Hypotheses regarding dawn singing in birds remain largely unsupported by quantitative data, especially for suboscine passerines (suborder Tyranni). During a study of the singing behavior of a suboscine, the Dusky Flycatcher (Empidonax oberholseri), in Alberta, Canada, spring storms in 2002 caused the disappearance and presumed mortality of several territorial males and some of their replacements, creating a serendipitous experiment. Overall, 15 males disappeared and, as a result, the number of territorial males in our study area declined from 13 prior to the storms to six afterward. Only one male maintained the same territory throughout the breeding season. During the period of inclement weather and social instability (20 May to 9 June), males sang at high rates during the predawn period. From 10 to 19 June, following this period of unstable weather and turnover in territorial males, males began dawn singing later in the morning, and sang shorter bouts at lower rates. The proportion of males engaging in dawn singing also decreased between the two periods. In contrast, daytime singing activity of paired males was low during both time periods. Playback of dawn songs to territorial males between 20 June and 21 July caused a resumption of some dawn singing. Singing rates were higher on the day of playback than on the day before playback, and the times when dawn singing was initiated were earlier on each of 2 d after playback than on the day before playback. In addition, the proportion of males engaging in dawn singing increased between the day before and the day after playback. Both the decrease in dawn singing when population density was reduced and its partial restoration by playback suggest that dawn singing in this species functions in male–male interactions and support the social-dynamics hypothesis as an explanation for the dawn chorus in this species.