TAXONO'vlIC CATEGOfllES IN FOSSIL 1l0MIl\IOS
ERNST ~lAYH
Thc\merican \luscu'" of Natural J1istory, Ne" lork
It is one of the most fruitful procedures of
modern science to bring specialists of various
fields together to discuss the problems that con­
cern the zone of overlap of their fields. ~ot pos­
sessing any first-hand knowledge of paleoanthro­
pology, my own contribution to the question of the
taxonomic categories of fossil man will be that
of a systematist. Significant prof!less has been
made wi thin recent years among biological! y
thinking taxonomists in the understanding of the
categories of subspecies, species, and genus, and
it is my hope that this knowledge may help in a
better understanding of fossil man.
The who] e problem of the origin of man depen,ls,
to a considerable extent, on the proper deflllition
and evaluation of taxonomic categories. gut, there
is less agreement on the meaning of the categories
species and genus in regard to nmn and the pri­
mates than perhaps in any other f!lonp of animals.
Some anfhropologists, in fact, imply that they use
specific and generic names merely as labels for
specimens without giving them any biological
meaning. The late Weidenreich, for example, stated
that in anthropology "it always was and still is
the custom to give generic and specifiC names to
each new type witbout mucb concern for the kind
of relationship to other types formerly known."
Broom (1950) likewise states, "I think it wil! be
much more convenient to split the different varie­
ties lof South African fossil ape-man! into dif­
ferent genera and species than to lump them."
The result of such standards is a simply be\\il­
dering diversity of names. In addition to various
so-called species of lIomo, the following names
for various hominid remains have been found by
me in the literature: Australopithecus, Plesi­
anthropus, Paranthropus, C'oanthropus, Gigan­
thopi thecus, Meganthropus, Pithecanthropus,
Sinanthropus, Africanthropus, favanthropus, Pale­
oanthropus, Furopanthropus, and several others.
No two authors af!lee either in nomenclature or
in interpretation. It seems to me tbat an effort
should' be made to give the categories species
and genus a new meaning in the field of anthro­
pology, namely, the same one which in recent
years bas become the standard in other branches
of zoology.
[109 ]
A re-evaluation of thc tcrminology of hominid
taxonomy is facilitated by the fact that in recent
years a magnificent body of new data has been
accumulated by anthropologists, partly bused on
comparative anatomical studies and partly on
significantncw discoveries of fossil man in south­
east Asia and in eastern and souther'l Africa.
The nomenclatorial dWiculties of the anthro­
pologists are chiefly due to two facts. The fIrst
one is a very intense occupation with only a very
small fraction of the animal kingdom which has re­
sui ted in the developmen t of standards that di fie I'
greatly from those applied in other fields of zool­
ogy, and secondly, the attempt to express every
di flerence of morpho] ogy, even the sl ightest one,
by a different name and to do this with the limited
number of taxonomic categories that arc availabl e.
This difference in standards becomes very ap­
parent if we, for example, compare the classifi­
cation of the hominids with that of the f)rosophila
flies. Then> ilre now about (,00 species of Oroso­
philo, known, all included in a single genus. If
individuals of these species were enlarged to the
size of man or of a gorilla, it would be app,U'ent
even to a lay person that they are probably more
different from each other than are the various pri­
mates and certairly more than the species of the
suborder Anthropoidea. What in the case of /Jroso­
ph.ila is a genus has almost the rank of an order
or, at least, suborder i:l the primates. The dis­
crepancy is equally great at lower categories, as
we shall presently see. It is not mere formalism
to try to harmonize the categories of anthropology
with those of the rest of zoology. Rather, the
evaluation of human evolution depends to a con­
siderable extent on the proper determination of the
categories of fossil man.
There are two recent developments in general'
systen,atics that will be particularly helpful in
our e rrorts. The first one is that the biological
meaning of the categories species and genus is
now better understood than formerly, and second,
that, in the attempt to close the gap between the
complexity of nature and the simplicity of cate­
gories, the nun,ber of existing categories has been
augmented by intermediate and group categories,
such as "local population" or "local races" and
....

no ERNST MAYR

"subspecies groups." The adoption of these as a truly collective category. I shall illustrate might easily
intermediate categories facilitates classification this need for the combining of genera by an ex­ the availabl~
without encumbering nomenclature. ample. Gorilla and chimpanzee are two excellent logical and (
species which, as Professor Schultz has shown, the hominids
THE TAXONOMIC CATEGORIES
differ from each other by a wealth of characters. the current n
The work in the new systematics has led to a At one time several species of gorillas and of This re-e\
far-reaching agreement among zoologists on the chimpanzees were recognized, but the allopatric fossil homil
meaning of the categories subspecies, species, forms within the two species are now considered categories E
genus, and family. ]n the following an attempt subspecies. Being left with one species of gorilla to a separat
shall be made to see how far the current usage of and one species of chimpanzees, we are con­ cal differen
these categories can be extended to fossil homi­ fronted by the question whether or not they are which the
nids and what such a reclassification means in sufficiently different to justify placing them ill liomo are sc
terms of human evol uti on. different genera. A specialist of anthropoids tification fo
The genus: The genus is a taxonomic category impressed by the many differences between these There is ev
for a group of related species. It is usually based species may \Hmt to do so. Other zoologists will African man
on a taxonomic group that can be objectively de­ conclude that the differences between the two opithecinae.
fined. However, the delimitation of these groups species are not indicative of a generic level of those of S(
against each other, as well as their ranking, is difi'erence when measured in the standards custom­ hominid cha
frequently subjective and arbitrary. A conventional ary in most branches of zoology. To place these others that
definition of the genus would read about as fol­ two anthropoids into two separate genera defeats as small si
lows: "A genus consists of one species, or a the function of generic nomenclature and conceals noteworthy,
group of species of common ancestry, which differ the close relationship of gorilla and chimpanzee investigator
in a pronounced manner from other groups of spe­ as compared with the much more difierent orang primitive s lJ
cies and are separated from them by a decided and the gibbons. Hecognizing a separate genus for that were fo
morphological gap." the gorilla would necessitate raising the orang ful canines
Hecent studies indicate that the genus is not and the gibbon to subfamily or family rank as has the first lo~
merely a morphological concept but that it has a indeed been done or suggested. This only worsens ment of tht
veiy distinct biological meaning. Species that the inequality of the higher categories among the powerful br
are united in a given genus occupy an ecological primates. able that m
situation which is different from that occupied by The same is true for the fossil hominids. After spe cializat
the species of another genus, or, to use the termi­ due consideration of the many di fferences between paid apes a
nology of Sewall Wright, they occupy a different \Iodern man, Java man, and the South African ape­ 1 he fact
adaptive plateau. ]t is part of the task of the man, I did not find any morphological characters izations ha
taxonomist to determine the adaptive zones oc­ that would necessitate separating them into sev­ dence to p(
cupied by the various genera. The adaptive pla­ eral genera. Not even Australopithecus has un­ a very iso
teau of the genus is based on a more fundamental equivocal claims for separation. This form ap­ This is by
difference in ecology than that between the ecolo­ pears to possess what might be considered the tation. Rat
gical niches of species. principal generic character of Homo, namely, typical ch~
Unfortunately, there is no such thing as a rec­ upright posture .... ith its shift to, a terrestrial !'lode well be a
ognized or absolute generic character. This was of living and the freeing of the anterior extremity response tl
known already to the earlier taxonomists, in fact, for new functions which, in turn, have stimulated now appear
Linnaeus stated, ''It is the genus that gives the brain evol ution. Within this type there has been the orang,
characters, and not the characters that make the phyletic speciation resulting in Homo sapiens. most of th
genus." The genus is a group category and it The claim that the many described genera of therefore,
defeats the object of binomial nomenclature to hominids and Australopithecines have no validity, available I
place each species into a separate genus, as has if the same yardstick is applied that is customary man may I
been the tendency among students of primates. in systematic zoology, is based on two major chiIlJpanze,
The acceptance of the new concept of biolog­ points. Both of these are admittedly somewhat orang or tc
ically defined polytypic species (see below) vulnerable. One is the overall picture of morpho­ in certain
necessitates the upward revision of all other cate­ logical resemblance with a deliberate minimizing be used t(
gories (Mayr, 1942). Often what was formerly a of the brain as a decisive taxonomic character. tionship. '
group of allopatric species is now a single poly­ To this point we shall return presently. The other poids are
typic species with numerous subspecies. To point is the assumption that all these forms, in­ and will tl
leave each of these polytypic species in a sepa­ cluding Australopithecus, "re essentially members verge nt, I
rate genus deprives the genus of its significance of a single line of descent. Additional finds upright pc
 1

TAXONOMIC CA. TEGORIES IN FOSSIL HOMINIDS III

I illustrate might easily disprove this. llowever, taking all
by an ex- the available evidence together, it seems far more
o excellent logical and consistent at the present time to unite
has shown, the hominids into a single genus than to con tinue
characters. the current multiplicity of names.
las and of This re-evaluation of the generic status of the
allopatric fossil hominids forces us to consider also the
considered categories above the genus. Does Homo belong
s of gorilla to a separate family Hominidae? The morphologi­
~ are con­ cal differences between Pongo, the genus to
lt they are which the chimpanzee and gorilla belong, and
Ig them in llama are so slight that there seems to be no jus
Illthropoids tification for placing them in separate families.
Neen these There is even less justification for placing South
ogists will African man in a separate subfamily, the Austral­
n the two opithecinae. The most primitive known hominids,
ic level of those of South Africa, combine certain typical
ds custom­ hominid characters, such as upright posture, with
lace these others that are usually considered simian, 8uch
,ra defeats as small size of brain and protruding face. It is
f conceals noteworthy, however, as pointed out by several
himpanzee investigators, that these hominids, even at this
rent orang primitive stage, lack certain other simian features
: genus for that were formerly considered as primitive: power­
the orang ful canines, large incisors, a sectorial form of
Illk as has the first lower premolar, an exaggerated develop­
.y worsens ment of the supra-orbitals, a simian shelf, and
among the powerful brachiating arms. It now appears prob­
abl e that many of these characters are functional
lids. After specializations which were acquired by the anthro­
s between poid apes after the hominid line had branched ofI.
ncan ape­ 1 he fact that the hominids fack these special­
characters izations has been used by some authors as evi­
into sev­ dence to postulate a very early human origin and
s has un­ a very isolated position of the hominid branch.
form ap­ This is by no means the only possible interpre­
dered the tation. Rather it seems to me that most of these
, namely, typical characters of the living anthropoids may
trial mode well be a single character complex evolved in
extremity response to a highly arboreal mode of living. It
;timulated now appears probable that the African anthropoids,
has been the orang, and the gibbons, may have acquired
:piens. most of these characters independently and are
genera of therefore, in a sense, a polyphyletic group. The
, validity, available evidence seems to indicate to me that
~ustomary man may be more closely related to the gorilla­
wo major chiplpanzee group than this group is either to the
somewhat orang or to the gibbons. The degree of similarity
f morpho­ in certain morphological traits cannot necessarily
inimizing be used to measure degree of phylogenetic rela­
:haracter. tionship. The arboreal, brachiating large anthro­
The other poids are exposed to a similar type of selection
'orms, in- and will therefore evolve in a parallel, if not con­
members vergent, manner. When tIle llama-line acquired
nal finds upright posture it entered a completely different
adaptive zone and became exposed to a severely
increased selection pressure. This must have re­
suI ted in a sharp accel eration of evolutionary
change leading to the well-known differences be­
tween man and the iiving anthropoids. This factor
must be taken into consideration when the phylo­
geny of man and the anthropoids is reconstructed.
It would therefore appear to be misleading from the
purely morphological-phylogenetic point of view to
separate man from the anthropoid apes as a spe­
cial family. It would be equally misleading to go
to the other extreme and to use the evidence of the
somewhat independent evolution of man and the
various anthropoids as a means to deny their
close relationship.
Denying the genus li omo family rank is based
on purely morphological considerations. It does
not take into account man's unique position in
nature. ~lan has undoubtedly found an adaptive
plateau that is strikingly difl'erent from that of
any other animal. There are some \\ho feel that
there is only one way by which to emphasize this
uniqueness of man, namely, by placing I!omo into
a separate family. The conventional standards of
taxonomy are insu fIicient to decide what is cor­
rect in this case.
From the purely biological point of view man
is certainly at least as diflerent as a very good
genus. We have thus the evolution of a new
higher category in the geologically short period
of one to two million years. This is another sig­
nificant illustration for the rapidity by which one
major taxonomic entity can be transformed into
another one, without any jumps.
The subspecies: Before we can attempt to
answer the question how many species of fossil
man have existed, we must say a few words on
infraspecific categories. The species of the
modern systematist is polytypic and multidimen­
sional. It has the geographical dimensions of
longitude and latitude and also the time dimen­
sion. It is polytypic because it is composed of
lower units, such as subspecies and local popu­
lations. Customarily in anthropology, distinct
local populatiorrs have been referred to as races,
and a similar custom exists in som,e branches of
zoology as, for exarr'ple, in ichthyology.
The amount of geographical variation and the
degree of difierence among the geographical sub­
divisions of a species are different from case to
case. Some species appear quite uniform through­
out their entire range; other species have a few
or many more or less well defined subspecies.
For instance, the two African forest anthropoids,
chimpanzee and gorilla, 8how only a moderate
112 ERNST iliA JR

amount of geographical variation, al though both can be interpreted as indicating that, in spite apparently
have well-defmed subspecies, and attempts have of much geographical variation, never more than and l'\eand
been IIlade to spl it the chimpanzee into several one species of man existed on the earth at any man IS n,
species. Ceographical variation is much more one time. We shall come back to this point later. Keanderthl
pronounced in the orang and even more so in some The species: As described in several publica­ hominids t
of the South American monkeys where geographi­ tions, the concept of the species has undergone P]eistocen
cal races are often different enough to be con­ a considerable change durinf!, recent years. The combe, an
sidered full species by conservative authors. morphological and typological species of the early modern rna
~Iodern man is comparatively homogeneous be­ taxonomists has been rep} aced by a l,iological primitive ,
cause there is much interbreeding between dif­ species. The species is now defmed "as a group They live<
ferent tribes and races. Still, we find in close of actually or potentially interbreeding natural more clos
neighborhood to each other such strikingly dif­ populations that is reproductively isolated from l'ieanderth.
ferent races as lmshmen and Bantus in South other SIl ch groups." When th is concept is applied If I un(
Africa, or the Congo pygmies and Watusi in cen­ to man, it is at once obvious that all living popu­ possible t,
tral Africa, or the \\edas and Singhalese in Cey­ lations of man are part of a single species. Not as remalOi
lon. There is much indirect evidence that primitive onl yare they conne cted everywhere by in ter­ cestra] bo
man was much more broken up into small scattered mediate populations bnt even where strikingly derthal an
tribes with little contact with each other, in­ distinct human populations have come in contact, by geogra
tensely subject to local selective factors. s!lch as Europeans and Hottentots, or as E~uro­ ing hypotl
In addition to this much greater geographical peans and Australian aborigines, there has been its classi
variation of primitive man, there is eviden ce also no sign of biological isolating mechanisms, only occurred il
of greater individual variation (including sexual social ones. Africa by
dimorphism). The variability of 'It. Carmel man The problem of species delimitation is much man, whil
has been conm;ented upon in the literature. It more dilTicult with respect to fossil man. lIow curred at I
seems possible, if not probable, that the various shall we determine which populations are "actu­ dertha]oid
South African finds, Australopithecus, Plesian­ ally or potentially interbreeding"? It is evident or 111 som
thropus, and Paranthropus, might well be age or that we must usc all sorts of indirect clues. The remalOS 0
sex stages of a few related tribes, notwithstand­ first concrete problem is what types of fossil man expand an
ing Broom's (1950) assertions to the contrary. should be included in the species homo sapiens. temporary
Differences between young and adult and be­ Cro-~lagnon man is so nearly identical with Homo the Austra
tween male and female appear to be greater in the sapiens that its inclusion in that species is not Indians. 1
goriIJa and orangutan than they are in modern man. c10ubted by any serious student. derthal ch
Variability may increase or decrease in the course The problem of Neanderthal man is much more surrably I
of evolution. Abundant proof for this staterr'ent ditlicul t. Should he be in cluded in the same of the cha
can be found in the paleontological Ii terature. I species as modern n'an or not? V,hen the fLfst fonds It IS v'
interpret the available literature to indicate that of "ieanderthal man were made there seemed to make the
primitive man showed more geographical as well be no problem. These fossils were characterized derthal ev
as individual variation than f'lodern man. by distinct morphological features and were low Dobzl
o •

Why primitive man should have been more vari­
able than modern rr~an is not entirely clear. A
study of the family structure of anthropoids might
shed son;e light on this problem. Perhars there
was a h'Teater functional difference between male
and fernaJe than in modern man. Perhaps the
ancestral hominids had a system of polygamy that
would favor the selection of secondary sex char­
acters in the male. \\e don't know. Whatever the
reasons, we should not use the variabil ity within
populations of modern man as a yardstick by-which
to judge the probable variability of extinct
popula tions.
This point is important because it bears on
the question whether or not more than one species
of hominid has ever existed on the earth at any
one time. Indeed, all the now available evidence
clearl y replaced by modern wan in l<'urope on a two forms
distinct chronological level. There is no morpho­ seems to
]0 gical or cui tural intermediacy. Additional finds, species.
however, have caused various dilliculties. Tn
llomo f
Palestine the ~.lt. Carmel finds belong to a popu­ ficiently (
lation that combines some features of Neanderthal have to h
with some of modern n an. It is in,material whether must be c
we interpret this as a hybrid popu]a tion, as an less of th
intermediate population, or as a population an­ and Wadj,
cestral to both. The fact remains that iV1t. Carmel unbroken
man makes the delimitation of modern man from t man to 11'
Neanderthal exceeding] y di fficult, if not impos­ pekinensi
sible, as pointed out by Dobzhansky (1944). man that i
Weidenreich supported the theory that modern man Ilcall y di
was a direct descendant of Neanderthal man. 1944).
130111 e and others have raised serious obj ections In spit,
to this theory. But how can we reconcile the pithecus
 TAXONOMIC CATEGORIES IN FOSSIL !iOMINIDS 11:3

~ in spite apparently incompatible views that Plod ern n,an Java ,lleganthropus to assure a correct classifIca­
PJore than and Neanderthal are conspeciflc and that modern tion. This is even more true of (;iganthopithecus
rth at any man is not a descendant of typical }~uropean whom some authors consider hominid and others
int later. Neanderthal? A possible clue is furnished by the anthropoid. One thing about Ciganthopithecus is,
:1 publica­ hominids that were widespread in Furope in n id­ however, very probable, naPlely, that it was not
undergone Pleistocene. The skulls of Steinheim, Swans­ necessarily a giant in spite of its giant teeth.
ears. The combe, and of Fontechevade con,bine features of Jaws and teeth of early fossil Dian were relatively
f the early modern man and of Neanderthal man, together with much larger than they are in modern man.
biological primitive and specialized features of their own.
as a group They lived apparently in inter-glacials and were Homo transvaalensis: South l\frican ape-man
19 natural more closely linked with a warm climate than again is one level further back and is suil'iciently
:ated from Neanderthal man. far removed from Java man to be considered a full
is applied lf I understood the evidence correctly, it is species. Actually, no less than three genera and
ving popu­ possible to interpret these early European fossils five species of South African ape-nan were de­
lcies. Not as remains of populations of !Jon;o that were an­ scribed which, in Broom's tenninolo!-,y, have the
by inter­ cestral both to sapiens and to "classical" Nean­ following names: Australopithccus a{ricanus 1925
strikingly derthal and from which these two forms evolved (Taungs), A. rrometheus 1917 C\lakapan), Plesi­
n contact, by geographical variation. Tentatively the work­ anthropus transvaal ens is 1936 (Sterkfontain),
as Euro­ ing hypothesis can be nlade that Neanderthal in Paranthropus robustus 1938 (Kromdraail, and
has been its classical form was a geographical race that Paranthropus crassidens 1949 (Swartkrans). \Jost
sms, only occurred in central Furope and was represented in of these names may not have any validity, accord­
Africa by fihodesian man and in Java by Njandong ing to the Hules of Zoological "iomenclature,
I is much man, while a more sapiens-like population oc­ Article 25A, as revised in 1930. According to
these Hules a name has validity only if the de­

j
nan. How curred at the same period as some of these Nean­
Ire "actu­ derthaloids either in north Africa or western Asia scription includes dia/2:nostic characters. Since
s evident or in some other area that has not yet yielded one of these names was based on a child, another
lues. The I remains of fossil man. When sapiens began to on an adult female, a third on an adult male, an

•
I

'ossil man expand and spread, he eliminated the other con­ enumeration of diagnostic differences is virtually
temporary races just as the white Polan drove out in~possible. The extant skulls are somewhat
) sapiens.
the Australian aborigines and the '\orth American altered in shape due to crushing, and the fact that
vith Homo
ies is not •
i Indians. The process of 1"1 imination of the Nean­
derthal characters in mixed populations was pre­

the cephalic index in the Taun/2:s child is 62.4

while it is R3.S in the Sterkfontein IT'ale is there­
]Uch more
the same
J sunably helped by selection preference in favor

of the characters of modern man.

fore not as signifIcant as Broom thinks. ~or is

the fact that the finds are associated with dif­
It is very probable that additional finds will
ferent faunas. Contemporary modern man can be
fIrst finds
make the delimitation of sapiens against Nean­
found associated with okapis or elephants or
leemed to
derthal even more difficult. It seems best to fol­
tigers or kangaroos, or South American edentates
~acterized
low Dobzhansky's suggestion and to consider the
or with polar bears. The various fwds of South
md were
two forms, as well as the ancestral group that
African "nan are presumably not contelrporary, but
'ope on a
seems to combine their characters, as a single
there is nothing in the evidence that has so far
o morpho­
species.
been presented (e.g. Broom, 1950) that would
,nal finds,
prove that more than one species is involved.
tlties. In
Homo erectus: Java and I'eking man are suf­ [ntil a real taxonorr;ic distinction has been
o a popu­
ficiently distinct from modern man so that they established, it will be safer and more scientific
anderthal
have to be considered a separate species, which to refer to the difTerent South African fossils by
1 whether
must be called /lama erectus. This is true regard­ vernacular names. There is no dan/2:er of confusion
In, as an
less of the fact that on Java, at least, Njandong if we speak of the Sterkfontain or \lakapan finds,
ation an­ and Wadjak man may have formed a practically while it implies an obviously erroneous con­
t. Carmel unbroken chain of hominids leading from Java clusion, namely that of generic distinctness, if
man from man to n'odern n,an. Peking man (Ilorho erectus we refer to them as Plesianthropus and Australo­
>t impos­ pekinensis) is, on the whole, so similar to Java pithecus. i':ew discoveries are still being made in
r (1944). man that it should be cons idered merely subspeci­ these cave deposits and many of those that have
,dern man ficall)' distinct, as I proposed previously (tvlayr, already been made have not yet been fully worked
hal man. 1944). out. There is good reason to believe that it will
bjections In spite of its obvious similarity to Australo­ be firmly established in the not-tao-distant future
Ilcile the pithecus too little is known of the still earlier how many different tribes, temporal subspecies,
 114 EfUvSTI1A YR

or even species of South African ape-man once Oakley (19S0) indicates that, indeed, jaw and derthal, i!
existed. To consider them all as one species is skull cap may be of the same geological age. biological
the simplest solution that is consistent with the It may take a long time he fore the Piltdown treat each
available evidence. puzzle is completely cleared up. As a working On the ot
A more important question is '" hether South hypothesis it might be snggested that Piltdm\n species c
African man is ancestral to modern man or merely man represents a geographical race of man that tween the
a specialized or aberrant sideline. The exact dat­ was restricted to northwestern Europe. Some of difference
ing of these fossils has not yet been achieved the characters, particularly in the jaw, appear to to the lev
but they are believed to be very cady Pleistocene be specializations rather than indications of Negro. lIo
or latest Pliocene, in fact, they presumably primitiveness, perhaps developed in cOllnection First of
ranged over a considerable period of time. There with the large size of the individual. The phylo­ classificat
is thus no defmite chronological reason why the genetic and chrono] ogical relationship of Pilt­ express a(
South African ape-man could not be considered down to the other hominid finds indicated by the of natural
a possible ancestor of modern man. The principal the words lleidelberg, Steinheim. Swanscombe, cies and g
objection that has been raised is that South and Fonteehevade still remains to be determined. have assi,
African man shows a combination of characters The simplified nomenclature of fossil man: \{e­ least an E
that "shonlcl not" occur in an early hominid. This ducing the bewildering assortment of genera and gories anI
argument is based on typological considerations. species of hominids to one genus with three solution a
Adherents of this concept believe that missing species results not only in simplicity lJUt it also from a rel
links should be about half-way between the forms makes certain conclusions 0bvious that were pre­ cate gories
the v connect and that thev should be half-way in viously not apparent. Before discussing these use such
eve'ry respect. This unclo:lbtedly is not the case conclusions, however, I might point out some of and "loca
with Australopithecus. It is apparently amazingly the disadvantages of such a simplified subspeeifi(
like modern man in its upright posture, structure classification. Hence, wt
of the pelvis, and other features, while it is very There have been two trends in human evolution practical 1'1
simian in its massive mandibles, large molars, as, indeed, there arc in the evolution of all or­ 1. Not t
prognathism, and small brain. J1ustralopitheclIs ganisms. First of all, there is a continuous evo­ population
lacks those specializations that stamp gorilla, I utionary change in time, the so-called phyletic specifiC raJ
orang, and gibbon as typical anthropoids. evolution, starting in the hon;inids with the most 2. To gi
The peculiar combination of cha~acters that is simian forms and ending wi th modern man. Simul­ deserve hij
found in ;lustralopithecus is due to the fact that taneously a centrifugal force has been operating, 3. To gr
rl uring evol ution of man di Herent characters namely, geographical and other local variation, those sub
evolved at different rates. If we would set the which tries to break up the uniform human spe­ either geoe
point where the human line branched off from the cies. Th is geographic variation leads to the Such suI
other anthropoids as zero and the !/omo sapiens formation of races and subspecies, and if this instance, n
stage as 100, we might give arbitrarily the follow­ trend would go to completion, to the formation of (a) ma
ing points to the various organs of /ll1stralopithe­ new separate species. There are all sorts of (b) Ne
cus: pelvis, 90; premolars, 75; occipital condyles, intermediate stages in both these trends and it (c) pn
RO; incisors, .5.5; the setting of the brain case, is obvious that all the Plany possible diITerences 4. Not to
70; shape of the tooth row, 70; the profile of the and gradations between the various kinds of to new foss
jaw, 30; the molar teeth, -10; the brain, 35; etc. hominids cannot be expressed completel y in the \' ernacular;
It is obvious that one type does not change into simple nomenclature of species, genus, and sub­ down man,'
r another 'type evenly and harmoniously, but that species. leading. T
some features run v.ay ahead of the others. For instance, man as he exists today, has pro­ specific Of
The inabil ity to understand this has been the nounced racial groups, such as the Whites, docs not e
reason for \\'eidenreich's insistence that t,'oan­ Negroes, and \'longoloids, which might well de­ justified pi
thropus was an artifact. lie maintained with re­ serve suhspecific recognition. But there are minor error as to
spect to l'iltdown man: "Form and individual racial differences within each of these subspecies. when the
features of the brain case are generally acknowl­ Furthermore, pre ceding modern man there have only to the
edged as those of modern []lan; those of the lower been types of !iomo sapiens that are now extinct, Anthropc
jaw have anthropoid characteristics. Therefore, like ero-Magnon man and his contemporaries. fact tha t tl
both skeletal elements cannot belong to the same This, no doubt, is a different level of subspecies lations,
skulL" As a matter of fact, the skull cap is not from those of living man. Neanderthal man is a
should nev
strictly modern nor is the jaw strictly anthropoid third level, and the pre-Neanderthal man, who
sumably m
and the recent ±luorine content determination by combines certain features of sapiens and Nean­

..

 TAXOAOMIC CA TEGORIES IN FOSSIL flOMINIDS 115

CONCL USIONS
ed, jaw and derthal, is a fourth level. It is unsatisfactory for
lcal age. biological, as well as for practical reasons: to The arranging of all finds of fossil hominids
the Piltdown treat each of these levels as a separate species. into a single genus with three species helps to
,s a working On the other hand, combining them into a single focus attention on the following conclusions.
llat Pil tdown species conceals the pronounced differences be­ The question of the "missing link." Ever since
of man that tween thesc levels and reduces the taxonomic there has been an appreciation of man's anthro­
.pe. SOlT'e of difference between Neanderthal and modern man poid origin there has been a s~arch for t.he
w, appear to to the level of difTerence between White \lan and "missing link." Some anthropologists may diS­
:Iications of Negro. How can this be avoided? claim this and say that they realize the gradual
I connection First of all, we must realize that no system of evol ution of mankind but the fact remains that
. The phylo­ classification and nomenclature can ever hope to accurate criteria of humanhood are elaborated
hip of Pilt­ express adequately the complicated relationships even in the most recent literature, such as Sir
:ated by the of natural populations. however, by giving spe­ Arthur Ke ith' s criterion of the brain volume
wanscombe, cies and genus the well-defined meanings that we of 750 cc.
determined. have assigned to these categories, we make at Th e anal ysis of this problem will be facil ita ted
~il man: He­ least an attempt to standardize taxonomic cate­ by the realization that it is an oversimplification

genera and gories and make them comparable. A possible to use in this case the uninomial al ternative

with three solution of our particular difficulties may come "ape" versus "man." Taxonomist~ kn?w. by

, but it also from a refmement of the levels of infraspecific experience the inadequacy of unmonlla!Jsm.

at were pre­ categories. In addition to the subspecies we may Classifying man binomially as llomo sapiens,

) . . " "
ssing these use such infrasubspecdic categofles as race it at once becomes apparent that we must look
out some of and "local pupulation," as well as the supra- for two missing links, namely that which connects
simpli fied subspecific category of tie I "sub '
species group. " sapiens with his ancestor and that which con­
Hence, we should be guided by the following nects Homo with his ancestor. Or, to express
lU evol ution practical rules: this differently, the two points of interest are the
n of all or­ 1. :\ot to assign a formal name to any local one on the phyl etic line of man where he reached
inuous evo­ population or race that does not deserve sub­ the sapiens level and second the place where the
ed phyletic specific rank. llomo line branched oft from the other primates.
th the nlost 2. To give trinomials to all forms that do not Let us look more closely at these two problems
man. Simul­ deserve higher than subspecies rank. of the origin of man. The branching oli of Homo
rr opera ting, 3. To group together as subspecies groups all frolil the other anthropoids was a case of orthodox
I variation, those subspecies within a species that form speciation distinguished only by the fact that
human spe­ either geographical or chronological groups. the new species simultaneously reached a new
ads to the Such subspecies groups in homo sapiens, for adaptive plateau. It is now evident, as has been
and if this instance, might be: stated by many authors, that a change in the
'ormation of (a) modern man mode of locomotion and a corresponding alteration
II sorts of (b) 1\ eanderthal group of the entire organization of the body, in other
~nds and it (c) pre-Neanderthal group words, the assuming of the upright posture, were
differences 4. Not to give formal generic and specific nan,es the essential steps that led to the evolution of
l kinds of to new fossil finds that are not sufticiently known. Homo. This evolutionary trend apparently afiected
tely in the Vernaculars, such as "Steinheim nlan" or "Pilt­ first the pelvis and poster-ior extremities, followed
s, and sub- down man," are just as useful and much less mis­ closely by the anterior extremities. The corres­
leading. The formal appl ication of generic and ponding re-organization of the skull lagged ap­
y, has pro­ specific names simulates a precision that often paren tl y behi nd. I t is therefore singular! y di flic ul t
fIe Whites, does not exist. To give the impression of an un­ to localize both in time and space this important
It well de- justified precision is as much of a methodological evolutionary step of the attainment of the upright
e are minor error as to make cal culations to the f.fth decimal posture with the help of jaw and tooth fragments,
mbspecies. when the accuracy of the original data extends such as constitute most of the primate and an­
there have only to the first decimal. thropoid remains in eastern Africa during Plio­
)w extinct, Anthropologists should never lose sight of the cene and Miocene.
~mporaries. fact tha t taxonomic categories are based on popu­ To determine the exact point in the phyletic
subspecies lations, not on individuals. Different names evol uti on of fl omo where the sapiens level was
I man is a should never be given to individuals that are pre­ reached, is quite impossible. It was a very
man, who sumably members of a single variable population.
gradual process leading from erectus to sapiens
and Nean­

116 ERNST MAYR
and no particular form can be singled out as the
missing link. However, there is a lower level in
the phyletic evolution of Homo that is of special
evolutionary interes t, namely, the level at whi ch
the hominids first displayed those intellectual
qualities that are considered distinctly human
rather than simian.
Attempts have been made to measure the attain­
ment of this llama level in terms of brain size.
This method is fraught with difiiculty. First of
all, brain size is to some extent correlated with
body size. If, for instance, a large gorilla should
have a brain of 650 cc. this is not at all neces­
sarily equivalent to the brain of a fossil hominid
of 650 cc., if that hominid were much smaller than
a gorilla. If the brain of the gorilla averages one­
fourth larger than that of the chimpanzee, it does
not mean ihat he is on the average 25 per cent
more intelligent. The correlation between br.ain
size and intelligence is very loose. There is good
evidence that the brain size of late Pleistocene
man may have averaged larger than that of modern
man. If true, this does not mean necessarily that
there has been a deterioration of man's intelli­
gence since the Pleistocene, for intelligence is
determined not only by brain size. It is, of course,
still unknown what neurological structures affect
intelligence but the folding of the cortex and all
sorts of specializations within the cortex appear
to be as important as size. It is therefore danger­
ous, in fact, outright misleading, to use size as
an absolute criterion and to say that the JJama
stage ~\(IS reached when brain size reached a
leve J of 700 or 750 cc.
It has been suggested to measure the attain­
ment of the human level by some cui tural achieve­
ment, such as the use of fIre, rather than by an
anatomical standard. like brain size. This is un­
questionably a superior approach, but has the
practical difiiculty that the flfst moment of fife
making was not fossilized and can never be dated
accurately. However, the first making of fife may
have occurred not much after the first use of tools
by hominids and some lucky finds illay shed light
someday on the period when that occurred. South
African :nan was presumably already a user of
tools, and the first use of tools may be coincident
with the evol ution of South African man.
Speciation in man: In the strict sense of the
word, speciation means the origin of disconti­
nuities through the origin of reproductive iso­
lating ,nechanisms. How often has man speciated?
The answer is that he has speciated only once if
our assumption is correct that never more than
one species of man existed on the earth at any
one time. This single event of speciation was the
--
branching ofT of Ilomo from the anthropoid stock. unique i
That some fairly distinct hominid remains have right. E'
been found in approximately contemporary deposits will con
does not prove t'heir specific distinctness. The species
sub-division of the human speci.es in.to independ­ appears
ent tribes favors diversification. If fossils of that is,
Congo pygmies and of lVatusi were to be found in species
the same deposit by a paleontologist, a million
years hence, he might well think that they be­
longed to two different species. As stated previ­ 1. The
ously, the known diversity (Jf fossil man can be than one
interpreted as being the result of geographic a given t
variation within a single species of Homo. This 2. It i:
led to the evolution of such aberrant types as hominids
Piltdown man of England, but apparently nowhere with thl
to the simultaneous occurrence of several species sapiens),
of Homo. What is the cause for this puzzling 3. The
trait of the hominid stock to stop speciating in the speci
spite of its eminent evolutionary success? It 4. The
seems to me that the reason is man's great ecol­ sl0wness
ogical diversity. Man has, so to speak, specialized lllechanis
in despecialization. Vlan occupies more diflerent llama int,
ecological niches than any known animal. If the
single species man occupies successfully all
the niches that are open for a llama-like creature, BROOM,
South
it is obvious that he cannot speciate. This con­ Anthfl
forms strictly to Cause's Uule. Also man is DOBZHAI\
apparently slow in establishing isolating mech­ living
anisms. This is indicated by the numerous in­ (N.S.)
MAYH, E.,
stances of incomplete speciation in the history
New'
of the hominids. In no case was this speciation 1944, 0
completed because the segregating populations subsp,
were either absorbed by intermarriage or exter­ 2: ll­
OAKLEY,
minated. Man is apparently particularly intolerant
of Pil
of competitors. The wiping out or absorption of STRAUS, I
primitive populations by culturally more advanced try. <)
or otherwise more aggressive invaders, which we
have witnessed so Illany times during the eight­
eenth and nineteenth centuries in Australia, North LASKER
America, and other places, has presumably hap­ justify c
pened many times before in the history of the to belong
earth. The elimination of Neanderthal man by the australopi
invading Cro-~1agrlon man is merely one example. one genus
There is one striking diHerence between man animal wit
and most of the animals. In animals whenever there cap, such
is competition between two subspecies the one to one for
that is better adapted for a specifIc locality seems paid brain
to win out. !\lan, who has reached such a high de­ Australopi
gree of independence from the environment, is less Incident
dependent on local adaptation, and a subspecies tomical rel
of lilan can quickly spread into many geographic­ with "Pil
ally distant areas if it acquires generalized NeanderthE
adaptive improvements such as are described by MA YR: S
the social anthropologist. Such improvements do acters (sel
not need to and probably often do not have genetic del imit gee
basis. The authors who have claimed that man is
 -r
1 TAXONOMIC CA TEGORIES IN FOSSIL flOJlIiVIDS 117

poid stock. I unique in his evolutionary pattern are undoubtedly \;]ONTAGU: I think it should be noted that con­
mains have right. Even though the phyletic evolution of man venience can be pushed to such a point that it
try deposits
lness. The
t! will continue to go on, the structure of the human
species at the present time is such that there
becomes a confounded nuisance. When, to para­
phrase A. E. Housman, the ambiguity of language
o independ­ appears to be very little chance for speciation, is brought in to add to the already existing
fossils· of that is, for the division of the single human confusion of thought, as has been done in the
I
be found in species into several separate species.
anthropological taxonomy of fossil man, confusion
t, a million
at they be­ i

i
SUYlMARY
is rendered worse confounded, and "convenience"
becomes an impediment to clear thinking and
tated previ­ t 1. There is no conclusive evidence that more further progress. The terminology of palaeoanthro­
man can be ! pology provides an unfortunate example of the
than one species of hominids has ever existed at
I
geographic
Homo. This
rlt types as
•
I
I
a given time.
2. It is proposed to classify fossil and recent
hominids tentatively into a single genus (Homo)
systematics of confusion. When types such as
Java man and Pekin man can be referred to by
generic names-Pithecanthropus and Sinanthropus

•
I
Itly nowhere with three species (transvaaIensis, erectus, -when, in fact, they represent no more than two
~ral species I sapiens ). subspecies or geographic races, it were high
is puzzling 3. The recognition of subspecies groups within time that we did something to bring the taxonomic
leciating in t
I
the species facilitates classification. practice of palaeoanthropology more into line with
mccess? It 4. The ecological versatility of man and his its own theory and the practice of the newer
, great ecol­ ~ sluwness in acquiring reproductive isolating systenl•• tists. I should therefore seriously suggest
I specialized mechanisms have prevented the breaking up of that a committee, consisting of such men as Drs.
ore different I llomo into several species. "'layr and Dobzhansky, and several palaeoanthro­
lima!' If the
~ssfully all
• HF:FERF:NCF:S pologists, be appointed to consider the matter of
revising the nomenclature which is at present
BROOM, R,. 1950, The genera and species of the

le creature,
~. This con­ •I
South African fossil ape-man. Amer. J. Phys.

Anthrop. (N.S.) 8: 1-13.

confusing the field of palaeoanthropology.

SCHULTZ: It has been long overdue and is most
Iso man is DOBZflANSKY. TIL, 1944, On species and races of welcome that expert taxonomists come to the aid
ating mech­ t living and fossil man. Amer. J. Phys. Anthrop. of anthropologists in modernizing the nomen­
(N.S.) 2: 251-265.
.umerous In- clature o~ the Hominidae which has reached such
MAYH, 1':., 19'~2, Systematics and the origin of species.
the history New York: Columbia University Press. absurdities as the generic separation of Java man
s speciation 1944, On the concepts and terminology of vertical and Peking man. Doctor Mayr's highly valued
populations subspecies and species. Natl. Res. Council 8ull. proposals for taxonomic reforms are rather start­
2: 11-16.
ge or exter­ t OAKLEY, l\.. P., 1950, New evidence all the antiquity
ling to anthropologists who still hope that their
I
ly intolerant of Piltdown man. Nature, Land. 165: :379-382. Latin or Greek binominal terms can imply con­
bsorption of STRAUS, W. L., Jr., 1949, The riddle of man's ances­ sistently degrees of distinction. To recognize
Ire advanced try. Quart. Rev. Fliol. 24: 200-223. only three species of llama, of which one is
:s, which we assigned to the Australopithecines and another
DISCUSSION
.g the eight­ to everybody from Neanderthal man to ourselves,
t
;tralia, North LASKER: I wonder how Professor iVlayr would is a very sudden jump from one, old extreme to an
umably hap­ justify considering men, gorilla and orang opposite, new extreme, even though the latter is
,tory of the to be long to three difl'erent genera, but man, the undoubtedly more consistent with-general usage
I man by the australopithecines and Piltdown man to belong to in modern systematics.
rle example. one genus. Considered purely morphologically, an Doctor MayI" s interesting suggestion that gorilla
)etween man anima) with "ape-like" jaw and "man-like" skull and chimpanzee represent only different species
neneverthere cap, such as Piltdown, if th'e parts really belong of one genus, because the skulls of the latter are
cies the one to one form, difflfrs greatly from one with anthro­ very similar to female skulls of the former, de­
cality seems poid brain case and more human jaw, such as serves comment as an excellent example to remind
I
:h a high de­ I Australopithecus. us, how unjustifiable it can be at times to assume
t
ment, is less Incidentally, I can think of no adequate ana­ close relationship between animals on the basis
L subspecies tomical reason for associating the Heidelberg jaw of close cranial resemblance alone (in one sex
geographic­ with "Pithecanthropus" rather than with the only). Even though the skulls of adult female
generalized Neanderthaloids. gorillas differ from those of adult chimpanzees to
I
:Iescribed by MA YR: Since there are no absolute generic char­ only limited degrees, the skulls of adult male
ovements do t acters (see above), it is impossible to define and gorillas are vastly difTerent, because sex difler­
have genetic delimit genera on a purely morphological basis. ences in gorillas are extremely pronounced in
I that man is

t
1
~ ­
..
 •
1
118 ERNST lvlAYR

contrast to those of chimpanzees which are Within the human family two groups may be recog­
generally less marked than in man. In other bodily nized, the small-brained man apes and large­
features the two African apes can differ enor­ brained, tool using man. It may be well to keep
mously. For instance, the proportionate size of the generic names Australopithecus and Homo for
the testes is extremely small in gorillas, whereas these. I thoroughly agree with Dr. :Ylayr on the
extremely large in chimpanzees, the female desirability of abolishing most of the genera he
sexual skin shows very difierent changes in the has mentioned. But too few names can be con­
two types, chimpanzees have relatively huge fusing, and I wonder if Dr. 'vIayr would feel this
outer ears, gorillas very small ones, the hand is suggestion is counter to his idea.
long and slender in chimpanzees, but relatively !'lelA): R: An unequivocal decision on the ranking It is tl
broader in gorillas than in any other simian pri­ of the groups that are included in the higher cate­ and to e1
mates, etc. If these and many other constant dis­ gories seems impossible. f!owever, the morpho­ should t
tinctions are mere species differences, then the logical difference between galagoes, lorises, and content ,
comparative Iy few known di fferences between monkeys seems much greater than that between sense th
Peking man and Hhodesian man, e.g., would de­ anthropoids and man. Furthermore, within the concepts
serve at most subspeciflc rank. single family Cercopithecidae (old world monkeys) There
With these remarks I wanted to point out chiefly there seems to be greater morphological variation
they are
that it will require more such profitable meetings (e.g., between macaques, baboons, langurs) than
however
between anthropologists, primatologists and there is morphological difference between man and
physical
taxonomis ts before we can expe ct a generall y the anthropoids. common;
acceptable, detailed, and much needed change in As far as brain size and use of tools is con­
nothing
the systematics of past and present higher cerned, there has been apparently a continuous
therefore
primates. line of development from the primitive hominids to
confusinl
WASHBURN: I feel that it is useful to have modern man. Java man (? or ape-man) is so com­ ment in
separate names to specify each of the major adap­ pletely intermediate between South African man McCown
tive groups of the primates. Traditionally, such and modem man, and the difference between these dents of
groups as galagoes, lorises, or monkeys were two terminal forms so much a matter of degree, ness " al
given family status. If this continues to be done, that it seems questionable whether these evo­ order to ,
then it is convenient to put apes and men in sepa­ lutionary stages justify generic separation. The archy. h
rate families, indicating the difference between difference is certainly not equivalent to a generic Chapelle
arboreal brachiation and bipedal ground living. di fterence in most groups of animals. demonstr:
culture f.
note of tI
to "pure
tionary I
geologic.
But he I
"good in
intuition
Mayr s
II says that
depends.
of taxonc
fined the
no such t
• Here prol
meet on
mechanis
may first
direction·
genetic s
that is I
logical d
assumpti,
ically u