Over the past decade, numerous epidemiological studies have reported a significant association between lack of male circumcision and risk for HIV infection, leading to recommendations for male circumcision to be added to the armamentarium of effective HIV prevention strategies. We review the epidemiological data from studies that have investigated this association, including ecological, cross-sectional/case-control, and prospective studies. We discuss problematic issues in interpreting the epidemiological data, including the presence of other sexually transmitted infections, age of circumcision, and potential confounders such as religion, cultural practices, and genital hygiene. In addition, we review studies of biological mechanisms by which the presence of the foreskin may increase HIV susceptibility, data on risks associated with the circumcision procedure, and available data on the acceptability and feasibility of introducing male circumcision in societies where it is traditionally not practised. Although the evidence in support of male circumcision as an effective HIV prevention measure is compelling, residual confounding in observational studies cannot be excluded. Taken together with concerns over the potential disinhibiting effect of male circumcision on risk behaviour, and safety of the circumcision procedure, randomised trials of male circumcision to prevent HIV infection are recommended. An individual's choice to undergo male circumcision or a community's decision to promote the practice should be made in the light of the best available scientific evidence. More knowledge is required to assist individuals and communities in making those decisions. We conclude with recommendations for future research.