Facial nerve motor neurography was performed at various times after the onset of Bell's palsy in 97 patients. Stimulation of the facial nerve was performed (1) electrically in the fossa stylomastoidea (ElStim), and (2) magnetically in the labryinthine segment of the facial canal (MagStim), evaluating different coil positions over the skull. Additionally, the face-associated motor cortex was stimulated magnetically in 47 patients (CxStim). A marked reduction of the amplitudes of the compound muscle action potentials (CMAP) evoked by MagStim on either m. nasalis or mentalis, or both, was observed which was clearly more pronounced than the amplitude reduction to ElStim. This discrepancy occurred very early during the disease, the mean amplitude (expressed in percent of the amplitude on the unaffected side) being 82% (S.D. 9.1) for ElStim and 1% (2.7) for MagStim at days 0-4. It persisted for several months, often when facial nerve function had recovered to normal, as assessed by clinical observation, ElStim, and CxStim. This amplitude decrease to MagStim, which appears to be related to a locally enhanced stimulation threshold of the facial nerve, is a very sensitive and reproducible finding in Bell's palsy. It may prove specific of the disorder, of diagnostic value, and of interest in the follow-up of patients during treatment trials.