What about Palau?

10 minute read

Lee Berger, Steve Churchill, Bonita De Klerk, and Rhonda L. Quinn have written a paper in PLoS ONE describing the skeletal remains of small-bodied humans recovered from two caves in the Rock Islands of Palau.

Full disclosure: I was the academic editor for this paper at PLoS.

OK, editor, what did you do for the paper?

The editor's role is to evaluate the manuscript's suitability for publication -- does it conform to the journal guidelines? Is it scientifically valid? Does it cite the existing literature appropriately? Do the observations support the claims made? A few manuscripts may be rejected immediately, because they fail to meet basic criteria of scientific value or readability.

Most manuscripts require the editor to seek out the opinions of additional experts in the process of peer review. PLoS ONE, unlike most journals, is committed to openness in the review and publication process (journal information).

In the case of this manuscript, I think it was a good fit to PLoS ONE because of the potential to report the new finds in an open access forum, where anyone can read the original research. It is not a monograph on the archaeology or skeletal biology of the sites, it is merely a preliminary report. However, unlike the kind of preliminary reports that we often see in journals like Nature or Science, in this case the journal provided more space for description and the potential to provide long lists of specimens. Many of those additional details were added to the manuscript in response to my editorial comments. If you read the reviewer reports for the paper (available at PLoS ONE), you can see that these additional details were essential to the scientific value of the manuscript, and that is why I required them. In addition, I suggested many other changes that would increase the value of the manuscript. The final version reflects the authors' responses to these changes: a preliminary report on the skeletal remains, in context, given the limitations presented by preservation and the need to conserve and prepare additional specimens.

Rex Dalton made the National Geographic Society's involvement with the research into a news story. Do you have any involvement with the media for a story?

Nothing at all. Sadly, most good manuscripts don't get any media attention.

Dalton emphasized the media attention to the find, particularly focusing on the role of the National Geographic Society. NGS produced a documentary about Berger's work on Palau (he is an NGS grantee).

In this case, National Geographic funded the work and apparently produced a documentary about it. Their production wasn't disclosed to the journal, and I view it as irrelevant to the scientific evaluation of the manuscript.

Paleoanthropologist Tim White is quoted in Dalton's story, saying that it appears that the "review process [was] driven by popular media." Since White was not involved in the review process of this paper, he obviously is just speculating.

I tend to give him the benefit of the doubt, since in this story it appears that Dalton was trying to play up any contrary quotes about the findings. Why else would he run otherwise-uninformed comments of the kind in the story?

I would tend instead to ask these questions: Does the Nature Publishing Group (NPG), in publishing Rex Dalton's piece, have a vested interest in the credibility of their own journals, in comparison to open access outlets like PLoS? Do NPG journals regularly receive manuscripts and publish them based on the associated media attention? Do they have an interest in pressuring grant agencies, like the NGS, into encouraging submission of manuscripts to NPG journals instead of alternate outlets? Does NPG have a well-established record of running stories questioning the value of open access publications?

In other words, consider the sources.

But aren't there other, normal-sized people on Palau at the same time?

In Elizabeth Culotta's article about the Palau specimens, she quotes archaeologist Scott Fitzpatrick:

But archaeologist Scott Fitzpatrick of North Carolina State University in Raleigh, who has worked in Palau for a decade, says he doesn't think the bone beds represent a true population. In a site only 4 kilometers from Berger's caves, he has excavated the burials of Palauans of similar age--and normal stature. That would seem to rule out isolation and island dwarfing, he says. "It would be very unusual to have a group of people living in close contact with a normal size population who evolved to be smaller." Instead, "the most parsimonious explanation is that they were Palauans with a genetic anomaly leading to small people who were buried in a clan or family plot."

Fitzpatrick has been excavating on Palau for a long time -- in comparison, Berger is a real Johnny-come-lately, who happened across his sites while on a vacation. So it's fair to say Fitzpatrick knows what he is talking about -- he has documented the earliest radiocarbon-dated cemetery on Palau, dating to approximately 3000 years ago.

The osteology of the skeletal remains from that cemetery, Chelochol ra Orrak, were reported by Fitzpatrick along with Greg Nelson, in 2006. That preliminary report is similar in form to this one, and they report measurements of the specimens. There are fewer specimens in that cemetery than are reported by Berger et al. from their caves, but the reported measurements are very comparable.

For example, Nelson and Fitzpatrick (2006:5) report a single femur complete enough to assess length; it has a maximum length of 392 mm and a maximum head diameter of 38.5 mm. This is smaller than the maximum femur length for the Khoisan sample reported by McHenry (1991), with an average length of 405.1 mm (S.D. 20.86), but longer than his (2) Akka Pygmy specimens (330 mm, S.D. 5.66). Berger and colleagues have no femora sufficiently preserved to estimate length, but their two femoral heads have diameters of 38.8 and 36.1 mm. These compare to Andamanese mean values of 37.3 mm and a San mean of 42.3 mm; McHenry's Khoisan sample has a mean of 37 mm.

Berger et al. report the proximal mediolateral diameter of two tibiae (63.1 and 53.1 mm); Nelson and Fitzpatrick (2006) report one specimen with a epiphyseal breadth of 64.8 mm and an estimated maximum length of 318 mm; the paired tibia has a length of 315 mm. By comparison, Flower (1885) reported a mean tibia length for Andamanese females of 321 mm.

In other words, the comparable remains published by Nelson and Fitzpatrick (2006) and Berger et al. (2008) appear to be consistent in size, and all within the range for small-bodied and pygmy human populations. One caveat is that the crania are as yet not directly comparable: Berger et al. cannot assess the crania from their caves because they remain to be prepared. Nelson and Fitzpatrick (2006) report two adult crania, the more complete of which (presumably male) is not a small skull, averaging larger in all preserved dimensions than Andamanese. Berger et al. (2008) report larger skeletal remains from areas that they believe are later in the chronology.

In any event, we don't need to posit two distinct populations living side-by-side to explain these remains. Working out the actual dynamics of this population over time is going to require a detailed understanding of several complicated stratigraphies, as well as detailed comparison of the skeletal remains. Whether they do in fact represent a single population must be determined by comparing the bones from these different sites with each other in a longer treatment. I hope that the analysts can get together to assess the sample as a whole.

Is this an extreme case of island dwarfing?

There's no question that the bones are small. However, I would not characterize them as extremely small compared to other small-bodied human populations. The paper provides a series of comparisons of linear dimensions of the Palauan remains to other small-bodied skeletal samples, including San, Onge, and Great Andamanese. In most cases, the Palauan remains average slightly smaller than these small-bodied samples, but within one standard deviation of the mean. A few adult specimens are substantially smaller, but it is not obvious that they are outside the range of living pygmy populations.

I should mention that is also true for the Liang Bua skeletal remains from Flores -- they are not obviously outside the range of living pygmy human populations -- despite the fact that none of the publications have reported comparisons with pygmy populations.

So, this would seem to be within the ordinary range of dwarfing in human pygmy populations. That raises the possibility the people may have been derived from such a pygmy population -- for instance, by colonization from the Philippines where small-bodied populations such as the Aeta and Batak are found today. There may be nothing exceptional about a relatively long-distance colonization of Palau by these peoples, who must after all have gotten to the Philippines!

I would say that the initial dispersal of the many small-bodied populations of mainland and island Southeast Asia is shaping up to be a very interesting anthropological topic. This population history has been partly obscured by the subsequent expansions of agriculturalists -- indeed, new colonizations like that of Palau may represent the effects of such interactions. Only traces of the ancient diversity remain, so it is difficult to reconstruct the ancestral population structure. But it is becoming increasingly clear that this was a cosmopolitan population, inhabiting several outlying island groups as well as areas of mainland Southeast Asia and the Sunda shelf.

In that context, small body size must be explained not merely as a consequence of inhabiting small islands, but more generally as an adaptive strategy for hunter-gatherers living in tropical ecologies. That is also true in the African context, so it should come as no surprise.

What do these small-bodied skeletons tell us about Homo floresiensis?

These ancient skeletons from Palau are not anything other than small-bodied modern humans. There is no question about that.

However, the bones share some interesting features with the Flores specimens. Here is what the authors say about the resemblances:

We feel that the most parsimonious, and most reasonable, interpretation of the human fossil assemblage from Palau is that they derive from a small-bodied population of H. sapiens (representing either rapid insular dwarfism or a small-bodied colonizing population), and that the primitive traits they express reflect possible pleiotropic or epigenetic correlates of developmental programs for small body size. In the comparisons drawn below, we note the shared possession of these traits with the Liang Bua fossils not to imply phylogenetic affinity or taxonomic identity, but rather to caution that some of the primitive features argued to reflect an ancestor-descendant relationship between H. erectus and H. floresiensis may also be homoplastically shared with modern humans from Palau, and thus that care must be exercised in interpreting their taxonomic and phylogenetic significance.


These results also suggest that the simple presence of additional small-bodied specimens with reduced chins (that cannot be shown to share all of the traits considered taxonomically significant in the Holotype Flores LB1) is insufficient to confirm the taxonomic validity of H. floresiensis.

I think that's pretty much the extent of what the Palau skeletal remains can say about the Flores sample. This is a reminder that the best comparative material is relatively local -- they should be compared to regionally similar populations, and populations of similar body size, rather than random people from European museum collections.

I think they also provide a cautionary note to the kind of trait-based typological classification that has been applied to the Flores specimens. Just because a sample lacks so-called "derived" features of "anatomically modern" populations, does not make it a member of a pre-human hominid population. "Anatomically modern" is itself a typological classification. Living people are variable and many express morphological features shared with fossil humans of various ages.

To be sure, the Flores LB1 specimen presents a large number of features that would be unusual in a living person, and the hypothesis that it represents a distinct human species is well-supported on this basis. On the other hand, it also has some features that are derived in recent human populations and others that are relatively common in small-bodied human populations in the region.

Nevertheless, the Palau remains do not provide any positive support to the idea that the Flores LB1 specimen is a microcephalic modern human. In particular, like other small-bodied human populations, none of the Palau skeletal remains indicate anything like the reduced brain size of LB1. There is nothing in this research concerning wrist morphology, humeral torsion, rotated premolars, or any of the rest of the odd features of LB1, nor is there any assessment of the paleopathology of the remains.


Berger LR, Churchill SE, De Klerk B, Quinn RL (2008) Small-Bodied Humans from Palau, Micronesia. PLoS ONE 3(3): e1780. doi:10.1371/journal.pone.0001780

Nelson GC, Fitzpatrick SM. 2006. Preliminary investigations of the Chelechol ra Orrak Cemetery, Republic of Palau: I, skeletal biology and paleopathology. Anthropol Sci 114:1-12.

Flower WH. 1885. Additional observations on the osteology of the natives of the Andaman Islands. J Roy Anthropol Inst G Br Irel 14:115-120.