1. The silent period evoked in the first dorsal interosseous (FDI) muscle after electrical and magnetic transcranial stimulation (TCS), electrical stimulation of the cervicomedullary junction and ulnar nerve stimulation was studied in ten healthy subjects. 2. With maximum-intensity shocks, the average duration of the silent period was 200 ms after electrical TCS, 300 ms after magnetic TCS, 43 ms after stimulation at the cervicomedullary junction and 100 ms after peripheral nerve stimulation. 3. The duration of the silent period, the amplitude of the motor-evoked potential, and the twitch force produced in the muscle were compared at increasing intensities of magnetic TCS. When the stimulus strength was increased from 30 to 70% of the stimulator output, the duration of the silent period lengthened as the amplitude of the motor potential and force of the muscle twitch increased. At 70 to 100% of the output, the amplitude of the motor potential and force of the muscle twitch saturated, whereas the duration of the silent period continued to increase. 4. Proximal arm muscle twitches induced by direct electrical stimulation of the biceps and extensor wrist muscles produced no inhibition of voluntary activity in the contracting FDI muscle. 5. The level of background activation had no effect on the duration of the silent period recorded in the FDI muscle after magnetic TCS. 6. Corticomotoneurone excitability after TCS was studied by means of a single magnetic conditioning shock and a test stimulus consisting either of one single magnetic shock or single and double electrical shocks (interstimulus interval 1.8 ms) in the relaxed muscle. A conditioning magnetic shock completely suppressed the response evoked by a second magnetic shock, reduced the size of the response evoked by a single electrical shock but did not affect the response evoked by double electrical shocks. Inhibition of the test magnetic shock was also present during muscle contraction. 7. Our findings indicate that the first 50 ms of the silent period after TCS are produced mainly by spinal mechanisms such as after-hyperpolarization and recurrent inhibition of the spinal motoneurones. If descending inhibitory fibres contribute, their contribution is small. Changes in proprioceptive input probably have a minor influence. From 50 ms onwards the silent period is produced mainly by cortical inhibitory mechanisms.