Abstract
Critical bands in the cat were measured by a behavioural psychophysical method. Pure tones were masked by noise by variable bandwidth but constant total power, geometrically centred on the test tone; the point at which the masked threshold began to fall as the masker bandwidth was increased estimated the critical bandwidth. At 2 kHz the critical bandwidth was also measured from the wideband masked thresholds of both tones and noise of variable bandwidth: this produced the same result as the first method. The measured critical bandwidth was greater than previously published values of the effective bandwidths of single fibres of the auditory nerve. The results do not fit in with the commonly accepted theory that the critical band represents the resolution of the cochlea.