Neuroleptic medication was abruptly discontinued in 24 male chronic schizophrenic patients who were subdivided on the basis of their history of illness into Kraepelinian (n = 8) and non-Kraepelinian (n = 16) subgroups. These patients were kept drug free for 6 weeks and then returned to treatment with haloperidol, 20 mg/day. Half of the non-Kraepelinian patients developed exacerbations of their symptoms, which quickly resolved when they were returned to medication, while none of the Kraepelinian patients showed a worsening of symptomatology. On-medication clinical severity failed to predict risk for exacerbation, with severity of exacerbation predicting the amount of improvement when returned to medication. The Kraepelinian patients were found to be much less variable than the non-Kraepelinian patients in their symptoms during both medication manipulations, suggesting that medication truly has a negligible effect on them.