Sexual selection and the geography of Plasmodium infection in Savannah sparrows (Passerculus sandwichensis)

According to Hamilton and Zuk's hypothesis of parasite-mediated sexual sele ction, host-parasite coevolution maintains variation in male genetic qualit y and allows for strong intersexual selection in species with high rates of infection. In birds, most interspecific tests of this hypothesis relate th e prevalence of blood para sites to some measure of the intensity of sexual selection. Such tests often rely on limited sampling of single populations to estimate species-wide infection rates, and many tests are thus vulnerab le to intraspecific (geographic) variation in the evolutionary ecology of d isease. Here, we used molecular techniques to examine variation in the prev alence of Plasmodium spp. across 14 populations of Savannah sparrows, Passe rculus sandwichensis, in eastern North America. Plasmodium could not be det ected in any of 68 island birds, but 34 of 119 (29%) mainland males, and 7 of 43 (16%) mainland females were infected. Among mainland birds, infection was common in southern populations but rare in New Brunswick, Canada. Over all, the prevalence of Plasmodium ranged from 0 to 60% across populations, although only 17.8% of birds were infected in the pooled (species-wide) sam ple. The extent of this geographic variation suggests that limited sampling of single populations is unlikely to yield accurate estimates of species-w ide infection rates. However, among mainland Savannah sparrows, the prevale nce of malaria correlated strongly with average male size and the degree of sexual size dimorphism. We speculate that either sexual selection leads to male-biased infection or, conversely, that high rates of infection promote the evolution of strong intersexual selection.