9513821 Nelson Vocal learning in song birds is a complex motor skill in which young males memorize the songs of adults, and then use their memory to control their own song output. One consequence of vocal learning in birds is the formation of vocal dialects in which males at one locality sing similar songs that differ from those of males at other localities. Until recently, it has been assumed that a single process is involved in the formation of dialects: the young bird imitates one or more adults at the time and place the youngster settles upon a breeding territory. Recent evidence suggests that an independent process of selective attrition of songs from a large repertoire acquired in whole or in part at a different time and place can also lead to song sharing between neighbors. The proposed research will examine the role of song overproduction and selective attrition in the formation of local song dialects. The PI will take a comparative approach by studying different subspecies of the white-crowned sparrow (Zonotrichia leucophrys). Field observations, and laboratory and field experiments will be conducted to compare processess of vocal development in one subspecies that does form local dialects (Z. l. oriantha or Z. l. leucophrys) with another that does not (Z. l. gambelii). In the field, the PI will observe vocal interactions between males as they settle upon territories, and determine the sizes of their song production repertoire early in Spring, and compare whether they selectively retain or reject songs that match their neighbors. This will test the hypothesis that developmental pathways are conserved during evolutionary diversification, and that changes often are made to terminal stages. The PI will then conduct field and laboratory playback tests contrasting the song matching behavior of males of two subspecies. These experiments are designed to reveal differences in the perceptural categorization of songs, and to identify possible genetically-based differences in song pe rception. The PI will also employ a new method for measuring song learning that is independent of the production of imitations to examine how songs are memorized during the sensitive phase and during the subsequent storage interval. Finally, the PI will conduct a song learning experiment to assess whether live tutors modiy or extend the early sensitive phase for song acquisition, as compared with tape tutoring.