Pliocene

Just noticing, in this John Noble Wilford article:

A new report, to be published Thursday in Nature, will review more skeletal evidence of the transitional aspects of the Dmanisi specimens.

More later...

UPDATE(2007/09/18): Wilford doesn't directly state the article's theme but it clearly has one: Why the heck can't these people agree about these fossils that have been out of the ground for thirty years?

The first answer that everyone has given him is about the "million year gap" between 3 million and 2 million years ago. People can't agree about early Homo because they can't decide what its ancestors looked like. Without any ancestors, they don't know which of the traits of early Homo are derived.

For a good example, we can turn to a feature Wilford doesn't mention: limb proportions. Recently, a lot of ink has been spilled discussing the evolution of arm size in later australopithecines and early Homo. OH 62 (probably Homo habilis) and A. africanus have been argued to have large arms compared to their legs. A. afarensis and Nariokotome (KNM-WT 15000, probably Homo erectus) have relatively small arms compared to their legs. Did H. habilis and H. erectus have different ancestors? Did H. erectus evolve from H. habilis, reverting its limb proportions to earlier A. afarensis? Or are all these comparisons just batty, since only three specimens have arm and leg elements whose length can be compared? There's no clear answer; but one of the most important specimens in the question (with sort-of-intermediate limb proportions) is the Bouri skeleton, BOU-VP 12/1, which at 2.5 million years old is right in the middle of that "gap."

The more you look at the "gap," the less gap-like it looks. For one thing, we have a pretty good idea of what was going on behaviorally during that million year span. The first stone tools are 2.6 million years old. The technology of these toolmakers -- although simple -- included all the basic manufacturing methods used before 1.5 million years ago. The tools were used to butcher animals and break bones for marrow; so we know that the toolmakers were depending on meat.

Second, we actually have quite a lot of fossils from this time period. The entire South African A. africanus fossil record, with the exception of a few early specimens like STW 573, come from this "gap." A fairly extensive record of the appearance and evolution of early robust australopithecines comes from this time period in East Africa.

And, here and there, a few specimens look Homo-like. Wilford's article discusses AL 666-1. To this we can add the Uraha mandible, Omo 75-14, an additional series of teeth from Omo, and possibly the Bouri BOU-VP 35/1 skeleton.

Properly considered, the rarity of early Homo in these contexts is not a problem; it is information. Wilford quotes Philip Rightmire to this effect, and we can easily expand on the basic concept. Early toolmakers did not undergo an immediate geographic expansion upon their origin. They spread across a relatively narrow strip of East Africa and stayed there for more than a half-million years. They were initially rare. That means that their adaptation was not immediately a barnburner of a success -- the early toolmakers took a while to perfect the adaptation of later Homo.

The middle part of the article takes in another reason for disagreement: whether H. habilis and H. erectus were ancestor-descendant:

Several scientists, notably Dr. White of Berkeley, took issue with the interpretation seeming to imply that evidence for the two species overlapping in time and exhibiting variable sizes was new. That, he said, had been recognized for a couple of decades.

Dr. Kimbel, who was not involved in the new research, defended the authors, saying that they had not "meant to imply that habilis could not have been ancestral to erectus, presumably on the basis of their being contemporaneous at Turkana," the site in Kenya where the fossils were found.

Susan C. Anton, an anthropologist at New York University who was a member of the Spoor-Leakey team, said, "My money is still on habilis as the potential ancestor, but there is a lot of room for additional knowledge, given the dearth of fossils."

None of these statements really disagree with each other. If anything, this particular question may have gotten easier to resolve lately, not as a consequence of new fossils, but as a result of new dates for many of the old ones. Susan Anton is later quoted saying that anagenesis "is the only option that is no longer on the table," and it seems to me that this is the clearest statement most likely to invite some hypothesis testing. But it is fairly clear that this problem cannot be resolved in terms of earlier fossils: I don't think there's any compelling evidence of H. erectus before 1.6 million years ago.

There is one significant word that doesn't appear in the article -- an absence that is especially interesting considering the quoted scientists:

Kenyanthropus

Remember, the dominant theme is about complexity and bushiness. And yet, here's that forgotten branch of the family tree; the one that was supposed to clarify everything by providing a different ancestor for KNM-ER 1470 from other H. habilis specimens, the one that showed a distinct line leading to Homo originating in the Early Pliocene.

I think our bush may have been pruned.

Gene Expression contributor p-ter expresses his (un)excitement about the Ventrome:

In terms of haplotype reconstruction, the authors make a number of dubious claims about the importance of their advances. It is not true, as stated in the introduction, that genome-wide association studies rely on phased haplotypes for analysis. In fact, most of the ones I have seen do nothing more than simply count up genotypes at each SNP in cases and controls and perform a chi-squared test. In most cases, haplotype-level analyses is simply not done. This may change in the future, of course, but it's difficult to see how what they've done (note they aren't even able to make ideal haplotype inference with the data) is that exciting.

Personally I can't understand why people are calling it the "Craigome". I suppose that's supposed to go along with the "personalized" medicine part. If it's personal, it must use your first name! That should give aid and comfort to the telemarketers, I suppose -- or maybe the genome muckety-mucks have been deploying cameras on the hucksters' sidelines!

Then again, somebody has surely thought it through. Maybe it's about Watson -- the "Watsonome" sounds like some kind of assisted living arrangement. Oooh -- you know George Church is next! I've got "Georgeome" on my mind!

Horsing around

Meanwhile, Brian Switek of Laelaps tells the changing story of horse evolution. This is a great article full of references and classic reconstructions of horse phylogeny, and Switek reviews some classics:

The 1966 edition of Romer's Vertebrate Paleontology fairs better overall, but is still found wanting. The same straight-line illustration I just mentioned is found in the section treating perissodactyls as a group, and the skeletons of Eohippus, Mesohippus, and Hippidion are shown left to right across pages 266 and 267. While the text does mention an overall diversity of forms, as well as using certain genera for the "type" from which modern horses evolved, the overall visual impression of simple anagenesis remains. Again, I doubt the casual reader picked up Romer's book for light nightly reading, but it is strange that the progressive ideas about evolution during that time are so poorly represented.

The versions of horse phylogeny from the 1970s and 1980s bear a strong resemblance to various hominid phylogenies that have emerged in the last 20 years. The theme of the essay is how the original unilineal picture of horse evolution became the current "bushy" version, and it's a good account. As for hominids --- well, keep in mind that the bushy horse phylogeny extends back to the Eocene. Horse phylogeny in the Pliocene and Pleistocene is not exactly a fractal version of the larger picture. Still, there were speciations of Equus during this time period, including the current diversity and some extinct horses.

Guts and sex

Carl Zimmer puts together three recent stories about gene duplications, including the human amylase-diet connection. I'll be posting on that one later this week, and Zimmer's account of erstwhile digestive proteins in the reproductive tracts of female Drosophila is a fascinating addition:

As gene copies build up in a genome, some of them can take on new jobs. And here is where the fly sex comes in. In some species of Drosophila flies, the males and females wage a biochemical war of the sexes. The males can boost their reproductive success by manipulating their female mates. They do so with a cocktail of chemicals mixed into their seminal fluid. Some of the chemiicals stimulate the production of eggs, some chemicals create a big mass in the reproductive tract that may make it harder for a female fly to mate with another male. Females produce chemicals of their own to counteract the male chemicals. You'd expect that mutations that gave either the male or the female an edge in this arms race would be strongly favored. And in the August issue of PLOS Genetics, scientists published a list of female reproductive proteins that show strong signs of natural selection. A dozen of these proteins are proteases--they are good at slicing other proteins apart. The flies only make these proteases in their reproductive tract, and yet the proteins are not closely related to other proteins made in the reproductive tract. Instead, their closest relatives are families of proteases that the flies make in their guts in order to digest food.

OK, maybe fascinating isn't the right word. I guess it depends how close you are to lunchtime. Or bedtime. Ewww.

"Previvors'" tales

Amy Harmon in the New York Times writes the story of breast cancer "previvors," women who have tested positive for one or more risk mutations, and must choose whether to take preventative measures such as prophylactic surgery:

Her father, who once feared he would lose his wife to cancer, encouraged the surgery. Her sister reminded her that cancer might be cured in a few years if she could wait.

Her aunt said she hated to see her niece embrace a course of action akin to "leechings of the Dark Ages." A cousin declined even to take the DNA test.

The article is illustrated by a pedigree that shows the affected family members. In this form, you can really sense the feeling of dread, seeing four generations of women ticked off the chart.

EyeOnDNA asks about the effects of the new BRCA1 and BRCA2 tests, which are intended to inform people of their carrier status for breast cancer susceptibility alleles:

This past week we saw the unveiling of a controversial general advertising campaign for the Myriad BRACAnalysis genetic test for breast and ovarian cancer susceptibility. One oft-quoted statistic in relation to the test states that "only 30,000 of more than 250,000 American women estimated to carry a mutation in BRCA1 or a related gene, BRCA2, have so far been tested." Myriad hopes to increase the number of mutation carriers detected. But what kind of counseling and support will women receive if they test positive? (emphasis in original)

Hsien puts the question in context by referencing Harmon's account.

A nice piece in The Guardian about the chimpanzee population near Bili, DRC. The lede is the suspicion of an apparent leopard kill -- that's chimpanzees killing a leopard -- but the other details are interesting:

[Cleve] Hicks said the animals also have what he calls a "smashing culture" - a blunt but effective way of solving problems. He has found hundreds of snails and hard-shelled fruits smashed for food, seen chimps carrying termite mounds to rocks to break them open and also found a turtle that was almost certainly smashed apart by chimps.

Like chimp populations in other parts of Africa, the Bili chimps use sticks to fish for ants, but here the tools are up to 2.5 metres long.

All these observations of chimpanzee behavioral diversity are pretty exciting, since they really provide an interesting model of early hominid diversification. Longstanding subspecies-level populations with strong behavioral differences involving food collection and diet may describe both A. afarensis (along with regional variants) and Late Pliocene Homo.

A mention is in the article about whether they should be considered a new (fifth) chimpanzee subspecies. I'd say that's probably a given at this point; doubtless the Fongoli chimpanzees will become a sixth.

Do they deserve it? Well, I suppose they're at least as distinct as the others behaviorally and anatomically, which isn't saying much. Genetically, it's an open question so far, but I wouldn't be surprised if they were as distinct as central and eastern chimpanzees, and certainly moreso than P. t. vellerosus. It's hard for me to see where ape subspecies taxonomy is going to end, since there are few scientific interests vested against further taxonomizing. Sure, there are conservatives, but none with enough firepower to roll back P. t. vellerosus, apparently. So, more subspecies lie in our future: I would guess two or three more for Bornean orangutan populations and who knows how many Western lowland gorilla populations will qualify?

(via Gene Expression)

Out of this week's Science Times special on evolution, I clicked into John Noble Wilford's article first, titled "The Human Family Tree Has Become a Bush With Many Branches".

Now, I don't know about you, but that seems like a boring headline to me. They've been talking about human evolution being a bush for going on 20 years now. It was an old idea when I was in graduate school. So it seems like, if this is all we have going on, the "new frontier" of paleoanthropology must be pretty dull.

The writer doesn't write the headlines, and the headline doesn't describe Wilford's story, which is basically a verbal slide show of fossil discoveries over the last decade or so. Some bone pictures (of the actual species discussed) accompany the article, and it's a good enough sort of account of new finds since 1990, framed around the tension between fossil finders and molecule mavens.

But I'll be a little critical. The thesis is that paleoanthropologists suffered a crisis of confidence after molecular data came online in the 1980's, and "a rapid succession of fossil discoveries since the early 1990's has restored" it.

Well, OK, maybe. But consider the listed discoveries: Kenyanthropus, Ardipithecus ramidus, Ardipithecus ramidus, Orrorin tugenensis, Sahelanthropus tchadensis, Homo floresiensis, and Australopithecus anamensis. Of all of these, only Ar. ramidus and Au. anamensis have gone without significant controversy.

We can set aside H. floresiensis for a moment -- the controversy about it being possibly the loudest, it also stands apart as the only species listed younger than 3.9 million years. All of these early Pliocene and Miocene species have also been challenged -- by the discoverers of the others, by old hands, and by young upstarts like me. At least one research group has claimed that all of the Miocene "genera" may actually belong to one species. Another has claimed that most of these "hominids" may actually be apes.

Whether there was any crisis of confidence among paleoanthropologists, all this disagreement is certainly business as usual.

And, contrary to the article, every one of these species would be thrown from the hominid line, if we believe the molecules. Here's the text from the article:

Genetic clues also set the approximate time of the divergence of the human lineage from a common ancestor with apes: between six million and eight million years ago.

Fossil researchers were skeptical at first, a reaction colored perhaps by their dismay at finding scientific poachers on their turf. These paleoanthropologists contended that the biologists' "molecular clocks" were unreliable, and in some cases they were, though apparently not to a significant degree.

...

The new finds have filled in some of the yawning gaps in the fossil record. They have doubled the record's time span from 3.5 million back almost to 7 million years ago and more than doubled the number of earliest known hominid species. The teeth and bone fragments suggest the form -- the morphology -- of these ancestors that lived presumably just this side of the human-ape split.

It is true that the new fossils date as far back as 7 million years; with Sahelanthropus just under that date, Orrorin at around 6 million, Ar. kadabba at 5.5, Ar. ramidus at 4.4, and Au. anamensis at around 4.1.

But it has been many years since a genetic comparison indicated a human-chimpanzee common ancestor as old as 6-8 million years. This year's study by Holbolth et al. (2007) estimated a human-chimpanzee speciation time of 4.1 +/- 0.4 million years. That makes Au. anamensis possibly too young to be a hominid. The rest of those species would presumably be just so many apes.

Now, I don't believe for a second that Au. anamensis is an ape and not a hominid. It just looks too much like Au. afarensis -- so much so that some would put them in the same species. The evolutionary transition between these two is well documented, and will be more so when some as-yet-unpublished fossils come out. So anything younger than 4.1 million years is almost certainly not right for the human-chimpanzee divergence.

But the 4.1 million year estimate is not unusual compared to other recent studies. My post from last May covers many of these recent studies, including last year's problematic "hominid-chimpanzee hybrid speciation" paper by Nick Patterson and colleagues. The conclusion in that paper about hybridization was certainly wrong, but the date of 5 million years was right in line with other estimates.

These genetic comparisons are not easily dismissed. Possibly there has been a rate deceleration of mutations in the human lineage that means that the estimated dates are too recent. Maybe 4.1 million years can be stretched into 6 million. Maybe it can even be stretched into 7 million. But all this stretching does have other effects -- on the estimated dates of earlier divergences -- and those are compounded by a large multiple of the few million years we may try to push the human-chimpanzee speciation date. That 4.1 million year estimate is calibrated from an African-Asian great ape divergence at 18 million years ago. Push the human-chimpanzee divergence to 7 million, and you push the orangutan-human divergence back into the Oligocene. Are silent sites in humans evolving more slowly than cercopithecines? Probably. Are they evolving that much slower than orangutans? I suppose nothing is impossible, but maybe we should take another look at those fossils.

All this is to point out that there really is a conflict between these Miocene "hominids" and genomic evidence about human-chimpanzee speciation time. I don't see any magic solution to this problem from the molecular side -- those dates keep coming up again and again from different regions, and from comparisons across many regions -- including estimates that are not calibrated by other fossil divergences. This is not an easy "the molecular clock must be wrong" kind of problem.

Nor are the fossils an easy problem. There is pretty good evidence for vertical posture or hindlimb-dominant movement in all of these "hominids." Up to now, we've accepted these kinds of features as de facto evidence of bipedality, and assumed that bipedality is such a unique character of hominids that it is unlikely to be any older. Yet few of these fossils provide really good evidence for obligate bipedality, and some of them provide none at all.

Is it possible that bipedal apes long preceded the divergence of humans and chimpanzees? Was the common ancestor of the two lineages a biped? Or was significant vertical posture a common feature of many Miocene apes -- making Sahelanthropus a possible homologue of Oreopithecus?

Which feature is the important one? The long nuchal plane of Sahelanthropus? The femur neck cortical bone distribution of Orrorin? The toe bone of Ar. kadabba? Heck, I can hardly convince my undergraduates about that toe bone!

I've talked to people about this. Some think that all the molecular stuff is just jibberjabbing, and any day now we will find out that the date estimates were wrong all along.

I think it may be time to start doubting our confidence again.

UPDATE (6/28/2007): I've gotten into rather an interesting e-mail discussion about whether I should have included Homo georgicus on the list of new species. Frankly it didn't occur to me: Wilford didn't mention it.

Actually if you start to think about all the new names that have been proposed in the last 15 years, it is a quite bushy list. It will be no surprise that I think this bushiness has more to do with the listers than the listees.

Anyway, there is something interesting about early Homo right now that goes beyond the simple splitter/lumper questions. I'll have more to say about it in a few days.

References:

Hobolth A, Christensen OF, Mailund T, Schierup MH. 2007. Genomic relationships and speciation times of human, chimpanzee, and gorilla inferred from a coalescent hidden Markov model. PLoS Genet 3:e7. doi:10.1371/journal.pgen.0030007

Patterson N, Richter DJ, Gnerre S, Lander ES, Reich D. 2006. Genetic evidence for complex speciation of humans and chimpanzees. Nature 441:1103-1108doi:10.1038/nature04789

What's behind the headline about "resurrecting an ancient virus"?

The study, which appears tomorrow in Science, focuses on Pan troglodytes endogenous retrovirus (PtERV1). More than 100 copies of inactive PtERV1 are sprinkled throughout the chimpanzee and gorilla genomes, whereas humans have none. "About 4 million years ago, this virus was active and independently infecting all these species, but not humans," says virologist Michael Emerman, who conducted the study with evolutionary biologist Harmit Malik and postgraduate student Shari Kaiser, all of whom work at the Fred Hutchinson Cancer Research Center in Seattle, Washington.

That's a pretty radical genome-wide difference between humans and the African apes. It's interesting that the chimpanzee and gorilla lineages were capable of exchanging viruses from early in their evolution. If they are right about the 4 Ma age, that is about halfway between the common ancestor of chimpanzees and gorillas (around 7--10 Ma) and today. Since they still exchange viruses today (including Ebola), you might think that it wouldn't be surprising that they did so in the Pliocene. But I think it's important because it establishes sympatry of the two apes: the ancestors of today's chimpanzees and gorillas lived in the same geographic region. If the two lineages originated in an allopatric speciation, then they had expanded their ranges by the Pliocene to overlap with each other. And sympatry means that they must have been adaptively differentiated by that time.

The now-extinct virus was interesting to this research group for another reason: apparently, a protein essential to HIV resistance in other primates lets humans down by being resistant to the extinct virus. This is a pretty tricky storyline, so I'll try to explain. The protein is named TRIM5α, it is key to the immune response to viruses. The human form of the protein does not do a good job of fighting HIV, and some other primates have a form that resists HIV infection must more effectively. So this protein has been a focus of HIV research.

When Kaiser and colleagues examined the sequences of the ancient retrovirus imprisoned in the chimpanzee genome, they found that one of the virus' genes interacts with TRIM5α. Prior work had established the variation among primates in TRIM5α's response to HIV. Now, they found that no known primate TRIM5α sequence provides an effective response to both HIV and the ancient chimpanzee virus.

In other words, this ancient chimpanzee virus, PtERV1, is like the immune system equivalent of Bizarro Superman -- it's just the opposite of the real thing.

In terms of hominid evolution, there is the obvious question of why humans never acquired the many endogenous copies of this viral genome. Clearly, ancient hominids did not have the virus in substantial enough numbers to result in its incorporation into our genome. But why not? In particular, did our TRIM5α sequence protect us?

Here's what Kaiser and colleagues have to say:

Although we cannot rule out the possibility that PtERV1 never infected human ancestors for other reasons (SOM Text, note 1), our data do suggest the possibility that TRIM5 was fixed in human populations because of its ability to confer protection against PtERV1 (Figs. 2 and 3) and that modern humans have descended from ancestors who resisted infection. Indeed, we know that there is very little diversity in the human population today in the part of TRIM5 that determines antiviral specificity (6, 16, 17). However, we find that chimpanzee TRIM5 is also capable of restricting PtERV1 and encodes an R332 (Fig. 3), yet chimpanzees contain multiple copies of PtERV1 in their genome and humans do not. Moreover, we find that R332 is monomorphic in the TRIM5 allele in all four subspecies of chimpanzees and in bonobos, which indicates that R332 is evolutionarily conserved through the chimpanzee radiation (in the past 1 to 2 million years). The most parsimonious explanation for the presence of R332 in humans and chimpanzees is that the mutation was fixed in our common ancestor, which presents a paradox because chimpanzee TRIM5 did not protect them against PtERV1. This suggests that TRIM5 alone does not determine retroviral invasion into the germline but that the combination of multiple retroviral restriction factors that are also rapidly evolving, such as the Apobec3 family (18), are necessary to explain ancient transmission events.

Well, no one can say for sure, but I think it seems pretty unlikely that the hominid TRIM5α gene is the result of selection against infection by this particular virus. The chimpanzee and gorilla genomes have more than 100 copies of the viral DNA in their genomes, which means that it was a long-term infectious agent in those lineages, and its genome was often incorporated into its hosts' germlines. It is probable that these viral genes evolved neutrally after being incorporated into the chimpanzee and gorilla genomes: if they were deleterious, they would be gone; if they were adaptive, they would probably be more highly conserved. Neither chimpanzees nor gorillas are known to carry the live virus today, which is therefore presumed to be extinct. If true, that means that both chimpanzees and gorillas lost the virus, probably sometime before 3 million years ago. This loss occurred either because of convergent genetic adaptations in both lineages (I say convergent because they may or may not have involved the same genes), or because of extreme bottlenecks during which the longtime viral parasites were lost by chance.

Now, consider some hypotheses:

Hypothesis 1: The virus was a longtime hominid pathogen, to which TRIM5α was an adaptation. If this were true, then the human genome ought to harbor at least some copies of the viral DNA. It has none. So this hypothesis probably isn't true.

Hypothesis 2: The virus was a severe short-term epidemic in ancient hominids, and we are descended from the only survivors, who happened to have a resistant TRIM5α allele. This is the hypothesis proposed in the quote above. If this were true, then we wouldn't necessarily expect to see copies of the viral genome in human DNA -- the infection and population crash may have happened too fast. But a single epidemic, or even a succession of several epidemics of the virus, would be unlikely to fix a variant allele. After all, neither the Black Death nor smallpox, nor any other historical epidemic has managed such a feat. And viruses with such exceptionally high death tolls do not tend to sustain epidemics through sparse populations like ancient hominids. Indeed, it seems likely that the virus' long survival in chimpanzees and gorillas implies that its hosts survived for a long time with the virus, and dispersed it as they encountered other individuals over time. If the virus let its hominid victims live a long time and thereby spread across low-density hominid populations, then there ought to be at least some copies of it in our genomes. And there aren't any. Also, the TRIM5α protein has a strong signature of positive selection in the human lineage, which means that there have been multiple selected substitutions. Multiple substitutions are very unlikely to have happened simultaneously; it is more likely that they occurred sequentially, taking a long time. So this hypothesis probably isn't true, either.

Hypothesis 3: The virus never infected hominids, who were, after all, allopatric from chimpanzees and gorillas. Instead, some other virus -- or more probably, several viruses -- infecting ancient hominids explain the evolution of the human TRIM5α gene.

I like hypothesis 3 the best; the data don't seem to reject it. Humans are not very susceptible to the PtERV1 virus. Indeed, our own TRIM5α variant, alone or with other genetic adaptations, have have helped to prevent the virus from infecting ancient hominids at all. Or maybe our ancestors never encountered the virus and our TRIM5α is a result of later events.

I should add, from a quick look at Sawyer et al. 2005, that chimpanzees and gorillas share at least one parallel amino acid subsitution in the rapidly-evolving SPRY domain of TRIM5α -- at position 340. The current paper (Kaiser et al. 2007) notes that humans and chimpanzees share a derived amino acid substitution at position 332. That position (332) is important because the biochemical work by Kaiser et al. 2007 and others has shown it is a critical site for human susceptibility to HIV:

For example, the amino acid at position 332 within this patch is a critical determinant of HIV-1 restriction (13). Humans and chimpanzees encode an arginine (R) residue at position 332, whereas the hominoid ancestral residue at this position is a glutamine (Q). Reversing this change (R332Q) had moderate effects on the ability of human TRIM5α to restrict MLV variants (fig. S2). Notably, changing the arginine to the ancestral glutamine abolished the ability of human TRIM5α to efficiently restrict PtERV1 infectivity (Fig. 2B). Unexpectedly, the R332Q mutation had the opposite effect on HIV-1, improving the ability of human TRIM5α to restrict this virus (15) (Fig. 2B). Thus, the R332Q mutation in human TRIM5α reveals a trade-off in TRIM5α's ability to restrict two retroviruses; a mutation that abolished restriction for PtERV1 results in a gain of restriction to other viruses such as HIV-1.

It would be interesting to see what difference the chimpanzee and gorilla alleles at position 340 make; perhaps their TRIM5α protein remains constrained by selection from yet another pathogen?

In any case, during the last few years we have learned a great deal about ancient selection in hominids associated with pathogens. And most of what has come out has been gross changes such as pseudogenization. In this instance, the key information comes from a genomic comparison showing the huge importance of an ancient pathogen in both chimpanzees and gorillas, but not humans.

It is hard to overstate just how glaring these comparisons are -- they are not at all subtle. If most genetic changes during human evolution have been like precision-aimed shots from a sniper rifle, these disease-associated adaptations we've been finding are like blasts from a cannon.

So there is a lot left to discover.

References:

Kaiser SI, Malik HS, Emerman M. 2007. Restriction of an extinct retrovirus by the human TRIM5α antiviral protein. Science 316:1756-1758. doi:10.1126/science.1140579

Sawyer SL, Wu LI, Emerman M, Malik HS. 2005. Positive selection of primate TRIM5α identifies a critical species-specific retroviral restriction domain. Proc Nat Acad Sci USA 102:2832-2837. doi:10.1073/pnas.0409853102

Well, I guess they've got a plot for the pilot of that caveman show:

Humans caught pubic lice, aka "the crabs," from gorillas roughly three million years ago, scientists now report.

Rather than close encounters of the intimate kind, researchers explained humans most likely got the lice, which most commonly live in pubic hair, from sleeping in gorilla nests or eating the apes.

"Sleeping in gorilla nests." Yep, that's the ticket.

The quote is from a LiveScience article by Charles Q. Choi. The article talks a lot about "monkey business" but really spends more time on the hominid-eating-gorilla scenario:

"Unfortunately, even today among modern humans there's a bush meat trade where gorillas are killed for their meat," he said. "If archaic humans were butchering or scavenging those animals 3.3 million years ago, it would be a simple thing to transfer those lice from prey to predator."

UPDATE (3/7/2006): Carl Zimmer's post is great (he's all about the parasites) and mirrors some of what I wrote below. He also includes probably the best snarky quote: "Is this evidence of a Pliocene love that dare not speak its name?"

To telegraph my conclusion a bit, I still think the flawed assumption is that the hominid-gorilla interaction occurred when the hominid and gorilla Pthirus diverged. The interaction works a lot better later, assuming within-gorilla parasite variation. Since there is a lot of within-human variation in the other louse genus, Pediculus, the idea of a couple million years of delay between louse genetic divergence and lateral transfer is not at all unlikely, even without invoking ancient gorilla speciations.

Thoughts

I downloaded the research paper in BMC Biology by Reed and colleagues. Here's the 'Conclusions' section of the abstract:

Reconciliation analysis determines that there are two alternative explanations that account for the current distribution of anthropoid primate lice. The more parsimonious of the two solutions suggests that a Pthirus species switched from gorillas to humans. This analysis assumes that the divergence between Pediculus and Pthirus was contemporaneous with the split (i.e., a node of cospeciation) between gorillas and the lineage leading to chimpanzees and humans. Divergence date estimates, however, show that the nodes in the host and parasite trees are not contemporaneous. Rather, the shared coevolutionary history of the anthropoid primates and their lice contains a mixture of evolutionary events including cospeciation, parasite duplication, parasite extinction, and host switching. Based on these data, the coevolutionary history of primates and their lice has been anything but parsimonious.

There is actually a much more interesting story here than is indicated in either press account or abstract. The genera Pediculus and Pthirus were thought to have diverged at the time that gorillas diverged from the chimpanzee-human clade. This would explain why gorillas have Pthirus and chimpanzees Pediculus. The fact that humans have both ... well, that remained unexplained. The purpose of the study was to test whether humans retained the two genera ancestrally, or if instead they picked one up later.

What they found is that the two genera didn't diverge at the gorilla-chuman split, but instead way earlier. Their estimate for the Pediculus-Pthirus divergence is 13 million years. Thirteen million is as much as twice the age of the human-gorilla common ancestor. This estimate is probably biased toward the recent side, since it is calibrated against a divergence between hominoid and baboon lice assumed at 22.5 million years ago -- probably more recent than the true hominoid-baboon divergence.

The paper considers it likely that the human-gorilla-chimpanzee common ancestor lineage maintained this pair of lice species for the intervening time period, with one genus being lost in each of the two (gorilla and chuman) descendant clades. This ancestral lineage would be similar to humans in that respect -- host to two distinct parasite lineages, both of which stemmed from a single ancestor species.

But much later than the chimpanzee-human divergence, humans apparently picked up the gorilla lice somehow. The paper doesn't belabor this point or attempt to explain it, beyond this:

Evidence suggests that Pthirus pubis has been associated with humans for several million years, and likely arrived on humans via a host switch from gorillas. Despite the fact that human pubic lice are primarily transmitted via sexual contact, such contact is not required to explain the host switch. Parasites often switch from a given species to a predator of that species [17], and are sometimes found to switch to unrelated hosts in communally used areas, such as roosting or nesting sites [18]. The host switch in question could have resulted from any form of contact between archaic humans and gorillas including, but not limited to, feeding on or living among gorillas. Regardless of how the transfer occurred, suitable habitat had to be available on the new human host for the host switch to be successful. For example, it is possible that the switch of Pthirus from gorillas to humans coincides with a change in available niche space in humans, such as the loss of body hair. Further study, however, is required to test such a hypothesis (Reed et al. 2007:7).

Hominids were certainly not hunting gorillas 3.3 million years ago. At least, not the hominids we know about. That date is a bit older than Lucy; it's 700,000 years older than the earliest evidence of flaked stone and 800,000 earlier than the earliest evidence of antelope butchery. Hominids weren't hunting gorillas because they weren't hunting any large mammal species then.

What's worse, gorillas and hominids weren't sympatric 3.3 million years ago. At least, not the gorillas and hominids we know about. Unless gorillas ranged into open woodland, and in particular the East African coastal forest, or hominids ranged into the central or west African rain forest, they never came into contact with each other at all.

If anything, we might expect that gorillas and chimpanzees would have been likely to come into contact and exchange parasites. They are currently sympatric, they eat the same foods, and they even build similar sorts of nests. It's like they share the same locker room. But they didn't have this parasite exchange.

It's all very strange. First we have this long period of divergence of the two great ape louse genera (orangutans don't have their own louse species). Then we have a divergence of the human and chimpanzee Pediculus species just exactly when it should have happened. And then there is this lateral transfer of lice from gorillas to humans 3 million years ago - when hominids and gorillas weren't apparently sympatric and had no credible mechanism for lice exchange.

Here's my hypothesis: cryptic African hominoids. The apparent craziness all comes from the assumption that the only species that existed are the ones we know about. For Africa 3 million years ago, that means two or three hominid species, one gorilla lineage and one chimpanzee lineage. We don't have any fossils that old for the apes; we can only infer their existence from the fact that they exist now.

Let's consider what we know. We know that 3 million years ago there weren't any chimpanzee or gorilla relatives in the Rift Valley, and plausibly (but not definitely) not in South Africa or the Sahel.

We don't know how extensively hominids ranged into the west African or central African forests, particularly from the north and southeast. We don't know how extensively gorillas and/or chimpanzees may have ranged outside the core forested areas where they have historically existed. In the absence of Homo, the competition between these apes and hominids at the forest boundaries may have been a close game.

We don't know how many species of ancient chimpanzees and gorillas there may have been. The present subspecific variation of chimpanzees seems to reflect recent colonization of the eastern range from central Africa, and some substantial population interchange between central and western ranges. Gorilla subspecies now seem to have emerged within the same time frame, with a possible colonization from their western range into their eastern range within the past million years.

Bonobos are only ca. 850,000 years old (Won and Hey 2005). To summarize, the current eastern chimpanzees weren't in East Africa half a million years ago, and the bonobos weren't south of the Congo a million years ago, and eastern gorillas weren't there a million years ago either.

Who was? It seems to me that the best candidates would be ancient species of gorillas and chimpanzees that no longer exist. A second-best (and maybe more interesting) candidate is some variety of hominid. A third-best (and even more interesting) candidate is an ancient ape lineage dating from before the G-C-H divergence.

Three million years ago, any one of those possibilities is credible. Here's my favorite: two gorilla species (or subspecies) became isolated enough for louse divergence 3.3 million years ago, and continued to coexist. Sometime after 2 million years ago, Homo encountered one of these species and picked up its lice. That gorilla lineage later became extinct, perhaps by range expansion from Homo.

Oh, and the long divergence time between the two lice genera? I like a long divergence and later lateral transfer from some pre-H-C-G Miocene ape lineage. There were likely several in Africa to choose from. Maybe it was Sahelanthropus...

References:

Reed DL, Light JE, Allen JM, Kirchman JJ. 2007. Pair of lice lost or parasites regained: the evolutionary history of anthropoid primate lice. BMC Biol 5:7. doi:10.1186/1741-7007-5-7

Won Y-J, Hey J. 2005. Divergence population genetics of chimpanzees. Mol Biol Evol 22:297-307.

The current Science has a paper by Eric Bazin and colleagues comparing mtDNA diversity with population size, history and ecology of 3000 animal species.

Here's the conclusion:

This study reveals that the mitochondrial diversity of a given animal species does not reflect its population size: No correlation between mtDNA polymorphism and species abundance could be detected, despite the large body of data analyzed. Nuclear data, in contrast, are fairly consistent with intuitive expectations. We conclude that natural selection acting on mtDNA contributes to homogenization of the average diversity among groups, in agreement with the genetic draft theory. mtDNA appears to be anything but a neutral marker and probably undergoes frequent adaptive evolution, e.g., direct selection on the respiratory machinery, nucleo-cytoplasmic coadaptation, two-level selection, or adaptive introgression, perhaps hitchhiking with a maternally transmitted parasite. mtDNA diversity is essentially unpredictable and will, in many instances, reflect the time since the last event of selective sweep, rather than population history and demography. Low-diversity mitochondrial lineages, typically disregarded as important from a conservation standpoint, might sometimes correspond to recently selected, well-adapted haplotypes to be preserved (Bazin et al. 2006:571-572, emphasis added).

This is a nice empirical comparison, and a very impressive exercise in data mining. To accumulate the dataset, they had to troll large data depositories for cases in which the same DNA segments had been sequenced in multiple individuals of single species, and then had to match those cases with ecological information about the species, as the accompanying perspective by Adam Eyre-Walker describes.

But, aside from the very persuasive presentation here, the fact has been obvious for years. I blogged about mtDNA selection last year. Finding such widespread mtDNA selection across taxa -- even into invertebrates -- is certainly strong support for the idea that it evolved adaptively in humans. And finding that the chance of adaptive evolution in mtDNA is proportional to population size enhances the likelihood of recent mtDNA selection in humans even more.

Eyre-Walker draws exactly the opposite conclusion than I do:

Interestingly, humans are an exception to the pattern seen by Bazin et al. If the authors are correct, then the effective population size estimated from mitochondrial DNA should be lower than that estimated from autosomal DNA. This is not what we see in humans; the effective population sizes estimated from autosomal DNA, Y-chromosome DNA, and mitochondrial DNA are all approximately 10,000. Does this mean that Bazin et al. are incorrect? Probably not. It may be that humans have such small effective population sizes that adaptive evolution in the mitochondrial genome is very rare; the neutrality index in human mitochondrial DNA, and perhaps nuclear DNA, certainly gives no indication of adaptive evolution. (Eyre-Walker 2006:538).

But of course, this is quite backward -- low mtDNA diversity cannot be evidence for neutrality; at best it can fail to refute a hypothesis of selection. With our long generation lengths, autosomal DNA would have to have coalescence dates in the Pliocene to make the low mtDNA diversity stand out statistically. It is not a question of them all being neutral, it is a question of packing most of human evolution into a space of 2 million years.

Supporting that is the observation that Eyre-Walker points out next:

Although nuclear diversity follows the expected pattern, with more diversity in organisms that are expected to have bigger population sizes, the differences are remarkably small; synonymous diversity varies by less than a factor of 10, and allozyme diversity by less than a factor of 4. This is striking given that the population sizes of marsupials and mussels, for example, must differ by many orders of magnitude, and one would expect diversity to be linearly related to population size. This observation is not new for allozyme data (4), but it is the first time this pattern has been so clearly illustrated for synonymous diversity in nuclear genes. The lack of a strong correlation between diversity and population size in nuclear DNA may also reflect the effects of genetic hitchhiking (ibid.).

In other words, selection has restricted mtDNA diversity, and it has also restricted nuclear DNA diversity -- just not as much. The "not as much" here is a function of recombination, which makes the nuclear genes true subjects of genetic draft.

This isn't news either. We've known about the restricted allozyme diversity since 1984. A few voices crying in the wilderness have been reminding us from time to time, like Gillespie.

I would note that some of their ecological substitutes for population sizes may themselves induce selective effects. For example, Bazin and colleagues note that marine molluscs have more allozyme variation than terrestrial molluscs, which they view as consistent with the greater dispersal of marine species. But greater dispersal might also involve the necessity to maintain diversity for dispersing into to different local environments, which would tend to drive frequency-dependent or balancing selection for traits responding to these local forces.

So there is more to be done on the nuclear DNA side of this equation, probably much more. But the mtDNA comparison is very important, and hopefully will drive some reevaluation of the use of mtDNA diversity as a proxy for genetic diversity in conservation and ecological studies.

References:

Bazin E, Glémin S, Galtier N. 2006. Population size does not influence mitochondrial genetic diversity in animals. Science 312:570-572. DOI link

Eyre-Walker A. 2006. Size does not matter for mitochondrial DNA. Science 312:537-538. DOI link

Rex Dalton has a great two-page article in Nature about the bush vs. ladder dispute. It keys off of the Middle Awash Australopithecus anamensis article by White and colleagues from a couple of weeks ago.

If you recall that one, White et al. posited that Ardipithecus was likely ancestral to Au. anamensis, and that the two did not overlap in time. Here's the key exchange in the Dalton piece:

This month's Nature paper makes a bold argument, and shows the Awash team seeking to put its mark on the record. Others in the
field are impressed. "When you find 30 new hominid fossils, you are allowed a certain amount of conjecture," says Bernard Wood, a palaeoanthropologist at George Washington University in Washington DC. "As always, they have done a fantastic job."

But he and others are unconvinced by the Awash team's conclusion: "This is only the first half of the rugby match," says Wood. Meave Leakey, lead author on the Au. anamensis discoveries in Kenya, is more blunt. "I don't believe this," she says. "We do not have the specimens to fill the gaps."

Leakey and Wood are among those who believe that other, as yet undiscovered hominid species may have lived at this time, from 4.4 million to 2.9 million years ago. The existence of other species would cloud or eliminate the argument for a direct lineage. "My prejudice is there are more lineages rather than fewer -- more diversity," says Wood. "I have to concede these new data are dramatic. But we should beware coming out with a complete explanation when we don't have all the
evidence."

This argument frustrates White. "There were Martians there back then too," he says. "And spacecraft all over the Pliocene -- we just haven't found them yet."

Waiting for Monte Cassino

In a series of articles since 2000, White and colleagues have laid out a systematic attack on the "bushy" phylogeny model. Their arguments have extended across four million years and seven species, with a breadth that rivals the Allies breaking the Winter Line.

Consider the angles of attack:

1. Au. anamensis -- Au. afarensis. Everyone basically accepts that Au. anamensis is a direct ancestor of Au. afarensis. And the two species are really not very different from each other -- for instance, they are more alike than either is to Ardipithecus. The transition between these species would look to be a simple case of anagenesis, except...

...for Kenyanthropus (Leakey et al. 2001). This small-toothed, flat faced hominid needs an ancestor, too. Au. anamensis might have been the common ancestor of Kenyanthropus and Au. afarensis. If so, then both these later species originated by cladogenesis from Au. anamensis. A similar argument might be made for other species, like Australopithecus bahrelghazali (Brunet et al. 1996) or the Sterkfontein Member 2 hominids. But Au. bahrelghazali is only known from a partial mandible and only differs from Au. afarensis by a three-rooted premolar, which is considered by many to be weak evidence, and the Sterkfontein Member 2 sample has not yet been taxonomically assigned -- they might turn out to be Au. afarensis, for example. Kenyanthropus remains the strongest case for cladogenesis (i.e., a "bush"). Yet...

...White (2003) denied that the Lomekwi skull KNM-WT 40000 was a distinct species. In particular, he argued that the extensive postmortem deformation of the skull made it impossible to substantiate an anatomical difference from Au. afarensis, and even if it was different, the anatomical diversity of living hominoid species is so great that it would probably encompass the difference between KNM-WT 40000 and known Au. afarensis crania.

2. Earliest hominids. At the moment, the earliest putative hominids include three genera: Orrorin (Senut et al. 2000), Sahelanthropus (Brunet et al. 2002), and Ardipithecus, represented in the Late Miocene by Ar. kadabba (Haile-Selassie 2001, Haile-Selassie et al. 2004). Evidence for obligate bipedality has been challenged (by different researchers) for each of these three (I'm one of those who has questioned bipedality for Sahelanthropus).

So far the only comparable anatomical parts from all three samples are teeth...

...which were examined by Haile-Selassie, Suwa and White (2004). They concluded that the variation among these three genera

is no greater in degree than that seen within extant ape genera. Despite claims of molar enamel thickness differences among these late Miocene fossils, we question the interpretation that these taxa represent three separate genera or even lineages. Given the limited data currently available, it is possible that all of these remains represent specific or subspecific variation within a single genus (Haile-Selassie et al. 2004:1505).

Additionally, Ohman, Lovejoy and White (2005) challenged the interpretation of the internal anatomy of the Orrorin femur, which had been suggested to be more derived than that of Au. afarensis. They wrote:

We agree that the Lukeino femur's external morphology suggests some form of bipedality. Yet the more detailed original scans appear to show a distinct superior cortex different from Australopithecus and humans, with the cortex distribution being more primitive than that seen in any other hominid, including Australopithecus.

The relevance of this argument to the phylogenetic diversity of early hominids depends on the anatomy of the Ardipithecus femur, which none of the rest of us are in a position to know. But one may speculate that if all these early "hominids" had femora with similar morphology, it would further reinforce the interpretation that they belong to a single lineage.

3. Ardipithecus -- Au. anamensis. This is the current example. Here's how Dalton discusses it:

The latest Afar discovery is exciting experts because it shows that the three hominids existing in the same area, but in successive time periods. Tim White of the University of California, Berkeley, co-leader of the Awash team, believes this points to a direct lineage between the three -- a process called phyletic evolution. The new Au. anamensis fossils are only 300,000 years younger than Ar. ramidus, meaning that if one became the other, the changes would have had to happen that fast. But the key point, says White, is that fossils of Au. anamensis and Au. afarensis have never been found in sediments the same age as those containing Ar. ramidus. If fossils of the different species were found together, that could show that they belonged to multiple lineages existing simultaneously.

Finding remains of all three species in the same area but not from the same time period suggests they did not coexist, says White.

...

The specimens also provide anatomical clues to evolutionary history. "The new Au. anamensis fossils are anatomically intermediate between the earlier Ar. ramidus and the later Au. afarensis," says White. For example, the teeth of the newly discovered Au. anamensis fossils seem adapted to chew tougher and more abrasive foods than Ar. ramidus. The researchers believe this shows that Au. anamensis had a broader diet. "All this strengthens the view that there is phyletic evolution from Ar. ramidus through Au. anamensis," says White. He believes he has nailed down the relationship between the two later species, although he says that further specimens are needed to prove the earlier link (Dalton 2006:1100).

Of course, it would help matters if we knew in more detail what Ardipithecus looked like. But one must imagine that the stage is being set for its revelation. The unilineal interpretation places Ardipithecus at the critical point as an ancestor to the major mid-Pliocene australopithecine lineage. Extending the unilineal interpretation earlier into the Late Miocene would make Ardipithecus the earliest hominid as well.

It is not necessary to think that taxonomic uniformity means anatomical uniformity, though. Ardipithecus already encompasses a trend of decreasing canine size and less sectorial P₃ for example. A trend toward fuller skeletal adaptation to bipedality may also be imagined. But in that context, it is important to note that the time interval between the Orrorin femur and the unpublished Aramis skeleton is longer than the time between Aramis and Hadar. Those relative times may become quite important in thinking about the evolution of those postcrania.

The Winter Line was broken at Monte Cassino, after many failed attempts from different approaches. The Aramis fossils are either the heavy shoe waiting to drop, or they are the uncomfortable foot that all this talk about phyletic evolution is meant to shoehorn into place.

Commentary

If all these cases are added together, they imply a single evolving lineage encompassing at least four anagenetic taxa, Ar. kadabba -- Ar. ramidus -- Au. anamensis -- Au. afarensis. This last would presumably be followed by a cladogenesis into a robust australopithecine species (Australopithecus aethiopicus) and Australopithecus africanus.

One could add Homo erectus to this list, since White and colleagues argued in their description of the Daka skull (Asfaw et al. 2002) that the Asian and African samples represent one cosmopolitan species.

But then one species sticks out as a surprising exception to the pattern: Australopithecus garhi (Asfaw et al. 1999). It will be interesting to see a close argument showing why this species is really different from South African Au. africanus. Say, more different than KNM-WT 40000 is from the Hadar crania. It's quite glaring, really, that this species should be there mucking up such a simple phylogeny.

I have to say, after reviewing all these papers in one sitting -- this entire bush vs. ladder thing is getting very tiresome! I mean, isn't there something else that we could organize early hominid discoveries by? These are all papers in the top journals, and this is the (fairly specialized) discussion that has been promoted as the central issue in the field!

The subtitle of the Dalton piece suggests that it is merely a philosophical difference:

Deciding whether our ancestors evolved as a single lineage may depend more on philosophy than fossils.

But that's not really true. There is a clear null hypothesis here, quite directly drawn from William of Ockham:

entia non sunt multiplicanda praeter necessitatem

Which of course means:

Sometimes fossil samples really do form ancestor-descendant relationships.*

(*) It doesn't really. It means "Entities should not be multiplied beyond necessity."

References:

Asfaw B, Gilbert WH, Beyene Y, Hart WK, Renne PR, WoldeGabriel G, Vrba ES, White TD. 2002. Remains of Homo erectus from Bouri, Middle Awash, Ethiopia. Nature 416:317-320. DOI link

Asfaw B, White T, Lovejoy O, Latimer B, Simpson S, Suwa G. 1999. Australopithecus garhi: A new species of early hominid from Ethiopia. Science 284:629-635. DOI link

Begun DR. 2004. The earliest hominins -- is less more? Science 202:1478-1480. DOI link

Brunet M. and 37 others. 2002. A new hominid from the Upper Miocene of Chad, Central Africa. Nature 418:145-151. DOI link

Brunet M, Beauvillain A, Coppens Y, Heintz E, Moutaye AHE, Pilbeam D. 1995. The first australopithecine 2,500 kilometres west of the Rift Valley (Chad). Nature 378:273-275. DOI link

Dalton R. 2006. Feel it in your bones. Nature 440:1100-1101. DOI link

Haile-Selassie Y. 2001. Late Miocene hominids from the Middle Awash, Ethiopia. Nature 412:178-181. DOI link

Haile-Selassie Y, Suwa G, White TD. 2004. Late Miocene teeth from Middle Awash, Ethiopia, and early hominid dental evolution. Science 303:1503-1505. DOI link

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. DOI link

Ohman JC, Lovejoy CO, White TD. 2005. Questions about the Orrorin femur. Science 307:845. DOI link

Senut B, Pickford M, Gommery D, Mein P, Cheboi K, Coppens Y. 2001. First hominid from the Miocene (Lukeino formation, Kenya). Comptes Rendus 332:137-144.

White T. 2003. Early hominids -- diversity or distortion? Science 299:1994-1996. DOI link

Tim White and colleagues (2006) report on new fossils from Aramis and a new site, Asa Issie, with estimated dates between 4.1 and 4.2 million years ago.

In addition to the paper, there are articles in the New York Times (by John Noble Wilford), the Associated Press (by Seth Borenstein), and BBC (by Paul Rincon).

The story is being played as another "missing link" -- this one between Ardipithecus and Australopithecus. From the Times:

Tim D. White, a paleontologist at the University of California, Berkeley, who was a leader of the team, and his colleagues said the 4.1-million-year-old fossils were anatomically intermediate between the earlier species Ardipithecus ramidus and the later species Australopithecus afarensis, the Lucy family. The newfound bones and teeth are the earliest remains of the most primitive Australopithecus, known as anamensis.

"This new discovery closes the gap between the fully blown australopithecines and earlier forms we call Ardipithecus," Dr. White said in a statement. "We now know where Australopithecus came from before four million years ago."

The fossil specimens are a partial maxilla from Aramis, ARA-VP-14/1; two partial maxillary dentitions from Asa Issie numbered ASI-VP-2/2 and ASI-VP-2/334; and a large femur shaft fragment, ASI-VP-5/154. There are also several postcranial bones -- phalanges, vertebrae, a metatarsal -- that are pictured in some of the press accounts and briefly discussed but not pictured or numbered in the paper. The postcanine teeth in the maxillary specimens are larger than the known sample of Ardipithecus, but the canines are larger and more mesiodistally elongated than in Australopithecus afarensis. The best anatomical match for these features is with the Kanapoi and Allia Bay samples assigned to Australopithecus anamensis, and White and colleagues assign the new fossils to that species.

So why are these fossils important? On the surface, there isn't very much to them. Three piecemeal upper dentitions don't tell much. They have big molars and big canines, both within the range of Au. anamensis. Neither they nor the femur shaft extend the known range of variation in early hominids.

Remembering that every fossil fragment is a precious relic of a bygone age, the main importance of these is that they may address hypotheses about the biogeography of Early Pliocene hominids. The maxillae show that a large-molared hominid existed in the same geographic location at a later time than the small-molared Ardipithecus. That could be interesting, and it is the hook for the news stories and the team's press statement.

The strongest part of this story is the geographic -- finding them in the Middle Awash instead of Kenya -- and the paleoenvironmental. There is some suggestion in the paper that there may be a paleoenvironmental difference at the sites that currently have evidence of Au. anamensis:

Palaeoenvironmental circumstances surrounding Au. anamensis ~1,000 km to the south in Kenya have been described for Allia Bay as a mixed assemblage sampling aquatic, forest, grassland and bushland. Nearby Kanapoi conspecifics were found in another mix of environments described as dry, possibly open, wooded, or bushland conditions with a wide gallery forest in the vicinity. Habitat preferences in such mixed assemblages are difficult to ascertain despite the assertion that the hominids "favored mosaic settings". In contrast, the Ethiopian occurrence of Au. anamensis described here allows its tight spatial and temporal placement in a vertebrate assemblage with taphonomic integrity. Its relative abundance suggests that it was a regular occupant of a wooded biome that appears to have persisted in this part of the Afar during the 200,000-yr interval subsequent to Ar. ramidus at Aramis (White et al. 2006:887-888).

This points to two salient facts about the Australopithecus lineage: they were able to disperse effectively across relatively long distances, and occupy at least those habitats where wooded cover and resources were available.

On the other hand, the fossils don't really "fill a gap" between Ardipithecus and Australopithecus, because they are pretty firmly within the time range of known Au. anamensis, being around the same age as the Au. anamensis sample from the Lake Turkana area -- the oldest Kanapoi hominids may be between 4.1 and 4.2 million years old also. The paper points out the other East African examples of Australopithecus at or above 4 million years ago; but it omits the Sterkfontein Member 2 remains, which are also conceivably in the age range of Au. anamensis. Or, for that matter, the Lothagam mandible, which might be the earliest australopithecine even if its date weren't as high as the >5 Ma estimate.

The paper attempts to close off -- for the moment -- the idea that there were allopatric species of early (ca. 4 Ma) australopithecines with differing dietary adaptations. But the paper cannot reject this hypothesis without caveats:

Two phylogenetic hypotheses concerning the origin of Australopithecus can be offered to account for the available data. The first hypothesis derives Au. anamensis phyletically from Ar. ramidus within a 200,000-yr interval [i.e., between 4.4 and 4.2 Ma]. The second involves cladogenesis of Au. anamensis from an ancestor (presumably Ardipithecus or some close relative) even deeper in the Pliocene or Late Miocene. Under the latter hypothesis, Ar. ramidus would represent a relict species in an ecological refugium (White et al. 2006:888).

This latter alternative is the only "bushy" interpretation -- the idea that known species of Ardipithecus can't really be the direct ancestors of Australopithecus, but that there must be some as-yet-undiscovered hominid (or better yet, hominids) that are the common ancestors, cousins, and other bushy relatives of the known species. White and colleagues cannot reject it, but they clearly do not favor it.

In its place, they suggest Ardipithecus ramidus as a lineal, possibly anagenetic ancestor of Au. anamensis, and Au. anamensis as the anagenetic ancestor of Au. afarensis. It's a ladder from primitive to derived, small-molared to big-molared, big-canined to small-canined.

I tend to think this is the null hypothesis -- we have sampled adaptations that differ because of evolution in what is essentially a single lineage of successive species. I say "essentially" because there was not necessarily a wholesale transformation of one species to another across its entire range. Instead, dispersals of new adaptive packages by population movements were probably important biogeographic aspects of evolution in these early hominids. But I think it important to recognize that one species can indeed be the ancestor of a later species.

People who like their phylogenies bushy and their speciations punctuated can take solace in that 200,000-year gap. The finding of Au. anamensis within the already-known time range of Au. anamensis means that the new fossils haven't really added much to the question of phylogenetic diversity in early hominids.

As a postscript, I have a nomination for "most significant sentence" in the paper:

At Aramis, the lone hominoid and largest primate was Ar. ramidus (109 of 6,156 identified specimens so far) (White et al. 2006:888, emphasis added).

References:

White TD and 21 others. 2006. Asa Issie, Aramis and the origin of Australopithecus. Nature 440:883-889. DOI link

A concise 4-paragraph article by Mathieu Schuster and colleagues reports on dune deposits that show the Sahara formed during the Late Miocene.

After the mid-Holocene humid period (6000 years ago), arid conditions developed throughout North Africa, culminating in the formation of the Sahara, which is the largest warm-climate desert on Earth (9,000,000 km2). However, earlier desert recurrences in the region are also documented. Direct evidence for eolian deposition is given by thermoluminescence dating for the Late Pleistocene; e.g., in Mauritania [25 to 15 thousand years ago (ka)] (1) or in Tunisia (86 ka) (2). The latter is currently considered as the oldest terrestrial record for desert conditions in the Sahara (2), even if firm evidence exists for a pre-Quaternary Great Western Sand Sea in Algeria (3). Some earlier arid episodes (Miocene-Pliocene) were also suggested by marine records off West Africa (4); but until now, no contemporary in situ eolian deposits were known in the Sahara region. In the northern Chad Basin, we recently identified and dated widespread outcrops of eolian dune deposits that are distributed over an area more than 2000 km². Our results testify that the onset of recurrent desert conditions in the Sahara started at least 7 million years ago (5-7) (Schuster et al. 2006:821).

The desert comes and goes, expanding and contracting -- and those vacillations are recorded by this earliest evidence, also:

In the Toros Menalla region, these eolian sandstones are conformably overlain by a horizon bearing abundant vertebrates fossils, including Sahelanthropus tchadensis, the earliest known Hominid [sic] (5, 7). In this horizon, named the Anthracotheriid Unit, biostratigraphic correlation of the mammalian fauna indicates an age of 7 Ma (5–7).

Now, this isn't news (which I'm sure Science didn't bother to check) since Vignaud and colleagues (2002) published the same evidence, complete with the wind direction chart:

The lower part of the section (at least 4 m thick) is composed of fine to very fine white sands, poorly cemented, and is mainly constituted by numerous quartz grains, without matrix. The grains are well sorted, well rounded, matt and frosted, and are strong evidence for aeolian modelling. The foreset laminations (avalanche laminations in front of the aeolian dune) represent a typically aeolian deposit. These sands show cross-beddings that progressively decrease in size from the bottom (1 - 2 m) to the top (20 cm). This facies exhibits typical alternations of grain-fall and grain-flow laminations, characteristic of aeolian dune deposits. Our interpretation is confirmed by frequent wind ripples at the foot of the fossil dunes, whose crests are perpendicular to the direction of dune progradation. These fossil dunes are, to our knowledge, the oldest evidence for desert conditions in the southern Sahara area (Vignaud et al. 2002:152).

I guess this is the science journal equivalent of getting "punk'd" -- "Ha ha! You published what we printed four years ago!"

I opened up the Vignaud paper to double-check the paleoenvironment in the fossil-bearing layer. From the faunal list, they conclude this:

The oldest known East African hominids (Ororrin [sic], Ardipithecus) are contemporary with faunas associated with wooded environments. Younger australopithecines lived in a wider range of habitats. In contrast, the TM 266 vertebrate fauna contemporary of the Toros-Menalla hominid suggests a mosaic of environments from gallery forest at the edge of a lake area to a dominance of large savannah and grassland. Determining the precise habitat of the TM 266 hominid locality among the mosaic of environments available to it constitutes a research challenge to be met by further laboratory and field studies currently in progress (Vignaud et al. 2002:155).

They (Vignaud et al. 2002) interpreted the succession of dune and lacustrine deposits to mean that the hominids lived in a mosaic environment near sandy desert, but locally including marshy/swampy, lake, and gallery forest. An alternative interpretation might be that the desert really receded (or disappeared) during the later time period when the hominids were there. In either case, the paleoenvironment is interesting, because it means that the Sahelanthropus-like primates colonized (and possibly repeatedly recolonized) areas that were periodically dune desert (and therefore probably not habitable by large primates). This may not mean much in terms of locomotion -- the hominid-bearing unit is clearly water-rich, and we can't refute the idea that the surroundings were as woodland-like as those preserved in the Late Miocene Middle Awash localities.

But I think it is a good hypothesis that all of these apes (or hominids) were very cosmopolitan compared to extant chimpanzees and gorillas. The question is whether their actual dispersal abilities were different from chimpanzees. Prehistorically, genetics would seem to indicate that chimpanzees had long-distance dispersal; the only fossil evidence of chimpanzees has been found in a region that historically did not support chimpanzees; and they today successfully utilize relatively open savanna at the eastern end of their range.

So it is by no means obvious that the cosmopolitan nature of these Late Miocene lineages would have required a specialized terrestrial adaptation -- at least not beyond the specialization of knuckle-walking. So why become bipeds?

References:

Schuster M et al. 2006. The age of the Sahara Desert. Science 311:821. Full text (subscription)

Vignaud P et al. 2002. Geology and paleontology of the Upper Miocene Toros-Menalla hominid locality, Chad. Full text (subscription)

An upcoming paper in Journal of Human Evolution by O. F. Huffman and colleagues reports on a possible location for the Mojokerto skull. A 1994 paper by Carl Swisher and colleagues dated rock from the supposed site to 1.81 million years ago.

This paper finds that the real site is a bit above that dated horizon. The abstract:

The fossil calvaria known as the Mojokerto child's skull was discovered in 1936, but uncertainties have persisted about its paleoenvironmental context and geological age because of difficulties in relocating the discovery site. Past relocation efforts were hindered by inaccuracies in old base maps, intensive post-1930s agricultural terracing, and new tree and brush growth. Fortunately geologic cross sections and site photographs from 1936-1938 -- not fully utilized in past relocation fieldwork -- closely circumscribe site geography and geology. These documents match the conditions at just one sandstone outcrop. It is situated on the southern margin of a topographic nose at the upper end of a 18 m-wide gully (0663760 m E, 9183430 m N, UTM Zone 49M), 15 m southeast of the Kumai et al. (1985) relocation. The relocated discovery bed is 3.3 m of fossiliferous pebbly sandstone, a river-channel deposit cut into tuffaceous mudstone. The sandstone and mudstone beds correspond to original site descriptions. Pebbly sandstone is also found within the skull.

The calvaria is well-preserved and taphonomically similar to large and fragile specimens found among several hundred vertebrate fossils excavated from the sandstone in 2001-2002. Since no well-preserved fossils were found intact at the surface of the sandstone, the good condition of the Mojokerto skull suggests that it was buried fully when discovered. The relocated hominin bed is the uppermost fluvial sandstone of a marine-deltaic sequence in the upper Pucangan Formation. The Mojokerto child probably died along the ancient seacoast, judging from the large extent of the deltaic facies and evidence that the calvaria experienced minimal transport. The relocated discovery bed is 20 m stratigraphically above the horizon from which the widely cited 1.81 +- 0.04 Ma ⁴⁰Ar³⁹Ar date for the skull (Swisher et al., 1994, Science 263, 1118) was obtained. Additional field and laboratory results will be required to determine the skull's age.

The paper gives a good history of attempts to find the original excavation site. An interesting heterogeneity of the matrix fill inside the skull (assessed by CT) also factors in the story.

After a long discussion of the complexities in dating the site, they conclude:

In summary, additional field and analytical results are needed to date the Mojokerto fossil more exactly than latest Pliocene or early-mid Pleistocene in age. The 0.3 Ma difference between the ⁴⁰Ar/³⁹Ar and fission-track age determinations must be resolved. For any of these radioisotopic dates to be considered other than a maximum age, better evidence must be advanced to show that the dated material was erupted shortly before deposition at Perning. Additional paleontological and magnetostratigraphic control and radioisotopic dating would seem to be required. Geochronological conclusions have to be evaluated further in terms of the potential for temporal stratigraphic breaks in the section, rates of deposition, and the regional stratigraphic (including sequence stratigraphic) context.

I don't suppose there will ever be a very good date for the specimen. But it's impressive the amount of work there has been on it in the past several years.

References:

Huffman OF, Zaim Y, Kappelman J, Ruez DR Jr, de Vos J, Rizal Y, Aziz F, Hertler C. 2006. Relocation of the 1936 Mojokerto skull discovery site near Perning, East Java. J Hum Evol (in press). DOI

Swisher CC 3rd, Curtis GH, Jacob T, Getty AG, Suprijo A, Widiasmoro. 1994. Age of the earliest known hominids in Java, Indonesia. Science 263:1118-1121. PubMed

Science seems to have had a stealth theme going last week on climate change, and it included this perspective by Anna Behrensmeyer on climate change in human evolution.

The central insight is the great difficulty of finding causes among temporally correlated events:

Another challenge is deciding what constitutes a strong case for a causal link between a climate change and an evolutionary event. We can't step into a laboratory to test the impact of climate change on the human genome, but we do have the results of natural experiments--the proxy evidence for environmental changes in continental rock sequences, as well as many fossils of hominins and other organisms that were evolving on different continents during that same time period. There is a rich body of data to draw upon, but hypotheses are often structured around an assumption that "synchronous" events in the geological and paleontological record constitute evidence for cause and effect. These hypotheses, while seductive in their simple explanation of how our species came to be, do not do justice to the complexity of the climate-evolution problem (see the figure) or to the full range of evidence and scientific methodologies that now can be brought to bear on this problem.

I'd say that sums up some frustrating problems very well. There is no climate-altering event during the past seven million years small enough that some paleophile hasn't offered it up as the key factor in human evolution. But how can you prove anything? How can you even test the hypothesis of causation for most of these?

I had to read this sentence a few times, but I think it circumscribes an essential problem:

The related notion that fluctuating lake levels provided environmental stress that drove speciation does not provide a mechanism for how this could have exerted selective pressure on the immediate ancestor of Homo and resulted in the emergence of a new genus and species.

This is the problem with almost any climate-driven hypothesis. How did the change in climate cause anything to happen? Especially considering the huge bias in the sites we have to sample. Sure the Rift Valley paleoenvironment was changing during the Late Pliocene, but how central was that area to the hominid range as a whole? It's like diagnosing the causes of the American Revolution only knowing what happened in Charleston.

What we sorely lack is mechanisms that would link climate change to fitness in hominid populations. So far, there are generally two serious options: "Trees Too Far To Walk Between", and "Volcanic Winter".

The story is especially bad for the origins of Homo in the Late Pliocene. The received wisdom is that the global climate got cooler, and East Africa generally got drier. We know the robust australopithecines appeared, as did stone tool manufacture and presumably Homo. But how are these linked? It would of course help if we knew what robust australopithecines ate! If we link stone tools to meat eating (reasonably), at least we have a mechanism for dietary change in Homo. But what does a drier climate have to do with that? More antelopes?

In that paragraph, Behrensmeyer is discussing the range of climate-change hypotheses for the origin of Homo, it continues:

Other proposals instead have linked human evolution with increasing aridity and climate variability. Finally, other paleoclimatic evidence indicates drier rather than wetter climatic conditions between 2.7 and 2.5 Ma [see the figure, land record (center)], bringing into question the extent of a prolonged high lake phase throughout East Africa. Although the multibasin approach to establishing regional paleoclimate trends is commendable, the proposed causal link between a wet climate phase and the origin of Homo is not yet supported by sufficient evidence to establish its credibility.

Aaarrggh! Are the lake levels a red herring? Are the global climate figures a red herring?

The stinky fish are our only trail. What we've got is the sites we've got and the climate records as they are. Is there a good reason to think they say anything interesting about the human lineage? That depends how much of the lineage we've sampled with those sites. Yuck!

The current (February 2006) issue of AJPA carries an article by Craig Stanford describing the context of bipedal posture for chimpanzees in the Bwindi Impenetrable National Park. When considering how bipedal locomotion evolved in early hominids, it is an essential comparison how chimpanzees (or other hominoids) use bipedal postures. Stanford writes:

As Hunt (1994, 1996) pointed out, hypotheses for the advent of bipedalism that involve behaviors in which prehominids may have frequently engaged offer the most plausible explanations for the adaptive shift from quadrupedal to bipedal posture (Stanford 2006:225).

Stanford was able to observe a large number of episodes of bipedal posture in the study group -- 179 cases in 247 observation hours. I find the context to be the most interesting result:

All 179 instances of bipedalism were recorded while chimpanzees were foraging in large trees. All but one instance occurred as postural rather than locomotor bipedalism, and 96% of all instances occurred in a feeding context....Chimpanzees appeared to forage bipedally most often when feeding in the upper portion of the crown, reaching up to branches emergent in the sunlight, and perhaps containing harder-to-reach ripe fruit. (Stanford 2006:227).

Studies of bipedal posture in wild chimpanzees have been rare, as Stanford reviews, but have typically found fewer instances of bipedality and have included some terrestrial cases. The key finding of all studies appears to be that foraging for fruit is the main reason why chimpanzees occasionally stand.

What do the chimpanzees tell us about early hominids? Here is the suggestion:

The behavior of wild chimpanzees suggests that several aspects of the positional behavior of earliest hominids may have been given less attention that they merit. First, arboreal bipedal posture is not dichotomous with arboreal quadrupedal posture. Bwindi chimpanzees moved fluidly between four-legged, three-legged, and two-legged postures while feeding in tree crowns. Their use of three-dimensional space in tree tops incorporated elements of positional behavior most often seen as binary states. This fluid quadrupedal-bipedal shifting may have occurred in the earliest hominids as well. Arguments about whether early hominids were fully adapted to bipedal walking, or facultatively arboreal, have been carried on for at least three decades (Susman et al., 1984; Lovejoy, 1988). Recent evidence suggests that knuckle-walking may have been employed by the immediate ancestors of the australopithecines (Richmond and Strait, 2000). Chimpanzee bipedal behavior suggests that early hominids likely engaged in a fluid variety of positional behaviors and postures, but provides little evidence for the adaptive advantage of terrestrial knuckle-walking in the last common ancestor of apes and humans (Stanford 2006:230).

Now, humans are fully adapted to bipedal walking, and we are facultatively arboreal (that is, we can climb trees), so there is no reason to think that early hominids were less facultatively arboreal than we are, and I would venture that they were probably a good deal more so.

The fundamental question about early hominids is why they abandoned the ability to be facultative quadrupeds. That is something that chimpanzee positional behavior isn't going to tell us -- after all, chimpanzees take on bipedal posture in ways that don't compromise their quadrupedal abilities.

The chief importance of the chimpanzee comparison is to illustrate the kinds of ways that locomotor diversity occur in hominoids. After a brief discussion of locomotor flexibility in gorillas, Stanford concludes with this:

Rose (1984) argued that there is no reason to view the origin of bipedalism as a progression from "poor biped" to "good biped." Instead, there was likely a diversity of forms of bipedalism in the earliest hominids. One such hominoid example may be Oreopithecus bambolii, a sup-
posed bipedal ape (Kohler and Moya-Sola, 1997). The bipedal evidence from Bwindi, Mahale, and Gombe supports this view of early hominid evolution. Instead of viewing the earliest bipedal adaptation as the lowest
rung on a posture/locomotion evolutionary ladder, it may be that early hominid species evolved a variety of forms of bipedalism in particular ecological contexts (Stanford 2006:230).

I guess that is one possibility to explain evidence of vertical posture in early hominids in the absence of good evidence of bipedality (from postcranial evidence).

The "diversity of forms" argument really suggests a stage during early hominid evolution when the ability to be effective quadrupeds had not yet been lost. Perhaps we will find these quadrupedal hominids. Perhaps we already have. On the other hand, this idea opposes the hypothesis that locomotor evolution may have either caused the origin of the hominid lineage or very closely followed it.

It seems to me that the level of species diversity of early hominids and this locomotor problem may be strongly linked. But I think they might be linked in the opposite direction than one might assume.

Suppose, for instance, that the hominid lineage arose as an adaptive radiation resulting from a significant new adaptation for bipedality. The "adaptive radiation" would be the origin of many new bipedal species spreading and adapting to different ecologies. Early hominid species diversity would be a consequence of their novel locomotor adaptation.

In contrast, if hominids originated as one among many quadrupedal apes in the Late Miocene, they might well have adapted over a long time as quadrupeds within a single ecology to which they remained limited. Perhaps the attainment effective bipedality would have spurred an adaptive radiation, but this event would have followed long after the origin of the hominids. Hominid species diversity might have always been low, or might have remained low until the Late Pliocene.

Now I don't think any of these arguments can be taken very far. It is always possible that bipedality arose early without any consequent adaptive radiation, or that there were multiple bipedal ape lineages other than hominids, or almost any other combination of events. There just isn't fossil evidence that could delimit hypotheses about hominid origins.

But I can't think about diversity without considering the mechanisms for it to have arisen. And while it is possible that many hominoid lineages were experimenting with bipedal posture and locomotion in diverse ways, I can't think of what would have caused the diversification of a large array of hominid species in the absence of bipedality.

References:

Stanford CB. 2006. Arboreal bipedalism in wild chimpanzees: Implications for the evolution of hominid posture and locomotion. Am J Phys Anthropol 129:225-231. Abstract

In Nature a couple of weeks ago, Robin Dennell and Wil Roebroeks had a provocative paper exploring the possibility that early humans (i.e. Homo erectus) originated in Asia rather than Africa.

The paper is all speculation of course; there is no evidence of any earlier hominid in Asia.

But it is the good kind of speculation. Although maybe not quite this big:

Most probably, we are on the threshold of a profound transformation of our understanding of early hominin evolution that might prove as far-reaching as the demise of the notion of Man the Hunter in the early 1960s (Dennell and Roebroeks 2005:1103).

Here's the abstract:

The past decade has seen the Pliocene and Pleistocene fossil hominin record enriched by the addition of at least ten new taxa, including the Early Pleistocene, small-brained hominins from Dmanisi, Georgia, and the diminutive Late Pleistocene Homo floresiensis from Flores, Indonesia. At the same time, Asia's earliest hominin presence has been extended up to 1.8 Myr ago, hundreds of thousands of years earlier than previously envisaged. Nevertheless, the preferred explanation for the first appearance of hominins outside Africa has remained virtually unchanged. We show here that it is time to develop alternatives to one of palaeoanthropology's most basic paradigms: 'Out of Africa 1' (Dennell and Roebroeks 2005:1099).

It is worth reviewing exactly what "Out of Africa 1" is supposed to be. The paradigm is that emergence of hominids from Africa required increases in brain size and/or body size, coincident with the emergence of hominids like KNM-ER 3733, KNM-WT 15000, and others. The motivation for this hypothesis is simple: australopithecines have not been found outside of Africa. Nor has anything like Homo habilis, which is australopithecine-sized but has larger brains.

Of course, it is questionable just how basic this paradigm is. Consider what I (and my colleagues) were able to write only seven years ago:

The problem is that significant range expansion out of Africa occurred a half million years or more later than the first H. sapiens [corresponding to others' H. erectus or H. ergaster]. Population size before then may have remained small, and this is not an inconsequential time span, being one quarter of the time H. sapiens has existed. An important date in behavioral evolution is 1.5 MYA because it is marked by the earliest appearance of the Acheulean, the ubiquitous hand-axe industry of the Early and Middle Pleistocene.... Before this time, humanity was limited to Africa and immediately adjacent sections of Asia such as the Levant (Hawks et al. 2000:7).

Evidence for large body size in Late Pliocene humans (notably KNM-WT 15000 but also many others) made it very plausible that larger bodies were necessary for dispersal from Africa. But without good evidence for such dispersal before around 1.4 million years ago (and arguably not before 1 million years), larger bodies could not be assumed to be a sufficient condition for dispersal. Writing about the origin of humans, we had to consider all these alternatives -- at a time when the Dmanisi sample consisted of a single uncertainly dated mandible and the Mojokerto date stood alone with very questionable provenience.

Now we know that hominids did leave Africa by at least 1.8 million years ago. Dmanisi has almost singlehandedly changed the perspective.

And in doing so, it made much more convenient the hypothesis that large body size was both necessary and sufficient for dispersal from Africa. If the date of dispersal and the date of human origins are the same, then it is natural to propose that the coincidence is more than chance.

I would say this is more of a convenient hypothesis (and an easy story to tell) than it is a basic paradigm. The idea that large body size caused dispersal from Africa may have been a local minimum in terms of parsimony (at least as long as the body size of the Dmanisi fossils was not known), but it was only one alternative among many still in play.

And it remains a plausible hypothesis -- after all, the Dmanisi remains are a bit larger than australopithecines, and they might well have shrunk from a larger early-human-like size after reaching Asia instead of before.

But Dennell and Roebroeks give motivations for examining some alternatives.

The only reason why the earliest tool assemblages in Asia are attributed to H. erectus s.l. is that palaeoanthropologists have already decided that, in effect, it was the only hominin capable of migration out of Africa, and with sufficient Wanderlust to do so (Dennella and Roebroeks 2005:1099).

Homo erectus sensu lato (s.l.) means Homo erectus "in the loose sense", which would include not only the "strict sense" (sensu stricto) H. erectus. from Java and China, but also hominids like OH 9 and KNM-ER 3733 from Africa, and presumably the Dmanisi hominids.

A long passage reviews the total faunal evidence from Asia during the Late Pliocene. The thrust of the passage is that there are very few sites with extensive fauna, and of these most preserve mainly large-bodied herbivores. There are a few hints that a hominid-friendly fauna may have existed, including the presence of baboons. But there are no hominids of any kind at the vast majority of Asian localities -- Dmanisi is a real exception in the Plio-Pleistocene record.

This is the key taphonomic argument: if we have only found Early Pleistocene humans from continental Asia within the past ten years, then how can we preclude there having been australopithecines there? Dennell and Roebroeks argue that if there were australopithecines, we shouldn't necessarily expect to have found them yet -- we just haven't looked extensively enough.

A close read of the section raises a caution, though. One of the main arguments for the incompleteness of the Asian record is that sites don't preserve each others' fauna.

It is also likely that the full range of taxa is incomplete for the Indian subcontinent, because Megantereon and Pachycrocuta are not recorded in India but are present in Pakistan; in Pakistan, there is no evidence of Camelus and small primates, and in neither country is Homotherium recorded, although this is present to the west at Dmanisi, to the north at Kuruksay, central Asia and to the east at Longuppo, south China (Dennell and Roebroeks 2005:1100).

Of course, all of these species are recorded in Asia taking all the sites in aggregate; this is hardly an argument for the overall weakness of the record -- just an argument that no individual site is an adequate record of the continent's fauna.

To me, the important question is not whether australopithecines as currently known from Africa were in Asia. A more troubling possibility is that the australopithecines that we now know from Africa were not the only (or main) manifestations of early hominids in Africa. Large parts of Africa that we might expect to be congenial to hominids, like the Zambesi basin, have few or no fossils at all. The recovery of the Bahr el Ghazal mandible (Brunet et al. 1994) certainly makes clear that hominids were living across a much larger area than we have adequately sampled. But that mandible is, although not identical, certainly very similar to known contemporary hominids in its adaptation.

The question is whether hominids had adapted to other ecologies that are much less satisfactorily sampled than the East African rift. They probably weren't living where chimpanzee and gorilla ancestors did, but where else might they have been? Some such ecologies -- like the coasts -- would make early dispersal very plausible.

(In this regard, early humans are not the only hominids who lack a satisfactory ancestor. Who was the ancestor of A. aethiopicus? In what ecology did the first robust hominid arise?)

So what is the broader set of hypotheses that we should consider? Dennell and Roebroeks suggest:

If the above taphonomic review suggests that we cannot show the absence of hominins from areas in Asia at a time before the little evidence we have indicates their presence, we need to consider alternatives to the current Out of Africa [that is, their "Out of Africa 1"] model. There are three issues here. The first is when hominin(s) first left Africa -- might they, for example, have left shortly after they acquired the ability to make stone tools, the earliest of which are currently 2.6 Myr old? Or could they have left even earlier, about 3.0Ð3.5 Myr ago, when some australopithecines were already living in the African grasslands? The second issue is whether we yet know the full range of hominins that inhabited both Africa and Asia in the Late Pliocene and Early Pleistocene. Even in east Africa, several new taxa have been claimed in the past decade (for example, A. anamensis, A. garhi, Ardipithecus ramidus and Kenyanthropus platyops) and doubtless more will be found. (An indication of how little we know about Pleistocene east Africa is that only recently has the first fossil evidence for chimpanzee been found.) In Asia, the recent discoveries of H. georgicus and H. floresiensis should make us very wary of assuming that H. erectus s.l. was the only player on the Asian stage in the Early Pleistocene. Third, Asia might not have been the passive recipient of whatever migrated out of Africa but might have been a major donor to speciation events, as well as dispersals back into Africa. Such two-way traffic is well documented for other mammals in the Pliocene and Early Pleistocene, such as Equus and bovids, with more taxa migrating into than out of Africa. There is no reason why hominin migrations were always from Africa into Asia, and movements in the opposite direction might also have occurred, as has been suggested for the Olduvai OH9 (refs 13, 58) and Daka specimens. We should even allow for the possibility that H. ergaster originated in Asia and perhaps explain its lack of an obvious east African ancestry as the result of immigration rather than a short (and undocumented) process of anagenetic (in situ) evolution (Dennell and Roebroeks 2005:1100-1101).

Of course, most of the evidence indicating the presence of hominids is not fossil but archaeological. On this topic, Dennell and Roebroeks have much to say:

Any stone tool assemblage in Asia dated as older than 1.9 Myr ago (the earliest date that Homo is supposed to have left Africa) is either dismissed or (more usually) ignored; undated Oldowan tools are assumed to