Skip to main content
Dryad

Data from: Classification tree methods provide a multifactorial approach to predicting insular body size evolution in rodents

Cite this dataset

Durst, Paul A. P.; Roth, V. Louise (2011). Data from: Classification tree methods provide a multifactorial approach to predicting insular body size evolution in rodents [Dataset]. Dryad. https://doi.org/10.5061/dryad.5v75d33g

Abstract

Many hypotheses have been proposed to explain size changes in insular mammals, but no single variable suffices to explain the diversity of responses, particularly within Rodentia. Here in a dataset on insular rodents we observe strong consistency in the direction of size change within islands and within species, but (outside of Heteromyidae) little consistency at broader taxonomic scales. Using traits of islands and of species in a classification-tree analysis we find the most important factor predicting direction of change to be mainland body mass (large rodents decrease, small ones increase); other variables (island climate, number of rodent species, and area) were significant although their roles as revealed by the classification tree were context-dependent. Ecological interactions appear relatively uninformative, and on any given island the largest and smallest rodent species converged or diverged in size with equal frequency. Our approach provides a promising framework for continuing examination of insular body-size evolution.

Usage notes