ECOMORPHOLOGICAL MODELS FOR ARBOREAL VERTEBRATES: THE CASE OF THE MARSUPIAL PHILANDER FRENATA

Abstract

Performance measurements can be used to infer the fundamental niche of species, allowing predictions of habitat or resource use in the absence of species interactions. We propose a standard set of tests to measure locomotory performance of arboreal vertebrates, and a procedure to develop ecomorphological models based on performance tests and path analysis. The proposed tests simulate habitat features such as support diameter and incline. Models of relationships between variables can be formulated, compared with path analysis, and the model of best fit chosen. This procedure was applied to the didelphid marsupial Philander frenata, comparing the effect of body shape on stride length and frequency in arboreal walking. The model including paths to stride length and frequency had a significant better fit than the model with paths only to stride length. An a posteriori model - obtained from the elimination of nonsignificant paths – suggested that the relative length of claws and tail were the more determinant of stride frequency, whereas stride length was more affected by the relative length of the limbs. These are hypotheses about an important aspect of the fundamental niche of didelphid marsupials, allowing inferences of niche similarity based on morphology. Independent data are necessary to test these hypotheses.