Elliptical Fourier analysis of symphyseal shape in great ape mandibles

The midsagittal profile of the mandibular symphysis has served as both a taxonomic marker and a phylogenetically salient character in debates over hominoid evolution. Nevertheless, the utility of symphyseal shape as an informative attribute for paleobiological reconstructions is suspect. Quantification of shape variation has proven to be particularly problematic; it has long been recognized that conventional linear measurements (and the indices derived from them), while replicable, summarize aspects of shape very poorly because of the vast amount of contour information that is lost in the process. In this study, a type of Fourier analysis is applied to cross-sectional contours of ape mandibles in order to provide a mathematical accounting of shape variation in a global sense; that is, by applying the totality of contour information in a comparative analysis. Shape variation in the mandibular symphysis is explored through the decomposition of coordinate data into elliptical Fourier coefficients. These coefficients are used to compute average taxonomic distances (ATD) among individuals of chimpanzees, gorillas, and orang-utans. The resulting shape-based distances are summarized via clustering (UPGMA) and ordination (principal coordinates analysis-PCO). Principal coordinate scores are subjected to analysis of variance in univariate and multivariate designs; these data are also applied to discriminant function analyses. Species and sex effects on morphology are statistically significant; however, no significant interaction of these factors is indicated. This would seem to imply that patterns of sexual dimorphism are not distinct among great apes; to the contrary, within-species sex comparisons reveal that significant size and shape dimorphism is present only in Gorilla. Despite significant size dimorphism in Pan and Pongo, significant shape differences between males and females are not apparent in these taxa. These results suggest that it is theoretically possible to sort taxa by a symphyseal shape criterion, but the discriminant function results suggest that there still exists a large potential for error in assigning particular shapes to a given species or sex. Thus, despite real shape differences among these species, the use of symphyseal shape as a character in species identification or in systematic arguments remains limited and problematic. (C) 2000 Academic Press.