The employment of a racial/ethnic perspective on crime is something that is taken for granted in much of the American-inspired social science literature. American criminologists appear to agree that there is an ethnic dimension to the crime problem in the USA, as is evidenced inter alia in the criminological literature on American gangs where there is apparent consensus that the majority of gangs are ethnically homogeneous (e.g. Knox 1991: 2, 18; Klein 1995: 105, 106). Gang formation is viewed as a coping mechanism employed by young people from ethnic minorities in the context of oppressive majority rule (Spergel 1995). But ethnic homogeneity in the choice of co-offenders in America is not limited to gangs. In other contexts too it is relatively unusual for individuals from different racial backgrounds to commit offences together. According to the statistics produced by the US Bureau of Justice (1984 table 49), individuals from different ethnic backgrounds were involved in only 6 per cent of reported violent offences. Reiss (1988: 135) compares these statistics with data on the gender composition of groups of perpetrators linked with these same offences and finds that it is less common for perpetrators of violent offences to come from different racial backgrounds than it is for them to come from the different gender groups, which is in itself also very unusual. Reiss and Farrington's study of youths in London (1991: 391) indicates that here too choices of co-offenders from different ethnic backgrounds are very rare.