Skip to main content

Burrowing Constrains Patterns of Skull Shape Evolution in Wrasses

Cite this dataset

Evans, Kory (2022). Burrowing Constrains Patterns of Skull Shape Evolution in Wrasses [Dataset]. Dryad.


The evolution of behavioral and ecological specialization can have marked effects on the tempo and mode of phenotypic evolution. Head-first burrowing has been shown to exert powerful selective pressures on the head and body shapes of many vertebrate and invertebrate taxa. In wrasses, burrowing behaviors have evolved multiple times independently, and are commonly used in foraging and predator avoidance behaviors. While recent studies have examined the kinematics and body shape morphology associated with this behavior, no study to-date has examined the macroevolutionary implications of burrowing on patterns of phenotypic diversification in this clade. Here, we use three-dimensional geometric morphometrics and phylogenetic comparative methods to study the evolution of skull shape in fossorial wrasses and their relatives. We test for skull shape differences between burrowing and non-burrowing wrasses and evaluate hypotheses of shape convergence among the burrowing wrasses. We also quantify rates of skull shape evolution between burrowing and non-burrowing wrasses to test for whether burrowing constrains or accelerates rates of skull shape evolution in this clade. We find that while burrowing and non-burrowing wrasses exhibit similar degrees of morphological disparity, for burrowing wrasses, it took nearly twice as long to amass this disparity. Furthermore, while the disparities between groups are evenly matched, we find that most burrowing species are confined to a particular region of shape space with most species exhibiting narrower heads than many non-burrowing species. These results suggest head-first burrowing constrains patterns of skull shape diversification in wrasses by potentially restricting the range of phenotypes that can perform this behavior.


Three-dimensional meshes were collected via micro-CT scanning for all 184 specimens. Shape data were collected using three dimensional geometric morphometrics.

Usage notes

All analyses were run in R and data files are saved as .R files