We conducted a double-blind crossover trial comparing gabapentin (up to 900 mg/day) to baclofen (up to 30 mg/day) as therapy for acquired nystagmus in 21 patients. We measured visual acuity and the nystagmus before, and at the end of, 2 weeks on each medication. For a group of 15 patients with acquired pendular nystagmus (APN), visual acuity improved significantly with gabapentin, but not with baclofen. Gabapentin significantly reduced APN median eye speed in all three planes, but baclofen did so only in the vertical plane. In 10 patients with APN, the reduction of nystagmus with gabapentin was substantial and 8 of these elected to continue taking the drug. In 6 patients with downbeat or torsional downbeat nystagmus, changes in median slow-phase eye speed were less consistent with both drugs, either increasing or decreasing, and being dependent on viewing conditions. Only 1 patient showed consistent reduction of median eye speed, and this was achieved by either drug. Our findings suggest that gabapentin may be an effective treatment for many patients with APN and that occasional patients with downbeat nystagmus will respond to gabapentin or baclofen.