THE CONCEPT OF A UNIQUELY 'CORRECT' GEOGRAPHIC PATTERN OF CRIME IS DISCUSSED. THE DEFICIENCIES OF THE ANNUAL CRIME STATISTICS PUBLISHED BY THE FBI BOTH AS STATEMENTS OF FACT AND AS BASES FOR SOCIAL COMMENT AND INTERPRETATION ARE ILLUMINATED. PROBLEMS ARISING IN THE SEARCH FOR EMPIRICAL ASSOCIATIONS BETWEEN CRIME AND OTHER CONDITIONS INVOLVED ECOLOGICAL CORRELATION ON DIFFERENT SPATIAL SCALES ARE ILLUSTRATED. ALTHOUGH NO ATTEMPT IS MADE TO ACCOUNT FOR GEOGRAPHICAL PATTERNS OF CRIME, IT IS SUGGESTED THAT THE CAUSE-AND-EFFECT LINKS ARE MORE COMPLEX THAN WOULD APPEAR FROM THE EXPECTED SPATIAL COINCIDENCE WITH OTHER SOCIAL PROBLEMS. RECOMMENDATIONS ARE OFFERED ON WAYS IN WHICH THE REPORTING OF CRIME AND ITS INTERPRETATION COULD BE IMPROVED. (AUTHOR ABSTRACT MODIFIED)