• Butterfly wing patterns;
  • developmental constraints;
  • genetic correlations;
  • nymphalid groundplan;
  • quantitative genetics


Developmental processes exert their influence on the evolution of complex morphologies through the genetic correlations they engender between traits. Butterfly wing color patterns provide a model system to examine this connection between development and evolution. In butterflies, the nymphalid groundplan is a framework used to decompose complex wing patterns into their component pattern elements. The first goal of this work has been to determine whether the components of the nymphalid groundplan are the products of independent developmental processes. To test this hypothesis, the genetic correlation matrices for two species of butterflies, Precis coenia and Precis evarete, were estimated for 27 wing pattern characters. The second purpose was to test the hypothesis that the differentiation of serial homologs lowers their genetic correlations. The “eyespots” found serially repeated across the fore- and hindwing and on the dorsal and ventral wing surfaces provided an opportunity to test this hypothesis. The genetic correlation matrices of both species were very similar. The pattern of genetic correlation measured between the different types of pattern elements and between the homologous repeats of a pattern element supported the first hypothesis of developmental independence among the elements of the groundplan. The correlation pattern among the differentiated serial homologs was similarly found to support the second hypothesis: pairs of eyespots that had differentiated had lower genetic correlations than pairs that were similar in morphology. The implications of this study are twofold: First, the apparent developmental independence among the distinct elements of wing pattern has facilitated the vast diversification in morphology found in butterflies. Second, the lower genetic correlations betweendifferentiated homologs demonstrates that developmental constraints can in fact be broken. The extent to which genetic correlations readily change, however, remains unknown. © 1994 Wiley-Liss, Inc.