To test whether the most rapid visually evoked reach adjustments are cortically organized in humans, we have measured their latency in a healthy subject with complete agenesis of the corpus callosum. This condition precludes direct communication between left and right cerebral cortices and so, in this subject, a purely cortical visuomotor process would be expected to produce longer-latency responses to a target that appears in the visual hemifield contralateral to the responding limb (crossed) compared with the ipsilateral hemifield (uncrossed). As predicted, when performing simple reaction time tasks that involved lifting a finger or an arm in response to a visual stimulus presented to either hemifield, this acallosal subject showed a significant crossed-uncrossed latency difference (mean 35.8 ms) that was not present in control subjects (group mean 2.2 ms). In contrast, when she reached for a target that unexpectedly jumped into either visual hemifield, the latencies of mid-flight adjustment were the same (approximately 120 ms) irrespective of either the target jump direction or which hand was used. This was not due to an early movement of the eyes bringing the target back on to the fovea since this subject's finger always deviated towards the new target position in advance of her eyes. Neither could it be explained by the use of ipsilateral corticospinal projections since transcranial magnetic stimulation over the motor cortex failed to evoke ipsilateral responses in arm or hand muscles. These results suggest that, even in humans, subcortical structures are involved in the fastest adjustments of the reaching arm made in response to fresh visual information. An additional finding in this subject was that, when reaching, the eye saccadic latency was greater by 36 ms on average when the target jumped right compared with left, irrespective of which hand was being used. This is the same value as the mean interhemispheric transfer time obtained in the simple reaction time tasks and may indicate right-hemispheric dominance for saccadic eye movement control.