I have been somewhat busy and have just read the Richmond and Jungers paper on Orrorin tugenensis, along with some of what has been written in the blogosphere about it. Unfortunately, I have to disagree with one post on the subject – that of Hawks. In his post Hawks is trying to argue that there isn’t that much difference between the results of Richmond and Jungers and the position of Senut and Pickford:
The overall morphological pattern of this femur, with its long neck and broad shaft, is much like known australopithecine femora. But to go a bit further, their metric comparisons show BAR 1002’00 to be the most Homo-like of the early hominid femora they examined, and their phenetic cluster puts it basal to the other australopithecines. That’s pretty much exactly what Senut et al. have consistently said [emphasis mine – afarensis]. So I have a hard time understanding how those observations refute the idea that Orrorin has a more Homo-like femur than later australopithecines!
Really?
The quote below is from Pickford and Senut’s 2002 paper on the Orrorin femur, like Richmond and Jungers, they focused mainly on BAR 1002’00:
In conclusion, from a systematic point of view Orrorin is a hominid sensu stricto, and in numerous features it is not chimp-like. In several features, Orrorin is closer to humans than australopithecines are which suggests that it may be more closely related to Homo than it is to Australopithecus and/or Paranthropus. If we are correct, then Australopithecus may represent a side branch in hominid evolution that became extinct without giving rise to Homo, a hypothesis that has already been suggested by Coppens [9] and Senut [32,37] among others.
Sounds different to me. Sounds like Pickford and Senut are claiming an Orrorin-Homo link to the exclusion of Australopithicines. Note also, that Senut, in the above quote, characterizes the paper describing the Orrorin finds as saying much the same thing.
Literature Cited
Pickford, M.H.L., B. Senut, D. Gommery & J. Treil (2002).
Bipedalism in Orrorin tugenensis revealed by its femora.
Comptes Rendus Palevol 1:191-203.
Filed under: Paleoanthropology | Tagged: Australopithecines, Orrorin tugenensis |
Afarensis: Anthropology, Evolution, and Science 
Nice catch, afarensis. Based upon several papers by Senut et al., I too understand that Senut et al. have tried to argue that Orrorin tugenensis is similar to Homo and that Australopithecus is not. Richmond and Jungers argue that Orrorin tugenensis is similar to Australopithecus.
Kambiz
I thought the same.
It would be nice to hear from him if he is sticking by that earlier post.
I think they agree on the anatomy and the behavioral implications of that anatomy. The big difference between Richmond and Jungers and the Senut camp are the conclusions they reach about the phylogeny. This is what most of the reports I have seen, correctly IMHO, focus on. Senut has, for years, been downplaying the role of australopithecines and Ardapithecus in human evolution. To focus on that is certainly fair game.
The question is not whether they are claiming different things about the phylogeny — they certainly are! The question is whether their data support their claims. Richmond and Jungers show that the Orrorin femur (one of three, by the way) is Homo-like in a way that other australopithecines are not. That is absolutely consistent with Senut et al.’s claims. The Senut et al. (2001) phylogeny shows Orrorin as a basal member of a clade including Praeanthropus (=A. afarensis) and Homo. Their phylogeny puts Australopithecus (excluding A. afarensis) and Ardipithecus in separate clades.
There is a rationale to split Australopithecus in this way — that’s the (now) long literature on limb proportions, with Lucy having much more Homo-like proportions than A. africanus. The complexity in this is where individual skeletons lie — OH 62 having strong forelimb dominance like STW 431, and BOU-VP 12/1 having relatively longer legs.
Richmond and Jungers wrote that their observations refute Pickford’s and Senut’s phylogeny. But they simply don’t. If Homo lacks many of the derived proximal femoral features of Australopithecus, and is similar to Orrorin in this respect, it is perfectly consistent with what Senut et al. (2001) claimed. I don’t especially subscribe to the Orrorin–Praeanthropus–Homo clade, but the femur is not an argument against it. I view this as a case of Richmond and Jungers wanting to make their phylogenetic position clear, and the reviewers letting them get away with doing so without any substantiation.
The question is not whether they are claiming different things about the phylogeny — they certainly are! The question is whether their data support their claims. Richmond and Jungers show that the Orrorin femur (one of three, by the way) is Homo-like in a way that other australopithecines are not. That is absolutely consistent with Senut et al.’s claims. The Senut et al. (2001) phylogeny shows Orrorin as a basal member of a clade including Praeanthropus (=A. afarensis) and Homo. Their phylogeny puts Australopithecus (excluding A. afarensis) and Ardipithecus in separate clades.
There is a rationale to split Australopithecus in this way — that’s the (now) long literature on limb proportions, with Lucy having much more Homo-like proportions than A. africanus. The complexity in this is where individual skeletons lie — OH 62 having strong forelimb dominance like STW 431, and BOU-VP 12/1 having relatively longer legs.
Richmond and Jungers wrote that their observations refute Pickford’s and Senut’s phylogeny. But they simply don’t. If Homo lacks many of the derived proximal femoral features of Australopithecus, and is similar to Orrorin in this respect, it is perfectly consistent with what Senut et al. (2001) claimed. I don’t especially subscribe to the Orrorin–Praeanthropus–Homo clade, but the femur is not an argument against it. I view this as a case of Richmond and Jungers wanting to make their phylogenetic position clear, and the reviewers letting them get away with doing so without any substantiation.
John, thank you for responding. I agree with you that the question is about whether the data support the claims being made. Where I disagree with you is over the nature of the claims being made by Senut and her research group (I’ll get back to that in a minute). Richmond and Jungers and Senut certainly agree on the fact that Orrorin was an australopithecine like biped with a certain amount of arboreality in its locomotor behavior. Richmond and Jungers does not address the thick vs thin enamel issue, but as you point out that is not a compelling argument anyway. The issue then boils down to whether or not BAR 1002’00 is, as you put it:
You claim Richmond and Jungers data answers this in the affirmative and that this supports Senut’s position. I have to ask where, because their Mahalanobis D2 indicates BAR 1002’00 is closer to early hominin taxa than to modern humans.
P.S. Off topic, but when are you going to take Luskin to task for quotemining Hawks et al 1999?
I’m glad you bring this up — I just saw an entire session at the meetings in which people consistently misrepresented the way that D2 distances apply to hypothesis testing!
In this paper, figure 1 has two subfigures: a 3-axis plot of canonical variates values for ape, human, and early hominid specimens, and a cluster tree. Clustering cannot answer the question, because it cannot test the hypothesis that the femur shares features with both Australopithecus and Homo, nor can it assess whether similarities are derived or plesiomorphic. But, even given these caveats, the clustering places Orrorin basal to australopithecine femora, the position predicted by Senut’s phylogeny.
The canonical variates are more clear. CV1 distinguishes humans from apes; CV2 distinguishes orangutans from African apes (and somewhat chimpanzees from gorillas), and CV3 early hominids from the rest (and somewhat chimpanzees from gorillas). Among the early hominids, BAR 1002’00 is clearly the closest to Homo, well within the Homo range on CV1 (the only one that distinguishes humans from apes). That means that it shares metric similarities with humans not present in other early hominids.
I will say again that I don’t favor this phylogeny; I just want the literature to engage these hypotheses correctly. I’m trying to read everything precisely (and to give Senut et al. their due) because the overarching problem is that all these Miocene hominids (Orrorin, Sahelanthropus and Ardipithecus included) predate the best estimates for human-chimpanzee divergence based on genetic comparisons. If Orrorin is actually Homo-like in relevant morphology, then that opposes the hypothesis that Orrorin is a stem chimpanzee-human relative instead of a member of the hominin clade.
Re: Luskin — I didn’t know about that. I’ll see what I can do…
No, it closer to KNM-ER 1503 – which Richmond and Jungers assign to Paranthropous boisei. In the plot of canonical variates that they present in the supplemental material of Can 1 and Can 2 BAR 1002’00 falls in an area of overlap between apes and humans on Can 1. On Can 2 it falls on the ape/human side rather than the orang side. Similarly, in the plot of PCA, agaon in the supplemental material (from which fig 1a is taken) the closest group to fossil Homo and modern humans is Australopithecus afarensis. BAR 1002’00 is, again closest to KNM-ER 1503.
On the Luskin thing, that isn’t the only time he has quotemined that paper. Seems to be one of his favorites – next to Wood and Collard.
I think we should be very precise about how to use Richmond and Jungers’ comparisons. They are phenetic measures, not cladistic tests.
Incorrect: BAR 1002’00 is closest to KNM-ER 1503 in the canonical variates plot. Therefore the two femora represent sister taxa, and Orrorin belongs to the Paranthropus clade.
Correct: BAR 1002’00 is between KNM-ER 1503 and recent humans in the canonical variates plot. Therefore it is more similar to humans than KNM-ER 1503 (or indeed any australopithecine femur) is to humans.
I am stressing the point that BAR 1002’00 is the most human-like of the early hominid femora in the comparisons that Richmond and Jungers present. There is no australopithecine femur that is more Homo-like than Orrorin in these comparisons. That’s what you expect for a specimen that predates the evolution of derived features in australopithecines, and is therefore perfectly consistent with Senut et al.’s interpretation.
This is a phenogram, not a cladogram, so one would be justified in taking the position that the data do not test any of these three hypotheses. If I were to take this position, I would have to conclude by the same logic that these comparisons are simply irrelevant to the position of Orrorin in the hominin phylogeny, and Richmond and Jungers were wrong to claim otherwise.
However, I tend to think that the phenetic comparison does provide some information about the cladistic relationships. The observation that BAR 1002’00 is the closest early hominin to modern humans does not argue against the idea that it is an early member (possibly a stem member) of the human clade.
Possibly there are other comparisons (such as ones considering character polarity) that would test Senut et al.’s phylogenetic hypothesis and come to different results. Senut et al. pretty much listed every possible derived character, however, so I think that these take the data about as far as they are going to go. What we would really want is a comparative assessment of the variation in Orrorin and other early hominin taxa.
I should mention that KNM-ER 1503 is placed by some workers into Homo…
Yeah, that is why I qualified it by saying Richmond and Jungers place it in Paranthropous. Unfortunately, it doesn’t group with the other early Homo they included in their analysis. I’m also not claiming that they are sister taxa. Nor would I characterize, based on the plots Richmond and Jungers provide, BAR 1002’00 as being between KNM-ER 1503 and recent humans, rather they are about the same distance from recent humans. I will also point out, again, that on both the principle components analysis and on the canonical variate analysis A. afarensis was more similar to humans than Orrorin. Yeah, a comparative assessment of variation would be good.
Figure S1.A shows CV1 and CV2, for which BAR 1002’00 is within the human range of variation, unlike all australopithecines. ER 1503 is the next closest to humans on this plot. Figure S1.B shows CV1 and CV3, for which BAR 1002’00 and A. afarensis are approximately equidistant to humans. Figure S4 confirms the position of BAR 1002’00 as the most human-like of the non-Homo femora, as assessed by D2 The plot for PC1, PC2, and PC3 in Figure S2 is the same representation (PC distances are D2 distances), and shows that BAR 1002’00 is the closest early hominid to humans on PC1 and PC2.
I don’t see how you can read the paper and conclude that A. afarensis is closer to humans than Orrorin. The two A. afarensis femora (AL 288-1 and AL 333-3) have the largest D2 distances from the human centroid. That’s one of the reasons I don’t really believe Senut et al’s phylogeny (which puts A. afarensis on the Orrorin-human clade).
What Richmond and Jungers really lack is a biomechanical model for why their measurements are meaningful. For instance, it seems reasonable that many of the apparent similarities between Orrorin and Homo come from the fact that Orrorin is larger than almost all of the other australopithecine femora. It’s not necessarily more like humans in its mechanics, just its size.
You are correct on the S1 and S2 graphs, I misspoke. Having said that, I don not see how you can claim that S2 supports you. On PC 1 A. afarensis and Orrorin are about the same distance from Homo. On PC 2 A. afarensis is much closer to Homo. On the D2, as I previously pointed out, Orrorin is closer to austalopiths than any other group.