Song structure without auditory feedback: emendations of the auditory template hypothesis.

Motor patterns of songs of swamp and song sparrows, Melospiza georgiana and M. melodia, deafened early in life display a significant degree of species-specific structure. Normal songs of the two species differ in the degree to which they are segmented. Swamp sparrow song consists of a single segment, and song sparrow songs are multisegmental. Song and swamp sparrows were deafened at 17 to 23 days, prior to the onset of song or subsong. The song sparrows developed more segments in their singing than the swamp sparrows. Species-specific trends were also evident in song durations and frequency characteristics. Abnormalities were found, however, in the morphology of the notes and syllables from which songs of early deafened sparrows are constructed. These results require emendation of the auditory template hypothesis of song learning in birds.