CAN A FRAGMENTED PAST BE TRUSTED? ASSESSING PHYLOGENETIC SIGNAL IN THE SQUAMATE FOSSIL RECORD

Alongside birds, squamates (lizards, snakes, and amphisbaenians) are the most diverse group of tetrapods, occupying every major biome throughout their ~257 million-year evolutionary history. Topological incongruence exists among current studies of squamate phylogenetic relationships, particularly between analyses using phenotypic data (skeletal morphology) and genotypic (mitochondrial, nuclear DNA) data. This conflict represents a major hurdle in understanding squamate evolution and necessitates a re-examination of how phylogenetic data is sourced. The present study: 1) characterizes the nature of the fossil phylogenetic data currently available from a pivotal period in squamate evolution (Late Cretaceous) in museum collections, and 2)assesses the phylogenetic “fidelity” of this data by measuring phylogenetic signal in fossil squamate skeletal elements “overrepresented” in museum collections. We sampled 480 skeletal specimens in Late Cretaceous (Campanian) squamate collections at four major North American natural history collections.We determined that 31.25% of observed specimens are tooth-bearing, 42% are dermal (osteoderms), 13.5% are axial (vertebrae), 11.25% are appendicular, and 2.1% are non-toothbearing cranial (skull). We used a recent comprehensive squamate morphological phylogenetic dataset (610 characters, 192 taxa) to assess homoplasy and synapomorphy (Consistency Index and Retention Index) in characters corresponding to underrepresented and overrepresented skeletal elements in the squamate fossil record. Our results showed that premaxilla and mandible characters possess statistically higher levels of homoplasy, and statistically lower levels of synapomorphy with respect to characters from skeletal elements that were underrepresented in fossil squamate specimens. These results reveal that skeletal elements commonly preserved in the squamate fossil record contain less phylogenetic signal than do other regions of the skeleton, and may negatively affect our ability to accurately infer the relationships of fossil taxa. This information can help determine how best to utilize a fossil dataset for a given group of organisms, and allows us to better characterize the limits and utilities of incomplete datasets.