Abstract A recruitment program for ‘normal’ control subjects is evaluated with regard to the yield of ‘healthy’ subjects, the degree to which subjects remain healthy over a number of years, and the rate of subject participation in research. Only 22.4% had no lifetime history of mental illness (Never Mentally Ill), 6.8% met the criteria for one episode of a minor mental disorder (MMD), 41.6% were currently healthy but had a more serious history of mental illness and 29.9% were currently mentally ill. Diagnostic follow-up found that subjects in the NMI and MMD categories did not differ with regard to the rate of interval episodes in themselves or in family members over an average of 27 months. However, subjects in the MMD category were more likely to have a positive family history of mental illness ( χ 2(1) = 21.34; P < 0.001). The longitudinal course of mental health in a combined group of NMI and MMD subjects was predicted by sex of the subject ( χ 2(1) = 4.04; P < 0.05), but not by age or family history of mental illness. These findings suggest that investigators selecting ‘healthy’ control subjects consider the probability that a currently healthy individual will have episodes of mental illness in the future.