Because of neuroleptics' potential long-term side effects, the authors conducted a pilot study of an alternative to maintenance medication for stable schizophrenic outpatients. The doses of 19 patients were gradually reduced to zero over 8 weeks, and medication was then given only when a patient experienced early signs of relapse. The patients attended weekly group therapy and were closely monitored for prodromal signs, especially at times of stress; significant others helped observe patients. Five patients experienced increased symptoms during the drug washout period and were dropped from the study; of the remaining 14, 10 remained stable on the intermittent medication protocol.