overshooting in response to stimuli placed on the side of the lesion (P < 0.001 for both animals, Fig. IC). In contrast to VIIIth nerve lesions, turn amplitude did not increase on the side opposite the lesion. Complete VIIIth nerve lesions disrupt input from all end organs which may result in more widespread turning deficits. For escape, preand post-lesion turn amplitudes were not significantly different (Fig. 1D). While lesions to the peripheral vestibular system did not affect the turning component of escape, they may affect the airborne phase and landing. Following PC nerve lesion, the amplitudes of prey capture and escape turns did not change (data not shown). This result implies that the turning deficits described here for prey capture are specific to horizontal canal nerve lesion and not a result of non-specific damage to the nervous system or to general surgical trauma. These data support the idea that on-line vestibular input is required for prey capture turning and not required for escape turning. This is consistent with the fact that turn accuracy is of greater importance for prey capture than for escape. Since the HC measures angular acceleration in the horizontal plane, it is well suited to supply the animal with on-line information about the actual position and velocity of the body in space. This information could be used to guide the centrally generated movement of the body and head toward the prey target. For escape turns, the spatial organization is more variable and the peak velocity is greater than for prey capture turns (4). This suggests that the speed of the escape turn is more important than accuracy; thus vestibular input may not be required. The initial phase of the escape response, the turn, may be truly ballistic in nature allowing the frog to maximize the speed of the response. This work was supported by the Grass Foundation.