Models considering sex ratio optima under single foundress strict local mate competition predict that female bias will be reduced by stochasticity in sex allocation, developmental mortality of males and limited insemination capacity of males. In all three cases the number of males per brood is expected to increase with brood size. Sex ratio optima may also be less female biased when several mothers contribute offspring to local mating groups or if non-local mating occurs between members of different broods; again more males are expected in larger broods. In the parasitoid wasp Goniozus legneri (Hymenoptera: Bethylidae), sex allocation has only a small stochastic component, developmental mortality is low and non-siblings are unlikely to develop in the same brood. However, the number of males per brood increases with the size of the brood (produced by a single mother). We investigated the further possibilities of limited insemination capacity and non-local mating using a naturalistic experimental protocol. We found that limited insemination capacity is an unlikely general explanation for the increase in number of males with brood size. All males and females dispersed from both mixed and single sex broods. Although most females in mixed sex broods mated prior to dispersal, these data suggest that non-local mating is possible, for instance via male immigration to broods containing virgin females. This may influence sex ratio optima and account for the trend in male number.