We studied the effects of longitudinal tendon vibration on the stretch reflex of the soleus and gastrocnemius muscles in 11 decerebrate cats. Vibration was applied at amplitudes (40-80 micrometer) and frequencies (120-250 HZ) sufficient to provide a strong tonic vibration reflex. In keeping with previous reports, we found that during an established tonic vibration reflex, the force and emg response to superimposed ramp and hold stretch are largely suppressed. This suppression is most obvious during the dynamic phase of stretch where it gives rise to a complex force response resembling that of active areflexic muscle. If stretch initiation is delayed until after vibration is terminated, the suppressed effects of vibration persist for 5 s or more. These suppressive effects are marked in the first 200 ms, and then decay gradually over the ensuing time period, paralleling the decline in emg and force which follows vibration offset. Simultaneous recordings from homonymous Ia afferents showed that this suppression persists even though the stretch responsiveness of primary spindle endings has returned to normal immediately following the end of vibration. When stretch is initiated shortly after vibration commences, the suppressive effects are first evident at 50-100 ms latency, but are not well established until 1 s or more after vibration onset. Tests of monosynaptic transmission using small amplitude tendon taps or electrical stimulation of synergist nerves to activate Ia fibers revealed that reductions in the magnitude of the response following vibration are usually modest (12% mean reduction at 50 ms, n = 5), and they are quite sensitive to the initial level of excitation of the motoneuron pool. These reductions were also rather shortlived, being largely completed within 500 ms of vibration offset. Although the relative contributions of presynaptic and postsynaptic inhibition are not readily dissociated in this type of experiment, it is likely that the magnitude of presynaptic inhibition is quite small. We argue that the effects of vibration on the stretch reflex are best explained by invoking an excitatory autogenetic projection from Ia interneurons to extensor motor neurons, which lies in parallel with the Ia monosynaptic projection. In order to account for the vibratory suppression, we propose that these interneurons are driven to saturation by vibration. When vibration ceases, the discharge rate of these interneurons declines with a prolonged time-course that coincides with the recovery of stretch responsiveness. This recovery would contribute to the return of stretch reflex force.