440

[m. n

3.

325. Other Examples of Quaternion Differentiation will be given in the

the two foregoing may serve sufficiently to exhibit the

following section but
nature of the operation, and -to show the analogy of its results to those of the
;

older calculus, while exemplifying also the distinction which generally exists
between them. And we shall here proceed to explain a notation, which (at
least in the statement of the present theory of differentials) appears to possess

some advantages

and

will enable us to offer a still

more

brief symbolical defi

the differential of a function fq, than before.

nition^ of

We

have defined (320, 324), that if dq be called the differential of a

the limit of the multiple,
(quaternion or other) variable, q, then
(1.)

of

an

indefinitely

when taken

variable

q,

number,

n, is

is

decreasing difference of the function, fq,

of that

(single)

of the multiplying
relatively to an indefinite increase

the corresponding or simultaneous differential of that function, and

by the symbol dfq.

and while that
before we thus pass to the limit, relatively to n,
the
is still considered and treated as finite,
multiple I. is evidently

denoted, as such,
(2.)

But

multiplier, n,

a function of that number, n, as well as of the two independent variables, gand dq.
And we propose to denote (at least for the present) this new function of the
three variables,
II.

of

n, q,

which the form depends, according

on the form of the given function, f,

III.

in such a

manner as

to write, for

by
.

and

d^,

to the law expressed

by the formula

I.,

the new symbol,

../;(?,

any two

d ? );
variables, q

and q, and any number,

n,

the equation,

IV. ../(?, sO =

which may obviously be

V.

and

is

"

/(? +

*)- /)5

also written thus,

f(q +

"Y)

=/? + n Vn

(<?,

4),

here regarded as rigorously exact, in virtue of the definitions, and without

as small.
anything]whatever being neglected,

DISTRIBUTIVE PROPERTY.

ARTS. 325, 326.]

(3.)

For example,

it

VI-

and from 324,

M*>

And

flO

calculation in 324,

little

= Vt + fa + try 1

if

fq =

that,

(1.),

that,

(2.),

VII.
(4.)

appears from the

441

fn (q,

= -

n^\^ T \

(q

may now

the definition of dfq

fq =

if

>.

be briefly thus expressed

VIII. ..Afy-f. (q,Aq);

or, if

the sub-index

be understood,

IX.

we may

dfq =f(q, dq)

this last expression,/^, dq), orf(q, q

),

still

write,

more simply,

denoting thus & function of two inde

pendent variables q and /, of which the form is derived* or deduced (comp. (2.)),
from the given or proposed form of the function/*?, of a single variable, q, accord

ing to a law which

one of the main objects of the Differential Calculus

it is

(at

least as regards Quaternions) to study.

326.

One

class f(q, q

of the

dent variable, q

which

called derivation,^

which

may

most important general properties, of the functions of this

all distributive with respect to the second indepen

that they are

is
),

is

introduced in the foregoing process of what we have

a theorem
given function fq, of a single variable, q

from some

be proved as follows, whether the two independent variables be, or

be not, quaternions.
f
and let n be any number;
(1.) Let c{ be any third independent variable,
then the formula 325, Y. gives the three following equations, resulting from
the law of derivation of fn (q, q

from fq

I.../fe +
II.

..f(q +

III.

w-y +

../(<?

-1

O-A +

w-y) =/(?

4-

/feO5

rl

n-y) + n~%(q +

+ w-y + w-y) =fq + n-%(q,

n^q",

q")

* It was
remarked, or hinted, in 318, that the usual definition of a derived function, namely, that
given hy Lagrange in the Calcul des Fonctions, cannot be taken as a, foundation for a differential cal
culus of quaternions: although such derived functions of scalars present themselves occasionally in the
applications of that calculus, as in 100, (3.)
which will be noticed soon. The present

and

Law

(4.),

and in some analogous but more general

cases,

of an entirely different kind, since it

conducts, as we see, from a given f uction of one variable, to a derived function of two variables, which
are in general independent of each other.
The function/,,^, q), of the three variables, , q, q , may
also be called a derived function, since it is deduced, by the^^^ law IV., from the same given function

fq, although

it

has in general a

less

of Derivation

simple form than

f Compare the note immediately preceding.

HAMILTON S ELEMENTS OF QUATERNIONS.

its

own

is

limit,

(q,

},

oi f(y, q

}.

ELEMENTS OF QUATEKKEONS.

442

by comparing which we
IV.

[III.ir.g3.

see at once that

./(<?,

q +

=f,

(q

q")

~Y,

Q")

/(<?,

the form of the original function, fq, and the values of the four variables, q, q ,
q \ and n, remaining altogether arbitrary
except that n is supposed to be a
:

number, or at least a scalar, while

q,

may

q"

(or

not) be quaternions.

may

=
(2.) For example, if we take the particular function/^
the form 325, YI. of the derived function/,^, q ), we have

<f,

which gives

VI.

q +

.fn (q,

q(<f

?")

0")

(q

?")

1
? + ir

^+

8
g")

and therefore
VII.

as required

.fn (q,

q +

q")

-fn (q,

q")

= qq + q q +

2
,r>(<?

+ q

q"

by the formula IV.

n,

Admitting then that formula as proved, for all values of the number
we have only to conceive that number (or scalar) to tend to infinity, in order

to

deduce this limiting form of the equation:

(3.)

VIII.

./.

(q,

q")

-/.

(g,

/) + /.

?)

5")

or simply, with the abridged notation of 325, (4.),

IX.

f(q,

q +

q")

=f(q, q

+f(q,

q")

which contains the expression of the functional property, above asserted to

exist.
(4.)

(4.)

For example, by what has been already shown (comp. 325,

(3.)

and

),

X.
and XI.

if

fq = q\

if

fq = q

in each of which instances

we

then

then

f(q, q

= qq + q q

f(q, q

= )

q^q q^

see that the derived function f(q, q

them that

is distributive

although
happens to be
distributive with respect to q also.
(5.) It follows at once from the formula IX. that we have generally*

relatively to q

it

is

only in the first of

XII.
*

../(<?,

We abstract here from some

it

0)=0;

exceptional cases of discontinuity, &c.

DIFFERENTIAL QUOTIENTS, DEBITED QUATERNIONS.

AnTS.326, 327.]

and

not

it is

As

(6.)

a result including this, that

difficult to prove, as

XIII.

/(#, ff/)

%f(q,

<?

),

if

scalar.

observe that the defini

be expressed by the following formula (comp. 324, IV.,

may

and 325, IX.)

% be any

we may

a confirmation of this last result,

tion of f(q, q

443

XIV.

.f(q, ?

= lim
n{f(q +
=

n^)

-fq]

oo

we have

therefore, if x be

XV.

.f(q, xq

finite scalar,

any

and

xr n,

= x .lim .m{f(q + m^q

-fq};

OT=<

a transformation which gives the recent property XIII., since it is evident

that the letter
may be written instead of n, in the formula of definition XIV.

Resuming then the general expression 325, IX.,

327.

i.

we

see (by 326,

IX.) that

tion

may

fq

be.

dfq, of q and d^,

in fact

it

is

variable,

being

scalar, x, the differential dfq,

any

f(q, xdq)

evident,

it

is

(as

= mf(q

dq), if

from the very

dq of

the

of the function fq,

scalar, or that

x be any

conception

scalar.

and

(320) of

definition

must admit

of

changed together to any system of equimultiples, or equisubmultiples, of

themselves, without ceasing to be simultaneous differentials

that

always

whatever the form of the given func

simultaneous differentials, that every system of such differentials

all

is

distributive with respect to that differential

be multiplied, at the same time, by the same

to

II.

And

=/(?, a?),

also (by 326, XIII.), that if the differential

variable, q, be multiplied by

comes

d/?

XL)

an independent

We see

this derived function,

VII. and

in the examples 324,

dq, considered as

or writing anew,

permitted

to

or

more generally,

of a system, by any common

multiply all the differentials

scalar.
(1.)

It follows that the quotient,

III.

..dfq: dq =f(q,

dq)

dq,

of the two simultaneous

differentials, dfq and d^, does not change when the
differential dq is thus
multiplied by any scalar ; and consequently that this

quotient III.

is

independent of the tensor Tdq, although

pendent of the versor

Ud^,

if

and

d<?

be quaternions

it is

not generally inde

except that

it
3

remains

444

ELEMENTS OF QUATERNIONS.

in general unchanged,

when we merely change

(or negative), or to
is

[III. n.

that versor to

own

its

because this comes to multiplying dq by -

TJd<?,

3.

opposite
1,

which

a scalar.

For example, the quotient,

(2.)

IY.
in which d^

and

Ud^

dq = q + dq

q*

l
.

dq~

Udg

denote the reciprocals of dq and

.Ud^"

is

Ud<?,

very far from

we

being independent of dq, or at least of

Ud^

the sum of the given quaternion

of a certain other quaternion,

and

q,

latter, in its geometrical interpretation

from

being derived

angle

2z.d<?

by a

q,

and vary ivith

In general we may,

the versor

(comp. 191,

conical rotation of

so that both the axis

Ud<?,

since

A#

represents, as

it

(5.)),

be considered as

may

round A.X

and the quantity of

see,

which

dq,

through an
on

this rotation depend

that versor.

if we please, say that the quotient III. is a

but
we
Differential Quotient
ought not to call it a Differential Coefficient
(comp. 318), becouse dfq does not generally admit of decomposition into two
(3.)

factors,

whereof one shall be the

differential dq,

and the

other

a,

function of q

alone.
(4.)

And

same reason, we ought not

for the

to call that Quotient

Function (comp. again 318), unless in so speaking

we

a Derived

understand a Function

of Two* independent Variables, namely of q and Ud^, as before.

(5.) When, however, a quaternion, q, is considered as & function of a scalar
variable,

t,

so that

we have an equation
Y.

it is

then permitted (comp. 100,

VI.

d?

where

ft,

denotes a scalar,

and

(3.)

dt = dft

of the form,

(4.)

d* = lim
ns

to write,

2.[f (* + **}
dtff \

nj

ft\
)

-lim.

and

to call this limit, as usual, a derived function oft, because

function of that scalar variable,

t,

alone,

and

is

independent of

tial, d*.

Compare the note

to 325, (4.).

it is

(in fact)

the scalar differen

ARTS. 327,328.]

DIFFERENTIAL QUOTIENTS, DERIVED QUATERNIONS.

We may

(6.)

VII.
and

dq =

Dq
t

formula, because the scalar

But

as regards the

ice

D ft,

or

quotient VI., obtained

X.

At,

(in fact) multiplied by

differential,

by
or

dft,

it,

dq or dft.

this derived function,

regard

IK. ..ft,
an actual

=ft

usual, a differential coefficient

differential, dt, is

coefficient,

as being

differential equation,

of the question (comp. again 100, (3.)

/00r*c

remember that

important to

dfq

or/% as

found for the quaternion

in the expression thus

(7.)

VIII.

or

dt,

call the derived quaternion, Dtq,

may

in. tf/aw

under these circumstances, the

also write,

445

),

it is

or differential

D q,

or

dividing an actual quaternion,

dq,

by an actual scalar, dt, of which the value is altogether arbitrary, and may (if
while the dividend quaternion
choose) be supposed to be large (comp. 322)

we
X.

depends, for

and on

its

value, on the values of the two independent scalars,

the form of the function ft,

according to the law

which

is

and

dt,

expressed by

the general formula 324, IV., for the differentiation of explicit functions of any
single variable.

328. It
tions

is

easy to conceive that similar remarks apply to quaternion func

and that when the differential of such & func

of more variables than one

tion is

expressed (comp. 324, II.) under the form,

I.

dQ

dF(q,

r, s,

.)

= F(q,

r, s,

dq, dr, ds,

.),

the new function jFis always distributive, with respect to each separately of the

being also homogeneous of the first dimension (comp.

with
in such a
to
all
those
differentials, considered as a system
327),
respect
manner that, whatever may be the form of the given quaternion function, Q, or

differentials dq, dr, ds,

F, the derived* function F, or the third member of the formula

this general functional property (comp. 326,

II.

where x
tials

F(q,

may

r, s,

xdq, xdr, xds

be any scalar

dq, dr, &c., of

depends, are excluded

q,

r,

.)

XIII., and 327,

= xF(q,

r, s,

must possess

II.),

dq, dr, ds,

.),

so that products, as well as squares, of the differen

&c. considered as so

from the expanded

many

Compare the note

variables

on which

expression of the differential

function Q.
*

I.,

last referred to.

dQ

of the

ELEMENTS OF QUATERNIONS.

446

For example,

(1.)

if

III.
it is

3.

the function to be differentiated be a product of two

quaternions,

then

[III. n.

Q=

F(q, r)

qr,

found from the general formula 324, II., that (because the
dq dr is null, when the number n -increases without limit) the

easily

limit of n~

differential of the function

IY.

= d qr =
dF(q,

dQ

with analogous

is,

r)

= F(q,

dq, dr)

r,

= q

dr + dq r
.

more than two quater

results, for differentials of products of

nions.
(2.)

if

Again,

we take

this other function,

Y.

= F(q,

= q~\

r)

then, applying the same general formula 324, II., and observing that we
have, for all values of the number (or other scalar), n, and of the four quater
nions, q, r, q

VI.

we

the identical transformation (comp. 324,

n[(q + n^q

)-

+ n r)

(r

find, as the required limit,

tial

of the function

VII.

which

dQ =

q~ r]

q~ r

when n tends

(q

(2.) ),

n~ q )~

l
q q~ (r +

wV),

to infinity, the following differen

q~ r

dF(q,

r)

= F(q

dq, dr)

r,

q~

dr -

q~

dq

q~ r

again, like the expression IV., distributive with respect to each of the
differentials d^, dr, of the variables q, r, and does not involve the product of
is

those two differentials: although these two differential expressions, IV. and
and VII., are both entirely rigorous, and are not in any way dependent on

any supposition that the tensors of dq and dr are small (comp. again 322).
329. In thus differentiating a function of more variables than one, we are

what may be called Partial Differentials of Functions of two

more Quaternions which may be thus denoted^

led to consider

d 3 Q,

or

which is here supposed to be

be a function, as above, of q, r, s,
differentiated with respect to each variable separately as if the others were
constant.
And then, if dQ denote, as before, what may be called, by contrast,

if

the Total Differential of the function Q,

II.

dQ

we

shall

have the General Formula,

= d Q + d r Q + ds Q +
q

PARTIAL AND TOTAL DIFFERENTIALS.

ARTS. 328, 329.]

447

and symbolically,

or, briefly

III.

d = d, + d r + d s +

denote the quaternion variables on which the quaternion function

depends, of which the total differential is to be taken whether those variables
be all independent, or be connected with each other, by any relation or relations.

if q, r, 8,

(1.)

For example (comp. 328,

IV.

Q=

if

qr,

then d q Q = dq

and the sum of these two partial

dQ, as otherwise found above.

Again (comp. 328,

(2.)

Y.

Q=

if

r,

and d r Q =

Q makes up

differentials of

dr

its total differential

then d q Q = -

q~

as that which

dQ

dq

d r Q = q~ l dr

q~ r

was otherwise found

before, for

form of the function Q.

this

and

the possibility of a relation existing between the variables

we
let those variables be now supposed equal to each other in Y.

To exemplify

(3.)

(2.)),

q~ r,

and d q Q + d r Q = the same

(1.) ),

r,

have

shall then

Q=

dQ

1,

and accordingly we have here d q Q = -

q~

dq

-d r Q.
= c = any constant quaternion
Again, in IY., let qr
= dQ = d q Q + d r Q-, and may infer that
again have
(4.)

VI.

dr = -

qr

dq.

if

r,

qr

const.

we

shall then

a result which evidently agrees with, and includes, the expression 324, XI.,
for the differential of a reciprocal.
(5.)

A quaternion

more scalar

q,

variables,

t,

u,

Article, its partial differentials,

may

to be expressed as a function of two or

may happen

and then
d

it

will have, as such,

But

d u q, &c.

q,

by the present
we

because, by 327, VII.,

in this case write,

VII.
where the
calculus),

coefficients

we

dtq

Dq
t

d uq =

d*,

D uq

dw,

are independent of the differentials (as in the ordinary

have (by

shall

II.)

an expression for the

total differential dq, of

the form,

VIII.

dq =

d# + d u q

Dq
t

dt + ~D u q

du +

ELEMENTS OP QUATERNIONS.

448

and

may

[III. n.

3-

at pleasure say, under the conditions here supposed, that the derived

quaternions,

IX.

D q,

are either the Partial Derivatives, or the Partial Differential Coefficients, of the

Quaternion Function,

X.

with analogous remarks for the

289) into a vector, p.
330. In general,

c,

when the

case,

320, that the differential of a constant

constant quaternion,

u,

.)

quaternion, q, degenerates (comp.

be considered as evident, from the definition in

may

it

= F(t,

is zero

L, then

in the equation 324,

so that if

And

dQ

he changed to any
to be replaced by 0,

is

there be given any system of

treat the
equations, connecting the quaternion variables, q, r, s, ... we may
the
for
as
system of
holding good,
system of differentiated equations,

in the differentiated equation, 324, II.

if

corresponding

simultaneous differentials, dq, dr, ds,

therefore, legitimately in

and may

theory, whenever in practice it shall be found to be possible, eliminate any

one or more of those differentials, between the equations of this system.
(1.) As an example, let there be the two equations,
I.

qr =

and

c,

II.

= r2

Then (comp. 328,

denotes a constant quaternion.
have the two differentiated equations corresponding,

where

we

III.

in which the points*

dr + dq

and when we substitute

IY.

The equation

I.

dr = -

gr

&q

ds = r

The former
.

r,

this

new

d<?

q~

dq

.r =

= (q^cf =

Compare the second note

l
<r

cq~

we thereby

the equation II. gives therefore this other expression,

.

(1.)

gives,

gives also the expression,

VII.

and 324,

dr + dr r

differential equation,

ds = - rq~ l

VIII.

VI.

as in 329,

this value in the latter,

and obtain
VI.

(2.)

might be omitted.
V.

differential dr,

(1.),

c,

to 324, (1.).

2
.

eliminate

the

ELIMINATION OF DIFFERENTIALS.

ARTS. 329,331.]

by

And

elimination before differentiation.

the expressions VII. and VIII.

equation,

IX.

= -

(qr c)*

q- cq~

dq

we

s,

q~ c

q~

dq

l
q~ cq~ c

331.

No

(1.)

and

(2.) ),

substitute

get this other differential

which might have been otherwise obtained (comp. again 324,

under the form,
d (q^eY = q- c d(q- c) + d(q~ c) q~ c.
X.
.

we

in the formula VI.,

if,

and

for r

449

special rules are required, for the differentiation of functions of

functions of quaternions

but

be instructive to show,

may

it

briefly,

how

the

consideration of such differentiation conducts (comp. 326) to a general property

offunctions of the

class f(q,

<?

and how that property can be

otherwise esta

blished.
(1.)

Let/,

$,

and $ denote any functional operators, such that

II.

r =fq,

and III.

we may

is,

we have IV.

<f>r,

whence V.
That

^=

then writing

ds =

differentiate the

(as usual)

i/#

compound function,

fy(fq), as iffq

were an independent variable, r; and then, in the expression so found, replace

obtained by differentiating the simple function, fq.

For this comes virtually to the elimination of the differential dr, or of the
symbol dfq, in a way which we have seen to be permitted (330).
the differential

(2.)

dfq by

But, by the

its

value,

definitions of

dfq

and/n (g,

),

we saw

that the differential dfq might generally be denoted

by

f(q, dq)

whence d^r and

may

dipq

also,

VIII. IX.)

(325,

f*(q, dq), or briefly

by

by an extension

of the

notation, be represented by the analogous symbols, 0*(r, dr) and

or simply
(3.)

by

We

0(r, dr)

and

same

\l>(q,

dq),

^(q, dq).

ought, therefore, to find that

VL...t.(q,dq)-1>.(fq,f.(q,W),

if

^=$

fq)

or briefly that

VII.

..i

for any tico quaternions,

and any two

if

-,

provided that the

fn (q, q ), n (q, q )^ ^ n (q q are deduced (or derived) from the func
tions fq, tyq, \f,q,
according to the law expressed by the formula 325, IV.
q,

f unctions

q>

functions, f,

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

450

and that then the

when

which these derived functions

limits to

the number n tends to infinity, are denoted

f

symbols, f(q, q

fn (q,

3, 4.

q}, &c. tend,

these other functional

by

&c.

),

To prove

(4.)

[III. n.

this otherwise, or to establish this general property VII., of

functions of this class f(q, q

),

without any use of differentials,

we may observe

that the general and rigorous transformation 325, V., of the formula 325, IV.
by which the functions fn (q, q } are defined, gives for all values of n the

equation

VIII.
but

also,

by

= (fq + n\ Tn (q, q
tf(q + irY)
$

tlie

hence generally, for

^ (q + wY) = $q +

of the number

all values

two independent quaternions,

q,

and

n^

(<l,

(5.)

n (q,

n, as

well as for all values of the

f,

tn (q,

<t>(fq,f,

(q,

),

^ - tfq

if

we
<f>,

oo, is

(with the notations used)

It is scarcely worth while to verify the general formula X.,

particular

example

an

yet, merely as

exercise,

it

may

by any

be remarked that

if

take the forms,

XI... fq = f,
which the two

first

XII.
the formula

give,

=
<t>q

^q = q\

q* 9

by 325, VI., the common derived form,

.fn (q,

=
}

1>

n (q 9

q)

= qf + qq +

n^\

X. becomes,
XIII.

= ?(qtf + q q +

wY)

+ H (q,

=
)

^ n (q\ qq + q q +

+ to + fa +

z
)

+ n

~l

nV)
(tf +

fa +

~l

n v

which agrees with the value deduced immediately from the function

by

q)

for all forms of the two functions,

an equation of which the limiting form, for n =

the equation VII. which was to be proved.

of

n^ n (fq,fn

write,

X.

we

= tfq +

same general transformation,

IX.

may

) )

the definition 325, IV.,

xiv.

z
)

-^q

or

namely,

feO-Mto+^W-^J-M^+ ^W+^+^T-b*)

!-

q*,

DIFFERENTIAL OF A PRODUCT OR POWER.

ARTS. 331, 332.]

rule, for differentiating as in (1.)

In general, the theorem, or

(6.)

451

of a Junction, of a quaternion or other variable,

cally expressed by the formula,

may

& function

be briefly and symboli

XV. ..d (*/)? = d*(/?);

and

if

we

supplied,

know

did not otherwise

by the recent proof

a proof of

it,

its

correctness

of the correctness of the equivalent

SECTION

would be

formula VII.

4.

Examples of Ctuaternion

Differentiation.

now

be easy and useful to give a short collection of Examples

of Differentiation of Quaternion Functions and Equations, additional to and
inclusive of those which have incidentally occurred already, in treating of the
332. It will

principles of the subject.

If c be

(1.)

constant quaternion (as in 330), then

any
1.

III.

if

dc = 0;

cfq

= cdfq

II.

IV.

d (fq.c] = dfq.

c.

In general,

(2.)

V.

&(fq +

<J>q

...)

= &fq +

d(j>q

or briefly,

VI.

dS =

2d,

be used as a mark of summation.

(3.)

VII.

Also,

and similarly

d(fq 00) = dfq

.

.
<j>q

+ fq

more functions than two the rule being

its own place, or without dis

for a product of

simply, to differentiate each factor separately, in

turbing the order of the factors (comp. 318, 319)

and then

to

add together ths

partial results (comp. 329).

(4.)

In

if

particular,

VIII.

be any positive whole number,

m = mq
dq +

m ~z

and because we have seen (324,

ix.

(2.)

dq q
.

+ qdq q m
.

~2

+ dq

m~ L
,

that

= -

d?

x
<r

>

3M2

ELEMENTS OP QUATERNIONS.

452

we have

[III.ii.

4.

this analogous expression for the differential of a power of a quater

nion , with a negative but whole exponent,

= (5.)

To

we

m = -

q~

l~

^ d?

m -

04

+ d

0*

0*

04

0~
.

we

differentiate a square root,

XL
if

q~

X.
m dq.q-

(10

md

m
.

q~

^ d?

q~*

m
q~

dq

q~

are to resolve the linear equation*

;

XI

or

rr + r r =

write, for abridgment,

XII.
(6.)

Writing

also, for this

XIY.

it

d0,

/=

04,

= dr.

Kr

will follow that

= ISY = Tr 2 = T0,

rs

purpose,

XIII.
whence (by 190, 196)

XV.

and

= 2Sr = 2S

the product and s?/w of these two conjugate quaternions, r and

scalars (140, 145),

we

have, by

XI

XVI.

rqs

0*,

being thus

s,

.,

r- q s

= r

whence, by addition,

XVII.
and

r
.

(r

+s) r +

(r

= 2/(r +

s)

finally,

XVIII

/-"

2(r +

or
s)

XIX

4S

an expression for the differential of the square-root of a quaternion, which

will be found to admit of many transformations, not needful to be considered
here.
it has
(7.) In the three last sub-articles, as in the three preceding them,
been supposed, for the sake of generality, that q and dq are two diplanar

Although such solution of a linear equation, or equation of the first degree, in quaternions, is
enough accomplished in the present instance, yet in general the problem presents difficulties,
without the consideration of which the theory of differentiation of implicit functions of quaternions
would be entirely incomplete. But a general method, for the solution of all such equations, will be
easily

sketched in a subsequent Section.

DIFFERENTIALS OF TRANSCENDENTAL FUNCTIONS.

ARTS. 332, 333.]

453

any application they happen, on the contrary, to be

complanar, the expressions are then simplified, and take usual, or algebraic
quaternions

but

forms, as follows

XX.
and

when

xxn.

in

if

XXI.

a.

m~
m =
mq dq

2*

q^dq,

XXIIL

if

m = m~
mq~ ^q
l

q~

<ty|||?(i23)

or with r, in the
complanar with g, and therefore with
the
numerator
of
that
expression XVIII.,
expression may be written as r~ q
r + s).

because,

is

<^,

More

(8.)

ordinary calculus, but

XXIV.
333.

may

if

generally,

still

The functions

x be any scalar exponent, we may write, as in the

under the condition of complanarity XXIII.
,

f = xq

x~ l

XXV.

or

?d

f = xqx &q.

which have been lately differentiated,

form the following are a few examples of

of quaternions,

be said to be of algebraic

what may be

differentials of

dq

by

called,

transcendental functions of

contrast,

quaternions: the condition of complanarity (&q\\\q) being however here sup

posed to be satisfied, in order that the expressions may not become too com

In

plex.

most

they will be found to assume, for the

fact, with this simplification,

part, the known and usual forms, of the ordinary differential calculus.

Admitting the definitions in 316, and supposing throughout that

q
a^HI^, we have the usual expressions for the differentials of t and \q, namely,
(1.)

I.

(2.)

We have
III.

(3.)

Also,

if

IV

also,

Mq

II. ..

by the same system

sin q = cos qdq

IV.

=
dl<?

r
.

= d

rl ?

q~ dq.

of definitions (316),
.

d cos q = -

and dr be complanar with q and

r
.

rlq

sin qdq

d?, then,

q (\qdir

&c.

by 316,

q~ r&q)

or in the notation of partial differentials

(329),

V.
(4.)

dq

In particular,

r
.

q
if

= rq r dq,
l

and

VI.

dr

r
.

q lqdr.

the base q be a given or constant vector,

exponent r be a variable scalar,

t,

then (by the value 316,

recent formula IV. becomes,

VII.

aM ITa +

Ua Wdt.

a,

XIV.

and

if

the

of \p] the

ELEMENTS OE QUATERNIONS.

454

If then the base a be a given unit

(5.)

we may

so that

line,

[HI.

ITa =

0,

and

Ua

4.

a,

write simply,

VIII.

a*

if

d*,

da =

0,

Ta =

and

1.

4>

This useful formula, for the differential of a

power of a constant unit
with a variable scalar exponent, may be obtained more
rapidly from the

(6.)

line,

equation 308, VII., which gives,

T-\7-

1A.

since

it is

pression

t
.

a*

tir

= COS -- + a

Sin

ITT
-

-,

if

Ta =

evident that the differential of this expression

2
- 1.
multiplied by J?rad^ because a =

is

equal to the ex

itself

(7.) The formula VIII. admits also of a simple geometrical interpretation,

connected with the rotation through t right angles, in a plane perpendicular to
or the versor ~UV, is considered
a, of which rotation, or version, the power a
,

(308) to be the instrument* or agent, or operator (comp. 293).

334. Besides algebraical and transcendental forms, there are other results
of
operation on a quaternion, q, or on a function thereof, which may be regarded
as

forming a new

class (or

kind) of functions, arising out of the principles and

rules of the Quaternion Calculus itself: namely those which we have denoted
in former Chapters

by the
I.

or

symbols,

TSq, Sq,

V?

Nq, 1g, Uj,

by symbols formed through combinations

of the

same

signs of operation,

such as
II.

And

it

is

these forms,

essential that

SU0, VU?, UV?, &o.

we should know how

to differentiate expressions of

which can be done in the following manner, with the help of the
and former Chapters, and without now assuming the

principles of the present

complanarity , dq

|||

q.

distributive symbol,
(1.) In general, let /represent, for a moment, any
the
have
that for any two quaternions, q and q we shall
equation,
,

in. ../
*

Compare the

first

=+
note to page 135.

so

DIFFEBENTIALS OF

AETS. 333, 334.]

and therefore

455

]fy.

also* (comp. 326, (5.)),

IY.

= fq,

f(xq)

if

^ be any scalar.

Then, with the notation 325, IY., we shall have

(2.)

and

AND

Sy, Vy,

therefore,

any such function

for

by 325, YIII.,

we

fq,

shall

have the

differential expression,

VI. ..dfq=fdq.

But

(3.)

S,

Y,

K have been seen to be distributive symbols

(197, 207)

we

can therefore infer at once that

VII.

dK<?

Kd?

VIII.

dS? = Sdg

IX.

..

or in words, that the differentials of the conjugate, the scalar,

conjugate, the

the

quaternion are, respectively,

scalar,

and

dV? = Yd?
and

the vector of a

the vector of the

differential of that quaternion.

To

(4.)
,

we

deduce an expression for

to

have (by VII. and- 145) the equation,

X.

dN? =

qKq =

but

and

N#, or

find the differential of the norm,

(1

4-

K)

= 2S q

Kq

qKq = dq Kq +
.

Kq, by 145, and 192,

Kq

XI.

therefore

= 2S(K? q
.

),

II.

by 196,

dN? = 2S(K?

Kdq

II.,

and 198,

I.

dj).

more im
(5.) Or we might have deduced this expression XI. for
mediately, by the general formula 324, IY., from the earlier expression 200,
VII., or 210, XX., for the norm of a sum, under the form,
dN<?,

XI

dNq =

lim n{N(q + n~ l dq) .

N<?)

CD

= lim
{2S(Kq d?) + n~ Ndq}
l

as before.
* In
quaternions the equation III.

is

not a necessary consequence of IV., although the latter

so of the former; for example, the equation IV., but not the
equation III., will be satisfied, if

assume /#
scalars.

qcq-

(/q,

where

and

are

any two constant quaternions, which do not degenerate

is

we

into

ELEMENTS OF QUATERNIONS.

456

The

(6.)

Tq and

N<?

tensor,

Tq,

[III. n.

the square-root (190) of the norm,

is

are scalars, the formula 332,

XXTI. may

N#; and

be applied

4.

because

which gives,

for the differential of the tensor of a quaternion, the expression (conip. 158),

XII.
a result which

is

more

dT ?

S(KU?

s^.

T?

The

(7.)

vcrsor

nion q divided by

whence follows

U^

its

is

- S

under the form,

easily remembered,

xm...

dj)

equal (by 188) to the quotient, q: Tq, of the quater

Tq hence the differential of the versor is,

tensor

at once this formula, analogous to

XIII., and like

it

easily

remembered,

XV
q

have generally
(8.) We might also have observed that because (by 188) we
=
therefore
we
have
also,
Tq Ug,
q
(by 332, (3.))
.

XVI.

d? = dTq

U?

+ Tq dU?,
.

and
_

XVII

dTTj.

7~T

any manner established the equation XIII., we can im

mediately deduce XV. and conversely, the former equation would follow at
once from the latter.
if

we have

then

in

(9.)

It

be considered as remarkable, that we should thus have

the formula:*
and

may

generally, or for any two quaternions, q

XVIII.
*

S(dU</

:U?) =

d<?,

or

XVIIF.

When the

dU?

:TJ?

= S^O

connexion of the theory of normals to surfaces, with the differential calculus of quater
have heen (even briefly) explained in a subsequent Section, the student will perhaps be
able to perceive, in this formula XVIII., a recognition, though not ?, ie?y direct one, of the geo

nions, shall

metrical principle, that the radii of a sphere are

its

normals,

DIFFERENTIALS OF TENSOR AND YEESOE,

ARTS. 334, 335.]

but this vector character of the quotient

dU?

~Uq can easily be confirmed, as

we

Taking the conjugate of that quotient,

follows.

457

have,

by VII. (comp. 192,

II.; 158; and 324, XI.),

XIX.

K (dU?

1
.

IT?

KUq~

dKU?

U?

d(Uq-

= -

1
)

dU?

ll?

whence

XX.
which agrees (by 196,

The

(10.)

II.)

+ K) (dU? U?-

(1

with XVIII.

scalar character of the tensor,

enables us always to write, as

T<?,

in the ordinary calculus,

XXI.
but IT? = SI?, by 316, V.
be thus written,

dlT? = dT?

When

equation

XV.

the recent formula XIII.

XXII. ..Sdl? = dSl? = dT?:d? = S(d?:?);

(11.)

Tq

may therefore, by VIII.

XXIF.

or

d?|||?, this last difference vanishes,

dty-^d^S^O.

by 333, II.; and the

takes the form,

XXIII.

dlU? = Vdl? = dVl?.

= VI?, by 316, XX.,

generally, lUq
although the differen
two equal expressions do not separately coincide with the members
of the recent formula XV., when q and d? are diplanar.
We may however

And

we have

in fact

tials of

these

write generally (comp.

XXIV.
335.

We

are formed

XXII.),

dlU? - dU? :U? = V(dl? - d?

:?)

= dl? - d?

have now differentiated the

?.

six simple functions 334,

the
of
the
six
by
operation
characteristics,

which

I.,

E,8,V,N,T, U;
and

compound functions 334, II., which

combinations of those former operations, it is easy on the same

as regards the differentiation of the

are formed

by

principles to determine them, as

(1.)

The

axis

represented by the combination

fore,

may

be seen in the few following examples.

Ax. q of a quaternion has been seen (291)

TJV?
by 334, IX. and XIV., be thus expressed

HAMILTON

ELEMENTS OF QUATERNIONS.

to

admit of being

the differential of this axis

there

may

ELEMENTS OF QUATERNIONS.

458
I.

d(Ax.

q)

= V(Yd? Vg)

dUV?

[III. n.

UV?

4.

whence
d(Ax.g)
"

dUVg ~

UV?

Ax.</

The

of the axis

differential

therefore, generally y a /me perpendicular to that

is

axis, or situated in the plane of

tJie

quaternion

(and therefore the axis) of that quaternion

and

its differential

(2.)

but

constant

is

It

dUVg

or

when the plane

when the quaternion

this complanarity

IV.

IV. ..dg|||g;

if

0,

may

be expressed by the equation III.

easy to prove, on similar principles, that

is

VdU?

V.

dVU?

VI.

dSU? = SdU? = S

V^
V
\

223,

vanishes,

Hence,

and conversely

(4.)

it

are complanar.

III.

(3.)

Ydg
Vq

Ug

But

(5.))

in general, for any two quaternions, q and g

the transformations,

VII.

S(Vg

g)

= S (Vg

V?) = S

.
<?

V?

we have (comp.

and when we thus suppress the characteristic V before

q, and insert it
before Ug, under the sign S in the last expression VI., we may replace the
:

d<?

new
or

by

outside
or

VU? by TVUg.UVUg (188), or by TVTJg.UVg (274, XIII.),

TVU? UVg (204, V.), where the scalar factor TVU? may be taken
also for q
UVg we may substitute 1 (UVg q),
(by 196, VIII.)

factor

:
<?UV<?,

because

UV<?|||

the formula VI.

may

therefore be thus written,

(5.)

Now

may

it

be remembered, that

among

the earliest connexions of

quaternions with trigonometry, the following formulae occurred (196, XVI.,

and 204, XIX.),

IX.

we had

also, in

SUg =

cos L

q,

TVU?

= sin L q

316, these expressions for the angle of a quaternion,

X.

.z

ARTS. 335, 336.]

we may

DIFFEBENTIALS OF AXIS AND ANGLE.

459

therefore establish the following expression for the differential of the

angle of a quaternion,

XL

The following

(6.)

d L q = dTVl? = dTlU? = S

is

another

way

dq

qUVq

of arriving at the

same

result,

through

the differentiation of the sine instead of the cosine of the angle, or through
For this purpose, it is only
the calculation of clTVU?, instead of dSUg.

necessary to remark that we have, by 334, XII. XIV., and by some easy
transformations of the kind lately employed in (4.), the formula,

XII

dTVU<7-S
U
b

?-

dividing which by

and attending

STJ<?,

to

IX. and X., we arrive again

at

the expression XI., for the differential of the angle of a quaternion.

Eliminating S (dq

(7.)

qUVq) between VIII. and XII., we

obtain the

differential equation,

XIII.
of which,

STJ0

dSU? + TVU? dTVU? =

.

on account of the scalar character of the differentiated variables, the

integrals evidently of the/orw,

XIV.

2
2
= const.
(SU?) + (TVU?)

and accordingly we saw, in 204, XX., that the sum in the

first

member

of

this equation is constantly equal to positive unity.

The formula XI. may

(8.)

XV.
with the verification, that

also be thus written,

.dZ2

S(V(d<?.):UV<ri;

when we suppose

dglH^, as in IV., and therefore

dUVg
by III., the expression under the ,sign S becomes the differential
of the quotient, VI? UVg, and therefore, by 316, VI., of the
angle L q itself.

336.

An

important application of the foregoing principles and rules

consists in the differentiation of scalar functions of vectors,

are defined

and expressed according

to the

when

those functions

laws and notations of quaternions.

It will be found, in fact, that such differentiations play a very extensive part,
in the applications of quaternions to geometry but, for the moment, we shall
;

treat

them

here, as

merely exercises of calculation.

The following

are a few

examples.
3

ELEMENTS OF QUATERNIONS.

460

Let p denote, in these

following equation be proposed,

sub-articles, a variable vector

(1.)

I.

r*

p*

by 279,
dp to

we

be,

II.

In

pp + p p

III.,

have,

fact, if r

Vr =

in which

0,

so that r is a (generally variable) scalar.

[III. n.

and

Differentiating,

and observing

by 322, VIII., and 324, VII., the

rdr + Spdp =

or

be supposed

III.

dr = -

equation,

r~

S P dp = rSp~

dp.

by 282, II., the tensor of p so

included in the general formula, 334,

positive, it is here,

that this last expression III. for dr

that,

we suppose

p be any second vector, such as

if

the

let

0,

2$pp

4.

is

XIII.
If this tensor,

(2.)

be constant, the differential equation II. becomes

r,

simply,

IV.

let

Again,

(3.)

V.
so that

more

or

2r 3

d>*

T (tp

+ p K ),

or

const.,

d =

with

K are here two constant vectors.

334,

XL,

because Kip

= idN( + K = S
tj0
p )
(pt +

if

dT/>

0.

K p) (idp

d* =

0,

0,

Then, squaring and differen

pi, &o.),

+ dp K ) =

2
(

+ K

briefly,

VII.
if

- Pz =

if

0,

the proposed equation be (comp. 282, XIX.),

we have (by

tiating,

VI.

and

Spdp

2r- J dr

v be an auxiliary vector, determined

VIII.

.r*v =

by the
a

(i

equation,

+ K)p+.2Viff/M;

which admits of several transformations.

For example we may

(4.)

IX.
or,

by 294,

X.

r*v

III.,

z
(i

r*v

(t

2
K-

write,

p +

by 295, VII.,

K pt

tp K

(ip

-f

pK)

K (pi

+ Kp)

and 282, XII.,

+ K z )p + 2( K Sip-S LK + L $ K

= (i-

K}

p + 2(iSK+KSip)

&o.

SCALAR FUNCTIONS OF YECTOES.

AETS. 336, 337.]

The equation Y.

(5.)

when squared without

(comp. 190, V.),

gives

461

differentiation,

XI.

f*

=
N(<p

by transformations

of the

recent expressions for

XII.

When

(6.)

(t

+PK)

+K

Z
)

p*

same kind

(ip

+ Kp)

/OK) (pi

IpKp

as before

we have

by the

therefore,

the following remarkably simple relation between

r*v,

the two variable vectors, p

and

.Sp

the scalar,

r,

v,

= l;
is

XII

or

we

constant,

Sv =

1.

by YIL, the

have,

differential

equation,

XIII.

Svdp =

whence

also

XIY.

Spdv

by XII.

0,

a relation of reciprocity thus existing, between the two vectors p and

the geometrical signification will soon be seen.

of

which

Meanwhile, supposing r again to vary, we see that the

(7.)

for 2r 3 dr

YI.

v,

either of the

may

two

last expression
be otherwise obtained, by taking half the differential of

last

expanded expressions XI. for

it

being remembered,

in all these little calculations, that cyclical permutation of factors, under the

permitted (223, (10.)), even

ever their number may be and that
sign S,

is

if

those factors be quaternions, and what

if they be vectors, and if their number be

then permitted, under the sign Y, to invert their order (295, (9.) ),
and so to write, for instance, V/pc instead of Yicpt, in the formula YIII.
:

odd,

it

is

(8.)

As

another example of a scalar function of a

vector, let

proximity (or nearness) of a variable point p to the origin o

XY.

- (- 2
p )-* = Tp-\

or

XY

denote the

so that

*
2
p~ + P =

0.

Then,

XYI.
v being here a

dp = Svdp,

new auxiliary

(YIII.), and having (as

vector p itself, but having

we
its

if

XVII.

vector, distinct

= p* P = /Up

from the one

lately considered

as the
see) the same versor (or the same direction)
tensor equal to the square of the proximity of~ptoo ,

or equal to the inverse square

of the distance, of one of those two points
the other.

337.

On

the other hand,

we have

from

often occasion, in the applications, to

consider vectors as functions

of scalars, as in 99, but

now with forms

arising

ELEMENTS OF QUATERNIONS.

462

[III.ii.

4.

out of operations on quaternions, and therefore such as had not been considered
And whenever we have thus an expression such as either

in the First Book.

two following,

of the

I.

or

II.

<t>(t),

p = 0(s,

t),

for the variable vector of a curve, or of a surface (comp. again 99), s

and

t) denoting any functions of vector

being two variable scalars, and 0() and
latter
is
the
to
whereof
here
be entirely independent* of the
form,
supposed
former, we may then employ (comp. 100, (4.) and (9.) and the more recent
(f>(s,

and 329,

sub-articles, 327, (5.), (6.),

partial

forms

and so

and

I.

may

(5.)

the notations of derivatives, total or

write, as the differentiated equations, resulting

II. respectively, the following

III.

IY.

dp =

dt = p dt = Dtp

dp = d s p + d p =

tft

from the

Dp
s

d^

ds +

Dp
t

fit

which the geometrical significations have been already partially seen, in

the sub-articles to 100, and will soon be more fully developed.
(1.) Thus, for the circular locus, 314, (1.), for which
of

Y.

we

And

= o

/3,

1,

Saj3

p = V.

Ta =

a*/3,

0,

derived vector,

for the elliptic locus, 314, (2.), for

VII.

we

by 333, VIII., the following

have,

(2.)

1,

which

but not Sa(3 =

0,

have, in like manner, this other derived vector,

(3.) As an example of a vector-function of more

resume the expression (308, XVIII.),

IX.
*

We

to which

>

= rk tjs kj- sk- t

are therefore not

employing here the temporary notation of

we

should have had, d0g

<)>(q,

dq).

scalars than one, let us

some recent

Articles, according

VECTOR FUNCTIONS, DERIVED VECTORS.

ART. 337.]

463

which we shall now suppose that the tensor r

is given, so that
p is the
of
the
is r,
which
radius
a
variable vector of a point upon
given spheric surface,
and the centre is at the origin while s and t are tico independent scalar variables,

in

with respect to which the two partial derivatives of the vector p are to be
determined.

The

(4.)

(295), and

X.
if

Ta =

derivation relatively to

we have

since

d a* =
.

Z
1,

and

is

easy

by 333, VIII.,

generally,

XI.

and therefore

a^dtf,

x be any scalar function of

if

for, since ijk are vector-units

t,

D<

a x = ^a* +1 D^,
&

we may

write, at once,

by

279, IV.,

XII.

and we

Dtp

=
&

(fy

-M)

TVfy

see that

XIII.

8pT>

0,

a result which was to be expected, on account of the equation,

XIV.
which

.^+r-0

by 308, XXIV., from the recent expression IX. for p.

expression of about the same degree of simplicity,

follows,

To form an

(5.)

other partial derivative of o,

(

vector part

(its

XV.

we may observe

scalar vanishing)
.

Dp
s

irVjk-tp

by the transformation 308,

that j** l kj~*

for the

equal to its

is

own

hence*

XVI.

or

(11.).

And

D,p

irtf

jp

because the scalar of

irjk^p,
tfjk"*

is

zero,

we

have thus the equation,

XVII.
which

is

SjoD.p

0,

analogous to XIII., and might have been otherwise obtained, by

XIV. with respect to the variable scalar s.

taking the derivative of

(6.)

The

partial derivative

p must be perpendicular
under the last form, is

D,p must be a

is

[Thus D, (/#-)

equivalent to

XV.]

=^0

* +1

hence,

by XV.

or/Af

or

XVI.,

a result which,

by the expression 315, XII. for p.

(3.) and (4.), and by the recent values

easily confirmed

In fact that expression gives,

by 315,
*

vector

to the vector tfjk-*, or k ztj,

^T +y *y-t-1 ) = y+

*;-

and therefore D,p =

TT

rk*jk-Wj*kj-*Jrt

which

ELEMENTS OF QUATERNIONS.

464

other forms for the

XII. XVI., these

been above considered

XVIII.

two partial derivatives

4, 5.

which have

of p,

- TrrrV./ 25

IV

[III. n.

XIX.

Dp

= irr(ArV.

zs + l

- V.

2s
7c

which might have been immediately obtained, by partial derivations, from

the expression 315, XII. itself, and of which both are vector-forms.

And

(7.)

315, XIII.,

hence, or immediately by derivating the expanded expression

we

new forms

obtain these

XX.
XXI.

D/jo

D p = Trr
s

for the partial derivatives of p

= trr(j cos
{

(*

cos

tir

tir

+/ sin

sin

fa) cos STT

sir

- k sin

STT

add that not only is the variable vector p perpendicular

the two derived vectors, D s p and D /j0 but also they are perpendicular

each of

to

to

each other

we may

for

XXII.

write,

S (Dsp

and the same conclusion

A vector may

(9.)

variables,

such as

r, s, t

by XII. and XVI.,

3V) = 7r S

XXIII.

may

k*jp*k

dp

=
;

XX.

and XXI.

be considered as a function of three independent scalar

or rather it must be so considered, if it is to admit of
;

= d r/ o + d sp +

dtp

T>

and then

dr +

IV

it

will

have a

derivatives.

XXIV.
may

also be thus

XXV.

Dp
r

more

total

&s

For example, when p has the expression IX., we have

partial derivative,

and we

be drawn from the expressions

have three* partial

will thus
(10.)

= TrVS

being the vector of an arbitrary point of space

differential (329) of the trinomial form,

which

sin

We may

(8.)

and

^TT)

Dp
r

r~

this third

Up,

fully written (comp. again 315, XIII.),

= h js 7(j-*k- t =
t

(i

cos

tir

+j sin

frr)

sin

sir

+ k cos

STT

see that the three derived rectors,

XXVI.

compose here a rectangular system.

* That

is to

have occasion to consider successive

say, three of the first order; for we shall soon
of
one or more variables, and so shall be conducted to the consideration

differentials, of functions of

orders of differentials

and derivatives, higher than the first.

SUCCESSIVE DIFFERENTIALS, DEVELOPMENTS.

ARTS. 337,338.]

SECTION

On

465

5.

Successive Differentials, and Developments, of Functions

of Quaternions.

338. There will

now be no

will be

found to be

difficulty in the successive differentiation, total

one or more quaternions

or partial, of functions of

as in the ordinary calculus, in connexion with

useful,

besides

developments of functions:

and such differentiation

that

it

is

necessary for

many

those

of

geometrical ami physical applications of differentials of quaternions, on which

few examples of successive differentiation may
we have not entered yet.

serve to show,

more

of the operation

than any general precepts, the nature and effects

and we shall begin, for simplicity, with explicit functions of

easily

one quaternion variable.

(1.)

Take then the

square,

We

be twice differentiated.

gave the equation,

I.

but we are

now

d fq of the function
of

and

dfq = d

tensor,

q*

= q dq + dq
q
.

(2.)

The

to

form the second

Td

2
<?,

dd</

= d 2 ?, where d 2
#

is

differential

still equal
a new arbitrary quaternion,

And

need not be small (comp. 322).

in general, this twice differentiated expression, or

differential

II.

is

treating the differential of the variable q as

(f,

in general writing

which the

which

saw, in 324, VII., that a first differentiation

to differentiate again, in order to

to d#,

of a quaternion, as a function fq,

(f,

tffq

= d2

= q dty + 2d? 2 +

&q

<f

thus

we

get,

of the second order,

.

q.

second differential of the reciprocal of a

quaternion

is

generally

(comp. 324, XL),

III.

d2

1
.

If p be a variable vector, then

(comp. 336, (1.))
second differentials of its square, the expressions

(3.)

and

we have,

for the first

IV.
(4.)

If fp be

d p 2 = 2S P dp
.

any

HAMILTON

d/jj

Y.

d 2 P z = 2S od 2/0 + 2d /0 2
.

other scalar function of a variable vector

p,

again the sub-articles to 336)

VI.

its first differential

2Svdp, when v

ELEMENTS OF QUATERNIONS.

is

and if (comp.
be put under the form,

another variable vector,

3

ELEMENTS OF QUATEKNIONS.

466

then the second differential of the same function

VII.
in

which we have written,

(5.)

[III. n.

may be

expressed as follows

briefly,

Sdvdp, instead of S (dv dp).

will be found useful, in the
.

The following very simple equation

theory of motions, performed under the influence of central forces

VIII.
(6.)

5.

dVpdp = V/odV

As an example

because

V. dp 2 =

0.

of the second differential of a quaternion, considered

as & function of a scalar variable (comp. 333, VIII.,

and 337,

ing may be assigned, in which a denotes a given unit

da = 0, but x is a variable scalar

(1.)

},

the follow

line, so that a

= -

1,

IX.

(7.)

nion q

d 2 a* = a
.

The second

differential of the product of any two functions of a quater

be expressed as follows (comp. II.)

may

X.

d 2 (fq

.
<<?)

tffq

.
<}>q

+ 2dfq

d^q +fq

d%.

The second

differential, d*q, of the variable quaternion q, enters

z
has
been seen) into the expression of the second differential d fq,
generally (as
of the function fq, as a new and arbitrary quaternion
but, for that very

339.

reason,

it is

permitted,

and

it

that this second differential d*q

is
is

frequently found to be convenient, to assume

or, what comes to the same
equal to zero
:

thing, that the first differential d^

is

And when we make

constant.

this

new

supposition,
I.

dq = constant,

or

d2 ^ =

0,

the expressions for d*fq become of course more -simple, as in the following

examples.
(1.)

With

this last supposition, I. or I

differentials, of the square

II.

d2

<

= 2d*

and the
III.

(2.) Again, if we suppose that

and take the function,
IV.
.

.,

we have the following

reciprocal of a quaternion

d2

~l
.

CQ , c l9 c 2 are

second

any three

fq = Wdqct,

constant quaternions,

SUCCESSIVE DIFFERENTIALS, DEVELOPMENTS.

ARTS. 338-340.]

we

under the same condition

find,

or I

I.

.,

that

its first

and second

467

differen

tials are,

Y.

dfq =

dq

c,qc 2

in writing which, the points*

The

(3.)

322,

c8

2
VI. .. d /? = 2c Q dq c,dq
.

cz

be omitted.

may

remaining still entirely arbitrary (comp.

its tensor Tdq is
no
that
supposition is made that
this differential dq to be constant (I.) we have

first differential, dq,

and 325,

(8.),

small,

c^qc.dq

(2.) ), so

although we now suppose

rigorously.

VII...
an equation which

may

(g

d^^ + d.if+id

And

in like

.?

be also written thus,

VIII. ..(q+ dqY =

(4.)

manner we

(1

+ d + Jd2 )

shall have,

q\

more generally, under the same

condition of constancy of dq, the equation,

IX. ..f(q + dq) =

if

(1

+ d + Jd )/?,

the function fq be the sum of any number of monomes, each separately of the
IV., and therefore each rational, integral, and homogeneous of the second

form

dimension, with respect to the variable quaternion, q

combined with others of the first

if

#1,

o,

i,

and

c lt

or of such monoines,

dimension, and with constant terms: that

<?

2,

o,

c/ 2,

is,

be rm# constant quaternions,

and

X.

easy to carry on the operation of differentiating, to the third

and higher orders remembering only that if, in any former stage, we have
2
we then continue so to
denoted the first differentials of q, dq,
by dq, d ^,
340. It

is

denote them, in every subsequent stage of the successive differentiation and

that if we find it convenient to treat any one differential as constant, we must
:

then treat

all its successive differentials as vanishing.

A.

few examples

may

be

given, chiefly with a view to the extension of the recent formula 339, IX., for
the function f(q + dq) of a sum, of any two quaternions, q and d^, to polynomial
forms, of dimensions higher than the second.

Compare the second Note

to

page 439.

302

ELEMENTS OE QUATEBNIONS.

468

The

(1.)

third differential of a square

I.

More

(2.)

d3

= q

q*

d 3 (/?

W)

<?

be thus expressed

may

= d 3/^

More generally

(3.)

d*q

dq).

generally, the third differential of & product of two quaternion

functions (comp. 338, X.)

II.

generally (comp. 338, II.),

+ d z q q + 3 (dq d 2 +

d*q

is

[IIT.ii.5.

2
+ 3d /?

.
<j>q

+ Sdfq

.
d<j>q

dz

+fq d^q.
.

(j>q

the n th differential of a product

still,

is,

as in the

ordinary calculus,
III.

d n (fq, $q) = d nfq.

..

~l

+ nd n fq. d$q + nA^fq. d z

<j>q

= n(n-l);
{

lf

n,

--!)(2%-

n(n
= -i

2)

&c.

-L,

(j>q

n
+fq. d

(j>q,

the only thing peculiar to quaternions

being, that we are obliged to retain
(generally) the order of the fact or in each term of this expansion III.
s>

(4.)

Hence, in particular, denoting

changing $q

IV.
(5.)

briefly

the function fq by

Hence

d n rq = d n r
.

under

also,

constant quaternion,

n l
q + nd r

this condition that

dq

d2q =

if

dq,

is

constant,

0.

if c

be any other

we have the transformation,

dr=0.

if

(6.)
t

by

Hence, by 339,

(4.), it is

the equation (comp. 316,

VI.

d
.

easy to infer that

-L d

= 1 + d + |d 2 +

we interpret this other symbol

formed from/?, by operating on

is, if

we

if

interpret the symbol

I.),

4/

that

and

r,

to q,

&

.,

fq as concisely denoting the

t

series

which

is

and

the function fq, thus operated on, be any finite poly nome, involving (like

if

it

with this symbolic development

the expression 339, X.) no fractional nor negative exponents

we
an
as
extension of a recent equation (339, IX.) the formula
write,
;

VII.

/? =f(q + dq),

if

d2? =

may

then

FUNCTIONS WHICH VANISH TOGETHER.

ARTS. 340, 341.]

469

here a perfectly rigorous one, all the terms of this expansion for a
function of a sum of two quaternions, q and dq, becoming separately equal to
zero, as soon as the symbolic exponent of d becomes greater than the dimension

which

of the

is

polynome.

(7.)

We

shall soon [342] see that there is a sense, in

transformation VII.
composed as above

which

this exponential

may be extended, to other functional forms which are not

and that thus an analogue of Taylor s Theorem can be

established for Quaternions. Meanwhile it may be observed that by changing

dq to A#, in the finite expansion obtained as above, we may write the formula
as follows

VIII.
which

*% =f(q + A?) = (1 + A)/?,

or briefly,

IX.

d= 1 +

may be operated on, or transformed, as in the

usual calculus of differences and differentials.
For instance, it being under
stood that we treat A ty as well as d 2 # as
vanishing, we have thus (for any
last symbolical equation

positive

and whole exponent m), the two following transformations

X.

Am =

f*

and

1),

XI.

dm =

(log (1

+ A) ) m

of IX.,

the results of operating, with the symbols thus

equated, on any polynomial
function fq, of the kind above described,
finite expansions, which

being always

are rigorously equal to each other.

341. Let Fx and x be any two functions of a scalar variable, of

vanish with that variable-, so that
they satisfy the two conditions,
I.

Then

jFO - 0,

00 =

which both

0.

the three simultaneous values,

II.

of the variable

x,

Fx,

and the two functions, are

three simultaneous
differences, as

fyx,

same time (comp. 320, 32.1)

with
this other system of three
compared
at the

simultaneous values,
III.
If,

then,

any

FO,

<pO.

equimultiples,

IV.

of the three values

II., can be
n,

0,

combined with a

nx,

nFx,

n$x,

made, by any suitable

increase of the number,

decrease of the variable, x, to tend

together to

any system of

ELEMENTS OF QUATERNIONS.

470
those limits

limits,

must (by the

[III. n.

5.

compare again 321) admit

definition in 320,

of being considered as a system of simultaneous differentials,

V.

cb,

answering to the system of initial values III. and must be proportional to the
ultimate values of the connected system of derivatives,
;

YI.

We may

F x, 0X

1,

when x

tends to zero.

therefore write, as expressions for those ultimate values of the two

last derived functions,

VII.

And

JF"0

= lim nF-,

000/1 ?/ tf/&00 last

values vanish, or

if

^0

VIII.

F x,

are satisfied, so that x,

simultaneous differences,

of quotients, which

IX.

is

lim (Fx

or

$x)

may

tend together

r -,

nd>

.PO = r

if

d>0

tfwo

new conditions

0,

0.

0,

and fix are now (comp.

we may

II.) a ??e# system of

s^7/ establish the following equation of limits

conditions VIII.,

last

= lim (F x

:
(f>

if

x),

.FO =

being understood that, in certain

with x

cases,

to infinity,

00

these two quotients

when x

may

both vanish

tends, as before, to zero.

This theorem

is so
important, that it will not be useless to confirm it
a geometrical illustration, which may at the same time serve for a geometri

(1.)

by

the

independent of these

Z=

it

= lim
d/0
r

cal proof

at least for the extensive case

are of scalar forms, and consequently

the corresponding ordinates,
and

XY

mon

abscissa

OX),

be represented, or constructed, by
(or ordinates answering to one com

may

XZ

0#Y and

of two curves

where both the functions fa and $x

OsZ, which are

in one plane,

and

set

out from (or pass through) one common origin 0, as in the annexed figure 75.
shall afterwards see that the result, so obtained, can be extended to quater

We

nion functions.
(2.)

Suppose then,

first,

that the ordinates of these two curves are pro-

portional, or that they bear to each other one fixed

the equation,

X.
is satisfied

XY XZ = xy
:

for every pair of abscissce,

corresponding ordinates

may

be.

OX and

and

constant ratio; so that

:
<rs,

Ox, however great or small the

Prolonging then

(if

necessary) the chord

GEOMETRICAL ILLUSTRATIONS.

ART. 341.]

Yy

of the first curve, to

determine a subsecant
corresponding chord Ziz
point,

meet the

we

some point

and

t,

so to

see at once

(by similar triangles) that the

of the second curve will meet the same axis in the same
tfX,

and therefore that

axis of abscissae in

471

it

will determine (rigorously] the

same subsecant,

tX..

the point x be conceived to approach to X, so that the

secant ~Yyt of the first curve tends to coincide with the tangent
to that
curve at the point Y, the secant Ztzt of the second curve must tend to coincide

Hence,

(3.)

if

YT

with the line ZT, which line therefore must be the tangent to that second
curve

or in other words, corresponding subtangents

coincide,

and

of course are equal,

under the supposed

condition X., of a constant proportionality of ordinates.

(4.)

Suppose next that corresponding ordinates

only tend to bear a given or constant ratio to each

other

or that their (now) variable ratio tends to a

given or fixed limit,

the common abscissa is

when the point X tends

when

indefinitely diminished, or
to

and

let

be

the variable point in which

still

the tangent to the first curve at

Y meets the axis,

so that the line

TX is still

Then

the first subtangent.

the corresponding tangent to the second curve at

not
will
in
Z
general pass through the point T, but will meet the axis in some
But the ratio of the two corresponding subtangents,
and
different point U.

TX

TJX, which had been a ratio of equality, when the condition of proportionality
X. was satisfied rigorously, will now at least tend to such a ratio so that we
;

shall have,

under

this

new

the limiting equation,

XL

whence the equation IX.

XII.
(5.)

We

might

condition, of tendency to proportionality of ordinates,

.lim(TX:UX) =

results,

lim (tan

also

under the geometrical form,

XTY

tan

have observed

corresponding areas* (such as

XUZ)

that,

= lim(XY
XZ).

when

#XY?/ and ^XZs)

exactly in the given ratio of the ordinates

OXY^/O OXZsO =
:

the proportion
of the

so that

proportion,

XIII.

l;

XY

rigorous,

XZ,

Compare the Fourth Lemma of the First Book of the Principia

which the reasoning of the present sub-article is virtually

Corollary, in

is

other equation, or

this

X.

two curves are then

and see especially

anticipated.

its

ELEMENTS OF QUATERNION S.

472

then

Hence

[m.

n.

5.

we only

suppose, as in (4.), that the ordinates

tend to some fixed limiting ratio, the areas must tend to the same
so that if

is

also rigorous.

if

the second member of the equation IX. have any definite value, as a
whereas the recent proof,
first memher must have the same
:

by

the

limit,

siibtangents,

served rather to show that if the first (or left hand) limit in IX. existed, then
the second limit in that equation existed also, and was equal to the first.
(6.)

Fx be

If the Junction

a quaternion,

we may (by 221)

express

it

as

follows,

XIV.
where W, X, T,

Fx =

W+ iX + jY + kZ,

Z are four scalar functions

which each separately can

and the recent geometrical*

of x, of

be constructed, as the ordinate of a plane curve

reasoning will thus apply to each of them, and therefore to their linear combi
;

nation

Fx: which quaternion function reduces

when

W= 0.
And

(7.)

if -tyx

itself to

a vector function of

x,

were another quaternion or vector function, we might first

and then eliminate the scalar function 0#; so that a

substitute it for Fx,

limiting equation of the form IX. may thus be proved to hold good, when
both the functions compared are vectors, or quaternions, supposed still to vanish

with

x.

The general

(8.)

considerations, however, on which the equation IX.,

lately established, appear

to be

more simple and

direct

and

is

it

was

evident

that they give, in like manner, this other but analogous equation, in which
are second derivatives, and the conditions VIII. are now supposed
F"x and
"x

to be satisfied

XV.

lim(.F^ far) = lim

:

(F"x

f V),

if

FO

0,

0.

* Instead of the
equation IX.,
to give one of the following

it has become usual, in modem works on the Differential Calculus,

form (deduced from principles of Lagrange)
:

if

some proper fraction, or quantity between

also & geometrical proof, when t hef^mctwns Fx and

6 denoting
is

and

1.

And

a geometrical illustration, which

(or conceived to be con

can be constructed
<f>x

is sometimes derived from the axiom (or geometrical

and plane arc must be parallel to the tangent, drawn at some
parallelism no longer exists, in general, when the curve is one of

structed) as the ordinates of two plane curves,

intuition], that the chord of

any

finite

But this
point of thai finite arc.
double curvature ; and accordingly the equation in this note is not generally true,
are quaternions
or even when one of them is a
quaternion, or a vector.
;

when

the functions

TAYLOR S SEEIES EXTENDED TO QUATERNIONS.

AKTB. 341, 342.]

And

so

we might

473

proceed, as long as successive derivatives, of higher orders,

continue to vanish together.

(9.)

Hence, in particular,

we take

if

this scalar form,

which evidently gives the values,

XVII.
and

if

00 =

4/0

0,

0"0

we suppose that the function Fx

XVIII.

.FWO has any

while

0,

FO =

in which the function Fx,

;

^ ^O =

value

0,

<0

1,

such that
F"0

0,

lim. (Fx

and the

0,

$x)

F^Q,

F^O

F( m ~^0 =

0,

this limiting equation

are here supposed to be generally

happen, in particular cases, to reduce

may

although they

we may then establish

finite value,

XIX.

quaternions

FO

0,

is

0,

them

selves (292) to vectors, or to scalars.

342. It will

VII.

and

to

now

be easy to extend the Exponential Transformation 340,

show that there

I.

which

is,

a sense in which that very important formula,

is

A =/(? +

known* mode

in fact, a

if

d?),

d3

0,

of expressing the Series or

Theorem of

and not merely for

those functions of finite &\\& polynomial form, vrith positive and whole exponents,
for which it was lately deduced, in 340, (6.).
For let fq and f(q + dq)
Taylor, holds

denote any two

quaternion

good

for Quaternion Functions generally,

states,

and

or values, of which neither

of the

II. .. dfq,

is infinite,

of any function of a

first differentials,

(\~fq,

& mfq,

dq =

in which

const.,

be supposed that no one is infinite, and that the last of them is different
them
zero; while all that precede it, and the functions/*? and/(<7 +
or
first
not
the
m
selves, may
terms, of the
may
happen to vanish. Let
let it

from

d<?)

Lacroix, for instance, in page 168 of the First Volume of bis larger Treatise on the Differential
(Paris, 1810), presents the Theorem of Taylor under the form,

and Integral Calculus

-\

d%

d-w

(\u

-i

1.2

where u denotes the value which the function u

increment d# (1 accroissement quelconque d.r).
HAMILTON

ELEMENTS OF QUATERNIONS.

d%
1

1.2.3

1.2.3.4

receives,

when

&c.

the variable x receives the arbitrary

ELEMENTS OP QUATERNIONS.

474

[III. n.

5.

l)fq, be denoted briefly by q^ q z

be
called
the remainder of the series, or the
may
correction which must be conceived to be added to the sum of these m terms,

exponential development of the symbol

qm

and

let rm

denote what

in order to produce the exact

III.

manner

in such a

(t

mine

of the difference,

A/? =f(q + A?) -fq

we

that

=/(<?

+ dq) -fq

have rigorously, by the notations employed,

shall

the equation,

IY.

f(q + dq) =fq + q +

qz

+ q m + rm

where

am

=
t

xi

w being different from

by what has been above supposed.
//&

erw

Y.

lim. (Trm

zero,

we please,

to the tensor

WZ

infinite 9

shall prove, as a Theorem, that

lim.

if

0,

or in words, that the tensor of the remainder

ratio as

but no one of the terms being

Then we

Tq m =

may

Tdq =

be

made

to

bear as small a

of the last term retained, by diminishing the tensor,

And this very

extend to functions of several

without changing the versor, of the differential (or difference) dq.

general result,
quaternions,

which we

is

which

Taylor s theorem to
be called, for the sake of

in the present Calculus that analogue of

lately alluded
"

Taylor

reference,

(in

340,

(7.)

Theorem adapted

and

it

may

to Quaternions"

Writing

(1.)

YI.

we

will soon be seen to

shall

Fx =f(q + ads) -fq - x&fq -

have the following successive derivatives with respect to

F x = df(q + xdq) -dfq- xd fq z

.- 77-,,- ^-r

HI

x,

VII.
(

m-Vx =

Wx

d.

=
d"/

-]f(q + ccdq)
(q

+ xdq)

- dm
~\fq

and

finally,

because, by 327, VI., and 324, IY.,

VIII.

Df(q + xdq) = lim.H{/(0 + xdq +

n=&

and in

like

n~ dq)

-f(q +

xdq)}

manner,

IX.

= d 2f(j + xdq], &o.

V*f(q + ady)

= df(q + xdq),

EXAMPLES OF QTJATEKMON DEVELOPMENT.

ART. 342.]

mark

the

on

of derivation

q alone,
(2.)

X.

referring to the scalar variable

a?,

while d operates

nor on dq.

a?,

have therefore, by VI. and VII., the values,

"We

and not here on

475

FO =

FQ

0,

whence, by 341, XIX.,

we have

XL

F"0

0,

0,

-l

FWQ

0,

= d mfq

this limiting equation,

..lim. *&:
x=Q

F( m

/O.O

-d%;

or

XII.
(3.)

and

qm ,

But

Km. (Fx fy) =

:

these two functions,

to

by changing dq

decreasing scalar, x,

We

versor

Udg.
+ ?*) qm
*

(qm

xdq

we may

Fx and

XIII.
and

it is

we have

XV.
and

th

differential is (for

dm

fq

= dm

l
.

q~

= 2

easy to prove, without

when

this
:

is

its

done, the quotient

q m to zero, as its limit

V. holds good.

dq =
.

by a

Tdq, without changing

example, let the function fq be the reciprocal, q

its

by IV. from q m + r m

of thus multiplying dq

tensor

its

therefore say that,

may

are formed

tends to unity, or this other quotient r m

As an

339, III.)

-^x,

and instead

diminish

or in other words, the limiting equation

(4.)

if

1,

1
;

then (comp.

const.),

]
.

differentials,

q~ (- r)

r = dq

if

1
.

q-

that

therefore here
..q,n

C (-

m
r)

rm =

- mr
q

(1

+ r)-\

this last tensor indefinitely diminishes

T (rm

qm)

= Tr

T (1

r)~

with Tdg, the quaternion q being

supposed to have some given value different from zero.

(5.) In general, if we establish the following equation,
i,l-m

,,1-2

XVI..

V. and if we suppose that

~
of
its
differentials as far as & m fq
one
nor
itself,
any
the result contained in the limiting equation XI. may then be

as a definitional extension of the equation 325,

neither the function fq

is infinite

expressed by the formula,

3

P2

ELEMENTS OF QUATERNIONS.

476

[III. n.

5.

which for the particular value m = 1, if we suppress the upper index, coincides
with the form 325, VIII. of the definition dfx, but for higher values of m
m
namely (when d /g is supposed neither to vanish, nor to
what we have called Taylor s Theorem adapted to Quaternions.

contains a theorem

become

infinite),

343. That very important theorem may be applied to cases, in which a

quaternion (as in 327, (5.)), or a vector (as in 337), is expressed as a function of
a scalar; also to transcendental forms (333), whenever the differentiations can

be effected; and to those new forms (334), which result from the peculiar
few such applications may here
operations of the present Calculus itself.

be given.
first this

Taking

(1.)

transcendental and quaternion function of a variable

scalar,

I...q = ft =

we

Ta =

with

a*,

da =

1,

d* = const.,

0,

have, by 333, VIII., the general term,

1

=
dividing then
correct,

d
.

a*

by

and convergent

a*,

we

obtain an

a d* = l

+ #a+

A(tfa)
-7T-

(2.)

(xa)-+
OQ
m
o
</

..

are obtained

when we

is

found

to

operate with

irdt
vdt
= cos+asm-7r
7r
4i

<>

V (q + dg), K (q + dq),

on Sq, Vy, Kg, and

N<?

and
;

example (dq being still constant), the third and higher differentials of
vanish by 334, XI., and we have
d

IV.

Ng =

+ d + |d 2 )

(1

be

a:a

Correct KoA. finite expansions, for S (g + d^),

N (q + dq),

which

(4.) ),

4>

o
2

lf

infinite series,

namely (oomp. 308,

2

III..

a* (ma)

Ng - N? + 2S (Kg

dg)

Ndg

N (g + dg)

for

Ng

an expression for the norm of a sum, which agrees with 210, XX., and with
200, VII.
let us
develop, on like principles, the tensor and versor of a sum,
this
of
quotient
again write r for dg g, and denote the scalar and vector parts
(3.)

To

by

and

(4.)

?;

so that,

by 334, XIII. and XV.,

Then writing

also,

for

abridgment, as in a known

factorials,

Y1L.. [-!]-(-!).

(-2). (-8). ...(-),

notation

of

EXAMPLES OF QUATERNION DEVELOPMENT.

AKTS. 342, 343.]

we

shall have,

by 342, XIII., dq being

still

treated as constant, the equation,

m
VIII.
of

which

it is

IX.
(5.)

We

d m (* +

= d m r - [-

v]

ds

==

have

- S

=-

(s

v)

(s

d lW*

XIII.

we

to = -

2
) ;

(*

BI+1
,

t?)

for example,

=-

t?)

2sv.

(s

+ d)

(0

+ d)

+ d) z

(t?+ d)

exponential formula 342,

XV.

Ttf
V

have, for instance, by IX.,

XIV, ..T(q+

or,

(s

Ta

the notation being such that

XII.

d"*

To

The

[- 1]

by V. and VI.,

also,

ml

(6.)

m
=

m+l
1] r

easy to separate the scalar and vector parts

.

477

I.,

d?)

+ d)

(s

(t?

d)

= s 2 + ds = 8

v=

t?

df>

gives, therefore,

=
~U(q + dq)

dividing and substituting,

XVI.

(1

s+d

1>)

XVII.

U (1 +

=
1?)

c+d l

being here a scalar and a vector, which are entirely independent of each
but
of which, in the applications, the tensors must not be taken too
other;
s

and

large, in order that the series

(7.)

The

may

symbolical expressions,

converge.

XVI. and XVII.,

for those

two

scries,

may

be developed by (4.) and (5.) thus, if we only write down the terms which do
not exceed the second dimension, with respect to s and v, we have by XII.
and XIII. the development,
;

XVIII.

T (1

v)

= 1 +

v,

W+

XIX.. .U
of

which accordingly the product

is 1

to the

same order

of approxi

mation.
(8.)

without differentials,

sum

two quaternions can sometimes be developed,

by processes of a more algebraical character and when

A. function of a

of

ELEMENTS OF QUATERNIONS.

478

happens, we

tliis

successive differentials of the function

m
342, XIII. for d
q

(9.)

And

and

from the

series

may we

verify the recent

not only

XX...T(l
XXI. ..TJ

processes,

the

if

also,

of these, for

first

which are in

we can

of the

infertile expression

XVIII. and

developments,

algebraical forms*

(1

for the reciprocal of a sum.

XIV-,

342,

0)

so deduce the values

for example,

XIX., by comparing them with the more

but

5.

compare the result with the form given by Taylor

may

as adapted to quaternions in 342,

Series,

[III. n.

+ ?)4(l +

s-t>)4,

example (when expanded by ordinary

have given us, without differen

this case applicable),

tials,

XXII.

we can then

(q

=
)

(1

\tf

Tq, where

and second

infer the values of the first

a quaternion, as follows

XXIII.

and

differentials of

V^V

1
,

the tensor of

dTq = S

T?

d z Tq = -

(V^
\
q

under the form,

it,

XXIV.
(10.)

agrees with 334, XII. or XIII,, and the second can be

first

deduced from

whereof the

S/^

d Tj = d

In general,

if

(g

( (S

T?)

we can only develop

term or terms which are of the

^J-

(^J)

a function f(q + q

first dimension relatively to

</

Tq.

as far as the

we

shall still

obtain thus an expression for the first differential dfq by merely writing dq in
the place of q
But we have not chosen (comp. 100, (14.) ) to regard this
9

property of the differential of a function as the fundamental one, or to adopt it as

the definition of dfq\ because we have not chosen to postulate the general

such developments of functions of quaternion sums, of which in fact

in many cases difficult to discover the laics, or even to prove the existence,

possibility of
it is

except in some such

(11.)

tions)

we

way

as that above explained.

This opportunity

may

be taken to observe, that (with recent nota

by VIII., the symbolical

** 1 = 1
XXV.
+s+v
or
have,

expression,

XXVI.

[These are equivalent

t [In fact by VIII., dr

d
.

<T

1+

to the transformations

=-

r*

and

(r

d)

(r

d)

r2

rz

0.]

r.t

SUCCESSIVE DIFFEBENCES AND DERIVATIONS.

ARTS. 343-345.]

479

344. Successive differentials are also connected with successive differences, by

laws which it is easy to investigate, and on which only a few words need here

be said.

We

(1.)

can easily prove, from the definition 324, IV. of dfq, that

if

dq

be constant,
I.

d 2/? = lim.

n* {f(q

M = oo

+ 2n~ dq} - 2f(q + n~ dq) +fq]

l

with analogous expressions for differentials of higher orders.

(2.) Hence we may say (comp. 340, X.) that the successive
2

II.

d /?,

dfq,

ay?,

=
frq

for

differentials,

0,

are limits to which the following multiples of successive differences,

III.
all

n~~fq,

nkfq,

when

simultaneously tend,

/<?,

the multiple

become equal

dq, or at least tends to

w3 A3

nAq

for

0,

either constantly equal to

is

number n

thereto, while the

increases

indefinitely.

And

(3.)

hence we might prove, in a

f(q + dq) can

new way,

that if the function

proceeding according to ascending and

whole dimensions with respect to dq, the parts of this series, which are of those
be developed, in a series

successive dimensions,
to

adapted

must follow the law expressed by Taylor

Theorem*

Quaternions (342).

345. It

easy to conceive that the foregoing results may be extended

(comp. 338), to the successive differentiations of functions of several quater
nions
and that thus there arises, in each such case, a system of successive
is

and partial:

differentials, total

as also a

system of partial derivatives, of orders

higher than the first, when a quaternion, or a vector, is regarded (comp. 337)
as a function of several scalars.

The

(1.)

general expression for the second total differential,

involves d 2 ^, dV,

but

it is

often convenient to suppose that all these second

differentials vanish, or that the first differentials dq, dr,

then

a*Q, or

m
d.

function of the

F(q

r,

.),

becomes a

th

(d q

Some remarks on

Lectures, pages 589, &c.

.)

are constant

and

and homogeneous
differentials dq, dr,
which

rational, integral,

dimension, of those
may (comp. 329, III.) be thus denoted,
m
II.
dm Q =
+d +
or
.

first

briefly,

III.

dm =

(d,

+ dr +

.)

",

the adaptation and proof of this

important theorem will be found in the

ELEMENTS OF QUATERNIONS.

480

[III. n.

in developing which symbolical power , the multinomial theorem of algebra

p

be employed

5.

may

because we have generally, for quaternions as in the ordinary

calculus,

IY.
(2.)

V.

For example,
.

= rqr,

r),

d q Q = rq r

and suppose

dr Q =

r>r

two equated symbols IV. gives, by

r + n

^dr)

F(q, r + n~ dr} F(q + n

l

which the numbers n and n are supposed

We

its

to

tend to

operation

I.),

dq, r)

+ F(q,

Q + &Q = F(q +

XL

or briefly,

+ dr,

dq, r

.)

* f +r+-

Finally, as an example of successive

XVIII.

the vector expression 308,

XII.
which has been seen

we may

} ;

infinity.

^ + ^"F(q

r)

with interpretations and transformations analogous to those which

occurred already, for functions of a single quaternion.
(4.)

r)

also write, for functions of several quaternions,

may

X..

r<?/;

/0V + r? /.

the limit of this other function, or product (comp. 344,

nn [F(q + n^dq,

(3.)

and

VII.

in

ff

d r d q Q = d,d r Q =

in general, each of the

on F(q,

IX.

VI.

V11I.

And

dr d q = d d r

we denote dq and dr by

if

then

and

and partial

derivation, if

have

we resume

(comp. 315, XII. and XIII.), namely,

= rtffkj-

k-

to be capable of representing the vector of

any point of

space,
by the principle
mentioned in 308, (11.), and by the sub-articles to 315, not only the form

observe that

XIII.
but

also, if

gives, without trigonometry,

it

p = rVj^k

as in 308,

-*,

XIX.,

a be any vector unit,

XIV.

rtt^ k*

r V.k

= rk*(k8

a 25

*S

*-

k~

f
;

whence

XV.
(5.)

articles

+ rk zf V.

zs
,

as in 315,

XII.

have therefore the following new expressions (compare the subto 337), for the two partial derivatives of ft& first order, of this variable
"We"

vector p, taken with respect to

XVI.

and

Dp

fa

irrtfj

y-

k =-

irptfjk-*,

with the verification, that

XVII.

P D sp

vrr

kj-

k*

tffif*k* = TrrWAr*

and

PARTIAL DIFFERENTIALS AND DERIVATIVES.

AET. 345.]

XVIII.

Dtp = irrV

= nr& jS

V.f

a^

r^pD p S

481

whence

XIX

T>

= -

;-D,p

r-

24 - 1

XX.

and

D.p

= TrVpS

D^

25

1
;

while

XXL
so that
first

we have

Dp
r

XXV.

= #/#/#-, as in 337,

the following ternary product of these derived vectors of the

order,

XXII.

Dp Dp

= TryS a 25

= 7rrD,S a 2 *
.

the scalar character of which product depends (comp. 299, (9.)) on the cir
cumstance, that the vectors thus multiplied compose (337, (10.)) a rectangular
system.
(6.)

It

is

easy then to infer, for the six partial derivatives of

second order, taken with respect to the

the expressions

same three

of the

p,

scalar variables,

r, s,

t,

XXIII.

XXIV.
(7.)

.T) r

D D p = r^D p
V&ip = D T) p . r/r^V. /

V*p = - TrV

The

DDp

T>

7T

T>

vector p, are the following

XXV.

d,./o

DDp=
r

DV = -

s+1
;

three partial differentials of the first order

tp

ir

rk tfV

rs

same variable

of the
>

r~

d s/o =

pdr

Dp
s

ds

d tp

D^

d^

with the products,

XXVI.
XXVII.
(8.)

These

d,p

d rp

d r p, like

radius of this sphere

D p,
r

= -

dtp

d sp

25
d*
TrrpdS a

dtp

d r p, d s]0

7rr

dr dS a zs
.

dt.

dtp, are (in the present theory)

a
line in the direction of p, or of the
being

differential rectors,

generally finite

d s p, like T) sp,

round the origin,

at least if dr, like

r,

be positive

while

(comp. 100, (9.)) a tangent to the meridian of that spheric

but dtp, like D/p, is on the contrary
surface, for which r and t are constant
is

a tangent
s

to the

small

circle (or parallel],

on the same sphere, for which r and

are constant.
(9.)

Treating only the radius r as constant, and writing p = OP, if we pass

P, or (s, t), to another point Q, or (s + As, t), on the same

from the point

meridian, the chord

HAMILTON

PQ

is

represented by the

ELEMENTS OF QUATERNIONS.

finite difference,

6j

and
3

in like

ELEMENTS OP QUATERNIONS.

482

[III.

5.

manner, if we pass from p to a point R, or (s, t + At), on the same parallel, the
new chord PR is represented by the other partial and finite difference, & p
while the point (s + As, t + A^J may be denoted by s.
t

now

(10.) If

come

the two points Q and R be conceived to approach

be very near

to

the chords PQ and

it,

of meridian

two corresponding arcs

same two circles at P,

drawn

so

and

to

will very nearly coincide with the

parallel

or with the tangents to the

as to have the lengths of those

and

or finally with the differential

and

PR

to p,

tangential vectors, d s p

and

d/p,

if

two

arcs

we suppose

we may, comp.

322) that the two arbitrary and scalar differentials, ds and

dt, are so assumed as to be constantly equal to the two differences, As and A,
(as

and consequently

Whether the

(11.)

d sp

XXYI. and XX.)

(as in

and dt be

differentials ds

Dp

like the product

dtp,

dimmish with them.

to

which the

of

large or small, the product

Dip, represents rigorously a normal vector

length bears to the unit of length

(comp. 281) the same ratio, as that which the rectangle under the two perpen
dicular tangents, d s p and dtp, to the sphere, bears to the unit of area.
Hence,

with the recent suppositions (10.), we may regard this product d sp dtp as
representing, with a continually and indefinitely increasing accuracy, even
.

in the

way

of ratio,

what we may

call the directed element

of spheric surface,

PQRS, considered as thus represented (or constructed) by a normal at P

the tensor of the

in which the negative sign

may

increases

call

and

same product, namely (by XXYI.),

XXVIH.
while

from

is

T(d sp d
.

p)

= -

S a

retained, because

to 1, is

dS

Tir

2s

a 2s

d*,

decreases from + 1 to -

1,

an expression on the same plan for what we

contrast the undirected element of spheric area, or that element

by

considered with reference merely

to

and not with reference

quantity,

to

direction.

all

to

and from

2,

this last differential

then,

(12.) Integrating,

to s

s l9

that

is,

expression

taking the limit of the sum of

the elements PQRS between these bounding values,

equation

XXVIII., from

we

find the following

XXIX.

Area of Spheric Zone =

27ir S(a

2s o

- a 2s
i)

whence

XXX.
and

Area of Spheric Cap

(s)

2irr*(l

- S a 2s =
.

47rr

finally,

XXXI.

Area of Sphere =

4?rr

2
,

as usual.

s 2
(TV. a )

SCALAR AND VECTOR INTEGRALS.

ART. 345.]

In

(13.)

manner the expression

like

account of the decrease of S

and

thus,

to

2s
,

XXVIL,

as in (11.)

),

with

483

its

sign changed (on

represents the element of volume

by integrating from r = r to r = r i9 from

= 2, we obtain anew the known values

to s

and from

1,

XXXII.

^(rf
o

Volume of Spheric Shell =

3
;

and
4-irr3

XXXIII,.

Volume of Sphere

(r)

= -

These are however only specimens of what

(14.)

as usual.

Integration, although connected with quaternion forms

characteristic of the present Calculus, if we
apply

be called Scalar

may
and

it

take the

briefly to

it

more

will be

Vector Integral, or the limit of the vector-sum of the directed elements

(11.) of
a portion of a spheric surface
a problem which corresponds, in hydrostatics,
:

to calculating the resultant of the pressures

having a normal

For

direction,

on that surface, each pressure

and a quantity proportional

to the element of area.

XX

we may employ

the expression
VI. with its
in
order
to
an
denote
inward
a
or
sign changed,
normal,
pressure acting from
without and if we then substitute for p its value XV., and observe that
(15.)

this purpose,

XXXIV.

VdJ
(VdJ

because
cause

0,,

7r

= -

1,,

Jo

and remember that V.

XXXV.
XXXVI.

7;S

*,

we

easily deduce the expressions

of Directed Elements of Spheric Cap

is

radius of the plane

TVAy>,

expression

XXXVII.

and

(*)

inward normals

?rr

(S

a 25 /

2s 2
= - wr z
k(l
(S a ) )
.

is

that of +

A-

TT

(VAy>)

hence,

if

and the
we still

by normals thus drawn inwards, we have

this

Sum

comparing which with

[Since

d/c 2

-1

of Directed Elements of Circular Base =

XXXVI., we
~k*t&(2t-

1)

2
7rA-(V/y>)

arrive at the formula,

by 333

(5.),

circular base, of the spheric segment

so that its area is in quantity =

direction of all its

Sum

represent the directed elements

new

of Directed Elements of Elementary Zone =

corresponding,

common

2s+1

Sum

But the

(16.)

7t

the integral [****d* = i

T
Jo

3Q2

ELEMENTS OF QUATERNIONS.

484

XXXVIII.
a result which

known

may

Sum

[III. n.

5, 6.

= Zcro\
of Directed Elements of Spheric Segment

be greatly extended, and which evidently answers to a

case of equilibrium in hydrostatics.

Quaternions (or of Vectors)

questions

may serve to show already, that Differentials of

may be applied to various geometrical and physical

These few examples

(17.)

and

that,

when

so applied, it

is

permitted to treat them as small9

if

any convenience be gained thereby, as in cases of integration there always is.

But we must now pass to an important investigation of another kind, with
which

differentials will

be found to have only a sort of indirect or suggestive

connexion.

SECTION

On the

6.

of Implicit Functions of Quaternions

Differentiation

and on the General Inversion of a linear Function, of a

Vector or a Quaternion with some connected Investigations.
:

346.
(5.)

and

We
(6.)

when

saw,
),

that

it

differentiating the square-root of a quaternion (332,

was necessary for that purpose to

equation* or an equation of the first degree

f

I.

in which r

rr +

resolve

a linear

namely the equation,

rr =

and q represented two given quaternions, q* and

while r
a
And
dr
d
or
represented
sought quaternion, namely
generally, from
d</,

<f

the linear or distributive form (327), of the quaternion differential

II.

of

any given and

explicit

.dQ = d/?=/(?,

function fq,

when

differential

dq of the quaternion variable q y

former differential to the latter, that is from

d,?),

considered as depending on the

see that the return from the

we

dQ

to dq, or the differentiation of

Q, requires for its accomplishment the

Solution of an Equation
or what may be called the
of the First Degree
Inversion of a Linear Function
are therefore led to con
of a Quaternion.
sider here that general Problem
to which accordingly, and to investigations

the inverse or implicit function

/"

We

connected with which, we shall devote the

however
present Section, dismissing
now the special consideration of the Differentials above mentioned, or treating
t Compare the Note

to

page 452.

DIFFERENTIALS OF IMPLICIT FUNCTIONS.

ARTS. 345-347.]

them only

which the

as Quaternions, sought or given, of

485

relations to

each other

are to be studied.

Whatever the particular form of the given linear

function, fq, may be, we can always decompose it as follows

or distributive

347.

I.

=/S? + /V? = Bq

.fq = f(Sq + V?)

./I + /V?

taking then separately scalars and vectors, or operating with S and

proposed

ii.

where r

on the

is

../<?

we can

a given quaternion, and q a sought one,

and so reduce the problem to the solution of a

Sq,

linear equation,

in general eliminate

linear

and

vector equation,

of the form,

III.

where a

is

shall

a given vector, but p (= ~Vq)

characteristic of a given linear

we

and

=
(ftp

cr

a sought one, and

is

is

vector function of a vector,

used as the

which function

throughout suppose to be a real one, or to involve no imaginary

But, to every such function $p, there always

constants in its composition.

corresponds

what may be

connected with

called a conjugate linear

and vector function

fy

p,

by the following Equation of Conjugation,

it

IV.

where A and p are any two

= Sp^ A

SAfto

vectors.

then, as

Assuming

we may,

that

ju

and v

are two auxiliary vectors, so chosen as to satisfy the equation,

V.

and therefore

Vfiv

<T,

also,

VI.

SA<T

SXfJLV,

S/LLCF

0,

Sv<7

0,

where A is a third auxiliary and arbitrary vector, we may (comp. 312) replace
the one vector equation III. by the three scalar equations,
VII.

And

these give,

by

Spf A

= SA/^v,

principles with

=
S/>0>

Spt v =

0,

which the reader

is

0.

supposed to be already

familiar,* the expression,

VIII.
*

mp =

i/,cr,

or

IX.

p =

$-*<!

= nr

\^a,

student might find it useful, at this stage, to read again the Sixth Section of the preceding
or at least the
early sub-articles to Art, 294, a familiar acquaintance with which is
presumed in the present Section.

Chapter

ELEMENTS OF QUATERNIONS.

486

if

be a scalar-constant, and

\p

an auxiliary

and

linear

[III.

vector function, of

the value and the/orw are determined by the two following equations

X.

6.

which

And

thus the proposed Problem of Inversion, of the linear and vector function
may be considered to be, in all its generality, resolved because it is always
;

(j),

possible so to prepare the second

members

of the equations

X. and XI.,

that

they shall take the forms indicated in the first members of those equations.*
a and
(1.) For example, if we assume any three diplanar vectors a, a
,

deduce from

XII.

them three other

.

j3

Saa

a"

Va

vectors

j3

oSaa

/3

a",

j3

0j

a"

j3"o>

by the equations,
jS^oSaa

Va"a,

a"

= Vaa

then any vector p may, by 294, XV., be expressed as follows:

XIII.
if

we

then

j3

Sa/>

j3

SaV +

j3"

Sa"p

write,

XIV.

|3

/3

</>j3,,,

tP

^)3"o,t

ft"

have the following General Expression, or Standard Trinomial Form,

for a Linear and Vector Function of a Vector,

we

shall

XY.
containing, as

we

.
ft>

= jSSap +

/3

Sa p +

/3"Sa"/[>

/r

see, three vector constants,

]3, /3

)3

or

wme

scalar constants,

such as

XVI.

SJ3, Sa

)3,

Sa^jS;

S|3

So

Sa^

r/
;

Saj3

Sa

Sa"j3";

(and generally will) all vary, in passing from one linear and
but which are all supposed to be
function $p to another such function
and given, for each particular form of that function.

which

may

sub-art.
[For a more elementary solution of the problem of Inversion, see

real,

(4.).]

three
f [The equations XIV. lead to a useful expression for a linear vector function in terms of
and
vectors
vectors
and
and
the
derived
#,
,
)8
/3".]
o,
o,
diplanar
"o,

INVERSION OF LINEAR AND VECTOR FUNCTIONS.

AET. 347.]

Passing to what we have called the conjugate linear function

(2.)

XY.

form

487

fy

p, the

by IV., the expression,

gives,

XVII.

= aS/3 P + a

.
0>

S/3

a"Sj3"p

f>

but

V. (aSfip + a Sj3 /i)

= Vaa S

]3

therefore the transformation

XVIII.

+ a

(aS]3i>

V. pVfiv = Vaa S

XI. succeeds, and

+

Y<i

a"S/3"/3>

I/Y>

= Vaa S

S/3 v)

= V/3

4-

i//

gives,
f

Vaa

S/373,0,

which the conjugate

of

V/3"/3Saa

/3"S"a>

/3

/3V/iv

j3

Va"aS/3/3>

as an expression for the auxiliary function

thus written,

XIX.

+ V/SjS Sa

>

a/o

may

be

r
by interchanging each of
changed to ^ when ^ is changed to
the three alphas with the corresponding beta.
we propose to do,
(3.) If we write, as in this whole investigation

so that

^/

is

XX.
the formulae XI. and

XXI.

X - V^y,

= VvX,

i/

= VA,u,

X. become,

= Y.

i,A

</>

XXII.

and

0Wv,

twSXX = S

Xl

with the same sort of abridgment of notation as in

coefficient of

SaaV

in this last expression

jj

+
the division by

SAX

7
,

or

by

which

may

may

and because the

by XVII. XVIII.,

and we

find the expression,

= Saa VSjS ^jS

when we

does not change

Sj3/3

Va,

j3"Sa

pass from

<

to

on which account we

write also,

XXIV.
because,

by

also be thus written,

XXIIF.
so that

is

^X^X

SX/zv, succeeds,

XXIII.

XXII.

(2.),

SXX = S 0X^ X
.

or

XXIV.

we can deduce from XI.

XXV.

.X

wSX^iv

= S

the conjugate expression,

= V.

ELEMENTS OP QUATERNIONS.

488

(4.)

We

ought then to find that the

XXVI
has

its solution

XXVII.

<r

6.

linear equation,

0p = jSSap +

Sa p +

|3"Sa"p,

expressed (comp. YIII.) by the formula,

pSaa

[III. n.

and accordingly,

if

a"S/3"/3

we

|3

Ya

a +

a"S/3"/3

Y"aS|3|3"<r

XXVI.

operate on the expression

for

+ Yaa S/373*

with the three

o-

symbols,

XXVIII.
we obtain the

p p,

0/3",

j3 /3,

three scalar equations,

XXIX.
from which the equation

8(3"

ft* =

XXVII.

S/3"/3

/3Sap, &c.,

follows immediately, without any intro

duction of the auxiliary vectors A,

/UL,

although these are useful in the

v,

theory generally.
(5.)

Conversely,

XXVII.

the equation

if

were

given,

and the value

of a

sought, we might operate with the three symbols,

XXX.
and

a,

(6.)

each

for

y,

found a useful check on formulae of

It will be

what we have

the first dimension

dimension, but

i//

called the Standard

for then

and

we may say

of the second,

i//

which these symbols

into

the expression

would follow.

ff

beta, in

j3 9

XXIX., from which

so obtain the three scalar equations

XXVI.

and

that

of

Form

this sort, to consider

(1.) of

<f>p,

as being of

and $ are also of the first

the third and every formula,
\

enter, will thus be homogeneous

a,

a,

a",

and

being not counted, in this mode of estimating dimensions, but

treated as of the first dimension, when it is taken as representing 0/j.*

v, p,

<r

(7.)

And

general, yet

although the trinomial form

if

XXXI.

we choose
. ..

to take the

f p - SjSSap,

XV.

has been seen to be

A,

JJL,

being

sufficiently

more expanded form,

which gives

XXXII.

>

- 2aSj3p,

any number of terms of 0p, such as /3Sap, /3 Sa p, &c., being now included in
the stm 2, there is no difficulty in proving that the equations VII 1. and IX.
are satisfied,

when we

write,
*

[Compare the

first

Note

to Art. 350.]

STANDARD TRINOMIAL FORM.

ARTS. 347, 348.]

XXXIII.

and

$p = SVaa

XXXV.

XXXIY.

With

S/373f>,

= 2Saa a

Sj3

/3 )3

SS/3/3

fp = SVjSjS SaV,

489

j3"Sa"a

a.

The important property (2.), that the auxiliary function ^/is changed
when is changed to $ may be proved without any
to its own conjugate
reference to the form Sj3Sp of
by means of the definitions IV. and XI.,
as follows.
Whatever four vectors ,u, v, jui, and vi may be, if we
and
of
(8.)

;//,

<j>p,

i//,

write

XXXVI.

\\ = Yun/i,

adopting here this

= V.

equation as a definition of the function

last

proceed to prove that

XXXVII.

and

it

conjugate to

is

i//,

we may

by observing that we have the

^/,

transformations,

XXXVIII.

SA

,i/A

which establish the relation in question, between

(9.)

And

when we
to

the not less important property

pass from

XV.

the/orm

or

<//,

XXXI.

of

and

(3.),

i//.

that

remains unchanged

manner be proved, without reference

$p by observing that we have by XXXVII.,

may

to

i//

in like
y

&c., the transformations,

XXXIX.

- S

= SA
=
t/^A

0Xi//A

<$>\<t>n<t>v

wSA

A = wSAjuv,

because the equations III. and VIII. give,

XL.

^p

whatever vector p

nip,

so that the value of this scalar constant

original linear function 0, exactly as

function

348. It

is

and ^ from

0/o

as its formula of solution

with the

We

0.
.

II.

gives

<T,

III.

X. or

in

be derived from the

from the conjugate

ntp

i//a,

general method,

have therefore the two

ma

=
(j)if;<r,

mp

above explained, of
identities,

=
\l/<f>p

and symbolically,
III

HAMILTON

was

found, then, that the linear and vector equation,

I.

or briefly

m may now

be

<

deducing

it

may

ELEMENTS OF QUATERNIONS

= 0^ =

i//0

ELEMENTS OF QUATERNIONS.

490

with which, by what has been shown, we

..m

III".

Changing then successively p and

definition of the auxiliary function

6.

may connect these conjugate equations,

=
i/,

and

v to if/p

in the equation

T//V,

of

or in the formula,

i//,

=v.0Vfv,

i/v// v

we

[HI.

347,

xr.,

get these two other equations,

IV.

- J.V.

m V.fi^v

v^fjL

in the former of which the points

accented

Y.

may

i//V.

i;

i>/,

m V^
2

be omitted, while in each of them

be exchanged with unaccented symbols of operation

may

see that the law of homogeneity (347, (6.)

transformations of the same sort

is

and we

And many

preserved.

other

be made, of which the following are a

may

few examples.
Operating on Y. by

(1.)

i/r

YI.

by wr

or

Y.

y
<//>//

^,

we

get this

= w*0Yjuv

new

formula,

comparing which with the lately cited definition of ^, we see that we may
to mty, and therefore also m to m~
to
if we at the same time change
change
$ being then changed to \f/, and i// to
;

i//

\fr t

m<j>\

(2.)

For example, we may thus pass from IY. and Y.

VII.

Vv

and

v,

VIII.

V.

to the formulae,

0>0

we see that the lately cited law of homogeneity is

The equation VII. might have been otherwise

in which

still

observed.

obtained, by inter
- wr
with
and the
or
with
and
in
and
v
i/T
^,
IY.,
operating
changing
of
of
definition
the
formula VIII. may be at once deduced from
equation
In fact, our rule of inversion, of the linear function
by operating on it with
(3.)

jj.

i/>,

<.

0,

may

be said to be contained in the formula,

IX.
where

(4.)

is

By

^V/xv = w^V.

0V<p

a scalar constant, as above.

similar operations

X.

2
.

XII.

XIII.

XL

Y.

and

substitutions,

= w^Vjui; = Y.

f/ju^ v

wj0
wi

Y.

/a

Y.

= iV/ii;

&o.

ARTS. 348, 349.]

SECOND FUNCTIONS, QUATERNION CONSTANTS.

But we have

(5.)

also,

XIV.

Atf>V.

\ = S

<j>p<j>

and
2

formed from

is

i//

= S

so that the second functions

hence, by XIII.,

pf

A,

are conjugate (compare 347, IV.);

as t// from
and generally it will
;

</>

be found, that iin be any whole number, and if we change

n
at the same time
to T// W and w to w* w
to
to ^
T//
i//

may

be remarked that the changes

also

to

<",

we change

",

It

(6.)

491

(1.)

conduct to the

equation,

XV.
and

to

many
The

349.

(S

^Ai/Y^v

expressions,
f/(j)jLL

with the significations 347,

1/<^>V,

XX.

of A
A,

/n,

each such

since

SA/^v,

A ^A +

when

are easily proved to vanish

by

2
0A0/^v) = SA/*vS

other analogous formulse.

A 0A +

divisible

//, i/,

v are

f/l^/LL

4-

lfjV

and others

of the

same type,

complanar* and therefore to be

expression involves each of the three

vectors A, /x, v in the first degree only

the quotients of such
divisions being therefore certain constant quaternions,
independent of A, ju, v,

auxiliary

and depending only on the particular form of

or on the (scalar or vector,
but real) constants, which enter into the composition of that
given function.
<f>,

Writing, then,

I.

..<?:

and
II.

we

^A +

f/^fi

(A

shall find it useful to consider

separately the scalar

and vector parts

two quaternion constants, q and q 2 which constants are, respectively,
fo& first and second dimensions, in a sense
lately explained.!
r
=
(1.) Since VA 0A
&c., it follows tliat the vector

these

^uSi^A

and

of

of

vSA<>,

parts

is
and therefore ^ to \ff. On
change signs, wlien
changed to
the other hand, we may change the
vectors
A, n, v to A //, i/, if
arbitrary
we at the same time change A to V^V, or to - ASA/zv, &c., and SX/av, or
SAA to - (SA^v) 2
then
we find these new

of #1

qz

dividing

III.

qi

IV.

q,

by

SA^uv,

expressions,

[^7 Pitting v = x\ +
t [It may be instructive to the student to reduce these quaternion constants by replacing
and v by xi + yj + zk, x i + y y + z k, and
+
+ k.~]
yft.."]

z"

"i

y"y

3R2

A,

yu,

ELEMENTS OF QUATERNIONS.

492

operating on which by S, we return to the scalars of the expressions

and t//.
and i// changed to
II., with
(2.)

Hence the

by passing

to the conjugate linear functions

VI.

K?

VIII.

(A

A +
<A

fo

n + v

^>

yv

and

I.

are obtained

write,

fv)

&c., in the dividends,

"Kq Z9

and thus we may

K?! = (A0 A + ,/0V + i,0V)

interchanging A with A

VII.

= (A

..?!

v.

or,

and

conjugate quaternion constants, ~Kqi

6.

[III.n.

A/A +

iV

where A = V^uv, &c., as before.

(3.) Operating with V
p on V#i, and observing that
.

V.

- X

pVA ^A

<(ASA

SAfp,

&o.,

while

^(ASA

/o

and
A SA^

/o

4-

with similar transformations for V.

IX.

we

find that

V.

|oV(? 2 ,

and

X.
(4.)

V.

Accordingly, since

the vector 0^ - ty p, if
therefore of the form,

do not vanish, must be a line perpendicular to

it

XI.

.
<pp

in

which y

is

some constant vector ;*

XII.

where the function

of

0jO

and

jO

<>

is *Vs o-ww

namely the

XIV.

and

2VyjO,

so that

we may

conjugate, or is the

write,

common

self-conjugate part

part,

XIII.

And we

<f)

p,

cj0

$(4>

P +

0.

see that, with this signification of y,

.

V(A 0X + fop + fo v ) - - 2 7 SA/iv,

*

or

XIV.

[This vector 7 has been called the spin-vector of the function

Vft - 2y
</>.]

QUATERNION CONSTANTS.

AET. 349.]

while

we have,

XV.

in like

V(AtyA +

493

manner,

fi^fi

+ i/iH = - 2SSA0V,

XVI.

Vy, - -

..

a confirmation, the part

has by (1.) no effect in Vg,
of
and
change 0A to VyA, &c., in the first member of XIV., we have thus,
<

4-

yS(AA +

Since VAty A = - ^VAtf/A

(5.)

by

vSyi/)

and

(1.), (2.),

juju

/)

ySAjuv

if

we

3ySA/ui>.

by 348, VII., while we may

&o.,

23,

P = 2V8p.

...i/y>-i/,

As

XV.

or

write,

(4.),

XVII.
XVIII.

or

XIX.
and

XX.
we have

between the two new vector constants,

this relation

XXI.
for 0,

that

form

is
;

and

have

We may

(6.)

all

the

0-y

sfl/we

^)

this

XXIII.
If

particular vector, 7.

8- JV(V/3

7
.

j3"

XV.,

we have

of 0^,

iV(/3a + |3V +

)3"a"),

VaV + V/3 73 V

r/

we denote by ^ and m

what

i//

and

the expressions,

a + V/3|3

r
.

become when

).

is

changed

easily find that

XXIV.

i/p

=
!/,/>

7 S r/0 +

so that the
self-conjugate part of

Vgp

XXV.

^p contains a
and

the vector 7, but

only in the second degree

fP

term,

7 again entering only in an even degree, because

pass from

to

or

from 7

[Expand

to

m
-

+ 878 =

7.*

= ^
-

in like

XXVI.
we

is of the first dimension, and

of the second dimension, with
respect to the betas of the standard

XXII.

we

= -

on

^<?c#,

in fact, with that /cm, 347,

(7.)

add that the vector constant 7

and

to 0o,

0/

T S 7jo

Vfy>

ySyp, which involves

manner,

87^7

remains unchanged, when

ELEMENTS OF QUATERNIONS.

494
It

(8.)

is

evident that

we have

XXVII.
and
and

[III. n.

6.

the relations,

and

that, in a sense already explained,

>

^o>

are of the

first, second,

third dimensions, respectively.

350. After thus considering the vector parts of the hco quaternion constants,
q

and

qz

new scalar

we proceed
constants,

to consider their scalar parts

m"

and

and

will lead to the

whence
conjugate auxiliary functions, %p and xp
ment of a certain Symbolic and Cubic Equation,

is

the

by

satisfied

will introduce two

employment

new

of two

also will result the establish

>

which

which

Linear Symbol of Operation, 0, and

is

of great

importance in this whole Theory of Linear Functions*

(1.)

Writing, then,
II.

we

= 8q l9

m"

and

III.

= S? 2

new constants changes value, when we

because, in such a passage, it has been

see first that neither of these two

pass from

seen that

denote by

to
</>

from 7

or

we only change

m\ and m

to

and

q 2 to

m and

\ what

and ~Kq 2

K<? t

become, when

m"

Accordingly,

is
</>

changed

to

if

we

we

<

easily find the expressions,

IV.
* [Or

directly,

m"

=
i"o

and

without introducing ^ or %
3

ASA</>A</>

will generally exist.

This

where the

may

\(p

\ +

y\

an arbitrary vector A the relation

2

^)AS^)

3
AA^ A +

2
4>

ASA(A<f>

be briefly written in the form,

3

coefficients

^ or
>

2
\S<p\(p

V.

mi<f>-\

can only depend on

<

mz\

mz(f)\

and

A.

0,

Operating on this by

and

2
<

<p

=
and

= 0.
But an arbitrary vector

may

be expressed in the form

and hence from the three equations, on multiplying by

tc,

y,

and

msp

z,

and adding, the equation

results.
This must be identical with the equation found by treating p directly, in the same manner
as A has been treated, and therefore the coefficients
must be independent of A. The suffixes here
See 347 (6.).]
printed serve to indicate the dimensions of the in.

SYMBOLIC AND CUBIC EQUATION.

ARTS. 349, 350.]

(2.)

It

and

be noted that

may

or

w",

w"

of the first dimension, but that

is

495

are of the second, with respect to the standard

form of

and

accordingly, with that form we have,

VI.

= S/3 + Sa j3 +

m"

Sa"|3"

and
VII.
(3.)

If

S(VaV

we introduce

V/3 73 +

Va"a

V/3/3"

and x

two new linear functions, xp

VIII.

+ Vim
/o,

V/3

/3).

such that

and
IX.
it is

each

easily

proved that these functions are conjugate to each other, and that
dimension in fact, with the standard form of $p, we have

of the first

is

the expressions,

X.

XP = V(aV/3 o + a

XI.

xp = V(/3Va P +

and S XaV/3/5 = S pjSVaA, &o.

will be found that

Also,

V/3>

]3

Va>

a"V/3"|o),

jS

To V)

formed from

XQ be

if

OJ

as

^ f rom

0>

it

XII.
where

is

xp

XP

and

Vyp,

XIII.

x p = X/ +

VT/

of the first dimension.

x<>

(4.)

Since

= 8
the expression II. gives,

by 349, V., the equation,

XIV.

m UU

= S

A and A being two arbitrary and independent vectors

our having ihe functional relation,

XV.
or briefly

=
$p + xp

"p

which can only

be,

by

and symbolically,

XVI.

x+

Accordingly it is evident that the relation

XP) combined with the standard form of

XV.

<j>p,

the constant

m" .

"

<j>

is verified, by the form X. of

and with the expression VI. for

ELEMENTS OF QUATERNIONS.

496
(5.)

The formula XVI.

[III. n.

6.

gives,

XVII.
and accordingly the identity
and v to
by changing

=
easily be otherwise proved,

in the

i//v

4>X

and $x may

x<

and

/u

*"0

x<

of

i//

f.i

definition

VIII. of
x>

remembering that

V
for thus

we

fv =

>

wjVuv

i/

and

wa,

have,

XVIII.

as required.

Since, then,

(6.)

the value III. of

?;/

by 349, VI., the equation,

gives,

XIX.

w SAA

A and X being independent vectors

XX.

= S

Aty +

ipp

And

m p,

(f>xp

which

/B".

VpVa

and thus the symbolic and

(8.)

combined with the value VII. of

is

for

i//p.

between the two equations XVI. and

=
= m, we find the symbolic expression,
i//-/>

mf

=
i//

coefficients,

XXIV.
may

also be

Sp

proved thus
3

(<>

and therefore vanishes.

2
m"0

;
<j>

cubic equation I. is proved.

because the

conjugate equation,

[Tins

+
m"(j>

m,

m",

seen to remain unaltered, in the passage from

V^oV

V/3"/3

x>

r/>^

a"

equation

XXIII.

(7.) Eliminating the symbol

XXI., and remembering that

And

/>x

XX., when it
expression 347, XVIII.

verifies the

and with the

V^

^+

form of $p we have

in fact, with the standard

XXII. ..^ XP = XtP =

hence,

or briefly,

XXI.

X )X

+m

This requires

to

<p

m) a

2
<

-I-

m}

have been

write also this

.*

tf>

vectors,

m"<j>

ty

(^>

3
m"

Sff

we may

m"<t>

and a are arbitrary

If p

(0

m tf +

of that equation,

cr

by 348 (5.),
$ - in) p,

0,

as p is arbitrary.]

SYMBOLIC AKD CUBIC EQUATION.

ARTS. 350, 351.]

Multiplying symbolically the equation

(9.)

by

\p(f>

In general,

(10.)

therefore
m"

defined

m<j>,

while \
is

XYI.

to

And

if

(11.)

other

2
w<

by

may

changed

XXI.

^/

m^

-t

in

and reducing
\f/,

\fs.

when we change $

to

same time changed

but also

to

to

0x>

mm",

and the

or to

to

and

i//,

x$>

we may

thus pass from the

and conversely, from the latter to the former.
to q 2

new

the two

mm

we change not only

at the

is

be remarked, that

may

it

as before,

to

to

quaternion
relation

$ may be substituted for

which

and

by

= m, we eliminate the symbol 0, and obtain this cubic

XXV.
in

I.

497

Accordingly,

auxiliary functions,

the equations VIII.

and IX.,

x and x be considered
,

as

their conjugate relation (3.) to each

be proved, without any reference to the standard form of $p, by

reasonings similar to those which were employed in 347,

corresponding conjugation of the functions ^ and i//

to establish the

(8.),

It

(12.)

may

be added that the relations between

<,

x,

and

m"

>

the following additional transformations, which are occasionally useful

XXYI.
XXYII.

give

with others on which we cannot here delay.*

351.

The

cubic in
I.

where p

is

be thus written

may
.

= mp -

an arbitrary vector.

m (pp

m"(j>

If then

$*p

happen that

it

3
</>

/o

0, &c.,

the constant
II.

if

0JO

must be

sero

and

0,

T/o

some particular
= 0,
= 0,

for

but actual vector, p, the linear function $p vanishes, so that

<j>p

<fp

or in symbols,

then

0.

>

0.

Hence, by the expression 347, XXIII. for m, when the standard form for
adopted, we must have either

fyp

is

III.

Saa

a"

0,

or else

IV.

Sj3"j3

so that, in each case, that

generally trinomial form, 347,
*

[Without introducing x since

A is arbitrary, that
,

follows, as

HAMILTON

for

w"V,ui>

any three vectors

= 0T/*y +

ELEMENTS OF QUATERNIONS.

V/t^j,

m"S\nv

V^>/*v.

= SA

This

is

)3

=
;

XV., must admit

(^ Y/tv

+ Y/^v + V^/tv
XXYI.]

equivalent to

of

ELEMENTS OF QUATERNIONS.

498

being reduced

a binomial.

to

[III. n.

when we have thus

Conversely,

6.

a function of

the particular form,

V.

we have

$p

-f

jSSap

j3

So>,

then,

VI.

0Vcm =

and a be actual and non-parallel lines, the real and actual vector
Vaa will be a value of p, which will satisfy the equation 0p =
but no
other real and actual value of p, except p = #Vcm will
if
that
so that

if

satisfy

/3

and

be actual, and non-parallel.

/3

In

every other vector to the fixed plane of

of

and since we have

<f>p

.
<t>

we

]3, ]3

which plane

p = aSfto + a

S/3>,

see that the locus of the functionally conjugate vector,

namely that of

which

is

by

p, is another fixed

Also, the normal to the latter plane

HCLQ

is

the

is

in like

VIII.

line

while the normal to

former operation, namely by

manner the line, which is annihilated by the
since we have

destroyed

the former plane

operation,

a,

$ reduces

therefore the locus

is

also,

VII.

plane,

equation,

this case V., the operation

latter

0T/3]3 =0,

or its
but not $ p
0, for any actual p, in any direction except that of Vj3/3
same.*
the
for
as
the present purpose, be regarded
opposite, which may however,
,

In

this case

we have

IX.

monomial forms for

also

$P = Vaa Sp pp,

so that the operation

X.

and

and ^p, namely

i/y>

.
i/,>

Sa

V/3j3

a,o

destroys every line in the first fixed plane (of |3, j3 ),

and

the conjugate operation

annihilates every line in the second fixed plane
a
On
the
other
hand, the operation $ reduces every line, which is
(of a,
).
out of the first plane, to the fixed direction of the normal to the second plane
;

and the operation

reduces every line which is out of the second plane, to that

i//
And thus it conies to
other fixed direction, which is normal to the first plane.
and then with $ or first with
pass, that whether we operate first with
;

i//,

and then with ^

with

t//

or

first

with

in all these cases,

we

$ and then with

or

first

with

<//;

<j>,

<

and then

arrive at last at a null line, in conformity with

the symbolic equations,

XL
which belong
*

^T//

i//0

i//

i//0

0,

to the case here considered.

Accordingly, in the present investigation, whenever we shall speak of a

of a given line," &c., we are always to be understood as moaning,

"direction

direction."

fixed direction,
"

or the

or the opposite of that

BINOMIAL FORM, FIXED LINES AND PLANES.

AETB. 351, 352.]

499

Without recurring to the standard form, of

the equation 348, YL,
= w^Y/iv, and the analogous equation Y.
=
i//,ui//v
i//^//v

(1.)

</>p,

namely Y.

w$

Y/uv,

might have enabled us to foresee that T///O and -^p, if they do not both
and therefore
constantly vanish, must (if m = 0) have each & fixed direction
;

must be expressible by a monome, as above the fixed direction of

that
of a line which is annihilated by the operation 0, and similarly
i///o being
and
for
$
p
that each

\fr

(2.)

And

because,

by 347, XI. and XXY., we have

= Y.

so that the line

Y/v = Y.

and

actual, is perpendicular to \fJVfiv,

and the

line 0/i per

pendicular to T//YJUI/, we see that #c/j of the ttco lines,
p and 0/o, must have
(in the present case) & plane locus; whence the binomial forms of the two
(p /m,

if

<J)

$p and Q

conjugate vector functions,

p,

might have been foreseen

\f,p

and

\j/p

being here supposed to be actual vectors.

(3.) The relations of r eci angularity, of the two fixed lines (or directions), to
the two fixed planes, might also have been thus deduced,
through the two
conjugate binomial forms, Y. and YIL, without the previous establishment of

the more general trinomial (or

standard) form of
The
existence
of
a
and of another for
(4.)
plane locus for
</>.

<pp,

case

when

0,

might
S

also

p, for

X41

= S

and the same equations might have enabled us to foresee, that the
constant m must be zero, if for
any one actual vector, such as X, either

scalar

0/\ or

becomes
(5.)

the

have been foreseen from the equations,

null.

And

the reducibility of the trinomial to the binomial

form,

when

this

last condition is satisfied,

might have been anticipated, without any reference

to the composition of the constant
m, from the simple consideration (comp.
294, (10.)
diplanar

),

that no actual vector

p can be perpendicular, at once, to three

lines.

352. It

may

happen, that besides the recent reduction (351) of the linear

function $p to a binomial form,

when
I.

exists

the relation
.

between the constants of that function, in which case the

symbolic and

cubic equation 350, I. reduces itself to the

form,

3 S 2

ELEMENTS OF QUATERNIONS.

500

thus losing

[III. n.

6.

absolute term, or having one root equal to zero, this equation

undergo a further reduction, by ttvo of its roots becoming equal to each

may
other

its

III.

or in another way,

V.

arise,

is

principal

it

2
.

and

0,

by the existence
w"

In each of these two

which

by our having

either

namely

may

4m

we

cases,

..<!>

=
\m")*

0.

tivo

fixed planes,

and

of

which the

which are the

loci

(j>p

satisfies at
(8.)

),

satisfies this

<

2
(3>

new

f-

Jw",

cubic,

i");

the same time a cubic equation with the same

coefficients

namely

X.

We saw

YIIL

and

symbol

IX.

(1.)

0(0

and functionally conjugate vectors namely, in

and fy p and in the case V. YL, the loci of

Jw",

so that, in this last case, the

(comp. 350,

shall find that certain neiv geometrical relations

=
4>

<!>

=
w")

if

YIL

while

(0

VI.

the mutual rectangular ity of

the case III. IV., the loci of

O p,

be interesting briefly to investigate

(comp. 351) of certain derived,

p and

of these other equations,

and

0,

IY.

in 351, (1.), (2.), that

(0 +

when

ij")-

the line

T///O

has generally

a fixed direction, to which that of the line $p is perpendicular and that in like
manner the line \frp has then another fixed direction, to which /o is perpen
;

If then the plane

dicular.

we must

also

loci

have the fixed

of

0p and $ p be at right angles

and T//JU rectangular, or

XL

= S

independently of the directions of A and n

XLI.
since p
(2.)

i/^

is

= Pn,

to each other,

lines i//A

whence

XIII.

or

= 0,

an arbitrary vector.

Now

in general,

by the functional

relation 350,

= m, we have the transformation,

XI Y.

z
.

=
<

XXI. combined

with

EQUAL ROOTS, RECTANGULAR LINES AND PLANES.

ART. 352.]

if then

^ must

the symbol

0, as in I.,

equation,

XV.
which

501

satisfy the depressed or quadratic

i/,

4*

accordingly & factor of the cubic equation,

is

XYI.

whereto the general equation 350,

\l?

XXY.

is

3
i//

this supposition of

by

reduced,

vanishing.

we have

If then

(3.)

the condition XIII.

to,

not only

by XY.

is satisfied,

but

0, as in I.,

and the two

also

0, as in III.,

planes, above referred

are generally rectangular.

(4.)

We

might indeed propose

xvn.
for every direction of

vanish (by 349, II.)

The

$ =

that

o,

but in

xvir.

is,

this case,

by sup

be equal to

still

particular supposition

$p =

o,

the quaternion constant, q z would

and therefore the constant

would

(by 350, III.),

(5.)

to satisfy that condition XIII.,

we had always,

posing that

as beiug its scalar part

zero.

XVII. would however

geometrical character of the question; for

it

alter

completely the

would imply (comp. 351,

(2.))

that the directions of the lines $p and

p (when not evanescent) are fixed,
iu stead of those lines having only certain planes for their loci, as before.
<j>

On

(6.)

functions,

we should thus

the side of calculation,

$p and

p,

<j>

have, for the two conjugate

monomial expressions of the forms,

XVIII.

.
</>

fiSap,

<t>

= aSpp

whence, by 347, XV1IL, and 350, VII., we should recover the equations,
f
=
and m = 0.
\pp

We

(7.)

XIX.
so that

(ftp

and in

like

should have
.

now

$p

also, in this particular case,

if

0,

vanishes, if

manner

fip

p_La,

p be any

(8.)

line

a null line,

is

at right angles to the other given

the two

and

if

XX.

fy

p and $p

0,

.
<p

_L j3

is

line, )3.
]3,

or the fixed directions of

will be rectangular (comp. (1.)

p be in that other fixed plane, which

),

if

equation,

xxi.

if

in the fixed plane perpendiculur to a

These two planes, or their normals a and

lines

tfp

o,

or

xxr.

.
<p

o,

we have

this

new

ELEMENTS OF QUATERNIONS.

502

for every direction of p

6.

and accordingly the expression XVIII. gives

Saj3

<f>~p

[III. n.

.
<j>p

if

0,

and

/3 _L a,

reciprocally.

Without expressly introducing a and (3, the equation 350, XXIII.

= 0, and therefore also m = 0, as in
shows that when
(4.), the symbol
(9.)

i//

satisfies

(comp.

(2.)

the new quadratic or depressed equation,

XXII.

m"$

<f

which

accordingly a factor of the cubic IV., but to which that cubic

reducible, unless we have thus ^ = 0, as well as m = 0.
is

The

is

not

and rectangularity of the tfwo

=
which we have respectively
and p = 0, without ^p
generally vanishing (a case which it would be useless to consider), is that the
(10.)

condition, then, of the existence

planes (7.), for

<j>p

four following equations should subsist

XXIII.

or that the cubic IV.,

0,

and

its

0,

m"

XVII.

and

0,

quadratic factor

i/,

XXII., should reduce them

selves to the very simple forms,

XXIV.
the cubic in
(11.)

having thus

We may

cubic 350,

3
.

its three roots equal,

also observe that as,

I. is zero,

that

is

XXV.

and

0,

when

XXVI.

0,

and

null,

when even

=
;

and

\^p vanishing.

one root of the general

the vector equation

.
<j>p

was seen (in 351) to be satisfied by one real direction of p, so when we have
also m = 0, or when the cubic in
has two null roots, or takes the form IV.,
then the two vector equations,

XXVII.

.
fr>

0,

are satisfied

$p =

0,

by one common direction of the real and actual

have, by 350, XVII. and XX., the general relation,

(12.)
it

And

follows that

because,

when

by 350, XV., we have

line

also the relation

the three roots of the cubic all vanish, or

xp

because

when

we

"

"p

01

>

the three

CASE OF DEPRESSED EQUATION.

ART. 352.]

XXIII.

scalar equations

are satisfied, then the three vector equations,

XXVIII.
have a common

and

(real

503

^=

0,

fc>

actual) vector root

Xp =

0,

0,

or are all satisfied

by one common

direction of p.

(13.) Since

m"

the cubic IV.

x>

be written under any one of

may

the following forms,

XXIX.

x =

=
<x0

X0 =

0X = &c

->

which accented

may be substituted for unaccented symbols and its

geometrical signification may be illustrated by a reference to certain fixed lines,
in

fixed planes, as follows.

o,nd.

from

first

Suppose

(14.)

that

zero, so that the cubic in

XXIV.

and
is

both vanish, but that

m"

different

is

reducible to the form IV., but not to the

and that the operation

which is here equivalent to

or
t//,
does not annihilate every vector p, so that (comp. (4.) (5.) (6.) )
f
and p have not the directions of two fixed, lines, but have only (comp. (1.)
and (3.) ) two fixed and rectangular planes, n and II as their loci; and let the
form
to -

0x>

x$>

<j>p

<f)

normals to these two planes be denoted by X and A

rectangular lines,
(15.)

line A

of the

Then

itself,

A and A

it is

while

form x\

common

IUL

0) to the plane

(j>~p

ju,

may

also

be denoted by x -1

XXX.

= xv =

=
;

II.

is,

direction of all the lines ftp, whatever the direction of p

2

and

destroys the

every line which is not

U _L A and that it reduces every
in the same plane, which is thus the

reduces* every other line (that

with Vo? =

the symbolical equation,

of

that these two

are situated respectively in the planes IT

line in that plane to & fixed direction,

And

so

shown (comp. 351) that the operation

easily

it

may

be.

expresses that this fixed direction

0,

or that

ni"n

fa,

which can accordingly be otherwise proved

and x
conjugate symbols,

we have

the equation,

if
fji

<fp,

with similar results for the

(16.) For example, we may represent the conditions of the present case by
the following system of equations (comp. 351, V. VII. IX. X., and 350, VI.
VII. X. XL)
:

=
w"

*
zero,

m
=

= S (Va

S]3

+ Sa

jj

V/3 0)

= S/3Sa

j3

Saj3 Sa

f3,

We propose to include the case where an operation of this sort destroys a line, or reduces it to
under the case when the same operation reduces a line to o.Jijced direction, or to a fixed plane.

ELEMENTS OP QUATERNIONS.

504

a V/3

v o=Yfl3Yao

XXXII.

and

may

"p

0p>

X =
.

Yaa

SXX =0,

||

0j3

fi

0V

||

fa,

= SX

SX/z

/-

which we easily find that

after

/\

10X =
/2
/j,

we have

Since

we have

manner,

is

xj

,/o
2

,/

II

p||,u,

0,

xV =

thus

o =

0,

where p

is

f\

x^=0.

a line in the fixed direction

-JXA>

/*

therefore a js?/0 perpendicular to the line p

in this plane J_ //

</>x

0,

in like

And

^?

or of X

results,

=
0>

it

X0

see,

destroys every line

are interchanged

by XXXII., that $xp has

= 0.
= 0, because
0x/
0X
r

and that

or X0 destroys every line in

reduces every line and that thus the

see also, that the operation

the plane n, to which the operation

symbolical equations, 0x
(19.) As a verification,

when accented

Accordingly we

with unaccented symbols.

Yau

and that the operation

with analogous

the fixed direction of

and

the locus of the line \p.

is

x =

may be interpreted as express

reduces every line in this wew? j9/cme of \p to the
.

direction of 0~K), or of

We

//

or

the normal to a plane, which

is
/UL

x/*

>>

(j)fj.=m/ui,

SpxV = VXP.

the symbolical equations,

ing, that the operation

(18.)

<M

also the equation,

XXXY.
the locus of

0>u/*,

-I

/-v

(17.)

(0V =0,

AAJLI V

^a?^

6,

then write (not 7^re supposing A = Y^uv, &c.),

XXXIII.

of

/o)

[III. n.

0x>

=
^>

may

^ e interpreted.

be remarked that the fixed direction X

may

r
,

of

two fixed planes

ought
of
and x/o and accordingly it is perpendicular by XXXIII. to their two
normals, \ and // with similar remarks respecting the fixed direction X, of
or

(f>Xp

to be that of the line of intersection of the

x0j>

<f>p

<f>

perpendicular to X and to /a.

= 0, and
Let
us
next
(20.)
suppose, that besides m

xp

or

but that

x$

which

is

p>

/r

is still

different

that the expression for

zero.

In

itself to

ft>

therefore that the operation

it

from

reduces

we have

this case, it

0,

we have ^ =

has been seen

the monomial form, jSSap

0,

(6.)

and

destroys every line in a fixed plane (_L a), while

reduces every other line to a fixed direction

that plane, because

not

now

Sa/3

0.

(||

]3),

which

is

not contained in

CASE OF MONOMIAL FOEM.

ART. 352.]

In

(21.)

we have by

this case

rrp

(/OO aV/3/o
quadratic

C!
/3 ^ A
= oap
r ^ U,

"

<

(m"

to 0, the expressions,

j3

XP

tip>

"V-

x/
/3Vap = (w

<>,

XVIII.

so that the equations

XXII.

Q/Q
= abpp,

X p = Y.

equating a or

(16.),

6p = ipoap,
rr
r

505

in 0,

may

are reproduced ; and the depressed cubic, or the

be written under the very simple form,

XXXVII.

= AY =

and

vd).

the operation $ here reduces an

while
|3,
x destroys every line in that
and conversely, the operation x reduces an arbitrary line to the

Accordingly (comp.

(5.)

(7.) ),

arbitrary line to the fixed direction of

direction

and ^ destroys every line in that fixed plane.

= 0, the fixed direction of
is

fixed plane perpendicular to a,

But because we do not

here suppose that

not contained in the fixed plane of

and

of
<J>p

the equation \pp

XXXI.

we have
=

We

<j>

0,

and

0,

m"

0, as in

have also

= Sa/3Sa

now

/3

Sa/3 Sa

= Sa/3 + Sa

]3,

~
xp =

xp = 0p,

becomes simply 3 =
to observe that we have not here (comp.
= 0, since we have now on the
tion
<

/3

0, as in
(9.)

<t>

XXIV.

>

but

it is

important

the depressed or quadratic equa

contrary the two conjugate expressions,

<f>*

XL.

.
<fp

= $p = Vaa

which do not generally vanish.

observing that

2
<

S/373/o,

And

fV

$p =

the equation

3
<

In

is

SaV
now

interpreted,

all lines in

<p~

which are

2
<

this last case (23.), in

and are
S

V/3/3

here reduces every line to the fixed direction of

while ^ reduces an arbitrary vector to that fixed plane,

destroyed by

HAMILTON

that

the expressions,

and the cubic in

(25.)

XXIII., but

p reappear, but with these two equations of con

whereof only one had occurred before

XXXIX.

equal,

(10.)) the directions

vector constants,

XXXVIII.

by

and

not satisfied for all directions of p, then the binomial

is

<j>p

(24.)

and

of

between their

dition

<j)p

(8.)

On

vanish, or that

forms

p are not rectangular to each other.

the other hand, if we suppose that the three roots of the cubic in

<f>

(23.)
(f>

\p

m"

and (comp.

null, the

theorem

ELEMENTS OF QUATERNIONS.

which

(12.), of

all the roots of

the cubic in

the existence of a common

are

vector root
3

506

ELEMENTS OF QUATERNIONS.

of the three equations

XXVIII. may
,

now,

XLI.

6.

ir.

be verified by observing that we have

;/Yaa =

0,

4>Vaa

[III.

X Vaa =

0,

we had supposed

the third of which would not have here held good, unless
fw"=Q.

This

(26.)

last condition allows us to write,

XLII.
the lines

n and

XLI1I.
this

new

fa =

0,

IT

(16.),

V/A

0,

8^

0,

in direction with the normals

0,

X and A

VAA

||

Vjiju

S^v =

so that

S//v

0,

vector v will be a line in the intersection of those two rectangular

In

/3,

/3

b,

may

a, b

which p

(in

be any four
may be any

XLV.
XLYI.

XLVII.
so that the scalar, So]3

nion* Vat/ .Vj3

j3,

(28.)
(26.), it

jSiSa./o

+ Sa

V
jS

/3

iSa

=
j/o

jSSa^ +

j3

Sa>

/
1

the vector, flSap +

/3

Sa p

and the quater

remain unaltered in value, when we pass from a grimi system

of four vectors aflafi

of the forms XLIY.

//,

4-

may be any four vectors, and

conduct to the following transformations

scalars,

/u,

and

= #a

vector]

p,

general, it is easy to prove that the expressions,

AljIV.

in which a,

<p

compose a rectangular system.

A, v) will

(27.)

to

0,

planes, which were lately seen (14.) to be the loci of the lines ^p and
are now (comp. (17.) ) the loci of %p and % p
and the three lines
(or A

then we write,

if

v =

VpX

0,

<>

and n thus coinciding

fjf

the planes

by

to another system of four vectors aijSio

ijS

i,

by expressions

With

may

We have,

the help of this general principle (27.), and of the remarks in

be shown, without difficulty, that in the case (23.) the vector

in these transformations,

examples of what

may

be called Quaternion Invariants.

VECTOR AND QUATERNION INVARIANTS.

ART. 352.]

constants of the binomial expression flSap +

real loss of generality, be supposed subject to

XLVIII.
which evidently conduct

So/3

= Sa

Sa p

(3

j3

507

may, without any

for fyp

ihefour following conditions,

= So

S/3/3

/3

to these other expressions,

~VT
^vJUl^X.

w^

T"V

(p

/3Q /Q Q
p = pfOCtp (3d

Ji3
(b

p,

A
= U

and thus put in evidence, in a very simple manner, the general non-depression
2
= 0.
of the cubic 3 = 0, to the quadratic,
<

The

(29.)

but

ff/.$0
i/<

case,

0,

and therefore

</>

linear function
<f>p

=
Sa/3

when we have not only

or sub-case,

between

reducing

0,

=
3

form of

0, as a depressed

<

0,

w"

0,

0,

by the

the monomial /SSap, with the relation

itself to

constants, has been already considered (in (10.)); and

thus the consequences of the supposition III., that there are (at least) two
its

equal but null roots of the cubic in 0, have been perhaps sufficiently discussed.
(30.) As regards the other principal case of equal roots, of the cubic

equation in 0, namely that in which the vector constants are connected by

the relation V., or by the equation of condition,

L.

mm

4m =

+ Sa /3

(Sa|3

4S(Vaa

=
it

may

remark that

suffice to

it

.V/3

by VL,

conducts,

j3)

- S

(Sa/3

j3

+ 4Sj3 S

by VII. and IX.,

or

/3,

to the

symbolical equation,

LI.

2
0<D

and that thus

its interpretation is

if

<

\m" ;

precisely similar to that of the analogous

equation,
2
X<

as given in (14.),

0,

\ =

where

and in the following

m"

sub-articles.*

*
[The following resume of the special cases discussed in recent
Assuming arbitrarily any three constant and diplanar vectors

function
<f>p

sion x, x

may

and

XXIX.,

0,

be resolved along these three vectors

are linear and scalar functions of

thus
<pp

articles
ft,

xft

fi

+x

ft

may not be superfluous

and

ft",

x"ft".

any linear vector

In this expres

and may consequently be replaced by Sap,

Sa p, and
= ftSap + ft Sap +
Hence the trinomial form
p.
is established, and the
function
is made to
When these are given, $ is deter
depend upon the three vectors a, a, and
mined and conversely, when
is given, the three vectors
and
are determinate, retaining
a, a
always the same set of vectors of reference ft, ft and
Special cases will arise when special
relations connect a, a and
,

x"

p,

So."

&"Sa"p

<pp

a".

<p

<p

If

$p =

a"

ft".

a".

for a particular value of

p,

Sap

= Sap =

Sa"p

are necessary consequences, and

3T2

ELEMENTS OF QUATERNIONS.

508

When we

353.

but not

0,

are all unequal,

of the cubic in

and the two

have

(347),
I.

and

when

(j)p,

these two scalar roots,

+ m

Ci,

is

4m

the three roots

still null,

as before

ff

c~2

to

and then dividing by

c,

We

real or imaginary.

shall

are real, there are then two real

p and p 2 in i^i fixed plane II which is the locus (351, 352) of the
possessing the property that for each of them the homogeneous and

directions,

line

m"

6.

scalar equation, with real coefficients

is

find that

nor

0,

formed from the cubic by changing

are also necessarily unequal, whether they be

which
c,

while one of them

of the quadratic

roots

[III. n.

vector equation of the second degree,

II.

or

0,

0p

p,

il

namely by our having,

for the first of these

directions, the equation

III.

= -

(j>pi

and

Y/00/o

without p vanishing

is satisfied,

two

for the second of

IY.

or

pi,

0^! =

if

0,

0!

= -

or

2/0-2,

p2

if

0,

<

4- Ci

and
But
o, a
(if actual) must be complanar.
nomial form, reduces on rearrangement to the binomial form
a"

+ c2

(j>

but that no other direction of the real and actual vector

(j)

them the analogous equation,

<t>p-

therefore

if a,

p, satisfies

and

o"

the equation

are cornplanar, the tri

provided a and # are the scalars determined by the relation of complanarity

versely, if the trinomial reduces to a binomial form, the three vectors o. a , and

a"

a"

aa

(if

Con
must be

a a.

actual)

complanar.
Further reduction to the monomial form will not be possible unless these three vectors are parallel.
In general, also, as typ - Ya a"Sp"& p 4 Va"oSj8"p + Yaa Sy8 /3p, typ will not vanish identically, or the

equation

ij/

This easily follows on replacing p

will not be true, unless the vectors are parallel.

by #, $ and
Remarking that, when $

successively

".

is expressible in a binomial form, it reduces those vectors which it does

not annul to a fixed plane, we may assume a plane containing a pair of arbitrarily chosen vectors
and consider all those functions
which reduce vectors to this particular plane. Just as in
j8 and /3
Sa p,
the case of the trinomial form, these functions fy may be expressed by the type typ = 0Sap +
5

<f>

and they depend on and may be determined by the vectors a and

served unchanged.

second root of the cubic will vanish

two ways

(1)

when Voa =

0, in

if

m =

SVoa

will vanish for all values of p, or ^ =

when the plane of a and a is at right

(2)

when Vaa

is

is

reducible to the

actual

angles to that of & and

the vectors

/8

/3

and

are pre

This may happen in

monomial form, and typ

equal to zero.

is

V/3

which case the binomial

if

that is,
and perpendicular to Y
In this latter case, the assumptions
,

ART. 353.]

CASE OF UNEQUAL KEAL ROOTS.

V., except that third

which has already been considered (351), as satisfying

509

the linear and vector equation,

V.

and

rectangular, to each other

is

^o =

VI.

by the

Tp

0.

>

directions, pi, p z , are not only real,

to the third direction p,

when

self-conjugate (349, (4.)), or

is satisfied

with

0,

shown that these two

It will also be

rfrp

= Sp^A,

S\<j)p

by the

given form of $, or

but

the linear function

the condition

VF.

or

0/o,

when

which enter into the

constants

composition of that linear symbol but that when this condition of self-conju
gation is not satisfied, the roots of the quadratic I, may happen to he imaginary:
and that in this case there exists no real direction of p, for which the vector
;

equation II. of the second degree

The most obvious mode

(1.)

actual values of p, except that

by

to satisfy the linear equation

V.

than
=
an(i ^
to assume an expression of the form, p
#|3 + # j3
=
=
+
the
+
with
0,
satisfy
equation,
$p
j3Sap
c)p

is

through V.,

of seeking to satisfy II., otherwise

>

seek thereby to

(<

two

satisfying separately the

by

is satisfied,

which has been seen before

one direction

VII.

= x(c +
S]3) + /S/3

scalar equations,

= x

+ Sa

(c

j3

+ *S

j3,

a
is
a =
-f bVftft and a
-f b Yftft are legitimate when a, a
b, and V are scalars, while
some vector in the plane of ft and ft and not, as before, diplanar to them. Replacing a and a, the
= (aft + a ft )Sft"p + (bft + b ft ) SV/3 p is obtained, and typ = (ab a b)
new binomial form,
a&"

ft"

ft"

<pp

V.

ft-

Vftft Sft ftp.

= Saft + Sa ft = S(aft + a ft ) = 0, or if
Again, a third root wil vanish if
V(0 + a ft )Vftft
Examining separately the case in which the symbolic equation of the binomial is depressed to a
2
= 0. It cannot be of the form
+
quadratic, it is seen at once that it must be of the form
in"

ft"

ft"

\\

<

x(j>

2
<

On

4xq>

0,

for
<$>p,

calculation of

it is

<p-p,

<p~p

and

if this

Xtpp

&c., are in the plane of

found that

vanishes for

ft(SaftSap

values of

all

The second
the
a

first

a\ftft

=-

pair of equations

pair x

- x

Sr
-,,

and a

=-

Of course
case, since

and

it is

=
when m

these x

xSap) +

ft

and

p is not generally in that plane.

+ Sa ft Sa

(Sa ftSap

a Vftft

jS

and

SajS

by assuming a
it

+X==r,

= Sa

= \r

ft

and

fyn",

is zero,

m"<p*p

indeterminate

if

a;

=+ m =

0,

-f

xSa p)

or

in addition

o-

m =

^in"

If

mp =
+

m"<pff

0,

0.

=-

x ( aft + a

ft

- -

\ftft I

\
.

easy to see that the general solutions are

is

and that Voa

2

the usual solution

<pp

and a - VrjS, and then from

ft

\ftft

3
<f>

Hence,

\ftft

From

So

satisfied

Srftft

Sa p

ft

p,

Sa.fi

is

ft ft

Saft

and

ft

/>,

<j>

x vanishes, the function becomes monomial.

4/<r

=
of the equation
= 0, the solution is
(j>p

<pp

the solution

is

m"p

ff

0- 0.]

<r

is

nugatory.

mp=

m"<r

<f><r

In this

]
</>"

0,

ELEMENTS OF QUATERNIONS.

510

which give, by elimination of x

VIII. ..(c +
which

by

Ci

real,

is

and

<?

Pl

p2

in like

and

write

Pl

(w"

I.

/3,

j3

Sa /3 = */3 +

/3

and IV.

it

j3

In

c z j3

)3, /3

Va V]3

/3

Ya

/3

Y/3

the expression IX. gives

fact,

or

1/

0,

thus,

= -

0)3

c,)/3

r2

/3

= -

0/3

2 j3

= -

0/3

wV ^
1

be thus written,

may

d|3

0/3

)]3

(m"

as above,

Denoting then,

0/3

==

0,|3

= -

0)3

zV/3,

gives,

(2.)

We may

XIII.

XIV.
and

PV=

shall then

(3.)

/3 (c 2

XVI.

we have
2

= -

or

c^p,

2j

o2

0.

Sa/3)

jSSa/^
r

j3Saj3

= cfi + YaY/3/3

+ VaV]3^ = /

c 2 j3

j3

II

pi

$$

\\

p2

have the equations,

J3

2 j3

||

frp\

0,

will be the

new

so will furnish no

Since

p( Cl + S/3)

but the directions of p\ and p

and

m fi

c 2 0/3

also write,

p\

XV.

it

Saj3

j3

manner, the expression X.

=
= + c2
,

0]0 2

pa,

= - dp!,
^0/3 + mfi

we may

jS

+ Sa

|3(f 2

0/>i

XII.

=
)

supposed for the present to be

I.,

j3(d + Sa

satisfy the equations III.

XI.

/3

6.

c,

these two real directions for p, in the plane II of

X.

because

+ Sa

(c

the roots of that quadratic

8,

IX.

and

Sa/3)

easily seen to be only another form of

we have

which

the following quadratic in

x,

[III. n.

same by VIII. as those of

solution of the

p^

and

just resolved.

problem

thus,
||

Pl

0r

\\

follows that the operation

0,

XVI.

and

reduces every line

0,/B

II

^//e

0i/3

II

P,

II

0^0,

fixed plane of

0p

to

the fated direction of 01^0; and that, in like manner, the operation 0i reduces
~
every line, in the same fixed plane of 0jO, to the other fixed direction of 2 0.
]

(4.)

Hence we may

write the symbolic equations,

XVII.
in which the points

XVIII.

may

0!

be omitted

=
0i0 2

0!

(0

0,

and in
C,)

(0

fact
Co)

0,0

we have the

W"0

so that

0102

0l

= ^0 =

0,

0.

transformations,
Wl

=
1/>,

CASE OF UNEQUAL EEAL ROOTS.

ART. 353.]

we propose

If

(5.)

XI.) by which

i//

XIX.
and

therefore,

XX.
and in

like

0V0V = V. (0V

+ dXP +

\l>p

+ dp) (0

XXI-

or

zero,

^ =$ +

and

if
d>

2X +

+ d2

<?,x

XXIII.

0il//i

vanishes, if

XXI.

+ d)

(0

cz\

be arbitrary scalar 8.

with 0u or symbolically multiply the expression

the symbolic product,
(<

what the

is

scalar

m$Xv

by the

= S

m
iX

when

becomes,
ii

= S

iv

operate on

by

i//x

we

0i,

T//I/O

get

Ci )

+ ^(0 + x) +

changed

or

to

^i

is

such that

</>i,

X +

definitions of

is

<

we

for

CiX +
<M

as appears

= ^ +

^i

(6.) Accordingly, without assuming that

XXIV.

+ dv),

manner,

be different from

where Wi

rule (347,

VIII. of x

XXII.
even if

by the same general

0i,

we have

definition 350,

fap

from

0,

= V.

t//iVjuv

by the

form

to

formed from

is

511

^, x?

w/
>

by the

an(l

Wi//
>

>

relations

between those symbols which have been established in recent Articles, or

appended to them.
= 0, and that d c* are the roots of the
(7.) Supposing now again that m
quadratic I. in c, we have by XXIII.,

in the sub-articles

XXV.
if

0i^i

be formed from

Comparing

(8.)

=
;

m by
XXV.
it

and

in like

XXVI.

manner

2
i/>

changing d

to c z

= nh =

0,

XXVII., we may

with

existence of an intimate connexion existing between

reduces an arbitrary vector to the fixed direction of 0f

these

be led to suspect the

and 2 0, since each
t//i
J

0,

or of pi

and

in fact

two operations are identical, because, by XXI., and by the known

between the symbols, we have the transformations,

relations

xxvu.

^ = ^ + d\
=

(?//

Ci)0

K=

^"0

C2

+ 2) +
= 00 2

and similarly

XXVIII.
=

while
<//

0i(/) 3 ,

as before,

^=

0i0

00i

c i( w

*"

0) + c

2
i

ELEMENTS OF QUATERNIONS.

512

We have

(9.)

which the

in

6.

thus the new symbolic equation,

XXIX.

00!0 2

0,

2 may be in any manner grouped

two equations XVII. and in which

three symbolic factors, 0, 0i,

transposed, so that it includes the

and

[III. n.

the subject of operation

is

an arbitrary

but we

vector p.

been

Its interpretation has

add, that while

may
already partly given
the fixed plane D, 0i reduces every line to another fixed plane, IIi, and
2
reduces to a third plane, II 2 thus 0i0 2 or 2 0i, while it destroys two lines p lt
p 2 and therefore every line in the plane n, reduces an arbitrary line to the
;

reduces every vector to

which intersection

fixed direction of the intersection of the tico planes IIiIIo,

must thus have the

direction of 0~ 1

and

in like

manner, the fixed direction

being that to which an arbitrary vector is reduced (3.) by the

0r
compound operation 2 0, or 00 2 must be that of the intersection of the planes
nn 2 and p or 2~ 0, has the direction of the intersection of nil! while on
1

pi of

0, as

2>

hand 00 2

the other

and 00! every

destroys every line in HI,

line in II 2

so

that these three planes,

with their three lines of intersection, are the chief

elements in the geometrical interpretation of the equation 00!0 2 = 0.

The

(10.)

conjugate equation,

xxx.
may

o,

be interpreted in a similar way, and so conducts to the consideration of

a conjugate system of planes and

are the

loci

of
(f>

p,

a^o,

p,

lines

namely the planes

while the operations

i0

II
2

2,

II

i,

and

i,

which

2,

destroy

three planes respectively, and reduce arbitrary lines to the

fixed directions of the intersections, U\U\, IT 2 rr, II II i, which are also those
all lines in these

-1

of

0V

0,

0,

It is

(11.)

0V1Q.
important to observe that these three

to the three first planes, II, II

lines

II"

and

last lines are

the normals

that, in like manner, the three former

To prove

are perpendicular to the three latter planes.

this, it is sufficient

to observe that

XXXI.
and

Sp fp = S p 0y =
1

similarly,

_L 0~ 0,

0,

if

0V =

or that

0,

(0
vector
c\)

any
form xQrp

&c.*

* [More

+
(<t>

J_

0/o

satisfies the symbolical cubic

sj^mmetrically, without assuming one root to be zero, if
+ cs) = 0, it is easy to show that pi, the result of operating by
+ c z ] ($ + c a ) on
cz)
For a second arbitrary vector a may be expressed in the
p, is parallel to a fixed direction.
<

+
-f

(<>

yq>p

symbolical cubic

+
(<j>

(<f>

zp,

and

so

ci)pi

(<p

0) is

c->)

(<p

likewise

C3)a
||

pi.

= X(^p\ +

yfypi

Thus the

zp\

operators

^i 2

(<p

cz)

yc\
(<p

z )pi (since

<?a),

(<j>

s)

by the

(<p

C
0>

TRIADS OF LINES AND PLANES.

AKT. 353.]

Instead of eliminating

(12.)

might have eliminated

XXXII.
also, if

x\

x and #

XXXIV.

XiXz

x z be the two values of x

^Sa /B + ^(Sa jS -

(x,x\

Pl

xi$ + ^]3

||

+ x 2 x\]

hence the condition of rectangularly of the two

0"Sa/3

x 9 then

Pz

this other quadratic,

Sa]3)

II

x\x\ = - Saj3

we

x between the two equations VII.,

which would have given

XXXIII.

and

of

513

y j3

x$

- Sa

(Sa/3

lines

p l9 p 2

j3

Sa

j3

^O, ^O,

or

is

expressed by the equation,

XXXV.

=-

/3

Saj3 + Sj3)3

(Sflj3

- Sa

jS

|3

Sa

j3

and consequently it is satisfied, if the given function

because we have then the relation,

XXXVI.
in fact the binomial form of

XXXVII.

$p

jp

.Vj3a +

(aS/fy

jSSa/o)

/3j3

V(/3

+ /3V)

be self-conjugate (VI.),

<

Vj3V=0;

gives (comp. 349,

<

=S

(a

S/3>

XXII.),
- Sa
= V. pVtfa +
p)
/3
/3V),

which cannot vanish independently of p unless the constants satisfy the con
9

dition

XXXVI.
With

(13.)

this condition then, of self- conjugation of $,

we have

the relation

of red angularity,

XXXVIII.

at least if these directions pi

The

to be, as follows.

XXXIX.

Spipz

and p z be

condition

^0

or

real,

which they can easily be proved

XXXVI.

= S aa V(|3a + j3V) =

^0

0,*

JL

gives,

a Sa|3 +

Saa^Sa

/S

Saj3)

- a /2
Sa/3

+ cz) reduce any vector to lines parallel respectively to three fixed directions pi, p?,
ci)
4- ci)p is a general expression for a
Further, by the property of the conjugate function
In the same way
vector perpendicular to pi.
+ c%)
+ c$) p is perpendicular to pz and also to ^3
and p z and p s similarly found and
and parallel to a fixed direction p i which satisfies
+ ci)p \ =
and
and

(</>

(</>

ps.

(<

(<

+ c^p

and

(<J/

(</>

are at right angles respectively to the planes of p 3) p lt and

Taking unit vectors through a common origin and parallel to these fixed vectors,
Up2,
and Ups determine a triangle on the unit sphere and Up i, Up g, and Up a are the vectors to the vertices
of the supplemental triangle. Again if y is the spin-vector defined in 349 (4.),
+ ci)y or its equal
satisfying
of pi, pz-

(<p

(<p

c 3 )p 3

U/>i,

TJ(<

U(cp

by

ci)y terminates at the pole of the great circle

UV7(/>7

is

the

common

through Upi and

orthocentre of the two triangles.

When

Up

i,

and the point determined

the function

the two supplemental triangles coincide, and consequently the solutions of Vp(pop

is

self-conjugate,
are mutually

perpendicular (16.).]
*
- cfiSpipz. So if pi is perpendicular
- )pipz =
[In general by 349 (4.), 2S7pip 2 =
to pa, 7, if it does not vanish, lies in their plane.
Conversely, if 7 lies in the plane of pi and pi,
<t>

S(<p

either Spipa

HAMILTON

0,

or

c\

(<b

=
GZ."]

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

514
8

hence

(a

aV

Sa

- a /2
Saj3

/3

(aV -

(Saa

and

XL.

(Saj3

aV

(Sa|3

- Sa
jS

{(Sa/3

]3

- Sa

(Sa/3

- Sa

- Sa

4m

= (Saa

2
)

(a

+ 4Saj3 Sa

|3

+ 4Sa/3 Sa

Sa

/3

/3

2
w"

j3)

/2

/3

6.

[Ill.n.

a Sa|3

4m"

2
>

0,

>

two quadratics, I. (or VIII. ), and XXXII., has real and

a conclusion which may also be otherwise derived, from the
= aa + ba, )3 = ba + a a, which the condition allows us to

so that each of the

unequal roots
expressions

/3

substitute for

/3

and

r
.

j3

(14.) The same condition XXXVI. shows

complanar, or that we have the relations,

XLI.
hence

when

Vaa

or

^O

the function

Saj3/3

8a

now normal

is

P p l9
,

compose a real and rectangular

In the present

]3Sa/>,

when

/o 2 ,

or

an arbitrary
which

is

^0, 0f

/3)

=
;

and therefore by

(13.),

0,

0,

(to 353),

we suppose

that the

are all unequal, the cases of equal roots (with

series (352)

a self-conjugate function

we must have

reduces

.V/3

are

si/stem.

the relation V/3a =

the fixed direction of which (comp. 352,

line in

V(Vaa

to the plane

having been discussed in a preceding

form

0,

series of sub-articles

three roots of the cubic in

in passing, that

/3

self-conjugate (VI.), the three directions,

is

XLII.

(15.)

0,

that the four vectors a]3a

(5.)

m=

0)

may be remarked,
reducible to the monomial
but

is

it

<f>p

and that thus the

and

(6.)

line /B, to

the operation

then

vector, is perpendicular to the fixed plane (352, (7.)), every

destroyed

by that operation

0.

In genera], if
be thus self -con jugate, it is evident that the three
n
IT
which
are (comp. (10.)) the loci of
2
FT,
i,
planes
$ z p, coincide
p,
with the planes n, Ed, Efo, which are the loci of 0p, 0ip,
o.
(16.)

<t>\p,

2J

(17.)

equal,

it

When is not self-conjugate, so that

and p are not generally
has been remarked that the scalar quadratic L, and therefore also the
<ftp

symbolical cubic in 0,

may have imaginary

roots

and

<j>

that, in this case, the

vector equation II. of the second degree cannot be satisfied

of p, except that one

which

by any

real direction

itself
the linear equation V., or causes
to vanish, while p remains real and actual.
As an example of such imaginary

scalars } as roofs of I.,

satisfies

and of what may be

<pp

called imaginary directions, or

CASE OF IMAGINARY ROOTS AND DIRECTIONS.

ABT.353.]

imaginary vectors (comp. 214, (4.)

themselves imaginary roots of II.,

which correspond

),

we may

515

to those scalars,

and are

take the very simple expressions

(comp. 349, XII.),

XLIIL
y denotes some

in which

Vy/o,

(j>p

real

and given

<j>

Vy/o

and which evidently do not

vector,

satisfy the condition VI., the function

conjugate, so that its self-conjugate part

p =

being here the negative of

is zero

<

its

own

We

(comp. 349, XIII.).

have thus,

XLIV.

m, =

0,

and consequently, by the sub-articles

XLV.

the quadratic

/-

0,

and

I.,

XLVI.
where

c~

1 is the

= - y2

roots c h cz

0,

to

in"

its

"

0,

0,

0,

Xo

= -

= -

\pp

ySy/>,

0,

\f>

Vy/o

therefore,

- + v/^T. Ty,

ci

fa

0,

349 and 350,

become

0,

= - v/ 1

c2

imaginary of algebra (comp. 214,

"!.

(3.)

Ty,

thus by

XX.

or

XXI., and XXII. we have now

XLVII.

= -

faff

ySyo-

hence

XLVIII.
and in
if

CiVya +

like

01^

(0

0,

XLVIII

then we take an arbitrary vector

from
fay

it

we

two (imaginary] vectors p l

shall

L.

as ones

nary

(y

7
.

o-,

Cl )

CL )

faff

Vy<r,

= (y -

C z ) Vycr

y^itr, &c.,

(y

d)Vy<r

^^o- =

and derive

= (y 2 cf)Vya =

0,

(or rather conceive as derived)

and p z by the (imaginary)

operations fa

and

have (comp. III. and IV.) the equations,

XLIX.
and

Vy(//i(T

+ cj^a =

manner

= (y -

<?iV

which are

directions,

p2

pi=
=

fa(T,

faff,

<jnp!

<^

2J

o2

0,

0,

<jo

= <j>pt

= -

at least symbolically true.

p and p z
}

=
Ci/Oi,

c2 /o 2 ,

V/Oi0/>i

V|0 2 0j0 3

We find then

==

0,

0,

that the two imagi

satisfy (at least in a symbolical sense, or as far as

concerned) the vector equation II., or that pi and p z are two

but that, because the scalar quadratic I.
imaginary vector roots of "VpQp =
has here imaginary roots, this vector
equation II. has (as above stated) no real
calculation is

vector root p,
except
satisfies

one in the direction of the given and real vector y, which

=
the linear equation V., or
gives Qp

0.

8 U 2

ELEMENTS OF QUATEENIOKS.

516

[III. n.

6.

This particular example might have been more simply treated, by a

We wish to satisfy the equation,
general method, as follows.

(18.)
less

LI.

which

when we

gives,

operate on

LIL
if

Y. p^Vyp = p$7p - p 2 7

by V. 7 and V.

it

Yrp

p, these others,

= p2
Y7p

S 7 p,

then we wish to avoid supposing 0p =

Yyp =

we must

0,

seek to satisfy the

two scalar equations,

LIIL
and conversely,
have

if

we can

S 7p =

2
p =

0,

satisfy these

by any

equation LIII.

LIY.
where o

is

when we assume

is satisfied,

p =

(c

the expression,

+ y)Vy(7 =

Yyd

an arbitrary vector, and c is any

and this expression LIY. for

laws of scalars

imaginary) p, we shall
equation LI. Now the first

(real or

satisfied (really or symbolically) the vector

(c

y),

scalar, or

p,

with

symbol subject to the

transformation just

its

assigned, gives

LY.
the quadratic
roots,

2
.

XLYI.

and imaginary

(19.)

p
is

(c

(Y 7

<r)

0,

therefore reproduced,

if

2
7 =

and we have the same imaginary

directions, as before.

Geometrically, the imaginary character of

the recent problem, of

pY7p = by any direction of p except that of the

line
or Y7p, is here a
given
7, is apparent from the circumstance that
vector perpendicular to
if
both
be
actual
and
that
the one
therefore
lines;
p,
satisfying the equation Y.

<p,

cannot be also parallel to the other, so

long as both are real*
354. In the three
preceding Articles, and in the sub-articles annexed, we
have supposed
is wanting,
throughout that the absolute term of the cubic in
or that the condition m =
is satisfied
in which case we have seen
(351)

Accordingly the two imaginary directions, above found for

seen to be those which

p, are easily
drawn in a given plane (perpendicular here to the
given line 7), to the circular points at infinity, of which
supposed directions the imaginary character
may be said to be precisely this, that each is (in the given plane) its own perpendicular.
in

modern geometry are

called the directions of lines

[As additional examples

= qpgr*, it is obvious that
l
is reciprocal, and it
This shows that the cubic of
$ = (jr
- 1) ($* - 2 cos 2w +
may easily be reduced to
= if u = L q. The real direction is \q, and
1)
the imaginary directions are
the lines to the circular points at infinity in the plane perpendicular to
- 1 = 0. The directions
\q. Again, if
into 7, and 7 into a, the cubic is
changes a into $,
are a +
+ a, 3 7, where is an algebraic cube root of unity.]
:

If

<pp

</>

(<j>

<J>

</>

<^

a>

<

VECTOR AND QUADRATIC EQUATION.

AKTS. 353, 354.]

517

always possible to satisfy the linear equation $p - 0, by at least one

real and actual value of p (with an arbitrary scalar coefficient)
or by at
least one real direction.
It will be easy now to show, that although con
that

it is

versely (comp. 351,

p, unless

(4.) )

we have thus

which the

the function
<j>p

yet there

0,

is

cannot vanish for any actual vector

always at least one real direction for

vector equation of the second degree,

I.

Vpfto

0,

which has already been considered (353) in combination with the condition
m = 0, is satisfied and that if the function be a self-conjugate one, then
;

this equation I. is always satisfied

by

at least three real

but not generally by more directions than three

tions,

and rectangular

direc

although, in this case

namely when

of self-conjugation,

II.

0V

IF.

or

and

for all values of the vectors p

A, the

and

true, for one real direction of p,

S\(j)p

<t>p,

&p<j>\,

may happen

I.

equation

to

become

for every direction perpendicular thereto

or even for all possible directions, according to the particular system of

constants,

show

also

We

which enter into the composition of the function $p.

that the scalar (or algebraic) and cubic equation,
III.

mc

2
m"c

shall

c*,

which is formed from the symbolic and cubic equation 350, I., by changing $
- c, enters
importantly into this whole theory and that if it have one real
and two imaginary roots, the quadratic and vector equation I. is satisfied by
to

only one real direction of p

be satisfied
roots

also

by

three right lines,

be formed

of any linear
<f>p

and

which thus

Op =

where

c is

to

may

Q
proved to be always real so that if m Q m\, and
349 and 350) from the self-conjugate part $,p
m"

vector function

itself,

then the new

IV.

then be said (comp. 353, (17.)

may

(as in sub- articles to

function

(1.) If

it

be said to represent generally a system of

For the case II., the
whereof one at least must be real.

so that this equation I.

scalar roots of III. will be

V.

but that

two imaginary directions, or to have two imaginary and vector

results, can never have

we

and

m"

formed from that

are

+ m\c +

imaginary

m \& +

3
,

roots.

write,

+
<j>p

as m,

<j>p,

cubic,

cp,

<$

an arbitrary

=
<}>

p+

scalar,

or briefly,

cp,

and

if

V.

we denote by

<

<

and

<

c,

what

c,

i/, -\j/ t

ELEMENTS OF QUATERNIONS.

518

and

+ c or
to
become, by this change of
353, (5.), (6.), show that we have the expressions,
t

VI.

= $ + c x + c\

and
VII.

with

VIII.

Hence

(2.)

may

it

=
+

m"c*

6.

the calculations in

<,

+ GX +

<//

M= m + m

[III.

z
,

3
,

M=

be inferred that the functions x x

an(* the constants

>

become,

m"

IX.

D*

= x +

2lf

lJr

D c Jf =

iD

2c,

D^

+ 3c

2m" c

Jf=m"

2c,

-t-

2
,

3tf;

with the verifications,

XI.
as

we

<P

X =

X = IT,

4>

4>X

4>

X +

= Jf

had, by the sub-articles to 350,

+ X =
(3.)

The

wew?

+ x =

^wear symbol

XII.
which accordingly can be

<P

w// ^

*X + ^ =

must

satisfy the

= Jf - -a/^ + Jf

fX
O

new
2

from the

at once derived

^=m

cubic,

3
;

old cubic 350, I.,

under

the form,

XIII.
(4.)

Now

it is

by substituting

m"(c

2
$) +

(c

XII.
=

and conducts

- Jf

^ + JfO

and thereby

to

2
>

to analogous
consequences,

352, II.,

w>,

which need not here be developed

in detail, since they can

easily be supplied by anyone
trouble to read again the few recent series of sub-articles.

to the form,

f>

case

precisely similar to the form,

(5.)

3
3>)

for c a real root of the scalar cubic III.

XV.
is

+
<J>)

always possible to satisfy the condition,

reduce the new symbolical cubic

which

m (c -

who

will take the

For example, unless it happen that

constantly vanishes, in which
= 0, and
which is
(if not identically null) takes a monomial form,
>

<&p

CASE IN WHICH NO BOOT

AET. 354.]

reduced to zero (comp. 352,

(7.)

IS ZERO.

519

for every direction of p in a given plane,

reduces (comp. 351) an arbitrary vector to a given direction

the operation
direction
so that, in every
and the operation
destroys every line in
;

<1>

there

case,

is

one real

/<?tfs

and therefore

also

of satisfying the vector equation

way

above asserted) the equation

(as

And

(6.)

since that equation I.

XYI.
see that

may

roots,

XVIII.
will generally

<Pp

two real and

distinct

while

0,

and these are

<

o2
2J

but

the cubic

The

satisfying

for if

=
3>!

have for

<E>

IT C + W,

we then

19

Ci,

a given plane, and there will be

its locus

one of which

for

for the other, so that each satisfies

3>

to

3>

and

I.,

write,

and p 2 in that plane,

<I>

2 jo,

XVI.,
and

if

or

I.

be

respectively.

of imaginary roots need not be dwelt on here

gp,

then

XX.

(g

the function $p have the par

if

0)

0,

XXI.

and

or III. having thus all its roots equal,

by

M=

(g +

cf

and the equation

I.

every direction of p, in this particular case.

general existence of a real

I.,

comp. 353,

being any scalar constant),

XIV.

satisfied

(8.)

C->

2
<7

new scalar and

vanishing, if this

be remarked in passing, that

<j>p

being

p,

||

4>p

precisely the fixed directions of Wip and

ticular form (g

XIX.

d,

by changing
Cases of equal and

may

it

directions pi

formed from
(7.)

or

0,

<&p

have real and unequal

<J>!pi

0,

without causing p

I.,

be thus written,

can be satisfied without

it

XYII.

the line

VpQp

quadratic equation,

Qp =

vanish.

itself to

we

when

the

condition

and rectangular system

II.

is

be proved as in

may

satisfied,

of three directions

and it is unnecessary to dwell on the case where, by two roots

353, (14.)
of the cubic becoming equal, all lines in a given plane, and /so-the normal
;

to that plane, are vector roots of I., with the

(9.)

And

because the quadratic,

same condition

m"c

II.

(353, I.), has been

=
proved to have always real roots (353, (13.)) when
p
must
have
XVII.
likewise
then
real
roots, (7i,
quadratic
<j>

the analogous

<j>p,

<7

whence

it

immediately follows (comp. XII. and XIII.), that (under the same con
dition of self-conjugation) the cubic III. has three real roots, c, c + Ci, c + C2

and therefore that

(as

above stated) the other cubic IV., which

is

formed

ELEMENTS OP QUATERNIONS.

520

from the
which

may

we denote

(10.) If

- S
7

<IVy

has

0,

manner by

in like

XXVI., comp.

(349,

whence, by comparing powers of

m
On

m\ - y

which ^ and v are arbitrary

-S 7

a>o

as in 350, (1.).*

m^ V/zv = V.i/^y

becomes,

J/^ V/iv = V.^u^v,

vectors,

the equation

XXI.) becomes,

m\

m"

c9

recover the relations,

a similar plan, the equation

.

the symbol

<J>

349,

we

c,

and

XXIV.
in

and

roots always real.

its

XXIII... Jf=Jf

(11.)

vector function 0,

6.

on that account be thus denoted,

XXII.

and

of the general linear

part

self-conjugate

[III. n.

and

c is

cornp. 348, (1.),

an arbitrary

scalar

or

more

fully,

XXV.

3
(m + m c + m"& + c )

+ c) V/*v = V.

(</>

(^ + GXH + c

/*)

(^v + c^y +

c*v)

whence follow these new equations,

XXVI.
XXVII.

(m +

XXVIII.
which can
to
</>

i//,

m
wY

(m +

<t>

(w"

)Vnv =

V(^

=
) YJUV

VGu^>

$v

vi/^

X^)>

+ XM

X")

=
)V,uv
V(^ X v VXM),

be otherwise proved, and from the last of which (by changing

&c.) we can infer this other of the same kind,
all

XXIX.
(12.)

As an

example of the existence of a rm^ and rectangular system of

ree directions (8.),

represented jointly by an equation of the form I., and
of a system of three real roots of the scalar cubic
III., when the condition II.
is satisfied, let

us take the form

XXX.
g being here any real and
[If

Qpi

From

this, c\ is

=
<opi

this cubic

#^w

VT/JI

=
gp + VApju
$

(f)p

=-

scalar,

cipi t

and

then

pi

A,

p,

^ any

real

and non-parallel

=-

a root of

(mo

and

must be

identical

- 87^7) + (w - 7 2 + w ^c 2 + c 3 = 0,
z
with m -f ;w +
+ c 3 = 0, as they have
)tf

<?

m"c

three roots

common.]

CASE OF SELF- CONJUGATION.

AET. 354.]
vectors

given

to

con jugate function

XXXI.

521

which form, indeed, we shall soon find that every selfo can be brought.
We have now (after some reductions),
oj

\f,p

XXXII.

Xp =
(\Sw +

+ 2gp

fjiSXp)

and

XXXIII. ..m = (g-

f
2 SA^u)^
XV), m = = - SA^u + 3g
m"

where the part of $p which

independent of g

is

Ay be put under several

may

other forms, such as the following,

XXXI Y.
=

and

<1>,

X(p$X/u +

changing g to

Y(AjO/^SAfi

SAjUjo)jU

X/LLSXp/j.)

= ^X(X/mp +

Xp/mSX/uL

=
pX/bLjfj.

-f

X(X$fjip

/zSAjo

Xpfj)in,

&G.

be formed from 0, ;//, ^, w, ;w m

by simply
= has therefore here three real and
The equation
/r

M\

X, Jf,

^P,

M."

g.

may

unequal roots, namely the three following

XXXY.

= - g +

and the corresponding forms

XXXVI.
Thus ^p,

Vp =

and

c+ C = -g +

SX/i,

of

Vp

VX/zSXjip,

TXju,

+ ft = -

TX/i;

are found to be,

,/<>

= - (XT/i

have in fact the three fixed and rectangular directions

of YA^u, AT^u + /zTA, and XT/a
juTA, namely of the normal to the given
of
and
the
bisectors
of the angles made by those two given
A,
^u,
plane
^ijo,

and

^o/o

accordingly the only directions which satisfy the vector

equation of the second degree,
lines

these are

XXXYII.

.
(Vp<j>p

= Y. pYA/o^u

so that this last equation represents (as

=) VpXS/i/o

+ V/o/iSAp =

was expected) a system of

three right

lines, in these three respective directions.

(13.)

In

general, if

M=

the cubic equation

and

0,

XXXVIII.

HAMILTON

.
,

if

we

$1 - $ +

the corresponding values of

XXXIX.

denote the three roots (real or imaginary) of

Ci, c 2 , c 3

dx +

write,
c l9

*I>

<

Cz,

<t>

c3 ,

will be (comp. VI.),

2

Cx

ELEMENTS OF QUATERNIONS.

= $ + c zX +

c2

2
,

= ^ +

c,

x+

c3
3

2
;

ELEMENTS OF QUATERNIONS.

522
also

we have the

6.

relations,

XL.
whence

[III.n.

easy to infer the expressions,

it is

XLI.

^=

Ca)-

(C,

$3 =

(3

CO

(C,

2 ),
1

(C*

<*>*

(2

1
Ci)"

(1

3),

~
1)

as binomials (conap.
which enable us to express the functions fyp,
3p
o,
351, &c.), when "^p, ^ 2 jo, ^? 3p have been expressed as monomes, and to assign
the planes (real or imaginary), which are the loci of the lines
$2/0, ^3/0.
3>

<&

2/

3>ip,

by which lines in the

Accordingly, the Mree operations,
three lately determined directions (12.) are destroyed, or reduced to zero and
(14.)

<J>,

4>i,

4>,,

which at

first

XLIL

present themselves under the forms,

a>9

are found to admit of the transformations,

where

Qp
z

"^i,

^P"

XXXVI., and

have the recent forms

the

loci

compose a system of three rectangular planes.

(15.) In general, the relations (13.) give also (comp. 353,

XLIV.

<J>

$3,

= ^3^1,

*3 = Ol0 2

of

4>/o,

<Pi|0,

(8.)),

and

XLV.
whence

$!-*,

4>.^ 2

= ^3^3 = *i$ a * 3 =

0,

also,

XLVL
the symbols

(in

^i^

any one system

= ^2^3 - ^Y^i =

0,

of this sort) admitting of being transposed

=
at pleasure
if then the roots of
be real and unequal,
there arises a system of three real and distinct planes, which are connected

and grouped

= 0, exactly as the
with the interpretation of the symbolical equation,
3
= 0.
three planes in 353, (9.) were connected with the analogous equation
2
said
then
be
(16.) And when the cubic has two imaginary roots, it may
^>!^>

<I>3

00i<

that there

one real plane (such as the plane _L 7 in 353, (18.), (19.)),

the
two imaginary directions which then satisfy the equation I.
containing
and two imaginary planes, which respectively contain those two directions,
is

and

intersect

cited),

each other in one real

namely the one

line

(such as the line y in the example

real vector root of the

same equation

I.

REAL AND RECTANGULAR SYSTEM.

ARTS. 354, 355.]

355.

Some

additional light

be thrown upon that vector equation of

may

by considering the system

the second degree,

I.

SX/x^ =

523

and

0,

II.

of the two scalar equations,

.

S\p =

0,

and investigating the condition of the reality of the two* directions, p and p z
by which they are generally satisfied, and for each of which the plane of p
l

and
<j)p

contains generally the given

We

line

X in

I.,

or

normal

is

to the

plane

two directions are always real and

that
become
rectangular (except
they may
indeterminate] when the linear
function
is its own conjugate
and that then, if A be a root p Q of the vector
locus II. of p.

shall find that these

equation,

III.

V/o^ =

0,

lines p and p z are also roots

the general existence (354) of a system of three real and

which has been already otherwise discussed, the

of that equation

rectangular directions,

thus proved anew

which

whence

scalar roots of the

cubic

when

satisfy this equation III.

also will follow a

M=

new proof of the

formed (354, IV. or XXII.) from the self-conjugate part

=
linear and vector function 0, as
was formed from

is

(1.)

all respects

IY.

X,

fj,,

yfj.

the equation II.

zv,

and

is satisfied,

V.

$
0,

and

which

of the general

<

<.

\p = yv

therefore,

and

I.

= f$v$/j. +

k.

(Sv^v

Writing then,

z/n,

=
<j>p

y^ + z$v,

becomes,

yz(Sv(j>v

S/x0/z)

the roots of which quadratic will be real and unequal,

VI.

v be a system of three rectangular vector units, following in

the laws (182, 183), of the symbols i,j,

p =

reality of the

0, for this case of self-conjugation of

therefore of the necessary reality of the roots of that other cubic,

Let

- $p, being

fy

S/x^) +

4S^vSy^

if

>

*
Geometrically, the equation I. represents a cone of the second order, with A for one side, and with
the three lines p which satisfy III. for three other sides and II. represents a plane through the vertex,
;

The two directions sought are thus the two sides, in which this plane
perpendicular to the side A.
= 1 where the
cuts the cone.
[The general equation of a quadric mp,y be written in the form
function
is self -con jugate.
The cone, through its intersection with a concentric sphere, is
+ r~ 2 )p = if r is the radius of the sphere. If this touches the plane 8\p = 0, it is geometrically
S/j0/>

<$>

Sp((j>

evident that the edge of contact is a principal axis of the plane section of the quadric as it passes
through the points of contact of the concentric sections of the quadric and the sphere. The condition
for contact is A.
+ r~ z )p, or SA/J^/J = 0, coupled with SA/J = 0. The directions of the principal
||

(<

axes thus determined are always real whether the plane cuts the quadric in a real curve or not.]

3X2

ELEMENTS OF QUATERNIONS.

524

and the corresponding directions

VII.
that

[III.

of p will be rectangular, if

is, if

VIII.

Sy0ju

at least for this particular pair of vectors,

the expression,
(2.)
conditions VI. and VIII. take the forms,

IX.

(Sv^ov

which are both

<j>p

v.

fap +

SufafiY + 4S(ju0

satisfied generally

and
=

/u

Introducing now

>

i>)*

4(Sy^v)

and

coefficients

X.

imaginary,
(3.)

when

On

condition

is

and when the two

when y

line

identity, by all
VI. and IX.) can

is actual,

or
<j>

p not generally = 0p, the

y happening
to

namely by

be in the assumed plane of

to

directions of p are thus not rectangular, or

Sypv does not vanish, we have only

becomes

the only ex

of rectangularity can only accidentally be satisfied,

the given or fixed

thus self-conjugate.

the other hand,

X.

(to

Sy^uv

the roots of that quadratic can never be

to say,

is

=
when y = 0, or =
to
become an
may happen

vanishing: but the opposite inequality

never hold good, that

(349, XII.), the

Vyp

ception being, that the quadratic V.

its

6.

ir.

when

/u,

the scalar

suppose that the square of this scalar

large enough, in order to render (by IX.) those directions coincident,

or imaginary.
(4.)

When

directions of p, or

y=

0, or

(5.)

shall

And

if,

/m

two rectangular
the very simple form yz =

and

v for the

we must
XT.

we

take

reduce the quadratic to

establish the relations,

may

but, for this purpose,

we may

0,

at the

S<v

same time, X

Svu

satisfies

0.

the equation III., so that

0A

||

A,

have these other scalar equations,

XII.

whence

W*

VvA

II

||

p,

and

0v

||

VA/*

v,

||

or,

XIII.

v thus forming (as above stated) a system, of

roots, of that vector equation III.
A,

ju,

(6.)

But

in general,

if

directions of p, the scalar

III. be satisfied

and

cubic equation

^/?rce

by even

real

tico

and rectangular
real

and

distinct

can have no imaginary

AUTS. 355, 356.]

root
GI

NEW PROOF

for the quotient

c zt

525

two directions give two unequal but real and scalar values,
p, then c\ and c z are two real roots of the cubic,
fyp

for if those

and

OF EXISTENCE OF THE SYSTEM.

of which therefore the third root

is also

and

real;

if,

on the other hand, the

^ and p z give one common real and scalar value, such as d, for
= - Ci/o, or ^p = ($ + CI)/Q = 0, for every line in the plane
that quotient, then
of pi, p z
so that $p must be of the form,
Cip + fl$p p2p, and the cubic will
have at least two equal roots, since it will take the form,
two directions

</o

xiv.
as is easily

shown from

o =

(c

Cl

y (c-

Cl

+ s^jS),

and formulae already

=

principles

if

/o

<pp

and therefore that the equation

established.

(7.) It is then proved anew, that the equation

has

all its roots real,

above stated) can

(as

never have an imaginary root.

(8.)

And we

at the

see,

same time, how the

351,

I.,

actual p
&c.,

M=

scalar cubic

have been deduced from the symbolical cubic 350,

I.,

as the condition for the vector equation III. being satisfied

namely by

and therefore

observing that

Mp

which

0, in

= -

if

cp,

<j>p

p,

by

then

<fp

supposition,

might

or from the equation

is

by any
= - c
p,
3

p,

<jpp

different

from

zero.

(9.) Finally, as regards the case* of inde termination, above alluded to, when
the quadratic V. fails to assign any definite values to y z, or any definite
directions in the given plane to p, this case is evidently distinguished by the
:

condition,

XV.

in combination with the equations

356.

The

XI.

existence of the Symbolic and Cubic Equation (350), which

is

by the linear and vector symbol 0, suggests a Theorem^ of Geometrical

Deformation, which may be thus enunciated
satisfied

by any given Mode, or Law, of Linear Derivation, of the kind above

denoted by the symbol 0, we pass from any assumed Vector p to a Series of
"If,

Successively Derived Vectors, p

pz p3
,

or
<fr

p, tfp, $*p,

structing a Parallelepiped, ^ve decompose

any Line of

three partial or component lines,

m pi,

mp,

m"

this Scries,

and

if,

by con

such as p 3 into
,

in the Directions of the three

p*,

* It will be found that this case

corresponds to the circular sections of a surface of the second order
less particular case in which ty p =
but not S^^yi = Sj^v, so that the two directions of p
are determined, real, and rectangular, corresponds to the axes of a non-circular section of such a surface.
;

while the

<pp,

f This theorem was stated, nearly in the same way, in page 568 of the Lectures and the problem
and vector function was treated, in the few preceding pages (559, &c.), though
with somewhat less of completeness and perhaps of simplicity than in the present Section, and with a
;

of inversion of a linear

ELEMENTS OF QUATERNIONS.

526
which precede

m"

p 3 bear

Mode

then the Three Scalar Coefficients, m,

Three Preceding Lines of the Series, will depend only on the given

Law

of Derivation, and

be entirely independent of the

icill

Length and Direction of the Initial Vector."

(1.) As an Example of such successive Derivation,
I.

pi

= - V/3p7,

=
(f>p

pz

and 7 are unit-lines. Treating them at

general method conducts to the equation,
/3

II.

with the following values of the

III.

may

in 354,
(2.)

= -

XXXI1L,

to 0,

when we suppose

that

as any two given vectors, our

m"p z ,

= -

j3y

m"

= Sj3y

by merely changing

calculation,

and

g, X,

/u,

7.

Supposing next, for comparison with 305, that

/3,

7 are unit

IV.
so that

new

and -

]3,

V/3/ory, &c.,

coefficients,

/3yS/3 7

be seen, without any

first

assumed

us take the law,

let

&c.,

(1.),

m pi

= mp -

/o 3

= -

which answers to the construction in 305,

as

6.

or the Three Ratios which these three Components of the Fourth Line

to the

or

as here of p, pi, p z

it,

[III. n.

)3

7*

and

1,

S/3 7

and /is the cosine

lines,

= -

/,

of their inclination to each

other, the values III. become,

The general form

slightly different notation.

be thus expressed

of such a function

which was there adopted may now

(j)p

the resulting value of

= 20 Sap + Vrp,

was found

to

and the auxiliary function which we

l
m<p-

ff

tyff

r heing a given quaternion

he (page 561),

now

= 2Vaa S#

denote by
<r

+ 2V.

\J>

was,

aV(Yj8<r.

4-

(VorSr

where the sum of the two last terms of

might have been written as orSr rSor. A student might
find it an useful exercise, to prove the correctness of these
expressions by the principles of the present
Section.
One way of doing so would be, to treat 2$Sa/j and r as respectively equal to
+ \yp and
c + e
which would transform m and $&, as above written, into the following,
i|/<r

<j>op

Mo - 8(7 +
that

is,

into the

new

e)

(0o

c]

(y

and

M and

0-

e),

values which the

<r

(7

e)

8(7 +

e)<r

of the Section assume,

Vcr(<f>

when $p

c)

(7

takes the

e)

new value,

THEOBEM OF SUCCESSIVELY DERIVED

ARTS. 356, 357.]

and the equation

II.,

connecting four successive

form,

YI.

p2

lp

pi

YII.

or

fa

we

a result which agrees with 305, (2.), since

was = -

the interval PiP 3

And

(3.)

pp 3

lines of

pa

pi

LINES.

527

the series, takes the

= -

--

I(p2

there found that

p)

if

= OP, &o.,

as regards the inversion of a linear

and vector function

(347), or

the return from any one line pi of such a series to the line p which precedes
our general method gives, for the example I., by 354, (12.),

VIII.

to -

|J3(/3

it,

Pl j3y) y ,

and

a result which

it is

easy to verify and to interpret, on principles already

explained.

357.

"We

now prepared

are

to assign

some new and general Forms,

which the Linear and Vector Function (with

vector can be brought, idthout assuming

simplest of

which forms
I.

$p

is

to

real constants) of a variable

its

self-conjugation

one of the

the following,

Y,,p +

with

YApjii,

F.

q = g +

two real and constant

beiug here a real and constant quaternion, and A,
which
all
be
when
the
can
is
vectors,
particular form of
definitely assigned,
given
except that A and
may be interchanged (by 295, YII.), and that
g

fj.

/LL

may be multiplied by any scalar, if the other be divided by the same.

It will follow that the scalar, quadratic, and homogeneous function of a vector,
either

denoted by Bp^p, can always be thus expressed

II.

Sp^p

2
gp +

or thus,

II

.
Sp<j>p

= g

p~

-f

2SApS/ip,

if

a general and (as above remarked) definite transformation, which is found to

be one of great utility in the theory of Surfaces* of the Second Order.
(1.) Attending first to the case of self-conjugate functions fap, from
which we can pass to the general case by merely adding the term Yyp, and

* In

the theory of such surfaces, the

of what are called the cyclic normals.

two constant and

real vectors,

A.

and

/*,

have the directions

ELEMENTS OE QUATERNIONS.

528

supposing

what precedes) that

(in virtue of

gular vector-units,

and

0i0 2

aia 2 a 3

[III. n.

three real and rectan

are"

three real scalars (the roots of the cubic

6.

0),

such that
III.

we may

(0o

0j.cti

0i)ai

0,

(0o

0.>ao

)a 2

0,

r<

(0

+ ^s)3 =

0,

write

IV.

p =

(oiSaijO

jO

0iaiSai/>

-f

+ a 3 Sa 3jo),

aoSao/o

and therefore

V.

a 2 Sa 2 p +

a 3 Sa3/o

so that

(03

VI.

-0 2 |0 =

C 2 )a 3 Sa 3/o

(03

(0i

-1-

(2.)

VII.

VIII.

= (Sa 1]0 ) 2 + (Sa 2f

p*

= Spfap =
+ 2 -

S/00/o

= ~

(0

(0 2

2f>

= -

03/o

(0 3

0!

<

<

+ (Sa
2

2
3/

o)

2
(Sa^) +

,)

0i) (Saj/))

)a 2 Sa 2/o,

(03

(Sa 2Jo)

(0 3

(03

3/ t>)*

0i)

(Sa 3jo)

2)

(Sa 3jo)

2)

(Sa 2 p)

2
:

assume that

X.

or that

3,

2
>)

Ci(8a,/t>)

(Sa

0i)

in general permitted to

it is

IX.

0,/a

which

)a 1 Sa, o,

= -

in

(0 2

2
3 being thus put in evidence.
have thus the general but scalar expressions

the binomial forms of 0i,

We

0i)a 3 Sa 3 p,

0i

= 20 2

= 20 /2

being real scalars, and the numerical coefficients being introduced for
a motive of convenience which will presently appear.
one of the expressions VIII. with IF., we
(3.) Comparing the last but

and

see that

we may bring SpQp

XI.

X =

because SX/x =
(4.)

But in

0ai

a3

ju

to the

= -

J(0i

0ai

general (comp. 349,

XII.

SoO

Sot

proposed form

a3 ,

II.,

by assuming,

g = SX/u

= - J(0 +
X

3 ),

3 ).

(4.))

we cannot

unless

XIII.

have, for all values of

.

/o,

unless the self- conjugate parts of

and (^ be equal; we can therefore
=
= its own conju
=
II.
from
that
Hbecause
m/6T
VXp/i,
V/i/oX
/o
VX^o/t
gate; and thus the transformation I. is proved to be possible, and real.

that

is,

<7p

RECTANGULAR AND CYCLIC TRANSFORMATIONS.

AET. 357.]

Accordingly, With the values XI. of A,

(5.)

XIY.

fap

ffp

+ VAp/i = p(g -

?,

/z,

SAju)

S(e o 3
eajp +

(e

529

the expression,

ASju/o

,uSA/>,

becomes,

XV

oj

= -

C 2p

(e aa

which agrees, by X., with VI.

if g, A, and
(6.) Conversely
- =
where
g
g
# VA/i,
^ VA^u

eai)

be constants such that

as before
hence -

fj.

c,, c 8 , c 3 of

three roots

SA/*,

the cubic J/ =

XVI.
we

are,

that of

is

by VI.,

A =

a3 ,

ai,

aa\

+ #

and the normal

0,

must have one of the three directions

that this plane

az -

of

01,

a 2 , a3

if

eai)

a3

S(<?

= gp +
VA/o//, then
must be one of the

<

0]

gr

to the plane of A,

then we assume, on

trial,

and write accordingly,

a 3,

= ba\ + & a 3 ,

/a

to seek for scalars

2
j

adW which

shall satisfy the three con

ditions,

XVII.

2ab =

d -

c2 ,

2a

c3

c2 ,

aV +

ba =

but these give

XVIII.

(2abJ = (Zba f =

(c*

so that if the transformation is to be a real one,

and

c3

words,

c 2 ) (c,

Cl ),

we must suppose

are either both positive, as in IX., or else both negative

c2

we must

the order IX., witli the values X.,

posing that

XIX.

a =

and are thus led back from XVI.

arid

we

V=

a =

be

may

Adopting then

XVII. by sup

satisfy the conditions

/,

cz

or in other

so arrange the three real roots of the cubic, that c z

(algebraically) intermediate in value between the other two.

for A,

that

-b = e;

to the expressions XI., as the only real ones

g which render possible the transformations

I.

and

II.

except
be interchanged, &c., as before.
see, however, that in an imaginary sense there exist two other
(7.)
for if we retain
solutions of the problem, to transform
as above
and
that A

/j.,

and

JJL

may

We

<j>p

$p<pp

the order IX., and equate cf in IF. to either - c or - c 3t we may in each case
conceive the corresponding sum of two squares in VIII. as being the product of
two imaginary but linear factors the planes of the two imaginary pairs of
t

vectors which result being real,

HAMILTON

ELEMENTS OF QUATERNIONS.

and perpendicular respectively

to ai

and
3

a3

ELEMENTS OF QUATERNIONS.

530
(8,)

And

if

the real expression

to pass

from

it

to the expression V.,

investigation in 354, (12.)

XX.
XXI.

d = -g ai

in which however

Tfi,

We

XXII.
XXIII.

be given, and

be required

it

cz

TA,u,

U(AT>

= - g +

SA/*,

<?

a 2 = UYX//,

M TA),

= - g + TA/u

a3

U(A

I>

permitted to change the sign of any one of the three

= cz-

2e*

(9.)

/o

Accordingly the expressions XI. give,

vector units.

TX =

for

6.

with the order of inequality IX., the

enables us at once to establish the formulae

it is

XIV.

[III. n.

A -

c l9

2<?ai,

have also the two

.

SXpup = P *T\n
SX/o/ip

= -

SX/u

= 2/ 2 =

cs

2^

<?

o2

SX/*

2,

A +

= g +

ji

cz

= 2/o 3

identical transformations,
8

-f

^TA/x

which hold good for any

- 2ee a 3 ai = +
YA^u =

=
JUL

TX/i

{(SX/i/o)

+ (SApT^u +

S^TA)

2
)

(TX/i

2
{(SA^) + (SA/oTV S MjoTA)

three rectors. A,

and may (among other ways)

and XXI., from II. and YIIL

/u, /o,

be deduced, through the expressions XX.

(10.) Finally, as regards the expressions YI. for fap, &c., if we denote
the corresponding forms of -^p by ^1^0, &c., we have (comp. 354, (15.) ) these
other expressions, which are as usual (comp. 351, &o.) of monomial form

and wliich verify the relations 354,

whole foregoing theory.
358.

The

general linear

and

XLL,

and

vector function

several other parts of the

of a vector has been seen

(f>p

and accordingly
(347, (1.))
the expression 357, 1. involves that number, namely four in the term Y^ /o, on
account of the constant quaternion q 0) and five in the other term YA/o^u, each of
to contain, at least implicitly, nine scalar constants

the two unit-vectors, UA and

Uju, counting as two scalars, and the tensor TA/z
as one more.
But a self-conjugate linear and vector function, or the self-conju

p of the general function 0p, involves only six scalar constants ;

either because three
or because the
disappear with the term YY/O of Qp
gate part

<p

FOCAL TRANSFORMATIONS.

ARTS. 357, 358.]

531

= 2y =
(comp. 349, XXII. and 353,
of self-conjugation, 2V/3a
form VfiSap (347, XXXI.),
the
take
for
we
arises
when
which
$p
XXXVI.),
scalar
is equivalent to a system of three
equations, connecting the nine constants.
condition

And

for the

same reason the general quadratic but

involves in like

manner only

scalar

function, 8/o^p,

Accordingly there enter

six scalar constants.

only six such constants into the expressions 357, II., II ., V., VIII., XIV.
for instance, being three such, and the rectangular unit system
3
Ci, c zy
;

The following other general transformations

not
of SpQp and
quite so simple as 357, II. and XIV., involve
although
the same number (six) of scalar constants, and deserve to be briefly considered
eii,

a2, a3

answering to three

others.

(j>op,

namely the forms,

II.

in which a, b are

= -

fap

two real

and

scalars,

We

are two real unit-vectors.

a, |3

the leading formulae, leaving the reader to supply the

merely set down

analysis, which at this stage he cannot find difficult.
(1.) In accomplishing the reduction of the expressions,

shall

8pi>p

= ^(Sa^) 2 +

C 2 (Sa 2/

c 3 (Sa 3/o)

>)

357, VIII.

and
i/D

c 2 a 3 Sa 2 /o

C z aS>a 3 p,

357, V.,

new forms I. and II., it is found that, if the result is to be a real one,
- a must be that root of the scalar cubic
= 0, the reciprocal of which is

to these

It is
algebraically intermediate, between the reciprocals of the other two.
therefore convenient here to assume this new condition, respecting the order of

mn

the inequalities,

which
c\>

02>

Ca,

(2.)

all positive,

"I

V\

->

-1
(

t*

>!,

/>

>

This being laid down

IV.

V.

~l

t/3

but will be incompatible

taken), the (real) values of

in

with the arrangement 357, IX.,

will indeed coincide

be

(or even,

a, b, a, /3

may

a = -

c%,

a = xai + Sa 3

we

if

witli it in

if

every other

)3

case.

choose, the opposite order being

be thus expressed

c\

the three roots

cz

= x

c3

a\

z a3

which

ELEMENTS OF QUATERNIONS.

532

VII.

VIII.

dci

=>

Y = b(xx +zz

^7

csb

X.
XI.

ctf +

C3 *

XII.

And

(3.)

XIII.

XIV.

M^c* =
i |3

bxfz

ft-

*"

#s

0i

xz +

if

ft

(say)

6.

+ y

= 1

(Saj3)

c lC8

= - fy3Saj3 = c^ai +

C 3 sa 3

= &c.

^)

XV.
which

= - Sa|3 =

IX.

there result the transformations

() 2

cV

tfx* +

[III. n.

last, if

XVI.

c^
.

"be

positive, gives this other real form,

S/00/o

-01^3

and

being determined by the expressions VI.

Those
(4.)
expressions allow us to change the sign of
determine a second pair of real unit lines, a and j3 which

a?

and $ in the forms I. and II.

opposite order), and the values IV.

for a

XVIII.

= -

.
Sp<t>p

.
0o/f>

Cz

(VapY +

c 2 aVa>

+ (d -

(c,

of a

C2

x,

may

and thereby

to

be substituted

the order of inequalities III. (or the

have therefore the double transformations

XVII.

and

b,

We

remaining unchanged.

c3 )

= -

(Sj3/>)

C2

two connected forms I. and II. had been given, we

might have proposed to deduce from it the values of CiC a c3 and of aia 2 a by
the general method of this Section.
We should thus have had the cubic,
(5.)

If either of the

XIX.

Jlf

(c

+ a)

{c

(a

b)c

aJ(Sa/3/};

3,

SECOND FOCAL TRANSFORMATIONS.

ART. 358.]

and because the quadratic

XX.

1
.

(c-

^)

(c

a)~

(Saj3)

fl-

(c-

be thus written,

may
l

(a-

533

8.(ap)>

b~

a- (Va/3)

0,

one positive and the other negative if

=
then we arrange the reciprocals of the three roots of
in the order III.,
Q
we have the expressions,
it

gives two real values of

XXI.

Cl

(b

c~

a~

a)

+ \ab

+ 2a^b~ B. (j3) 2 + &-);

l

</(a-

-|

(c,

(b

a)

\ab */(a* + 2a-

b~

2
.

(a/3)

GI~

c3

~l

The

(GI

c3 )

z
:

ab(Safi)

results,

roots c l9 c2

for the directions of

their negatives)

Accordingly these expressions for the roots agree evidently

IV. and XI., because S. (a/3) 2 = 2(Sa]3) 2 - 1.

0.

>

with the former

(6.)

= - a

+ b~

the signs of the radical being determined by the condition that

cz

<n,

being thus known, the same general method gives

c3

a 2 , a 3 the versors of the following

expressions (or of

Sc a
3

XXII.

-f

6j3So)

of which the monomial forms

XXir.

oi

U(c3 a +

may

again be noted, and which give,

a3

^/3Sa^),

Accordingly the expressions in

(7.)

XXIII.

c,a

bfiSafi

(c,

(2.)

give

Yaj3

Cl )xa l9

a3

UYaj3,
(if

(x &

= + U(c,a

we suppose

a 3ai

= + a 2 ),

and

as

an additional verification of the

consistency of the various parts of

whole theory,

XXIV.

it

ac^(c,a
=

As

2
-f

Z>j3Sj3)

c2 ) (c2

(c,

Cl )

Cl ),

(c,

we may

XXV.

(c,

^(Va/3)
+ ^]3Saj3) 2 =

c,),
l

ac{- (c ia

regards the second transformations,

to prove that

this

be observed (comp. 357, XXIV.}, that

may

(c*

(8.)

C3

( Cl

c3 ) (c,

XVII. and XVIII.,

write,

Cl )a

bftafi

aa,

(c3

c^

aa(3a

bfr

it is

c,).

easy

ELEMENTS OF QUATERNIONS.

534

XXYI. ..-(*so that

we have the following

XXVII.

which

aVa

true for

is

expressions

XXVIII.

it is

made by

(c3

ci)o

(fy3a|3

(a*

any two unit lines a, ]3, and any two scalars a, b.

evident from (4.), that ai, a 3 must be the bisectors

any vector

XXY. may

p,

a,

and

(c 8

whence, by XXIII., we

(10.)

XXX.

( Cl

ca) (j3

j3

(11.)

and the

and a 3 the external angle,

of the lines a, a

these other expressions,

to the simple forms,

.

]3

with the recent meanings of the

And

j3

2( Cl /3

which can easily be reduced

XXXI.

|3,

c3 ),

a - a = 2sa 3

a + a = 2#ai,

we have

made by
- a - d +

write,

so that GI bisects the internal angle,

the same time

ci)o

may

XXIX.

At

also of those

be thus written (because

6.

equation,

+ b(8p*)

Accordingly

(9.)

of the angles

ctf

[III. n.

j3

2o/ ai ,

|3

coefficients

a?

j3

and

we

although, for the sake of obtaining real transformations,

have supposed (comp. III.) that

XXXII.

1
.

(cr

c^)

(cr

cs- )

>

0,

because the assumed relation a = xa + sa 3 between the three unit vectors

whereof the two latter are rectangular, gives x* + s 2 = 1, as in IX., so

that each of the two expressions VI. involves the other, and their comparison
afters,

gives the ratio,

XXXIII.

a8

= (cf 1

cz

-1

(c,~

c3

~l
),

we see that, without this inequality XXXII. existing, the foregoing

transformations hold good in an imaginary (or merely symbolical) sense so
that we may say, in general, that the functions
and 0/o can be brought
S/o^/o
yet

<

PASSAGE FROM CYCLIC TO FOCAL FOBHS.

ARTS. 358, 359.]

and

to ihe forms I.

II. in six distinct ways,

whereof two are

real,

535

and the four

others are imaginary.

(12.)

It

be added that the

may

first

XXII. admits

equation

of being

the following,

replaced by

XXXIV.

= -

^p

with a corresponding form for

bc{-

\^ 3

(c$

and that

S(d)3

tfaSa/3)

ffaSo/3)p,

thus, instead of

XXIF., we

are

at liberty to write the expressions,

XXXY.

= U(dj3 -

a:

for the rectangular unit system,

359. If

we

call, as

I.

we

= UVa/3,

deduced from

I.

- U(c 3 j3 -

a3

ffaSa/3),

or II.

naturally may, the expressions

fap

a2

rtaSa/B),

CiOiSaijo

CtfijSazp

C 3 a 3 Sa 3 |0,

357, V.,

and
II.

the Rectangular Transformations of the Functions

another geometrical analogy, which will be seen

briefly

of the

theory

<f>

and

SJO^/Q,

when we come
we may

Surfaces of the Second Order,

of

then by
to

speak

call

the

expressions,
III.

p = gp + VA/o/i,

357,

XIV.,

and
IV.

=
Sp<j>p

gf

357, II.,

SA/ojup,

the Cyclic* Transformations of the same two functions

and may say that the

two other and more recent expressions,

V.

VI.

<

oj

= -

aaVap +

358, II.

flj3Sj3/o,

and
.

Spp

aVa

are Focal^ Transformations of the same.

358

+ b(S3Y,

We

>

have already shown (357) how

exchange rectangular forms with cyclic ones and also (358) how to pass
from rectangular expressions to focal ones, and reciprocally but it may be
worth while to consider briefly the mutual relations which exist, between

to

cyclic

and focal expressions, and the modes of passing from either

Compare the Note

to Art. 357.
a, a , of 358, are the tivo real focal lines of the real or
= const.
asymptotic to the surface of the second order, Sp^p

f It will be found that the two real vectors

imaginary cone, which

is

to the other.

ELEMENTS OF QUATERNIONS.

536

pass from IV. to VI., or from the cyclic to the focal form, we
accomplish the rectangular transformation II., with the values 357,

first

XX., and
assumed

XXL,

of

and of

c z , c3 ,

c-i,

a2, a3

01,

and then

Y1-L1.

c3

c2

>

VIII

{V. p(d

(UX + Up) +

IX.

(cs

c2 )

Sp^p = (V.

4
f>(c

it is

c/p

which

of

4
(- c 3 )

VXp +

3)

(UX
(UX -Up))j

c^ (XTp +

(-

pTX))

\^ ~\T\

C2 )

4
<?

YAp

fe-d

to be

XI.

(UX - Up)

(UX + Up) + (-

Cl )

CJ
//
+ (o
p((-

X.

c3

4
4Sp0p = {S.p((- d) (UX -Up) +

IX.

.L

+ {V.p((-

which

as in 357,

Ci,

>

4bp^)p

the order of inequality being

have (comp. 358, XV.) the following expressions

shall

T7TTT

and

to be

VII.

in

6.

To

(1.)

may

[III. n.

remembered that (by 357, XX.),

d =

all

T\fi

= - g +

cz

c3

SXju,

= - ^ + TX/z

are symbolically true, or give (as in IV.) the real value

+ SA/o^up for Sp^p,

if

g,

X,

/u,

And

p be rm/.

down

although they have been written

in this symbolical sense,

count as three

as four, they only

and r0J

form because the

and other formulae
expression VIII is an immediate consequence of VIII.
IX and
in
IX. and X.
like
from
manner
at
be
once
derived
might
distinct focal transformations,

of a given

cyclic

But

(2.)

if

we wish

to confine ourselves to real focal forms, there are then

four cases to be considered, in each of which some one of the four equations
VIII. VIII IX. X. is to be adopted, to the exclusion of the other three.
.

Thus,

if

XII.

the form VIII.

XIII.
If

XIV.

cz

is

>

c2

>

>

0,

and therefore

the only real one.

c3

cz

>

>

c,

>

>

>

cz

>

cr

1
>

ci

l
>

ci

l
>

0,

If

c l9

ci

d,

ci

l
>

ci

l
>

>

1
>

>

>

Ct"

cr
ci

1
,

then X.

is

the real form.

the only real form

is

IX.

PASSAGE FROM CYCLIC TO FOCAL FORMS.

ART. 359.]

Finally
that

XV.

if

is, if

all

>

c3

>

c2

the roots of the cubic

c l9

>

>

c,~

>

c2

~l

ci

>

in VI.

|3

with

0i

c2

-f

l
,

be negative, then VIII

to be adopted, under the same condition of reality.

all the roots c are positive, or in the case
(3.) When

real focal form, the unit lines a,

may

XVir.
so that

if

we

4S P 0p = N{(c s* +
48/0^0

4
j

is

the

3
(<j

cfy P .TJX +

d4 )

UA

(c 3

d4)

may

be written

TJ/i

write, for abridgment,

XVIII.

we

the form

as before (358, IV.).

4$p<j>p

is

TJ/0 +

In the same case VIII., the expressions for

(comp. 358, XVI.) under either of these two other real forms
.

when VIII.

be thus expressed

(4.)

XVII.

537

- J (^ + ^)UA,

Ko

shall have, briefly,

XIX.

and

= i (cf 4 +

shall then

which

Spfc,

= N(
+
ioJ

jO^)

- N(|0

+ K

0/

0).

Or we may make

(5.)

XX.

may

)UA,

i ( Cl

-*

4
cs"

which form, or

of its

two

was lately shown

(in 358, (8.), &c.)

ELEMENTS OF QUATERNIONS.

how

same kind, with

in the same plane as the two old or given

S

XXIX.

vector constants,

of the ellipsoid,
i,

K, it is

and

necessary

of the cubic should all be positive as above.

real focal form to a second of the

HAMILTON

whence

)!!/!,

be compared with the equation 281,

that the roots

It

K -

have the transformation,

for the reality of

(6.)

c3

lines, a,

to pass

its
]3

from a given and

new real unit

;

but

lines

a,

f3

we have not yet

3

ELEMENTS OF QUATERNIONS.

538

shown how

from a focal form

to pass

[III.n.

6.

although the converse

to a cyclic one,

passage has been recently discussed. Let us then now suppose that the/orm
YI. is real and given, or that the two scalar constants a, b, and the two unit
vectors

a,

have real and given values

]3,

and

let

us seek to reduce this

expression VI. to the earlier form IV.

We

(7.)

might, for this purpose, begin by assuming that

XXII.

or

1
>

as in 358, III.

ci\

>

which would give the expressions 358, XXI. and XXII., for c&c* and aia 2 a 3
and so would supply the rectangular transformation, from which we could pass,

as before, to the cyclic one.

But

(8.)

focal form

XXIII.

XXIV.

in each of

vary a

to

some one

is

Spfp =

(S/3 of

Spjp =

(SjSorf

which a and

(Va of

XXIir.

2
;

>)

>)

XXIV

+ (V

o)

oj

fact the

that the given

8p$p = (V

Bpjp =

2
OJ

(Va

o)

a)

(Sj3 0/o)

o]

(S/3 0/o)

and

real vectors, but not

cases

included under the

are conceived to be given

|3

and which are in

generally unit lines;

now suppose

the analysis, let us

little

of the four following (comp. (1.))

four
+ b($f3p) z according as the scalars a and b are positive
It will be sufficient to consider the two cases, XXIII. and
2

general form, a(Vap)

or negative.

XXIV., from which

(9.)

For the

the two others will follow at once.

XXIII. we

case

XXV.

easily derive the real cyclic transformation,

2

(Sj3 oP )

Sp<t>p

(S oP )

+ a

where

XXVI.

X=

and the equations 357,

cyclic

+a

]3

fl ,

n =

i(/3

fl

),

g = i(a

j3

enable us to pass thence to the two imaginary

(9.)

forms.

(10.)

For example,

XXVII.

if

the proposed function be (comp. XIX.),

Spfp = N(,

oj

pKo )

(S(i

Ko )pY

(V(*

/coV)

we may write
o

*o

/3

<

KO ,

X =

2t

n =

KO ,

g =

8
t

2
;

PASSAGE FBOM FOCAL TO CYCLIC.

ART. 359.]

and the required transformation

XXVIII.
To

(11.)

is

>

XL),

(comp. 336,

+ pBo )

N( (rf

(,

539

KO

+ 2

XXIV.

treat the case

for simplicity the subin dices

by our general method, we may omit

and write simply (comp. V. and VI.) the

expressions,

XXIX.

.
<j>

= -

aVap +

in which, however,

now unit

not

it is

lines (8.).

the expressions

XXXI.

S^p

2
(Vap) +

(S/3/o)

= a 2 (Sa/3) 2

XXXII.

= a 2 (a 8 -

VajSS^ap

Va^Sa/3

/3

a\aVap +
-

-f

(Sa/3)

a(a

w"

- 2a 2

|3

j3V/3/>)

a>

)S af

/3

>,

and therefore

XXXIII.

M=

- a2
)

(c

(c

- a2 )c

(j3

(Sa/3)

),

and

XXXIV.
=

(a(a

j3

(12.)

()3

- aS
a/o)

(c/o

cfi)Sf3p

-f

^(aSo/Q

(c*

(a

a- (a

+ 4(Sa/3) 2 =

2 2

j3

(a/3)
3

(/3a)
2

jSaj^J

(a
4

j3

j3- (/3

)c

()3

)3

/3S/3/o)

Introducing then a real and positive scalar constant,

= a4 +

in

- a2
)

Va/o/3Sa]3 +

-e) + ^Sa]3)Sap + (aSa/3

XXXV.

r,

2
.

(a/3)

2
.

(a]3)

the roots of

M=

(Sa]3)

XXXVI. ..*

+ (Vaj3) 2 = 2(Sa/3) 2 - a 2]3 2 = 2(Vaj3j 2 +

= i(a 2 -

j3

),

cz

= a2

and when they are thus arranged, we have the

XXXVII.

ci

>

>

c3

>

c2 ,

cf

c,

|3

j3

= *(a 2 -

|3

2
>*

inequalities,
-

1
>

>

c2

1
>

C3-

such that

+ aj3a) 2 = &o.,

admit of being expressed as follows

+ 4(Va)3) 2

= a 4 + 2S

(SajS)

which (by 199, &c.),

to be observed that a and )3, though real vectors, are

2
Hence, because - aVap = a$ap - a p, we easily form

XXX.

and

jSSfto,

Z 2

ELEMENTS OF QUATERNIONS.

540

The corresponding forms

(13.)

[m. n.

6.

p are the three monomial expressions,

of

^ p = Vaj3S/3a/>,
2

XXXVIII...
j3Sa/3)

which may be variously transformed and

verified,

and give the three follow

ing rectangular vector units,

XXXIX.

01

in connexion with

which

T(a Cl +

We

]3Saj3)

- c^
-

c a )i (c s

Cl i(c3

c 2 )^

^^

(ci

= r

c 3 )i

(c,

(c,

c$

(- cz }\

^3)^

(c*

c^ c

XXXVII.

the radicals heing all real, hy

(14.)

c 3 )i (ci

(d

/3Sa/3)

= TJ(ad +

a3

easy to prove that

it is

J TYj3 =

= TJVaj3,

a2

|3Sa/3),

= (T(ac, + /3Sa]3)

XL.

= U(aC 3 +

have thus, for the given focal form

XXX.,

the rectangular trans

formation,

XLI.

sp

= (VapY +

<t>p

Cl

(&(ac 3 + j3Saj3) o)

c 2 )r*

cz ) (c3

(G!

cz )

Ci(c s

- ^ V2
;

or briefly,

XLII.

S/00/o

= d(S .pU(aC 3 +

2
(Vap) +

(S/3/o)

+ a 8 (S pUVajS) 2 +

/3Sa/3)/t>)

in which the first term

is

positive,

C 3 (S

2
pUCaC! + j3S/3))

but the two others are negative, and

c i9

are the roots of the quadratic,

XLIII.

We

(15.)

- az }c

(|3

2
.

(Saj3)

have also the parallelisms,

XLIV.

ac,

+ ]3Saj3

||

j3c!

aC,

aSajS,

+ /3Saj3

||

j3c 3

aSaj3,

because
c,c3

and may therefore

XLV.

Sp^p =

= -

(Saj3)

write,
2

(Va/))

= d(S pUfjSc! .

(S/3/>)

+ o*(S /oUVa/3) 2 +
.

aSa/3))

6 3 (S

^Uf/Bcs

aSa/3))

PASSAGE FKOM FOCAL TO CYCLIC.

ABT. 359.]

541

while

XLYI.

and r =

TQ3ci

c 3 )i,

(ci

= rcjfa -

aSajS)

cs )l,

T(j3c 3

XLYII.

new
.

c z ) (c

(c,

obtain this

C8 )*

(<?

C 8 )i,

with real radicals as before.

2
2
Multiplying then by r (TYa/3) or by

(16.)

= r
(-

aSa/3)

c 3) (c3

we

c 2 ),

equation,

( Cl

C,]

(C3

(dSjS/o

2
- a2
(Saj3p)

{(TYa/3) ((Yap)

(SjSp)
2

SajSSap)

(c,

- a2
)

(cpp

oSajSj

only another way of expressing the same rectangular transformation

as before, but has the advantage of being freed from divisors.
- CA
(17.) Developing the second member of XLYII., and dividing by c

which

is

we

new

obtain this

XLYIII.

transformation

(TVapyfytp = -

(Ya]3)

= a 2 (Saj3 o) 2 -

we have

in which

(Sap)

+ 2a 2 Saj3Sa oS/3 o
j

The expressions

C=

XXX YI.
L.

C=

2
dcs - a (c! +

cs ).

for c i9 c3 give thus,

2
- a4 (Yaj3)

and accordingly, when this value is substituted

equation becomes an identity, or holds good for all
a, )3,

as

may

for

in

XLYIII., that

values of the three vectors,

be proved* in various ways.

this result,

Admitting

(19.)

the equation

written for abridgment,

XLIX.
(18.)

2
((Ya^) +

(Sa]3)

XLYII.,
XLIII.

it is not-

particular quadratic

we

see that for the

ct

and

sufficient,

for

necessary that
It

is

c2

mere establishment of
should be roots of the
purpose, that they

this

should be roots of any quadratic,

LI.

between

+ Ac +
its

with the relation

0,

LII.

Ac? +

But when we combine with

coefficients.

B+a

+ (Ya/3) 2 =

0,

this the condition of

rectangularity, a 3 JL ai, or

LIIL
*

= S

( Cl )3

aSj3)

(c 3 /3

Sj3)

-4(Sa/3)

Bp

(Sa]3)

Many such proofs, or verifications, as the one here alluded to, are purposely left, at this stage,
as exercises, to the student.

ELEMENTS OE QUATERNIONS.

542

we

[III. n.

6.

obtain thus a second relation, which gives definitely, for the two coefficients,

the values,

LIV.
and

so conducts, in a

In

(20.)

this

new way,

manner, then,

- a2

]3

of

= a2

(Saj3)

to perceive the truth

with the quadratic equation

XLVIL,

are roots, without having previously found the

of which the quadratic is a factor, and of which the other

which

XXXIII.

B=-

to the equation XLIII.

we might have been led

rectangular transformation

of the

XLIII.
cubic

and

Ci

c3

But if we had not employed the general method of the present

which
conducted
us to form first that cubic equation, there would
Section,
have been nothing to suggest the particular form XLVIL, which could thus

root

is c 2

have only been by some sort of chance arrived

The values

(21.)

LV.
that

is,

LVI.

a^aa give

= (S P
U(j3 Cl

Cl c,( Cl

also (comp. 357, VII.),

2

Saj3))

by XL. and
.

of

at.

2
+ (S pUVajS) 2 + (S pTJtfc, - aSa/3))
.

XL VI.,
2

C3 ) (p

(Vaj3)

(Safto)

C,(c,

- a2
) (c&fo
- a2
) (c 3
cfa

SajSSa/o)

SajSSa/o)

S/3/>

XXXVI.

and accordingly the values

member
(aS(3p

of

this

last

of

common

under the

equation

jSSap)

as in

357,

LVII.

-S

of the

is

((SajSj

p(cfi

2
.

&c.).

>

2
((Vap) +

aSa/3 + c^VajS) S

c2

(XXXVII.) with

c3 )

(S^)

be perpendicular to

is to

p(cfi

to

eliminate

which gives

((Sa/3)

it

c/3*}

will be

(C L

aSajS

- a2
)

2
(Saj3)>

C^VajS)

found that the following equation,

2
(Vap) +

(c

the

for the real cyclic trans

We are therefore
XLVIL and LVI.,

real transformation

Cl j3 )

>

kind required.

Accordingly

(23.)

(3.),

between the equations

SajSSap)

few reductions) the

(after a

LVIII.

c\c$(c\

(and not to a 2

which

form,

c
Comparing the recent inequalities c
that
we
arrangement 357, IX.,
see, by 357, (6.),
formation (6.) at present sought, the plane of X, ^

(c 3 Sj3/>

enable us to express each

c3

Ci,

(22.)

a3

2
)

Saj3S a/

>)

vectors

|3,

543

holds good for all values of the scalar

2
2
since, by addition of c(Va/3) p on both sides,

identity, or that

an

is

AND IMAGINARY.

IDENTITIES, KEAL

ART. 359.]

it

and

c,

of the

takes this

it

obviously identical form,

LIX

8
.

so that if

Cij3

((Sa)3)

Ci

be

cfr) (Sap)

of the quadratic

e//7?er rootf

LYII.

the transformation

2
- a2
= (rSfto - S/3Sap) 2
) (S|3p)

c(c

XLIIL,

or

at least symbolically valid

is

take, as above, the positive root of that quadratic for d,

if

we

formation to be a real one, as regards the constants which

if

its

we had happened (comp.

transformation

quadratic

XLIIL,

but

(Saj3)

we must

wish that trans

it

And

employs.

LIX., and to see

that way led to form the

(20.)) to perceive this identity

LVIIL, we might have been

in

XXXIII.

without having previously formed the cubic

we

Already, then,

(24.)

- a2 =
)

if Ci(d

see

how

to obtain one of the

two imaginary

cyclic

XXX., namely by

changing c\ to e s in
LYII. and the other imaginary transformation is had, on principles before
2
between XLVII. and LVL; a process
explained, by eliminating (Sa/3/))
transformations of the given focal form
;

which

LX.

easily conducts to the equation,

.

2
2
(Yap) + (Sfto)

-f

aY

(Ci

C 3 )-

Cl -*(cSpp

where the second member

"

SajSSap)

is the sum of tic o squares

being
LYII. would also become, if c\ were
2
2
Accordingly, each member of LX. is equal to (Sap) + (Sj3/o)

>

(<?i

as the second expression

0,

but

c2

<

replaced by
,

if
<?i,

CA

0),
c3 .

be the

roots of any quadratic LI., with only the one condition,

LXL

2
.

.0x<%-.B-

(Sj3)

which however, when combined with the condition of rectangularity LIU.,

2
2
suffices to give also A - |3 - a
as in LIY., and so to lead us back to the
which
had
been
deduced by the general method, as a
quadratic XLIIL,
,

factor of the cubic equation

And

(25.)

member

of

XXXIII.

since the values

LX.

XXXVI.

of c lt

c-3

reduce, as above, the second

form (Sap) 2 + (Sj3p) 2 we may thus, or even

the following expression for
3 at all, deduce
Ci,

to the simple

without employing the roots

the last imaginary cyclic transformation
<?

LXII.

Sp<fto

2
(Ya o) +
(

(Sj3p)

= -

ay

+ 8 (a +

J^j3)p 8 (a
.

ELEMENTS OF QUATERNIONS.

544

where
d

r of

J-

1 is the

imaginary of algebra (comp. 214, (6.)) while the real scalar

at the same time receive the connected imaginary form,
a -

Filially, as regards the

(26.)

second real focal form (comp. 358,

LXIV.
in

which a and

(Va P

passage from the given form

(4.)), or the transformation,

(S/3 P )

)>

+ a and

be sufficient (comp. 358,

may

it

XXX.,

to a

= (Va
pf

are real vectors, distinct from

j3

same plane with them,

the formulae

6.

XXXV. may
LXIII.

[III. n.

j3,

(8.)), to

but in the

down

write

LXV.

rV =

(a

|3a|3),

2
with the same real value of r as before

fi

= -

(]3

so that (by

a)3a),

XXXV.,

&c.)

we have

the relations,

LXVI.
(a
2

r (a

T/3

= a(r2 - a 2 +

)3

2 a) = a(r + a

)3

+ a

To = Ta,

)
2

Sa

T/3,

2/3Saj3

/3

= Sa/3

= - 2(aC 3 +

/3Saj3)

+ 2/3Sa]3 = 2 (ad +

j3Saj3)

+ 2Sa/3 = -

||

||

a3

LXVIII
2

1 r (/3

(27.)

We

LXIX.

with which

LXX.

/3

/3(r

- a2 +

3
(V(a +

jSajS)^)

(S(|3

(V(a

which enables us

2
/3aj%) +

its

(S(/3

to pass

2(j3c 3

aSa/3)

||

a3

ajSajp)

two new

lines

(a*

ajSa)?)

- 2S

2
.

(a/3)

)30

form XXIII., to a second real focal

and it
in the same plane as the two old ones

from the

may be noted that we can pass

a to a J - J
.

be combined this other of the same kind,

form, with

]3

have then the identity,

may
.

focal

from

LXIX.

to

LXX., by changing

BIFOCAL AND MIXED TRANSFORMATIONS.

ARTS. 359, 360.]

545

360. Besides the rectangular, cyclic, and focal transformations of

which have been already considered, there are others, although perhaps of
but we shall here mention only two of them, as specimens,
less importance
whereof one may be called the Bifocal, and the other the Mixed Trans
:

formation.

The two

(1.)

lines

LXV., being called focal

them both may be called on that

a, of 359,

expression which shall introduce

lines* an

account a

bifocal transformation.
4
Eetaining then the value 359, XXXV. of r and introducing a new
auxiliary constant e, which shall satisfy the equation,

(2.)

- a 2 = rV,

I.

]3

and therefore

IT.

2
4(S/3) -

r*(l

<?),

so that

III.

the

first

z
r (ea - a

4e (Saj3)

LXY.

equation 359,

IV.

=
)

(1

=
=
in which a 2 = a

(e

=
e*)8p<t>p

(]3

- a2 2
)

r\e8ap

Sap) = 2Saj3S|3p

e*)

LXVL,

(V^) +

- Sa
(Vap) + (eSap
(Sap)

of S/o^/o,

(1

l)aV +

by 359,

XXX.

(1

(1

V.

2/3Sa/3,

gives,

and therefore, with the form 359,

VI.

(Sfto)

2e8ap8a p + (Sa pY

so that a

and a may be considered

enter symmetrically into this last transformation, which

is

of the bifocal

to

kind

above mentioned.
(3.)

For the same

reason, the expression last found for

again (comp. 358) six scalar constants; namely,

involved in the two unit lines, Ua, Ua

e,

Ta(= Ta

SpQp involves
and the four

),

In

foregoing transformations, the scalar and quadratic

function SpQp has been evidently homogeneous, or has been seen to involve no
terms below the second degree in p.
may however also employ this
(4.)

all

the

We

apparently heterogeneous or mixed form,

VII.

.
Sp<j>p

HAMILTON

= g (p -

2
)

Compare the Note

ELEMENTS OF QUATERNIONS.

+ 2SA( /0 -

to Art.

359

[p. 535].
4

ELEMENTS OF QUATERNIONS.

546

which g

in

new

[III.

have the same significations as in 357, but

subject to the two conditions of homogeneity,

X,

constants,

//

VIII.

6.

% are three

e, e,

and
z

IX.

c/e

in order that the expression VII.

X.

Spjp =

</y

may admit

of reduction to the form,

+ 2S\p$fip,

IF.

as in 357,

Other general homogeneous transformations of

which are them
selves real, although connected with imaginary* cyclic forms
(comp. 357, (7.)),
because a sum of two squares of linear and scalar functions is, in an
imaginary
(5.)

Sp<f>p,

sense, a product of

XI.

XII.
in

two such functions, are the two following (comp. 357,

.

Spjp = gf + 8\ pM p =

Spjp = gf + SXppp =

which (comp. 357,

XIII.

XIV.

and

(2.)

g*p*

(SX 3f

(Sfi lp

Y
2

(S/is/O

>)

(8.)),

= g +
= T\fi

g,

+ (SX 1)0 ) 2 +

g^

(9.jj

c ly

g,

= g - TA/x = -

cs ,

Xa

= VA/.(TA^ -

SX/i)^,

Ml

= (XT^ +
^uTX) (TX/.

X 3 = VX/.(TX^u +

SXju)-*,

/i 3

= (XT^u - ^TX) (TX^ +

and

XV.

so that #!, Xi,

(6.)

/ii,

We have
XVI.

XVII.

with these two

and ^ 3 X 3
,

are rm/,

/z 3

if

g,\,

JJL

be such.

therefore the two new mixed transformations following

Sp^p =

new

g,( P

E ,)

(SX 3 ( P

&))

(8^,( P

&))* 4

pairs of equations, as conditions of homogeneity,

XVIII.

XIX.

XX.

0,6,

+ A.8&A, + (U.S^, =

0,

2
2
^ ll + (S&A,) + (SS 1JU1 ) +

*,

0,

and

XXI.

</

3 63

- A 3 S^ 3 X 3 -

g^ -

/u 3 S^3ju s

2
(S^A 3 ) -

(8,1,2,)

0,

0.

1 /ti, and As
A.I
\/
\/ 1 jits, may here be said to be two pairs of imaginary cyclic
= const.
normals, of that real surface of the second order, of which the equation is, as before,
Compare the Notes to pages 527, 534.
S/></>

RECIPROCITY OE FORMS.

ABTs.360, 361.]
361.

We

547

saw, in the sub-articles to 336, that the

scalar function of a vector,

in general be expressed

may

I.

d/p

differential, Afp,

of a

under the form,

= nSv&p,

a derived vector function, of the same variable vector p, and n

And we now propose to show, that if
scalar coefficient.

where v

is

II.

still
<j)p

and

have the equation

I.

which

of

=
Sp(j>p,

so that the part

and seeking the

the self -con jugate part,

may

n =

v =

2,

$0/0

two vectors p and

each other.

thus be deduced from $p by operating with ^dS p,

dp under the sign S. in the result while there
.

relations of reciprocity (comp. 336,

which are in

v,

we

coefficient of

exist certain general

We

is still

jo

with the values,

III.

(1.)

denoting the linear and vector function which has been considered

in the present Section,

shall

fp

is

this

way

have here, by the supposed

between the

(6.)),

connected, as linear functions of

linear

form

of

the differential

typ,

equation (comp. 334, VI.),

IV.

d^p =

<dp

also

and

p),

by 349, XIII., we have,

hence,

V.
(2.)

of

00/0

let
(}>p,

.
<j>dp)

S(#

/o

dp)

as asserted,

As an example

from

S(/o

of the

employment

of this formula, in the deduction

us take the expression,

VI.

.
<}>p

= SjSSap,

347,

XXXI.,

which gives,

VII.

.fp

and therefore
VIII.

Comparing

this

dfp

with the general formula,

IX.

idyo = Si dO = S

A2

ELEMENTS OF QUATERNIONS.

548

we

form VI.

find that the

X.

of

$p has

j2(/3Sap +

<j)(,p

and

we saw

in fact

(347,

expression,
.

aSjSjo)

that this form gives, as

XXXII.)
XI.

fp

Supposing now, for simplicity, that the function

self-conjugate, by taking (if necessary) the semisum of

we may

write

the

its conjugate,

2aSj3p.
is given,

(3.)

conjugate function,

6.

for its self-conjugate part,

[III. n.

itself

and

instead of
>

and

or made,
its

own

shall thus have,

simply,

XII.

v =

XIII.

0/>,

whence

= 8vp,

XIV.

dfp = 2Svd /0

also (comp. 348, I. II.),

XV.

..>

rv

XVI.

and

Writing, then,

(4.)

XVII.
we

.fp

shall

Fv =

8v<t>-

have the equations,

XVIII.

XIX.

Fv =fp,

dFv = SSpdv = 2S

tff

and generally, as
so that p may be deduced from Fv, as v teas deduced from fp
above stated, there exists a perfect reciprocity of relations, between the vectors
p and v, and also between their scalar functions, fp and Fv.
;

As regards

(5.)

Fv,

it

may

the deduction, or derivation, of v from//o, and of p from

*
occasionally be convenient to denote it thus
:

XX.

v =

dod

XXI.

* [Hamilton suggested the notation

= (S.d/>)~ d in page 291 of a paper published in the Pro
(j
On the same page he introduced the "more general
ceedings of the Royal Irish Academy," vol. iii.
characteristic of operation,
]

d
*T~
.

da;

which

and

+JT- +
dy

"

d
K-dz
7

<,

ordinary rectangular coordinates," while i, j, and A; are unit vectors parallel

has been printed v, and in accordance with the notation for
Elements v =
partial differentiation used in the
+j^y + kD z Now if p = ix + jy + kz, for
in

x, y,

z are

to the coordinate axes.

More recently

<

^x

"

"

>

any system of rectangular axes,

dfp

Comparing

this

with dfp

=
(dxT>

it is

nSvdp,
v. f(p) must be true.
particular set of coordinate axes.]
equation nv

+ &yn v +

dzl>

z )fp

it

Sdpv. fp.

may have any direction whatever, that the

may be inferred that v is independent of any

evident, as dp

Hence

=-

IlECIPEOCITY OF FORMS.

ART. 361.]
in fact, these last

549

be considered as only symbolical transformations of the

may

expressions,

XXII.

dFv = 2S(dv

v),

2S(d/>

d//>

/o),

which follow immediately from XIV. and XIX.

(6.) As an example of the passage from an expression such
resume the

equal expression of the reciprocal form Fv, let us

as

/jo,

cyclic form

to

an

357, II.,

writing thus,

XXIII.

= SP

fp

and supposing that g, X, and ^ are real.

shown (in sub-articles to 354 and 357), if
in

we

(3.),

+ 8\pnp,

0t>

<i>p

Here, by what has been already

typ be supposed self-conjugate, as

have,

XXIY.

XXY.
XXYI.

T//V

v =

(g

$p = gp +

- SX
M)

(g*

~V\p/uL

= -

Ay)

and therefore

XXVII.

which

last,

when compared with

a bifocal form

Us focal

what may be

lines a,

mFv =

360, VI.,

is

seen to be

is

and

bifocal transformations are therefore reciprocals of

(7.)

As another example

former,

us

now

it

of this reciprocal relation

From

.fp

[Since

follows that

Kvp.

The

^u,

cyclic

between cyclic and

from the

latter to the

begin with the focal form,

XXVIII.
*

XXIII.

each other.

focal lines, in the passage from//o to Fv, or conversely

let

called

a (360, (1.)) having here the directions of X,

called the cyclic lines-f of the form

that

of

what we have

= g\pp + \V 2 P

this

"

= Spjp = (Yap) 3 +
=

gp

SA.V/*

and, on substitution, equation

+ \pp.
= (\V2 - 9 z )p +

A^SAp/x

XXVI. may

(g

(ff

(Sfto)

359,

A/i)Sty/i

XXX.,

gv.

SA,u)SA./tyi,

at once be found,

1
remembering that ^ssm^r

."]

t They are in fact (compare the Note to page 527) the cyclic normals, or the normals to the
= const. while they
cyclic planes, of that surface of the second order, which has for its equation fp
;

are, as above, the focal lines of that other or reciprocal surface, of

the equation

is

Fv =

const.

which v

is

the variable vector, and

ELEMENTS OF QUATERNIONS.

550
in

which a and

by 359,

|3

XXIX.

have now,

= 0o = - a Yap +

jSSjSp,

= a 2 (Sa/3) 2

2 l^n = Vav/3Sa/3 + a(a

)Sov,

)3

therefore,

XXX.

mFv =

2
2
a (Sa/3) JFv =

Si^v
2 (a

= SayjSySajS +
2
- v2
(Saj3}

|3

(Say)

+ Sav((a 2 -

/3

- n
(Saj3j + SavS(a +
a

an expression which
focal line a of
3

(a

real

We

real vectors.

6.

(11.),
(

and

and

are supposed to be given

[III. n.

]3aj3),

and

of cyc&c

is

and the

/orm

/3a/3)i;,

Fv being

owe cyclic line of

Fv

other cyclic line of

fp
and consequently (by 359, LXY.)
;

)Sav + 2Soj3Sj3i/)

of

the given

having the direction of

+a

where a

is

the second

focal line of fp.

(8.) And to verify the equation XVIII., or to show by an example that

the two functions fp and Fv are equal in value, although they are (generally)

different

inform,

to substitute in

it is sufficient

XXX.

the value -XXIX. of v

which, after a few reductions, will exhibit the asserted equality.

362. It is often convenient to introduce a certain scalar and symmetric
function of two independent vectors, p and p, which is linear with respect to
each of them, and

is

deduced from the linear and

$p, of a single vector p, as follows

I-

-f(p>

With

this notation,

VI.

and

= f(p

p}

=
$p<t>p

8/

we have
II.

IV.

self-conjugate vector function

.f( p + p

f(p, p)

.. f(xp,

yp

= fp
)

=fp

+ 2f(p, p

V.

= xyf(p,
p

),

+ fp

dfp

= 2f( P dp)

if

NX

V/

=
;

as a verification,

VII. ..f(vp) =

v*fp,

a result which might have been

obtained, without introducing this
function I.

new

DERIVED LINEAR EORMS.

ARTS. 361-3G3.]

551

at this stage, to write down proofs of

(1.) It appears to be unnecessary,
but it may be
the foregoing consequences, II. to VI., of the definition I.
worth remarking, that we here depart a little, in the formula V., from a
notation (325) which was used in some early Articles of the present Chapter,
;

although avowedly only as a temporary one, and adopted merely for con
venience of exposition of ihe principles of Quaternion Differentials.
(2.)

In that provisional notation (comp. 325, IX.) we should have had,

for the differentiation of the recent function/^ (361, II.), the formulae,

f(p,

d./f>

f(p>

V>)>

)>

%&p

;
<t>p

the numerical coefficient being thus transferred from one of them to the
But there is a
other, as compared with the recent equations, I. and V.

convenience now in adopting these last equations V. and

= 2f( P dp),

dfp

I.,

namely,

/fo/O-S^p;

because this function Sp Qp, or S/o0p occurs frequently in the applications

of quaternions to surfaces of the second order, and not always with the
,

coefficient 2.

Ketaining then the recent notations, and treating dp as constant, or

null, successive differentiation of fp gives, by IV. and V., the formulae,

(3.)

d 2 p as

VIII.
so that the

theorem 342,

IX.

typ

d 2/p = 2/(dp)

I. is

(1

d!/p

&c.

here verified, under the form,

+ d + jd2 )/p = /p + 2/(p, dp) +fdp

or briefly,

x.

yp=/(p

+ dp),

an equation which by II. is rigorously exact

(comp. 339, (4.)), without any
supposition whatever being made, respecting any smallness of the tensor, Tdp.
363. Linear and vector functions of vectors, such as those considered in the
present Section, although not generally satisfying the condition of self-conju
gation, present themselves generally in the differentiation of non-linear but
vector functions of vectors.

linear function

by

In

fact, if

w(p), or simply

we denote

for the

by up, the general

moment

such a

non

distributive property

(326) of differential expressions allows us to write,

I.

has

where
<j>

all

dw(p)

=
<(dp),

or briefly,

dwp

the properties hitherto employed,

including that of not being

ELEMENTS OF QUATERNIONS.

552

There

generally self -con jugate, as has been just observed.

an extensive and important

shall soon see,

is

function, wp,

for such a function

exists,

self-conjugation

itself

[III. u.

however, as we

is,

which the property of

case, in

namely when the

<

6.

differentiated

the result v of the differentiation of a scalar function fp

vector p, although not necessarily a function of the second

or more fully, when
such as has been recently considered (361)

of the variable

dimension,
it is

the coefficient of dp, under the sign S., in the differential (361,

I.)

of

that scalar function fp, whether it be multiplied or not by any scalar constant
And generally (cornp. 346),
(such as n, in the formula last referred to).

the inversion of the linear and vector function

differentiation of the inverse (or implicit) function

the equation I. or I
II.

be written under

may

dw^o- =

dv = m~

As

(1.)

manner

in such a

that

this other form,

if

^d(j,

<t>~

in I. corresponds to the

<

w"

<r

up.

a very simple example of a non-linear but vector function, let us

take the form,

III.

This gives,

if

dp

= p

IY.

Y.

(T

(t)(p)

= pap,

a is a constant vector.

= $dp

fyp

dwp = pap + pap =

=S

SX0 o =2SX
j

VI.

where

fX

/ t>ap

/ >VX;

2YX oa = 2Va oX,

/

2V/oa|C/

0V = 2Vapp

and $ p are unequal, and the linear function $p

so that fyp

is

not self-

con jugate.
(2.) To find its self-conjugate part fap
are to form the scalar expression,

VII.
of

which the
VIII.

differential,

WP

by the method

iS^V =

..///=

>S

do
j

On

S|o;
,

is

giving

and accordingly this is equal to the semisum

VI., for $p and its conjugate.
(3.)

2
i

taken with respect to p

= S
00/dp = 2S

of the

IX.

.
/>

2p 8aP

two expressions, IV. and

the other hand, as an example of the self-conjugation of the linear

and vector function,

X.

we

of Art, 361,

d. v

dwp =

when
<}>dp,

dfp

= 28vd o = 2S
/

DERIVED LINEAR FORMS.

AKT. 363.]

even

the scalar function

if

fp

be of a higher dimension than the second,

let

have the form,

this last function

XI.

yjo

553

S0/o0

q,

/D0"p,

being three constant quaternions.

q"

q>

Here
XII.
XIII.

dv =

fydp

v =

wp = ^f(qpq

fyp

yV(qp

pq"

pq"

f>f(\qp

+ q pq pq +

pq"p

q"pqpq }

q)

and

XIV.
=

so that
([>

(4.)

SA0/ = JS

+ p q\) + &o. =

Sp>X

0, as asserted.

In general,

used as a second and independent symbol of

8 be

if

we may

differentiation,

write (comp. 345, IV.),

XV.

8a/0 = a8/0,

where/0 may denote any function of a quaternion in

by the principles of the present Chapter (comp. 344,
;

fact,

each

member

is,

and 345, IX.), an

I.,

expression for the limit*

XVI.

lim. nn {f(q
W=

??

(5.)
first

n~

M - f(q + n~ dq) - f(q + w M 80) + /0).

dq + w 80)
l

QO

z:a>

As another statement

of the

differentiation of /0, with each

same theorem, we may remark that a

symbol separately taken, gives

results of

the forms,

XVII.
and then the

assertion

is,

a/0 =/(0,

that

if

we

d0),

#0 = /(0,

differentiate the

80)

first

of these with 8,

and

the second with d, operating only on q with each, and not on

d0 nor on 80,
we obtain equal results, of these other forms,

xvni.
For example,

We may

sa.# =/(?, dq,

.fq = qcq,

80)

= /(0,

80, a0)

as/0.

+ xdq +

where

c is

a constant quaternion,

two symbols XV. represents the coefficient of x l y l in the

y$q) according to ascending powers of x and y, when such development

also say that each of the

development of f(q
is

if

XIX.
*

possible.

HAMILTON

ELEMENTS OF QUATERNIONS.

ELEMENTS OF QUATERNIONS.

554

common

the

value of these last expressions

XX.

6.

is,

Ufq

[III. n.

Writing then, by X.,

(6.)

XXI.

d/jb

and

XXII.

with

0fy>,

<$d)p

dw/o

0djo, as before,

we have the general equation,

XXIII.

S(d o

which dp and $p may represent r/y tivo vectors the linear and vector
function, $, which is thus derived ham a scalar function fp by differentiation, is
in

therefore (as above asserted

and exemplified)

The equation XXIII. may be thus

(7.)

XXIY.
and

will be

it

three

known

found

is

[If n defined

DJ>

=
<^>dp

is

arbitrary,

and the conjugate of

is

Hence the spin-vector

of 6

This vanishes only

when

is self- con

0(

Z I),

DjL.f

nSvdp

(361, I.)

generally ceases to be self-conjugate.

For

2Jf(p)S/idp,

2Svd/>

/ndF

d/j..F

=
(<>(

(^(

some function

/u.$\dp

and

SAdp,

V\mF-

is

F is

the function

p,

d^ =

this case

system of

Differentiating this again

ftF(p).

df

D Dy

y T) z

d/p

dfp

arbitrary, v

again, as dp

T>

by the equation

if

Here

= D,,D,

example, comparing
is

Sd /0 8v = SS t)dv;

to be virtually equivalent to the following

not a constant scalar but a function of

since dp

equations, in the calculus of partial differential coefficients,

XXV.
*

always, self-conjugate*

briefly written,

d/t

SA(

_ A S u(
y

0dp

F,

0dp.
7

yli

-1

)J

J*- 1
)

or
of /(p), or a constant as

may

be easily verified, and in

jugate.]

t [In terms of the characteristic of operation v, defined in the Note

to

page 548,

it is

easy to see

that

5d/p

d5/p

= - 8Sdpv./=
= - dSSpv./= SdpvSSpv./-

In the transformation of functions

involving

v>

and operating on a single function /(p), or

LINEAB FUNCTION OF A QUATEBNION.

AETS. 363, 364.]

At

364.

commencement

the

(347) the

of a linear (or distributive) quaternion function

of the inversion (346)

problem

we reduced

of the present Section,

555

of a quaternion, to the corresponding problem for vectors and, under this

reduced or simplified/orw, have resolved it. Yet it may be interesting, and
;

it

will

now

be easy, to resume the linear and quaternion equation,

I...fq =
and

with

r,

..f(q +

II.

<f)

= fq +

/<?

to assign a quaternion expression for the solution of that equation, or for

the inverse quaternion function,

l

III.

.q=f~ r,

with the aid of notations already employed, and of results already established.
(1.)

The

conjugate of the linear

and quaternion function/*? being denned

(comp. 347, IV.) by the equation,

IV.

p and

in which

Spfq = Sqfp,

q are arbitrary quaternions,

if

we

set out (comp. 347,

XXXI.)

with the /om,

Y.
which

in

which

is

6-,

more than

and

t,

fq =
,

tqs

qs

S.tqs,

are arbitrary but constant quaternions,

sufficiently general,

we

shall

have (comp. 347,

and

XXXII.)

the conjugate form,

VI.

fp

spt + s pt

Sispt

whence
VII.

./I =

and

Sfe,

VIII.

./I =

Ss*

then possible, for each given particular form of the linear function/^, to
assign one scalar constant e, and two vector constants, c, e such that
it is

IX.

f(ix

+jy +

kz)

v =

*Dj;

..!

+jD y + kD z may

<?

! =

be treated as an ordinary vector since D*,

obey symbolically the ordinary laws of scalar multiplication as expressed by

Comparing

sa//

2SSp^d/j

the vector function

<j>(

= ^vS(

Dj,,

and

XXV.

Sd/aySSpV-/,
)v-/,

since dp and Sp are both arbitrary. Of course v operates onjT and not on the vector operated on by
- 2v, #( ) = - vS( )/,
This expression for
shows again that it is self -conjugate. Again, as
and in this v operates on v and not on the subject of
<p.

v/=

<f>

(/>.]

ELEMENTS OF QUATERNIONS.

556

and then we

[III. n.

have the general transformations (comp. 347,

shall

qfl =

X.

Sfq = S

XI.

V/0 = s$q +

e&q + Se

Y./Y?

6.

I.)

and
XII.
in which Sc ^ =

.fq =

Yg, and

0Y<?

(e

or

is

Y/Y<?

a /wear and

of the kind already considered in this Section

Y<?,

the form Y. of fq,

As

function of

being also such that, with

we have
XIII.

(2.)

t?ecfor

regards the number of independent and scalar constants which enter,

at least implicitly, into the composition of the quaternion function/^,

ways be shown

in various

XII., the scalar

one

e is

to

be sixteen

the two vectors,

it

may

and accordingly, in the expression

count each as three and
e and E
,

the linear and vector function, 0Yg, counts as nine (comp. 347, (1.)).
(3.) Since we already know (347, &c.) how to invert a function of this
last

kind

XIY.

0,

we may
r =

in general write,

Sr +

Yr = Sr 4

0^,

XY.

where

the sm/ar constant, m, and the auxiliary linear

deduced from the function

then to express

and

or Sq

q,

and

t/

tfT

Yr =

gcfor function,

by methods already explained.

in terms of r, or of Sr and p,

Y<?,

It

is

^/,

being

required

so as to satisfy

the linear equation,

XYI.
the constants

e, c,

(e

+ f )S? + Se

and the form

0Y^ = Sr +

$f>

of 0, being given.

*
[By a method analogous to that of the Note on page 507, if any three diplanar vectors )3i, 2,
and #3 are chosen, any quaternion function fq may have its vector part resolved along these three
= fiix\ + faxi + /33#3 + #4, in which the coefficients x are scalar functions of q,
vectors, so that fq
and are moreover linear jlfq is linear in q. So for a linear function,

fq

and in

this expression p\,

z,

= faSpiq + faSpzq +

p3

and

constants and determining the function/.

fq =

and

(/3iSai

4-

/8 2

Sa 2 4

this is manifestly of the type,

/3 3

S^ 3 2 + Sptf,

are four constant quaternions involving sixteen scalar

Denoting Sp by

a,

and

\p by

a,

on rearrangement,

PROBLEM OF QUATERNION INVERSION.

ART. 364.]

for this purpose the expression,

Assuming

(4.)

XVII.
in which q

is

557

new sought
XVIII.

= q +

p,

we have the new

quaternion,

= Sr + ^p -fp = S(r -

fgf

equation,

p)

whence

XIX.

S(r-./- l,
l

and

XX.
in

it

p + S(r

./- !

- e
(by supposition) a known vector, and S(r
p) is a known scalar;
l
remains
to
determine
the
unknown
but
constant
only
quaternion, f~ l,

which p

so that

is

or to resolve the particular equation,

XXI.
c

fq =

in

1,

XXII.

which

+ y = f~

\,

being a new and sought scalar constant, and j being a new and sought vector

constant.

Taking scalar and vector parts, the quaternion equation XXI. breaks
up into the two following (comp. X. and XL)
(5.)

XXIII.

1 = Sf(c + y)

=ec + Sty

which give the required values of

XXV. ..c=(e- SeV *)

1

XXIY.

and

y,

V/(<?

+ y] =

tc

fy

namely,

XXVI.

and

whence

7 = c^e

-,

and accordingly we have, by XII., the equation,

XXVIII.
(6.)

of

The problem

tracfor inversion,

/(I

0-

*)

of quaternion inversion is therefore reduced

and

sotai? thereby

but we can

farther, in the elimination of inverse operations,

of

<f/rec

ones.

SfiV* = V^O.

Thus,

if

we

observe that
<j>~

ow advance some steps

and in the

substitution for

them

= m~ ^, as before, and write for

1

abridgment,

XXIX. ..n

anew to that

= me Se^e = f(m

t//e),

ELEMENTS OF QUATERNIONS.

558
so that n

is

we

a new and known scalar constant,

[III. n.

shall have,

6.

by XV. XX.

XXVII. XXIX.,

XXX.

mp = ^Vr

XXXI.

nf

=m-fc;

and

XXXII.

mnq = n\pVr + (mSr

Sc i/YV)

(m - ^e),

an expression from which all inverse operations have disappeared, but which
still admits of being simplified, through a division by m, as follows.
in the term n-jNr of XXXII., the value
(7.) Substituting (by XXIX.),

me - Se

T//e

for n 9

and changing (by

are not obviously divisible

XXXIII.

by
.

\fjVr to nip, in the

XXX.)

terms which

such a division gives,

-m,

nq = (m

- Se
)Sr + e$Vr
i/TW +

<r,

where

XXXIV.
But (by

(r

* -

+ rSc

Se

348, VII., interchanging accents)

XXXV.

= V.

we have

the transformation,

because ^p = Vr, by XIV. or XV.

everything inverse therefore^ again dis
with
this
new
elimination
of
the auxiliary vector p, and we have
appears
;

this final expression,

XXXVI.

.nq =

nf

= (me SE^C) ./
=
(m ^e)Sr +

which each symbol of operation governs all that follows it,

except where a
point indicates the contrary, and which it appears to be impossible further to
in

reduce, as the formula of solution of the linear equation L, with the form
of the quaternion function, fq*

[The following solution

the two equations

is

possibly

fg=(e+
e$q + Se

are found.

e)S?

+ SeT? +

V? =

(e

Equating the scalar and vector parts of

direct.

Sr,

<f!q

and

Operating on the second equation by

given by

Novr

more

XII.

<p-\

= Sr + Vr,

eS# + tyVq

= Vr

and replacing Nq in the

- S^-^Sg- = Sr -

first,

S?

is

seen to be

PBOBLEM OF QUATERNION INVERSION.

ART. 364.]

559

(8.) Such having been the analysis of the problem, the synthesis, by which
an a posteriori proof of the correctness of the resulting formula is to be
given,
may be simplified by using the scalar value XXIX. off(m - fa) and it is
sufficient to show
(denoting Vr by w), that for every vector w the following
;

equation holds good, with the same form XII. of/:

XXXVII.

f(e^

Accordingly, that form of

(9.)

employed

in

Sify w )

= (me - Se

/VeVVew

w.

gives, with the help of the principle

XXXV.,
= -

A.A.A. V11I.

because

fa)

(e

8wiY =

Se ^w, &c.

XXXYI.

and thus the equation

is

proved, by

actually operating with/.

(10.)

As an

we take

if

example,

XXXIX.
in

the particular form,

r = fq =

which

XL.
we have

p =

pq +

<^,

+ a = a given quaternion,

then,

XLI.

=/

/I

whence by the theory of

XLII.

! =

linear

= 2a,

2y;,

and

= 2a,

$p -

2^

rccfor functions,

$ p = 2ap,

\fjp

4<r/o,

= 8

3
,

and therefore,

XLIII.

so that,
dividing

XLIY.

fa

by

= 8
<//

a )q

(^

XXXYI.

formula

8a, the

2a(a*

8fl a,

= a (a - a)Sr + a

n =

a),

16rt

2
(

so on substituting the value of

2a(a + a]q =

Vr -

flS

only remains to replace

tfr

(a

)_(s

by m- ^ in order
1

Bc

-s e y.y,.)

^*
1

to recover

aYr - aY. aVr,

+ a)Vr - Sor,

Sy just found,

(iIt

Sr +

becomes,

or

XLY.

2
a"

XXXVI.]

ELEMENTS OF QUATERNIONS.

560

[III. n.

6.

or

XLYI.

2pq$p = S rKp + pVr = rSp + V(Vp

.

.Vr),

or

XLYII.

4pqSp = 2H8p + (pr

rp)

= pr +

rKp

or finally,

XLVIII.

rr =

-1

Accordingly,

XLIX.
In

(11.)

rKp) +

(pr +

so simple

(rp

Kp

an example as the

ourselves of special methods

r)

last,

= 2r(p +
Kp) = 4rSp.

we may with advantage avail

we may use that which

for instance (comp. 346),

was employed in 332, (6.), to differentiate the square root of a quaternion, and
which conducted there more rapidly to a formula (332, XIX.) agreeing with
the recent
(12.)

L.

XLVIII.

We

.pr-

also

might

rp =

have observed, in the same case

= 2V(V(jo 2 )

qp*

= 4Sp

.V<?)

XXXIX.,

.~V(~Vp .V?)

that

= 2Sp (pq - qp)

.

whence pq - qp, and therefore pq and qp, can be at once deduced, with the
and generally it is possible
same resulting value for q, or for/
as before
th
root of a quaternion.
to differentiate, on a similar plan, the n
1

We

on Linear Functions, of the kinds

above considered, by proving the general existence of a Symbolic and Biqua
365.

shall conclude this Section

dratic Equation, of the form,

which is thus satisfied by the Symbol (/) of Linear and Quaternion Operation
on a Quaternion, as the Symbolic and Cubic Equation,
I

+
4>

was

satisfied

coefficients, n,

by the symbol
n

n",

ri",

(0) of linear

and

m"tf

350,

f,

vector operation

on a vector

I.,

ike four

being four scalar constants, deduced from the function

in this extended or quaternion theory, as the three scalar coefficients m,

And

f
,

m"

at the

were constants deduced from 0, in the former or vector theory.

same time we shall see that there exists a System of Three Auxiliary Functions,
F,

O,

H,

functions,

-^

of the Linear

and Quaternion kind, analogous

and x which have been

to the two vector

so useful in the foregoing theory of vectors,

SYMBOLIC AND BIQUADRATIC EQUATION.

AKTS. 364, 365.]

and

them connected with each

like

function/, by

and with the given quaternion

other,

and useful

several simple

561

relations.*

(1.) The formula of solution, 364, XXXVI., of the linear and quaternion
equation/*? = r being denoted briefly as follows,
>

II.

so that (corap. 348, III

.)

l
nq = nf~ r = Fr,

we may
III.

it

may

write, briefly

and symbolically,

.fF=Ff=n,

next be proposed to examine the changes which the scalar n and the
Fr undergo, when/r is changed to fr + cr or / to /+ c where c is

function

scalar constant

any

to

that

by 364,

is,

and

(2.)

e is

changed

to e

same time changed, according

at the

being
^,
laws of the earlier theory.
;

XII., when

c,

and

to the

Writing, then,

IY. ..fc

=f+c,

ec

c,

tf,

(j>

c,

<j>

and
Y.

we may

= 4 + C +
X
1

\f,

new form

represent the

VI.

nlc

*>

VII.

or

? +

f(

XXXVI.

of the equation 364,

F r,

n e fi r =
l

./C FC = n e

as follows

where
VIII.
*

Fr =

(m c

fa}8r + e^Jfr Se

i/,

Vr -

Wf VeVr,
c

[That a linear quaternion function satisfies a symbolic quartic may be established as follows
it is possible to determine a scalar c and a
quaternion q o ihatfq t eg = 0, the
:

On inquiry whether
two equations

(e

are found

Vg =
+c-

($

by equating

c)~

Se
(4>

c)Sq

+ Se Vj =

to zero

the

0,

scalar

and

and vector

eS^, and, on substitution in the

l

c)

0.

It

may

eS#

it

first,

(0

c)Yq

0,

Hence from the second equation

parts.
appears that c must satisfy the relation

be shown without difficulty, as in the text, that this leads to a

quartic equation in c.
If c n is any root of this quartic, and if a = + cn }- ] e, the quaternion qn = 1 -f a will satisfy
(/+ cn }qn = 0. Corresponding to the four values of cn are four quaternions, and in terms of these
(<J>

any arbitrary quaternion may

Assuming

in general be expressed.

and deriving from

x\q\

this the equations

Yq =

2#nctn,

and

Sq

2,xn ,

and again from these the equation

V<?

HAMILTON

aiSq = #2(02

ELEMENTS OF QUATERNIONS.

01)

a; 3

(o 3

01)

-f

0-4(04

01),

ELEMENTS OF QUATEBNIONS.

562

[III. n.

6.

and

IX.
(3.)

In

this

manner

it is

nc =

ec

- Sg

we may

seen that

X.

i// c

6.

write,

and

XT.
where F, G,

9i,

n",

&+

"&

4
;

ri"

- S Vr (e x + ^)Vr
x

)Sr +

(m"

(e

SetyYr

+ x )Vr - SeV

are four scalar constants, namely,

em - S

n =

n =

given,

Wf WW

Gr = (m - X )Sr +

XIII.

is

ri

c^r -

the scalar #4

Fr = (m - ^ )Sr +

llTr =

and

nc = n + n

are three functional symbols^ such that

H>

XII.

(as in

i//

+ em - St vc
+

on operating by SV(oa

em"

--

- S

01) (a 3

364,

XXIX.)

r
;

ai),

by

040102

and the values of the other scalavs may be written down from symmetry (comp. p. 48). In general
and #4 are uniquely determinate provided the four vectors ct n do not terminate on a common
l
As c varies, the curve traced out by p =
-f c)~ e is a twisted cubic and upon this curve
plane.
the vectors a terminate, and consequently their four extremities do not lie on a plane.

#1, #2, #3,

(<f>

To

verify that p

1
-(((/>

e is

<?)~

SA(<|>

a twisted cubic, the equation

c)- 6

=1,

or

SAi|/ c e

=m

found determining the values of c for the points in which the curve cuts the arbitrary plane
SAp +1 = 0. As this is a cubic equation in c, the curve cuts the plane in but three points.
In general then

is

Operating on

from which

which

is

this

q\

by/+

c\,

has disappeared.

equivalent to I.]

xiqi

and

Similarly operating

by/+

cz

destroys the term in #2, and finally

SYMBOLIC AND BIQUADRATIC EQUATION.

ART. 365.]
(4.)

Developing then the symbolical equation VII., with the help of X.

of c we
XVI. XXI. XXIII.)

and XI., and comparing powers

tions (comp. 350,

new symbolical equa

obtain these

XIV.

"

-fH =

- n

ri

F= n -fG = ri and

563

"f+f*
a

n"f+

-/>

"/

finally,

XV.
which

n =

Ff = rif-

n"f*

+ *

-/S

"f*

only another way of writing the symbolic and biquadratic equation I.

Other functional relations exist, between these various symbols of

is

(5.)

we cannot here delay

operation, which

and

as in the theory of linear

to develop

we may remark

but

mixture of functions with their conjugates (comp. 347, XI., &c.).

(6.) This seems however to be a proper place for observing, that
write, as

temporary notations, for any four quaternions, p,

XVI.
XVIII.

XVII.

=pq-qp;

[pj]

[>r]

that,

vector functions, these usually introduce a

(pqr)

= (pqr) +
(rqlfr + [pr]8j +

q, r, s,

= S

we

the equations,

.jp|>]

|>]Sr

if

and

XIX.
so that [pq]
ternion,

we

(pqrs)

S.jp[gr],

a vector, (pqr) and (pqrs) are scalars, and [pqr]

is

shall have, in the first place, the relations

XX.
XXI.

XXII.

(pqr)

[_pq]

= -

(qpr)

= -

[pqr]

= - [qp],

[qpr]

=
=

(qrp)

\_qrp]

[pp]

=
=

is

a qua

=
;

&c.,

(ppr)

&c.,

[ppr]

=
;

and

XXIII.
(7.)

(pqrs)

In the next

XXIV.

= -

(qprs)

place, if

(qrps)

= -

(qrsp)

&c.,

(pprs)

0.

be any fifth quaternion, the quaternion equation,

= p(qrst) +

q(rstp)

+ r(stpq) + s(tpqr) + t(pqrs)

4C2

ELEMENTS OF QUATERNIONS.

564

which

may

is

q(prst)

= p(qrst] +

r(pqst)

+ s(prqt) + t(prsq)*

analogous to the vector equation,

XXYI.

aSjSyS

/3S 7 3a

-ySSa/3

SSaj3 7 ,

or to the continually f occurring transformation (comp. 294,

XXVII.

Saj3 7

because

is satisfied generally,

XXVIII.

= aSS/3 7 +

it is satisfied

= p,

XXIX.
is

q(prst)

7 + 7 Sj3S,

for the four distinct suppositions,

r,

s,

t.

\_rst]$pq

\_stp~\$rq

[tpt

^Ssq

[prs^Stq,

analogous to this other % useful vector formula (comp. 294, XV.),

XXX.

because the equation

four distinct symbols

SSaj3 7

XXIX.

V/3 7 Sag + VyaS/38

gives true results,

when

it is

operated on by the

(comp. 312),

XXXI. ..S.p,
*

/3Ser

XIV.),

In the third place, we have this other general quaternion equation,

(8.)

which

6.

also be thus written,

XX Y.
and which

[III. n.

S.*.

S.,

S.r,

[Or again as a determinant

Xl

XZ

#3

#4

X$

y\

y*

2/3

2/4

2/5

o,

if
p = w\ + ixi +jyi + fai, &c.]
t The equations XXVII. and XXX., which had been proved under slightly different forms in the
sub-articles to 294, have been in fact freely employed as transformations in the course of the present
Chapter, and are supposed to
familiar to the student. Compare the Note to page 485.
\>e

Compare the Note immediately preceding.

GENERAL QUATERNION TRANSFORMATIONS.

ART. 365.]

Assuming then any four quaternions, p,

(9.)

r, s,

565

which are not connected

t,

the relation,

by

XXX.II.
and deducing from them four

others,

pf(pr8t}

xxxin

(prsf)

/,

by the equations,

=/M

0,

r (prst] = ~

supposed to be a symbol of linear and quaternion operation

on a quaternion, the formula XXIX. allows us to write generally, as an
expression for the f unction /g, which may here be denoted by / (because r is
in

which /is

now

still

otherwise used)

XXXIY.
and

its sixteen scalar

involved in
(10.)

= /? = p Spq + /Sr? +

constants

(comp. 364,

Ssq +

Stq

now

are

(2.))

those which are

four quaternion constants, p

Operating on this last equation with the four symbols,
its

XXXY.
we

..<?

S [rW],

S OT/],

obtain the four following results

S [_fpY],
.

[>W],

xxxvi
and when the values thus found

XXXYII.
are substituted in the formula

quaternion inversion

Irst] (q r

S^?,

Sr?,

Ssq,

XXIX., we have

&tq,

the following new formula of

XXXVIII.
=

for the four scalars,

s f)

(pVT)

[stp~] (q s

(prst)q
)

= (p

[ipr] (q

f)

(prst)f->q

pY) +

[>]

(//rV)

which shows, in a new way, how to resolve a linear equation in quaternions,

when put under what we may call (comp. 347, (1.)) the Standard Quadrinomial Form,

XXXIV.

m. n.

ELEMENTS OP QUATEKNIONS.

566

6.

Accordingly, if we operate on the formula XXXVIII. with /,

attending to the equations XXXIII., and dividing by (prst), we get this new
(11.)

equation

XXXIX.

(p r

)fq =

s t

rs

(q

- r

(q s t

p +
)

tft p r

(q

} ;

whence
fq = g

(12.)

It has

been remarked

by

XXV.

that p,

(9.),

r, s,

in recent formulae,

t,

any four quaternions, which do not satisfy the equation

XXXII.

may

he

we may

therefore assume,

XL.

p =

r =

1,

i,

with the laws of 182, &c., for the symbols

XLI.
and then

it

will be

i t j, k,

(lijk)

= - 2

found that the equations

/-/,

XLIL ../=/!,
so that the standard quadrinomial

form

= /,

k,

because those laws give here,

XXXIII.

*--/*.;

=-//,

XXXIY.

give simply,

becomes, with this selection

of prst,

XLIII.

fq

=/l Sq -fi.&iq - fj
.

and admits of an immediate

expressed (comp. 221) by

Conversely,

if

verification, because

we

Skq,

any quaternion,

q,

may

be

= Sq - iSiq -jSjq - kSkq.

set out

XLY.
which

-fk

Sjq

the quadrinomial,

XLIY.
(13.)

with the expression,

.q =

w+

ix

+jy +

221, III.,

kz,

gives,

XLYI.

..fq = wfl + xfi + yfj +

zfk,

or briefly,

XLYII.

= aw + bx + cy + dz

GENEEAL QUATERNION TRANSFORMATIONS.

ART. 365.]

567

the letters abcde being here used to denote five known quaternions, while icxyz
are/owr sought scalars, the problem of quaternion inversion comes to be that of
the separate determination (comp. 312) of these four scalars, so as to
satisfy
the one equation XLYII.
and it is resolved (comp. XXY.) by the
of

system

the four following formulae

(w(abcd)

XL V 111.

(ebcd)

x(dbcd)

(aecd)

= (abed)

(y(abcd)

z(abcd)

(abce)

being retained.
(14.) Finally it may be shown, as follows, that the biquadratic equation I.,
for linear functions of quaternions, includes* the cubic I ., or 350, 1., for vectors.
the notations

(6.)

Suppose, for this purpose, that the linear and quaternion function,/*?, reduces
itself to the last term of the general expression 364, XII., or becomes,

..fq = jVq,

XLIX.

the coefficients n,

?/,

LI.

and the biquadratic

n",

I.

is

operation

now

is

a vector, by

now

* In like manner

it

n = m,

n"

= (-

XLIX., and

be

/I

=0,

=/

=
;

"

m"

for

ty

m =

which case

= M,

m=

0,

and

m"<{>

when we

also

\]/p

the

the

confine

0, if

p be

or

0,

fact, the cubic gives here (because

(<J>

m"<J>

In

understanding as to the operand.

it

equation includes a quadratic one,

<f>

)/

<,

<f>x

may be any vector, p also

denoted by
when the subject of

said, that the cubic

this

0,

m f- my- +/

ourselves to the consideration of vectors in one plane

=
a line in the given plane for we have then

with

take then, by XIII., the values,

equivalent to that

may

becomes,

LII.

But fq

n"

n = 0,

L.

so that

=
)<t)p

m=

0),

Q;

and therefore
3

(</>

a- be
already the result of an operation with
a line in the given plane.

m"$

if

(f),

+m

}<r

on any vector p

that

is if it be,

as above supposed,

ELEMENTS OF QUATERNIONS.

568
operation

is

a vector

we may

LIIL
which agrees with 351,

\q~

therefore, in the case here considered, write

= (-

iri

tf

<t>

<t>*)p,

and reproduces the symbolical

I.,

cubic,

when

the

of the operand (p) is suppressed.*

[A few additional remarks may be made concerning the

0, and the relations connecting them.

solutions of

"Vq^fq

0,

and of

fq=

It is easy to see,
dratic.

If, for

and its conjugate

in various ways, that
and q are any arbitrary quaternions

3
"/

I.,

satisfy the

ri

same symbolic biqua

instance, q
4

by

6.

LIT. under the form,

this last equation

symbol

[III. n.

"/

- n

= Sj (/ 4 -

n}q

-f

"/

f* - n f + n)q =

and therefore as the quaternions are arbitrary,

(/

~ n

3
"f

+ n

T~

"

/ +

n)q

0.

Again, the same property follows from the equation

e

Se (0

same

as the roots c are the

necessary to have Sq\g

If then

^ = 14

= Se(0 +

c)~ f

(See the Note to page 561.)

Now if, as in the Note just cited, q\, #2,
T
l
= 0,
i
# 2, g a, and q\ are those of \ q~ f q

and q

#3,

are the solutions of

Vq~ fq

0,

Hence if c\ is not equal to &,

/ and for its conjugate /
and, in general, Sqnq n = 0, where n is different from

for
;

c)~

and

it is

01,

and

q\

i,

&c.,

Sqiq

Saia

0.

shows that if vectors are drawn through the origin equal

o 2, o s, and 0*4 will terminate on the polar plane of ai
to ai, 02, 03, 04, and to o i, 0/2, a 3, and cu
In other words, the tetrahedron determined by the
with respect to the unit sphere p 2 + 1 = 0.
In
extremities of 01, 02, 03, and 04 is the polar reciprocal of that determined by o i, a z, a z, and a\.
the particular case in which/ is self -conjugate, 1 + 80102 = 0, and the tetrahedron is self -reciprocal
with respect to the unit sphere or, without reference to a sphere, the tetrahedron may be said to be
orthocentric as the perpendiculars ( or 1 &c.) from the origin on the faces pass through the
Interpreted geometrically this property

corresponding vertices.

Hence, any quaternion q

may

be expressed in the form (compare again the note to page 561)

and
Sqq

may

Sqq

Sqg

be regarded as a canonical form of a function /.

from the properties of the reciprocal tetrahedra that the vector

It is easy to see

V(o3d4

4-

0402

-f

23)

being the negative of the reciprocal of the vector perpendicular on the plane through the extremities
2, a sj and 0*4.]

of a

INDEX TO VOLUME
(

The Numbers refer

to the

I.

Additional references are occasionally given

Pages.

to sections

in which the subjects are specially treated.}

Abstraction of symbol from subject of operation, T,

168; K, U, R, 142; K, S, V, 204.
Irish, first

Academy, Royal

communication on quater

construction of ellipsoid, note,

Anharmonic

of vector and point, 5.

equation of curves,

commutative and associative,

not commutative, 156

is

as

of,

133, 224,

253, 258,

289, 316.

use of signs

in,

108, 123, 256.

5, 6,

of, II. 11.

5,

277.

Alternation and inversion, equidifference of points, 4.

equality of vector-arcs, 144.
equality of geometric quotients, 118.

Amo,

32

on

of surfaces, I.

quaternion, 294, 350.

130,

165,

191

Arc, vector-, II.

i.

9,

cyclic

143; associative addition

II. n.

Arcual sum, 156, 303, 369.

addition.

Area, sign

See Addition of vector-arcs.

of, 18.

directed, 482.

spherical triangle, 364.

principal amplitude, 263.

Amplitude of quaternion,
S

6,

of parallelogram, 246.

for amplitude, 262.

HAMILTON

n.

cyclic, 185, 308.

extended,

Algebraic form, equations

263

I.

representative, 143.

149.

in,

6, 87.

304.

principles adopted from, 108.

Z.n,

in space, 294

cone, 181.

Algebra, imaginary symbol

Am,, or

in.

Apollonius of Perga, locus,

spherical, 406.

Am, symbol

Anti -parallel sections of a cone, 183.

304, 408.

of vector- angles, 406.

paradox

in space, 55.

function or quotient, 15.

176, 204, 207.

sociative,

construction of cubic curve, 37.

of amplitudes, 264.

is

111.

a circle, 297.

of quaternions, definition of, 116.

of vector-arcs,

coordinates, 23

of four points on a line, 15

6.

is

5, 120,

Euclidean, 120.

110.

of quaternions,

i.

vector-, 151, 406.

Acceleration, vector of, 100.

3,

II.

representative, 151.

230.

spherical quadrilaterals, note 392.

Addition of vectors,

264.

of,

Angle of quaternion,

differential of, 458.

nions, note, 160.

Actual vector,

Amplitudes of quaternions, addition and subtraction

3,

262

ELEMENTS OF QUATERNIONS.

note, 120.

spherical polygon, or curve, 368, 370.

spherical cap, 482.

of,

ELEMENTS OP QUATERNIONS.

570

Centre of conic inscribed to triangle, 36.

Arithmetic, illustration of differential from, 434.

Aspect, of plane, 112

of homology, 60.

note, 207.

Associative law for addition of vectors, 7

nions, 207

quater

of ruled hyperboloid, 92

vector-arcs, 304.

for multiplication of

245; II. in.

2,

i,

308

/,
;

159

k,

of involution, 16.

quaternions,

vector

1,

301.

308, 340.
Circle, equation of, square of right radial, 134.

Asymptote, of hyperbola, 34.

inverse of line, 296.

At wood

vector expression for, 417.

machine, 100.

Auxiliary functions, linear vector, 495

quaternion,

560.

vector equation of, 349, 355.

quaternion equation

Ax., symbol for axis of quaternion, 120.

IUV,

replaced

to spherical triangle, 401.

Axes

203.

by UV,

touching

of ellipsoid, principal, 238

of quadric, 536.

of section of quadric, principal, 238, 525.

of,

small

three

circles

on

sphere,

427.

334.

Axis of quaternion, 112; II.

of, 133.

inscribed or exscribed to triangle, 33.

examples on, 121.

equals

96.

to,

See Symbol.

Chasles, referred to in notes, 16, 31, 72, 89, 183, 300,

vectors, 337.

for multiplication, enunciations of, II. in.

Characteristic of operation.

i.

5,

Circular group of four points, 297.

sections of cyclic cone, 184.
ellipsoid, 232, 239.

119; differential

458.

parabola, 34.

points at infinity, note, 516.

logarithmic spiral, 419.

successions, 297, 305, 311.

Cis (symbol), 260.

Barycentres, 85.
Barycentric calculus, notes, 22, 50, 61, 62, 85.
Bicouple, 289.
Bifocal form of linear vector function, 545.

Binomial form of quaternion, 254.

of linear vector function, 498.

Biquadratic equation of linear quaternion function,

560.

Biquaternion, 133, 225, 289, 316.

Class, of a curve, 42, 93.

surface, 88.
Coefficients of vectors, 9.
differential, 444, 99.

Collinear, condition that three points should be, 14,

52

three right quaternions, 247.

quaternions, 210.

Commutative law

for addition of vectors, 6

of qua

ternions, 176, 207.

Biscalar, 225, 289.

does not hold for addition of vector arcs, 156.

Bisecting sides of spherical triangle, triangle, 358.

Bisectors of a triangle, 18.

does not hold for multiplication of quaternions,

Bivector, 225.

Booth, tangential coordinates, note, 40.

147, 159.

Complanar, vectors, 14, 340

proportion

of,

250, 256.

termino-, 45, 344.

points, 14, 45.

quaternions, 116, 211, 250.

Calculus, Barycentric, notes, 22, 50, 61, 62, 85.
of functions, 205, 202.

of finite differences, 83.

Cap, area of spherical, 482.

Carnot, on transversals, note 65, note 377.
Cartesian expressions, 242.
coordinates, 248.

Cassinian, 281, 285.

Complanarity, sign
condition

of, 117.

of, of

quaternions, 148.

of vectors, 14, 338.

of points, 14

45, 52.

Complex mean of vectors, 85.

Composition of two quadrantal rotations,
Concurrence of three

lines, 18.

of four planes, 57, 342.

149.

INDEX TO VOLUME
Condition of contact of line and sphere, 224, 427,

Construction, points

plane net, 17, 22.

of, for

for net in space, 61.

428.
of parallelism of

two

vectors, 10, 194, 325.

of cubic curve, 38.

perpendicularity, 180, 325, 345.

of ellipsoid 234.
of fourth proportional to three diplanar vectors,

collinearity of three right quaternions, 247.

complanarity of three vectors, 14, 338, 345.

362.
of series of spherical parallelograms, 390.

of versors, 148.

Continued proportion of complanar vectors, II. n.

of concurrence of three lines, 18.

on a

that three points should lie

quinary symbols, 52.

that four points should

\ 2,

256, 251.

homosphericity, 352, 354.

line 12, 14

in

Convention respecting sign of area, 18

of volume, 48.

sense of rotation, 119; notes, 111, 152, 369.

in a plane 14, 45

lie

in

position of operator relatively to operand, 110, 147.

Convergency of

quinary symhols, 52.

he termmo-collinear,

that three vectors should

series, 269, 424.

Coordinates of a couple, 254.

anharmonic, 23, 55

14, 343.

that four vectors should be termino-complanar,

of a plane, 60.

idea of, foreign to quaternions, note 112.

tangential, 40.

45, 344.

that four planes should concur, 57, 342.

Cone, cyclic or quadric, 95, 181, 304.

Cartesian, 248.

Cosine of quaternion, 275, 424.

expressed by S, 181.

Couple, 254.

vector expression for, 95, 101.

Criterion as to nature of conic inscribed to triangle, 34.

of revolution, 183.

expressed by L, 121

cubic, vector expression for, 95.

Congruence, formula of, for quinary symbols, 51.

Conic, anharmonic equation of, 32.
triangle, 32, 33, 34, 36.

Conical rotation, 154, 359, 398, 429.

ellipse,

95

Cubic curve, anharmonic construction

ellipsoid, 95.

of linear vector function, symbolic, 494.

Curve, vector equation

of, 94.

with conjugate point, 37, 41.

Cyclic form of linear vector function, 520, 528, 535,
cubic,

cone, 181, 309.

planes, normals, of cone, 183, 549.
of ellipsoid, 232, 235.

harmonic, 16.
point of cubic, 41.

arcs, 185, 308.

quaternions, 115, 123.

quadrilateral, 296, 347.

of vector, 346.

of linear vector function, 485.

of linear quaternion function, 555.

Conjugation, characteristic of,

K, 124.

equation of (linear vector function), 485.

(linear quaternion function), 555.
(pole

and

of, 38.

cone, 95.

538.

spherical, 182, 239, 303, 309.

Conjugate diameters,

Cube -root of quaternion, 256 principal, 257.

Cube-roots of unity, nine, 291.
;

S, 180.

equation of one sheet of, 121.

tangent, to sphere, 225.

and

571

I.

polar), 229, 428.

under sign S, 350.

Cyclical permutation of vectors
of quaternions under sign S, 248.
law of

i,j, k, 159.

Cylinder, of revolution expressed

by TV, 195

by V,

199.

tangent to sphere, 201

spheroid, 201

ellipsoid,

202.

Constants determining a linear vector function, 486,

530

quaternion function, 556,

vector-, of ellipsoid, 201, 236.

or invariants of linear vector function, 491.

Constituents of a quaternion, 242.

Decomposition of vector, 193, 338, 339.

= T? ,Vq, 169
q
quaternion q
;

Sq +

193.
4

ELEMENTS OF QUATERNIONS.

572
Definite integral, 368, 482.

Differentiation, partial,

101

Definition of vector, 3.

sum

of vectors, 7

of quaternions, 109.

addition and subtraction of vectors,

differential of

a vector, formula

for,

35

of a vector,

4,

451.

Dimensions, theory of, applied to linear functions of

a vector, 488.

of, 98.

Diplanar quaternions, 116.

Direct similitude, 115.

simultaneous differentials, 431.

,

symbol

of a quaternion, 446.

examples of quaternion, III. n.

5,

of quaternion, note, 35.

422.

circular succession, 297.

Deformation of sphere, 232.

theorem of geometrical, 525.

Directed area, 482.

Degenerate quaternions, 120, 178, 333.

Directions of linear vector function, 508, 515, 517.

Degree of plane curve, 32

Distributive property of multiplication of vectors

Direction, relative, 110, 138.

of surface, 87.

of curve, 93.

DeMoivre

De Morgan,

by

coefficients, 9.

theorem, 264.

of quaternions, 212, 219.

of sign I, 206

note, 278.

Denominator, reduction of two quaternions to a com

Division of vector

mon, 116.
Depressed equation of linear vector functions, 501,
505.

K, 176, 207

S, 185

V, 204.

of differential, 441.

by

parallel vector, 10.

of quaternions, defined, 109, 116.

homographic, 16.

Derivative of a vector, 99.

Descartes, notes 394, 404.

Effective vector, 3.

Desk, illustration of quaternion by, 113.

Elimination of a vector, 342, 355.

of differentials, 448.

Developable surface, 100.

Development of functions of a quaternion, 475.

Diacentric sphere, 234, 241.

Diameters, conjugate of

of

two quaternions,

finite,

section of cylinder, 196, 199, 418.

ellipse,

Difference of two points, 3

II.

of
i.

95

of ellipsoid, 95.

two

Ellipsoid, equation of, 201, 230, 313, 325, 328, 535,

537

vectors, 5.

bifocal, 545

focal, 201, 230,

principal axes of, 238.

sections of, 238.

Differential quotient, 443.

homologies

of,

and sphere, 315, 232.

201, 236. 537.

of implicit functions of quaternions, 484.

vector-constants

of n th root of a quaternion, 560.

tangent right cylinder

Differential of vector, I. HI.

of square root, 452

of a product or power, 451,

of

K0, V?, and S0, 455

U0, 456.
finite, 99,

432.

coefficient, 99, 444.

quotient, 444.
Differentials

and differences, 434.

distributive property of, 441.

successive, 465, 479.

simultaneous, 431.

to,

202.

Elliptic logarithmic spiral, 419.

Equality of points,
;

of,

vector expression for, 95.

7, 96, 98.

476.

439

cyclic,

construction of, 234.

differentials, 434.

0-i,

circular sections of, 232, 235.

102.

successive, 479.

of

531

527.

12, 176.

Differences, finite, equations in, 84.

and

Ellipse, vector expression for, 95, 417.

of

Tq and

3, 13.

vectors, 3.

quaternions, 109, 115.

vector-arcs, 144.

121
Equation of loci involving signs Ax. and Z,
T, 165,
V, 195, 199
K, 127 S, 180, 190

167, 190; U, 142.

powers of a vector, 417.

anharmonic of curve, 32 local and tangential, 39.
;

of surface, 87.

INDEX TO VOLUME
Equation, vector-, of curve, 94
See Sphere, Ellipsoid, &c.

Formula

of surface, 94.

(A), 160.

of congruence, 51.

of differential, 98, 438.

exponential, for spherical triangle, 404.

of relation between

in finite differences, 84.

of algebraic form in quaternions 277

of,

573

I.

w 4 roots

+ and

-, 5.

of commutation, 7.
of association, 7.

292.

of conjugation

and

(poles

229

polars),

(linear

of collinearity of three right quaternions, 247.

two

of perpendicularity of

functions), 485, 555.

of second degree, homogeneous and vector, 508.

parallelism of

two

vectors, 325.

vectors, 325.

complanarity of three vectors, 338, 247.

Formulae of spherical trigonometry, fundamental, 400.

of six segments, 21, 18.

symbolic, for linear functions, 494, 560.

Four constituents of quaternion, 242.

Equi-difference of points, 4.

of, linear, 15

circular, 297.

Euclid, angle, 120.

points,

Euler, note 244,

points, complanarity of, 14, 45.

Examples, geometrical, on signs Z, Ax. 121

142

S,

vectors, linear equation between, 44, 338.

Evolutionary quaternion, 295.

K, 296;

group

K, 127

proportion of, 250.

180, 190; T, 165, 167, 190; U,

V, 195, 199.

Fourth proportional to three vectors, complanar, 250,

293 diplanar, 356 rectangular, 377.
;

depending on power of vectors, 417.

unit in space, 394, 380.

of quaternion differentiation, 451.

Excess, spherical, 364.

Exponential form for sine and cosine, 266, 274.

under Ponential.

See

Fraction, geometric, 107.

Function, anharmonic quaternion, 294.
of vectors, a quaternion, 332, 394.
of three vectors, linear, 338.

equation for spherical triangle, 404.

transcendental, of quaternions, 421, 453.

transformation of Taylor

trigonometrical, of qiiaternions, 424.

Exponents,

s series,

473.

scalar, 264.

linear.

quaternion, 274, 421.

Exscribed or circumscribed conic, 36.

Extensions of

algebraic

See Linear functions.

Functional notation, 205, 202.

notation,

5,

Functions, calculus
6,

108,

123,

of,

205, 202.

Fundamental formulae of spherical trigonometry, 400.

256.

Gauche
Factor, or operator, 108, 135.

quadrilateral, 82.

hexagon in sphere, 305.

Factorials, notation of, 476.

Generation, double, of ruled hyperboloid, 90.

Finite difference, 102.

of ellipsoid, 241.

differences, equation in, 84.

Generatrix of ruled hyperboloid, 89.

differential, 99, 432.

Geometric quotient, 107.

inversion and alteration

Five quaternions, identical relation connecting, 563.

Fluxions, note 431.

Z, 121;

Focal property of sphero-conic, 310, 393.

535, 538.

equations of surfaces of the second order, 535.

lines of cones, 545, 549.

Foci of involution, 16.

I.

in.

illustration

of

space,

531,

S,

180

Ax. and

K, 127; R, 296; U, 142; S and T, 190.

nets, plane, I.

relations, notation for, 310.

transformation of linear vector functions,

of, 118.

Geometrical examples on V, 195,199

vectors, 47.

n.

3,

20

I.

n.

5,

29

in

4, 61.

differential,

436

of

ratio

of

vanishing quantities, 470.

deformation, theorem
"Geometrie superieure,"

of,

525.

of Chasles,

notes, 16, 72, 89, 300.

referred to in

ELEMENTS OF QUATERNIONS.

574
"Geometrie

de

of Carnot, referred to in

position,"

five quaternions, 563.

Graves, J. T., note, 276.

six spherical arcs, 377.

Gravitation,

344,

375,

four vectors, 376.

Graves, C., note, 308.

Group

337,

426.

note, 335.

Grammarithm,

three vectors,

Identity connecting

notes, 377.

Newton

law

of, 99.

of four points, circular, 297.

See Anharmonic.

on square root of quaternion (p^q*)^, 315.

and difference, 434, 436.

Illustration of differential

ratio of vanishing quantities, 470.

Guide-points, 239.

a quaternion, 113.

Half -line or ray, equation

of, 121, 142.

Handle, versor compared to

Harmonic

a,

336

Imaginary, intersections, note, 87, 223.

roots of equation of algebraic form, II. n.

note, 345.

mean

of

two

vectors, 231, 298

of

interpreted as geometrically real, 133.

300.

of algebra, 224.

property of quadrilateral, 20.

of pole and polar, 229.

roots

and

directions of linear vector functions,

515.

polar of point to a triangle, 21.

Helix, 419.

Imponential, 274.

Heptagon, inscribed to sphere, 354.

Indeterminateness of interpretation of

Hexagon,

Higher Plane

Salmon

Curves,"

notes, 37, 40, 41,

206, 258, 262, 278.

to,

Hodograph, 99.

Homographic

null quaternion, 120.

Index of right quotient,

symbol

nets in space are, 79.

sum

and sphere, 315, 232.

ellipsoid

centre of, 60

Homosphericity, equation

i.

of,

product

122.

i.

4,

331

of,

of,

329.

circular points at, note, 516.

354,

Inflexional tangents to cubic, 37.

Hyperbola, 33.

Inscription of polygons in a sphere, 354.

Integral, definite, 368, 482.

Hyperbolic paraboloid, 93, 96.

Hyperboloid, ruled, anharmonic equation

scalar

of, 88.

Hypotenuse, proof of theorem of square

Interpretation of a product of vectors as a quaternion,

of, 212.

III.

symbol for index of right quotient, 187.

of

distributive, 206.

of

242.

of,

of a right
4,

planes, 338.

Inverse or reciprocal of a vector, II. n.

quaternion, with

331

two

Invariants, quaternion, 491, 506.

early use of, notes, 160, 345.

i.

321, 337, 394.

of line and plane, 47.

10, 157, 344.

vector in terms of, 344.

index, III.

1,

133, 253.

real, 220.

quaternion in terms

Identification

i.

V^T,

Intersection, imaginary, note, 87, 223.

= IUV, 203.
IV = V, 335.

Ax.

of, II. i.

and vector, 482.

Integration, 482.

vector expression for, 95.

laws

note, 193.

206.

Hydrostatics, 483.

i,j, k,

175.

Infinity, line at, 27.

plane of, 60.

of,

5,

Indices of right quotients, quotient

property of ruled hyperboloid, 89.

Homologies of

II.

of, 187.

equals right quotient, III.

division, 16.

figures, plane nets are, 31

is

133.

directions of linear vector function, 501.

History of quaternions, notes

I,

1,

vector-arcs, 145.
s,

42.

Homology,

V -

of versor, 139.

spherical, 303.

inscribed to sphere, 305, 354.

"

5,

289.

section, 16.

note, 193.

a quaternion, 122
its

own

similitude, 115, 129.

of line, 296

7,

293.

differential of, 439.

of circle, 296.

INDEX TO VOLUME

575

I.

Linear vector function, inversion

Inversion and alternation, 4, 118, 144.

form

geometrical, 293.

of,

of,

485.

standard trinomial, 486

monomial, 501, 505

binomial,

of linear vector function, 485.

498

of linear quaternion function, 557.

528, 535, 538; bifocal, 545; focal, 531,

cyclic,

520,

533, 535, 538.

Inversoi , 135.

conjugate, 485.

semi-, 135.

self -con jugate, 513, 519, 525.

Involution, 16.

non-conjugate part of, 492.

495.
auxiliary (if,), 485
( x ),

double, 72.
in space, 295, 300.

derived, 561.

symbolic cubic

introduced, 124.

differential of

of,

494.

with equal roots, 500.

-examples on, 127.

unequal real

K?, 455.

imaginary

roots, 509.

roots, 515.

Lacroix, referred to in notes, 473.

with depressed equation, 501, 505.

Lagrange, referred to in notes, 441, 472.

number

Law
"

Norms, 173, 244.

of the

Lectures"

powers

referred to in notes, 160, 206, 219, 304,

311, 315, 345, 354, 380, 479, 525.

See Convention respecting rota

Left-handed, 111.

of,

of constants in, 486, 530.

491.

system of lines and planes related

dimensions of, 488.

number

of constants in, 556.

Lemniscata, 286.

standard quadrinomial form

Length, 163.

symbolic biquadratic

of

two

Limiting

ratios,

note 311.

Quaternions,"

Loci,

equations

L and

Line,* expressed by
142 V, 195.

Ax., 121

K, 127

of,

involving Z, Ax, 121

K, 127;

167; U, 142;

V, 194, 199.

U,

powers of a vector, 417.

anharmonic equation
symbol of ternary, 56

of,

26

coordinates

anharmonic, 57

of, 27.

quinary,

62.

Locus, Apollonian, 130, 165, 191.

Logarithm of quaternion,
Logarithmic

vector expression for, 15, 94

to,

335

II.

IT.

4, 288, 275, 421.

spiral, 418.

equation of, 195.

Mean

point, 5.

point, projection ot, 19, 81.

at infinity, 27.

of gauche quadrilateral, 82.

half-, 121, 142.

complex, 85.
of partial systems, 83.

rational, 29.

proportional between

reflexion of, 129.

intersection of,

565.

of,

560.

R, 296; S, 180, 199; T, 165,

469.

number added

of,

Local equations, 39.

lines, relative, 111.

on

512.

quaternion function, 485, 555.

tion.

"Letters

to,

and plane, 47

and sphere, 220,

223,
of intersection of

two

Lines, relative length and direction of two, 111.

Linear equation between two vectors, 12.

of

n vectors, simple, 81

of n, 300.

complex, 85.

of linear vector function, 545.

Mobius, referred to in notes, 22, 31, 65, 79, 61, 52,

66, 77.

of linear vector function, 501, 505.

equations of circle and sphere, 355.

four vectors, 44, 338.

five quaternions, 563.

The word

vectors, 251.

vectors, harmonic, 298

Monomial form

three vectors, 13.

two

Mixed transformation

planes, 338.

two

of

Mourey, note, 278.

line is

frequently used for vector in the Elements.

ELEMENTS OF QUATERNIONS.

576

Oblique cone, 183; note, 181.

Multiplicand, 147, 110, 159.

Multiplication of vectors

quotient, 321, 337.

coefficients, 9.

by

of quaternions by coefficients, 119.

Ohm,

note, 276.

Opening, spherical, 366.

definition of, 116.

is

not commutative, 147.

Operation, characteristic of.

is

doubly distributive, 219, 212.

Operations, calculus of, 202.

is associative,

245, II. in.

2,

308.

Operator or factor, 108, 135.

reduced to multiplication of versors, 172.

of versors illustrated

by

notes, 548, 554.

vector-arcs, 147.

Opposite quaternions, 126.

337.

Order of curve, &c.

of i,j, k, 159.

of vectors

See Symbol.

vectors, 3.

is associative,

Multiplier, 147, 110, 159.

See Degree.

when one

of factors, indifferent

is scalar,

119.

generally important, 147, 153, 158.

N, symbol
Ng,

for

norm, 130.

Origin of vectors, 12.

differential of, 455.

Oval, 279.

Negative unity, square of right

square root

of,

radial, 132, 203.

P, symbol for ponential, 268.

133, 224, 253, 289.

square of vector, 345.

Parabola, inscribed to triangle, 34.

Paraboloid, ruled, anharmonic equation

Net, plane geometrical, 20, 29; in space, 61.

Nets are homographic figures, 31, 79.

Paradox, apparent, 149.

Newton, Principia, note, 97, 431, 471.

Non- commutative multiplication, 147.

Parallelepiped,

addition, 156.

Non- conjugate

part of linear vector function, 492.

Parallelogram, area

130, note 128.

246.

of,

Parameters in vector equations, 94.

Partial differentiation of a vector, 101.

differential of, 455.

Normal,

247, 338.

of,

spherical, 388.

of sum, 189, 219, 476.

Norms, law of

volume

Parallelism, condition of, 325.

Non-scalar, 110.

Norm,

of, 93.

vector expression for, 96.

of a quaternion, 446.

the, 173, 244.

sign for, 35.

cyclic, of cones, 183.

successive, 479.

of ellipsoid, 232, 235.

Parts, right, or right quotients, or vector parts, distri

Notation, functional, 205, 202.

butive addition

simplification of, 334.

of,

Pencil of lines, 23, 307

extended meaning of algebraic,

5, 6, 108, 123, 256.

for focal relations, 310.

204.
planes, 57.

Pentagon, inscribed to a sphere, 351.

Permutation, cyclical, of quaternions under S, 248,

of factorials, 476,

350; of i,j,

See Symbol.

k, 159.

Perpendicular, from point on line, 179, 194, 427.

Null quaternion, 125, 139.

on a plane, 180, 342, 353.

vector, 3.

Perpendiculars, of triangle, spherical, 217.

vector-arc, 146.

Perpendicularity of two vectors, condition

Number

of points of construction, 22, 73.

Plane, expressed by symbol Z, 121

signless, 111, 170.

added to

line, 335.

127; T, 165, 167; U,143;

of constants involved in linear vector

function,

note, 130.

"Norm"

borrowed from theory

vector expression for, 24, 94.

anharmonic equation

486, 530; quaternion, 556.

Numbers, name

of,

325,

345.

of,

quinary symbol

of,

for, 53.

scalar equation of, 180.

56.

Ax.,

S, 180.

121

K,

INDEX TO VOLUME

577

I.

Pressure, hydrostatic, 483.

Plane, through three points, 344.

Principal root of quaternion, 259.

polar, of point to sphere, 228.

geometrical nets, 20, 29.

cuhe-root, 257.

rational, 54.

amplitude,

of quaternion, 111.

axes of ellipsoid, 238.

and

Principles adopted from algebra, 108.

line of intersection of two, 338.

system

of, related to linear

Point, symbol of

a,

Product, of quaternions, defined, 109.

vector function, 512.

quinary, 51

ternary, 25

qua

of

conjugate

of, 173.

differential of, 451.

reciprocal of, 173.

ternary, 55.

mean, of

263.

Principia, referred to in notes, 97, 431, 471.

of ellipsoid, 232.

section of quadric, 525.

line, intersection of, 47.

Planes, condition of concurrence of four, 57, 342.

cyclic of cones, 183

am

scalar of, 187, 245.

triangle, 19.

gauche quadrilateral, 82.

tensor, 171.

vector of, 245.

in general, 81.

addition of vector to, 5.

versor, 171.

of

two quaternions, 109, 116, 171.

equi- difference of two, 4.

of

two

condition for three on a line, 14.

of vectors, 337, 346, 394.

Points, difference of two, 3.

four in a plane, 14, 45.

five

of

on a sphere, 354.

first

types, 75

group of

diagram

linear, 15

62

tahle of

circular, 297.

in space, 54, 79

types

of, 55.

mean

point, 19.

of closed figure, 8.
of line, 179.

commutative, of addition,

of plane curves, 35.

207, 304

6,

176, 207.

harmonic of quadrilateral, 20.

inscrihed in sphere, 347.

Proportion of vectors, 118, 175, 250.

of angles of, 415.

exponential equation for spherical, 404.

continued, of vectors, 251, 256.

mean, of vectors, 251.

Proportional, fourth, to three complanar vectors, 250.

Polynomial equations, 277.

Poncelet, Traite des Proprietes Projective, referred to
in notes, 60.

diplanar vectors, 362.

rectangular vectors, 377.

Ponential of a quaternion, II. n.

268.

Provector,

3,

146.

Pyramid, volume

Position, vector of, 100.

Positive or signless numher, 170

note, 111.

of, 247, 338.

sign of volume, 48.

vector, a quaternion, 396, 399.

equation of loci involving a, 417.

transformations of, 420
development of, 476.
;

differential of a, 451.

Powers of quaternions, 264, 274, 421.

of linear vector functions, 491.

HAMILTON

distributive, of multiplication, 9, 212, 219.

Polygon, area of spherical, 368.

sum

7,

of multiplication, 159, 245, 308, 337.

Pole and polar of a sphere, 228.

Power of a

Projection of

Property, associative, of addition,

circular, at infinity, note, 516.

spherical

vectors, 333.

stereographic, 311.

in space, 294
;

321.

of sides of an inscribed polygon, 347.

space,

of, 78.

rational, in a plane, 29

1,

of sides of a triangle, 348.

construction in space, 61.

of second construction in

two rectangular

i.

of indices, 329.

of plane construction, 17, 22.

of

vectors, interpreted, III.

ELEMENTS OF QUATERNIONS.

Quadrantal rotations, 149.

triangle, 377.

Quadratic equation in quaternions, 281.

vector equation, 508.

ELEMENTS OF QUATERNIONS.

578

Quaternion, functions of a single:

Quadric, ruled hyperboloid, 88, 95.

focal

and

cyclic equations of

a,

535.

axes of plane section of

See Ellipsoid

a,

imponential, 274.
inverse, 122.

bifocal equation of a, 545.

linear function of, 485, 555.

525.

and Linear vector function.

logarithm, 275, 421.

opposite, 126.

Quadrilateral, cyclic, 296, 347.

ponential, 268.

of a plane net, 31.

power, 268, 274, 421.

gauche, 82.

harmonic property

reciprocal, 122.

of, 20.

area of spherical, 368.

root, 259.

product of sides

sine, 275, 424.

347.

of,

Quadrinomial form for quaternion, 242.

linear quaternion function,

square, II.

i.

132, 141, 170, 187.

7,

square root, 188; differential of , 452.

565.

Qualitative element of a quaternion, note 167.

Quantitative element of a quaternion, notes 167, 138.

transcendental, functions of

Quaternions, operations on two

Quantities, ratio of vanishing, 470.

addition, 176, 116.

Quaternary symbols, 55.

conjugate of product, 173.

Quaternion, as a factor or operator, 135.

division, 116.

as a fourth proportional, 357, 362.

as a

power

of a vector, 396, 399.

as quotient of

two

is a,

quotient, 109, 116, 171.

332, 394.

reciprocal of product, 173.

subtraction, 116.

oblique quotient, 321, 337.

product of vectors, 321, 337.

number added

multiplication, 116, 173.

product, 109, 116, 171.

vectors, 110.

general function of vectors

a,

sum, 176.
Quaternions, two

to line, 335.

scalar plus vector, 11, 335.

are collinear, 116.

square root of a positive plus square root of a

negative, 203.

condition of complanarity, 148.

+ V?, 193.
TqUg, 169.

equality of, 109, 115.

qp, 274, 421.

S^

Quaternions, special cases of

Quaternion, anharmonic, 294, 296, 350.

binomial, 254.

radial, 131.

development, 473.

right part, 192.

equations, 243.

right quotient, 121.

algebraic, in complanar, 277.

scalar, 120.

general algebraic, 292.

vector, 335.

evolutionary, 295.

null, 125.

versor, 143.

Quaternions, parts of

exponents, 274, 421.

identities, 426, 563.

amplitude, 262.

integration, 482.

angle, 119.

invariants, 491, 506.

axis, 119.

inversion, 557.

four elements of, 112, 113.

quadrinomial, 242.
Quaternion, functions of a single

index of right, 122.

:

plane, 111.

conjugate, 123.

scalar, 177.

cosine, 275, 424.

tensor, 167.

cube root, 256.

vector, 192.

421, 453.

INDEX TO VOLUME
Quaternions, parts

Rectangular vectors, fourth proportional to three, 377.

137.

of, versor,

Quaternions, any number of

versors, multiplication of, 149, 157

addition, 207.
collinear, 210.

513, 519, 525.

cyclical permutation

under

S, 248.

function, 528.

Reduction of two geometric quotients to a

multiplication, 245, 219, 301, 308.

Quaternions, calculus of

common

denominator, 116.
of quaternion multiplication, 171.

applications to plane trigonometry, 178, 197, 214,

208.

of a quaternion to a

power of a

vector, 399.

Reflexion of a line, 129, 358.

spherical trigonometry, 216, 209.

See History.

history of.

successive, 361.

Regression, edge

Letters on, note 311.

s series

adopted

to,

four vectors, 44, 376.

473.

five vectors, 47.

Quinary symbols, 50.

six spherical arcs, 377.

types, 55.

five quaternions, 563.

Relative length and direction, 111, 138.

Quotients, of vectors, 107.

Remainder of a

right, 121.

quotient

sum

of,

index

100.

of, 93,

Relation connecting three vectors, 337, 344, 375, 426.

See Lectures.

Lectures on.

of,

206.

of,

arc, 143.

122, 331.

point, 143.

Resolution of vector along and perpendicular to given

radial, 131.

line, 193.

along three given lines, 338.

geometric, 109.

two, with

474.

series,

Representative angle, 151.

175.

oblique, 321, 337.

common denominator,

perpendicular to three given planes, 339.

109, 116.

inversion and alternation of, 118.

differential, 443.

of quaternion.

See Decomposition.

Revector, 3.

two quaternions, 109, 116, 171.

scalar of, 187.

Re versor,

139.

Revolution, cone of, 183

spheroid

R, symbol for reciprocal, 141, 293, 328.

Right-hand

Radial quotient, 131.

Right

of,

cylinder of, 195, 199.

201.

rotation, 119.

part, 193.

quotient or quaternion, 121.

right, 132.

Radical plane, 328.

index

Ratio of vanishing quantities, 470.

identification of, with its

Rational points, lines and planes, 29, 54, 79.

Ray

vectors, 333.

transformations of self -conjugate linear vector

diplanar, 116.

of

self- con jugate linear vector function,

system of

complanar, 250, 116, 211.

Taylor

579

I.

or half -line, equation of, 121.

Reality of roots of self -conjugate function, 513, 519,

525.

i.

4,

331

own

index, III.

note, 193.

quotients, quotient of, 175.

index of

sum

of,

206

radial, 132.

Reals, 11, 258.

versor, as unit-line, 335.

Reciprocal of quaternion, 122; development of, 475;

versor of, 138.
vector, II. n.

of, 122.

7,

293.

Reciprocity of forms, 547.

Rectangle, spherical, note, 388.

Root of quaternion equation of n th degree, 292.

square, of quaternion, 188, 452.
nine of unity, 291.
cube, of quaternion, 256
;

n ih roots of unity, n geometrically

n(n

1)

real, 259.

imaginary, 290.
4

ELEMENTS OF QUATERNIONS.

580
Eoot, principal, 259.

of negative unity, imaginary symbol of algebra,

convergency of, 269, 424.

remainder of, 474.

Series,

of spherical parallelograms, 388.

224.

Sign of area of plane triangle,

of zero, 316.

of equations of algebraic form, II. n.

5,

277.

Rotation, convention respecting, 111, 119; notes, 49,

of

volume of a pyramid,

of spherical area, 370.

Signless number, tensor

369.
conical, 154, 359, 398.

a,

170, 111.

Similitude, 115.

Simplification of notation, 334.

of solid, 361.

Rotations, composition of

two quadrantal, 149.

Simultaneous

differentials,

431.

Sine, exponential form for, 266, 274.

in, 415.

theorem

18.

48, 342.

Ruled hyperboloid, 88,

95.

of a quaternion, 424.
Solution of the exponential equation, 409.

paraboloid, 96, 93.

Space, scalar unit in, 380.

symbol

S,

for scalar,

examples on

177.

S, 180, 190.

V, 199

distributive, 185.
cyclical permutation under, 248, 350.

i(l

=
S- 1

+ K), 177.
-V, 193.

W.

Sadleir, Rev.

Scalar,

in terms of,

of,

N, T,

165, 167

S2

K, 130; S, 180;
- V a 200.
,

square of a vector, 327.

Apollonian locus, 130, 165, 191.

intersection of right lines and, 220, 223.

Sg, differential of, 455.

394.

of,

monomial, 355.

202.

Salmon

symmetry

Sphere, equation

of cone and, 181.

D., note 375.

Higher Plane Curves, notes, 37, 41, 42.

why

so called,

1 1

of ellipsoid and, 240.

201.

tangent cylinder

to,

tangent cone

225.

to,

of a quaternion, 177, 186.

poles and polars of, 228.

of a product, 245.

homologies of ellipsoid and, 315, 232.

of a sum, 185.

diacentric, 234, 241.

unit in space, 380.

Spherical parallelogram, 388.

integration, 482.

addition, 406.

plus vector equals quaternion, 11, 335.

area, 364, 368, 482.

Screw

surface, 419.

cap, 482.

conic, 182, 239, 309, 310.

Section, harmonic, 16.

homographic, 16.

excess, 364.

Sections of cone, 181.

hexagon, 303.

of cylinder, 196, 199.

opening, 366.

of ellipsoid, 238.

polygon, 414.

of quadric, principal axes of, 525.

sum, 156, 406, 415.

circular, of cone, 183.

of ellipsoid, 232.

Segments, equation of

six, 18, 21.

tetragonometry, 417.
See Triangle.
trigonometry, 209, 216, 325, 358, 400.

triangle.

Self -conjugate functions, 513, 519, 525.

Spheroid of revolution, 201.

Semi-inversor, 135.

Spiral, 418.

Sense of rotation.

Square of a quaternion, II.

See Convention respecting rotation.

Series, ponential, 268.

exponential, for sine and cosine, 274, 421.

Taylor

s,

102, 473.

187.

of a vector, 327, 345.

of a right-radial, 132.

i.

7,

132,

141, 170,

INDEX TO VOLUME

Symbol of inverse

Square root of zero, 316.

of-

similarity

for vector-arc

133, 224, 253, 289.

1,

581

I.

(r\),

<x

),

115.

144.

for spherical addition [(+)], 406.

of quaternion, 315, 367, 452.

on hypotenuse, proof of theorem of, 212.

Standard form for quaternions, binomial, 254

v/
;

quad-

rinomial, 242.

indeterminate, 133; uninterpreted, 289.

Symbols, quaternary, 55.

quinary, 50.

for vectors, 344.

ternary, 25, 56.

for linear vector functions, 486.

of algebra, imaginary, 224.

extended use

Stereographic projection, 311.

and

of, 5, 6, 108, 123,

Sub-contrary sections of a cone, 183.

4-

Subtraction of vectors,

notes on, 262, 334.

5.

256.

formula of relation between,

5.

of quaternions, 116.

Symbolic cubic of linear vector function, 494.

of amplitudes, 264.

Succession, direct or indirect, 297.

biquadratic of linear quaternion function, 560.

Symbolical or imaginary, 133, 289.

Successive differentiation, 465, 480.

Symmetry

Sum

Syn typical

of vectors, 7.

System of

arcual, 156, 369*.

spherical, 156, 406, 415.

tensor and

norm

lines

of a, 189, 219.

Supplementary

three right versors, II.

focal

Tg, differential

See Quadric and Ellipsoid,

and cyclic forms of equations

developable, 100.

L, 120;

u>

I,

Am Am n
,

263.

263.

Tangents, to sphere, 481.

Tangential coordinates, 40.

equations, 39.

replaced

334.

by UV,

187; suppressed, 334.

j, k, note,

K,

535.

cylinders, 201.
;

260.

Ax. 120

of,

Taylor

s series,

102, 473.

Tension, act of, 164.

160.

Tensor of quaternion, 167.

124.

of vector, 163.

of a scalar, 168.

276.

N, 130, note, 128.

of a sum, 189, 219.

P, 268.

development

R, 141.

Term,

S, 177, 166, note, 127

S-

O,

UO, 140.

V, 193, note, 124

of complanarity

|||

of intersection

V^O, 202.
117.

),

(OA BC),

of focal relation

of similarity

oc),

202.

of, 476.

3.

Termino-collinear vectors, 14, 343.

complanar vectors, 45, 344.

Ternary symbols, 25, 56.

T, 163, note, 131.

U, 136

of, 456.

Table of types of points of construction, 75.

Tangent to a curve, 97.
cone to a sphere, 225.

screw, 419.

Symbol, Am. 262

examples on, 165, 167, 190.

of, 94.

of second order.

t,

10, 157.

i.

T, symbol for tensor, 163.

triangle, 217.

vector equations

I,

and planes related to linear vector function,

5.

Surfaces, anharmonic equations of, 87.

cts,

points, 55.

six planes, 302.

512.

scalar of, 185.

Summand,

of space, 394.

),

393.

115.

17.

types, 76.

products of vectors, III.

i.

6,

337.

Tetragonometry, spherical, 417.

Tetrahedron, pyramid or gauche quadrilateral, 82
Total differentiation, 479.

ELEMENTS OF QUATERNIONS.

582

Transcendental functions of a quaternion, 421, 453.

of

V, symbol for vector, or right part, 193

See Ellipsoid,

= Sq +

of powers of vectors, 420.

Vg, differential

of quaternions, 356.

Vanishing quantities, ratio

exponential, 473.

Vection,

197.

-K),

455.

of,

470, 483.

3.

Transvector, 3, 147.

difference of points, 3.

Transversal, triangle cut by, 21.

origin and term of, 3.

resolution of, 194, 338, 339.

of spherical triangle, 362.

Triangle, plane, harmonic

of,

5.

Vector, definition of,

Transport, 4.

as a factor, 335.

relations, 21.

trinomial form, 344.

conies and, 32.

product of vector sides of plane, 348.

arc, 3.

determination

spherical, area, 364,

exponential equation, 404.

of, 144.

arcs, addition of, 147.

of angles, 406.

is associative,

Trigonometry, plane, fundamental formula, 214; 178,

197, 265, 272.

303.

representative of a versor, II.

357, 325, 400.

associative property of, 304.

sum

S and V, 209.

angles,

solution of the exponential equation of, 409.

expression of curve, 94

Trigonometrical functions of a quaternion, 424.

Trinomial form for vectors, 344, 242.

Types of points of construction,

of,

156, 406.
;

of surface, 94.

of developable, 100.
tensor

linear vector functions, 486.

of,

163.

versor of, note, 137.

55.

scalar of, 346.

reciprocal of, 293.

U, symbol for unit -vector, 136.

conjugate

for versor, 137.

of, 346.

plus scalar

examples on, 142.

is

Uq, differential of 456.

is

quaternion, 11, 335.

right part of quaternion, 335.

differentiation of, 98.

469.

partial differentiation of, 101.

\/~l,

224, 289.

differential of, 96.

Unit-vector, 120.

327, 345.

square

of,

conception of right versor as, 335.

power

of, 396, 399,

sphere, 120.

called imaginary part, note, 316.

in space, scalar, 380.

trinomial form

Unity, particular case of radial quotient, 132.

negative, inversor, 135.

291.

th
geometrically real n roots of, 259.
th
geometrically imaginary n roots

344.

special case of quaternion, 335.

addition of, to point, 5.

square root of, 133, 224, 253, 258, 289.

cube root of, principal, 258.
of,

for,

420, 476.

or vector-part, of a product, 245.

square of right radial, 132.

nine cube roots

9, 143.

representative of a versor, 161.

spherical, fundamental formula, 216

Uninterpreted symbol,

i.

angle, 3.

S and V, 208.

ratio,

335.

Vg, 193.

of linear vector functions, 528.

V= -S = |(l

Ultimate

= IV.

distributive character of, 204.

U, 141.

of equation of ellipsoid.

sum

examples on, 195, 197, 199, 208.

Transformations from S to T, 190.

of,

of conjugate, 197.

equation of circle, 349, 355.

of second degree, 508.
Vectors, addition,

290.

is

commutative,

coefficients of, 9.

6.

INDEX TO VOLUME
Vectors, complanar, 340.

Velocity, vector of, 99.

complanarity of three, 338.

continued proportion of complanar, 251.

Version, 164.

Versor or radial quotient regarded as a factor, 135.

of quaternion, 137; note, 124.

determined by three scalars, 341.

division

by

11.

by numbers,

of vector, note, 137.

parallel vector, 10.

by any

of scalar, 139.
of null quaternion, 139.

vector, 107.

elimination of, 342.

of conjugate, 138.

fourth proportional
diplanar, 357

to three,

complanar,

250

See Linear vector function,

function, linear.

harmonic mean of two, 298.

identities connecting.

of,

139.

of unit vector, cf., 261.

depends on relative direction, 138.

any

linear relations connecting

mean between two,

any

four, 44

of sum, 476.

five, 47.

of product, 171.

251.

development

multiplication, 323.
is associative,

of,

476.

Versors, condition of complanarity of, 148.

equation between, 409.

Volume, sign of a, 48, 342.

337.

3.

325

power

right, as unit -line, 335.

See Identity,

linear relation connecting

parallel, 10,

conjugate
is

integrals, 482.

null,

of reciprocal, 138.

reciprocal of, 138.

rectangular, 377.

583

I.

perpendicular, 325, 345.

product of two, interpreted as quaternion, 321.

of parallelepiped, 247, 338.

of sphere, spheric shell, 483.

product of any number, interpreted as quaternion,

337, 346, 394.

product of two rectangular, 333.

"Warren,

note, 278.

proportion of, 118, 175, 250.

subtraction, 5.

Bum

of, defined, 7.

termino-collinear, 14, 343.

tennino-complanar, 45, 344.

Zero, square root of, 316, 291.

null quaternion

versor

Zone, area

END OF VOLUME

I.

of, 139.

of,

482.

is,

125.

^
RETURN
TO*

ALL

Astronomy/Mathematics/Statistics Library

100 Evans Hall

BOOKS MAY
DUE AS

j*iiiKS^t;

642-3381

BE RECALLED AFTER 7 DAYS

STAMPEDLOW

DEC 05 19975*

DEC

23

1998

b 19S9
Kff
d UCB A
Rec
,-J

reca ||

DEol

2000
UNIVERSITY

FORM NO. DD

19

OF CALIFORNIA, BERKELEY

BERKELEY,

CA 94720