For a brief moment in the 1980s, a number of local authorities across England and Wales declared themselves ‘custody free zones’ (Rutherford, 2002), symbolising an intent that children should not be incarcerated for their offending behaviour within the local authority boundary. While this lofty aspiration was not always fulfilled, on occasion it was. More importantly, such declarations can be seen as manifestations of a widespread commitment, extending well beyond those ‘zones’, to the idea that the use of custody for children should be avoided wherever possible. During the course of the decade, that commitment found expression in an unprecedented fall in the number of young people deprived of their liberty by the criminal courts. The 1990s proved to be very different. A sea-change in the treatment of children in conflict with the law led to an escalation in custody every bit as sharp as the decline that had preceded it. This thesis seeks to explain that shift in order to understand the implications for practice if the imprisonment of children is to be reduced. The explanatory account is set against a longer term background of social, economic and political change that yielded what Garland (2002) has characterised as the ‘crisis of penal modernism’. It invokes what Foucault (1991) calls a ‘history of the present’ in which patterns of incarceration since the late 1960s are analysed to shed light on the systemic determinants of the current high level of youth custody. Local variations in the use of child imprisonment are also interrogated in the context of the historical experience to ascertain the nature of political, systemic and cultural factors that are consistent with lower rates of detention. The thesis concludes that the recent rise in custody cannot be understood in isolation: the same underlying dynamic that fuelled the carceral explosion impacted equally on other aspects of the youth justice system; and necessitated a significant cultural shift on the part of the those who might previously have been expected to resist the use of detention. Without an understanding of these corresponding changes, strategies for custody reduction that rely heavily on the provision of ‘robust’ community based alternatives, or those that seek to reduce the population of the secure estate simply by ‘nipping in the bud’ (Straw, 1997) children’s offending, are unlikely to lead to the desired outcome.