In an earlier paper [J. Acoust. Soc. Am. 50, 1116–1122 (1971)] I presented a lateralization model of binaural masking‐level differences (MLD’s), according to which detection under conditions which produce an MLD is based on the average value of an interaural parameter, Δ. This factor is made up of interaural time (Δt) plus interaural intensity (ΔI) weighted by a binaural trading ratio (TR). Two figures in that paper depicted application of the model to MLD’s found as a function of frequency [(Hirch and Burgeat, J. Acoust. Soc. Am. 30, 827–832 (1958)]; those figures are partially incorrect. They show the lateralization equation with a time‐intensity trading ratio of 20 μsec/dB to be the best fit, when in fact, the fit to TR=0 μsec/dB is excellent. The lateralization model with a trading ratio of 0 is very much like the vector model of Webster [J. Acoust. Soc. Am. 23, 452– 462 (1951)]and Jeffressetal. [J. Acoust. Soc. Am. 28, 416–426 (1956)]. It is suggested here that variability in the noise process may be significant factor, allowing binaural neurons tuned to high frequencies to recover and then respond again to interaural differences of time.