In a previous study, we presented psychophysical evidence that time-shrinking (TS), an illusion of time perception that empty durations preceded by shorter ones can be conspicuously underestimated, gives rise to categorical perception on the temporal dimension (Sasaki, Nakajima, & ten Hoopen, 1998). In the present study, we first survey studies of categorical rhythm perception and then describe four experiments that provide further evidence that TS causes categorical perception on the temporal dimension. In the first experiment, participants judged the similarity between pairs of /t1/t2/ patterns (slashes denote short sound markers delimiting the empty time intervals t1 and t2). A cluster analysis and a scaling analysis showed that patterns liable to TS piled up in a 1:1 category. The second and third experiments are improved replications in which the sum of t1 and t2 in the /t1/t2/ patterns is kept constant at 320 ms. The results showed that the 12 patterns /115/205/, /120/200/, . . ., /165/155/, /170/150/ formed a 1:1 category. The fourth experiment utilizes a cross-modality matching procedure to establish the subjective temporal ratio of the /t1/t2/ patterns and a 1:1 category was established containing the 11 patterns /120/200/, /125/195/, . . ., /165/155/, /170/150/. On basis of these converging results we estimate a domain of perceived 1:1 ratios as a function of total pattern duration (t1 + t2) between 160 and 480 ms. We discuss the implications of this study for rhythm perception and production.