The Insurance-Egg Hypothesis and Extra Reproductive Value of Last-Laid Eggs in Clutches of American Kestrels

Abstract

Many species of birds lay more eggs than they normally rear to independence. This is illustrated well by species with obligate siblicide, such as eagles (Meyburg 1974) and boobies (Anderson 1990), where the youngest nestling in a brood is killed by its older sibling in a physical struggle soon after hatching. A common explanation for such behavior is that parents can raise only one offspring, but lay the extra egg for insurance against infertility (the insurance-egg hypothesis; Dorward 1962); the surplus offspring is eliminated if its older sibling is viable. Field experiments testing the insurance-egg hypothesis have been performed for a few species with obligate siblicide (e.g. Cash and Evans 1986, Anderson 1990). Until recently, the insurance value of offspring was not considered to be important for birds with facultative siblicide (i.e. where survival of youngest offspring is conditional on prevailing environmental conditions). Such offspring usually were thought to be an adaptation to an environment where food was unpredictable at the time of laying. Parents could rear the entire brood if food proved to be abundant, but if food was scarce, they could eliminate the extra offspring through brood reduction (Lack 1947, Ricklefs 1965; called by Forbes 1991). Mock and Parker (1986) emphasized that the youngest nestling in a brood may serve two simultaneous functions in species with facultative brood reduction: resource tracking and insurance. The costs and benefits of insurance offspring have been modelled in a general way by Forbes (1990, 1991); extra offspring may provide multiple benefits simultaneously (Mock and Forbes 1995). Here, I examine the value of the lastlaid egg in broods of American Kestrels (Falco sparverius), a species with facultative brood reduction. Methods.-From 1988 to 1992, coworkers and I studied a population of about 200 pairs of American Kestrels breeding in nest boxes at Besnard Lake, Saskatchewan (55?20'N, 106?W). Details about the study site and nest boxes are summarized in Bortolotti et al. (1991) and Bortolotti (1994). Each year, adults in the population were trapped, banded, and measured (see Wiebe and Bortolotti 1993); by visiting boxes regularly, we recorded the reproductive performance of most parents. In many nests, eggs were numbered according to laying sequence with nontoxic felt markers, and nestlings were similarly marked according to hatching order (Wiebe and Bortolotti 1995a, b). The