Excessive olivo-cerebellar synchrony is implicated in essential tremor. Because synchrony in some networks is mediated by gap junctions, we examined whether the gap junction blockers heptanol, octanol, carbenoxolone, and mefloquine suppress tremor in the mouse harmaline model, and performed an open-treatment clinical study of mefloquine for essential tremor. Digitized motion was used to quantify tremor in mice administered harmaline, 20 mg/kg s.c. In mice the broad-spectrum gap junction blockers heptanol, octanol (350 mg/kg i.p. each), and carbenoxolone (20 mg/kg) suppressed harmaline tremor. Mefloquine (50 mg/kg), which blocks gap junctions containing connexin 36, robustly suppressed harmaline tremor. Glycyrrhizic acid (related to carbenoxolone) and chloroquine (related to mefloquine), which do not block gap junctions, failed to suppress harmaline tremor in mice. Clinically, tremor was assessed with standard rating scales, and subjects asked to take 62.5, 125, and 250 mg mefloquine weekly for 12 weeks at each dose. None of the four human subjects showed a meaningful tremor reduction with mefloquine, likely because clinical levels were below those required for efficacy. In view of recent genetic evidence, the anti-tremor mechanism of these compounds is uncertain but may represent a novel therapeutic target, possibly involving gap junctions other than those containing connexin 36.