THE DISCUSSION OF CLIENTS AND CASELOADS RAISES SUCH QUESTIONS AS TO WHO IS PLACED ON PROBATION; WHETHER CASELOAD SIZE MAKES ANY DIFFERENCE IN PROBATION EFFECTIVENESS; AND WHETHER PLACEMENTS ON PROBATION ARE IN ANY RESPECT MORE EFFECTIVE, FOR ANY CLASSIFICATION OF PERSONS, THAN ALTERNATIVE SENTENCING DISPOSITIONS. IN CONSIDERING THE PROBLEM OF THE PREDICTION OF PROBATION OUTCOMES, METHODS AVAILABLE FOR DEVELOPMENT AND VALIDATION OF PROCEDURES FOR CLASSIFICATION OF PERSONS WITH RESPECT TO RISK OF PROBATION VIOLATION ARE REVIEWED, AND THE POTENTIAL USES OF SUCH PROCEDURES ARE DISCUSSED. IT IS CONCLUDED THERE IS LITTLE AGREEMENT ABOUT THE MOST USEFUL DEFINITION OF 'RECIDIVISM' AS APPLIED TO PROBATIONERS, AND A MODEL IS PROPOSED FOR THE USE OF 'RECIDIVISM' AS A CONCEPT IN PROBATION INFORMATION SYSTEMS IN SUCH A WAY THAT CONTINUOUS GUIDANCE CAN BE GIVEN TO PROBATION PROGRAM DEVELOPMENT AND MANAGEMENT. IN THE AREA OF TREATMENT IN PROBATION, SOME PROMISING RESULTS THROUGH TREATMENT PROGRAMS ARE REPORTED, BUT THE REVIEW MAINLY ILLUSTRATES THE COMPLEX ISSUES PRESENT IN THIS ARENA AND THE LIMITED KNOWLEDGE AVAILABLE. NOTES AND TABULAR DATA ARE PROVIDED.