The evolution of demographic tactics in lizards: a test of some hypotheses concerning life history evolution

We analyze, with an augmented data base, patterns of covariation of the three primary demographic parameters (age at maturity, fecundity, adult survival, all measured in the same unit of time) in lizards. This also constitutes a first attempt to use all three of these parameters for this group of species. We attempt to place these analyses in the framework of recent theories on life history evolution (Ferrière and Clobert, 1992; Charnov, 1993). Life history data were collected from the literature and from our original work, and a composite phylogeny was assembled, based on a variety of published sources. Using a phylogenetically based statistical method (independent contrasts), the allometric (log-log) relationship of fecundity (and of clutch size) in relation to snout-vent length was found to differ significantly between the two major clades of extant lizards, Iguania (43 species in our data set) and Scleroglossa (47 species). We therefore emphasize analyses done separately for the two clades. Without removing correlations with body size, the relationships between fecundity and survival, and between fecundity and age at maturity, were also found to differ between clades, which differs from Charnov’s (1993) predictions. When correlations with body size were removed statistically, however, the two clades did not differ significantly in these relationships. In a principal components analysis (PCA) of the three demographic variables plus snout-vent length, the first axis explained the majority (53–57%) of variation in both clades, while the second axis explained 27–31% of the variation and loaded mainly on fecundity. In a PCA of size-adjusted demographic variables residuals (from log-log regressions on snout-vent length), the first axis explained 66–68% of the variation and was clearly interpretable as the classical ‘‘slow-fast’’ continuum, which has been described in birds and mammals. The PCA of residuals did not provide clear evidence of additional significant patterns of covariation. However, the rate of evolution of