The last three decades have seen an expansion of morphological disparity analyses, which is now routinely indexed alongside species diversity in clades. Strikingly, species diversity is often a poor correlate of morphological diversity across clades and time bins. Small clades with only a handful of species can encompass spectacular morphological diversity, whereas species rich groups may be morphologically highly conservative and vice versa. Higher morphological disparity should ultimately result from changes at the DNA level, and we might expect it to be associated with greater divergence between species at the molecular level. However, genetic or molecular disparity has never been quantified explicitly in a manner analogous to morphological disparity, and its possible relationship with both morphological disparity and species diversity has never been assessed. To investigate this relationship mammals represent an ideal study group, since their diversity is strongly heterogeneous across higher clades, there is a well-studied conflict between morphological and molecular phylogenetic analyses, and extensive genomic data are available for many species. Here, we explore one possible interpretation of the concept of genetic disparity, along with indices for its quantification. For major mammal subclades, we find a significant positive pairwise correlation between genetic disparity inferred from each of 9 different genes, suggesting that there is a consistent signal in genetic disparity. In marked contrast, we find no significant correlation between any indices of genetic and morphological disparity. Our results support the notion of genetic disparity as an evolutionary index that complements those of morphological disparity and diversity.

This data is publicly available from:

https://morphobank.org/index.php/Projects/ProjectOverview/project_id/773

We have pruned the data for compatability with certain R. packages: Ape, Claddis, dispRity  and vegan

We have removed Rhyncoyon cirnei from the data as it's genetic coverage was too low for our methods. Multistate discrete morphological characters have been replaced with unknown character states in order to accomodate the Claddis function morphdistMatrix().