A PUBLICATION OF THE COOPER ORNlTHOLOGICAL SOCIETY

-I examined three factors that may limit the clutch size of the Glaucous-winged Gull (Larus gluucescens) to three or fewer eggs: the energetic cost of egg formation, the shelf-life of eggs, and the incubation capacity of adults. Incubation capacity was found to have a significant effect on the success of large clutches but it cannot explain the absence of 4-egg clutches. Energetic limitation following the initiation of the clutch may be a more important factor limiting clutch size to three. I examined several aspects of the brood reduction hypothesis to determine whether the presence of brood reduction adaptations is compatible with evidence that clutch size is not limited by the brood rearing capacity of the adults. Asynchronous hatching was found to be beneficial regardless of the number of young that could be raised and thus is consistent with evidence that brood-rearing capacity does not limit clutch size. The small size of the third egg, generally considered to be another brood reduction adaptation, was found to be a result of energetic shortages during laying and thus may not be an adaptive mechanism for brood reduction. Members of the family Laridae exhibit modal clutch sizes ranging from 1 to 3 eggs, 4-egg clutches occurring infrequently (see Conover 1984). Some of the 4-egg clutches reported are produced by female-female pairs (Conover 1984). Clutchsize distributions with modal and maximal values of 3 eggs are found in at least 11 larid species: the Herring Gull (Larus argentatus), Laughing Gull (L. atricillu), Glaucous-winged Gull (L. gluucescens), Common Tern (Sterna hirundo), California Gull (L. culifornicus), Lesser Blackbacked Gull (L. fuscus), Black-headed Gull (L. ridibundus), Common Gull (L. cunus), Western Gull (L. occident&), Ring-billed Gull (L. deluwurensis), and Great Black-headed Gull (L. ichthyuetus) (Samorodov and Ryabov 1969, Conover 1984). This group encompasses a diverse range of both body sizes and feeding habits, though most species are relatively large and all breed in temperate zones. The reason for the truncation of the clutch size frequency distribution at three eggs is not clear. Three factorsthe energetics of egg formation, incubation capacity, and brood rearing capacity-have received attention as factors potentially limiting clutch size to three or fewer eggs. The energetic cost of egg formation is thought to explain patterns of variation in larid egg size, clutch size, and nesting phenology (Nisbet 1973, 1977, Mills 1979, Pierotti 1982, Schreiber et al. 1979, Mills and Shaw 1980, Boersma and Ryder 1983, Houston et al. 1983, Winkler 1983, 1985, Pierotti and Bellrose 1986). The energetic cost of egg formation, however, does not place a strict upper limit on egg production at 3 eggs because protracted laying can be induced in at least 4 of the 11 species exhibiting a truncated clutch-size frequency distribution (Herring Gull: Paludan 1951, Harris 1964, Parsons 1976, Pierotti 1982; I Department of Zoology NJ-15, University of Washington, Seattle, Washington 98195. Glaucous-winged Gull: this study; California Gull: Winkler 1983, 1985; Black-headed Gull: Weidmann 1956). Similarly, the incubation capacity of gulls and terns may not impose a fixed upper limit on clutch size. Most large gulls have three brood patches (see Table 10) and it is possible that this broodpatch configuration results in a 3-egg limit (Vermeer 1963, Pierotti and Bellrose 1986). There are no studies, however, that support this hypothesis. Experimental manipulation of clutch size during incubation has shown that more chicks hatch from artificially enlarged clutches than from 3-egg clutches (Coulter 1973a, b). There is also no evidence that clutch size is limited to 3 or fewer eggs by the brood-rearing capacity of the adults. In at least 4 species, adults are capable of rearing more than 3 young (Herring Gull: Haymes and Morris 1977; Glaucouswinged Gull: Vermeer 1963, Ward 1973; Lesser Black-backed Gull: Harris and Plumb 1965; Western Gull: Coulter 1973b). To further complicate the question of clutchsize determination, many larids possess traits that potentially conflict with the observation that gulls are capable of raising more than 3 young. Two traits characteristic of all larids with 3-egg clutches are the presence of asynchronous hatching, and size reduction of the third (c-) egg relative to the first 2 (aand b-) eggs. These traits are frequently cited as evidence supporting the brood reduction hypothesis (Lack 1968, O’ Connor 1978, Clark and Wilson 198 1, Hahn 198 1, Slagsvold et al. 1984). Asynchronous hatching and the small size of the c-egg place the third chick at a disadvantage with respect to its siblings in competition for food. During years of food shortage, these traits are thought to facilitate the early mortality of chicks that could not be raised, thereby increasing food available to the surviving chicks. The presence of brood reduction traits is interpreted as circumstantial evidence that