1. In micro-electrode recordings from the human peroneal and tibial nerves, the responses of thirty-two primary spindle endings, thirteen secondary spindle endings and three Golgi tendon organs were studied during vibration of the tendons of the receptor-bearing muscles in the leg. The amplitude of the applied vibration was 1-5 mm and the frequency was varied from 20 to 220 Hz. As checked with e.m.g. and torque measurements, the muscles of the leg were relaxed during the sequences analysed. 2. Providing that the vibrator was accurately applied, all endings responded with discharges phase-locked to the vibration cycles, the discharge rates being at the vibration frequency or at subharmonics of that frequency. The response to vibration was of abrupt onset and offset, was maintained for the duration of vibration, and was not subject to fluctuation with changes in attention or with remote muscle contraction. 3. The maximal discharge rate that could be achieved varied from one ending to the next, and increased with the length of the receptor-bearing muscle. For endings driven at their maximal rate an increase in vibration frequency produced a decrease in discharge rates as the ending changed to a subharmonic pattern of response. The converse occurred on decreasing vibration frequency. 4. For any given muscle length, primary endings could generally be driven to higher rates than secondary endings but there was a wide range of responsiveness within each group and a significant overlap between the groups. At medium muscle length, the most responsive primary endings could be driven up to 220 Hz but secondary endings did not reach discharge rates higher than 100 Hz. 5. With combined vibration and passive movements, primary endings exhibited maximal vibration responsiveness during the stretching phases, sometimes firing twice per vibration cycle. During the shortening phases, however, they usually ceased responding to the vibratory stimulus. The vibration responsiveness of secondary endings was not potentiated to the same extent by on-going muscle stretch or reduced to the same extent by on-going muscle shortening. Thus, during shortening, secondary endings may be more responsive than primary endings. 6. The responses of primary endings to tendon taps were reduced during muscle vibration, a reduction which probably contributes to vibration-induced suppression of tendon jerks. Additionally, as the muscle shortened after tendon percussion, there was a transient pause in the response to vibration.