Abstract
Around-the-clock behavioral observations indicated: (1) unmated males sang more than mated males, (2) only unmated males sang at night, (3) amount of singing by mated males decreased when females incubated, and increased during the nestling stages, (4) males sang during copulations, and (5) males achieving the highest reproductive success were among those singing the least. Behavioral sampling of mated, unmated, and female-removed males indicated: (1) new females were not available to replace those removed, (2) male attack rates against conspecifics and heterospecifics were not correlated with the amount of time spent singing, and (3) heterospecific intrusion rate into a personal space of 5 m from males was greater than the rate of attack by males against heterospecifics. Interspecific aggression appears to be parental investment rather than territorial defense. Song analyses indicated that new song types may continually be added to the vocal repertoire, and there may not be a definable upper limit of song repertoire size. Results of this study are consistent with the hypothesis that song functions ultimately in male-female bonding and reproductive interactions, rather than in male-male competitive interactions. Song learning, vocal mimicry and large song repertoires may have evolved as mechanisms whereby males advertise their learning abilities to potential mates. Once a pair is mated, song appears to coordinate the timing of reproduction.