More important implications of Flores: fossil damage

It’s been a long day, more than a full work-day of fossils. Which is nice. Foolishly checking updates to Journal of Human Evolution instead of leaving to eat and relax, I find yet another paper about the Flores fossils. My interest in Flores has strongly declined since last year, yet I had to look. Too tired to read the paper, but rather skimming through the figures, I find out the following disappointing information:

An attempt was made to cast the LB1 mandible, which resulted in damage to surface bone. Filler was used to correct the visible damage, and external dimensions were altered due to the separation of cracks in the corpus and a corresponding increase in bigonial breadth. The deep cut marks now running along the lateral inferior corpus of the mandible are an artifact of the mold-making process (Fig. 5). The dimensions and morphology reported here were recorded prior to the LB1 mandible being removed from Jakarta, and are supported by CT scans, stereolithographic models, radiographs, and photographs.

During an attempt to cast this mandible, it was irreparably damaged. The symphysis was broken, the medial surface of the right ramus removed and poorly repaired (Fig. 16), and cut marks, glue, and filler are now a feature of the lateral corpus and ramus. This has altered the original arch dimensions, occlusion, and morphology of the symphysis and ramus. Data reported here are based on a limited number of dimensions and photographs recorded before LB6 left Jakarta, supplemented by measurements, photographs, radiographs, and CT scans recorded after the mandible was returned. Unfortunately, many of these observations can no longer be verified through reference to the original specimen.

Damage to fossils is a very unfortunate consequence of their study. I’m told that the Taung child (first Australopothecus africanus discovered) cannot be studied anymore, and that so many people have taken measuring calipers to the thing that its gross morphology is drastically and irreversibly altered. Fortunately, non-invasive methods of study, such as CT and laser scanning, are becoming more common. However, until scanning becomes cheaper and scans of fossils are made widely and easily available to all scientists, our fossils will continue to be subject to the abuses of study. Or they become unavailable. And one important criterion for research to be “Scientific” is for it to be reproducible. Reproducability is undermined by damage to fossils, witholding of fossils for fear of possible damage (or being data-greedy), and the inherent difficulties in making CT data easily available. Quite a shame.

I suppose I should end on a positive note. What I did like from my brief glimpse of the paper is that they published all their raw data. This is an important step to Scientific repeatability (although sometimes it is good to verify others’ work on your own).

Oh, and on a less positive note, there’s a principle components figure that is one of the worst things I have ever seen. The figure is supposed to demonstrate how modern humans, fossil humans, Australopithecus afarensis, and the LB1 and LB6 chins are distinguished base on their side-view outlines. Modern humans overlap every single group, including A. afarensis. So the significance of the Flores chins falling closest to some A. afarensis specimens is not so shocking: some modern and fossil humans fall closest to afarensis, too. Anyway, a terrible figure, which I refrain from posting here simply out of concern for copyright. But just imagine something painted by Jackson Pollock, or an explosion, or what Oreos look like inside your stomach.


Brown, P and Maeda, T. In press. Liang Bua Homo floresiensis mandibles and mandibular teeth: A contribution to the comparative morphology of a new hominin species. Journal of Human Evolution. doi:10.1016/j.jhevol.2009.06.002

More LB1 stuffs, but no real news

New analyses of the Flores “hobbit” material have been published recently, including a geometric morphometric (GM) study of the LB1 cranium (Baab and McNulty, in press). Readers will recall that this particular small-brained specimen has fueled the controversy over whether the Flores material is a unique, small-bodied hominin species, or whether this and other specimens were pathological (humans). At this stage in the game, I still do not know what we can say exactly is going on until we get more fossils. Admittedly, I haven’t yet read the recent analyses of the postcranial material. Maybe they will change my mind. So, I’d like to keep this more descriptive and comparative, than speculative on phylogeny versus pathology. Some neat things from the study:

Coordinates of landmarks were recorded off a “stereolithographic model” of LB1. This is a neat technology in which a plastic 3D model is ‘printed out’ based on data from a CT scan. While this might not give the same resolution as the real thing (or a good cast of the real thing), it’s a great way to make physical models of delicate fossils. Additionally, measuring/collecting landmark data from fossils can always potentially damage specimens, and stereolithography is a way to circumvent this problem. However, I don’t know that it would have been any more difficult to digitize landmarks from the actual CT scan (i.e. with a computer software package), and it almost certainly would have been cheaper. Oh well, smoke them if you’ve got them, right?

The study used GM to compare the shape of LB1 cranium to the shapes of apes, fossil hominins and modern humans. I like that the study included the analysis in Procrustes shape space (with nuances of location, rotation and scaling omitted) and in form space (omits only location and rotation, leaving size/scaling in the analysis). By examining the specimens in form space, the authors were able to study the relationship between shape and size, i.e. the effects of allometry: what cranial shape would “an LB 1-sized individual” of each taxon have? The results of the form space analyses are kind of neat, although I can’t interpret them too readily since I’m not sure what exactly the first two PC scores indicate, other than size. In a nutshell, though, LB1 appears most similar to what would be expected of a small-bodied fossil Homo individual. Of course, this is nothing new: Gordon et al (2008) came to the same conclusion last year, but using inter-landmark distances as variables (as opposed to landmarks as variables). Additionally, last year we found that the microcephalics in our lab appear more similar in shape to modern than fossil humans (I wonder if we should have tried to publish that…).

The authors also examined asymmetry as a way to address the allegation of pathology in LB1 (left-right asymmetry is believed to be indicative of developmental disruptions arising from environmental or genetic stress). LB 1 is more asymmetrical than the human mean (but within the range of variation), and roughly in the middle of the range of variation for fossil hominins. This suggests to the authors that LB1 is not pathologically asymmetrical, and that asymmetry in the fossil is rather likely due to effects of fossilization. Unfortunately, however, the issue of pathology-related asymmetry is not wholly adequately addressed, so this result must be taken with a grain of salt (fluctuating asymmetry is generally believed to indicate developmental disruptions, and there are two other types of asymmetry that must be factored out…). It is at this point that some GM qualms I have come to the fore. Really there’s one issue specifically: homology and measurement error.

As mentioned above, GM statistically analyzes shapes using landmark coordinates as variables. Of course when comparing things, you want to make sure you’re actually comparing equivalent things (“homologous” structures). So on a cranium there are a few, which Bookstein identified as “Type I,” and are often the intersections of bony sutures, such as the point where your two nasal bones contact your frontal bone. But beyond that, many landmarks are extrema (“Types II-III), i.e. points of widest breadth, which are not necessarily (ever?) homologous structures. The trouble, then, is a large risk of comparing non-homologous structures, which can render biological analyses nearly meaningless. Additionally, many of these such points must be found fairly arbitrarily, which means that even if given landmarks are homologous (i.e. biologically meaningful), there is a good chance of measurement error confounding the study–this is especially true in fluctuating asymmetry studies.

Use of these Type II-III landmarks is not terrible, it’s just a reason to be wary. One must ask, based on how arbitrarily a landmark was found, and the likelihood of measurement error, how meaningful will the results be? One landmark that comes to mind from this study is alare, “the most lateral point on the margin of the nasal aperture.” How significant is ‘lateral-most’ as a criterion–is there something functionally important about this point, and is it very variable within a given taxon (I really do not know, but now I’m interested…)? Moreover, the margins of the nasal aperture (“nose hole”) are not always sharply delineated, but can be rounded, making the margin itself difficult to identify: for example, OH 5 has very rounded margins, but KNM-ER 406 seems to have sharp margins–both specimens are Australopithecus boisei. In this study, alare contributed second most to individual asymmetry in all taxa. Is this because of measurement error, or the fact that this is not a biologically significant point, or both/neither? Other fun landmarks include “malar root origin,” and “Frontomalare temporale… The point where the frontozygomatic suture crosses the temporal line (or outer orbital rim)” (my emphasis). So throughout the study, you took either one intersection or the other?! When did you use which one, and why? Homology FAIL. Maybe I (or someone else) will come up with a new (hopefully better) way to study asymmetry… Now I’ve lost my train of thought.

Baab K and McNulty K. Size, shape and asymmetry in fossil hominins: The status of the LB1 cranium based on 3D morphometric analyses. J Hum Evol, in press.

Gordon A, Nevell L, and Wood B. 2008. The Homo floresiensis cranium (LB1): Size, scaling, and early Homo affinities. Proc Nat Acad Sci 105: 4650-4655.

I, or someone, have drawn a brown, orange, and blue Gwenhidwy

You’re probably thinking, “I thought zacharoo was dead,” because I’ve been completely MIA for the past few weeks. My apologies, but I was trying to wrap up this past semester, the terminus of my first year in grad school. And I must say I think I did a pretty good job, not to toot my own horn. This is as good a time as any to ask, “What the eff have I learned this year?”

1. Milford is awesome. Probably the past few decades have shown this, but I’ve only known the guy for less than a year. Given my more ‘arts’ educational background, Milford (and the Big Chief and the rest of the bios) have taught me how to do ‘science,’ formulating and testing a hypothesis. Though I’m certainly no Chung-I Wu, my mentors and colleagues have certainly gotten me started. Also, I was a bit unhappy with Milford last semester for pushing me and my peers to take a heavy course load. But I must say it was worth it, I’ve learned a lot this past year, and if I’d taken another (i.e. the non-bio) way I would not have learned nearly as much. He can also improvise a wicked country-twangy song (“I wish I grew up on a pig farm”). Great advisor, great man.

2. Steer clear of the hobbit. That situation is messier than the van-ride back from DC. I talked about LB 1 (the hobbit skull) in a few posts earlier this year. It’s clearly not a cretin, and at the AAPA meetings in Columbus a few weeks ago, Dean Falk defensively countered the Laron Syndrome hypothesis. Bill Jungers reported at the meetings that the foot of LB 1 was not that of a runner (I forget the specifics, but it was missing one or both of the plantar arches). It’s overall cranial shape based on various measurements show it has striking affinity with Homo habilis (in a broad sense) <!–[if supportFields]> ADDIN EN.CITE Gordon200810310317Gordon, Adam D.Nevell, LisaWood, BernardThe Homo floresiensis cranium (LB1): Size, scaling, and early Homo affinitiesProceedings of the National Academy of SciencesProc Nat Acad SciProc Nat Acad Sci07100411052008March 20, 2008 10.1073/pnas.0710041105<![endif]–>(Gordon et al. 2008)<!–[if supportFields]><![endif]–>. Chief, her husband Adam, Pappano, and I looked at UM’s collection of modern human ‘microcephalics’ (there are myriad ways to be microcephalic), and found that LB 1 is still more similar to the habilines (cf. Gordon et al 2008). This really suggests to me that maybe some early Homo or Australopithecus species made it out of Africa to Flores early in human prehistory; however, I don’t think we can say yet whether it’s a real case of insular dwarfing in a hominin or pathology or what. Still very messy.

3. Molding and casting is neat but difficult. One of the projects Milford got me started on a cranial reconstruction. Sounded simple enough at first, but it has required me to make molds and casts of the individual cranial bones: the two temporals, occipital and the paired parietals were not too difficult, but the face is really giving me grief. It was also difficult fitting my busy schedule to Bill Sanders’s lab schedule. So, long story short, I didn’t finish the project (should be done before June . . .), but Bill has taught me a ton about molding and casting, as well as proffered his wisdom. It has also reinforced my desire to be a paleontologist. Cool beans.

4. Genetics sucks. For decades now, paleoanthropology has come to be not just about fossils, but also about molecules. Today, genetic studies are incredibly influential in studies of human evolution, i.e. supporting models of migration and introgression. But I took a course this past semester in the department of Ecology and Evolutionary Biology, and it seems to me that molecules are not really any less unequivocal than fossils. Really genetics is all comparing predictions of models with various parameters (i.e. effective population size, population expansion, etc.) with actual empirical data, and it’s all about probability. So you can say something like, “There is a high probability of seeing this type of data given that type of hypothesis/model.” But different sometimes data have the same probability given different parameters. So genetics can tell a lot, but you have to take what they say with a gram of coke, I mean granary of salt. And all the nucleotides in the world probably won’t help resolve robust australopithecine phylogeny.

Now I’m tired, so I’ll stop there for now. I’ll post more pearls (of wisdom) I learned this semester as I recall them. So that’s where I’ve been o’er the past few somethings. Weeks. Oh, and I just received a possible job offer working with crash-test dummies (or something, I’m not exactly sure) with the UM Transportation Research Institute, hopefully that works out.


<!–[if supportFields]> ADDIN EN.REFLIST <![endif]–>Gordon AD, Nevell L, and Wood B. 2008. The Homo floresiensis cranium (LB1): Size, scaling, and early Homo affinities. Proceedings of the National Academy of Sciences:0710041105.

The Little Hominin That Could (Fuel Publications)

Two and a half months into 2008 and we already have three (if not more) papers out about (or with bearing on) the diminutive Flores hominin material. Indeed, interested paleoanthropologists are like the “first family” at AL 333, and the papers an interesting catastrophic event that is burying us all (not that they’re terribly bad, or anything).

As mentioned in an earlier post (“Cretins and Omomyids”), the past four years have borne witness to a huge debate on whether the ‘Homo floresiensis‘ is a new species of insular dwarfed hominin, pathological early human or what. There is no doubt about the fact that the material from Liang Bua cave, dating to as recent as 12 ka, is small. What has been most hotly debated is the etiology of the unique features of the LB-1 skull, such as (From Brown et al. 2004):

  • Cranial capacity estimated ~400 cc
  • Reduced parietal lobe development in the brain (relative to H. sapiens)
  • Long, low cranial vault
  • Maximum cranial breadth just above the mastoids
  • Prominent nasal pillars
  • Mesially rotated mandibular P4s
  • Posteriorly inclined mandibular symphysis (i.e. no chin)

Arguably these traits distinguish LB1 from modern humans (although one of my mandibular premolars is fairly mesially rotated), and more closely aligns it with earlier Homo taxa, e.g. H. habilis/erectus. Last year, Hershkovitz et al. made a pretty strong case for many of these ‘distinguishing’ characters being the result of Laron syndrome.

The first floresiensis-related paper this year was Obendorf et al.’s diagnosis of myxoedematous cretinism. According to them, thyroid malfunction resulted in the anamolous Liang Bua morphology. Their basis for this diagnosis was what they perceived to be an enlarged pituitary fossa, based on a picture of a CT scan. Though interesting, this hypothesis was nevertheless blown out of the water, as many parties–many of whom have seen the actual specimen–denied existence of an enlarged pituitary fossa. Bummer there.

A few days later, Berger et al. reported on diminutive humans (dating to 3-0.9 ka)from caves on the island of Palau. Authors noted some similarities to the Liang Bua material. They suggested that, since their material is undoubtedly human, but similarly tiny like floresiensis, many of the unique features of floresiensis might really be a function of their diminutive size, and that new species designation might not be appropriate. Of course, most of the LB 1 anomalies are cranial, and Palau crania are not yet observable, so Palau, at the moment, has little bearing on Liang Bua (viz. LB 1 and 6).

Finally, Gordon et al. today published a paper in which they used craniometric morphology of LB 1 to establish its taxonomic affinities. Their analysis shows that LB 1 falls outside their human sample, and clusters well with H. erectus, in a broad sense, and to a lesser extent with H. habilis. Thus, contra those who posit that LB 1 is human, but unique because of pathology or pleiotropic effects of small size, Gordon et al. support the idea that LB 1 (and therefore, all diminutive Liang Bua hominins) represents a new species, derived from H. erectus or habilis (cf. African or Dmanisi erectines). Their compartive sample included 2500 modern human crania from all over the globe and 30 hominin crania.

I suppose that I have to take issue with the sample. Comparing LB 1 to modern human crania from all over the globe is a bit like comparing a 15 ka Javanese apple to a modern, international fruit basket (I admit I’m embellishing). On the other hand, comparing it to a wide diversity of human crania might be generous, as it increases the range of variation, making it easier for LB 1 to fit in. But they did not address the issue of pathology very well. They write:

With regard to microcephaly, it should be noted that in the shape analysis performed here, LB1 cranial shape is shown to differ significantly from the modern human comparative sample (and from fossil H. sapiens and Neanderthals) and to be very close in morphological space to non-Asian H. erectus specimens (D2700 and KNM-ER 3733). And H. habilis specimens (KNM-ER 1813 and OH 24). LB1 and the non-Asian H. erectus specimens are much closer than the average pairwise distance between modern human crania, and standardized residuals of LB1 from the estimated non-Asian H. erectus and H. habilis scaling relationshps average 1.38 and 1.21 standard deviations away from expected shape, respectively . . . well within the range of population-level variation (Table 1). Thus, if microcephaly is responsible for the extremely small size of the LB1 cranium, of all possible ways that microcephaly could cause LB1 cranial shape to differ from that modern humans and for these six variables, it happens to differ in the same way that earlier Homo species differ from modern humans [p. 4654].

Ok, so here they sort of test the hypothesis that LB 1 is craniometrically like a microcephalic by comparing it to a large sample of non-microcephalics–now that’s comparing ancient apples to a Harry and David exotic fruit basket! To my knowledge, Argue et al. (2006) are the only ones to compare craniometrics of Liang Bua to any microcephalics. Their findings (see their Figs. 3 and 4) show that, as in the Gordon et al. study, LB 1 is most close to ER 3733 and OH 24. But, this study looks at only two microcephalics (though they compare LB 1 to these in separate analyses…), and these two individuals are on the extreme margins of a wide range of human variation. Moreover, there are myriad ways to be microcephalic, and it is unclear how different ways might affect craniometric variation.

That Gordon et al. and Argue et al. agree that LB 1 is craniometrically very similar to ER 3733 and OH 24 might not be insignificant. But Gordon et al. make the assumption that modern human microcephalics are craniometrically no different from non-pathological humans. While the Argue study suggests microcephalics might not be too different, they do show they are nonetheless on the outskirts of human variation (when also compared to fossil hominins…); moreover their microcephalic sample is very small. In order for Gordon et al. to reject microcephaly, they need to satisfactorily establish that microcephalics are not too different from humans cranimetrically. And that requires a larger sample of microcephalics. Hey, there’s a handful of microcephalic skulls (get it–“handful,” cuz they’re micro…) in my lab, maybe I’ll undertake this endeavor, after I get done with my billion other projects….

Anyway, apparently Hershkovitz et al. didn’t officially settle the Liang Bua issue. It’ll be interesting to see what Falk et al. bring to the table in Columbus in a few weeks. A never-ending saga . . .
Argue D, Donlon D, Groves C, and Wright R. 2006. Homo floresiensis: Microcephalic, pygmoid, Australopithecus, or Homo? J Hum Evol 51: 360-374.

Berger L, Churchill S, de Klerk B, Quinn R. 2008. Small-bodied humans from Palau, Micronesia. PLoS ONE 3: e1780

Brown P, Sutikna T, Morwood M, Soejono R, Jatmiko, Wayhu Saptomo E, Awe Due R. 2004. A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia. Nature 431: 1055-1061.

Gordon A, Nevell L, Wood B. 2008. The Homo floresiensis cranium (LB1): Size, scaling and early Homo affinities. Proc Nat Acad Sci 105: 4650-4655.

Hershkovitz I, Kornreich L, Laron Z. 2007. Comparative skeletal features between Homo floresiensis and patients with primary growth hormone insensitivity (Laron Syndrome). Am J Phys Anthropol 134: 198-208.

Obendorf PJ, Oxnard CE and Kefford BJ, in press. Are the small human-like found on Flores human endemic cretins? Proc R Soc B xx: 1-10.