Experimentation with early release of dangerous offenders as early as 1837 is described. Public outrage at the commission of serious, violent offenses by released offenders is seen as stemming from confused policies of certain Australian States concerning the prediction of dangerous and habitual offenders. It is argued that parole should not be granted if dangerousness cannot be predicted. In addition, probation should not be used for purposes of therapy or counseling of still dangerous offenders. It is also contended that the evaluations of dangerousness rendered by prison officials and guards, who are in almost daily contact with prisoners, are of more value than are the opinions of mental health professionals. Classification committees, consisting of prison administrators and mental health personnel, are urged to extend their investigations of prisoners beyond the prisoner's file in advising the prison chief on whom to release to community treatment programs. An analysis of dangerous offenders remaining within Australian prisons indicates that long-term offenders (those doing 5 years or more) caused most prison superintendents less trouble than all other classes of prisoners. However, trouble has resulted among dangerous long-term offenders due to inadequate and/or overcrowded prisons and from decsions concerning parole and work release. Consequently, it is recommended that 1) parole and work release programs be established as nonjusticiable issues, and 2) penal institutions be modernized and constructed to house smaller number of prisoners, thereby facilitating the formation of programs to assist prisoners in making constructive use of their prison time. It is also recommended that prison officials be educated in the psychological and sociological determinants of human behavior to promote understanding of the reasons people commit violent crime. Thirty-six references are provided.