While many studies have looked at selection for large repertoire sizes in birds, few have explored the potential for selection on the ability to sing each type in the repertoire consistently. I explored this issue by studying the consistency with which two closely related species of mockingbirds, i.e. the northern (Mimus polyglottos) and the tropical mockingbird (M. gilvus), repeated each syllable type. I hypothesized that if there is selection for singing ability and not just for an increase in repertoire size, it could be expected that syllable consistency should be (1) higher in older, more experienced individuals, (2) higher in species with higher potential for sexual selection via female choice (i.e. higher in the northern versus the tropical mockingbird) and (3) higher in males expected to be of higher quality (i.e. in dominant versus subordinate birds). I found that song consistency could be improved with practice because tropical mockingbirds became more consistent with age. When comparing across male types, I found that consistency was higher in northern and dominant tropical mockingbirds than in subordinate tropical mockingbirds (although the data suggests that dominant tropical mockingbirds that compete with subordinates for breeding may not be much better singers than these subordinates). I suggest that the lack of significant differences between northern mockingbirds and dominant tropical mockingbirds could be a product of a trade-off between repertoire size and singing consistency because as the number of types in a repertoire increases, the opportunities to practice each type are reduced.