The precision with which a listener can lateralize a binaural-sound image about his medial plane was measured using three separate techniques, the method of constant stimuli, Mood and Dixon's up-and-down method, and a new sequential procedure (Wetherill's UDTR method). The latter technique was found to have several important advantages over the other two. An application of the UDTR method is described, in which the precision of lateralization for one of two mutually interfering speech sources was measured.