It would seem that mentoring--matching youths with a caring and committed adult--would fit hand in glove with the needs of young people who are transitioning out of the foster care system. A stable, consistent, and caring adult presence is precisely what many such youths lack as they reach the age of legal adult maturity (18 in most states, 21 in others) and may no longer have access to foster care services. It is not surprising that mentoring programs targeting foster care youths have been cropping up across the United States and abroad (Clayden & Stein, 2005; Mech, Pryde, & Rycraft, 1995), as mentoring programs have enjoyed unprecedented growth in recent years (DuBois & Karcher, 2005). However, the existing empirical literature on the conditions associated with effective formal youth mentoring relationships and the potential for harm in their absence should give us pause, as meeting these conditions may be especially challenging when working with transitioning youths. The needs of transitioning youths and the efficacy of mentoring programs are of central concern to social work. Child welfare has been a major field of practice since the beginning of the profession. In addition, relationship-based approaches to intervention are a core technology of the profession, both through clinical intervention and community-based programming. In this article, consistent with the social work profession's attention to the empirical evidence base for interventions, we identify and critique the research literature on the effectiveness of mentoring programs for youths more generally and the implications of this evidence for programs serving youths leaving foster care and for policies guiding and governing these programs. We use the ecological approach (for example, Germain & Gitterman, 1996) in our analysis, partially out of concern that mentoring has tended to focus intently on the interpersonal relationship to the neglect of both mezzo and macro issues (see Keller, 2005, for an exception).