Patterns of Variance and Covariance in Anthropoid Limb Proportions: Implications for Interpreting the Hominin Fossil Record Open Access
Downloadable ContentDownload PDF
Interpreting the taxonomic and behavioral implications of variation in the inferred limb proportions of fossil hominin taxa is contingent upon assessing how much variation exists in extant primate taxa and, by extension, how much of that variation is associated differences in their locomotor behaviors. However, the majority of evidence linking limb proportions to behavior in extant primates is based on taxonomically-restricted samples, or on species means as opposed to individual values, or does not account for field observations that capture the complexity of locomotor behavior in a primate taxon (see Napier & Napier, 1967; Fleagle, 1988; see also Preuschoft, 2002). With regards to extinct taxa, the problem is compounded by a necessary reliance on relatively few associated skeletons, most of which are incomplete, or fragmented or both. This thesis addresses the aforementioned issues using a) multivariate methods to quantify the relationships between limb proportions and behavioral repertoires in extant anthropoids; b) machine-learning methods to select relevant extant models with which to interpret the limb proportions of extinct taxa; and c) resampling methods to evaluate hypotheses regarding major adaptive shifts in inferred locomotor behavior.