Comparative Methods at the Species Level: Geographic Variation in Morphology and Group Size in Grey-crowned Babblers (pomatostomus Temporalis).

We show that a new comparative method that sheds light on evolutionary trends among species may also illuminate trends within species. This finding comes from a phylogenetic autocorrelation analysis of morphological traits among individuals sampled from ten populations of a cooperatively breeding songbird, the Grey-crowned Babbler (Pomatostomus temporalis). Highly variable mitochondrial DNA (mtDNA) sequences from both the eastern (Pomatostomus temporalis temporalis) and western (Pomatostomus temporalis rubeculus) lineages were used to define genetic distances among 120 individuals and to estimate correlations among individuals in wing length, tarsus length, and body weight via an intraspecific weighting matrix. The autoregressive model effectively removed intraspecific correlations for all three morphological variables, and the proportion of the total phenotypic variance due to genealogical relationships varied from 0.68 (weight) to 0.23 (tarsus). The analysis revealed correlations among the specific components of traits, in which none were previously detected (type-I error) and diminished correlations that appeared strong when phylogeny was ignored. Group size was the only trait for which the autoregressive model failed to remove intraspecific correlations, a result likely due to the plasticity, convergence, and clinal variation in this trait in both the eastern and western lineages. The autocorrelation analysis weakened significant negative correlations between group size and total values for wing length and body weight, but the interpretation of this result depends on the adaptive significance ascribed to the "phylogenetic component" of trait values removed by the analysis. Comparative methods employing distance matrices are one useful way of summarizing the pattern of nonhierarchical relationships among conspecific individuals ("tokogenetic" relationships, sensu Hennig).