In this article, research on the natural course of illness among first-admission schizophrenic patients is reviewed from an epidemiological perspective. Three types of studies are considered: statistical reports dating primarily from the preneuroleptic era; long-term followback studies; and more recent prospectively designed cohort studies. Although relatively more first-admission patients have a positive course than do multiple admissions patients, the findings confirm the substantial heterogeneity in course and outcome. Methodological improvements in first-admission research are suggested, including separating the analyses of first-episode patients from those with past psychotic experiences; employing longitudinal methods for deriving diagnosis; including patients with drug and alcohol problems where appropriate; enlarging the samples either by multisite or multicenter collaborations or by pooling data across studies; and obtaining better data on treatment experiences in naturalistic research.