Dysfunction of spinal motor conduction during surgical procedures may not be reflected by changes in somatosensory evoked potential waveforms. A method of monitoring that allows direct and continuous assessment of motor function within the central nervous system during surgery would be useful. This paper describes one such method utilizing noninvasive electric cortical stimulation to evoke muscle activity (the motor evoked potential, or MEP) during surgery. The effect of isoflurane (superimposed on a baseline of N2O/narcotic anesthesia) on MEP's in response to cortical stimulation is specifically examined. Eight patients undergoing elective neurosurgical operations were included in the study. All patients received a background of general anesthesia and partial nondepolarizing neuromuscular blockade. The motor cortex was stimulated electrically via self-adhesive scalp electrodes. Electromyographic responses from multiple muscles were measured with subdermal electroencephalograph-type needle electrodes. Motor responses to stimulation were continually recorded on magnetic tape for off-line analysis. Once closing of the surgical incision was begun, a series of four to five stimuli of constant magnitude were applied to obtain "baseline" MEP responses. Patients were then ventilated with isoflurane for up to 8 minutes, during which time stimuli were continued every 15 to 20 seconds. Comparison was made of MEP responses for trials before, 1 minute after, and 5 minutes after the addition of isoflurane. All patients demonstrated reproducible motor responses to cortical stimulation during surgery. Addition of isoflurane [isoflurane)exp, less than or equal to 0.5%) to pre-existing anesthesia caused marked attenuation of MEP amplitudes in all patients within 5 minutes of its application, without affecting neuromuscular transmission as judged by direct peripheral nerve stimulation. It is concluded that: 1) monitoring motor system integrity and function with electric transcranial cortical stimulation during surgery is feasible when utilizing an N2O/narcotic anesthetic protocol; and 2) the quality of data obtained will likely suffer with the addition of isoflurane.