To describe the prevalence of DSM-IV disorders and the pattern of comorbidity in a population-based sample of 7- to 14-year-old Brazilian schoolchildren. Random sampling of schools (stratified into private, public rural, and public urban) was followed by random sampling of pupils from school lists. In 2000-2001, a total of 1,251 children were assessed for DSM-IV diagnoses using the Development and Well-Being Assessment, a structured multiinformant assessment supplemented by verbatim reports reviewed by clinicians. The response rate was 83%. The overall prevalence of DSM-IV disorders was 12.7% (95% confidence interval = 9.8%-15.5%), with 3.5% of children being assigned as not otherwise specified rather than operationalized diagnoses. The overall prevalence of psychiatric disorder was significantly higher than in a British study with the same measures and diagnostic procedures (12.7% versus 9.7%, p =.02). Approximately one in eight schoolchildren in the study area in the southeast of Brazil have psychiatric disorders involving a level of distress or social impairment likely to warrant treatment. The gulf between need and provision is currently vast.