This experiment used cortical microstimulation to probe the mapping from primary motor cortex to the biceps and triceps muscles of the arm in monkeys. The mapping appeared to change depending on the angle at which the elbow was fixed. For sites in the dorsal part of the arm and hand representation, the effects of stimulation were consistent with initiating a movement of the elbow to an extended angle. Stimulation evoked more triceps activity than biceps activity, and this difference was largest when the elbow was fixed in a flexed angle. For sites in the ventral part of the arm and hand representation, stimulation had the opposite effect, consistent with initiating a movement of the elbow to a flexed angle. For these sites, stimulation evoked more biceps activity than triceps activity, and the difference was largest when the elbow was fixed in an extended angle. For sites located in intermediate positions, stimulation evoked an intermediate effect consistent with initiating a movement of the elbow to a middle, partially flexed angle. For these sites, when the elbow was fixed at a flexed angle, the evoked activity was largest in the triceps, and when the elbow was fixed at an extended angle, the evoked activity was largest in the biceps. These effects were obtained with 400-ms-long trains of biphasic pulses presented at 200 Hz and 30 microA. They were also obtained by averaging the effects of individual, 30-microA pulses presented at 15 Hz. How this stimulation-evoked topography relates to the normal function of motor cortex is not yet clear. One hypothesis is that these results reflect a cortical map of desired joint angle.