A revised chronology for early Homo

In case you haven't been paying attention, the chronology of early African Homo has been completely turned upside-down this year. Well, "upside-down" isn't precisely right; "displaced younger by a quarter-million years" is better.

The redating has come from Frank Brown's group, which in a series of papers has defined and dated stratigraphic units between the major tuffs of the Koobi Fora formation, between the KBS Tuff at 1.87 Ma and the Chari tuff at around 1.38 Ma. Gathogo and Brown (2006) outline the consequences of this redating for fossils of early Homo. Their paper focuses on the fossils from area 123 at Koobi Fora, but discusses the likely consequences of redating on other localities.

Fossils of Homo now estimated to be 1.65 +/- 0.15 myr in age in the Koobi Fora region are currently assigned to at least two taxa on the basis of both crania and mandibles. Homo habilis is represented by specimens KNM-ER 1501, 1502, 1805, and 1813, and H. ergaster is represented by specimens KNM-ER 730, 1812, and 3733 (for attributions, see Wood, 1991, 1992; Wood and Richmond, 2000). The ages of specimens KNM-ER 1501, 1502, 1812, and 1813 have been discussed above, and although not the main focus of this paper, a few notes are offered below on the others.
Specimen KNM-ER 730 derives from a level 5 m below the Koobi Fora Tuff Complex in Area 103 (Feibel et al., 1989), and is thus ca. 1.6 myr old. Feibel et al. (1989) gave an age of 1.85 myr for KNM-ER 1805, but this specimen lies "just below the base of the Okote Tuff" in Area 130 (Leakey et al., 1978), and is more likely closer in age to that of the base of the Okote Tuff Complex (ca. 1.6 myr) than it is to that of the KBS Tuff (1.87 myr). On the basis of mollusc-packed sandstones and algal horizons correlated from Area 102 to Area 104, Feibel et al. (1989) estimated that KNM-ER 3733 was 1.78 myr in age. Although the age of KNM-ER 3733 cannot be confirmed without additional fieldwork, the White Tuff, with an estimated age of 1.63 myr (Brown et al., 2006), is the nearest unequivocally identified unit in the local section in Area 104. This tuff is exposed <300 m from the location of KNM-ER 3733, and Tindall (1985) records only 8 m of section below the White Tuff nearby. Therefore KNM-ER 3733 should be approximately the same age as KNM-ER 1813. Indeed, all specimens from Koobi Fora assigned to H. aff. H. erectus by Wood (1991), many of which are now referred to H. ergaster (Wood and Richmond, 2000), are now estimated to be 1.45 to 1.65 myr old with the exception of KNM-ER 2598. The latter specimen, which is a partial occipital bone from Area 15, was placed 4 m below the KBS Tuff by Feibel et al. (1989) and estimated to be about 1.9 myr old. This age estimate is reasonable because strata do not extend more than 7 m above or below the KBS Tuff at the recorded location of KNM-ER 2598 (Gathogo and Brown 2006:7-8, emphasis added).

This raises a question: Just how much evidence is left for large-bodied H. erectus-like hominids earlier than 1.65 Ma?

Wood (1991) didn't diagnose postcrania, and Gathogo and Brown (2006) don't comment on them. At least KNM-ER 1808 would seem to fall under this umbrella, since Wood (1991) did diagnose that. But more important in bracketing the evolution of large body size is KNM-ER 3228, a hip bone previously dated to 1.95 Ma. It's pretty big for a human, let alone an australopithecine. On the other hand, McHenry and Coffing (2000) suggested that KNM-ER 3228 might belong to H. rudolfensis. To my eyes, this would make it a pretty big specimen compared with femora like KNM-ER 1472 and KNM-ER 1481, but who knows?

Another uncomforable fit in an H. rudolfensis would be KNM-ER 2598. It sure looks like a large-brained, thick-boned specimen. It doesn't look much like KNM-ER 1470. But then, maybe 1470 is the unusual specimen...

Gathogo and Brown (2006) take on directly the issue of KNM-ER 1470 and KNM-ER 1813. The two were formerly considered contemporaries at around 1.89 Ma, but now KNM-ER 1813 is only 1.65 Ma.

KNM-ER 1813, lateral view

The real offshoot of this is that there are no longer any early small-skulled habilines. The question of whether KNM-ER 1470 and KNM-ER 1813 were too different to belong to a single species has drawn a lot of ink, but it was always a non sequitur, because the two weren't the only crania to consider. The more interesting observation had been that Olduvai Gorge preserved only small-skulled habilines, while Koobi Fora had both small and large ones. This was not only a geographic problem but also a temporal one, since the Olduvai habilines were all relatively late (less than around 1.8 Ma) and the Turkana habilines were mostly earlier.

Now the situation has changed. The small Turkana habiline, KNM-ER 1813, is now contemporary with the Olduvai sample. There are no longer any small-skulled early Turkana habilines. KNM-ER 1805 makes sense as a male of the later, small-skulled sample because it is relatively small-brained but robustly built (e.g., with a sagittal crest). That leaves KNM-ER 1470, KNM-ER 1590, KNM-ER 3732, and KNM-ER 3735 as plausible habilines before 1.85 Ma.

This seems like a nice sample as a possible ancestor for both later large-bodied Homo and later habilines. Heck, Wood (1991) even wrote this in his description of KNM-ER 3735:

Some features (e.g. vault thickness) ally it with a Homo erectus-like hominid, but in other areas (e.g. the frontal) it is more like crania such as KNM-ER 1813, a conclusion endorsed by Walker (1987) and by Leakey et al. (1989). Tobias (1989) includes KNM-ER 3735 within H. habilis (Wood 1991:134-135).

What more could you ask of a common ancestor? But then if some of this ancestral population would be expected to resemble later H. erectus-like specimens, then why not KNM-ER 2598?

And what, exactly, would make such a population -- with its mixture of H. erectus-like and habiline-like features -- different from Dmanisi? The answer, of course, is KNM-ER 1470. It's still the odd one in this lineup. But then, it does have the largest brain in this set, which might help to explain the rounded occiput.

Looking at what is left in the early part of the sequence is certainly interesting, but just as interesting is how all the H. erectus-like specimens are all bunched together between 1.65 and 1.45 Ma. This is the time interval that already held KNM-WT 15000, KNM-ER 3883, and KNM-ER 42700, and is just older than OH 9. Now we can add KNM-ER 3733, KNM-ER 730, KNM-ER 1808, and KNM-ER 1821. Isn't this an interesting sample? Don't you wish we knew about the other postcrania?

It seems to me that the hypothesis that H. erectus-like hominids first appeared in Africa around 1.65 Ma has interesting archaeological consequences. This isn't long before the appearance of the earliest Acheulean, and it plausibly makes the Developed Oldowan-Acheulean sequence a correlate of this evolution.

It is markedly not coincident with the earliest such evidence in Asia. But that raises the Dmanisi question again, doesn't it?


Brown FH, Haileab B, McDougall I. 2006. Sequence of tuffs between the KBS Tuff and the Chari Tuff in the Turkana Basin, Kenya and Ethiopia. J Geol Soc 163:185-204.

Gathogo PN, Brown FH. 2006. Revised stratigraphy of Area 123, Koobi Fora, Kenya, and new age estiamtes of its fossil mammals, including hominins. J Hum Evol (in press) DOI link

McDougall I, Brown FH. 2006. Precise 40Ar/39Ar geochronology for the upper Koobi Fora Formation, northern Kenya. J Geol Soc 163:205-220.

Wood B. 1991. Koobi Fora Research Project, Volume 4, Hominid Cranial Remains. Clarendon Press, Oxford.