Ardipithecus challenge explication: the molecular clock

I've had a chance to mull over the exchange between Esteban Sarmiento and Tim White and colleagues in Science this week (Sarmiento 2010, White et al. 2010). It is not really fair to rely on brief technical comments to straighten out the meaning of a fossil skeleton. Each set of authors had less than 1000 words to put forth their arguments, which means that there were doubtless many pieces of support that they had no space to include.

But when I write a technical comment, I spend a lot of time and effort to make the important points in that small space. If we can't agree on the basic outlines of the issues in a thousand words, I expect that ten thousand wouldn't settle anything, either.

I have a wide array of reactions to the points in these comments, but I think it will be most useful for me to focus on just three issues. I'm going to include a lot more text than I would for a technical comment, both so that I can include the direct quotes from Sarmiento and White and colleagues, and so that readers with less direct knowledge of the issues can follow along. And I'll divide each issue into its own post, so that it doesn't take a week to get something posted.

Let's start with the molecular clock argument. Sarmiento puts it briefly, depending on citations to do the lifting:

Over the past 40 years, a multitude of independent biomolecular studies based on different methods, some analyzing millions of DNA base-pair sequences, have arrived at a minimum human/African ape divergence date of ~3 to 5 million years before the present (1926)a date that accords well with those based on comparative anatomical studies of living and fossil hominoids (15). With a 4.4-million-year geologic age (1), Ar. ramidus probably predates the human and African ape divergence.

As I mentioned earlier this week, I discussed the issue in some depth last fall. The same argument originally was made by Vince Sarich, when the biomolecular evidence was based on antibody reactions to blood albumin, and the question was whether Ramapithecus was too old to be a hominin. Sarich (1971:76) memorably wrote:

[O]ne no longer has the option of considering a fossil specimen older than about eight million years a hominid no matter what it looks like.

David Pilbeam and others had claimed Ramapithecus as a hominid mostly because of its dental similarities to Australopithecus. Later, it became clear (especially thanks to David Frayer and Leonard Greenfield) that Ramapithecus wasn't even a valid taxon; the remains were females of Sivapithecus. Later it was shown that Sivapithecus itself had the forelimb of an arboreal quadruped; it apparently did not have a locomotor strategy like that of living great apes.

Sound familiar?

Sarmiento is correct. Over the past ten years, the human-chimpanzee divergence time has usually been put around 4 million years ago. Two things make this a deeper problem than it may appear. This estimate refers to the population divergence, and is a function both of the average genetic divergence and the variance among genetic loci in that divergence. That means that a simple recalibration to a lower mutation rate may not be enough to raise the estimate substantially.

Second, the date of Ardipithecus isn't 4.4 million years -- it's 5.5 million, the time assigned to Ardipithecus kadabba. Unless they want to sunder the genus, White and colleagues really need a much higher population divergence time than the range most studies have been reporting.

It's a complicated issue. So I was very interested to see which parts of this problem White and colleagues were especially focused on. How would they respond to the Sarich argument?

[Sarmiento] argues that biomolecular studies accurately converge on a divergence date of approximately 3 to 5 million years ago, concluding that Ar. ramidus "probably predates the human and African ape divergence." However, his cited estimates vary widely and all rely on inadequate calibration. Indeed, the strongest calibration is now from hominids themselves: Late Miocene fossils from Chad, Kenya, and Ethiopia whose derived characters effectively falsify late divergence estimates (2).

I found this really disappointing. There's no attempt here at any sensible critique of the molecular divergence time. Why is the calibration inadequate? What is the maximum human-chimpanzee divergence date we get by assuming that Chororapithecus is on the gorilla clade? Do they have a candidate for a significantly earlier pongine than Sivapithecus indicus? Do White and colleagues advocate an Eocene divergence of hominoids and cercopithecoids? Do they claim a mutation rate slowdown in humans, or in hominoids?

Instead of giving a sensible response, White and colleagues resort to circular logic. In their description, molecular comparisons can never show that Ardipithecus is too early to be a hominin, because we can never accept a calibration that shows Ardipithecus is too early to be a hominin.

References:

Frayer DW. 1976. A reappraisal of Ramapithecus Yearbook Phys Anthropol 18:19-30.

Greenfield LO. 1979. On the adaptive pattern of "Ramapithecus". Am J Phys Anthropol 50:526-548.

Sarich VM. 1971. A molecular approach to the question of human origins. In (P. Dohlinow & V.M. Sarich, Eds.) Background for Man: Readings in Physical Anthropology, pp. 60?81. Boston: Little, Brown.

Sarmiento EE. 2010. Comment on the paleobiology and classification of Ardipithecus ramidus. Science 328:1105. doi:10.1126/science.1184148

White TD, Suwa G, Lovejoy CO. 2010. Response to Comment on the paleobiology and classification of Ardipithecus ramidus. Science 328:1105. doi:10.1126/science.1185462