Data from great egrets and great blue herons were used to test a fundamental assumption of Lack's brood-reduction hypothesis, that mortality is brood-size dependent. This was confirmed for the largest brood sizes (4 and 3), which, in egrets, also have the highest sib-fighting rates. Broods of one, however, experienced paradoxically high mortality, especially early in the season. The hypothesis is advanced that parents desert unprofitably small broods when sufficient time remains for production of a larger brood. A simple game-theory model shows that this parental desertion may hinge primarily on the overall costs of renesting. Egret brood reduction caused by sibling aggression (siblicide) occurred later than less aggressive forms of brood reduction. The inclusive fitness of senior broodmates is maximized by the successful fledging of all sibs, and the physical superiority of seniors (in food-handling for herons; food-handling and aggression for egrets) usually suffices to guarantee their own welfare in brood competitions. Finally, it is shown that the last chick in asynchronously hatching broods represents two kinds of reproductive value (RV) to the parents-"extra RV" (obtained despite the survival of elder sibs) and "insurance RV" (obtained only when at least one elder sib dies first)-which can be distinguished from field data. This approach can be used in comparisons with other asynchronous species for partitioning the fitness contributions of marginal offspring.
© 1986 The Society for the Study of Evolution.