To examine the functional significance of song in a tropical resident wood-warbler (Emberizidae:Parulinae), the vocal behavior of the Adelaide's Warbler was studied in relation to its social behavior and ecology. Breeding was seasonal although individuals maintained monogamous pair bonds and exclusive pair territories year-round. When residents disappeared, territories and pair bonds were replaced by floaters or neighbors. Paired birds gave duets consisting of male songs plus female calls (females song was uncommon). Both sexes used the same non-song display in close-range territorial encounters, but only males maintained territories when unpaired. The large song repertoire of each male (about 23 song types) was organized into two categories, A and B, distinguished by structural characteristics, manner of singing, and temporal use. Relative to A songs, B songs were more complex and emphasized lower frequencies, and were sung at more rapid rates and with greater sequential variety. Males sang A songs regularly throughout the day and year. In contrast, males sang B songs mainly during the breeding season, in bouts, regularly and intensively at dawn and sporadically and less intensively during the day. Social patterns of song use suggested A songs were more important in communication with females, whereas B songs were more important in communication between males. Males sang A songs in male-female interactions and increased their output of A songs when unpaired. Paired males sang B songs during male-male interactions only when females were not nearby, especially at dawn and during the incubation stage. Song sharing patterns indicated repertoires were learned from neighbors after dispersal. Micro-geographic song variation was extreme, with little sharing at distances $\geq$500 m. Repertoire similarity between males was negatively correlated with the distance and frequency of social interaction between them. In addition, B songs were more often shared than A songs, suggesting different timing or modes of learning. In field song playback experiments, males responded significantly different towards: (1) conspecific and congeneric songs, (2) familiar and unfamiliar conspecific songs, and (3) conspecific A and B songs. Both males and females responded differently to playback during the breeding versus non-breeding seasons.