Successes and Failures in Avian Egg and Nestling Recognition With Comments on the Utility of Optimality Reasoning

Abstract

SYNOPSIS. Interactions between parasitic birds and their hosts provide some of the best examples of recognition that has obvious and strong adaptive value. But these interactions also involve numerous failures of recognition that result in enormous losses in fitness. The failures are more significant than the successes because they are exceptions to basic evolutionary theory which predicts that organisms behave optimally. The egg recognition behavior of North American passerines separates species into two discrete groups. In “rejecter” species nearly all individuals are intolerant of nonmimetic parasitic eggs. In “accepters” nearly all individuals tolerate such eggs. Rejecters possess a true recognition of their own eggs, apparently learning this in an imprinting-like process. No adaptive explanation accounts for the lack of rejection in accepter species. Many of these species are victimized heavily by parasitic cowbirds and have eggs divergent from those of the cowbird. They readily accept clutches containing only cowbird eggs. But accepters do show egg recognition in some contexts. They abandon clutches whose total volume of eggs falls below a critical value; they eject and visually discriminate among eggs that are broken, thus showing that they are capable of the critical behaviors needed to reject parasitic eggs. Since most accepters occur in taxa that also contain rejecters, I propose that stochastic processes are largely responsible for their nonoptimal behavior. Even though many birds reject parasitic eggs, all or nearly all accept nonmimetic parasitic nestlings, a situation for which there is no reasonable optimality explanation. Although these recognition studies demonstrate the value of optimality reasoning because they are based on clearly defined predictions arising from the expectation of adaptiveness, they also demonstrate some general problems inherent in optimization models. It is argued here that while optimality reasoning is the best and most all-inclusive paradigm in biology, its value will be enhanced if its limitations are recognized. When warranted, conclusions of nonoptimality should be stressed and optimality itself should be tested and not employed only as an assumption used to formulate hypotheses.