Leakey, Tobias and Napier on the definition of our genus

13 minute read

Leakey, Tobias and Napier (1964) defined the species, Homo habilis. A simple species diagnosis was not enough: Leakey and colleagues had to argue for an expanded definition of our genus Homo that could accommodate the set of very primitive skeletal remains from Bed I and Bed II of Olduvai Gorge.

It has long been recognized that as more and more discoveries were made, it would become necessary to revise the diagnosis of the genus Homo. In particular, it has become clear that it is impossible to rely on only one or two characters, such as the cranial capacity or an erect posture, as the necessary criteria for membership of the genus. Instead, the total picture presented by the material available for investigation must be taken into account.

This paper was the first and last formal attempt to expand our genus based on morphological considerations. That makes it a very important document in considering how we currently apply the genus concept to human evolution.

Consider the famous simplification of hominin taxonomy proposed by Ernst Mayr. Mayr argued that Homo should be expanded in usage to encompass the entire hominin fossil record – from “Homo africanus” to Homo erectus to Homo sapiens. That taxonomic simplification in the heyday of the Modern Synthesis was in fact what gave us Homo erectus in the place of Pithecanthropus and Sinanthropus. That is, Mayr really did expand the genus Homo beyond its earlier and traditional usage. He did not do so by mounting a morphological argument but instead a phyletic one – Homo in his conception was simply a name that encompasses the human lineage. Later, Morris Goodman would go further based on analogy with the level of genetic divergence found in genera of mammals more generally, urging that chimpanzees and bonobos be admitted to Homo as well. These arguments both are fundamentally cladistic, but not specifically diagnostic. In both cases, the solution for Homo is simply to eliminate Australopithecus and every other genus of hominins.

Leakey and colleagues faced the more difficult task of outlining a principled way to separate Homo from Australopithecus based on morphology.

The Olduvai sample gave them two big challenges. First, the brain sizes represented by OH 7, OH 13 and OH 16 are substantially smaller than those known at the time for Homo erectus. Further, the endocranial volume of the robust OH 5 craninum approaches these Homo-like specimens. Brain size no longer seemed a suitable criterion, at least not by itself. With any closely related taxa, it is inevitable that most characteristics will overlap in range, probably extensively. If we attempt to separate early Homo from other hominins based on brain size only, we will have to exclude small specimens (like KNM-ER 1813) or include large australopithecines (like OH 5). Although it is conceivable for a character to have a truly disjoint distribution (like tiger stripes versus lion manes), we should not expect to find such a character in a limited sample of fossil evidence.

Second, the sample included skeletal parts – notably the hand and foot – that were not previously well represented in fossil Homo or Australopithecus. Leakey, Tobias and Napier had the luxury of looking beyond the skull and teeth, for the first time incorporating them into the morphological diagnosis of our genus. But they had no way to ascertain whether these features were really diagnostic beyond their Olduvai sample.

In other words, they were forced to abandon the one traditional feature of Homo erectus relative to Australopithecus, the brain. They had the opportunity to add features that had not previously been important in the definition of Homo species, in particular the hand and foot. But those features were not at the time well evidenced in early hominins. So they proposed a combination of features as the essential qualities of our genus.

Reading through their paper today, it is striking how the fossil record has changed. I thought it would be useful to annotate their definition from 1964 with notes about how today’s fossil record has (mostly) made it obsolete.

In some cases I’ve drawn from descriptions of later specimens, especially those carried out by Wood (1991). Wood would later (Wood and Collard 1999) mount an argument that none of the fossils attributed to Homo habilis or Homo rudolfensis should really be placed in Homo, but should instead be assigned to Australopithecus. I’ll return to that argument later, but for the moment am just using the best descriptions of these Koobi Fora specimens to focus on how their anatomy contributes to an understanding of Leakey, Tobias and Napier’s definition of Homo.

Revised diagnosis of the genus Homo. A genus of the Hominidae with the following characters: the structure of the pelvic girdle and of the hind-limb skeleton is adapted to habitual erect posture and bipedal gait;

We now know that this criterion does not separate Homo from earlier hominins.

the fore-limb is shorter than the hind-limb;


the pollex is well developed and fully opposable and the hand is capable not only of a power grip but of, at the least, a simple and usually well developed precision grip;

The Hadar hand remains are capable of a simple precision grip, and the Malapa MH 2 hand remains are certainly capable of a well developed precision grip. Indeed, MH 2 has a more human-like wrist morphology in some respects than does the OH 7 hand. So as we know it today, the hand of Australopithecus would qualify as Homo under this definition.

the cranial capacity is very variable but is, in the average, larger than the range of the capacities of members of the genus Australopithecus, although the lower part of the range of capacities in the genus Homo overlaps with the upper part of the range in Australopithecus; the capacity is (on the average) large relative to body-size and ranges from about 600 c.c. in earlier forms to more than 1,600 c.c.;

Cranial capacity is still the definition of Homo in practical terms. Most of the debate about whether early specimens can be attributed to Homo has revolved around what we know (or don’t know) about the size of their brains.

But things have changed to some extent since 1964. The Dmanisi sample of Homo erectus includes three specimens with cranial capacities that are 600 ml or less, including the robust D4500 skull with a volume of around 550 ml. Since 1964, a broader sample of hominin fossils from East Africa have been referred to Homo habilis by one or another anthropologists. These include KNM-ER 1813, with a volume of only slightly more than 500 ml.

the muscular ridges on the cranium range from very strongly marked to virtually imperceptible, but the temporal crests or lines never reach the midline;

This is basically saying that some skulls of Homo are very robust in muscle markings and attachment and others are not, with the exception that there should never be a sagittal crest. We can set aside as exceptions a very small fraction of modern human skulls in which the temporal lines do meet at the midline. The main effect of this restriction in the definition is to exclude KNM-ER 1805, a skull that has some Homo-like features and that Bernard Wood (1991) assigned to the genus Homo, but that has a clear sagittal crest.

the frontal region of the cranium is without undue post-orbital construction (such as is common in members of the genus Australopithecus);

Together with the insistence on temporal lines that do not meet, this emphasizes a relation in which the brain is relatively large and the jaw muscles relatively small in proportion to each other. That is a tricky one only at the boundary of small brain size; however in reality the postorbital constriction present in specimens like MH 1 or even Sts 5 is not too extreme to fit into Homo, at least as represented by KNM-ER 1813 – or even the much larger but similarly shaped OH 9 cranium.

the supra-orbital region of the frontal bone is very variable, ranging from a massive and very salient supra-orbital torus to a complete lack of any supra-orbital projection and a smooth brow region;

That’s our second “very variable” qualifier. This is a problem character. Obviously a supraorbital torus is rare in living humans, but has been nearly universal in many ancient populations of Homo, including some that are uncontroversially recognized as Homo sapiens. So the trait has to be variable.

But at the earliest end of the fossil sample attributed to Homo, we also have that variation. KNM-ER 1470 has no supraorbital torus. KNM-ER 1813 does have a clearly marked one with a supratoral sulcus above it, one of Wood’s (1991) strongest arguments placing it into Homo instead of Australopithecus. The supraorbital torus of MH 1 contributes to its Homo-like appearance and is a derived similarity with Homo in the comparisons carried out by Berger and colleagues (2010).

the facial skeleton varies from moderately prognathous to orthognathous, but it is not concave (or dished) as is common in members of the Australopithecinae;

Together with the supraorbital region and the occipital region, the face bears a great deal of weight in assignment of fossils to Homo. Leakey, Tobias and Napier adopt a conservative definition here. Homo does not have a dish-shaped or concave face. That excludes most of the australopithecines that were then known. But none of the Olduvai sample that Leakey, Tobias and Napier considered in this 1964 paper had enough facial skeleton to assess these aspects of their shape. OH 13 has maxillary remains but none of the upper face or nasal region.

Ironically, a specimen found later at Olduvai and assigned to Homo habilis after its discovery, OH 24, has a face with a concave profile – although in reconstruction, the face is not concave transversely (side-to-side), but only sagittally, with a flat or slightly concave nasal region and moderate subnasal prognathism.

the anterior symphyseal contour varies from a marked retreat to a forward slope, while the bony chin may be entirely lacking, or may vary from a slight to a very strongly developed mental trigone;

This isn’t very helpful; not clear that Sts 52 or even Early Pliocene mandibles like KNM-KP 29281 would be excluded from Homo by this definition. It is only expressing that Homo is variable. This definition is no help at all for contentious cases like KNM-ER 1802, or Omo 75-14, or the Uraha mandible.

It might be more helpful to consider the exact angle of the symphysis or its curvature, or to add that Homo lacks some degree of posterior buttressing of the symphysis present in many mandibles of Australopithecus. But the practical problem is differentiating early Homo from robust mandibles on the one hand – which are more vertical in symphyseal profile and have a thick, tall corpus – from earlier hominins on the other hand, which have a sloping symphyseal profile and often match OH 7 in corpus dimensions.

Leakey, Tobias and Napier concentrated most closely on the dentition, which would become crucial to understanding OH 7. The dentition is the clearest anatomical region in which these authors expanded the definition of Homo in order to accommodate their concept of Homo habilis.

the dental arcade is evenly rounded with no diastema in most members of the genus; the first lower premolar is clearly bicuspid with a variably developed lingual cusp;

This corresponds to the full loss of a canine-P3 honing complex and the shift to a molar-like function for the P3. But already in 1964, that was already known to be true of mandibles assigned to A. africanus such as Sts 52 or MLD 2. And of course it would include A. boisei and A. robustus. From today’s point of view, this form excludes some of the Hadar mandibles, which have strongly unicuspid and rotated P3s.

the molar teeth are variable in size, but in general are small relative to the size of the teeth in the genus Australopithecus; the size of the last upper molar is highly variable, but it is generally smaller than the second upper molar and commonly also smaller than the first upper molar; the lower third molar is sometimes appreciably larger than the second;

Much had been made of the taxonomic relevance of the size proportions of the molars. Generally in Homo erectus, the third molars were smaller than the second molars. The Olduvai adult dentitions, OH 13 and OH 16, had larger third molars on the bottom – and in OH 13 these were markedly elongated. The type specimen of Homo habilis, OH 7, was a juvenile and did not have third molars. The upper dentition of OH 13 has smaller third molars than second molars and even first molars – as stated here. But the OH 16 upper dentition has larger third molars; they are the broadest of the upper teeth and may be the largest (the first molars are more elongated but narrower).

Australopithecine dentitions almost universally have their third molars as the largest upper and lower teeth, and second molars as the next largest. In other words, the molars get larger further back in the dentition. As defined by Leakey, Tobias and Napier, Homo disrupts this relation.

in relation to the position seen in the Hominoidea as a whole, the canines are small, with little or no overlapping after the initial stages of wear, but when compared with those of members of the genus Australopithecus, the incisors and canines are not very small relative to the molars and premolars; the teeth in general, and particularly the molars and premolars, are not enlarged bucco-lingually as they are in the genus Australopithecus; the first deciduous lower molar shows a variable degree of molarization.

Of these, the most important is the buccolingual expansion of australopithecine molars and premolars relative to those in Homo. The Olduvai sample of Homo habilis has elongated molars mediodistally, enhancing this contrast with australopithecines, despite the overall similarity in size between the Olduvai teeth and A. africanus. In their definition of the species Homo habilis, Leakey, Tobias and Napier further note that the dimensions of the molars in Homo habilis fall within the ranges of both Australopithecus and Homo erectus.

What to make of this definition?

Later workers have largely been more restrictive rather than more expansive. Wood (1991) applied a more restrictive set of features as important to assigning Koobi Fora cranial specimens to Homo, emphasizing the basicranium, frontal and occiput. No one has made use of a more expansive definition based on postcranial evidence. Wood and Collard (1999) and Wolpoff (1999) both concluded that a principled division between Australopithecus and Homo could be defended on adaptive grounds only if “Homo habilis” and “Homo rudolfensis” were included in Australopithecus instead of Homo. That distinction in both cases did use postcranial features in addition to cranial features, which tended to restrict Homo to a smaller set rather than expand it to a broader one.

The definition proposed by Leakey, Tobias and Napier in 1964 is obviously unworkable today. Arguably, some of the ambiguity of the definition of Homo can be resolved by studying their definition of the species Homo habilis that directly follows. After all, they clearly believed it defensible to place the Olduvai fossils in Homo despite the problems that they note at the outset.

But really the problem is a phylogenetic one. There is no feature listed in their diagnosis of the genus Homo that is a shared derived feature of all the samples attributed to Homo, that is not also found in some samples attributed to Australopithecus or other genera. The example of MH 1 is instructive – setting aside the cranial capacity of the specimen, its other features nearly all align with those ascribed to Homo by Leakey, Tobias and Napier. It is very difficult to include a specimen like KNM-ER 1813 without bringing in many other specimens of A. africanus by the same argument.


Leakey, L. S., Tobias, P. V., & Napier, J. R. (1964). A new species of the genus Homo from Olduvai Gorge. Nature, 202(4927), 7-9.

Wood, B. A. (1991). Hominid cranial remains. Koobi Fora research project, Vol. 4. Clarendon, Oxford.

Wood, B., & Collard, M. (1999). The human genus. Science, 284(5411), 65-71.

Wolpoff, M. H. (1999). Paleoanthropology. New York: McGraw-Hill.

Berger, L. R., de Ruiter, D. J., Churchill, S. E., Schmid, P., Carlson, K. J., Dirks, P. H., & Kibii, J. M. (2010). Australopithecus sediba: A new species of Homo-like australopith from South Africa. Science, 328(5975), 195-204.