Woronso-Mille: A ladder not a bush

In a new paper, Yohannes Haile-Selassie and colleagues describe new hominin fossils from Woranso-Mille, Ethiopia. A good thing: It gives somebody like me a rationale for describing early hominins from the point of view of Hadar. You see, Hadar is the first sample to include a really complete skeletal representation. You can present earlier sites as a series of “firsts”, but that’s kind of misleading. Now, the simple Ardi-Lucy comparison carries a lot of water for teaching early hominins, and if we can assume that the samples intermediate in time are mostly A. afarensis-like, so much simpler.

Oh, and one more really good thing: Standard dental measurements are provided in the text of the paper. Thank you, AJPA! We may not get all the specimens, but at least we can check the statistics.

But there’s a chance that things are not so simple as they seem, that there are mysteries still waiting to jump out of this sample and scare us at night. I imagine that some people are less than thrilled about this paper, which explicitly rejects the reality of one Leakey-named species and ignores another into obscurity. One might expect me to welcome our new lumping taxonomic overlords. And yet, this little paper doesn’t provide some information and comparisons that seem like curious omissions. Which makes me wonder…

The fossils from Woranso-Mille are between 3.6 and 3.9 million years old – basically older than Laetoli and younger than Kanapoi. Since the Laetoli sample is A. afarensis, and the Kanapoi sample is A. anamensis, we can expect that the Woranso-Mille sample would say something about how these two species were related to each other. The fossils might be one species or the other, they might be intermediate between them, or they might even be something altogether different.

What is there?

The sample as described is almost exclusively dental, with only a fragment of mandible and another of maxilla tossed in the mix.

Some readers may have been under the impression there’s more at this site, and indeed I am as well. I think I’ve even seen them for 500 milliseconds at a meeting once upon a time. Of course, maybe that was a dream. Much in paleoanthropology seems to be fading into a unicorn fairyland these days…

Wait a minute! It’s for occasions like this that I have a blog! As it turns out, I took some notes on Woranso-Mille back in 2007.

Now, I have to warn you: These notes were so snarky that I didn’t dare hit “publish”. But there’s no sense shirking responsibility for them now. Next thing I know, some crank will be hacking my server to bring all this snark into the open!

Along with many other people, I got to see the hominids from Woronso-Mille this spring. Then again, see is probably an overstatement. I mean, when you see something, generally light waves from the objects actually have time to strike your retinas. I couldn't swear that anyone actually had that experience during Yohannes Haile-Selassie's talk to the Paleoanthropology Society. Sure, there was a subliminal impression that the pictures were there. And yet, Powerpoint and automatic timing can do magical things. I experimented a bit later, to try to estimate just how long the pictures had been up there. The 500 millisecond setting seemed about right. Definitely automated. Too short for microsaccades to bring in the edges of the fossils properly. And many of them were in situ photos, with a lot of brown-on-brown. Hard to pick out edges at all, and some edges were still in the ground. I mean, really, work out the time that Santa Claus has to spend in each kid's house on Christmas Eve to make it to all the world's children in one night. That's the kind of time we're talking about.

See what I mean. I mean, that’s over-the-top snark. Still, it’s better material than I usually work with, so I can’t for the life of me figure out why I didn’t publish it. It goes on:

Don't get me wrong. I think it's entirely appropriate to hide the images, dim them, heck, don't even show them if they don't want to. Think of all the yokels like me who could tell immediately from a decent picture whether the fossils were A. afarensis or not, and go shooting off their servers to the rest of the world. Hard work in the field, with the high risk of failure, deserves every possible reward -- certainly the right to take the necessary time to make a careful analysis. I hardly ever make any comments after I hear a public talk, unless the material is already well-known or described elsewhere. And there are other practical reasons not to talk about it -- for one thing, people sometimes change their minds! But why should I feel any compunction about prognosticating on fossils that are announced in the press? Hey, if they didn't want the attention, they wouldn't have a press conference, right? I'll bet they didn't make the press sit through the half-second slide show! Haile-Selassie announced several of the Mille fossils in 2005, notably the partial skeleton -- of which they are still trying to find more parts. At the time, I wrote about it, Rex Dalton wrote about it, Ann Gibbons wrote about it, twice, the AP wrote about it. Good times were had. Oh, those good times. Sure, no descriptions. Granted, in situ brown-on-brown photographs with buried edges. But good, good times.

How could I have forgotten those good times?

Now, there is a second press offensive underway. The best stories are at National Geographic News and The Cleveland Plain Dealer (Haile-Selassie's in Cleveland). It's an important site, with dozens of specimens. Hey, maybe they're like the Laetoli footprints and they rebury them when they're done looking at them. Kind of like catch-and-release.

The stealth mandible

All this was nearly three years ago. Which is a bit strange, considering that the current paper still doesn’t include all the specimens. Assuming the 2007 illustrations were correct, the current paper doesn’t even include all the Woranso-Mille dental specimens, as at least one mandibular dentition appears to have been omitted. It is, of course, possible that the news reports had the wrong picture.

Here are my 2007 thoughts on the matter:

We can probably answer this already: the National Geographic story includes a picture of the most complete mandible, and it looks an awful lot like LH 4, maybe with a more sloping symphysis. It's a rotten view - artistic, sure, but a lousy angle for comparison.

This mandible is not included in the current paper. It is pretty obviously the most diagnostic of the mandibular/dental specimens, if it’s from Woronso-Mille. I wonder if National Geographic really may have credited the wrong photo to Haile-Selassie? Very strange. In any event, it’s an important question since the sample of other postcanine teeth in the paper is generally 2-3 specimens. A missing postcanine dentition would make a lot of difference to our picture of the variation.

OK, continuing on:

But still, the teeth appear to fall into the Laetoli-Maka-Hadar sample, the postcanine rows diverge from each other, and the symphyseal morphology in A. afarensis is certainly variable enough to encompass this mandible. Really the only missing feature that would be helpful is the P3, but unless other specimens turn out to be outside the Hadar range, I would assume this is going to be assigned to A. afarensis.

Which does make me wonder how much the hidden mandible has driven the paper’s conclusion. On the basis of the specimens they published, the majority of dental features seem to argue in favor of A. anamensis, as they explicitly write. They mention only a few “derived features” also present in later sites. Given the date, one might just as easily argue that these “derived” features were actually low-frequency variants heretofore unrecognized in the small A. anamensis sample, so that Woranso-Mille extends the range in that species while maintaining its overall anatomical pattern. The stealth mandible, if indeed it exists, looks more persuasive fitting in the pattern of A. afarensis.

The unanswered phylogenetic questions are chiefly about what other lineages there may have been at the time. Mille might answer that question if substantial hominid diversity were found there, or at least something really different from the other sites. But no apparent evidence of such diversity was apparent in the public lecture. Maybe there are surprises waiting, but this team in the past has argued pretty strongly for taxonomic conservatism. On the other hand, this is what Haile-Selassie told the Cleveland Plain Dealer:
"The current hypothesis, which so many people seem to accept, is that they were ancestral descendents [sic, I'm assuming that's a misquote]- that anamensis gave rise to afarensis," Yohannes Haile-Selassie, expedition co-leader and anthropology curator at the Cleveland museum, said in a phone interview from Addis Ababa. "To test that, we need fossils. That's why we think these specimens are really, really important."

Sinking A. anamensis

In their current paper, Haile-Selassie and colleagues conclude the following:

The Woranso-Mille hominids are signi?cant for understanding the evolutionary history of early Australopithecus, particularly due to critical placement within a previously poorly known time period, 3.5 and 3.8 Myr. They are of paramount importance in testing hypotheses about the ancestordescendant relationship between Au. anamensis and Au. afarensis. The Woranso-Mille hominids shed some light on whether Au. anamensis and Au. afarensis are two distinct species, or parts of a single evolving lineage undergoing morphological change through time. Dentally they are more similar to Au. anamensis from Allia Bay than to Au. afarensis from Laetoli. However, they also share some derived characters with Au. afarensis from Hadar and Laetoli. Based on the currently available evidence, the Woranso-Mille hominids are temporally and morphologically intermediate between the more primitive Au. anamensis from Allia Bay and the slightly derived Au. afarensis sample from Laetoli (Ward et al., 2001; Kimbel et al., 2006). They appear to potentially represent a transitional population within an anagenetically evolving Au. anamensis-Au. afarensis chronospecies (White et al., 2006; Kimbel et al., 2006) providing further support to the well-established hypothesis of ancestordescendant relationship between the two species. To test this and other alternative hypotheses rigorously, and elucidate the evolutionary history of early Australopithecus, more complete fossil specimens are needed from the critical time period between 3.6 and 3.8 Myr. However, what appears to be evident with the discovery of new fossils spanning the 4- to 3.5-Myr interval is that morphological differences between Au. anamensis and Au. afarensis do not warrant a species level distinction (emphasis added).

Buh-zaaaaaaam! All your species are belong to us! Kenyanthropus? We won’t even dignify it by using the word. A. anamensis? Sunk like the Bismarck.

The fundamental debate here is semantic. Everyone seems to agree that anagenesis (that is, gradual evolution over time) is a likely hypothesis for this lineage. Where they disagree is how to handle the taxonomy.

  1. Strict cladists want to name species based on the appearance of unique features (that is, phylogenetic species), in which case A. anamensis is a species, A. afarensis is a later species with new characters, and very possibly we need to resurrect Praeanthropus africanus for the Hadar sample, even if it mostly overlaps, since it has a few characters never found earlier and represents a broader sample of postcranial anatomy, which is entirely unknown at earlier sites.

  2. Strict users of a Wiley-like Evolutionary Species concept always place anagenetic lineages into one species. So, the single lineage hypothesis lumps A. afarensis and A. anamensis together. And as Haile-Selassie and crew go on to point out, we might even lump Ardipithecus, if it’s the lineal ancestor of the later hominins.

  3. Not-so-stickly people, which is most everybody, pretty much recognize species along with the crowd. A. anamensis has a history now. It’s not just early A. afarensis, because, well, lots of people said so. And after all, you can tell the difference between them if you look carefully.

What’s interesting (at least to me) is to read Kimbel and colleagues’ 2006 paper, keeping in mind the “following the crowd” scenario 3. In this light, much of that paper is boundary defense. A. anamensis had already elbowed its way into the textbooks, and the paper recognizes the existing taxonomy without attempting any revision. But the demonstration of anagenesis within A. afarensis would be sure to provoke some strict cladists to name some more species – a species for Hadar, for example. Kimbel and colleagues reiterated that anagenesis within A. afarensis is expected – it’s part of the species’ literature, now. So the paper tried to draw two lines in the sand: on the one hand, A. anamensis is real; and on the other hand, no further distinction within A. afarensis is warranted. Taxonomic containment.

Just a pelvis away...

But, drawn in this way, both lines in the sand might be washed away by a single discovery. The present pattern of evidence is mostly dental and mandibular. Woranso-Mille may be only one postcranial specimen away from lacking a bunch of derived postcranial characters that are well-evidenced at Hadar.

After Ardi, I think this is a serious possibility because of the scope of postcranial innovations at Hadar that are not evidenced anywhere earlier. It could be that all the postcranial traits of Lucy and her kin are lineage-typical, going back all the way to Kanapoi (and don’t forget the A. anamensis from the Middle Awash). But we don’t know this. Given Ardi, it appears that the adaptive package appeared rapidly, after 4.2 but before 3.5 million years ago. It seems to me that there’s every chance that A. anamensis, and possibly the Woranso-Mille sample, hadn’t built the whole package yet.

Ah, now I’ve gotten down to the end of my notes. I think I’m starting to remember why I didn’t put them up at the time:

While the evidence for bipedality in the earlier A. anamensis is not nearly so extensive as that in A. afarensis, nevertheless it is quite compelling, particularly the KNM-KP 29285 tibia. You'd get pretty long odds betting that the Mille pelvic bones looked very different from Lucy's. I have no information about the pelvis at all, certainly no photos, but it would indeed be a surprise for it to be outside the A. afarensis-A. africanus range of variation. But then, all it would take is one funky-looking pelvis to throw the whole question of pre-4.0-million-year-old hominids wide open. So maybe we should hope that it's strange.

Well, we certainly got one funky-looking pelvis, didn’t we? I’m beginning to think I should republish old notes more often. What are the chances that another funky pelvis is waiting to be published?

Could it be that Woranso-Mille could represent an intermediate postcranial form at 3.7 million years? That would be one good reason to nail down the question of anagenesis from the craniodental perspective.

I think we may already have a hint at the answer. It’s a little hard to imagine that Haile-Selassie and colleagues would propose sinking A. anamensis if they already knew that their skeleton has a different postcranial anatomy than represented at Hadar.

There’s one more thing worth mentioning: this paper doesn’t include any discussion, comparison – it doesn’t even breathe the name – of the other 3.5-million-year-old hominin. It’s not just a skull; there is a sample of teeth from that unmentioned site, which of course may or may not represent the same taxon. Like a forgotten stepchild of paleoanthropology. Is it possible that peer reviewers have already forgotten it’s existence?

When I wrote about what I’m wondering, well, this isn’t the only paper to have recently omitted this obvious comparison. I’ll have more on that little problem later on…


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