Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution.

TLDR

Heliconius butterflies are renowned for Müllerian mimicry with diverse wing patterns, and geographic variation has been used to support a Pleistocene refugium hypothesis, yet evolutionary rates and relationships among races have not been explicitly examined. This study constructs a mitochondrial DNA–based phylogeny of 14 divergent Heliconius erato races to test whether similar wing patterns have arisen rapidly and convergently. Clade ages were inferred by calibrating sequence divergence against a linear plot of mitochondrial divergence in arthropods with independently dated splits. The phylogeny shows a basal east–west Andean split, low sequence divergence with shared haplotypes across allopatric races, a simultaneous radiation within the last 200,000 years, and supports a clocklike mitochondrial evolution in arthropods when divergence is low.

Abstract

The neotropical Heliconius butterflies are famous examples of Müllerian mimicry, due to the diverse array of shared, brightly colored wing patterns that advertise the butterflies' unpalatability. The parallel geographical variation in these patterns within several widespread species has been invoked to support the controversial Pleistocene refugium hypothesis of tropical diversification. However, in no Heliconius species have either evolutionary rates or relationships among geographical races been explicitly examined. I present a phylogenetic hypothesis based on mitochondrial DNA sequences for 14 divergent races of Heliconius erato, which reveals that similar wing patterns have evolved rapidly and convergently within the species. There is a basal split between groups of races from east and west of the Andes, reflecting a vicariant separation at the base of the Pleistocene. Within each of these clades, sequence divergence is very low, and some haplotypes are shared between allopatric races with radically different wing patterns. The topology implies a simultaneous radiation of races in these two areas within the last 200,000 years. Ages for the clades are estimated by comparing sequence divergence to a plot of mitochondrial divergence in several arthropod taxa with independently dated divergence times. This plot is linear and suggests that mitochondrial DNA in arthropods evolves in a clocklike manner, at least initially, when sequence divergence is low.