THE PAPER EXAMINES INITIALLY THE GENERAL RULES OF DISTRIBUTION OF AN ABSTRACT SOCIAL GOOD THAT WOULD HAVE TO BE ADHERED TO UNDER ALTERNATE FORMS OF SOCIAL WELFARE. IT THEN CONSIDERS THE IMPLICATIONS THAT THIS MIGHT HAVE FOR A SPECIFIC URBAN SERVICE, THE POLICE SERVICES IN NEW YORK CITY. THE AUTHOR SUGGESTS WAYS IN WHICH THE PRODUCT OF POLICE SERVICES MIGHT BE EVALUATED AND EXAMINES APPROACHES TO THEIR QUALIFICATION AND THE DEGREE TO WHICH PRESENT DISTRIBUTION IS ALSO ASSESSED. A THREE-EQUATION SIMULTANEOUS EQUATION SYSTEM IS DEVELOPED TO DESCRIBE THE PRODUCTION OF CRIME, THE PRODUCTION OF POLICE SERVICES, AND THE PRESENT PRACTICES DETERMINING THE DISTRIBUTION OF INPUTS. EACH IS DEVELOPED SEPARATELY, CONSIDERING THE PREVAILING THEORIES ON THE CAUSATION OF CRIME, PARTICULARLY THE ECONOMIC THEORY OF CRIME. THE ESTIMATION OF THE MODEL FOR VARIOUS CRIMES AND THE ECONOMETRIC DIFFICULTIES ENCOUNTERED ARE DISCUSSED. THE PRIMARY FINDING IS THAT POLICE DO HAVE A SIGNIFICANT IMPACT ON ROBBERIES, BUT A LESS CLEARLY OBSERVABLE EFFECT ON OTHER CATEGORIES OF CRIME. THE FINAL CHAPTER ASSESSES THE IMPACT OF A CHANGE IN THE DISTRIBUTION OF MANPOWER ON CRIME RATES IN VARIOUS PARTS OF THE CITY. THE AUTHOR RECOMMENDS A SHIFT IN MANPOWER DISTRIBUTION THAT IS MORE EFFICIENT IN TERMS OF MINIMIZING CRIME AND MORE EQUITABLE IN THAT IT WOULD RESULT IN A MORE EVEN DISTRIBUTION OF CRIME. TABLES OUTLINE PREDICTED ROBBERY RATES IN DIFFERENT AREAS WITH THE ALTERNATE MANPOWER ALLOCATIONS. APPENDIXES PROVIDE DESCRIPTIONS OF SOURCES OF THE DATA USED; DEFINITIONS OF THE VARIABLES USED IN REGRESSION; A GLOSSARY; AND A MAP OF THE PRECINCTS OF NEW YORK CITY. REFERENCES ARE SUPPLIED. (VDA)