A MAJOR ASSUMPTION IS THAT CRIMINALS ARE NOT VERY DIFFERENT FROM NONCRIMINALS AND THAT BOTH PERCEIVE AND RESPOND TO THEIR ENVIRONMENT IN TERMS OF PARTICULAR, INDIVIDUAL MOTIVES. THE STUDY SAMPLE WAS BASED ON KNOWN PROPERTY OFFENDERS CURRENTLY RESIDING IN THREE OF OKLAHOMA'S CORRECTIONAL INSTITUTIONS AND HAVING BEEN CONVICTED OF BURGLARY, ROBBERY, OR LARCENY. A STRATIFIED SAMPLING PROCEDURE WAS USED TO COLLECT A MATCHING NONCRIMINAL SAMPLE. GESTALT, FIELD, AND PERSONAL CONSTRUCT THEORIES WERE USED AS A BASIS FOR MEASURING INDIVIDUAL AREA PERCEPTIONS, AND THE SEMANTIC DIFFERENTIAL WAS ADOPTED AS A PERCEPTION MEASUREMENT TOOL. FINDINGS INDICATE THAT CRIMINALS' AREA IMAGES ARE SHAPED BY THEIR VOCATIONAL INTERESTS SO THAT THEY PERCEIVE MORE CRIME ACTIVITY, POLICE PROTECTION, AND GREATER DIFFICULTY OF COMMITTING CRIME IN THE CITY THAN DO NONCRIMINALS. IN ADDITION, CRIMINALS, PARTICULARLY BLACKS, ARE LIKELY TO CHOOSE A CRIME TARGET IN A FAMILIAR AREA. WHITES ARE MORE CONCERNED WITH THE RISK AND DIFFICULTY OF COMMITTING THE OFFENSE THAT AN AREA PRESENTS. BOTH GROUPS OF CRIMINALS TENDED TO AVOID HIGH STATUS NEIGHBORHOODS. STATISTICAL ANALYSES INDICATE THAT THE PATTERNING OF CRIME IS SYSTEMATICALLY RELATED TO AREA IMAGES HELD BY CRIMINALS, THAT THESE IMAGES ARE THEMSELVES DIFFERENTIABLE, AND THAT CERTAIN IMAGE COMPONENTS ARE MORE POWERFUL THAN OTHERS IN INFLUENCING CRIMINAL BEHAVIOR. IMPLICATIONS FOR PUBLIC POLICY AND THE CRIME PREVENTION THROUGH ENVIRONMENTAL DESIGN (CPTED) MOVEMENT ARE DISCUSSED. TABULAR AND GEOGRAPHICAL DATA, A BIBLIOGRAPHY, CHAPTER NOTES, AND AN APPENDIX SHOWING THE GEOGRAPHICAL AREAS STUDIED ARE INCLUDED. (AOP)