You are on page 1of 3

Gudermann function

A definition for the Gudermann function is:


Z x
gd x = secht dt.
0

The function is also called the Gudermannian function or the hyperbolic amplitude function. Its domain is
all x.

Using the substitution u = sinht, for which du = cosht dt, and 1 + u2 = 1 + sinh2 t = cosh2 t we find an
alternate definition for this function:
Z x x Z x Z sinh x
1 cosht du
Z
gd x = secht dt = dt = 2
dt =
0 0 cosht t=0 cosh t u=0 1 + u2
 sinh x
= arctan u u=0
= arctan sinh x.

So Z x
gd x = secht dt = arctan sinh x. (1)
0
The graph of gd is similar to the graph of arctan.

By the Fundamental Theorem of Calculus:


d
gd x = sech x. (2)
dx
Since gd x is an angle whoseptangent is sinh x = (sinh x)/1 so we have a triangle as follows, where the
hypotenuse of the triangle is 1 + sinh2 x = cosh x:
cosh x 



  sinh x

 gd x

1
From this triangle we read off the following relationships. They are significant since through the Gudermann
(and without using complex numbers) they express hyperbolic functions in terms of trig functions.
1
cos gd x = = sech x (3)
cosh x
sinh x
sin gd x = = tanh x (4)
cosh x
tan gd x = sinh x (5)
sec gd x = cosh x (6)
cosh x
csc gd x = = coth x (7)
sinh x
1
cot gd x = = csch x (8)
sinh x
Here is yet another alternative expression for gd:
π
gd x = 2 arctan ex − . (9)
2
To prove this it suffices to show two things:

• Both sides are 0 at 0: gd 0 = arctan sinh 0 = arctan 0 = 0, while on the right side,
2 arctan e0 − π2 = 2 arctan 1 − π2 = 2 π4 − π2 = 0.

• The derivatives of both sides are the same. The derivative of the right side is
1 2
2· x 2
· ex = −x = 1/ cosh x = sech x,
1 + (e ) e + ex

which happens to be the derivative of gd x by equation (2).

The inverse Gudermann function

The inverse function to x = gd y is y = arcgd x, the arc Gudermann function. Its domain is (− π2 , π2 ) and its
graph is similar to the graph of tan on that interval. Its derivative, using equation (6), is

darcgd x dy dx d
= = 1/ = 1/ gd y = 1/ sech y = cosh y
dx dx dy dy
(6)
= sec gd y
= sec x. (10)
Rx
Integrating 0 and noting that arcgd 0 = 0 (see (9)) gives:
Z x
sect dt = arcgd x. (11)
0

Next we determine several alternate expressions for arcgd, that is, for the integral of sec.

The six equations (3) to (8) give, respectively:

arcgd x = arcsech cos x = arctanh sin x = arcsinh tan x (12)


= arccosh sec x = arccoth csc x = arccsch cot x, (13)

the proof of which we illustrate by using equation (4). The equation gives arctanh sin gd x = x, from which,
on replacing x with arcgd x, we get arctanh sin gd arcgd x = arcgd x, that is, arctanh sin x = arcgd x. Note the
pleasing symmetry:
gd x = arctan sinh x and arcgd x = arctanh sin x. (14)
Similar formulas hold for the other five cases.
There are expression for arccosh and arctanh in terms of ln which give these formulas:
 p 
arcgd x = arccosh sec x = ln sec x + sec2 −1 = ln(sec x + tan x), (15)
 
1 1 + sin x
arcgd x = arctanh sin x = ln . (16)
2 1 − sin x

Equation (9) gives arctan ex = 12 gd x + π4 , which rewrites as x = ln tan 1



2 gd x + π4 . Replacing x with arcgd x
and using gd arcgd x = x gives: x π
arcgd x = ln tan + . (17)
2 4

In summary, from equations (11), (15), (16), (17):


Z x x π
sect dt = arcgd x = arccosh sec x = arctanh sin x = ln(sec x + tan x) = ln tan + .
0 2 4
These formulas for the integral of sec, especially the last one, have an interesting history, as they were
discovered numerically in connection with map making by Mercator, Gunter, and Bond and proved by
Newton’s teacher Barrow and another famous mathematician of the time (Gregory).

Note that the antiderivative of sec is usually given with an absolute value:
Z
sec x dx = ln | sec x + tan x| + c

but in our case the domain is (−π/2, π/2) making the absolute value superfluous since there sec x + tan x =
1
cos x (1 + sin x) is a product of positives.

You might also like