All cases of sexual aggression found in the Montreal police computerized system from January 1993 to May 1996 were included in the study, yielding a sample of 2,636 adult men. The second dataset used in the study was based on the characteristics of aggressions/aggressors who were transferred to the Centre Regional de Reception (CRR) in the Province of Quebec. Between February 1995 and July 1996, 209 sexual aggressors were processed by the CRR. Of this number, 199 agreed to participate in the study. The third dataset was based on a sample of Federal inmates who were treated in a specialized unit for sexual aggressors at the Institut Philippe Pinel de Montreal (IPPM) between 1978 and 1991 (n=93). Generally, the study findings support the idea that substantial differences exist between samples of aggressive sexual offenders selected along the penal continuum that ranges from police detection to Federal imprisonment to psychiatric treatment. The principal factors involved in the filtering process were the seriousness of the offense and the treatability of the offender, both estimated indirectly by using victim characteristics and the objective seriousness of the crime. There was clearly an overrepresentation of older offenders in penal institutions, presumably because of the more extensive prior criminal record of the older participants. The treatment sample involved younger offenders, perhaps because their offenses were more violent than the older offenders, or because treatment was more likely to be available to younger offenders. Also, offenders who committed a sex crime against children were more likely to be imprisoned, and they were more likely to be selected for treatment in a psychiatric facility while imprisoned. The objective seriousness of the offense, as measured by the presence of a weapon, was important in determining the processing of cases. Such offenders were more likely to be selected and sent to a penitentiary. Suggestions are offered for future research. 4 tables and 36 references