The interhemispheric transfer of visual discriminations in split-chiasm monkeys (Macaca nemestrina) was assessed by training with one eye to a criterion level, then testing either with that same eye (control) or with the other eye (transfer). The difference between these two values was the loss due to transfer. A computer simulation suggested that the usual savings score could grossly misestimate transfer ability. In addition, stimuli with comparable left and right halves were used to minimize the effect of the bilateral hemianopia caused by chiasm section. Performance with the untrained eye was slightly, but statistically significantly, poorer than with the trained eye. No evidence of the phenomena of "learning to transfer" was found (i.e., there was no improvement in transfer ability in relation to concurrent intrahemispheric controls).