Twenty patients with hemiplegia and 13 patients with motor neurone disease were studied with electrical and magnetic transcranial stimulation. Motor evoked potentials were recorded from the biceps, thenar and tibialis anterior muscles. In both groups of patients magnetic stimulation with a Novametrix stimulator revealed fewer abnormalities than electrical stimulation with a Digitimer D180 stimulator. In patients with hemiplegia, motor evoked potentials after electrical stimulation were absent in 70% of muscles, delayed in 22% and normal in 8%; after magnetic stimulation, they were absent in 53% of muscles, delayed in 28% and normal in 19%. In patients with motor neurone disease, motor evoked potentials after electrical stimulation were absent in 62% of muscles, delayed in 10%, and normal in 29%; after magnetic stimulation, they were absent in 45% of muscles, delayed in 15%, and normal in 40%. The reason why magnetic stimulation reveals fewer abnormalities than electrical stimulation could be that magnetic stimulation repetitively discharges the pyramidal cells and, because of temporal summation mechanisms, produces more powerful excitatory potentials at the lower motoneurone synapse.