This study arises in response to Cook (1987) and Karno and Konecni (1992), who call into question the ability of listeners to perceive tonal closure in the tonic key, a primary structural principle of tonal composition. Participants in their experiments heard unaltered compositions and variants that destroyed the compositions' tonal structure by transposing the second half to a new key or by reordering formal units. Neither study found a correlation between the listeners aesthetic preferences and the tonal closure of the original compositions. The blurred distinction between aesthetic response and music-structural judgments in these designs is misleading; the experiment reported here asks the music-structural question explicitly. Our stimuli consisted of 12 unaltered excerpts from classical CD recordings in three categories: nonmodulating, modulating to the dominant, modulating to a key other than the dominant. For each excerpt, 36 students at a professional music school answered six questions, one of which asked whether the concluding key was the same as the initial key. Participants' accuracy was significantly above chance in this task. Musicians in scholarly disciplines and performers discriminated tonal closure equally well, although there was a significant effect for number of years of theory training.