[SLIDE 1 — TITLE SLIDE]

[SLIDE 2 — HADAR REGION] In the early 1970s, very early australopithecines, dated to around 3.5 million years, began to emerge from the Hadar area in the Afar Triangle of Ethiopia. The first descriptions of this material by Don Johanson and Maurice Taieb identified THREE different species in the Hadar remains—

[SLIDE 3 — THREE SPECIES] — a small gracile Australopithecus represented by the "Lucy" skeleton, a larger Paranthropus form represented by the A.L. 166-9 temporal, and an early Homo with front teeth too large for an australopithecine. Tim White studied these finds and concluded that the Paranthropus temporal could be interpreted as a big male "Lucy," and that the big front teeth of A.L. 200 could be read as primitive retentions from a more apelike ancestry. And so—

[SLIDE 4 — 1978 paper] — in 1978, White, Johanson, and Coppens lumped all the Hadar and Laetoli hominins together into a single new species, which they christened Australopithecus afarensis.

[SLIDE 5 — 1979 PHYLOGRAM] Johanson and White portrayed this species as earlier and more primitive than the previously described australopithecines. THOSE creatures, they said, lay on a sterile side lineage leading to Paranthropus; but A. afarensis was the sole australopithecine ancestor of the genus Homo. There were questions and dissention. Stern and Susman accepted afarensis as a single species, but concluded that the big males and the small females had different locomotor adaptations -- a dimorphism unique among primates, if not among mammals. Others went further and proposed that the big and small morphs were different species.

[SLIDE 6— OTHER SPP.] And some claim to have discovered other contemporary hominins living alongside, or sympatric with, A. afarensis. But the afarensis concept was powerfully defended, and a consensus emerged that A. afarensis was a single species — with a lot of geographical and temporal variation and sexual dimorphism — and that it encompassed all the East African hominins dated to between 3.9 and 2.9 million years.

[SLIDE 7 — BURTELE FOOT 1] In 2009, this partial foot skeleton was found at the Burtele-2 locality in Ethiopia. Dated to around 3.4 million years, the Burtele foot falls right in the middle of the temporal and geographic range for "A. afarensis." But it belies all the descriptions of that species as a dedicated terrestrial biped.

[SLIDE 8— BURTELE FOOT 2] Its lateral metatarsals have domed heads, showing that the fulcrum of the foot at toe-off was at the metatarsal heads as in humans, not further back as in apes. But the lateral metatarsal bases are dorsoventrally shallow, suggesting that the foot was not yet a really stiff propulsive lever like our own.

[SLIDE 9— BURTELE FOOT 3] The first digit is robust, and ITS metatarsal base IS tall, as in bipeds. But it is short, like those of nonhuman catarrhines; and its metatarsal head isn't domed, so it wasn't involved in producing propulsive thrust at toe-off. Most importantly—

[SLIDE 10— BURTELE FOOT 4] —The configuration of the joint surface at its base shows that the hallux was mobile and opposable—

[SLIDE 11— BURTELE FOOT 5] —and the second digit was rotated medially to oppose it in grasping. In short, the foot of the Burtele hominin was a handlike grasping organ, like that of an ape — or of its possible precursor Ardipithecus.

[SLIDE 12— SIX INTERPRETATIONS] There are six possible interpretations of these facts:

[SLIDE 13] (1) The Burtele foot represents a previously undiscovered species, and all the other hominin fossils from this time and area are sampling a different species, A. afarensis. This is possible, but we doubt it — because of the odds against it, and because of the exceptional variation seen in some other parts of the skeleton in the "afarensis" collection.

[SLIDE 14] (2) A second option: the Burtele foot belongs to A. afarensis, and shows that this species had a divergent, grasping hallux.

[SLIDE 15— LATIMER & LOVEJOY QUOTES] We reject this and accept the conclusion that some "A. afarensis" had a permanently adducted hallux. The defenders of the "afarensis" concept seem pretty definite on that point.

[SLIDE 16] (3) The Burtele foot belongs to A. afarensis, and shows that some individuals in this species had a hallux like an ape's, while others had a longer, permanently adducted big toe like a human's. Is this possible?

[SLIDE 17—ARDI QUOTES] Well, the As Duma "Ardipithecus ramidus" had a propulsive hallux, and that from Aramis did not. [PAUSE TO ALLOW READING]

[SLIDE 18— StW QUOTES] And hominin foot bones from Member 4 at Sterkfontein have been reported to show similar functional differences. Perhaps Ardipithecus ramidus, A. afarensis, and A. africanus all represent populations with huge variability in locomotor adaptations, representing an unstable transitional stage in the evolution of a stiff, propulsive foot. But we doubt it. No living species we know of shows this sort of variation, and it is hard to believe that such variability would have persisted for over a million years. We regard this interpretation — in all THREE of these cases — as possible but unlikely.

[SLIDE 19] (4) The differences between the Burtele foot and the later A.L. 333 foot bones represent anagenetic evolution within A. afarensis. Again, this is possible, but we doubt it, because the two sites are separated by only some 200,000 years.

[SLIDE 20] (5) The Burtele foot belongs to A. afarensis, and demonstrates that there was marked locomotor dimorphism in that species, with the females being more arboreal than the males — as suggested by Stern and Susman. Again, the difficulty here is that no living analogs are known.

[SLIDE 21] (6) The most probable interpretation is that these two sorts of foot morphologies represent different species. This was also the conclusion of the describers of the Burtele foot. This has several implications.

[SLIDE 22 — READ THE TEXT]

1. There were at least two hominin species in the Afar region between 3 and 4 million years ago.

2. There are (probably) at least two hominin species represented in the "afarensis" hypodigm.

3. Some of the fossils conventionally assigned to "afarensis" may belong to the same species as the Burtele foot bones.

4. But some other foot bones do not. 5. For the rest of the skeleton, we don’t know which.

6. [SLIDE 23 — LH 4, W/ CAPTIONS] Laetoli Hominid 4 is the type specimen of afarensis, and so it certainly belongs to that species. But does it belong with the contemporary Burtele foot, or with the later, more humanlike foot bones from A.L. 333? We can no longer feel certain about this. No doubt some of the other material currently assigned to A. afarensis belongs to the same species as LH 4 here — but probably not all of it. This may explain some or all of the anomalies that have bedeviled the study of A. afarensis for decades. Let's list a few.

[SLIDE 24— KADANUUMU AND LUCY] A. afarensis is unique among primates, and perhaps among mammals, in combining SLIGHT canine dimorphism with high BODY-SIZE dimorphism — some say as much as in gorillas or orangutans. These facts hamper our efforts to make inferences about its social organization. For example, it has been suggested that the short-legged females and the long-legged males must have been organized in separate bands, with different traveling speeds. Perhaps so. But perhaps they belong to different species.

[SLIDE 25 — PALATES] Molar dimensions and the sizes and shapes of palates and dental arcades vary widely in afarensis. Some have flat, apelike palates, whereas others have arched palates like those of Homo. Some suggest that this reflects sexual dimorphism. But perhaps they belong to different species.

[ SLIDE 26— THORAXES] Australopithecus sediba from South Africa had a funnel-shaped upper thorax, like those of apes. Schmid's reconstruction gives Lucy's thorax a similar shape. But the describers of Kadanuumuu reconstruct ITS thorax as more humanlike in shape — and conclude that this must also have been true of Lucy. But perhaps they belong to different species.

[SLIDE 27 — LOVEJOY PELVIS RECONSTRUCTIONS] Here are two strikingly different pelves, both attributed to A. afarensis and reconstructed by Owen Lovejoy, who attributes their differences to sexual dimorphism in size and shape. But in many respects, the larger pelvis on the left seems more humanlike than the smaller one. Perhaps they belong to different species.

[SLIDE 28 — CALCANEI AND GRAPH] The tuber calcanei is fat and broad in humans, and skinny in apes. Both have fat tubers when young, and they remain fat in adult humans. But the Dikika infant's is skinny and apelike, whereas that of this A.L. 333 adult is fat and humanlike. Perhaps, as DeSilva and others suggest, afarensis had a unique ontogenetic pattern, unlike those of either humans or apes. But perhaps the two feet belong to different species.

[SLIDE 29— SCAPULAE] The Dikika scapula looks a lot like a baby gorilla's, whereas the Kadanuumuu scapula looks more like an adult human's than like an adult gorilla's.

[SLIDE 30—CHURCHILL ET AL., #1] Some multivariate analyses represent Dikika's scapula as apelike and arboreal, whereas similar comparative analyses of the Kadanuumuu scapula —

[SLIDE 31— CHURCHILL ET AL. #2] — place it with modern humans. Steve Churchill and others suggest that this may reflect defects in the analyses or samples, or an ontogenetic pattern in A. afarensis.

[SLIDE 32— THIRD POSSIBILITY] But there is a third possibility: perhaps they are different species.

[SLIDE 33— LOVEJOY & HAILE-SELASSIE QUOTES] Perhaps Kadanuumuu has a scapula that looks more like Homo — and a pelvis that looks more like Homo, and a ribcage that looks more like Homo, and limb proportions more like those of Homo — than other australopithecines have, because it is more closely related to Homo. Why not? We can't tell from the skull and teeth; it has none. If it is to be assigned to Australopithecus, it must be on the basis of postcranial anatomy.

[SLIDE 34— AUSTRALOPITHECINE TRAITS] Its main postcranial similarities to australopithecines are plesiomorphies, irrelevant to phyletic affinities. And early Homo was not so different from Australopithecus in many of these features as we once thought. And if the Ledi-Geraru mandible is Homo at 2.8 million years, then the possibility that Kadanuumuu is on that lineage isn’t as far-fetched as it might seem. Where does that leave things? We have no pet or pat answers to suggest: only a string of prescriptions. We should abandon the default assumption that every East African hominin from this time span belongs to the same species as Laetoli Hominid 4 unless proved otherwise. We should no longer reconstruct these early hominins by piecing together fossils from different sites, or assume that what we know about (say) Kadanuumuu tells us anything much about the missing parts of Lucy. It seems likely that what we have been calling "A. afarensis" incorporates more than one species, that the boundaries between them are unclear, and that they are going to stay that way for at least the immediate future. They may stay that way indefinitely. This may seem like a counsel of despair. But if we free ourselves from the cladistic compulsion to start by constructing an optimal cladogram, we can actually say quite a lot about the evolution of early hominins without knowing just how many species there are, where the boundaries between them lie, or exactly how they are related.

[SLIDE 35 — PHYLOGRAM] Here is an evolutionary tree diagram of the old-fashioned kind, with a time dimension and a lot of dotted lines and question marks, that sums up some established facts.

[SLIDE 36] First fact: no matter how the "robust" australopithecine species are related, they stand out as separate and distinctive morphs. Their specializations make them unlikely candidates for human ancestry.

[SLIDE 37] No matter whether all the "A. anamensis" fossils from East Africa belong to a single species, they are the oldest Australopithecus known. They can be taken as representing the ancestral population(s) from which all later hominins arose.

[SLIDE 38] Wider distribution and more diverse morphologies appear at about 3.5 Mya, indicating a radiation of "gracile" types into new ecologies.

[SLIDE 39] Fragmentary jaws and teeth regarded as possible representatives of very early Homo begin to show up in East Africa from 2.8 Mya onward. Their relatively small molars and premolars suggest that they branched off from an unspecialized mid-Pliocene population of Australopithecus. Taken together, these conclusions show that it is possible to say a good deal about the evolution of a group of species, even when many species boundaries and relationships cannot be precisely made out, by putting together information about time, place, and adaptation: things that are systematically and deliberately left out of consideration in constructing cladograms. We suggest that a healthy approach to this fossil record should neither insist on overlooking things like these that we actually know, nor insist that we must always begin by pretending that we know more than we are ever going to.