Phylogenetic analysis of coadaptation in behavior, diet, and body size in the African antelope

Citation
Js. Brashares et al., Phylogenetic analysis of coadaptation in behavior, diet, and body size in the African antelope, BEH ECOLOGY, 11(4), 2000, pp. 452-463
Citations number
87
Categorie Soggetti
Animal Sciences","Neurosciences & Behavoir
Journal title
BEHAVIORAL ECOLOGY
ISSN journal
10452249 → ACNP
Volume
11
Issue
4
Year of publication
2000
Pages
452 - 463
Database
ISI
SICI code
1045-2249(200007/08)11:4<452:PAOCIB>2.0.ZU;2-#
Abstract
Several authors have suggested that African antelope (family Bovidae) exemp lify coadaptation of ecological, behavioral, and morphological traits. We t ested four hypotheses related to the ecology and behavior of 75 species of African antelope using both conventional statistical techniques and techniq ues that account for the nonindependence of species by considering their ph ylogenetic relationships. Specifically, we tested the hypotheses that (1) d ietary selectivity is correlated negatively with body mass, (2) dietary sel ectivity is correlated negatively with group size, (3) gregarious species e ither flee or counterattack when approached by predators, but solitary and pair-living species seek cover to hide, and (4) body mass and group size ar e correlated positively. Each of these hypotheses was examined for the glob al data set (family Bovidae) and, when possible, within the two antelope su bfamilies (Antilopinae and Bovinae) and within 7 of the 10 antelope tribes. The results of our conventional and phylogenetically corrected analyses su pported the hypotheses that group and body size vary predictably with feedi ng style and that antipredator behavior varies with group size. The hypothe sis that body mass and group size are correlated positively was supported b y conventional statistics, but these two traits were only weakly related us ing a phylogenetically corrected analysis. Moreover, qualitative and quanti tative comparisons within each of the eight major African antelope tribes g enerally gave little support for the four hypotheses tested. Thus, although our analyses at the subfamily level provided results that were consistent with prior hypotheses, our analyses at the level of tribes were equivocal. We discuss several possible explanations for these differences.