Meeting Abstract

42.1  Sunday, Jan. 5 10:15  Evaluating the impact of anatomical partitions on morphology-based phylogenetic reconstructions BURROUGHS, R. W. ; The University of Texas at Austin, Jackson School of Geosciences RBurroughs@utexas.edu

The nature of biology is such that the total amount of potential phylogenetic data cannot be used to construct phylogenetic hypotheses. This is particularly true for the fossil record, which most often is limited to morphological data for phylogenetic inference. In systematic analyses that include fossils this often results in an implicit assumption that individual anatomical regions each provide complimentary phylogenetic signal and hypotheses. I chose to explicitly evaluate this assumption, by testing whether or not two separate character sets each provide congruent and complimentary phylogenetic hypotheses. I used a dataset of extant turtles from the Emydidae and partitioned it into two anatomical regions, head and shell. My results indicate that these anatomical regions did not independently yield complimentary or congruent phylogenetic hypotheses. The results of a partition-homogeneity test and comparison of consistency index scores were inconclusive as to whether the two partitions could be combined. Results from Bayesian partition analyses indicated that the best models of evolution for my data are ones that allow independence among the partitions in the form of unlinked branch length priors. My results support the idea that head and shell characters for emydid turtles can be combined, but need to be modeled with some degree of independence. These results also indicate that it is not appropriate to assume that anatomical regions provide complimentary phylogenetic signal and hypotheses without testing this assumption. My work provides a framework by which others may investigate their own morphological datasets to evaluate similar problems.