We divided 19 patients with essential tremor into two subtypes according to clinical characteristics of the tremor. Ten patients had pure postural tremor distributed in the hand(s), head, and face (group A). Nine patients had tremor extending to the voice or leg(s), associated with resting tremor and/or hyperkinesie volitionnelle of the hand(s) (group B). Their ages, the age of onset, and the duration of illness were not different between the two groups. Electrophysiologically, the tremor of group A patients had higher frequencies than that of group B patients, and had synchronized activities for antagonistic muscles. Four of group B patients had reciprocal antagonistic activities of the tremor. Inactive phase of tremor induced by an electrically-evoked muscle twitch was invariably within the range of the physiological silent period for group A patients, and prolonged beyond the range for four of group B patients. Pharmacologically, 78% of group A patients responded well to beta-blocker, which was effective for 25% of group B patients. Sixty per cent of beta-blocker-resistant group B patients responded well to phenobarbital. In conclusion, a peripheral mechanism, presumably beta-adrenergic drive, is important for the tremor in group A patients, while central pathogenic mechanisms are more important for the tremor of group B patients.