PATTERNS OF SEX-RATIO, VIRGINITY AND DEVELOPMENTAL MORTALITY IN GREGARIOUS PARASITOIDS

Theory considering sex ratio optima under 'strict local mate competiti on with offspring groups produced by a single foundress' makes a suite of predictions, one of which is a mean female bias. Treating individu al offspring as discrete units, theory further predicts sex ratios to have low variance (precise sex ratio) and to equal the reciprocal of c lutch size (one male per clutch). The maternal decision may be complic ated by imperfect control of sex allocation, limited insemination capa city of sons and offspring developmental mortality: each can lead to v irgin daughters (with zero fitness) and consequently select for less b iased sex ratios. When clutches are small and/or developmental mortali ty is common, appreciable proportions of virgins are expected, even wh en control of sex allocation is perfect and the mating capacity of mal es is unlimited. This suite of predictions has been only partially tes ted. We provide further tests by examining sex ratios and developmenta l mortalities within and across species of locally mating parasitoids. We find a wide range of mean developmental mortalities (6-67%), but m ortality distributions are consistently overdispersed (have greater th an binomial variance) and sexually differential mortality appears to b e absent. Sex ratios are female biased and have low variance, but are not perfectly precise and variance is increased by mortality within sp ecies and (equivocally) across species. Sex ratios less biased than th e reciprocal of clutch size are observed; probably due to a maternal r esponse to developmental mortality in one species, and to limited inse mination capacity in others. Cross species comparisons indicate that m ean proportions of mortality and virginity are positively correlated. Virginity is more prevalent than predicted among species with higher m ortalities but not among lower mortality species. Predicted relationsh ips between virginity and clutch size are supported in species viith l ower mortalities but only partially supported when mortality rates are higher. (C) 1998 The Linnean Society of London.