Sexual selection for spatial-learning ability

Sex differences in spatial learning have been thought to be universal among mammals, but their adaptive significance has been neglected. Spatial-learning skills are hypothesized to evolve in proportion to nagivational demands, and it is predicted that sex differences in spatial ability will evolve only in species where range expansion contributes differentially to the reproductive success of males and females, This prediction was tested via field studies of ranging behaviour and laboratory studies of spatial ability in two congeneric rodent species whose mating systems differ. Radiotelemetric studies showed that, in a polygynous species (meadow voles, Microtuspenns),lt:anicus), males expanded their ranges only during the breeding season and only when they attained full reproductive status. Females showed neither response. This suggests that range expansion was a male reproductive tactic. In contrast, a monogamous congener (prairie voles, M. ochrogaster) showed no sex differences in ranging, regardless of reproductive status, This probably reflects the relative inability of monogamous males and females to benefit from increased exposure to members of the opposite sex. When subsequently tested in a series of seven symmetrical mazes, subjects from the field studies exhibited the predicted sex-by-species patterns of spatial ability: only meadow voles showed consistent male superiority on these spatial tasks. Sex differences in the ability to solve spatial problems are among the most reliable of human sex differences (Maccoby & Jacklin 1974; Harris 1978; McGee 1979; Wittig & Petersen 1979). Tests of spatial ability are diverse, but all require the mental manipulation of spatial data. Each test yields a characteristic male advantage: on some tests the male-female differential is slight, but on others the average man performs a full standard deviation above the average woman (e.g. Wilson & Vandenberg 1978; see Fig. 4 in Linn & Petersen 1985). Factor-analytic studies suggest two or three components to spatial ability, but such data-reduction techniques leave the male advantage intact, and meta-analytic approaches confirm its validity (McGee 1982; Rosenthal & Rubin 1982; Linn & Pulos 1983; Linn & Petersen 1985). Sex differences in spatial ability are not restricted to Homo sapiens. In both wild and laboratory rodents, males consistently outperform females on spatial tasks such as maze learning (e.g. McNemar & Stone 1932; Barrett & Ray 1970; Joseph et al. 1978). The apparent generality of this sex difference invites explanation. Psychologists have explored the ontogenetic bases of sex differences in spatial ability. For example, androgens elevate spatial ability in * Present address: New Jersey School of Conservation, Montclair State College, Branchville, NJ 07826, U.S.A. 0003.3472/89/020322 + 10$03.00/O ( humans and laboratory rodents and thus are a proximate cause of such sex differences (Dawson et al. 1973, 197.5; Stewart et al. 1975; Joseph et al. 1978; Buchsbaum & Henkin 1980; Hier & Crowlty 1982; Rovet & Netley 1982; Resnick et al. 1986: Gaulin & Hoffman 1988). But regardless of what developmental pathways influence spatial ability, a complete analysis would explain why selection has differentially favoured spatial ability in the two sexes. Sexual selection theory (Bateman 1949; Williams 1966; Trivers 1972; Wade 1979) provides a general model for the evolution of sex differences in morphology, physiology and behaviour. Briefly, the strength of disruptive selection (moving male and female phenotypes apart) depends on the ratio of male to female variance in reproductive success. When this variance ratio is near unity, disruptive selection will be weak. Otherwise, competition for mates will be more intense in the sex with higher variance in reproductive. success, causing disruptive selection for the traits most useful in this competition. Monogamous mating systems tend to produce variance ratios around unity and are thus characterized by minimal sex differences; polygamous systems produce variance ratios that can deviate markedly from unity and thus foster the evolution of sexual dimorphism. We postulate that selection should favour spatial 1989 The Association for the Study of Animal Behaviour