CLUTCH SIZE, BREEDING SUCCESS, AND PARENTAL SURVIVAL IN THE TREE SWALLOW (IRIDOPROCNE BICOLOR).

Abstract

Lack (1947) hypothesized that clutch size in nidicolous birds has evolved by natural selection to correspond with the maximum number of young that, on average, the parents can feed. Although the hypothesis gained wide acceptance in subsequent years, the evidence is equivocal and inconsistencies remain (Klomp, 1970; Cody, 1971; von Haartman, 1971; Hussell, 1972). Those cases in which the most productive brood size is larger than the most common do not support the implied concept of direct limitation of clutch size by food supply. Furthermore, the interpretation of brood manipulation experiments that support Lack's hypothesis is open to question, since the results do not distinguish between food supply limits in the environment and possible adaptive limits upon parental feeding behavior (Cody, 1971; Hussell, 1972). Mountford (1968) suggested that incorrect formulation of predictions was responsible for some apparent contradictions. Lack recognized that selection should favor the clutch size that maximizes fitness, but overemphasized the direct influence of environmental factors upon clutch size as a single trait. A more comprehensive perspective incorporates interactions between clutch size and other life history features as well (Fisher, 1958; see review in Stearns, 1976). Pertinent models predict a most common clutch that is smaller than the most productive, under the assumption that rearing larger broods places greater stress upon parents and reduces their chances of surviving to breed again (Williams, 1966; Charnov and Krebs, 1974). The assumption of a trade-off between clutch size and adult survivorship within populations is largely untested. Observations from naturally occurring brood sizes have not yielded any consistent relationship between clutch size and parental survival (e.g., Kluyver, 1963; Perrins, 1965; Lack, 1966, p. 109). However, since variation in parental ability (e.g., efficiency in gathering food for egg formation or for feeding nestlings) may contribute to adaptive modification of clutch size (cf. Klomp, 1970), the search for such a relationship is confounded. If parents that normally initiate larger clutches are more capable of rearing them, the young in.those broods are not necessarily disadvantaged (cf. Perrins and Moss, 1975), nor are those parents necessarily less likely to survive than parents raising smaller broods. Greater weight losses among parents with larger broods (Hussell, 1972; Winkel and Winkel, 1976; Bryant, 1979) and lower probabilities of initiating a second brood after a large first brood (Kluyver, 1963; Pinkowski, 1977) provide indirect evidence that rearing large broods is stressful physiologically, but have yet to be linked to differential survival. Recently, Bryant (1979) found differences in survival between singleand double-brooded female House Martins (Delichon urbica), but these differences were not related to the brood sizes reared. Experimental approaches to testing the hypothesis that birds adaptively limit clutch size for the sake of enhanced survival are likely to provide the strongest inferences (cf. Ricklefs, 1973, p. 426; Stearns, 1976, p. 42). Artificial manipulation can extend brood sizes beyond limits currently observed within a population. However, if parental ability is reflecterd in individulm] clutch sizes. nqirpnts 278