The consensus among influential AIDS experts that heterosexual transmission accounts for 90% of HIV infections in African adults emerged no later than 1988. We examine evidence available through 1988, including risk measures associating HIV with sexual behaviour, health care, and socioeconomic variables, HIV in children, and risks for HIV in prostitutes and STD patients. Evidence permits the interpretation that health care exposures caused more HIV than sexual transmission. In general population studies, crude risk measures associate more than half of HIV infections in adults with health care exposures. Early studies did not resolve questions about direction of causation (between injections and HIV) and confound (between injections and STD). Preconceptions about African sexuality and a desire to maintain public trust in health care may have encouraged discounting of evidence. We urge renewed, evidence-based, investigations into the proportion of African HIV from non-sexual exposures.