THE AUTHOR CONDUCTED A 1972 STUDY IN PONTOISE, FRANCE, INVOLVING TWO SAMPLES, THOSE WHO HAD RECEIVED A SUSPENDED SENTENCE AND THOSE WHO HAD SERVED A FULL PRISON TERM; HIS RESULTS SHOWED THAT 3 PERCENT OF THE FIRST GROUP AND 2.7 PERCENT OF THE SECOND GROUP BECAME RECIDIVISTS. THE PERCENTAGE INCREASED TO 27 PERCENT WHEN THE SAMPLE CONSISTED OF THOSE WHO RECEIVED A PROBATIONARY SUSPENSION ON TOP OF A SIMPLE SUSPENDED SENTENCE (I.E., THOSE WHO BECAME RECIDIVISTS WITHOUT EVER SERVING A PRISON TERM). HOWEVER, A 1972 STUDY BY THE CENTRE NATIONAL D'ETUDES ET DE RECHERCHES PENITENTIAIRES (CNERP: NATIONAL CENTER FOR PENITENTIARY STUDIES AND RESEARCH) CLAIMED A RECIDIVISM RATE OF 56.1 PERCENT FOR THOSE SERVING PRISON SENTENCES FOR THEIR CRIMES. THE AUTHOR OF THIS ARTICLE TAKES ISSUES WITH THE CNERP FINDING ON THE BASIS OF BOTH THE UNDERLYING CONCEPT OF RECIDIVISM AND THE METHODOLOGY OF THE STUDY. THE CNERP USED A RETURN TO PRISON WITHIN 10 YEARS OF THE DATE OF COMPLETING SENTENCE AS A DEFINITION OF RECIDIVISM AND MADE NO DISTINCTION BETWEEN FIRST OFFENDERS, RECIDIVISTS, AND MULTIRECIDIVISTS. THE FRENCH PENAL CODE, HOWEVER, VIEWS ONLY A 5-YEAR PERIOD AFTER INITIAL OFFENSE, CONSIDERS RECIDIVISM FOR CASES IN WHICH NO PRISON TERM IS SERVED, AND DISTINGUISHES BETWEEN THE THREE CLASSES OF OFFENDERS. BECAUSE THE CNERP COMBINED ALL THREE TYPES OF OFFENDERS IN ITS SAMPLE, THE AUTHOR OF THE ARTICLE DOES NOT CONSIDER THEIR SAMPLE REPRESENTATIVE AND POINTS OUT THAT, IF ONE MIXES FIRST OFFENDERS WHO PRACTICALLY NEVER BECOME RECIDIVISTS WITH RECIDIVISTS WHO PRACTICALLY ALWAYS RECIDIVATE, ONE INDIVIDUAL OUT OF TWO CAN BE SHOWN TO RETURN TO PRISON WITHIN 10 YEARS. THE AUTHOR'S SECOND STUDY SET OUT TO EXPLORE WHETHER THE PENAL CODE'S DISTINCTION BETWEEN OFFENDERS WAS VALID AND WHETHER THE SENTENCING INCREASES PROVIDED FOR RECIDIVISTS WERE JUSTIFIED. THE SAMPLE CONSISTED OF MOST CASES HEARD IN PARIS 'POLICE COURTS' (TRYING MISDEMEANORS AND PETTY OFFENSES) DURING THE WEEK OF NOVEMBER 22-28, 1965, WITH A 5-YEAR PERIOD SET FOR DETERMINING RECIDIVISM. VARIABLE ANALYSIS WAS PERFORMED FOR ALL THOSE WHO SERVED PRISON TERMS, AND FINDINGS FOR BOTH RECIDIVISTS AND FIRST OFFENDERS WHO DID NOT REPEAT WERE COMPILED. SIGNIFICANT CONCLUSIONS INCLUDED: (1) IN RELATION TO AGE, FIRST OFFENDERS WHO DO NOT BECOME RECIDIVISTS ARE DISTRIBUTED ALONG A NORMAL CURVE, IMPLYING THAT 'DELINQUENCY OF OCCASION' (NONREPEATED) IS NOT SPECIFICALLY JUVENILE IN NATURE; (2) THE MINIMUM PENALTY FORESEEN BY THE PENAL CODE OFFSETS THE PROPENSITY TO RECIDIVATE FOR MORE THAN 95 PERCENT OF FIRST OFFENDERS; (3) THE POPULATION OF HABITUAL OFFENDERS COMES INTO THE OPEN FOLLOWING A HALF-LIFE PERIOD OF 7 YEARS BEGINNING WITH 14.5 YEARS-OF-AGE, I.E., OVER THE COURSE OF 7 YEARS, HALF OF THOSE WHO WERE RECIDIVISTS WILL RECIDIVATE, ETC; AND (4) THE PROPENSITY TO RECIDIVATE AMONG HABITUAL OFFENDERS DIMINISHES BY 2 PERCENT PER YEAR BEGINNING WITH 14.5 YEARS-OF-AGE. THE AUTHOR FEELS THAT, AS FIRST OFFENDERS RARELY BECOME RECIDIVISTS, SUSPENDED SENTENCES SHOULD BECOME THE GENERAL RULE FOR THESE OFFENDERS AND ATTENTION SHOULD BE TURNED TO REPARATION OF DAMAGES TO THE VICTIM; HOWEVER, AS RECIDIVISTS OFTEN BECOME HABITUAL RECIDIVISTS, THE LAW SHOULD APPLY ALL OF THE FORCES WITHIN ITS POWER TO DISCOURAGE RECIDIVISM THROUGH REPRESSION AND STIFFER SENTENCES. FINDINGS ARE PRESENTED IN TABULAR AND GRAPHIC FORM. -- IN FRENCH.