Phylogenetische Untersuchungen an Schädeln der Neuweltaffen (Platyrrhini)

Abstract

In the present study, the statistical procedure of Q-factor analysis with a Varimax rotation (see e.g. BORTZ, 1993) was applied to detect complex characters from skull measurements of New World monkeys (Platyrrhini) and related taxa. The idea behind this approach was that similarities in complex characters are more unlikely to evolve independently multiple times. Therefore, they are more suitable for phylogenetic reconstructions on the basis of parsimony. Skull measurements were taken from 160 Platyrrhini, 54 Catarrhini, and ten Tarsius, predominantly at the Field Museum, Chicago. To solve the problem of missing data, the method of multiple imputation (SCHAFER, 1997) was utilized. Five different completed datasets were generated, separately analyzed, and the results combined. To scrutinize the suitability of this method for morphometric data, datasets with missing data were artificially generated from a complete set of measurements taken from ten Colobus. A comparison of the results from the complete and incomplete data showed that multiple imputation yields reliable results. For the Q-factors that divided the investigated individuals into separate clusters of taxonomic groups, an outgroup comparison was used to identify synapomorphies, which were then taken to detect monophyletic groups. Four substudies were performed using different ingroups, namely the clawed monkeys (Goeldi s monkey, tamarins, and marmosets), the Pitheciines, the prehensile-tailed monkeys, and the entire Simians. Results regarding the phylogeny of New World monkeys were a sister group relationship between Saguinus and Callithrix, between Callimico and all the remaining tamarins and marmosets, between Chiropotes and Cacajao, and between Lagothrix and Ateles. Furthermore, as an unusual result, there where indications that Alouatta and Brachyteles might be sister groups. Relationships on higher levels could not be revealed. This probably points out to the limits of phylogenetic analyses using morphometric data. The results of the present study are compatible with recent molecular findings. Although occasional difficulties with factor rotation and finding the number of factors to extract are to be discussed, Q-factor analysis was found to be a suitable tool to construct complex morphological characters which yield more reliable results than simple characters. Furthermore, factor analysis guarantees character independency, an important prerequisite for phylogenetic analyses.