The evolution of sexually selected signals has been a major topic of scientific research since Darwin. In recent years, scientists have focused on how elaborate signals can indicate honest information about the quality of their bearers, as predicted by reliability theory. A key concept relating to how mating displays could reliably reveal quality is "performance." Animals face limits in display production, and producing high-‐performance displays depends on the adept coordination of multiple motor systems. Thus, by observing motor performance, signal-‐receivers can assess the quality of signalers. Birdsong is a prime example of a display that involves motor challenges in its production. In my dissertation, I examined the connections between signal reliability and vocal performance in the swamp sparrow, Melospiza georgiana, and addressed three main questions. First, does vocal performance signal the level of aggressive threat during territorial defense? In wild male swamp sparrows, I measured aggressive responses to playback of various performance levels. Males responded with decreased aggression to low-‐performance stimuli, suggesting these stimuli indicated low threat. Males were as aggressive to control-‐as to high-‐performance stimuli, but avoided approaching high-‐performance stimuli as closely. Additionally, I found that males who possessed high vocal performance were more aggressive. Second, does developmental stress affect adult vocal performance? I found that birds experiencing poor early nutrition had lower vocal performance than did control birds, indicating lasting effects of early stress. Also, males in both groups significantly improved their vocal performance of learned songs between years. Together these results suggest that vocal performance can indicate early condition and age. Third, what factors influence the development of song preferences in females? In two experiments, I raised and tutored females with songs of normal performance. When tested as adults, females displayed stronger preference behavior to tutor than to novel songs, indicating the effects of learning. Females also gave the fewest displays to low-‐performance and responded more strongly to normal-‐and high-‐performance songs, indicating an influence of sexual selection. These experiments provide the first evidence that the development of female preference for sexually selected traits can be guided by an interplay of learning and innate biases.