This article argues that the traditional focus of work on prisoners' families is misleadingly restrictive, considering that a high proportion of prisoners have neither an active spousal relationship nor children for whose support they are responsible. It seeks to connect thinking about prisoners' families with recent work on family ties and obligations, and to explore the criminological implications of what is known about the sources of social support available to people leaving prison. Drawing on a small-scale research project by one of the authors, it suggests that the family ties of many prisoners are diverse and complex, and that this has implications for penal policy and practice and more generally for the development of control theory in criminology.