Early isolation experiments indicate that male songbirds learn their songs during an early sensitive period, although later work has shown that some open-ended learners modify songs in later years. Recent isolation experiments suggest that in some species song has a stronger genetic basis than previously thought. This study raised domestic canaries under different combinations of acoustic and social isolation and followed song development into the second year. Males raised alone in acoustic isolation developed songs with normal syllables, but larger repertoires and also produced syllables with lower repetition rates when compared to controls. The smallest repertoire occurred in males raised in a peer group. Isolate males had a smaller song control nucleus HVC than controls, but there was no effect on nucleus RA or on brain weight in general. In the second year, after introduction into a large normal colony, isolate and peer group males adjusted their syllable repertoire to normal size. In particular, the isolates reduced their repertoire even though the size of HVC showed a significant increase in volume. However, songs of isolate and peer group males still differ in repetition rate and number of single syllables in the common aviary. In contrast, control males showed low syllable turnover and no significant change in repertoire size. Nor did they show any significant change in the volumes of song control nuclei. It seems that complete isolation affects only some aspects of song and brain development, and later socialization corrects some but not all of these in the second year.