Rhythms are fundamental to behavior, but the control mechanism for timed responses is not known. Many theorists have assumed that there is a central clock coordinating behavior in all sensory modalities and response modes. We tested this hypothesis using a rhythmic tapping task in which university undergraduates first attempted to synchronize responses with brief auditory, tactile, or visual stimuli and then continued to tap at the same rate on their own. Performance was most variable with visual stimuli and least variable with auditory stimuli. The detailed results suggest that performances are not based on a common clock, but, rather, different strategies are employed when the task is presented in different modalities. We reject the hypothesis of a single timing mechanism as controlling behavior and, in doing so, question the validity of information processing models that are formulated without regard to temporal relations among their conjectured processes.