Constraint is a fundamental concept in evolutionary theory. Morphology and ecology both are limited by functional, historical, and developmental factors to a subset of the theoretical range species could occupy. Cat-like carnivorans (Feliformia) offer a unique opportunity to investigate phenotypic constraint, as several feliform clades are purported to be limited to generalized ecomorphological roles, while others possessing extremely specialized durophagous (bone-crushing) and sabertooth morphology. We investigated the evolutionary history of feliforms by considering their phylogeny, morphological disparity and rates of evolution. We recover results that show a mosaic pattern exists in the degree of morphological disparity per anatomical region per clade and ecology. Non-hypercarnivores, such as viverrids (civets and genets), Malagasy euplerids and lophocyonids (extinct hypocarnivores) have the greatest dental disparity, while hypercarnivores (felids, nimravids, many hyaenids) have the lowest dental disparity but highest cranial and mandibular disparity (excluding dentition). However, high disparity is not necessarily associated with high rates of evolution, but instead with ecological radiations. We reveal that relationships between specialization and disparity are not as simple as past research has concluded. Instead, morphological disparity results from an anatomical mosaic of evolution, where different ecologies correlate with and likely channel unique patterns/combinations of disparity per anatomical partition.

Morphological information was taken from observation and measurment of museum curated fossil and modern skeletal specimens and published datsets from scientifc literature. This was combined with molecular sequence data (nuclear and mitochondrial) for extant and recently extinct taxa downloaded form GenBank. The best partition and evolutionary models for phylogenetic analysis was determined using Stetting Stone analysis for the morphological data, while PartitionFinder 2 for the moecluar data. From these data, a total-evidence Bayesian phylogentic analysis was run in Beast. Rates of evolution and morphological disparity were then calculated from the morphological character dataset in the R Claddis package across a sample of the posterior trees in the phylogentic analysis, as well as the MCC tree.