The first 2 experiments examined the ability of monkeys (Macaca fascicularis) to learn a series of visuomotor conditional discrimination problems for food reward. In each discrimination problem there were 2 visual stimuli, which were different from those in any previous discrimination problem. Each trial within a problem presented either one visual stimulus or the other, and the stimulus indicated which of 2 motor responses, tap or hold, was correct; the motor responses were defined in such a way as to be mutually exclusive. Successive problems were each trained to a criterion of correct performance. In the first experiment, it was found that fornix transection did not impair monkey's learning rate in this task. This result contrasts with previous results showing an impairment of learning rate following fornix transection in visuo-motor conditional discriminations in which the 2 motor responses were differentiated from each other by their spatial position or direction. The present result shows that the earlier demonstrated impairments are specific to spatial responses. In the second experiment, learning rate in the present task was found to be impaired by the combination of a unilateral inferotemporal ablation, contralateral to the hand used in the task by the animal, with a transection of the anterior corpus callosum. This result suggests that cortico-cortical interaction between the inferotemporal area and the frontal lobe contralateral to the hand in use is necessary for efficient learning in this task. The third experiment examined simultaneous 2-choice visual discrimination learning in the animals which had participated in experiment 2. In contrast to the results of Expt. 2, learning rate in this task was unimpaired. Experiment 3 shows that the impairment observed in Expt. 2 is not a general impairment of visual learning.