In experiment 1, thresholds for detecting FM were measured for carrier frequencies f[inf c] from 0.25 to 6 kHz and for modulation rates f[inf m] from 2 to 20 Hz, using a 2AFC task. To disrupt cues based on changes in the excitation pattern, thresholds were also measured when the carriers in both intervals of a trial were amplitude modulated at the same rate as the FM with a modulation index of 0.333. The phase of the AM relative to the FM was randomized. For f[inf c]<4 kHz, the deleterious effect of the AM increased with increasing f[inf m]. For f[inf c]=6 kHz, the effect was independent of f[inf m]. In experiment 2, psychometric functions were measured for detecting combined FM and AM, with f[inf m]=2 Hz, as a function of the relative phase of the modulators, using modulation depths for AM and FM that would be equally detectable if each were presented alone. Relative modulator phase had no effect for f[inf c]=0.25 kHz, small effects for f[inf c]=1 kHz, and large effects for f[inf c]=6 kHz. The results suggest that a temporal mechanism dominates FM detection for f[inf c]<4 kHz, and for f[inf m]=2 Hz. A place mechanism dominates for high carrier frequencies, and for lower carrier frequencies when f[inf m](greater than or equal to)10 Hz.