Supplementary files for: Why should we compare morphological and molecular disparity?

1. Indices of morphological disparity seek to summarise the highly
multivariate morphological variation across groups of species within
clades, time bins or other groups. Morphological variation can be
quantified using geometric morphometric, outline or surface-based methods.
These are most effective when morphological differences are relatively
modest and there are numerous ubiquitous landmarks and phase-aligned
features of shape variation. The most disparate samples, such as those
across classes and phyla, typically necessitate the use of discrete
characters. Unfortunately, such characters are often compiled subjectively
in a manner reflecting the level of morphological and taxonomic focus and
the intensity of taxon sampling. 2. Sampling intensity is often highly
variable within a single data set, especially in repurposed and
amalgamated cladistic matrices. Here we propose indices of molecular
disparity analogous to those of morphological disparity. Molecular
sequence data can be obtained in a more objective, automated and scaleable
manner than morphological data. 3. Comparisons of the morphological and
molecular disparity of subclades in sixteen large data sets suggest that
molecular disparity is less susceptible to sampling biases than
morphological disparity. Moreover, distance matrices inferred from
individual genes tend to correlate strongly with each other and with
distances from all concatenated genes. By contrast, morphological and
molecular disparity are typically not significantly correlated across
subclades, such that comparisons for groups can help to give a fuller
picture of their evolution. Within mammals, Afrotheria have conspicuously
high morphological disparity but modest molecular disparity, suggesting
unusually high morphological plasticity. Even more strikingly, the
molecular disparity of rodents is over five times that for Artiodactyla,
despite having only half of their morphological disparity. These contrasts
suggest the differential operation of geometric, biomechanical,
ontogenetic and environmental constraints on form. 4. Given the increasing
abundance of total evidence data sets in the literature and the widespread
and sometimes uncritical repurposing of discrete morphological characters,
we propose the comparison of morphological and molecular disparity as a
useful tool to understand subclade evolution more fully.