In many species of birds, males have complex song repertoires, which are used in the context of breeding. The succinct rationale of the research done in this thesis is the following paradox; assuming that song repertoires are costly to produce and store, why do males have repertoires of songs rather than a single song? A possible reason is that, if only males of good quality or in good condition are able to produce these repertoires, good quality males would outweigh the costs of the repertoire by increased benefits in mating success. I studied this issue in a population of willow warblers (Pbylloscopus trocbilus). I found some evidence that repertoire size was correlated with male quality. Male age was found to correlate with repertoire size, and there was a positive correlation between repertoire size and survival. The probability of a male having an offspring recruiting into the population was also correlated with repertoire size. Number of fledglings was positively correlated with repertoire size, even when the effect of arrival date was taken into account. However, female choice bore no relation to repertoire size. Females patted first with early arriving males. Females did not use repertoire size either when choosing extra-pair partners. Another characteristic, song length, seemed to be behind this choice, and males with short songs were more likely to be cuckolded. The effects of song repertoires in male-male competition were examined by means of a playback experiment. The results did not provide conclusive evidence of an effect of repertoire in male-male competition. Taken together, the evidence gathered in this thesis suggests that, although repertoire size correlates with several measures of male quality in this species, the maintenance of this trait does not seem to be based on a disproportionally higher mating benefit.