ABSTRACT: Variation in larval size has been shown to be an important factor for the post-metamorphic performance of marine invertebrates but, despite its importance, few sources of this variation have been identified. For a range of taxa, offspring sizeis positively correlated with maternal size but the reasons for this correlation remain unclear. We halved the size of colonies in the bryozoan Bugula neritina 1 wk prior to reproduction (but during embryogenesis) to determine if larval size is afixed or plastic trait. We manipulated colonies in such a way that the ratio of feeding zooids to reproductive zooids was constant between treatment and control colonies. We found that manipulating colony size strongly affects larval size; halved coloniesproduced larvae that were ~13% smaller than those produced by intact colonies. We entered these data into a simple model based on previous work to estimate the likely post-metamorphic consequences of this reduction in larval size. The model predictedthat larvae that came from manipulated colonies would suffer ~300% higher post-metamorphic mortality and ~50% lower fecundity as adults. Colonies that are faced with a stress appear to be trading off current offspring fitness to maximize their ownlong-term fitness and this may explain previous observations of compensatory growth in colonial organisms. This study demonstrates that larval size is a surprisingly dynamic trait and strong links exist between the maternal phenotype and the fitness ofthe offspring. The performance of settling larvae may be determined not only by their larval experience but also by the experience of their mothers.