Since 2004 the taxonomic status of Homo floresiensis has been one of the more hotly contested issues in paleoanthropology. I have 34 articles on the subject and there are some that I haven’t acquired yet. The Journal of Human Evolution has four more papers on Homo floresiensis. One discusses the Liang Bua faunal sequence, a record that spans 95,000 years, and fills in the paleoenvironmental context of the finds. Two other papers describe the postcranial anatomy of all the Liang Bua homins. The fourth paper which looks at the LB1 cranium is, by far, the most interesting.
The paper (Size, shape, and asymmetry in fossil hominins: The status of the LB1cranium based on 3D morphometric analyses) is authored by Karen Baab and Kieran McNulty. The study took 35 midline and bilateral landmarks from a wide variety of ape, fossil hominin, and modern human crania. The material was divided into two samples for analysis. The first sample included neurocranial and facial landmarks, while the second, larger sample, included neurocranial landmarks only. Taxa sampled include the bonobo, three subspecies of chimp, three subspecies of gorilla, Australopithecus africanus, Paranthropus boisei, H. habilis, H. Erectus (African, Asian, and Dmanisi), Neanderthal, Kabwe, Petralona, Omo 2, Dali, and Skhul. The data was analyzed using principal components analysis (PCA) and Procrustes analysis. Additionally, the authors used multivariate regressions of the landmark data to create simulations of scaled down versions of modern humans and fossil hominins and LB1 was compared to each (more about this below). Finally, the authors looked at the issue of asymmetry. Hawks summarizes the asymmetry argument here.
So, what are the results? In both the PCA and the Procrustes analyses LB1 was closest in morphology to H. erectus. In terms of asymmetry, LB1 was less asymmetrical than Kabwe, Bodo, and Arago and more asymmetrical than OH 5, Petralona, Abri Pataud, and Fish Hoek. In terms of extant taxa, LB! was within the range of asymmetry displayed by bonobos, chimps, and gorillas but not modern humans. This brings us to the simulations. Basically, as mentioned above, the simulations were created by multivariate regression of all the shape variables. From the article:
Separate simulations were generated from the modern human sample and the fossil hominin sample. Unlike regressions on single eigenvectors, this procedure incorporated all of the shape variation correlated with size to estimate the landmark configurations of
hypothetical individuals with a cranial centroid size equal to that of LB1. Simulations were then compared to the real LB1 specimen by computing the Procrustes distance between them; this metric summarizes overall shape differences as the distance between two configurations in Kendall’s shape space … Distances between LB1 and the simulations were evaluated against intraspecific variation in Procrustes distances calculated from extant taxa … If the distance between a simulated configuration and LB1was within the 95% distribution of Procrustes distances in a modern species, this supported the hypothesis that LB1′s morphology could be derived from a known human group. If the distance between a hypothetical configuration and LB1 was beyond the 95% distribution, this constituted evidence that size reduction in modern or fossil humans was not sufficient to account for the LB1 morphology.
Turns out that scaling down a modern human is a poor predictor of LB! morphology. Scaling down fossil hominins does a much better job of predicting LB! morphology. One other item of interest is that LB1 was close to the Dmanisi material and some of the African Homo erectus as well as the Indonesian H. erectus.
Bonus Material! There is one more interesting point to be extracted from this paper. In my recent post on the Palau skeletons I quoted this bit from Berger’s paper:
A number of the individual traits observed in the Palauan sample are seen also in specimens from Flores (although the form of these traits may differ in the Palauan sample), some of which have been argued to support the unique taxonomic status of H. floresiensis: small body size, reduction of the absolute size of the face, pronounced supraorbital tori, non-projecting chins, relative megadontia, expansion of the occlusal surface of the premolars, rotation of teeth within the maxilla and mandible, and dental agenesis. These last two features are not argued to be taxonomic markers, but their occurance in specimens from both Palau and Flores is notable, as they may be parallel results of founder effects, genetic isolation and a high inbreeding coefficient, or may simply be a factor of evolutionarily rapid reduction in body and craniofacial size…
Baab and McNulty discuss what happens when both archaic humans and modern humans undergo size reductions:
Richards (2006) also suggested that the seemingly archaic nature of LB1 might be the direct result of a modern human population becoming smaller in an island setting. Results of the sizeshape analyses suggest otherwise; small modern humans do not more closely resemble H. erectus or LB1 than do larger modern humans. In fact, our results for modern H. sapiens neurocrania (captured by PC 1 of Procrustes form space) indicate that those with smaller skulls are relatively taller at bregma, less projecting posteriorly, and have wider frontal bones compared to those with larger skulls. This is opposite to the pattern observed in the LB1 cranium. Rather than size reduction causing modern humans to look more archaic, it is clear from results presented here that size reduction instead causes archaic facial morphology to look more modern, at least in certain respects. [bolding mine - afarensis]
In other words, the cause of whatever primitive traits the Palau material displays, and I’m somewhat skeptical they display any, has nothing to do with small body size.
Update 1: I should have mentioned that in one way or another these four papers respond to most of the criticisms that have been raised concerning H. floresiensis. Some of the responses are quite effective. It would still be nice if we had some more crania though…