1. The number of reproductive events per lifetime affects all levels of ecological complexity, including the structure and dynamics of species communities. Here we investigate a sexual parasite-host system with two water frog species of which one (Rana esculenta, E) is a hemiclonal hybrid that depends on one of its parental species (R. lessonae, L) for successful reproduction. Previous theoretical models have shown that relative fecundities are important for the species composition and stability of this system, but empirical data from natural mixed LE-populations were lacking. 2. We used three different methods to estimate the proportion of annually reproducing hybrid and parental females: (a) screening for eggs through a tiny incision into the belly skin, (b) measuring plasma hormone titres from blood samples and (c) counts of egg clumps deposited into the pond. 3. After accounting for the respective adult population sizes of the two female types - calculated from mark-recapture analyses - we found that all three estimates revealed a higher proportion of reproducing L- than E-females. Ratios ranged from 1·43 (method c), to 1·90 (a) to 2·38 (b), with an average of 1·90. 4. We also investigated three potential causes for the fecundity difference (age structure, hybrid sterility and differences in reproductive costs) but could not find compelling evidence for any of them. We hypothesize that size-related differences in energy budgets offer the most probable explanation. 5. The consequences of the observed species differences in annual reproductive success were analysed with a theoretical model. The higher proportion of reproducing R. lessonae females increases the possibilities for stable coexistence in mixed LE-populations under a variety of combinations of mating patterns, relative fecundity and larval performance. The results, however, also indicate new limits for coexistence and highlight the paramount importance of larval competition for explaining the composition and dynamics of mixed LE-populations.