LAUSR

Proctophyllodid feather mites (400+ species) are permanent (full-time) symbionts commonly occurring on passerine birds. Phenotypic evolution of these mites appears to be greatly influenced by characters related to reproduction (>87.5% of a total of 32 taxonomically important discrete characters) and male genitalic characters (21.9%). Because sexual selection could the major evolutionary driver in this system, we test the theoretical expectation that genitalic or sexually dimorphic characters should evolve more rapidly and divergently then other characters. We inferred a time-calibrated molecular phylogeny (6 genes, 8571 nt aligned, no missing data) for 133 taxa of proctophyllodid mites and 40 outgroups. Comparisons of the average number of character state changes inferred on 10,696 Bayesian stationary trees indicate that male genitalic or sexually dimorphic characters do not evolve significantly faster than other characters (p=0.537 and p=0.819, respectively). However, among the male genitalic characters, a trait related to the relative length of the aedeagus experienced extremely fast rates of evolution and was detected as a statistical outlier. In this character, the transitions between short, long, and several intermediate states occurred in both directions. In contrast, the evolution of extremely long aedeagi (nearly as long as the body) occurred unidirectionally and irreversibly. This surprising result may be due to constraints imposed by the female spermathecal canal, which, in species where males have extremely long aedeagi, is also very long and may impede pumping sperm by short aedeagi. In proctophyllodid mites, extremely long aedeagi evolved independently five times in five different monophyletic lineages. Several of these lineages were lumped together by taxonomists to form easy-to-distinguish but apparently artificial species-groups. Male genitalic characters, thus, can introduce false synapomorphies that could affect morphology-based phylogenetic inference. For the most species-rich genus, Proctophyllodes, we develop a predictive classification of species-groups that reconciles molecular and morphological data.