Reduced parasitism by retaliatory cuckoos selects for hosts that rear cuckoo nestlings

We present a model to investigate why some bird species rearthe nestlings of brood parasites in spite of suffering largereductions in their own immediate fitness. Of particular interestis the case in which hosts rear only the parasite's young, allof their own offspring having been ejected or destroyed by theparasite. We investigate the conditions for the evolution ofretaliation by brood parasites against hosts that eject theiryoung, as well as the evolution of nonejection by hosts. Retaliationby cuckoos can evolve, despite potentially benefiting otherbrood parasites, if rates of ejection by hosts are neither toohigh nor too low, and if depredated nests are reparasitizedat a high rate by the depredating cuckoo. The presence of aretaliatory cuckoo then eases the conditions for the evolutionof hosts to accept and rear cuckoo offspring. A key conditionfavoring the evolution of non-ejection is that nonejectors enjoylower rates of parasitism in later clutches compared to ejectors.This requires that cuckoos reparasitize the clutches of ejectorsat relatively high rates and that nonejectors can rear a clutchof their own following the rearing of a cuckoo nestling. Ifthese conditions are not met, it pays hosts to eject cuckoonestlings even if the cuckoo retaliates. The model can explainwhy nonejection is relatively easy to evolve in cases in whichthe host young are reared alongside those of the cuckoo, suchas in cowbirds, and shows how hosts can resist invasion by parasiticcuckoos. The model predicts that retaliatory brood parasitessuch as the cuckoo have good memory for the location and statusof nests in their territory. Hosts of retaliatory cuckoos whosenestlings destroy the host clutch are predicted to have longbreeding seasons or the ability to attempt more than one clutchper season. Our model of retaliation may have wider applicationsto host-parasite relationships, virulence, and immunity.