SEXUAL COMPETITION AMONG BROTHERS MAY INFLUENCE OFFSPRING SEX RATIO IN SNAKES

Most species of animals produce approximately equal numbers ofsons and daughters, apparently because the individual fitness ofparents will usually be maximized by equal investment into offspring of each sex (Fisher, 1930; see review by Charnov, 1982). Among taxa with genotypic sex determination, the main departures from a 50150 sex ratio in the progeny are in small populations where dispersal is limited, so that sibs compete with each other either for mating or resources (e.g., Hamilton, 1967, 1979; Clark, 1978; Werren, 1980; Bulmer and Taylor, 1980; Wilson and Colwell, 1981). If sons compete with each other whereas daughters do not, natural selection may favor genes that cause mothers to produce an excess of daughters. This phenomenon is well-known in insects and mites (Hamilton, 1967; Waage, 1982), and possible examples have been described in mammals (Clutton-Brock and Albon, 1982; Cockburn et aI., 1985). Our data on adders show that brothers compete for matings, that female adders produce more daughters than sons, that the variance in litter sex ratios is less than expected under the binomial distribution, and that females facultatively adjust their progeny's sex ratio in a manner predicted by sib-competition theory. Sib-competition (or 'local mate competition') should influence parental investment into sons versus daughters only in small isolated populations, because only in such cases can a single female's offspring influence the population's adult sex ratio (e.g., Hamilton, 1967, 1979; Charnov, 1982). Theoretical models predicting female-biased sex ratios also require occasional dispersal and the founding of new populations by dispersing individuals, but the predicted equilibrium sex ratio is affected only slightly by the number of generations of within-population reproduction prior to dispersal (Bulmer and Taylor, 1980). Adders (Vipera berus) appear to fulfil the assumptions of these models. Even in large ares of suitable habitat, these small venomous snakes are highly philopatric, occurring in small sub-populations with only occasional interchange of individuals (Vitannen, 1967; Vozenilek, 1991; Madsen, unpubl. data). For example, during seven years of study in the Genarp region of Sweden, we have recorded only a single instance of inter-population migration among three small populations ofadders (each consisting of 8 to 40 adult snakes) using hibernacula only 500 m apart. We studied a population of adders at Smygehuk, in extreme southern Sweden (55°20'N, 13°22'E: see Madsen, 1988 for a description of the study area). The area contained an average of approximately 30 reproducing adult snakes (range 21 to 46, of which around one-third were females) over the period 1984 to 1990. The Smygehuk adders are separated by > 20 km from any other adder populations, and our mark-recapture and radiotelemetric studies reveal no evidence of any dispersal to or from the local population during the last seven years. Models oflocal mate competition rely upon significant migration rates among populations, a condition that is not currently satisfied by the Smygehuk adders. However, the isolation of this population is relatively recent, and is due to agricultural activities. We infer that the adder populations in this region have evolved under conditions (small, semi-isolated populations) that fulfil the assumptions oflocal-mate-competition models, but recognize that this inference is an important caveat to our conclusions. Since 1984, one ofus (T.M.) has observed the snakes on every sunny day during the three-week mating period each spring. Adult females reproduce, on average, every second year, but adult males reproduce every year (e.g., Prestt, 1971; Nilson, 1981). All reproductive females have been followed using radiotelemetry, and all adult snakes have been scale-clipped and paintmarked for individual identification. Gravid females are collected in late summer, a week or two prior to parturition, and released (with their offspring) after the young have been born, weighed, measured, sexed, and marked. The gender of neonatal adders can be reliably determined by dorsal coloration, shape ofthe tail-base, and relative tail length (female tails -s13% of snoutvent length, males ~ 14%). Adult male adders engage in vigorous combat bouts during the mating season, with larger males driving smaller individuals away from receptive females (Andren, 1986; Madsen, 1988). Hence, there is the potential for brothers to compete with each other for mating opportunities. We have observed combat between brothers in each of the last two mating seasons (1989, 1990), the first seasons since marked snakes ofknown parentage have attained sexual maturity and, thus, the first in which we could recognize brothers. Two 'pairs' of male sibs were present as adults in 1989, and one 'pair' in 1990. Brother-brother combat was recorded