Retaliatory mafia behavior by a parasitic cowbird favors host acceptance of parasitic eggs

Abstract

Why do many hosts accept costly avian brood parasitism even when parasitic eggs and nestlings differ dramatically in appearance from their own? Scientists argue that evolutionary lag or equilibrium can explain this evolutionary enigma. Few, however, consider the potential of parasitic birds to enforce acceptance by destroying eggs or nestlings of hosts that eject parasitic eggs and thereby reject parasitism. This retaliatory "mafia" behavior has been reported in one species of parasitic cuckoo but never in parasitic cowbirds. Here we present experimental evidence of mafia behavior in the brown-headed cowbird (Molothrus ater), a widely distributed North American brood parasite. We manipulated ejection of cowbird eggs and cowbird access to predator-proof nests in a common host to test experimentally for mafia behavior. When cowbird access was allowed, 56% of "ejector" nests were depredated compared with only 6% of "accepter" nests. No nests were destroyed when cowbird access was always denied or when access was denied after we removed cowbird eggs, indicating that cowbirds were responsible. Nonparasitized nests were depredated at an intermediate rate (20%) when cowbirds were allowed access, suggesting that cowbirds may occasionally "farm" hosts to create additional opportunities for parasitism. Cowbirds parasitized most (85%) renests of the hosts whose nests were depredated. Ejector nests produced 60% fewer host offspring than accepter nests because of the predatory behavior attributed to cowbirds. Widespread predatory behaviors in cowbirds could slow the evolution of rejection behaviors and further threaten populations of some of the >100 species of regular cowbird hosts.

Fig. 1.

Effect of cowbirds on rates of nest predation and reproductive output in nests of prothonotary warblers. (A) Rates of nest predation were significantly different among the five categories of nests (χ₄² = 42.22; P < 0.001). Bars shown with the same letter above them were not statistically different (P > 0.05) from each other based on pairwise comparisons using χ² tests. (B) Mean + SE numbers of warbler offspring produced per nest were significantly different among the five categories of nests (Kruskal–Wallis test; H₄ = 40.22; P < 0.001). Bars shown with the same letter above them were not statistically different (P > 0.05) from each other based on pairwise comparisons using Mann–Whitney U tests. Sample sizes for nest categories are 46, 72, 32, 16, and 16 for categories 1–5, respectively. See Table 1 and Materials and Methods for a description of each category.