We studied the characteristics of combined eye-head gaze shifts in human subjects to determine whether they used similar strategies when looking at visual (V), auditory (A), and combined (V + A) targets located at several target eccentricities along the horizontal meridian. Subjects displayed considerable variability in the combinations of eye and head movement used to orient to the targets, ranging from those who always aligned their head close to the target, to those who relied predominantly on eye movements and only moved their head when the target was located beyond the limits of ocular motility. For a given subject, there was almost no variability in the amount of eye and head movement in the three target conditions (V, A, V + A). The time to initiate a gaze shift was influenced by stimulus modality and eccentricity. Auditory targets produced the longest latencies when located centrally (less than 20 degrees eccentricity), whereas visual targets evoked the longest latencies when located peripherally (greater than 40 degrees eccentricity). Combined targets (V + A) elicited the shortest latency reaction times at all eccentricities. The peak velocity of gaze shifts was also affected by target modality. At eccentricities between 10 and 30 degrees, peak gaze velocity was greater for movements to visual targets than for movements to auditory targets. Movements to the combined target were of comparable speed with movements to visual targets. Despite the modality-specific differences in reaction latency and peak gaze velocity, the consistency of combinations of eye and head movement within subjects suggests that visual and auditory signals are remapped into a common reference frame for controlling orienting gaze shifts. A likely candidate is the deeper layers of the superior colliculus, because visual and auditory signals converge directly onto the neurons projecting to the eye and head premotor centers.