Using the first four sweeps of the British Crime Survey, patterns of property and personal crime victimization are investigated, and the interdependence of incidents is established: the probability of a further victimization increases with each subsequent victimization. The authors compare the observed distribution and a hypothetical distribution which assumes that crime incidents are randomly distributed throughout a small number of chronic victims. The authors also make recommendations for the way official crime statistics are recorded: (1) Make short-term provision for incorporation of data on crime concentration; (2) Incorporate data which identify repeat events suffered by the same people or at the same location; (3) Give due weight to the phenomenon and circumstances of repeat victimization; (4) Use information about crime concentration in the rapprochement of crime prevention and victim support services; and (5) Encourage crime prevention endeavors concentrating on repeat victimization as having the virtues of prioritizing effort upon places and people at high risk of crime. Tables, references