Palaeont. afr., 23, 1-17 (1980) 

"AUSTRALOPITHECUS AFARENSIS" AND A. AFRICANUS: 
CRITIQUE AND AN ALTERNATIVE HYPOTHESIS 

by 

Phillip V. Tobias 

Department of Anatomy, University of the Witwatersrand, Johannesburg 
Honorary Professor of Palaeo-anthropology, 

Bernard Price Institute for Palaeontological Research 

ABSTRACT 
During the seventies, a succession of East African discoveries has been claimed to represent 

the "true" ancestral line of modern man, thus relegating A. africanus, and especially its Trans­
vaal subspecies, to a subordinate role in hominid phylogeny. The latest such attempt has 
been the claim of Johanson and his co-workers that the 3, 7-2,6 My-old hominids of Laetoli in 
Tanzania and of Hadar in Ethiopia represent a new species, "A . afarensis", which led to H . 
habilis, whilst A. africanus represents early stages in a specialized side-branch leading to A. 
robustus and A. boisei. A critique of the diagnostic criteria of "A. afarensis" reveals that on the 
available evidence, the Laetoli and Hadar fossils cannot be distinguished at specific level 
from A. africanus transvaalensis. Furthermore, it is by no means clear that the pooling for statis­
tical and comparative purposes of the Hadar and Laetoli fossils is justified. Hominids from 
the two sites are separated by about 800 000 years and about I 600 km as well as by morpho­
metric differences. As an alternative hypothesis, it is proposed that the Laetoli and Hadar 
hominids belong to the same lineage as that represented by the hominids of Makapansgat 
Members 3 and 4 and of Sterkfontein Member 4. Moreover, it is hypothesized that the Lae­
toli and Hadar hominids cannot be separated morphologically from A . africanus and that they 
represent two new subspecies of that species. Since "A . afarensis" is tied to a Laetoli specimen 
as holotype, only the Laetoli specimens should be designated A. africanus afarensis (though A. 
africanus tanzaniensis suggested by the author in 1978 would have been a more appropriate no­
men) and the Hadar fossils A. africanus aethiopicus. These newest East African discoveries af­
ford strong confirmation of the hypothesis that A. africanus is the common ancestor of the two 
later hominid lineages, A. robustuslboisei and Homo, leading from H. habilis through H. erectus to 
H. sapiens. 

CONTENTS 
Page 

THE EMERGING PARADIGM ON HOMINID PHYLOGENY (1950-1975) ............................................. 2 
CHALLENGES TO THE PHYLOGENETIC ROLE OF AUSTRALOPITHECUS AFRICANUS....... .... .. .. . 3 

The reality of A. africanus ... .... ..................... .... ... .... ...... ............... .. ........ .. ..... . .... .... ... ... ... ..... ....................... 3 
The East Turkana early Homo . . .. . .. . . . . . .. . . ... . . . . . . . . . . . . . . . .. . . . .. . . . . . . . . .. . . . .... . .. . . . . . . ... . . . .. . . . .. . . .. . . .. ... . .... . .. .. . . . .. . . . . . . .. . 3 
The supposed early Homo of Hadar and Laetoli.... .. .. .. .......... .. ............ .................. ............ .............. .. .. .. .. ... 4 

"AUSTRALOPITHECUS AFARENSJS"................................................................. .. ..................................... 5 
Is the "diagnosis" diagnostic? ............ .... ... ... ... .. .... .... ................................... ............... ... ................ ............. 5 

Cranium.............. . ................. ............ ...................... ... .. .. ........... ... ...................... ... .............. ...... . ... ......... 5 
Mandible............ . .......... ... ....... ...... . ............ .. ............... .. ... .. .... .. ....... ........ .. ... ...... ...... ... ....... .... .. ... .......... 6 
Dentition..... .... ... ..... ....... ....... ... ........... ... .. ..... ........ .. ... .. .. .. .. .... ... ... .................. ..... ............ ......... . .... ..... .... 6 
Summation on diagnosis . . . . . . . . . . . . .. . .. . .. .. . .. . ... . . .. . . .... . . .. . . ... . . . ... . . . ... . . ... . . . .... . . . .. . . . . .. . . .. . . . .. . .... . ... . .. . . . .. . . ... . . .. . . . 9 

Is the pooling of the Laetoli and Hadar fossil hominidsjustified? ............................ .................................. .... 10 
THE PHYLOGENETIC SYSTEM ERECTED BY JOHANSON A D WHITE .......................................... 12 

The dating of the Transvaal sites........................................................................... ............. .. ... ............... .. ... 12 
The supposed "robust" trend in A. africanus transvaalensis and at Hadar .............. .............................. .. .. .... 13 

CONCLUSION AND PROPOSAL OF AN ALTERNATIVE HYPOTHESIS .............. ...... ........................ . 14 
Laetoli ..... ... .. .. .. .. .. ........... . .. .. ................... ... .......... ........... ...... ..... .. ... ............. .... ........... . .......................... ... 14 
Hadar ........................................................................................................................................................ 14 
Summation . .. . . . . . . . . .. . . . . . .. . .. . .. . .. . . .. . . .. . . . . .... . .. . ... . . ... . . ... . ... . . .... . . ... . .. .. . . . ... . . . .... . . .... . ... . . . .. . . . .. . ... . . ... . . .. . . ... . ... . . . . . . . 14 

ACKNOWLEDGEMENTS...... ...... ...... ... ..................................................................................................... 15 
REFERENCES .. ... .. .. .... .. . .. .. ..... . ... ..... .... .. ... .... . ........ .. .. .. ... ...... ... ...... ........ .... .... ............. .................. .......... .... 15 

INTRODUCTION 
In he first 25 years after R.A. Dart ( 1925) pub­

lished an account of the Taung australopithecine 
skull, few scientists believed that this fossil, or 
those ancient hominoid specimens discovered 
shortly before and after World War II at Sterkfon­
tein, Kromdraai, Makapansgat and Swartkrans, 

represented species that should be classified in the 
Hominidae. Fewer still were those who accepted 
that these South African forms included any that 
were directly ancestral to later taxa of man, 
namely Homo erectus (known then as Pithecanthropus 
erectus) and Homo sapiens. 



2 

THE EMERGING PARADIGM ON 
HOMINID PHYLOGENY 

A change of attitude came gradually in the late 
forties and early fifties. The cardinal factors her­
alding the new consensus of the ensuing quarter of 
a century (1950-75) may be listed as follows: 
(a) the accumulation of new australopithecine 

specimens from the four Transvaal caves and 
from a growing number of East African sites; 

(b) the publication of the Transvaal Museum 
Memoirs in which R. Broom and his co-work­
ers, J.T. Robinson and G.W.H. Schepers, pre­
sented for the first time detailed accounts of 
the fossils' structure; 

(c) the studies of S. Zuckerman and his co-work­
ers, especially E.H. Ashton, on the cranial and 
postcranial skeletal morphology of the great 
apes - and, though their aim seems to have 
been to show that the Transvaal australopithe­
cines were no more than apes, in the event 
nothing proved more effectively than their data 
that the australopithecines fell outside the fam­
ily Pongidae and within the Hominidae; 

(d) J.T. Robinson's systematization of the gracile 
and robust australopithecines in respect of 
form, function, systematic and phyletic status 
and, especially, their dental morphology; and 

(e) W.E. Le Gros Clark's meticulous appraisal of 
the cranial, dental and pelvic morphology of 
the australopithecines. 

As a result, a substantial consensus was attained 
and has been sustained for several decades. On 
this paradigm all australopithecines are seen as 
hominids and, for a majority of palaeo-anthropolo­
gists, as belonging to a single genus of micren­
cephalic hominids called Australopithecus. To this 
genus are assigned those fossils that have pre­
viously been named Australopithecus, Paranthropus, 
Paraustralopithecus, Plesianthropus, Praeanthropus and 
Zinjanthropus. A second aspect of the paradigm is 
that some populations of Australopithecus are con­
sidered to have been ancestral to the genus Homo, 
particularly those of the less-specialised species, A. 
africanus. 

Mayr ( 1970: 360) states the concept in broad 
terms: " ... it is now quite evident that man's an­
cestors must have passed through an Australopithe­
cus-like stage." More explicitly, Campbell (1974: 
94-5) speaks of " ... our ancestor, Australopithe­
cus", and he adds, "The fossils that fall into this 
genus represent two species (A. africanus and A. ro­
bustus) . . . One lineage, Australopithecus africanus, 
leads on toward man ... " The ultimate expression 
of the closeness of this relationship was the sub­
mission of Robinson ( 1966, 1972) that the fossils 
assigned to A. africanus should be taken out of the 
genus Australopithecus and reclassified in the genus 
Homo as Homo africanus. 

Further elaboration of the pattern of hominid 
phylogeny has resulted from finds of fossil homi­
nids at Olduvai, Peninj and other East African 
sites and of their placement in chronological se­
quence by the K-Ar and 40Arf39Ar dating tech-

niques, fission-tracking and palaeomagnetic deter­
minations. It has been shown that the robust and 
hyper-robust australopithecines, as well as the 
earliest species of Homo, called by Leakey et al., 
(1964) H. habilis, were later in time than A. afri­
canus (fig. 1). Hence, there emerged a third aspect 
of the paradigm, namely that for considerable peri­
ods of time (from about 2,3 to about 1,0 My B.P.) 
at least two hominid lineages co-existed: indeed, 
they were not only synchronic but also sympatric 
in both South and East Africa. 

As a corollary, a fourth feature of the paradigm 
was that an early and little-specialised species of 
Australopithecus, such as A. africanus, provides the 
likeliest claimant to have been the common ances­
tor to both the Homo lineage and the A. robustus! 
A. boisei lineage. In other words Australopithecus af­
ricanus, or a species close to it, was probably the 
common ancestral hominid (Tobias 1967, 1975, 
1978a) (fig. 2). · 

My 

20 

30 

40 

50 

DISTRIBUTION OF HOMINID 
TAXA IN TIME (1979) 

H.erectus H. habilis A.africanus A.robustus A.boisei 

Figure I. The approximate distribution in time of five major 
extinct taxa of the Hominidae. This chart shows 
the dating of three species belonging to the genus 
Australopithecus and two species of the genus Homo. 
The two species on the right- A. robustus in the 
Transvaal and A. boisei in East Africa- are con­
sidered to be on a lineage which became extinct 
about a million years ago. The species on the 
left- the earlier Homo habilis and the later Homo 
erectus - are believed to be chronospecies on the 
lineage leading to the third species of Homo, namely 
H. sapiens. Homo sapiens is not shown on this chart, 
but its beginnings are set by various workers at 
500 000 to 250 000 years B.P. Australopithecus afri­
canus, in which the author believes the fossils of 
Laetoli and Hadar should be included, is the oldest 
species of hominid widely recognised as such today. 
It is the probable common ancestor to the pro­
gressive lineage on the left leading to modern man, 
and to the conservative lineage on the right that led 
to an evolutionary cul-de-sac. 



My 
0 

05 

10 

1·5 

20 

25 

30 

35 

4·0 

45 

50 

5·5 

PROVISIONAL SCHEMA OF HOMINID PHYLOGENY 1979 

H SAPIENS 

[Australial 
Amencaj 

[Europe] 
H.ERECTUS 

[Asia] 

H HABILIS 

[Africa] 

A AFRICANUS 

[Africa] 
I 
I 
I 
I 
I 
I 
L 

Figure 2. Provisional phylogenetic tree of the Hominidae 
according to the latest information available in 
mid-1979. The lightly shaded lower part of the 
common trunk represents a part of the fossil record 
where specimens are rare and not absolutely diag­
nostic, namely Kanapoi and Lothagam in northern 
Kenya. The geographical zones indicated on the 
Homo lineage are areas into which hominids moved 
from their presumed African source-area. 

Until recently, it had seemed that the pattern of 
hominid evolution had been progressively revealed 
by the researches of the fifties and sixties. A major 
degree of consensus was evident at international 
gatherings, though there remained differences of 
opinion, mainly on nomenclature and systematics 
and on relative dating, especially of the South Afri­
can sites. For little more than the last five years, 
however, new African finds and especially some of 
the interpretations offered have led several workers 
to cast serious doubts on what a few of them are 
already calling "the classical view" or "the con­
ventional wisdom", that Australopithecus is man's 
ancestor. These alternative views have had the ef­
fect also of minimizing the. role of the Transvaal 
australopithecines in hominid phylogeny. There 
follows an examination of this new development of 
the seventies, which we may jocularly dub "the as­
sault on Australopithecus africanus" . 

CHALLENGES TO THE PHYLOGENETIC 
ROLE OF AUSTRALOPITHECUS AFRICANUS 

The reality of A. africanus 
After the revelation ofthe existence of Homo ha­

bilis as a form intermediate in some key morpho­
logical characters between A. africanus and H. erec­
tus, the reality and validity of A. africanus 
transvaalensis was on several occasions questioned 

3 

by one or two colleagues. They claimed that some 
of the Sterkfontein Member 4 specimens of A. afri­
canus were, in fact, H. habilis; furthermore, the oc­
currence at Makapansgat of some robustly-built 
specimens of A. africanus was considered by them 
to connote the presence of A . robustus in Member 3 
at that site. The argument then followed these 
lines: "Take away from the supposed hypodigm of 
A. africanus those elements that are really H. habilis 
and A . robustus, and what is left of A. africanus?" 
However, this line suggested a measure of unfamil­
iarity with the immense collection of fossils of A . 
africanus transvaalensis, and we have not heard it re­
peated in the last few years. It may be noted that, 
save for A. robustus crassidens of Swartkrans, A. afri­
canus transvaalensis from Sterkfontein Member 4 and 
from Makapansgat Members 3 and 4 remains the 
largest, well-described taxon of very early fossil 
hominids, documented in great detail in numerous 
articles and monographs by Broom, Dart, Robin­
son, Schepers, Le Gros Clark, Sperber, Wallace, 
Clarke, Tobias and others. The abundance of the 
hypodigm of A. africanus transvaalensis has been 
made clear in the section on South African fossils 
(Tobias et al., 1977) in the second edition of the 
Catalogue of Fossil Hominids, Part I: Africa (Oakley 
et al., 1977). 

The East Turkana early Homo 
A more serious problem was posed by the dates 

originally assigned to the cranium KNM-ER 1470 
found in the lower member of the Koobi Fora For­
mation in 1972 (Leakey et al., 1978). This cranium 
was assigned to H . habilis (Leakey 1973, 1976). 
The dates claimed initially were "pre-2,6 million 
years" and even 2,9 My. 

Although Cooke ( 1970) had assigned a faunal 
date of 2,5-3,0 My to Sterkfontein and 
Makapansgat, being supported in this by Maglio 
( 1973), it was not generally known for some time 
that so high an antiquity had been proposed for 
the Transvaal australopithecines. Hence, the sup­
posedly 2,9 million-year-old Homo specimen from 
Koobi Fora was deemed by many workers to have 
antedated the Transvaal samples of A. africanus. If 
Homo was already in existence before A. africanus 
transvaalensis, the latter was unlikely to have been 
on the direct human line. This led an anonymous 
correspondent to claim in Nature ( 197 5) that "all 
previous theories of the origin of the lineage which 
leads to modern man must now be totally re­
vised", a claim which has recurred recently. 
Oxnard (1975) proposed specifically to deny 
to Australopithecus "a direct place it) the human 
lineage". 

The validity of this supposed refutation of cur­
rent hypotheses on hominid evolution depended 
almost entirely on two premises, namely that ER-
1470 was a member of Homo and that it was as old 
as had been claimed. On the first point there is 
much agreement that the calvaria-facial mor­
phology and the cranial capacity of 770-775 cc 
(Holloway, cited by Day et al., 1975) mark ER-



4 

1470 as a member of Homo habilis (e.g. Walker and 
Leakey 1978, Tobias l979a). However, some 
newer estimates of its age have claimed a dating of 
2,4 My for the overlying KBS tuff (Hurford et al., 
1976, Fitch et al., 1976), instead of the 2,6 My 
earlier proposed. Moreover, the work of Curtis et 
al. ( 197 5) and the recent demonstration that more 
than one tuff has, apparently in error, been called 
the KBS tuff (Gerling et al., 1979) have produced a 
date of l ,8 My for the tuff that overlies the dis­
covery-site ofER-1470. On this basis early Homo at 
Koobi Fora is no older i:han H. habilis from Oldu­
vai (Hay 1976), nor than that from the Shungura 
Formation at Omo (Boaz and Howell 1977). This 
is in keeping with the faunal evidence from Koobi 
Fora, notably that furnished by the suids (White 
and Harris 1977). Since 147-0 man is not as old as 
the dates arrived at on faunal and palaeomagnetic 
grounds for Makapansgat and Sterkfontein, the 
claim that ER-14 70 seriously challenges the place 
of A. africanus on the human lineage has fallen 
away. As Boaz and Howell (1977) put it, pre-erec­
tus Homo appears in the fossil record no earlier 
than 2,3 My. 

The supposed early Homo of Hadar and 
Laetolil 

The next claims for the existence of much earlier 
Homo in East Africa were made in 1976. The fossils 
in question came from the Laetolil Beds at Laetoli 
in the southern Serengeti Plains of northern Tan­
zania and from the Hadar site in the Afar depres­
sion of north-eastern Ethiopia. The hominid-bear­
ing Laetolil Beds have been provisionally placed 
stratigraphically between volcanic strata dated by 
the conventional K-Ar method and, in one in­
stance, the 40Arf39Ar method, to average ages of 
3,77 My and 3,59 My (Leakey et al., 1976). The 
geochronology and palaeomagnetism suggest for 
the Hadar Formation and its fauna a Gauss Epoch 
age that spanned a period from somewhat more 
than 3,1 My ago (just older than the Mammoth 
Event) to somewhat less than 2,6 My ago (Aron­
son et al., 1977). This palaeomagnetic sequence co­
incides with that for upper Member 2 to Member 
4 at Makapansgat (Brook 1977, Partridge 1979, 
McFadden et al., 1979). 

From both Laetoli and Hadar fossils have 
emerged that were initially claimed to show affini­
ties with Homo. The first publication on the Laetoli 
fossils claimed, "Preliminary assessment indicates 
strong resemblance between the Laetolil hominids 
and later radiometrically-dated specimens 
assigned to the genus Homo in East Africa. Such 
assessment suggests placement of the Laetolil spec­
imens among the earliest firmly dated members of 
this genus" (Leakey et al., 1976). However, the 
same work likened the Laetoli fossils to hominid 
fossils from South Africa, including Sts 3, 24, 50, 
51, 52 and MLD 11. All of these fossils belong to 
the hypodigm of A. africanus though this escaped 
mention in the paper cited. Subsequently, a care­
ful, detailed description of the Laetoli fossils was 

published by White (1977). No attempt was made 
in this anatomical account to compare the fossils 
with other early hominids nor to develop the 
theme of their systematic status. 

In their preliminary account of the Hadar fossils 
Johanson and Taieb ( 1976: 297) thought they 
could detect the presence of three synchronic and 
sympatric taxa: "On the basis of the present homi­
nid collection from Hadar it is tentatively sug­
gested that some specimens show affinities with A. 
robustus, some with A. africa nus (sensu stricto), and 
others with fossils previously referred to Homo". 

Subsequently, Johanson et al. (1976: 129) sug­
gested that the Hadar hominid remains could be 
"split into two clear groups. On the one hand, 
something resembling but most probably more 
primitive than A. africanus, is represented by the 
partial skeleton and the A.L. 128-129 specimens. 
On the other hand, the presumed Homo group may 
represent some of the earliest evidence known thus 
far, for the lineage which ultimately led to modern 
man". Three points are worthy of note. First, they 
prefaced their interim conclusions by the 
statement that the Hadar hominid remains had 
not at that stage been subjeCted to extensive de­
tailed studies. Secondly, they supported their 
claim that Homo was represented at Hadar by "the 
closeness in morphology of certain mandibles ... 
with OH 7 (jaw of the type specimen of H. habilis 
from Olduvai) and the new Laetolil specimens ... " 
Qohanson et al., 1976: 129) (italics mine). The 
case for the latter belonging to Homo was then, and 
is now, unproven; so the comparisons with Laetoli 
could not be taken to strengthen their claims for 
the presence of Homo at Hadar. Thirdly, they 
again drew attention to robust elements in certain 
of the Hadar fossils U ohanson et al., 1976: 128-9): 

" ... it is possible that A.L. 211-1 and A.L. 
166-9 may represent an early occurrence of a ro­
bust australopithecine lineage. The somewhat 
atypically robust character of the temporal may 
be considered as consistent with the idea that 
the early stages of the robust lineage are being 
sampled." 
These remarks are especially interesting in view 

of the later claim of Johanson and White ( 1979) 
that the Transvaal A. africanus (but not the Hadar 
fossils) is already on the robust lineag·e. 

However, Johanson et al. (1976: 129) went on to 
say, "For the moment a strong case cannot be 
made for the presence of a robust hominid in Ha­
dar and should be considered as a preliminary 
suggestion". 

Thus the original claim that there were three 
elements, including Homo, at Hadar, had in the 
same year been replaced by a proposal that there 
were effectively "two clear groups". A year later 
Coppens and Johanson ( 1977) indicated that they 
were prepared to entertain the hypothesis that all 
of the Hadar hominid material could have be­
longed to a single taxon though they did not spe­
cify which taxon. 

The view that Homo was present at Laetoli and 



Hadar was strongly contested at a number of in­
ternational meetings at Cambridge ( 1976), Nice 
(1976), Paris (1977), Nairobi (1977), Hamburg 
( 1977) and Karlskoga ( 1978) (Tobias 1978a, 
1978b, 1979b, 1979c, 1979d). Detailed analysis of 
all the published data on the Laetoli and Hadar 
fossil hominids as well as a preliminary study of 
the casts and some of the originals kindly made 
available by the discoverers, led me to several con­
clusions: 
(a) Hadar: The published evidence does not allow 

one to confirm that the 3 million years old 
hominids of Hadar include among them any 
that should be referred to Homo; the picture 
afforded by these fossils rather suggests an 
Ethiopian population of A. africanus. 

(b) Laetoli: The published evidence, including the 
descriptive detail of White ( 1977), leads to the 
conclusion that the early hominid fossils of 
Laetoli closely resemble those of A. africanus 
and can be accommodated comfortably within 
that species, perhaps with some minor modifi­
cation of the sample ranges of variation for sev­
eral dental metrical traits. 

(c) Both the Laetoli and Hadar fossils of the 3 to 
3, 75 My time-range provide us with East Afri­
can samples of A. africanus very similar to and 
approximately contemporary with, or slightly 
older than, the Transvaal populations of that 
spectes. 

(d) The place of A. africanus in time and space is 
now confirmed and strengthened by the Hadar 
and Laetoli finds, which have added greatly to 
the probability that A. africanus was the ances­
tor of both the later forms, the robust australo­
pithecines and the earliest species of Homo 
(Tobias 1978b). 

By 1978 Johanson, Coppens and White had 
dropped the idea that Homo was represented at 
either Laetoli or Hadar and had accepted the view 
that the hominids from the two sites belonged in 
the genus Australopithecus and that they were close 
to, if not identical with, the southern African taxon 
A. africanus transvaalensis. 

"A USTRALOPITHECUS AFARENSIS" 
Although recognising that the Hadar and Lae­

toli fossils are very similar to A. africanus, Johanson 
and Coppens ( 1978) formally proposed that these 
fossils be placed in a separate species, "A. afaren­
sis". The first preliminary account of the pos­
tulated new species was published on 1st June 
1978 by Hinrichsen (1978), but its formal diag­
nosis was subsequently published over the names 
of Johanson, White and Coppens ( 1978). Further 
details were published by Johanson and White 
( 1979). The remainder of this article is devoted to 
a critique of this proposal. 

Is the "Diagnosis" diagnostic? 
An anonymous "Palaeoanthropology Corre­

spondent" pointed out in Nature (29 March 1979) 
that the success of Johanson and White "in 

5 

demonstrating the morphological distinctiveness of 
A. afarensis is by no means clear". The correspon­
dent added that "very few characters of A. afarensis 
are not also found in A. africanus". Moreover, 
"there are very few features said to be characteris­
tic of A. afarensis which are, in fact, distinctive of 
that group". 

This coincides with the view I reached after I 
had examined every feature in the published diag­
nosis and compared it with the corresponding trait 
in the Sterkfontein and Makapansgat fossils of A. 
africanus. A suite of eight cranial, five mandibular 
and eight dental features from the published diag­
nosis is examined here. 

Cranium 
It is claimed that "A. afarensis" has "strong al­

veolar prognathism with convex clivus". Marked 
alveolar prognathism is a well-known feature of 
Sterkfontein crania as exemplified by TM 1511, 
TM 1512, Sts 5 and Stw 13. Convexity of the cli­
vus is a continuously varying trait, and it is pres­
ent in diverse degrees in Sterkfontein maxillae 
such as TM 1512, Sts 17, Sts 71 and Stw 73. The 
convexity in the latter specimens is both transverse 
and vertical, as in the Hadar maxillae of A.L. 
200----l a (fig. 3). 

"Dental arcade (of "A. afarensis") long, narrow, 
straight sided". There are few complete palates 
from Sterkfontein or Makapansgat, but it is clear 
from those preserved that the same description 
would have pertained to these. There are of course 
markedly varying degrees of prognathism and so of 
palatal lengthening within early hominid taxa. 

"Compound TIN crest in larger specimens". 
Such a crest is well-known in A. boisei (Tobias 
1976, Day et al., 1976) and its probable devel­
opment in larger Transvaal gracile australopithe­
cines was hinted at by the position in MLD 1, the 
first hominid specimen from Makapansgat, in 
which the least distance apart of the inferior tem­
poral and superior nuchal lines is l mm. As the 
latter specimen is probably mature though not 
aged, it is likely that with further development of 
the individual a small compound TIN crest would 
have formed, just as it is possible that this cranium 
might have developed a sagittal crest had the sub­
ject lived longer into middle life. Tobias ( 1967: 
23-4) pointed out that compound T I N crests 
should be expected in australopithecines generally, 
especially adult males with big teeth and heavy 
musculature. The presence of a compound T I N 
crest in the strongly-muscled A.L. 333-45 (though 
not on the less robust A.L. 288-l) is precisely 
what one would have expected in a well-muscled, 
small-brained australopithecine and what is likely 
to have developed in MLD l and others of the 
Transvaal A. africanus in the event of survival to 
older adulthood. This developmental variable can­
not be accepted as a specific diagnostic feature. 

"Shallow mandibular fossae, with weak articu­
lar eminences placed only partly under brain­
case". All three of these features are absolutely 



6 

Figure 3. Maxillae ofaustralopithecines seen from above. On both specimens the floor of the nasal cavity is present in the centre, 
flanked on either side by the floor of the maxillary sinus. 
Left: Stw 73, a specimen of A. africanus from Member 4 at Sterkfontein. 
Right: A.L. 200-la, a specimen of "A. afarensis" (or in the author's view A. africanus aethiopicus) from Hadar. 
In both specimens the sub-nasal or nasa-alveolar clivus is well-rounded, from above downwards and from side to side. A 
convex clivus has been included by Johanson et at. (1978) as diagnostic of "A . afarensis". 

characteristic of australopithecines in general and 
indeed do not serve to distinguish among A. afri­
canus, A. robustus and A. boisei (Tobias 1967, 1979a) 
let alone "A. afarensis". 

"Occipital condyles with strong ventral angu­
lation". Such high convexity of the condyles 
characterises A. boisei and other australopithecines 
(Tobias 1967: 27). 

Mandible 
"Ascending ramus of mandible broad, not 

high". Unfortunately, in this as in other such 
"diagnostic" traits, no measurements have been 
given in the diagnosis. Nevertheless, broad rami of 
moderate height are characteristic of A. africanus 
as, for example, are shown by Sts 36 (and see Dart 
1954, 1955). 

"Moderate superior transverse torus" . This very 
variable character is to be found in a number of 
Transvaal A. africanus jaws (e.g. Broom et al., 
1950, Dart 1954). 

"Low rounded inferior transverse torus". This 
description fits extremely well most A. africanus 
mandibles. 

"Anterior corpus rounded and bulbous". This 
variable trait characterises the symphyseal region 

of a number of the Transvaal australopithecine 
mandibles. 

"Strong posterior angulation of symphyseal 
axis". Again, no measurements of the symphyseal 
angle are cited; once more this is a variable trait. 
Yet strong posterior angulation of the symphysis 
certainly characterises most A. africanus and A. ro­
bustus jaws (e.g. Broom et al., 1950: 35, Dart 1955, 
Tobias 1976). 

Dentition 
"Upper central incisors relatively and absolutely 

large": Johanson and White ( 1979) have published 
means, standard deviations and ranges of both 
crown diameters for each tooth type in the combined 
Hadar plus Laetoli sample. Sample sizes are given, 
but it is not clear whether this refers to the number 
of individuals, or the number of teeth. If both teeth 
are present in a single jaw (e.g. the two upper cen­
tral incisors in the maxilla of A.L. 200-1a), then in 
an overall small sample this one specimen may 
bias the mean if both teeth are included in the 
sample each in its own right. A common practice 
in such cases is to use the mid-values for the calcu­
lation of the means, though of course the maxi­
mum or minimum individual values are used in 



statemen_ts of the sample range. Furthermore, the 
explanatiOn of Table 1 of Johanson and White 
(1979) makes clear that the M.D. diameters for 
postcanil?~ tee\~ have ~een corrected for interprox­
Imal attntwn, except m cases where that was im­
possible". It would appear from this wording that 
the authors have included data for some post­
canine teeth that they were unable to correct for 
interproximal attrition: if so, their inclusion would 
lower the means for postcanine teeth. Moreover 
since it is usually possible without difficulty to cor~ 
rect for interproximal attrition on incisors and can­
ines, the authors' failure to do so might have intro­
duced a small bias to the mean, which may affect 
comparisons with mean metrical traits of other 
samples. 

Table 1 compares the published mean data for 
the ~o~bined ~aetoli-Hadar sample of upper cen­
tral mCisors with the corresponding means for the 
A. africanus transvaalensis sample based upon Sterk­
fontein Member 4 and Makapansgat. In the 
Transvaal sample all teeth have been corrected for 
interproximal attrition. Where a pair of antimeres 
is involved, the mid-value between the left and 
right teeth is used for computation of the mean 
and the pair of such teeth counts as only one item 
in the given "n" where "n" refers to individuals. 

TABLE 1 
Size of Upper Central Incisors 

Sample Ranges and Means of Metrical Characters 

A. africanus 

( 4 teeth) 
"A. afarensis" 

n 

M.D. B.L. Module 
Diameter Diameter (M .D. + B.L. ) 

(mm) (mm) 
9,3-cl2,0 8,2-c9,6 

10,60 8,93 
3 2 

9,0-11,8 7, 1-8,6 

2 
8,80-c I 0, 70 

9,77* 
2 

x I 0,36 8, 16 9,26* 
n 4 5 (4-5) 

* The mean module for A. africanus refers to the mean of the 
modules of individual teeth, but for " A. afarensis" it is the 
module of the means, since the diameters for all individual 
teeth have not been published . 

Small as are the samples, the data in the Table 
show that the supposed large size of the Laetoli­
~adar I_1 does n_ot distinguish it from that of A. af­
ncanus; If anythmg, on the available evidence the 
P of the latter is slightly larger. If the M.D. diam­
eters of ~~e Hadar P were corrected for approxi­
mal attntwn, and the published illustration sug­
gests that at least _some correction is necessary 
U?hanson and White 1979: fig. 2), this would 
b~mg the Hadar mean slightly nearer to parity 
with that of the Transvaal series though the mean 
B.L. diameter would remain in arrears. In a word 
the supposed relatively and absolutely large size of 
the I 1 does not distinguish "A. afarensis" from A. 
africanus. 

"Diminutive lateral incisors". Table 2 shows 
that small upper lateral incisors characterise both 
the S~erkfontein-Makapansgat sample and the 
Laetoh-Hadar sample. On the available samples, 

TABLE2 
Size of Upper Lateral Incisors 

Sample Ranges and Means of Metrical Characters 

A. africanus 
9 teeth 
(I immature, 
excluded) 
"A . afarensis" 

n 

M.D. B.L. 
Diameter 
5,8-c7,5 

6,77 

6 
6,7-8,2 
(worn) 

Diameter 
5,6-c7,3 

6,89 

6 
6,2-8, 1 

Module 
5,70-c7,90 

6,83* 

6 

x 7,65 7,18 7,41 * 
n 6 8 (6-8) 

7 

* " Mean module" for A . africanus refers to mean of modules 
and for " A . afarensis" module of means. 

the A. africanus lateral incisors are somewhat 
smaller than those of "A. afarensis". 

"S_trong variat!on in cani?e size, canines asym­
metnc, lowers with strong lmgual ridge": To take 
the two l~tter ~orphological features first, both 
asymmet~Ic. canmes and strong lingual ridges are 
charactenstics of A. africanus (fig. 4) which Robin­
son ( 1956) long ago brought to notice. For exam­
ple, he illustrated and pointed out the asymmetry 
of the cro""ns of maxillary canines (Robinson 1956: 
45) and of mandibular canines (p . 49) of A. afri­
cqnus; an~ he drew. sp,~cial attention to the lingual 
ndge which he said appears to be a prominent 
feature of the Sterkfontein mandibular canines" 
(pp. 49-50). Strong variation in size has been 

Figure 4. Lingual aspect of the crown of mandibular left can­
ine Stw 21, of A . africanus from Sterkfontein Mem­
ber 4. Note the obvious asymmetry of the crown 
and the strong lingual ridge, both of which features 
are said to be diagnostic of "A . afarensis" (Johanson 
et al., 1978) . 



8 

shown to be a mark especially of canines, in most 
primates, including modern man (Gonda 1959, 
Mijsberg 1931) and the australopithecines (Wol­
poff 1975, 1976); and it is known that the canine is 
generally the most sexually dimorphic tooth. 
Therefore the amount of variability in canine size 
in a fossil hominid sample may be influenced 
greatly by the sexual composition of the sample. 
This would be especially likely with the small sam­
ples that palaeo-anthropologists generally have at 
their disposal. 

Using the means and S.D.s furnished by Johan­
son and White ( 1979) and mindful of the afore­
mentioned uncertainty whether bilaterally repre­
sented teeth have been counted twice or only as a 
mid-value in the calculation of the mean, I esti­
mated the coefficient of variation for crown diame­
ters of upper and lower canines in the Laetoli­
Hadar sample, in A. africanus transvaalensis and in 
A . robustus crassidens ( Swartkrans Member 1) 
(table 3). 

TABLE 3 
Canine Variability 

Upper Canines Mean S.D. C.V. 
mm % 

M.D. diameter 
A. africa nus transvaalensis 9,61 (6) 0,60 6,24 (6) 
"A. afar ens is" 9,92 (10) 0,74 7,46 (10) 
A. robustus crassidens 8,54 (12) 0,40 4,68 (12) 

B.L. diameter 
A . africanus 9,90 (6) 1,16 11,72 (6) 
"A. afar ens is" 10,94 (10) 1, 11 10,15 (10) 
A. robustus crassidens 9,38 (12) 0,94 10,02 ( 12) 

Lower Canines 
M.D. diameter 
A. africa nus 9,11 (10) 0,45 4,94 (10) 
"A. afarensis" 9,16 (5) 1,54 16,81 (5) 
A. robustus crassidens 8,01 (4) 0,33 4,12 (4) 

B.L. diameter 
A. africanus 9,93 (II) 0,64 6,45 ( II) 
"A. afarensis" 10, I 7 (9) 1,15 11 ,3 1 (9) 
A. robustus crassidens 8,53 (4) 0,62 7,27 (4) 

The "A. afarensis" and A. africanus means are 
very close to each other, differing by a small frac­
tion of a millimetre in all instances save for the 
B.L. diameter of upper canines where the differ­
ence of the means is 1,04 mm. The means for A. ro­
bustus crassidens reveal the small size of its canines, 
upper and lower. A glance at the differ.ences of the 
means and at the S.D.s shows that none of the dif­
ferences between the means of "A. afarensis" and 
A. africanus is statistically significant. 

For the upper canines the Laetoli-Hadar C.V.s 
are comparable to those of the Transvaal samples: 
in M.D. diameter "A. afarensis" is slightly more 
variable than A. africanus, and the reverse is true 
for the B.L. diameters. In keeping with previous 
studies that have shown the greatest degree of 
dimorphism to be in the B.L. diameter, for each of 
the three taxa the C.V. for B.L. diameter is greater 
than for M.D. diameter in the maxillary canines. 

When one turns to the lower canines, however, 
"A. afarensis" shows an astonishingly high value 
for the C.V. of the M.D. diameter, not only higher 
than any other C.V. in the entire table, but con­
trary to expectation higher also than the C.V. for 
the B.L. diameter of the same sample. This high 
C.V. of 16,81 flows from the very high S.D. of 1,54 
for a sample of M.D. diameters stated to be only 5, 
whereas the sample for B.L. diameters of lower 
canines from Laetoli-Hadar is given as 9. The high 
S.D. is obviously related, in part, to the small sam­
ple size. There is a clear-cut discrepancy within 
the Laetoli-Hadar sample, the M.D. diameters be­
ing more highly variable than the B.L. diameters, 
whereas the opposite is the case for all other sets of 
data in Table 3 and in most previous studies. It is 
difficult to resolve this problem, since with a few 
exceptions the original measurements for individ­
ual Hadar teeth have not been published. Sec­
ondly, we are told that the M.D. diameters of the 
front teeth have not been corrected- and it is a 
set of M.D. diameters of canines that have yielded 
the discrepant results. Thirdly, we do not know 
whether and to what extent the results have been 
influenced by the inclusion of both members of a 
pair of antimeres in the calculation of the mean, 
such as the two lower canines in the Hadar jaw 
A.L. 400-la. Fourthly, one questions whether the 
high C.V.s for the mandibular teeth of Laetoli-Ha­
dar do not betray that more than one population 
has been sampled, a not unreasonable query, since 
the data quoted by Johanson and White are for 
two site-samples that have been pooled- vide in­
fra. Lastly, the internal discrepancy compels one to 
ask whether a computational or typographic error 
might have crept in to their published table of 
means. 

Save for this discrepant value, the morphology 
and variability of the canines do not distinguish 
between "A. afarensis" and A. africanus. 

"P3 occlusal outline elongate oval in shape with 
main axis mesiobuccal to distolingual at 
45 °-60 o to tooth row, dominant . . . buccal 
cusp, small lingual cusp often expressed only as in­
flated lingual ridge": The shape as described is ex­
actly as we find it in many of the Sterkfontein and 
Makapansgat P3s. Here are Robinson 's ( 1956: 
72-73) very words as he describes the Sterkfontein 
P3s: "The crown thus has a skewed appearance in 
occlusal view, with the longest diameter running 
from the distolingual to the mesiobuccal angle". 

It would have been helpful to know what num­
ber of Laetoli-Hadar P3s had moderate, small and 
very small lingual cusps: since this is a most vari­
able feature in populations of early hominids. Of 
seven excellent casts of Hadar mandibles kindly 
made available by Dr. Johanson , all of which show 
the unworn lingual cusps of P3, two show moder­
ately high lingual cusps: of these it is highest in 
A.L. 333W-la where the buccal and lingual cusps 
are sub-equal; the pattern is similar to that in the 
Sterkfontein jaw SL 52b. The lingual cusp is mod­
erately high on both left and right in A.L. 400-la. 



Only two of the seven mandibles have P3s with 
very low lingual cusps: A.L. 128-23 and A.L. 
288-11. Coppens (1977: 1301) singles out the latter 
as being "pratiquement monocuspide". These ob­
servations suggest that the expression of the lin­
gual cusp in the Hadar population is very variable, 
as in A. africanus transvaalensis. On the three Laetoli 
P3s Wolpoff ( 1979) speaks of "a dominant buccal 
cusp, a weaker somewhat lower lingual cusp", 
which implies a better developed lingual cusp than 
the form Johanson et al. ( 1978) describe as "often 
expressed only as inflated lingual ridge". P3s with 
very small lingual cusps are to be found in the 
Transvaal samples; for example the P3s of MLD 2, 
the first jaw of A. africanus found at Makapans­
gat (Dart 1948a). Of course there may be a differ­
ence in the frequency of small lingual cusps be­
tween the three populations , but this would have 
to be very marked before it could be regarded as a 
diagnostic trait. On presently available evidence 
the range of characteristics of the P3s of Hadar and 
Laetoli do not distinguish between "A. afarensis" 
and A. africanus. 

"Postcanine (mandibular) teeth aligned in 
straight rows". This feature is to be found in A. af­
ricanus transvaalensis, for example Sts 7 (Broom et 
al. , 1950), MLD 18 (Dart 1962) and Stw 14 
(Tobias 1973b). It is therefore not diagnostic. 

Summation on diagnosis 
Feature after feature listed in the "Diagnosis" of 

"A. afarensis" by Johanson et al., ( 1978) has proved 
to be not diagnostic; and working systematically 
through the remaining half-dozen enumerated 
traits, I have been led to the same conclusion 
about them. In a word "A. afarensis" cannot be 
separated from A. africanus on the characters that 
have been adduced as the diagnosis . 

From the tooth measurements for Laetoli pub­
lished separately by White ( 1977) the author was 
able to show that there are indeed some slight 
odontometric differences between the fossils of 
Laetoli and those of A. africanus. The only such dif­
ferences are that the teeth of Laetoli are slightly 
larger on the average, and the mandibular pre­
molars of Laetoli, though larger, are longer and 
narrower than those of A. africanus (Tobias 1979d). 
These tenuous differences seem to me to be insuffi­
cient evidence on which to separate the Laetoli 
hominids from Transvaal A. africanus at the species 
level. There may well be a case to separate them at 
the subspecies level, since, aside from their small 
dental differences, they are 0,6-1,0 My earlier than 
Makapansgat Member 3 and Sterkfontein Mem­
ber 4 and 2 200--2 500 km apart. 

Other descriptive details have been placed on 
record by Johanson et al., (1978) though without 
these points being listed as part of the formal diag­
nosis. These traits, too, do not strengthen the case 
for separate specific status. As was pointed out in 
the anonymous comment in Nature, such ad­
ditional features as the "waisted" capitate bone, 
the lack of a styloid process on the third metacar-

9 

pal and the features of the pelvis of "Lucy" (A.L. 
288-1) are extremely similar to corresponding 
traits from Sterkfontein. The same applies to the 
strong molar size gradient M3> M2> M 1 which 
typifies Australopithecus. Strong dimorphism in 
body size is said to characterise the Hadar collec­
tion, but it is true too of the Transvaal australopi­
thecines. Wolpoff (1975) was led to conclude that 
in the gracile australopithecines of the Transvaal 
the male is considerably larger than the female. 
Apart from the evidence adduced by him we may 
point to the great variation of vertebral size be­
tween the very small vertebrae of the putative fe­
male Sts 14 and the much larger ones comprising 
the still articulated spine Stw 8/4 1, derived from 
the same Member 4 at Sterkfontein: these ver­
tebrae point to a strong degree of sexual dimor­
phism in body size (bulk) in A. africanus transvaalen­
sis (fig. 5). 

Figure 5. These two partial vertebral columns of A. africanus 
from Sterkfontein Member 4 reveal the great size 
dimorphism present within the population. 
Left: Sts 14, probably female . 
Right: Stw 8/ 41, probably male. 
Marked sexual dimorphism has been claimed to be 
a special feature of the Laetoli-Hadar fossils for 
which the nomen "A. afarensis" has been proposed. 

In short the Laetoli and Hadar h01;ninids seem 
to have provided us with the first good East Afri­
can populations of A. africanus, for it has been a 
curious feature of the record of discovery that the 
luxuriant fossil sites of East Africa have until lately 
not yielded any series that bears direct and close 
comparison with A. africanus, save for isolated teeth 
from Omo and possibly for the incomplete man­
dible of Lothagam and humeral fragment from 
Kanapoi. The claims for their distinctiveness 
notwithstanding, the Laetoli and Hadar fossils 
seem at last to have filled this gap. 



10 

Is the pooling of the Laetoli and Hadar fossil 
hominids justified? 

In creating the supposed taxon "A afarensis" and 
in further justifying it, Johanson and his co-work­
ers have pooled the two site-samples of Hadar and 
Laetoli. In vindication of this procedure, they 
speak of "The strong morphological and chrono­
logical continuity" between these two site-samples 
Oohanson and White 1979: 321). 

The supposed chronological continuity is not 
immediately apparent. For the Hadar Formation 
Aronson et al., ( 1977: 327) have arrived at a range 
of dates which extends "from somewhat more than 
3,1 My ago to somewhat less than 2,6 My ago". 
The ages of the hominid-bearing strata at Laetoli 
have been bracketed between 3,59 and 3, 77 My 
B.P. (M.D. Leakey et al., 1976). 

Thus the latest hominid at Laetoli and the old­
est at Hadar are separated in time by something 
like 0,4 My. If we take the mid-value of the range 
of datings for each site, we have values of about 
3,68 My at Laetoli and of about 2,85 My at Ha­
dar, these mid-range values being separated by 
0,83 My. 

Even within the Hadar Formation the hominids 
are distributed in three Members of different ages. 
At a rough estimate based on the lists of specimens 
published by Joh.anson, White and Coppens 
( 1978), some 11 % of Hadar hominids are derived 
from the Sidi Hakoma Member. This is the oldest 
member, and its dating is said to be older than 
c3,0 My, but less than c3,3 My. Thus, only a small 
minority of the oldest Hadar fossils are within 
about 0,4 My of the most recent date for the Lae­
toli fossil hominids. 

Most of the Hadar hominids are derived from 
the two younger Members, Denan Dora and Kada 
Hadar, both of which are dated to younger than 
c3,0 My though not more recent than c2,6 My. 
The older of the two, the Denan Dora Member, is 
the one that is richest in hominids, including the 
supposed family group that constituted one of the 
more remarkable finds at Hadar. At a rough esti­
mate, 85 % of the Hadar hominid specimens come 
from this Denan Dora Member. The remaining 
Kada Hadar member seems to be 2,8 My old, or 
just less, and has yielded one hominid individual, 
the 40 % complete skeleton of A.L. 288-1 known 
as Lucy. 

There is thus a range in time of at least 0,5 My 
between the oldest and the youngest hominids 
from Hadar. This is comparable with the mini­
mum estimate of the lapse of time between Mem­
bers 4 and 5 at Sterkfontein and much less than 
the interval between Members 1 and 2 at Swart­
krans. Although the Hadar hominids are dispersed 
through these three Members, they have been 
pooled and treated as one sample by Johanson and 
his co-workers. Then, this entire site-sample has 
been pooled with the still earlier hominids of Lae­
toli, the oldest of which may be 1,0 My older than 
the youngest from Hadar. That is, the combined 
sample encompasses a million years of possible 

morphological change- and even within a single 
evolving lineage this may be appreciable. For such 
pooling to be justified, it would be necessary for 
the hominids from each Member in the Hadar 
Formation to be first analysed separately; it would 
need to be shown that no appreciable morphologi­
cal differences distinguished the hominids in the 
earlier and later parts of the half million years or 
more that elapsed during the deposition of the 
various Members of the Hadar Formation. This 
might have been done by Johanson and his col­
leagues, but nothing to this effect is mentioned in 
any relevant publication. Also, the data from each 
Hadar Member sub-sample have not been pre­
sented separately with a statistical evaluation of 
the mean measurable characters and of the signifi­
cance of any differences that may be detected by 
such analyses. Once it had been shown that the 
fossil hominids from the three Members at Hadar 
were indeed "poolable", one should then by a 
similar procedure evaluate the morphology and 
measurable differences between the Hadar sample 
in toto and that of Laetoli: if the differences were 
sufficiently small, it could be adjudged that the 
two site-samples could be pooled. In this way each 
of the site-samples of Sterkfontein and Makapans­
gat was analysed separately, and only after i.t was 
proved that they shared a common morphological 
pattern did Robinson ( 1954) formally propose that 
they be grouped as two sub-sets of a single sub­
species, A. africanus transvaalensis. 

With "A. afarensis", there are no published indi­
cations, data nor statistical analyses to show that 
such a procedure has been followed. Early and late 
sub-samples and site-samples have simply been 
pooled and treated as one population, comprising 
a supposedly distinctive lineage and taxon. Had 
the describers of the proposed new taxon manifestly 
followed this time-honoured and well-validated 
procedure, and, had statistical tests not militated 
against such pooling, the case for a new taxon 
would have been more convincing. 

To the chronological difficulties just mentioned 
should be added their geographical separation 
by 1 600 km, a sufficiently great distance for geo­
graphical subspeciation or "raciation" to have 
occurred even if the two groups had been syn­
chronic. With major differences of time and space 
between the Hadar sub-samples and the Laetoli 
assemblage, the chances of the two site-samples 
belonging at least to different subspecies are very 
great. However, only the morphological analysis of 
the various sub-samples will be able to prove 
whether subspeciation had indeed occurred. 

The manner In which the preliminary morpho­
logical accounts of the Hadar fossils have been 
published makes it difficult to determine what dif­
ferences exist between the Tanzanian and Ethio­
pian samples. Although some Laetoli fossils have 
been described in detail (White 1977), the data 
published separately for the earlier-discovered Ha­
dar specimens do not allow metrical comparison 
with the Laetoli fossils. For example, it is not clear 



whether the published mesiodistal diameters of the 
Hadar teeth had been corrected for approximal at­
trition U ohanson and Taieb 1976, Johanson, Cop­
pens and Taieb 1976); yet without such cor­
rections comparisons of dental measurements of 
hominids are largely vitiated (Wolpoff 1971, 
Tobias l978b). 

The present appraisal has already referred to 
the large coefficient of variation of the mesiodistal 
diameters of lower canines of the Laetoli-Hadar 
combined sample, and the query was raised 
whether in fact more than one population has been 
sampled. 

It has proved possible to calculate approximate 
means for some dental traits for the Hadar popu­
lations alone (table 4) from published data (White 
1977, Johanson and White 1979). In the Table only 
metrical traits based on cheek-tooth measurements 
have been used , since these are the only ones for 
whichJohanson and White (1979) have made cor­
rections for approximal attrition. The approximate 
Hadar means in the Table differ appreciably from 
the Laetoli means, whether we consider mesiodis­
tal or buccolingual diameters, crown areas or the 
sum of crown areas, or shape index values 
(M.D./B.L. X 100 %) . It is noteworthy that for 
all metrical characters cited in the Table the 

TABLE4 
Some Estimates of Odontometric Contrasts Between 

Laetoli and Hadar* 

Mandibular 
Tooth Material 
(P3-M3 ) 
M.D. diam . P3 
B.L. diam. M 1 

B.L. diam. M 3 

Crown Area P3 
Crown Area M 1 

Shape Index P3 
Shape Index P4 

A . africanus Laetoli Hadar Laetoli-Hadar 
transvaalensis " A. afar ens is " 

860,66 868,61 c732,52 748,61 
9,61 (5) 11 ,60 (2) c9, 16 9,51 (14) 
13,8 ( 15) 14,5 (3) cl2 ,60 13,23 (9) 
14,2 ( 14) cl4,2 ( 1) cl 3, 14 13,23 (12) 

114,10 (15 ) 120,08 (2) c95, 12 102,25 (7-8) 
181 ,01 (9) 181 ,02 (3) cl58,13 162,17 (16-18) 

82,57 (5) 11 2,45 (2) c85 ,93 89,72 (14) 
83,04 (8) 94,46 (3) c85,38 87 ,65 (12-13) 

* The means for A . africanus transvaalensis are the latest results 
estimated by me and based upon enlarged samples that in­
clude most of the newest finds from the Transvaal sites . The 
means for Laetoli above have been prepared by me from 
mean diameters given by White ( 1977). The mean diam­
eters for Laetoli-Hadar are taken from Johanson and White 
( 1979), and the derived measures and indices have been 
based upon the published means . The values for Hadar 
have been estimated from those for Laetoli-Hadar com­
bined and Laetoli alone, since data for individual Hadar 
teeth are generally not available, and those published might 
not have been corrected for interproximal attrition . Man­
dibular tooth material is my modification of Howes's (1954) 
tooth material concept: it is the sum of the crown areas 
(M .D . X B.L. ) of the five mandibular permanent cheek­
teeth (P3-M3). In parentheses " n " refers in the first two col­
umns not to the total number of teeth available, but to the 
number of individuals from each of whom one or two 
measurable representatives of the tooth-type in question are 
available. For Laetoli-Hadar "n" is as published; it is not 
known if it refers to teeth or to individuals. 

BP-8 

II 

means for Laetoli, as computed for White 's ( 1977) 
individually recorded tooth measurements, give 
higher values than the means cited by Johanson 
and White ( 1979) or derived from their mean 
values for the Laetoli-Hadar combined sample. 
Thus, the mean values for measures of size of the 
Hadar teeth alone are all smaller than the means 
for the combined sample. 

The relatively small size of the Hadar teeth in 
turn must have a major effect on means for the 
combined sample, at least 80 % of which seems to 
comprise Hadar teeth. Their smallness is vividly 
protrayed when one compares the Cheek-Tooth 
Material (summed crown areas of P3 to M3) for 
various fossil populations (table 5). Laetoli alone 
has a mandibular tooth material of 868,61, very 
close to the mean for A. africanus (860,66). 

TABLE 5 
Cheek-Tooth Material of Hominids 
Summed Crown Areas of P3 toM3 

A . boisei (several E . African sites) 
A. robustus crassidens (Swartkrans) 
A. robustus robustus (Kromdraai ) 
Laetoli 
A . africanus transvaalensis 
(Sterkfontein and Makapansgat) 

Maxillary 
1 230,88 

990,67 
929,08 

* 
819,70 

Mandibular 
1 312,00 

960,13 
882,10 
868,61 
860,66 

H. habilis (several E. African sites) 743,48 786,59 
Laetoli-Hadar 734,39 748,61 
Hadar 686,61 t 732,52t 
H. erectus erectus 641, 13 695,34 
H. erectus mauritanicus 665,03 
H. erectus pekinensis 577,07 607,66 
H . erectus heidelbergensis 543,65 
H. sapiens sapiens 471 ,80 484,83 
* No measurements for a P4 from Laetoli have been pub­

lished, so it has not been possible to compute its maxillary 
tooth material. 

t Figures for Hadar have been computed from estimates based 
on the means for the combined Laetoli-Hadar sample and 
for the separately-published Laetoli teeth. 

The values for Laetoli-Hadar combined and 
those estimated for Hadar are so low as to set 
them below the tooth-material of H. habilis. This 
position is decidedly odd when one notes that the 
Laetoli component of the combined sample has 
tooth material somewhat larger than that of A. af­
ricanus. These relative placements make the com­
bined sample look positively heterogeneous. Again 
one has cause to ponder the "poolability" of Ha­
dar and Laetoli and to consider whether they do 
not represent different populations. 

In five out of six size parameters (table 4), the 
means for Laetoli alone are closer to the A. afri­
canus transvaalensis means than are the estimated 
means for Hadar alone. There is near identity be­
tween the Laetoli means and the Transvaal means 
for mandibular tooth material, buccolingual dia­
meter of M 3 and the Crown Area of M 1; the values 
are close for the buccolingual diameter of M 1 and 
the Crown Area of P3• Only for the mesiodistal di­
ameter of P3 is the Hadar estimated mean closer to 
the A . africanus mean. On the other hand, as men-



12 

tioned above, in shape measures the mandibular 
premolars of Laetoli differ appreciably from those 
of A. africanus, and the values estimated for Hadar 
are very similar to those for the Transvaal. 

Approximate as is this method for the testing of 
metrical similarity between Laetoli and Hadar, it 
does at least confirm what the variability had 
earlier shown, namely that there is a strong prima 
facie case to examine more closely the characteris­
tics of the Laetoli and Hadar samples and sub-
samples separately. It provides us with some justi­
fication for the feeling already voiced that the two 
site-samples may be samples of different popu­
lations. How different those populations might 
have been is altogether another question; the den­
tal metrical differences may bespeak no more than 
geographical and chronological subspeciation 
within a single lineage. Nevertheless the differ­
ences between Laetoli and Hadar are great enough 
to deter us from easily accepting the pooling of the 
two site-samples for purposes of statistical evalu­
ation or for taxonomic appraisal. 

THE PHYLOGENETIC SYSTEM ERECTED 
BY JOHANSON AND WHITE (1979) 

Johanson (1978a) and Johanson and White 
( 1979) have proposed that "A. afarensis" is directly 
ancestral to (a) H . habilis and (b) a branch leading 
via A. africanus to A. robustus. That is, they consider 
that A. africanus is off the main line of hominid evo­
lution leading to Homo, and they regard it as 
already differentiated in the direction of the robust 
australopithecines. This view involves judgements 
on both the morphology and the dating of the 
Transvaal hominids. Attention will here be drawn 
to these two aspects. 

The dating of the Transvaal sites 
Opinions on the dating of the Transvaal sites of 

Sterkfontein and Makapansgat have been shown 
in the family trees published by Johanson (1978a, 
1978b) and by Johanson and White (1979). Al­
though aware of the known difficulties in dating 
these sites in the absence of radiometric ages, Jo-

. hanson and White place Sterkfontein at about 
2,3 My and Makapansgat at about 2,4-2,5 My 
B.P. In one of the phylogenetic charts A. africanus 
is actually shown as a direct contemporary of H . 
habilis U ohanson 1978b) . In the light of all the evi­
dence that has accumulated the promoters of "A . 
afarensis" have assigned too young an age to the 
Transvaal gracile australopithecines. 

As a result of numerous faunal analyses (e.g. 
Cooke 1970, 1974, Cooke and Maglio 1972, Mag­
lio 1971 , Collings et al. , 1976, White and Harris 
1977) and of the newest palaeomagnetic dating for 
Makapansgat (Brock 1977, Partridge 1979, Mc­
Fadden et al., 1979), Sterkfontein Member 4 is 
dated on faunal grounds to 2,5 to 3,0 My (fig. 6) 
with " a fairly probable minimum of about 
2,5 m.y." (Cooke 1974). Palaeomagnetic results for 
Sterkfontein are not yet available. Makapansgat 
Member 3, the stratum that has yielded all but 

STERKFONTEIN FORMATION 

FOREIGN STONE 

-s 
HOMO C F HABILI S 

PLUS STONE TOOL 
AND FOREIGN S TON E 

AUSTAALOPITHECUS 

NO STONE TOOLS 
NO FOREIGN STO N 

? 

PMF/ JAN 1979 

AFRICANUS~ 

E 

... y MEMBER BED 

6 
8 

f-:;:-

? 2 0 5 

_£__ 

c 
? 2 · 5 -
-J 0 • 4 -

A 

3 ------G.4P IN SEC TION 

3 

.... UNCONFO RMABLE , ABRUPT, ERODED 

.... UNCONFO RMABLE. ERODED 

... EVEN AND ABRUPT 

.... ERODED AND OI SCONFOAMABLE 

... INTERDIGITATING 

STRATIGRAPHIC 
COLUMN OF THE 

STERKFONTEIN 
FORMATION 

(AFTER T. C PARTRIDGE) 

TOTAL DEPTH 

CA 27 · 5M 
CA 90FT 

Figure 6. Schema of the stratigraphic sequence at Sterkfon­
tein based on Partridge ( 1978). The A . africanus­
rich M ember 4 is dated by the associated fauna to 
3,0-2,5 M y B.P .; palaeomagnetic data are not yet 
available. 

one of the australopithecines from that site, is 
dated faunistically as somewhat older than Sterk­
fontein Member 4 and palaeomagnetically as 
either just over 2,9 My or just over 3,06 My 
(fig. 7), according to which sequence applies (Mc­
Fadden et al., 1979, and this volume). 

The latest palaeomagnetic dating for Maka­
pansgat was announced by Brock as long ago as 
1977 and supersedes the preliminary datings pub­
lished by Brock et al. ( 1977). 

The faunal datings are based independently on 
a variety of faunal elements, including suids, pro­
boscideans, felids and hyaenids, whilst bovids have 
not so far provided definitive indications (Vrba 
1974, 1975). It is noteworthy that White and Har­
ris (1977: 19) have equated the Makapansgat suids 
(of Member 3) with Omo Shungura Members B 
and C though on p. 20 they mention only Member 
C. The approximate time equivalent of Shungura 
B and C spans a period from just under 3,0 My to 
about 2,4 My (Brown and Shuey 1976). Cooke 
( 1978) has contested the suid phylogenetic and 
correlative scheme of White and Harris ( 1977) as 
oversimplified, and he has proposed an alternative. 
None the less, he has made clear that the "taxo­
nomic and phylogenetic differences do not have 
any significant effect on the stratigraphic interpre­
tations offered by White and Harris" (Cooke 1978: 
460). In view of all this faunal and palaeomagnetic 
evidence, some of which White himself has ad-



MAKAPANSGAT FORMATION 

STRATIGRAPHIC 
COLUMN 

(AFTER T. C PARTRIDGE) 

PALAEOMAGNETIC 
PHENOMENA 

(AFTER BROCk , 
MC FAOOEN AND PARTR IDGE) 

AUSTRALOPITHECUS AFAICAMJS .. 

(MLD 37/38) 

.AUSTRALOPITHECUS AFRICANUS ~ 

(ALL BUT ONE) 

MY MEMBER 

5 

4 

2· 80 

3 ·06 

)J -06 3 

2 

3 ·32 

BED 

D 

I 
• 

-

• 

• 

f--
• 

• 
r--

• 

POLARITY 

T 
? 

.... UNCONFORMITY 

? 

KAENA EVENT _} 

MAMMOTH EVENT-? 

? 
..... UNCONFORMITY 

~ -+-
) 3 · 32 1 

;:::;:=, DOLOMITE ~ 

PMF/JAN 1979 

Figure 7. Stratigraphic and palaeomagnetic sequence of 
Makapansgat Formation based on Partridge ( 1979) 
and McFadden et al. ( 1979). The palaeomagnetic 
dating column shown here represents one of two 
possible models that fit the palaeomagnetic data, 
according to McFadden et al. ( 1979). On this 
model the A. africanus-rich Member 3 is just older 
than 3,06 My B.P. ; on the other model not por­
trayed here, Member 3 is just over 2,9 My B.P. 
Both models ,. which are considered to be the only 
ones that fit the palaeomagnetic sequence, yield an 
age for Member 3 of about 3,0 My B.P. 

duced, it is surpnsmg to see so young an age as 
2,3-2 ,5 My being accorded Sterkfontein and 
Makapansgat by Johanson and White (1979). The 
use of the newest estimates and determinations of 
the age of the A. africanus sites in the Transvaal 
would invalidate some aspects of the phylogenetic 
reconstruction offered by the proponents of "A. 
afarensis''. 

The supposed "robust" trend in A. africanus 
transvaalensis and at Hadar 

In 1967 the author drew attention to some 
robust-like elements among the small sample of A. 
africanus fossils from Makapansgat. Subsequently, 
Aguirre ( 1970) in a study of early hominid man­
dibles from South Africa went further and claimed 
that one or two specimens from Makapansgat, 
notably the MLD 2 mandible, should be reclassi­
fied as A. robustus. His view did not gain support. 
Later one of my Ph.D. students, J.A. Wallace, re­
turned to this theme and claimed that in its P3 

13 

MLD 2 showed "a curious amalgam of Sterkfon­
tein-Swartkrans dental characteristics" (1975 : 
216). On several occasions I have speculated on 
the possible explanations for these robust-like ele­
ments in the Makapansgat gracile australopithe­
cines and most recently have inclined to the view 
that they are an expression of polymorphism in A . 
africanus transvaalensis (Tobias 1967, 1969, 1973a). 

From scattered references in the literature about 
MLD 2 and the possibility that the bigger 
Makapansgat crania might have been furnished 
with sagittal crests, Johanson and White ( 1979) 
have generalised to the entire population of A. afri­
canus transvaalensis. They even include in the con­
cept the large sample of Sterkfontein Member 4 
among which, however, robust-like elements have 
not been noted. Speaking of the entire "South Afri­
can gracile australopithecine group", they ( 1979: 
327) opine that "It seems to share several distinc­
tive, derived characters with later robust australo­
pithecines". In making this claim, they have disre­
garded the possibility that the robust-like features 
might simply have represented one extreme in t'he 
range of variation of traits which differed in degree 
but not in kind. The development of the sagittal 
crest, for example, is the consequence of powerful 
temporalis muscles spreading their area of pur­
chase over a small-brained calvaria. A large male 
A. africanus is just as likely to have developed one 
as an average-sized A. robustus. Such continuously­
varying traits do not provide support for the idea 
that A. africanus is already specialized along the 
line towards A. robustus. 

A further difficulty with the phyletic system hy­
pothesized by Johanson and White is that, if it 
were correct that A. africanus were already on a line 
of specialization leading to A. robustus, it could be 
predicted that the earlier population on that line 
should show fewest robust elements and the later 
population the most. The facts are at variance 
with the prediction: the older Makapansgat site 
has those few robust-like elements; the somewhat 
younger Sterkfontein lacks them. The deduction 
from the hypothesis is thus disconfirmed. This 
amounts to a refutation of their hypothesis on the 
place of the Transvaal gracile australopithecines in 
hominid phylogeny. 

Thus, it seems that both the morphology and 
the chronology of the Transvaal A. africanus fossils 
have been misinterpreted by the authors of "A. 
afarensis''. 

It is interesting: to note that not long ago Johan­
son and Taieb (1976) and Johanson et al. (1976) 
were so impressed with certain robust elements in 
the Hadar population as to have suggested the 
possible presence of "an early occurrence of a ro­
bust australopithecine lineage". Among other fea­
tures they drew attention to the heavily-pneuma­
tised temporal (A.L. 166-9) and a right proximal 
femur (A.L. 211-1). Although they subsequently 
abandoned this notion for their East African fossils, 
it does at least emphasize that both at Hadar arid 
its near-contemporary Makapansgat the popu-



14 

lation of gracile hominids included polymorphi­
cally some robust-like features that seem .to have 
been part and parcel of the same populatiOn ~nd 
doubtless expressions of a markedly sexually dtm­
orphic phase of the hominid lineage. 

CONCLUSION AND PROPOSAL OF AN 
ALTERNATIVE HYPOTHESIS 

Laetoli 
The published evidence, including the det~iled 

description by White ( 1977) and the observatiOnal 
contact I have been allowed to enjoy with the 
originals and casts, have led me t<;> conclude that 
the early hominid fossils of Laetoh closely ~ese.m­
ble those of A. africanus. Indeed, no convmcmg 
morphological evidence has been presentec;I that 
the fossils of Laetoli belong to a dtfferent lmeage 
and are any more than subspecifically di~tinct 
from A. africanus transvaalensis. From . the mfor: 
mation at my disposal I believe that the Laetoh 
hominids can be accommodated comfortably 
within the species A. africanus, perhaps "':'it~ some 
minor modifications of the ranges of vanatwn for 
several dental metrical traits. On the basis that the 
Laetoli teeth are very slightly larger and the man­
dibular premolars somewhat longer and narrower 
than those of the Transvaal A. africanus, and since 
the Laetoli fossils are 0,6--1 ,0 My older and geo­
graphically well-removed from A. africanus transvaa­
lensis, there may well be a case to separate the Lae­
toli hominids at the subspecies level. At the same 
1978 Nobel symposium at which Johanson for­
mally announced the proposal to est~blish a new 
species, "A. afarensis" , I suggested m my paper 
that the Laetoli hominids should be regarded as 
the Tanzanian subspecies of A. africanu~ under the 
nomen of A. africanus tanzaniensis (Tobtas 1979d). 
Appropriate as this name would be for the Tanza­
nian geographical subspecies of A. africanus, the 
claimed new species "A. afarensis" has since been 
published. Even though its authors. drew its name 
from the Ethiopian Afar depressiOn, they hav~ 
chosen as its type specimen one of the Laetoh 
nb.andibles (L.H. 4). Thus the name a[arensis re­
mains inseparably linked to the Laetoh holotype. 
This would remain true even were "A. afarensis" to 
be relegated to the status of a subsp.ecies within A. 
africanus. For the proposed subspectes represented 
by the Laetoli fossils the correct name would th.en 
be A. africanus afarensis. The o~her two subspectes 
of A. africanus formally recogmsed are A. afr~canus 
africanus for the Taung type skull and .A. afncanus 
transvaalensis for the gracile australoptthecmes of 
Sterkfontein and Makapansgat (Robinson 1954, 
Campbell 1962)* 

* Campbell ( 1972, 1973) has more recently propounded a 
hominid classification in which the Transvaal robust aus­
tralopithecines, Homo habilis and the Mojokerto hominid are 
regarded as representing other subspecies of A. afric~nus, 
namely A. africanus robustus, A. a. habzlls and A. a. mod;~ker­
tensis, respectively. His proposed scheme has not yet gamed 
acceptance. 

Hadar 
Neither the published evidence, nor the personal 

acquaintance I have had with the originals and 
casts of some of the fossils , allows me to confirm 
that the hominids of Hadar are so different that 
they - or even some of them - should be re~erred 
to any new species of Australopithecus. The ptcture 
afforded by most of these fossils suggest.s an. Ethi­
opian population or sequence of populatiOns m the 
A. africanus lineage. We ne.ed close f~uther ~tudy, 
including statistical analysts of metncal .trmts, of 
the Ethiopian remains (from both the earher mem­
bers of Omo and the various members at Hadar), 
before it can be determined whether the various 
sub-sets of Hadar hominids, severally or collec­
tively, should be referred to (a) the same s~bspe­
cies as their southern African contemporanes, A. 
africanus transvaalensis; (b ) the same subspeci~s as 
the earlier Tanzanian hominids from Laetoh; or 
(c) a separate fourth subspecies of~· africanus, fo.r 
which a name other than A. afncanus afarenszs 
would have to be sought if the latter name re­
mained tied to the Laetoli fossils. 

One morphological trait to which attention has 
not been drawn is incurvation of the M 3s of the 
superb Hadar palate and dent~l arca?e, A.L. 
200-la manifest on the cast and tllustratwns (e.g. 
Johans~n and Whit~ 1979: fi~ . 2): T~is gives a 
maxillary bidental wtdth at M whtch 1s less than 
that at M 2 • This feature has not been found on a 
small series of australopithecine dental arcades of 
A. africanus, A. robustus or A. boisei (Tobias 1980), 
nor does it appear to characterize the more fully 
adult specimen A.L. 199-1. If it is a real adl:'lt 
trait of A.L. 200-1a and not the consequence oftts 
young adult status, 'it may follow that this individ­
ual possessed the structural basis to h~ve devel­
oped a helicoidal occlusal plane (Tobtas 1 ~80). 
The presence of this structural and functiOnal 
complex, along with the probable tendency to 
smaller teeth, and perhaps a few other odonto­
scopic special features, may signal sufficient ~i~er­
ences to justify recognition of the Hadar hommtds, 
and perhaps the A. africanus-like teeth of .Omo 
(Howell 1969, Coppens 1975), as representattve of 
a distinct geographic subspecies of :4·. africanus. ~f 
further close morphological and stattsttcal analysts 
confirms the presence of these small mark~ of dis­
tinction, it may be necessary to recogmze and 
name a separate subspecies within the evolving 
and polytypic lineage of A. ~fricanus. As the f.lame 
afarensis would be pre-occu pted by t~e Laetoh fos­
sils, then it would be most appropnate to suggest 
an alternative regional or territorial soubriquet: I 
propose Australopithecus africanus aethiopicus as a 
suitable name for this Ethiopian taxon. 

Summation 
No adequate morphological, chronological or 

other evidence has been provided that would re­
fute the alternative hypothesis, namely that both 
the Laetoli and Hadar samples and the Omo teeth 
allocated to A. africanus or A . a.ff africanus belong to 



15 

the same lineage as A. africanus transvaalensis and, more-diversified programme of cultural behaviour. 
moreover, constitute earlier and later populations To this derivative lineage has been given the name 
of the long-elusive East African representative of Homo, and the varied morphological grades it tra-
A. africanus. It is here proposed that the Laetoli versed are designated H. habilis, H. erectus and H. 
and Hadar fossils and those of Omo of the sapiens, more or less in succession. 
3, 7-2,5 My time-range provide us, at last, with This alternative hypothesis underlines a view 
East African samples of A. africanus very similar to that I have long held: it is not the South African 
the Transvaal populations of that species, partly fossils alone, nor yet the East African fossils alone, 
contemporary with and partly older than the that may be expected to lay bare the secrets of hu-
Transvaal counterparts. man evolution. Accidents of preservation and of 

On the view propounded here the place of A. af- discovery have decreed that our early African fos-
ricanus in time, space and phylogeny is confirmed sils are sundered into these two geographical 
and strengthened by the valuable new finds of groups with a palaeo-anthropological no-man's-
Laetoli and Hadar. These early East African land separating them. None would doubt, how-
hominids have added greatly to the probability ever, that the distributions of the early hominids 
that A. africanus was a polytypic species, part of an embraced the intervening region: unfortunately, J. 
evolving lineage of ancestral hominids from at Desmond Clark's searches in the vicinity of Lake 
least 3, 7 to about 2,5 My B.P. These hominids Malawi about half-way between the northern Tan-
occupied a late-Pliocene time-slot, earlier than the zanian sites and those of the Transvaal have not 
appearance of both the robust australopithecines yielded the hoped-for early hominids. It is, how-
(A. robustus and A. boisei) and the earliest members ever, most likely that they are there and will still 
of the genus Homo which first emerge in the fossil be discovered. My point is that only by the careful 
record about 2,3 My B.P. In this sense A . africanus study of all South and East African early hominids 
from the Transvaal, Tanzania and Ethiopia is may we expect to find an approximation to the 
most likely to have been the common ancestor of truth, for the birth of the hominids was a pan-
both later lineages of hominids. African phenomenon. The uncovering of those 

One derivative lineage retained micrencephaly, birth-stages requires a pan-African approach, free 
developed its teeth and masticatory apparatus to a of regional or territorial preconceptions and pre-
very marked degree, became on the average dilections. 
heavier though not necessarily taller, evolved per- ACKNOWLEDGEMENTS 
haps a greater degree of sexual dimorphism and I am grateful to Dr. Mary D. Leakey, Professor A. Bethwell 
apparently did not show marked dependence on Ogot, Mr. R.E.F. Leakey, Dr. D.C. Johanson, Dr. Y. Cop-
. 1 b d 1 b cr. h pens and Dr. T.D. White for allowing me access to many of 
Imp ements: A. ro ustus, a mo est Y ro ust ous oot the original specimens as well as casts of the Laetoli and Ha-
known thus far only from the Transvaal; and A . dar hominids. I thank F.E. Grine for helpful comments. 
boisei, a hyper-robust branch so far represented For the excavation of the new Sterkfontein specimens illus-
only in the end-Pliocene and early-Pleistocene fos- trated and cited herein I am indebted to Mr. Alun R. Hughes, 
sil assemblages of Tanzania, Kenya and Ethiopia. whose patience, indefatigability and field experience have 

The other seemingly derivative lineage was been an unlimited source of strength over the 13 years since 
characterized by allometric brain enlargement: the latest Sterkfontein dig began. The financial support of this 
meso-encephaly in its early stages, as in Homo habi- programme has been generously provided by the Council for 
lis; macrencephaly in the subsequent grade, H. Scientific and Industrial Research, the University of the Wit-

watersrand Council Research Committee, and the Bernard 
erectus; and gigantencephaly in H. sapiens. It was Price Institute for Palaeontological Research. 
marked, too, by reduction of the dentition and The illustrations were provided by Mrs. Lynn Rough, A.R. 
masticatory apparatus, a probable diminution in Hughes, P.M. Faugust, c. Richards and J. Bunning. Mrs. 
the degree of sexual dimorphism and increasing Kay Copley and Miss Christel Eckert typed the manuscript 
dependence on implemental activity and an ever- with care and thoughtfulness. 

REFERENCES 

AGUIRRE, E. (1970). ldentificacion de " Paranthropus" en 
Makapansgat. Cronica. del XI Congreso Nacional de Arqueologia, 
Merida I969, 98-124. 

ANONYMOUS CORRESPONDENT (1975). Ethiopian fos­
sil hominids . Nature, Lond., 253, 232-233. 

ARONSON , J.L. , SCHMITT, T.J. , WALTER, R.C. , 
TAIEB, M., TIERCEL! , J.J. , JOHA SO , D .C., 
NAESER, C.W. and NAIRN, A.E.M. (1977) . New geo­
chronologic and paleomagnetic data for the hominid-bear­
ing Hadar Formation of Ethiopia. Nature, Lond., 267, 
323- 327. 

BOAZ, N.T. and HOWELL, F.C. (1977). A gracile hominid 
cranium from Upper Member G of the Shungura Form­
ation, Ethiopia. Am. ]. phys. Anthrop., 46, 93-108. 

BROCK, A. ( 1977). Paper presented to Pan-African Congress 
on Prehistory and Quaternary Studies, Nairobi, Sept. 
1977. 

BROCK, A., McFADDEN, P.L. and PARTRIDGE, T.C. 
( 1977). Preliminary palaeomagnetic results from 
Makapansgat and Swartkrans. Nature, Lond., 266, 
249-250. 

BROOM, R ., ROBINSON, J.T. and SCHEPERS, G.W.H. 
( 1950). Sterkfontein ape-man Plesianthropus. Transv. Mus. 
Mem. , 4 , 1-117. 

BROWN, F.H. and SHUEY, R.T. (1976). Magnetostrati­
graphy of the Shungura and Usno Formations, Lower 
Omo Valley, Ethiopia. In: Coppens Y., Howell F .C ., Isaac 
G .L. and Leakey R.E.F., Eds., Earliest Man and Environ­
ments in the Lake Rudolf Basin. University of Chicago Press, 
64-78. 

CAMPBELL, B.G. (1962). The systematics of man. Nature, 
Lond., 194, 225--232. 

---- ( 1972) . Conceptual progress in physical anthro­
pology: fossil man. Ann. Rev. Anthrop., 1, 27-54. 



16 

---- ( 1973). A new taxonomy of fossil man . Yearbook of 
Phys. Anthrop., 17, 194-201. 

---- ( 1974). Human E volution, an Introduction to Man 's 
Adaptations (2nd ed .) . Chicago , Aldine Publishing -Co., 
1-469. 

CERLING, T.E. , BROWN, F.H ., CERLI G , B.W., CUR­
TIS, G .H . and DRAKE, R.E. (1979). Preliminary correla­
tions between the Koobi Fora and Shungura Formations, 
East Africa . Nature, Lond. , 279, 118-121. 

COLLINGS, G .E., CRUICKSHANK, A.R.I. , MAGUIRE, 
J.M. and RANDALL, R.M. (1976). Recent faunal studies 
at M akapansgat Limeworks , Transvaal, South Africa. Ann. 
S . Afr. Mus. , 71 , 153-165. 

COOKE, H.B.S. (1970). 1 otes from members. Bull. Soc. Verte­
brate Paleont. , 90, 2. 

---- ( 1974) . Plio-Pleistocene deposits and mammalian 
faunas of Eastern and Southern Africa. Proc. V Congres 
N eogene Miditerranien, Ly on 1971. Paris: B.R.G.M ., Memoir 
78, 99-108. 

---- ( 1978) . Suid evolution and correlation of African 
hominid localities: an alternative taxonomy. Science, 201 , 
460-463 . 

COOKE, H.B.S. and MAGLIO, V.J . ( 1972) . Plio-Pleistocene 
stratigraphy in East Africa in relation to proboscidean and 
suid evolution . In : Bishop, W .W . and Miller, J.A., eds ., 
Calibration of Hominoid E volution. Edinburgh, Scottish Aca­
demic Press , 303-329. 

COPPENS, Y. · ( 1975) . Evolution des hominides et de leur en­
vironnement au cours du Plio-Pleistod:ne dans Ia basse 
valee de l'Omo en Ethiopie. C.R. A cad. Sci. Paris, 281 , 
1693-1696. 

---- ( 1977) . Evolution morphologique de Ia premiere 
premolaire inferieure chez certains Primates superieurs . 
C.R. Acad. Sci., 285, 1299-1302. 

COPPE S, Y. and JOHANSO , D.C. (1977). Paper pres­
ented to Pan-African Congress on Prehistory and Quater­
nary Studies , Nairobi , Sept. 1977. 

CURTIS, G .H. , DRAKE, R.F. , CERLING, T .E., CER­
LING, B.L. and HAMPEL, J.M. (1975). Age of KBS tuff 
in Koobi Fora Formation, East Rudolf, Kenya . Nature, 
258, 395-398. 

DART, R.A. ( 1925) . Australopithecus africanus: the man-ape of 
South Africa. Nature, Lond. , 115, 195-199. 

---- ( 1948a). The adolescent mandible of Australopithe­
cus prometheus. Am.}. phys. Anthrop. , n.s . 6 , 391-412. 

---- ( 1948b). The Makapansgat proto-human Australo­
pithecus prometheus. Am.}. phy s. Anthrop., n .s. 6 , 259-284. 

---- ( 1954). The adult female lower jaw from 
Makapansgat. Nature, Lond. , 173, 286. 

---- (1955) . Extinct and extant human mandibles . S. 
Afr.J. Sci., 51 , 258-262. 

---- (1962) . A cleft adult mandible and the other nine 
lower jaw fragments from Makapansgat. Am. }. phys. An­
throp., n.s. 20, 267-286. 

DAY , M .H ., LEAKEY , R .E.F. , WALKER, A.C . and 
WOOD, B.A. (1975) . 1 ew hominids from East Rudolf, 
Kenya , I. Am.}. phys. Anthrop. , 42, 461-476. 

---- ( 1976) . ew hominids from East Turkana , 
Kenya. Am. ]. phys. Anthrop. , 45, 369-436. 

FITCH, F.J. , HOOKER, P.J . and MILLER, J.A. (1976). 
40 Ar/ 39 Ar dating of the KBS tuff in Koobi Fora Formation, 
East Rudolf, Kenya. Nature, Lond ., 263, 740-744. 

GONDA, K . (1959). On the sexual difference in the dimen­
sions of human teeth.}. Anthrop. Soc. Nippon, 67, 151-163. 

HAY, R.L. ( 1976) . Geology of the Olduvai Gorge. University of 
California Press , 1-203. 

HINRICHSEN, D. ( 1978) . How old are our ancestors? New 
Scientist, 78, 5 71. 

HOWELL, F.C. ( 1969). Remains of Hominidae from Plio-

cene/ Pleistocene formations in the lower Omo basin, Ethi­
opia. Nature, Lond., 223, 1234-1239. 

HOWES, A.E. (1954) . A polygon portrayal of coronal and ba­
sal arch dimensions in the horizontal plane. Amer. }. 
Orthod., 40, 811-831. 

HURFORD, A.J ., GLEADOW, A.J.W. and NAESER, C .W . 
( 1976) . Fission-track dating of pumice from the KBS tuff, 
East Rudolf, Kenya . Nature, Lond., 263 , 738-740. 

JOHANSON, D .C. (1978a). A new species of man, Australopi­
thecus afarensis from Eastern Africa stimulates major revi­
sions in understanding human origins . Further Evidence, 1, 
5-8. 

---- ( 1978b). Our roots go deeper. Science Year 1979, 
Worldbook Childcraft International Inc., 43-45. 

JOHANSON, D .C ., COPPENS, Y. and TAIEB, M. ( 1976). 
Pliocene hominid remains from Hadar, Central Afar, 
Ethiopia. In: Tobias, P.V. and Coppens , Y. eds ., Les Plus 
A nciens H ominidis, U.I.S.P.P., Nice. C . . R.S. , 120-1 37. 

JOHA1 SON, D.C. and COPPENS, Y. (1978). Paper pre­
sented to 1 obel Symposium on " Current Argument on 
Fossil Man", Karlskoga (in press) . 

JOHA SON, D.C . and T AIEB, M . ( 1976). Plio-Pleistocene 
hominid discoveries in Hadar, Ethiopia. Nature, Lond. , 
260, 293-297. 

JOHANSON, D.C. , WHITE, T.D. and COPPENS, Y. 
( 1978). A new species of the genus Australopithecus (Pri­
mates : Hominidae) from the Pliocene of eastern Africa . 
Kirtlandia, 28, 1-14. 

JOHANSON, D .C . and WHITE, T .D. ( 1979). A systematic 
assessment of early African hominids. Science, 202, 
321-330. 

LEAKEY, L.S.B., TOBIAS, P.V. and APIER, J .R. (1964). 
A new species of the genus Homo from Olduvai Gorge. Na­
ture, Lond. , 202, 7-9. 

LEAKEY, M .D., HAY, R.L. , CURTIS, G .H ., DRAKE, 
R.E. , JACKES, M.K. and WHITE, T .D. (1976). Fossil 
hominids from the Laetolil Beds. Nature, Lond. , 262, 
460-466. 

LEAKEY, R.E.F. (1973). Evidence for an advanced Plio­
Pleistocene hominid from East Rudolf, Kenya. Nature, 
Lond. , 242, 447-450. 

---- ( 1976). Hominids in Africa . A mer. Scient. , 64, 
174-178. 

LEAKEY, R .E.F. , LEAKEY, M .G . and BEHRE SMEYER, 
A.K. (1978) . The hominid catalogue . In: Leakey, M.G. 
and R.E.F. , eds ., Koobi Fora Research Project, Vol. 1, Oxford , 
Clarendon Press, 86--182. 

M cFADDE , P .L. , BROCK, A. and PARTRIDGE, T.C. 
( 1979) . Palaeomagnetism and the age of the Makapansgat 
hominid site. Earth Planet. Sci. Lett. , 44, 373-382. 

MAGLIO, V.J. (1973) . Origin and evolution of the Elephan­
tidae. Trans. Amer. Phil. Soc., n.s. 63, 1-149. 

MA YR, E. ( 1970) . Populations, Species and Evolution. Cam­
bridge, Mass ., Harvard University Press, 1-453. 

MIJSBERG, W.A. (1931 ) . On sexual differences in the teeth 
of the Javanese. Konink . Akad. Wetensch. , Amsterdam, 34, 
1111-1115. 

OAKLEY, K .P ., CAMPBELL, B.G . and MOLLESON, T.I. , 
eds . ( 1977). Catalogue of Fossil Hominids. Part I: Africa (2nd 
ed. ) . London, British Museum (Nat. Hist. ), 1-223. 

OXNARD, C.E. (1975) . The place of the australopithecines 
in human evolution , grounds for doubt? Nature, Lond., 
258, 389-395. 

PALAEOANTHROPOLOGY CORRESPOND£ T (1979). 
Difficulties in the definition of new hominid species . Nature, 
Lond. , 278, 400-401. 

PARTRIDGE, T.C . (1978) . Re-appraisal of lithostratigraphy 
ofSterkfontein hominid site. Nature, Lond. , 275, 282-287. 

---- ( 1979 ). Re-appraisal of lithostratigraph y of 



Makapansgat Limeworks hominid site. Nature, Lond., 279, 
484-488. 

ROBINSON, J .T . ( 1954). The genera and species of the Aus­
tra1opithecinae. Am.]. phys. Anthrop., 12, 181-200. 

---- ( 1956). The dentition of the Australopithecinae. 
Mem. Tramv. Mus. , 9, 1-179. 

---- ( 1966). Comment on "The distinctiveness of Homo 
habilis". Nature, Lond., 209, 957-950. 

TOBIAS, P .V . ( 1967). Olduvai Gorge, Vol. 2, The Cranium 
and Maxillary Dentition of Australopithecus (Zinjanthropus) 
boisei. Cambridge University Press, 1-264. 

---- (1969). Commentary on new discoveries and in­
terpretations of early African fossil hominids. In: Genoves, 
S., ed. , Yearbook of Phys. Anthrop. 1967, 15, 24-30. 

---- ( 1973a). New developments in hominid paleon­
tology in South and East Africa. Ann. Rev. Anthrop., 2, 
311-334. 

---- ( 1973b). A new chapter in the history of the Sterk­
fontein early hominid site. J. S. Afr. bioi. Soc., 14, 30-44. 

---- (1975). New African evidence on the dating and 
phylogeny of the Plio-Pleistocene Hominidae. In: Suggate, 
R.P. and Cresswell, M.M. Eds., Quaternary Studies, Bull. 
No. 13, Royal Society of New Zealand, Wellington, 
289-296. 

---- ( 1976). Incomplete or delayed closure of the sym­
physis menti in early hominids and a possible mechanism. 
Yugoslav Anthrop. Soc. Special Publicatiom, Vol. 3 , 23-32. 

---- ( 1978a). The place of Australopithecus africanus in 
hominid evolution. In: Chivers, D.J . and Joysey, K .A., 
Eds. , R ecent Advances in Primatology, Vol. III, Evolution. Lon­
don, Academic Press, 373-394. 

---- (1978b). Position et role des australopithecines 
dans Ia phylogenese humaine, avec etude particuliere de 
Homo habilis et des theories controversees avancees a pro­
pos des premiers hominides fossiles de Hadar et de Laeto­
lil. In : Les Origines Humaines et les Epoques de / 'Intelligence. 
Paris , Masson, 38-77. 

---- ( l979a). Olduvai Gorge, volume 4, H omo habilis 
(in preparation). 

- --- (1979b). Recent palaeontological studies in south-

17 

ern Africa: new discoveries at Sterkfontein, 1968-1976. 
Proc. IXth Congr. U. I .S.P.P. , Nice (in press). 

---- ( l979c). Australopithecus and early Homo. Proc. VIJJ 
Pan-Afr. Congr. Prehist. Quatern. Stud., Nairobi (in press). 

---- ( l979d). A survey and synthesis of the African 
hominids of the late Tertiary and early Quaternary 
periods. Proc. 41st Nobel Symposium, Karlskoga (in press). 

---- ( 1980). The natural history of the helicoidal occlu­
sal plane and its evolution in early Homo. Am. J. phys. An­
throp. , 53, 173-187. 

TOBIAS, P .V ., COPLEY, K . and BRAIN, C.K. (1977). 
South African entries in Catalogue of Fossil H ominids, Part I: 
Africa (2nd edn.), Oakley, K .P ., Campbell, B.G. and Mol­
leson, T.I., Eds., London, Brit. Mus. (Nat. Hist.); 95-151. 

VRBA, E.S. ( 1974). Chronological and ecological implications 
of the fossil Bovidae at the Sterkfontein australopithecine 
site. Nature, Lond. , 250, 19-23. 

---- ( 197 5). Some evidence of chronology and palaeo­
ecology of Sterkfontein, Swartkrans and Kromdraai from 
the fossil Bovidae. Nature, Lond., 254, 301-304. 

WALKER, A. and LEAKEY, R .E.F. ( 1978) . The hominids of 
East Turkana. Sci. Amer. 239, 54-66. 

WALLACE, J.A. (1975). Dietary adaptations of Australopithe­
cus and early Homo. In: Tuttle, R.H., Ed., Paleoanthropology, 
Morphology and Paleoecology. The Hague, Mouton, 203-223. 

WHITE, T.D. ( 1977). New fossil hominids from Laetolil, 
Tanzania. Am. J. phys. Anthrop., 46, 197-230. 

WHITE, T.D. and HARRIS, J.M. ( 1977). Suid evolution and 
correlation of African hominid localities. Science, 198, 
13-21. 

WOLPOFF, M.H. ( 1971 ). Interstitial wear. Am. J. phys. An­
throp., 34, 205-228. 

---- (1975). Sexual dimorphism in the australopithe­
cines . In: Tuttle, R.H ., ed., Paleoanthropology, Morphology 
and Paleoecology. The Hague, Mouton, 245-284. 

---- ( 1976). Some aspects of the evolution of early 
hominid sexual dimorphism. Curr. Anthrop. , 17, 579-606. 

---- ( 1979). Anterior dental cutting in the Laetolil 
hominids and the evolution of the bicuspid P3. Am. J. phys. 
Anthrop., 51 , 233-234.