1. Vibration was applied longitudinally to the fully innervated soleus muscle of the decerebrate cat by attaching its tendon to a vibrator. Vibration at frequencies of 50-500/sec with amplitudes of 10 mu upwards caused the muscle to contract reflexly for as long as the vibration was maintained. The response was recorded myographically by a myograph mounted upon the vibrator, and electromyographically by gross ;belly-tendon' leads. The reflex contraction produced several hundred g wt. of tension and involved too many motor units for their discharges to be separable. The maintained reflex was abolished by making the preparation spinal or by anaesthetizing it with pentobarbitone, but it persisted after removing the cerebellum.2. The minimum latency for the appearance of the reflex response at the beginning of a period of vibration was about 10 msec. The latency of cessation of the response at the end of vibration was similarly short.3. On increasing the amplitude of vibration at any particular frequency in the range 100-300/sec the resulting reflex tension increased to an approximate plateau for amplitudes of vibration of 100-200 mu. Further increase in the amplitude decreased the size of the contraction, though there was no such reduction in records of the ;integrated' electromyogram.4. Such large amplitudes of vibration also reduced the tension, and shortened the duration, of a twitch contraction of the muscle elicited by stimulating its nerve. The strength of a tetanic contraction was much less affected by vibration than was that of the twitch contraction, and the muscle action potential elicited by stimulation of the nerve was unaffected. Thus, large-amplitude vibration influenced the contractile mechanism of the muscle (cf. Buchtal & Kaiser, 1951).5. Increasing the frequency of vibration increased the value of the plateau tension reached on increasing the amplitude. The effect was, however, relatively small and the largest increase seen was 3 g wt. of contractile tension per c/s increase in vibration frequency.6. The primary afferent ending of the muscle spindle is considered to be the receptor whose excitation leads to the reflex response to vibration. The vibration reflex thus appears to be the well-known stretch reflex, elicited by a rather unusual form of stretching. The size of the vibration reflex and its variation with frequency are discussed in relation to the servo theory of muscular contraction.