BACKGROUND For parasites with complex life cycles, size at transmission can impact performance in the next host, thereby coupling parasite phenotypes in the two consecutive hosts. However, a handful of studies with parasites, and numerous studies with free-living, complex-life-cycle animals, have found that larval size correlates poorly with fitness under particular conditions, implying that other traits, such as physiological or ontogenetic variation, may predict fitness more reliably. Using the tapeworm Schistocephalus solidus, we evaluated how parasite size, age, and ontogeny in the copepod first host interact to determine performance in the stickleback second host. METHODS We raised infected copepods under two feeding treatments (to manipulate parasite growth), and then exposed fish to worms of two different ages (to manipulate parasite ontogeny). We assessed how growth and ontogeny in copepods affected three measures of fitness in fish: infection probability, growth rate, and energy storage. RESULTS Our main, novel finding is that the increase in fitness (infection probability and growth in fish) with larval size and age observed in previous studies on S. solidus seems to be largely mediated by ontogenetic variation. Worms that developed rapidly (had a cercomer after 9 days in copepods) were able to infect fish at an earlier age, and they grew to larger sizes with larger energy reserves in fish. Infection probability in fish increased with larval size chiefly in young worms, when size and ontogeny are positively correlated, but not in older worms that had essentially completed their larval development in copepods. CONCLUSIONS Transmission to sticklebacks as a small, not-yet-fully developed larva has clear costs for S. solidus, but it remains unclear what prevents the evolution of faster growth and development in this species.