PHYLOGEOGRAPHY OF THE JUMPING SPIDER HABRONATTUS PUGILLIS (ARANEAE: SALTICIDAE): RECENT VICARIANCE OF SKY ISLAND POPULATIONS?

Open Access
Publisher Website
Google Scholar
Abstract
In island systems with diverging populations, the history of island formation and genealogical estimates of divergence dates can be mutually informative. In the "sky islands" of southeastern Arizona, climate-induced contraction of woodlands appears to have fragmented populations of woodland-dwelling species onto disjunct mountain ranges. Montane populations of the jumping spider, Habronattus pugillis, display striking amounts of phenotypic divergence among ranges. Paleoclimatic estimates date woodland fragmentation at approximately 10,000 years ago, suggesting that phenotypic divergence has been extraordinarily rapid in these spiders. This phylogeographic study of populations of H. pugillis attempts to clarify the species' history of isolation and divergence and to address the suitability of available paleoclimatic data for dating divergences among populations of the region's woodland-dwelling organisms. Mitochondrial sequence data of spiders from 13 mountain ranges was used to reconstruct genealogical relationships. Gene trees show that small mountain ranges tend to have populations whose sequences form monophyletic groups, whereas larger ranges do not. Paraphyly among genes from larger ranges could result from either recent migration or incomplete lineage sorting. I use phylogenetic and geographic information to test these alternatives, and conclude that incomplete lineage sorting best explains the observed paraphyly. Gene trees are concordant with some of the predictions of vegetation history generated by examination of topography. Dates estimated for divergence of populations vary from 30,000 years to more than 2 million years ago, suggesting multiple vicariance events that are older than would be inferred from paleoclimatic studies. These findings illustrate that use of any single paleontological dataset to calibrate molecular clocks can potentially greatly underestimate actual divergence times.