The study investigated the ability to detect and discriminate frequency glides under a variety of experimental conditions. The subjects distinguished between a comparison signal that either was level in frequency or was swept across a fixed frequency span, and a target signal that changed more in frequency than the comparison signal. Tone durations were 50 and 400 ms. Nominal center frequencies were 0.5, 2, and 6 kHz; actual center frequencies were varied randomly, or roved, over a range equal to 0.1 times the nominal center frequency. Up- and down-glides were used. The transition span of the comparison signal was either 0, 0.5, 1, or 2 times the equivalent rectangular bandwidth of the auditory filter at the nominal center frequency. Discrimination thresholds were obtained for all combinations of center frequency, direction, and span. Overall, thresholds expressed as ΔHz/ERB varied little as a function of center frequency. Glide duration had no effect on discrimination. The 50-ms down-glides were more difficult to detect than the 50-ms up-glides; otherwise, the effect of direction was not significant. With the exception of the 50-ms down-glides, detection/discrimination thresholds were similar for the 0-, 0.5-, and 1-ERB transition spans, but increased significantly for the 2-ERB span. The absence of significant variation across frequency supports a place mechanism for the detection of frequency change in gliding tones, based on the detection of changes in the excitation pattern. An excitation pattern model cannot account for the asymmetry noted for glide detection, however.