1. The effect of muscle stretch on the EMG response from the stretched muscle to transcranial magnetic stimulation of the motor cortex was studied in eight subjects. Muscle stretch was produced by increasing the torque of a motor acting through a lever which was held at constant position by a flexion force of the index and middle fingers. EMG responses were recorded from fine-wire electrodes inserted into flexor digitorum profundus muscle in the forearm. They consisted of a spinal latency component and a long-latency component which could in some subjects be separated into an early and a late phase. 2. In four subjects, four intervals between the stretch and the cortical stimulus were explored using three intensities of cortical stimulation. At all three intensities, when the magnetic cortical stimulus was timed to produce an EMG response in the period of the later part of the long-latency stretch reflex the response was larger than when it was timed to produce a response in the period of the short-latency spinal reflex or when superimposed on the tonic muscle activity used to resist the standing torque of the motor. 3. When the intensity of magnetic cortical stimulation was reduced so that it was just below threshold to produce an EMG response in the short-latency reflex period or on the background tonic EMG activity, it still was capable of producing a response when superimposed on the long-latency stretch reflex. 4. In four subjects the time course of this effect was studied in more detail using only one intensity of magnetic cortical stimulus set to be just above threshold to produce a response in tonically active muscles. The time course of the facilitatory effect was similar to the time course of the later part of the long-latency stretch reflex. From these data it was not possible to determine whether the early part of the long-latency stretch reflex also was accompanied by the facilitatory effect since this component was present in only one of the four subjects. 5. The facilitatory effect persisted after the ulnar and median nerves were totally blocked at the wrist by injections of local anaesthetic. This suggests that inputs from muscle receptors of the stretched muscle contribute to the effect.4=