Contrasting levels of extra-pair paternity in mainland and island populations of the house sparrow (Passer domesticus): is there an 'island effect'?

Despite the many studies that have investigated the genetic mating system of socially monogamous birds, very little is known about the underlying causes of extra-pair paternity and few studies have attempted to test those hypotheses which have been suggested. This study describes the analysis of the genetic mating system of two populations of the house sparrow (Passer domesticus), and uses the results from four other populations to test existing hypotheses using an intra-specific comparative approach. The parentage analysis was conducted using a combination of published and newly presented microsatellite loci isolated from the house sparrow. One population in Kentucky, U.S.A. was found to contain what may be considered to be a typical level of extra-pair paternity for this species (10.5%, 19/ 185 offspring). The second, a population on the island of Lundy, UK, exhibited a ver) low level (1.3%, 4/305 offspring), significantly lower than that in all the other populations studied so far. The finding of such diverse rates of extra-pair paternity, along with the existing estimates from other populations, has allowed us to test the effects of breeding density and genetic variation on the level of extra-pair paternity. We found no effect of either factor on the frequency of extra-pair paternity in the house sparrow, leaving the cause of this variation open to fresh ideas. (C) 1999 The Linnean Society of London