In a study to determine binaural thresholds for azimuth difference [Schmidt, Van Gemert, De Vries, and Duyff, Acta Physiol. Pharmacol. Neerl. 3, 2–18 (1953)] it was found that, at high enough frequencies, one or more regions of low localization sensitivity exist at the higher azimuth values. This suggests the use of a “null” method as a useful technique to gather information on the localization process. Two loudspeakers, whose separation could be changed, were sounded alternately. The subject, seated in a rotating chair, could change his azimuth at will. The regions in which the subject seems to hear the sound coming from only one source are then determined. For frequencies below about 1200 cps, only two such regions exist, at the expected 90° and 270°. For frequencies above about 1400 cps, more than two such regions exist. At intermediate frequencies, two or more regions may exist, depending on speaker separation. Possible explanations for these results will be discussed.