Rapid increase of host defence against brood parasites in a recently parasitized area: the case of village weavers in Hispaniola

Passerine hosts of brood parasitic birds usually vary in their ability to d iscriminate and reject alien eggs. Two main hypotheses have been suggested to explain the persistence of acceptor individuals in species that are expl oited by brood parasites. The evolutionary lag hypothesis postulates that s ome hosts have not yet evolved the ability to discriminate against alien eg gs. Once the ability to recognize the parasitic egg has appeared by mutatio n, it spreads because of the selective advantage of rejection. Parasites ar e then selected to produce more mimetic eggs, in order to escape host discr imination, which eventually ends up in an arms race between the parasite an d the host. The evolutionary equilibrium hypothesis is based on the assumpt ion that rejection behaviour is costly in the absence of parasitism, becaus e of recognition errors. Acceptor hosts can persist when parasitism rate fl uctuates or is consistently low Indirect evidence for costs of rejection in the absence of parasitism has been provided by Cruz & Wiley, who reported low rejection rate for a population of village weavers (Ploceus cucullatus) introduced from Africa to Hispaniola (West Indies) more than a century ago . In Africa the species is parasitized by Chyrosococcyx cuckoos and shows h igh levels of egg discrimination. Since no brood parasite was present in Hi spaniola, Cruz & Wiley suggested that rejection was selected against in the absence of parasitism due to recognition costs. Introduction of village we avers in Hispaniola, therefore, provided a unique opportunity to test the d ecline of an adaptation. During the past century the shiny cowbird (Molothr us bonariensis) has expanded its range, invading most of the West Indies fr om South America. It was first observed in Hispaniola in 1972, and it start ed to exploit village weavers as a host. Given that shiny cowbirds substant ially reduce the reproductive success of weavers, we should expect higher r ejection rates nowadays compared to those reported by Cruz & Wiley 16 years ago. In agreement with this prediction, we found a high rejection rate of cowbird model eggs (89.3%, 95% CI = 81.1-97.5%), moderate levels of rejecti on of non-mimetic weaver model eggs (67.5%, 95% CI = 52.5-82.5%) and rather low levels of rejection of mimetic weaver model eggs (25%, 95% CI = 4-46%) . The rejection rate of artificial cowbird eggs has therefore increased fro m 13.8% (95% CI = 5-22.6%) to 89.3% in 16 years. To check whether this rapi d increase in host resistance is compatible with a genetic microevolutionar y change, we built a population dynamics model where, as an upper bound, re sistance is inherited by the progeny with a probability of one. This simple model shows that observed changes of rejection rate are compatible with a genetic microevolutionary shift only under the most favourable scenario for rejecters to spread. Relaxing one or several of these assumptions (e.g. hi gh parasitism rate, absence of rejection costs) considerably lengthens the period needed for rejecters to spread. We suggest that both genetic and lea rning processes might be involved in the observed changes.