The forelimb and hindlimb remains from Liang Bua cave

18 minute read

Jungers and colleagues (in press) provide a medium-length description of the lower limb remains from Liang Bua cave. In a second paper with much overlap of authors, Susan Larson and colleagues (in press) describe the upper limb remains. Both papers present quite a bit more information about the hindlimb remains than the earlier presentations of the postcranial remains in Nature by Brown et al. (2004) and Morwood et al. (2005). That is no surprise, since even a 6-page Nature paper is very superficial in terms of description and comparison, certainly when we have more than a single bone to consider. Unlike many Nature papers of recent vintage, the 2005 paper by Morwood and colleagues didn’t include a very extensive data supplement (compare, for example, the 2006 description of the Dmanisi postcrania by Lordkipanidze and colleagues). This is not really a knock on the previous papers, which conformed to the Nature limitations. But for non-insiders, it can be hard to understand why so many disagreements have arisen over basic anatomical facts. In the case of the Flores bones, one important reason for these disagreements has been the brevity of the descriptions and lack of specialized review, which has allowed some errors to be propagated into print.

I’m beginning to think it would be better to go to an online refereed monograph. Just put everything out there, let the world pick out the problems, respond to them in an iterated fashion, and let libraries buy hard copies if they want. That would make just as much news as the traditional journal-driven model, it would count as a major monograph instead of a journal article for the major authors, and it would facilitate the distribution and provision of comparative materials.

Heck, I’d be happy to underwrite the whole thing – although I’m sure I wouldn’t be alone, since this would be a pretty simple operation.

Meanwhile, the Journal of Human Evolution appears to have taken a monographic approach by publishing a set of papers related to the Flores hominids in an upcoming issue. The paper on the hindlimb elements by Jungers and colleagues is part of that issue, and comprises 15 pages of description on hindlimb elements alone (pelvis, leg and foot). Likewise, Larson et al. give us 15 pages of description on the forelimb elements. What more could one ask for?

As it turns out, plenty. A special journal issue does not a monographic treatment make.

Descriptions: somehow never quite enough

I applaud the effort that has gone into preparing these descriptions; they clearly add much to the record. I have some hesitation in criticizing the papers, since they really do answer some questions that have remained mysterious up to now. I’d rather have them than nothing at all.

But as I began to read through them, I was missing many of the kinds of obvious comparisons that a descriptive paper ought to include. Where are the data tables with ranges and means for human populations? The obvious comparison here would be with African Pygmies, Andamanese, or other small-bodied humans. Collecting such data from scratch would be a lot of work (and not possible in some cases), but we might at least expect to see a literature review.

Indeed, the authors give an impression of having already performed a review – for instance, there are passages like this:

At 280 mm in total interarticular length (Brown et al., 2004), LB1/9 is shorter than any modern human femur of which we are aware, including African pygmies and Andaman Islanders; it is, however, remarkably close in length to the reconstructed femur of AL 288-1 ("Lucy") (e.g., Jungers, 1982) (Jungers et al. in press:6).

And yet, there are no references to sources for femur length in either population. Another example:

The humerus is 243 mm in total length (Morwood et al., 2005), which, while short in an absolute sense, can be matched to the lower extremes of small-bodied African pygmies and Andaman Islanders (WLJ, pers. obs.) (Larson et al. in press:3).

Now, I was taught that anthropology is a comparative science, and that description can only take place in the context of comparison. So I’m really missing the comparisons here – as basic as means and ranges, for goodness’ sake! Nothing against “personal observation” as a source, but I wish there were numbers.

I don’t want to take this criticism too far. The papers obviously were not intended to provide comparative information. Presumably, the authors (or others) have plans to provide comparisons elsewhere (as Larson et al. 2007 have done for the shoulder anatomy). Possibly we will be looking at one or more monographs over the next few years.

Well-described elements

As I mentioned, there’s a lot of description packed into these 30 pages of text. That provides many opportunities for solid descriptions. Several anatomical elements are described with reference to early hominids, modern humans, or other apes. A number of these passages include references to literature on the variation of these comparative models. Although I wish there were more, I want to give lots of credit to the authors for the instances that provide sufficient information to pursue further.

Jungers comes into his own when describing the foot. If you want to know about foot and ankle anatomy of LB1 (and to a minor extent, LB6 and LB10), there’s a lot in this description to work with.

Likewise, the description of the carpals in Larson et al. (in press:7-9), presumably written by Matt Tocheri, is material that would have been helpful as accompaniment to the 2007 Science paper on the same topic by Tocheri and colleagues. The description includes photographs of the bones and selected articulations.

Now, maybe anatomical description is not your bag. These are not primarily interpretive papers – in fact, neither paper includes a discussion. But assuming that we don’t see a monographic treatment of the remains very soon, these are the best anatomical descriptions we can expect for a while. Elements like the scapula and clavicle don’t often get their due in discussions of human evolution. Here they are.

At last, a chronology!

The first things I looked at in the papers were the chronologies. Each paper includes a list of specimens, including each element of LB1 separately and all other forelimb and hindlimb elements. Much has been made of the notion that the deposits in Liang Bua cave span 95,000 years – a date that would supposedly be earlier than modern humans could have occupied the island. So the anatomy of the earliest specimens in the cave should be of very great interest.

Despite many intimations that the anatomy of the earlier specimens is “indistinguishable” from the later ones, I haven’t yet been able to sort out exactly which of those specimens are supposedly early. It has been evident that the major specimen LB1 and the partial skeleton LB6 are both late in the sequence, younger than 20,000 BP. But nothing published to date has given me a clear idea of which specimens date to the earlier time intervals in the cave.

The chronologies here allow us to sort it out, at least as far as hindlimb and forelimb elements go. Almost all the bones are younger than 20,000 BP, including LB1, LB6, the juvenile partial LB4, the LB8 tibia, LB5, LB7, LB9, LB12, LB13 – all of them are less than 20,000 years old. LB14, a fragment of acetabulum, is bracketed between 18.2 and 41±10 ka.

Four specimens are assigned to the date interval 74+14/-12 ka. This may be an exaggeration of the date – for example, LB2 was previously bracketed between this early date and an overlying date of 37.7±0.2 ka (Brown et al. 2004). Still, the anatomy of the five specimens older than 20,000 years ago – LB2/1, LB3, LB10, LB11 and LB14 – are of great interest. None of them are crania – so they can’t tell us any more about the very small brain size of LB1 and whether it was characteristic of the population generally. But these specimens alone can tell us whether the later specimens represent a long-lasting small-bodied population on the island.

Larson et al. (in press) provide a good description and measurements of the two forelimb elements from these five specimens. LB2/1 is a proximal ulna, including both the coronoid and olecranon processes and an intact semilunar notch with measurable diameter. It is associated with a left P3, as mentioned in the initial H. floresiensis publication by Brown et al. (2004). In its measurements, the ulna is slightly smaller than the right ulna LB1/52 and slightly larger than LB6/3. LB3 is a radius shaft, lacking both proximal and distal ends. In its extant dimensions, it is a bit bigger than the LB6/2 radius (although Larson et al. (in press) didn’t include neck diameters for LB6/2).

In contrast, Jungers et al. (in press) provide a perfunctory description of the early hindlimb elements. The lack of description is attributed to the condition of the remains, which evidently gave them little to work with:

LB11 includes badly fragmented pelvic and metatarsal bones, ...[omitting LB13]... and LB14 is a fragmentary acetabulum. These fragments add little to the anatomical description of the hind-limb remains of H. ?oresiensis; therefore, they are not ?gured here nor discussed further (Jungers et al. in press:15).

That’s literally the entire description. Sure, there may not be much there, but after 14 pages of detail on LB1, would a picture have hurt? The fragmentary acetabulum may not be enough for any meaningful size estimate, but if not, could we at least hear about it?

Anyway, the exception is LB10, which is a proximal big toe phalanx. Jungers et al. (in press) describe it as basically human in form but quite small; they illustrate it articulated with the corresponding metatarsal from LB1 and it appears to fit.

From these descriptions, I think we can conclude that the early specimens from Liang Bua are consistent with the post-20-ka remains. As it stands the early parts of the sample are consistent with the explanation proffered: that they represent a single small-bodied hominid population. They were apparently already small as preserved in time horizons chronologically near the arrival of seagoing peoples. The date associated with these specimens, with a maximum as recent as 62,000 years and possibly younger, is not older than modern humans might have occurred in the region. The earlier, 95,000 year date is not pertinent to these remains.

It would be nice to have a broader comparison. For example, how does the variability within this sample compare to the variability within small-bodied human populations?

I don’t think we can yet test the hypothesis that modern humans themselves generated the archaeology at the site, or even that they may have been the ancestors of the later small-bodied islanders. We really want more detail from the early part of this sequence. That goes to the issues of chronology that I’m discussing in a separate post.

The australopithecine connection

I first raised the comparison in 2004, reiterated it in early 2005, and disavowed it by the spring of that year. With the small body and brain size of LB1, and the small body sizes evidenced by LB6 and LB8, the Liang Bua remains are within the size range of known australopithecines. Further, their proportions and some elements of their anatomy (early on, I considered the pelvic anatomy and endocast size) look similar to known australopithecines.

Now, why did I stop thinking that the Liang Bua hominids were connected to australopithecines? First, was persuaded that significant evidence of pathology in LB1 might eliminate the small brain size as a character relevant to the population.

Second, I became less convinced that a small brain was good evidence for an australopithecine relationship, even if the brain was shrunk by evolution and not pathology. In short, I came to think that a dwarfed human population might well have evolved a relatively small brain.

Aside from the brain size, none of the other “australopithecine-like” features of the Liang Bua hominids are very compelling. The pelvis of LB1 has some interesting similarities with that of AL 288-1 (“Lucy”), but maybe that’s what we should expect of a small-bodied hominid. Other features of the Liang Bua (and in the wake of the Gona pelvis, probably even the pelvic features) wouldn’t distinguish them from early Homo. Heck, most of the supposedly primitive Liang Bua features are within the range of modern humans (Jacob et al. 2006).

If someone really wanted to test the hypothesis that the Liang Bua hominids had a close phyletic relation with australopithecines (and not Homo), they would need to do very extensive comparisons of the variation and distribution of metric and non-metric features in living people and recent humans. Living people have a tremendous store of variation that is Middle Pleistocene in origin. Small body size creates a tremendous potential of convergence with the features of australopithecines and other small-bodied early hominids (e.g. Homo habilis).

Jungers et al. (in press) and Larson et al. (in press) are not out to test this hypothesis. But their comparisons skew toward Australopithecus. They often provide direct descriptions (and in some instances, photographs) of the way the Liang Bua specimens compare to “Lucy” (AL 288-1), Sts 14, Dmanisi, or other small hominds. In contrast, they provide few comparisons with modern humans, and none at all with modern humans who live in Southeast Asia, Indonesia, or nearby regions today.

I readily admit, this is my short impression of over 30 pages of detailed text. The early hominid comparison is obviously relevant – the very small LB1 and LB6 skeletons may be close to these small-bodied hominids in size. On the other hand, comparisons with small-bodied human populations would be equally (or possibly more) relevant.

For better or worse, those of us who study fossil hominids have a tendency to reference our knowledge of anatomy from those few specimens. It’s natural to compare the pelvis of LB1 to Lucy, and describing it in those terms is economical since most of us are very familiar with Lucy’s anatomy. But it will be even more helpful to have a full range of comparisons with modern humans.


After a brief introduction to the history of the specimens, Jungers et al. (in press) include synopses of two papers discussing the hypothesis of pathology for LB1 – these include the contribution by Argue et al. (2006), arguing against one diagnosis of primordial dwarfism, and another by Hershkovitz et al. (2007) arguing in favor of a diagnosis of Laron Syndrome (I discussed that paper here).

We should remind ourselves why the pathology hypothesis is relevant, given that there are other small-bodied specimens from the site. No one disputes that a small-bodied population existed on Flores. The time depth of this small-bodied population has not been clear, but that has nothing to do with pathology.

What some people dispute is whether this small-bodied population was also a small-brained hominid species. So the argument for pathology is principally about the brain size of LB1. If LB1 was an unusual individual in which some pathological condition resulted in small adult brain size, it wouldn’t by itself be sufficient to evidence a small-brained hominid species. Some critics (e.g. Jacob et al. 2006, Hershkovitz et al. 2007) have examined other parts of the skeleton, including the postcrania, looking for possible correlates of developmental abnormalities in the brain. This process is essentially the paleopathology method of diagnostics – looking for a suite of characters that result from known disorders, and comparing those to the traits of a skeletal individual.

The postcranial descriptions provided by Jungers et al. (in press) and Larson et al. (in press) are potentially very relevant to this process. With substantially more space than earlier descriptions, they could provide the comparative basis to conclude that the morphology of LB1 is within a normal range of variation for those features previously described as consistent with pathology. By themselves, the ranges of variation might not be sufficientto test the hypothesis – after all, many living people are diagnosed with pathologies based on a combination of factors, not because any single factor lies outside the normal range of variation. But such information would certainly be a start.

The papers approach this issue cautiously. Jungers and colleagues (in press:2) end their introduction by noting that they do not intend to test the hypothesis of pathology:

When appropriate, we critique speculations about pathology (e.g., degrees of left-right asymmetry, cortical bone thickness), but a full rebuttal to these diverse claims of pathology is beyond the scope of this contribution.

Likewise, Larson et al. (in press:2) include this statement:

Although a detailed rebuttal to each of these assertions is beyond the scope of this paper, some claims have been based on inaccurate depictions of H. ?oresiensis morphology, perhaps due in part to the necessary brevity of the original descriptions of the material. The following detailed descriptions of the upper limb elements from Flores provide a more complete characterization of this material. For more detailed and comprehensive comparative analyses, the reader is referred to Argue et al. (2006), Larson et al. (2007), and Tocheri et al. (2007).

Both are fair enough. Neither paper intends to provide a detailed comparative perspective on the range of normal variation in humans or other relevant species. They are not monographs. So we should read the descriptions as giving possible information about pathology, while noting that they were not compiled for that purpose.

Along those lines, Jungers et al. (in press) include four arguments relevant to earlier assessments of the hypothesis of pathology. First, they dispute the conclusions of Jacob et al. (2006), who asserted that the LB1 long bones had unusually thin cortical thickness. Jungers et al. (in press:8) write that “the cortical indices in LB1 are all in the normal human range, and the metric claims and attendant conclusions of Jacob et al. (2006) can be firmly rejected.” Unfortunately, Jungers et al. (in press) provide no references or tabular information about the human ranges of such values. I’m quite willing to believe them, but given the adversarial tone of the contribution, it would be appropriate to actually back these assertions with relevant facts!

Second, Jacob et al. (2006) had asserted that many of the postcranial elements of LB1 had left-right asymmetries that might indicate a developmental disorder of some kind. In contrast, Jungers et al. (in press) claim that the extent of any asymmetries of the pelvis, femora, and fibulae of LB1 are “trivial” or “minor.” In some instances, apparent asymmetries are explained as the result of postdepositional processes or breakage. within the range or low compared to asymmetries of the same elements in modern humans.

Third, Hershkovitz et al. (2007) had suggested that the tibial curvature of LB1 was consistent with Laron Syndrome. Jungers et al. (in press:8) write that this curvature is “slight” and “comparably curved tibia [sic] can be found among modern humans.”

Finally, Jungers et al. (in press:8) cite Hrdlicka (1898) in support of the observation that the tibial shaft of humans is quite variable in cross-section. Jacob et al. (2006:13425) had observed that LB1 has an oval tibial cross-section, interpreting that shape as “suggesting compromise between the need to support and move body mass and generally weak muscle development.” In contrast, Jungers et al. (in press) argue that the oval cross-section is a sign of robusticity, concordant with the relatively large shaft diameters of all the long bones, a condition they relate to other early hominids. Jungers et al. (in press:5) also emphasize that the femur shaft is robust, explaining its external anatomy in terms of robusticity instead of developmental abnormality.

Larson et al. (in press) do not discuss the arguments about pathology to any notable extent. They do provide additional description and information about the pathology of the LB6/2 radius, which appears to represent a break that had healed with the distal end of the bone displaced from the long axis. Larson et al. (2007) had described the humeral torsion of the specimen and placed it in the context of the variation of modern humans and fossil hominids, so that aspect of the forelimb anatomy, raised by some as a possible symptom of developmental pathology, was already addressed in the earlier publication.

That’s more or less it, in terms of information relevant to the possibility of pathology, either in LB1 or the sample generally. Maybe the anatomical descriptions will give others more to go on, and as Larson et al. (in press) suggest, may resolve some of the disagreement about the details of the specimens that may have caused confusion. I am skeptical, though. If we really want people to understand the details of the anatomy, we’re going to need a real description that includes comparative evidence, ranges of variation and definitions of states in reference samples.

In any event, the focus for the pathology hypothesis remains the brain. Postcranial observations may be relevant to the diagnosis of a developmental disorder, but the test must ultimately return to the skull.

Last words

That’s more or less what I got out of my reading of the papers. I’ll return to these over the next few weeks, I’m sure, as I think about the issues related to the Flores hominids.

It’s disappointing that the few early specimens don’t preserve more anatomy. They are insufficient to compare earlier and later portions of the sample; except insofar as the earlier specimens fit within the size range of the later ones. All are small, which is useful to know.

Jungers et al. (in press) discuss some anatomical features of the postcrania that have been suggested by some authors to be diagnostic or symptomatic of developmental disorders or other pathological conditions. In each case, Jungers et al. (in press) conclude that the anatomy is either expected for a robust early hominid, or is within the range of variation of humans. These observations could be better documented.

Finally, some of the groundwork for a link between the Flores hominids and australopithecines can be found in these papers. They do not push this hypothesis, and do not really provide the data to test it. But they present many of the anatomical features using explicit comparisons with early hominid specimens.


Jungers WL, Larson SG, Harcourt-Smith W, Morwood MJ, Sutikna T, Rokhus Awe Due, Djubiantono T. In press. Descriptions of the lower limb skeleton of Homo floresiensis. J Hum Evol (early online) doi:10.1016/j.jhevol.2008.08.014

Larson SG, Jungers WL, Tocheri MW, Orr CM, Morwood MJ, Sutikna T, Rokhus Awe Due, Djubiantono T. In press. Descriptions of the upper limb skeleton of Homo floresiensis J Hum Evol (early online) doi:10.1016/j.jhevol.2008.06.007