An essential element in predicting the maximum range at which an underwater sound source (signal) may be heard over a given sonar listening system is the signal to noise ratio for a prescribed probability of recognition. This ratio, measured in the frequency band presented to the listener, and expressed in decibels, is termed the recognition differential. The usefulness of this quantity rests on the fact that with sonar equipment properly engineered and operated, signals ordinarily are masked by noise picked up by the hydrophone; thus both signal and noise receive the same amplification at each frequency, making their ratio independent of the gain of the system. Recognition differentials for many different combinations of underwater signals and background noises were measured experimentally by presenting recorded samples of the sounds to a crew of five observers. Although this method yielded a reliable value of the recognition differential for each pair of sounds tested, the wide variation in the value from signal to signal, and the observed dependence on the frequency band passed by the system led to a search for a more general measure of signal audibility. By making suitable spectrum analyses it was possible to interpret the observed phenomena in terms of current masking theory utilizing Fletcher's critical bands. This led to a useful criterion for predicting the threshold level for a signal masked by a background noise where the spectra of both are known. Application of this criterion is restricted to: (1) background noises having continuous spectra, (2) signal and noise spectra differing in shape, (3) receiving systems producing adequate loudness. This critical band criterion of recognition has been found useful in predicting the effects on sonar operation of such factors as: (1) system frequency response including filters, (2) system gain, (3) self-noise spectrum, (4) hydrophone directional characteristics, (5) frequency selective sound propagation. This represents one of the results of research carried out by the University of California Division of War Research under contract OEMsr-30 with the Office of Scientific Research and Development (Section 6.1 NDRC). This work is continuing under the Psychophysics Section of the U. S. Navy Electronics Laboratory.