Auditory masking of one pure tone by another.—Using an air damped telephone receiver supplied with current with a proper combination of two frequencies, as source, the amount of masking by tones of frequency 200 to 3500 was determined for frequencies from 150 to 5000 per sec. The magnitude of a tone is taken as the logarithm of the ratio of its pressure to the threshold value, and masking is taken as the logarithm of its threshold value with masking to that without. The curves of masking as function of magnitude are approximated straight lines as a rule except for rounded feet, of slope $s$ intersecting the magnitude axis at minimum masking magnitude $m$. For a given masking frequency $n$ the slope increases from zero through nearly 1.0 for a frequency near $n$, then more slowly, approaching about 3 to 4 for the highest frequencies measured. The intercept is small or zero below $n$, then increases rapidly, approaching the value 3 for high frequencies. Except when the frequencies are so close together as to produce beats, the masking is greatest for tones nearly alike. When the masking tone is loud it masks tones of higher frequency better than those of frequency lower than itself. When the masking tone is weak, there is little difference. If the masking tone is introduced into the opposite ear, no appreciable masking occurs until the intensity is sufficient to reach the listening ear through the bones of the head. At intensities considerably above minimum audibility, there is no longer a linear relation between the sound pressure and the response of the ear. Data are given showing combinational tones resulting from this non-linearity when two tones are simultaneously introduced in the ear. The presence also of subjective overtones in a loud tone accounts for the large amount of masking of tones higher than itself by a loud masking tone.

Dynamics of inner ear.—The data on masking together with Knudson's data on frequency sensibility are interpreted in terms of the dynamical theory of the cochlea which ascribes its frequency selectivity to a passing of vibrations along the basilar membrane and a shunting through narrow regions of the membrane at points depending on the frequency. Conjectured curves are given for a few single frequencies of the amplitude of vibration of this membrane as a function of the distance along it.

• Received 23 July 1923

DOI:https://doi.org/10.1103/PhysRev.23.266