In the fossil record, a species is a hypothesis. We can't test that hypothesis in the way we do with living animals. Even in the dark, after all the paleontologists have left, the fossil bones just won't get it on. No reproduction, no test.
So, sometimes we have to live with hypotheses that we can't immediately test. Because many hypotheses are wrong, we have to keep juggling in our minds the names of more species than probably existed.
All the juggling frequently leads to confusion. One may reasonably wonder how we know that species X evolved into species Y, when half the field rejects the hypothesis that X was a real species. Often we don't disagree about the "evolving into", but we do disagree about the boundaries and other relationships of the populations -- which we can understand only indirectly from the fossils. The fossils don't change, but our hypotheses about them do.
That brings us back to A. anamensis. Here's a hypothesis about an ancient lineage of hominins, based on a certain number of differences from later fossil samples assigned to A. afarensis. That's a precarious place for a hypothesis, because from the beginning, the definition of A. afarensis has encompassed geographic and temporal variability. How hard would it be to recognize a little more temporal variability? Not very.
So, as I described last week, Haile-Selassie and colleagues (2010) propose sinking A. anamensis ("Woranso-Mille: A ladder not a bush"). I should mention, if I haven't already, that I have great sympathy for this viewpoint. Absent some compelling evidence that the lineage includes a speciation event, I prefer slow gradual anagenesis to be categorized into one evolving species, not an arbitrary set of chronospecies.
In the same post, I described the work of Kimbel and colleagues (2006), who had argued for anagenesis in the same sample of A. anamensis and A. afarensis-referred fossils, but retained the two distinct names for them. One thing stands out as a mystery to me in that paper. Why didn't they let Kenyanthropus make the argument for them?
If you want to establish that A. anamensis is taxonomically valid, the simplest way to answer all critics is if it has more than one descendant. You don't have to demonstrate the phylogeny beyond all doubt, I would say, you just have to take the hypothesis seriously.
In this instance, we seem to have two good candidates for a non-A. afarensis descendant of A. anamensis. The more obvious of them is Kenyanthropus. Why didn't they advance the hypothesis of a A. afarensis-Kenyanthropus clade? Here's what Kimbel and colleagues wrote about the latest Leakey find:
A more significant concern is the presence of Kenyanthropus at 3.5Ma. Kenyanthropus may demonstrate cladogenesis prior to this time, but this taxon is only directly relevant to the analysis if any of the samples share derived character-states with it. At present, the Kenyanthropus hypodigm does not match the others in the availability or quality of character data for the mandible and anterior teeth, while the evidence that exists (from the maxilla, for example) does not suggest a close relationship of Kenyanthropus to any of the phena considered here (Leakey et al., 2001).
So, they took the hypothesis off the table. And again, they've refocused the question solely upon dental and mandibular evidence.
This is a very large hole. At 3.5 million years, KNM-WT 40000 is earlier than any other comparably complete hominin skull, except for Ardi and Toumaï. It doesn't look like them, that's for sure. Not the same phenum at all. If you're going to insist that KNM-WT 40000 isn't A. afarensis, it's hard to see a better hypothesis than that it's descended from A. anamensis.
If it doesn't have a close relationship to either A. afarensis or A. anamensis, then I'm at a loss to figure out what they think it is related to!
I'm willing to believe White's (2003) argument that it just is a member of the A. anamensis-A. afarensis lineage, but I don't see the contrary argument that it's so different from this lineage that it must be derived from some as-yet-undiscovered hominin. Keep in mind that the argument was formed by people, many of whom already knew basically what Ardi's skull was going to look like. I just can't feature why this phylogenetic problem didn't raise itself to a higher profile.
As I wrote above, we juggle more hypotheses than can be true. In this instance, the null hypothesis is that all these hominins belong to a single evolving species, which would be called A. afarensis. But one alternative, in which A. anamensis existed as the ancestor of A. afarensis and at least one other species, has some utility. It lets us refer clearly to phylogenies with late-diverging sister taxa. As Yoda might have said, there is another possible sister taxon for A. afarensis: A. africanus might be derived from, or might itself be, a South African contemporary of the Hadar-Maka-Laetoli sample. The earliest Sterkfontein dates go up and down; one or more of the remains might be contemporaries of Laetoli or even earlier East African localities.
Anyway, I'm not so interested in this question of bushes versus ladders, or "Pliocene diversity". I'm more focused on the curiously non-hominin-like features of Ardipithecus. If we suppose a late molecular divergence of hominins and chimpanzees (say after 4.4 million years ago), then Ardipithecus might be an ape or ancestral (stem) hominid, not a hominin. If so, then samples now referred to A. anamensis, including later specimens from Aramis, Asa Issie, and Kanapoi, are in fact the earliest-known members of the human lineage.
A. anamensis is no afterthought in that case, it may be the stem hominin.
That is, if it hasn't been sunk into A. afarensis.
UPDATE (2009-12-10): A reader writes to remind me about A. garhi, which presents itself as very much like A. africanus and is quite a lot later than any known A. afarensis samples. One alternative is that A. africanus is simply the latest element of the single A. anamensis-A. afarensis-A. africanus anagenetic sequence.
I don't think that's very easy to reject, considering the lack of good cranial remains between 3 and 2.5 million years ago (or for that matter, even later) in East Africa.
Some of the Sterkfontein specimens, particularly those from Jakovec cavern and from Member 2 (including Little Foot) have been suggested to be older than 3 million years. Partridge and colleagues (2003) put them at older than 4 million years ago, which would make them rivals of A. anamensis as the earliest hominins. But Berger and colleagues (2002) argued (apparently preemptively) that these early dates are not necessary on faunal or magnetostratigraphic grounds.
A second reader wonders why I didn't mention A. bahrelghazali as a possible sister to A. afarensis. Well, nothing's impossible, but I'd say the case for the Bahr el Ghazal mandible being distinct from the Hadar-Maka-Laetoli sample isn't very strong. Still, we're only talking about hypotheses here, I suppose.
Berger LR, Lacruz R, de Ruiter DJ. 2002. Revised age estimates of Australoipithecus-bearing deposits at Sterkfontein, South Africa. Am J Phys Anthropol 119:192-197. doi:10.1002/ajpa.10156
Haile-Selassie Y, Saylor BZ, Deino A, Alene M, Latimer BM. 2010. New hominid fossils from Woranso-Mille (Central Afar, Ethiopia) and taxonomy of early Australopithecus. Am J Phys Anthropol (in press) doi:10.1002/ajpa.21159
Kimbel WH, Lockwood CA, Ward CV, Leakey MG, Rak Y, Johanson DC. 2006. Was Australopithecus anamensis ancestral to A. afarensis? A case of anagenesis in the hominin fossil record. J Hum Evol 51:134-152. doi:10.1016/j.jhevol.2006.02.003
White T. 2003. Early hominids -- diversity or distortion? Science 299:1994-1996.
Suwa G, Asfaw B, Kono RT, Kubo D, Lovejoy CO, White TD. 2009. The Ardipithecus ramidus skull and its implications for hominid origins. Science 326:68e1-68e7. doi:10.1126/science.1175825
Leakey MG. Spoor F, Brown FH, Gathogo PN, Kiarie C, Leakey LN, McDougall I. 2001. New hominin genus from eastern Africa shows diverse middle Pliocene lineages. Nature 410:433-440.
Partridge TC, Granger DE, Caffee MW, Clarke RJ. 2003. Lower Pliocene hominid remains from Sterkfontein. Science 300:607-612.