Masked thresholds for spondees were measured during 37 binaural listening conditions covering homophasic, antiphasic, parallel time‐delayed, and opposed time‐delayed listening in the presence of one to four competing maskers at three nominal masker levels (65, 75, and 85 dB SPL). Two of the maskers used were white noise modulated 4 times per second by 10 dB with a 50% duty cycle; the modulation patterns were such that when the noises were combined they produced a continuous white noise. The other two were whole sentences spoken by different male talkers. Masking‐level differences (MLD's)rehomophasic performance in the same masker complex emerged for every one of the 28 conditions of dichotic listening. MLD's were slightly smaller in the presence of a single competing talker (with or without noise also present) than with two talkers, and MLD's were always largest for the antiphasic conditions (range 4.4–6.8 dB). MLD's for parallel delayed listening (entire masking complex given 0.8‐msec lead to a single ear) averaged about 1 dB less than antiphasic conditions, ranging from 3.5 to 6.6 dB. MLD's for opposed time‐delay conditions (0.8‐msec delays given to all portions of the masker complex but not all masker signals delayed to the same ear) reduced the MLD another 0.6 dB from the antiphasic conditions on the average (range 2.6–6.4 dB). Since subjective separation of competing signals was more definitive during opposed time delays than during either parallel time delays or antiphasic presentation, but since MLD's were smaller during opposed time delays than in the companion dichotic conditions, this study indicates anew that escape from masking and lateralization of signal origin, while depending on common input information, are separate capacities involving two different neurological mechanisms. These findings also suggest that listening in interaurally complex conditions brings into play phenomena that are not covered by contemporary theories of masking.