1. Statistically significant criteria distinguishing good, guarded and poor prognoses from each other were determined on the basis of the clinical records of 1,036 drug addicts admitted to the U. S. Public Health Service Hospital at Lexington, Kentucky, during the fiscal year July 1, 1936 to June 30, 1937.2. Case histories of composite individuals most representative of their respective prognoses are presented.3. Follow-up studies made of the patients after release from the hospital reveal that 43.8 per cent with a good prognosis, 36.9 per cent with a guarded prognosis, and 20.7 per cent with a poor prognosis were still abstinent.4. For practical purposes the prognosis may be reduced to two categories, poor and guarded.