Modules, semi-autonomous subsets of highly-correlated
traits, provide a practical framework for summarizing general patterns of trait
interactions, which are often cited as a major influence on morphological
variation. However, little comparative
data on modularity exists, hindering the application of hypotheses of
modularity’s evolutionary significance to large-scale evolutionary patterns. Our research focuses on a broad, comparative study of cranial modularity and
trait integration in mammals, encompassing over 100 fossil and extant species
of monotreme, marsupial, and placental mammals. Recent projects have also examined integration in diverse vertebrate clades, including pterosaurs, birds, and amphibians.
As part of her Ph.D. research in the Committee on Evolutionary Biology at the University of Chicago, Dr. Goswami collected 3-D landmark data from more than 2000 osteological specimens and developed new analytical tools to assess landmark correlations and their relationship to phylogeny, diet, encephalization, and sequence heterochrony. Using taxa with convergent ecologies and morphologies to isolate the relationships among trait integration, phylogeny, and diet, she demonstrated that modularity does evolve during mammalian evolution, with significant differences between monotremes and therians (marsupials and placentals). The strength of within-module trait correlations varies across the skull and across taxa, probably reflecting developmental influences. The recognition that modularity is an evolving feature has great significance for understanding and testing how trait correlations influence morphological evolution.
In recent years, we have expanded this research in several ways. We have used computer simulations, based on empirical and theoretical models of modularity, and analyses of morphological disparity within modules to assess the long-standing hypothesis that trait correlations influence morphological variation and, thus, morphological evolution (with P.D. Polly, Indiana University). We have also simulated the effect of character correlations on discrete character states, with relevance for cladistic analyses. We are also developing techniques of 3-D model building and digitization to gather quantitative data from ontogenetic sequences for a broad sample of mammals. Using these new data, we are testing hypotheses of ontogenetic variation and integration and their relationship to adult morphology, particularly as it relates to the marsupial-placental dichotomy in mammal evolution.
Most recently, we have begun developing methods to analyse morphological rates of evolution using 3D landmarks and other multivariate metrics (with C. Soligo, UCL Anthropology and J. Smaers, SUNY Stony Brook).