Attitudes about the criminal sentence have changed. Until recently, the ideal of treatment dominated; the sentence was intended to rehabilitate, and sentencing and parole boards were supposed to have wide discretion so that they could tailor the sentence to the offender's needs. However, the advocates of treatment are now on the defensive. A survey of most of the major rehabilitation programs between 1945 and 1967 concludes that with a few isolated exceptions, the rehabilitative efforts reported so far have had no appreciable effect on recidivism. In keeping with this new trend in thinking, Harvard's James Q. Wilson suggests that most serious crimes are committed by a relatively small number of repeaters who, because of the large number of crimes they perpetuate, sooner or later are caught and convicted. Wilson argues that if prison sentences, rather than probation, were imposed in such cases, the incapacitative payoff would be substantial. However, Wilson's preoccupation with crime control excludes certain aspects of justice. Instead, sentences should be decided according to a principle of commensurate deserts. Penalties must be scaled in accordance with the gravity of the offense, and departures from the deserved sentences should be impermissible. A sentencing system based on this concept would have the following principal features: (1) the likelihood that the offender might return to crime would be irrelevant to the choice of sentence, (2) indeterminancy of sentence would be abolished, (3) sentencing discretion would be sharply reduced, (4) imprisonment would be limited to serious offenses, and (5) less severe penalties than imprisonment would thus be imposed for the majority of the criminal justice system's caseload. One footnote is inlcuded.