INDEPTH PERSONAL INTERVIEWS WITH 49 ARMED ROBBERS SERVING AT LEAST THEIR SECOND PRISON TERM WERE THE PRIMARY SOURCE OF DATA. THE INMATES' CRIMINAL RECORDS WERE USED TO VALIDATE SOME OF THE SELF-REPORTED INFORMATION. THREE TIME PERIODS IN THEIR CRIMINAL CAREERS -- JUVENILE, YOUNG ADULT, AND ADULT -- WERE CONSIDERED IN CHARTING THE OFFENDERS' BEHAVIOR. QUESTIONS COVERED SUCH AREAS AS FAMILY RELATIONSHIPS, SOURCES OF INCOME, EMPLOYMENT, FREQUENCY OF CRIMINAL ACTIVITY, MOTIVATIONS, ATTITUDES, ARRESTS AND CONVICTIONS, METHODS OF PLANNING AND EXECUTING CRIMINAL ACTS, INVOLVEMENT WITH DRUGS AND ALCOHOL, USE OF VIOLENCE, AND POSTRELEASE BEHAVIOR PATTERNS. THE SAMPLE REPORTED A TOTAL OF 10,500 CRIMES, OR AN AVERAGE OF 214 PER OFFENDER. BASED ON AN AVERAGE CRIMINAL CAREER OF 20 YEARS (APPROXIMATELY HALF OF THIS TIME IN PRISON), THE FIGURE OF 20 MAJOR FELONIES PER YEAR IS DERIVED. THE NUMBER OF SELF-REPORTED CRIMES COMMITTED PER MONTH OF 'STREET TIME' DECLINED NOTICEABLY AS THE SAMPLE POPULATION GREW OLDER. INTERVIEWEES WHO WERE INVOLVED WITH DRUGS AND ALCOHOL WERE THE MOST SERIOUS OFFENDERS IN EVERY CAREER PERIOD; THOSE INVOLVED WITH ALCOHOL ALONE TENDED TO COMMIT FEWER AND LESS SERIOUS CRIMES. THE TENDENCY TO COMMIT CRIMES PRIMARILY AS A SOURCE OF MONEY FOR DRUGS OR ALCOHOL SHOWED A MARKED INCREASE AS THE CAREER ADVANCED. TWO BROAD CATEGORIES OF OFFENDERS EMERGED FROM THE SAMPLE: THE 'INTERMITTENT' AND THE 'INTENSIVE.' THE INTENSIVES, ABOUT ONE-THIRD OF THE SAMPLE, WERE MORE CRIMINALLY ACTIVE AND MORE SKILLFUL AT AVOIDING ARREST; INTENSIVES COMMITTED 10 TIMES AS MANY CRIMES AS THE INTERMITTENTS BUT WERE 5 TIMES LESS LIKELY TO BE ARRESTED FOR ANY SINGLE CRIME. THE PRELIMINARY EVIDENCE SUGGESTS THAT INCAPACITATION, BY IMPRISONMENT, MAY BE THE MOST DIRECT ALTERNATIVE FOR REDUCING THE SOCIETAL COSTS CAUSED BY HABITUAL OFFENDERS. (AUTHOR ABSTRACT MODIFIED).