Postsynaptic potentials were recorded from motoneurons in the facial nucleus in response to stimulation of the vestibular and trigeminal nerves. The motoneurons were identified by antidromic activation from their peripheral axons. Disynaptic excitatory and inhibitory postsynaptic potentials (EPSPs and IPSPs) and mixed EPSP/IPSPs were recorded in response to vestibular nerve stimulation, ranging in latency from 0.9 to 2.1 ms, with most at 1.5 ms. Activity in secondary vestibular axons recorded within the facial nucleus occurred at a latency of 0.7-1.1 ms. The amplitudes of the vestibular postsynaptic potentials were small, generally less than a millivolt, but double shocks produced marked summation. The average time to peak of ipsilateral vestibular EPSPs, 1.1 ms, was faster than that of either ipsilateral IPSPs, 1.6 ms, or contralateral EPSPs, 1.4 ms. The double-spiked vestibular activity was detectable in double-peaked PSPs. Disynaptic EPSPs, ranging in latency from 2.0 to 3.0 ms, were recorded in response to trigeminal nerve stimulation. The average time to peak was 1.3 ms. The multiple-spiked activity of the trigeminal neurons was detectable in multipeaked EPSPs. Inhibitory ipsilateral effects (Vi IPSPs) were recorded twice as often as excitatory ipsilateral effects (Vi EPSPs), being found in 29% versus 15% of the motoneurons. Contralateral effects were found in 13% of the motoneurons studied, and almost all were excitatory. Analysis of synaptic potential shapes suggested that the excitatory and inhibitory vestibular synapses probably contact distal dendrites preferentially, with the excitatory connections being somewhat closer to the soma. The trigeminal inputs probably contact the facial motoneurons more extensively near the soma. Horseradish peroxidase was injected into the facial nucleus, and retrograde uptake by vestibular neurons was studied. The majority of filled vestibular neurons was ipsilateral to the injection site, especially in the medial vestibular nucleus, ventral y group, and supravestibular nucleus. On the contralateral side, filled vestibular cells were found almost exclusively in the medial nucleus. Filled cells were also noted in the trigeminal nucleus, predominantly ipsilaterally at all rostrocaudal levels. We have demonstrated monosynaptic projections to facial motoneurons from both vestibular and trigeminal nuclei. The trigeminal input is likely to be involved in facial reflexes, especially blinking and grimacing. The afferent vestibular population overlaps that going to the oculomotor and cervical motoneurons; these projections may be collaterals of single vestibular neurons.4+.