Body Size Evolution in Snakes: Evidence from Island Populations
The current literature reports divergent conclusions on the patterns of body size change in island snakes. I reviewed body size data in the published literature and tested the effects of island biogeographic variables on such changes. I found that none of the physiographic variables (island area, island age, distance to mainland, and latitude) was important in determining changes in size of island snakes. Additionally, a current hypothesis of phylogenetic history had no effect on changes in body size. Rather, the proportional change in body size of island snakes was bimodal, consistent with a diet alteration hypothesis that suggests that snake body size is principally influenced by prey size and that island snakes encounter prey that are larger or smaller in size compared with those on the mainland. Also, snakes that became small on islands did so to a relatively greater degree than those that became large. Ontogenetic changes in foraging strategies appeared to explain this pattern. The distribution of gigantic and dwarf snake populations on islands differed significantly between the families Viperidae and Colubridae. The foraging style of colubrids, specifically nest-robbing behavior, may predispose these species to become larger on islands. Numerous colubrid (and one elapid) species attain their largest sizes on islands that also support nesting seabirds, whereas dwarfed populations consume mainly squamates.Abstract
Phylogenetic hypothesis of the relationships among snake taxa used in this study. The branching pattern is based on published molecular phylogenies (Lawson, 1987; Kluge, 1991; Densmore et al., 1992; De Queiroz and Lawson, 1994; Nilson et al., 1994; Heise et al., 1995; Lopez and Maxson, 1995; Keogh, 1998; Crother, 1999; Parkinson, 1999; Rodriguez-Robles and DeJesus-Escobar, 1999; Parkinson et al., 2000; Murphy et al., in press). The graph indicates mean percent size change for each species: y-axis = mean percent change in maximum size on island compared with maximum size of mainland population. For species with multiple island populations, the range in percent size change (about the mean) is indicated with T-lines
Path diagram generated from the analysis of 30 snake species and the predictor variables listed in Appendix 1. Correlations between predictor variables are to the left of these variables, path coefficients are to the right. Bolded double-headed arrows and * indicate significant correlations (with Bonferoni adjustment P < 0.003). Each direct path (depicted by one-headed arrows) has two values. The first value is the unadjusted path coefficient, and the second is the “phylogeny-free” path coefficient using network autocorrelation. Solid lines indicate positive paths; dotted lines indicated negative paths. U represents the path for unexplained variation
Frequency distribution of island body size changes of the 30 snake species used in the study. A = all species, B = Viperidae, C = Colubridae