AN INDEX GROUP OF 66 FEMALE FELONS WAS SELECTED FROM THE ENTIRE CASELOAD OF THE MISSOURI STATE BOARD OF PROBATION AND PAROLE IN JULY 1969, PLUS ALL NEW CASES ADDED IN JULY AND AUGUST. THE SEMI-STRUCTURED INTERVIEW CONCERNED HISTORY OF CURRENT AND PAST ILLNESSES AND OPERATIONS, A DETAILED SYMPTOM INVENTORY AND FAMILY HISTORY, AND SCHOOL, JOB, AND MARITAL HISTORY. SUBJECTS WERE CONTACTED 5 YEARS AFTER THE LAST INDEX INTERVIEW FOR A FOLLOWUP THAT USED THE QUESTIONS. THE INTERVIEWER REMAINED BLIND TO THE INDEX INTERVIEW, RECORDS, AND DIAGNOSES UNTIL ALL FOLLOWUP INTERVIEWS AND DIAGNOSES WERE COMPLETED. DIAGNOSTIC CONCORDANCE WAS THEN ANALYZED ACCORDING TO A GENERAL CAUSAL MODEL. AS IN THE INDEX STUDY, THE DISORDERS FOUND IN THE GENERAL POPULATION RATES AT FOLLOWUP WERE ANTISOCIAL PERSONALITY, ALCOHOLISM, HYSTERIA, DRUG DEPENDENCE, AND HOMOSEXUALITY. DIAGNOSTIC STABILITY WAS SHOWN BY THE SIGNIFICANT CONCORDANCE BETWEEN THE INDEX AND FOLLOWUP DIAGNOSES FOR INDIVIDUAL SUBJECTS. PATH ANALYSIS ACCORDING TO A GENERAL CAUSAL MODEL REVEALED SUBSTANTIAL HETEROGENEITY NOT ONLY REGARDING THE SOURCES OF DIAGNOSTIC DISAGREEMENT BUT ALSO IN TERMS OF CLINICAL STABILITY. DISAGREEMENT ON HYSTERIA, ALCOHOLISM, AND DRUG DEPENDENCE RESULTED MOST FREQUENTLY FROM THE EFFECTS OF INCONSISTENT SELF-REPORTS. BORDERLINE SEVERITY MOST INFLUENCED THE DIAGNOSIS OF ANTISOCIAL PERSONALITY, AND TEMPORAL FACTORS THE DIAGNOSIS OF ANXIETY NEUROSIS. TAKING ERRORS OF MEASUREMENT INTO ACCOUNT, THE TEMPORAL STABILITY OF CLINICAL STATUS WAS HIGH FOR HOMOSEXUALITY AND DRUG DEPENDENCE, MODERATE FOR ANTISOCIAL PERSONALITY AND ALCOHOLISM, AND LOW FOR HYSTERIA AND ANXIETY NEUROSIS. TABLES AND 25 REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED)