Molecular phylogeny and systematics of the Pieridae (Lepidoptera: Papilionoidea): higher classification and biogeography

Abstract

The systematic relationships of the butterfly family Pieridae are poorly understood. Much of our current understanding is based primarily on detailed morphological observations made 50-70 years ago. However, the family and its putative four subfamilies and two tribes, have rarely been subjected to rigorous phylogenetic analysis. Here we present results based on an analysis of molecular characters used to reconstruct the phylogeny of the Pieridae in order to infer higher-level classification above the generic level and patterns of historical biogeography. Our sample contained 90 taxa representing 74 genera and six subgenera, or 89% of all genera recognized in the family. Three complementary approaches were employed: (1) a combined analysis of a 30 taxon subset for sequences from four gene regions, including elongation factor-1 alpha (EF-1α), wingless, cytochrome oxidase subunit I (COI), and 28S (3675 bp, 1031 parsimony-informative characters), mainly to establish higher-level relationships, (2) a single-gene analysis of the 90 taxon data set for sequences from EF-1α (1066 bp, 364 parsimony-informative characters), mainly to establish lower-level relationships, and (3) an all available data analysis of the entire data set for sequences from the four genes, to recover both deep and shallow nodes. Analyses using maximum parsimony, maximum likelihood and Bayesian inference provided similar results. All supported monophyly for the four subfamilies but not for the two tribes, with the Anthocharidini polyphyletic and the Pierini paraphyletic. The combined and all available data analyses support the following relationships among the subfamilies: ((Pseudopontiinae + Dismorphiinae) + (Coliadinae + Pierinae)), corroborating Ehrlich's 1958 phenetic hypothesis. On the basis of these analyses, and additional morphological and life history evidence, we propose a reclassification of the subfamily Pierinae into two tribes (Anthocharidini s.s., Pierini s.s.) and two informal groups (Colotis group, Leptosia), with the tribe Pierini s.s. subdivided into three subtribes (Appiadina, Pierina, Aporiina) and three genera (Elodina, Dixeia, Belenois) of uncertain status (incertae sedis). The combined and all available data analyses support the following relationships among the Pierinae: (Colotis group + Anthocharidini s.s. + Leptosia + (Elodina + ((Dixeia + Belenois) + Appiadina + Pierina + Aporiina))). Application of a molecular clock calibrated using fossil evidence and semiparametric rate smoothing suggests that divergence between the Pierina and Aporiina occurred no later than the Palaeocene (> 60 Myr). The minimum estimate for the age of the crown-group of the Pieridae was 112-82 Myr, with a mean of 95 Myr. A historical biogeographical hypothesis is proposed to explain the present-day distribution of the clade Pseudopontiinae + Dismorphiinae, which argues for an origin of the two subfamilies in western Gondwana (Africa + South America) during the Late Cretaceous.