SPACE-FREE CORRELATION BETWEEN MORPHOMETRIC AND CLIMATIC DATA - A MULTIVARIATE-ANALYSIS OF AFRICANIZED HONEY-BEES (APIS-MELLIFERA L) IN BRAZIL

Morphometric patterns of Africanized honey bees and their relationship with climatic variation from forty Brazilian localities were investigated using a set of multivariate analyses. There is a strong canonical correlation between morphometric and climatic data, suggesting that adaptation and natural selection are occurring. However, both morphometric and climatic data were spatially autocorrelated and causal explanations can be wrong if the influence of space is not removed. In this paper, we removed this effect using a Canonical Trend Surface Analysis (CTS). We then recomputed the canonical correlation between the two sets of data using the residuals of each CTS. The results of this procedure, which we called space-free canonical correlation, show no significant relationship between morphometric and climatic data. Instead, other explanations for morphometric variation of Africanized honey bees were considered. The historical records of the migration of these bees, together with spatial analysis of morphological and molecular data suggest that the most parsimonious explanation is related to a demic diffusion process.