Recognizing species diversity among large-bodied hominoids: a simulation test using missing data finite mixture analysis

A persistent problem in paleoanthropology is the recognition of intra-vs. i nter-specific differences within fossil samples. Exacerbating this situatio n is the often fragmentary nature of the fossils themselves, thus precludin g rote applications of many multivariate approaches designed for complete c ase analyses. In this paper we apply finite mixture analysis to samples of large-bodied hominoids to test this procedure's efficacy in clustering indi viduals by species without a priori knowledge of group membership. In addit ion, we stochastically remove individual specimens and measurements,, simul ating small, incomplete fossil samples, and re-apply the finite mixture pro cedure to test how often it correctly assigns these "fragmentary" specimens . Finite mixture analysis can be highly accurate, even when confronted with s mall sample sizes and missing data. For example, a combination of 124 chimp anzees and humans are correctly identified in one analysis, and the accurac y drops only 2% to become 98% when the total sample size is reduced to 16 a nd missing data patterns are applied. In comparisons to better known method s that have been used to recognize groups in the fossil record, such as k-m eans, the benefits of finite mixture analysis are readily apparent. First, k-means is unable to accommodate missing data, an obvious deficiency when i nvestigating the fossil record. Second, in direct comparisons of their abil ity to accurately assign "unknowns" to taxa, finite mixture performed at le ast as well as, and often better than, k-means in our analyses. A potential test that can be used to identify species in the fossil record, derived fr om comparisons of results generated from a general vs. a restricted (isomet ry-corrected) finite mixture analysis, is presented. (C) 1999 Academic Pres s.