Infants 6, 9, 12, 15, and 18 months were seated in a dark room facing a semicircular array comprising 10 loudspeakers, 5 to their right and left at 18 degrees, 36 degrees, 54 degrees, 72 degrees, and 90 degrees from midline. Each infant received 2 types of trials: auditory-alone trials (only a sequence of 8 clicks played), and auditory-visual trials (following the first 4 clicks, a light display at the location of the loudspeaker was activated and remained on for the remaining 4 clicks). Calibration markers on the infant's head were used to measure azimuth error (i.e., the discrepancy between the angle of head turn and loudspeaker location). For auditory-visual trials, all infants turned their head within 4 degrees-6 degrees of the sound source; there was no effect of age or loudspeaker location. For auditory-alone trials, there were significant age differences in performance, with a systematic decrease in discrepancy angle with increasing age. By 18 months there was no difference in performance as a function of localization cue.