Breast feeding has been recognized as a mode of HIV transmission since 1985. It is estimated that infants run 1-in-3 risk of being infected with HIV when breast fed by mothers who were initially infected with HIV after delivery. Breast feeding however is unlikely to confer such as a high risk to the child when the mother is infected before delivering the child. Acknowledging the risk of infection the World Health Organization (WHO) in 1987 nonetheless recommended that mothers continue to breast feed. An analysis subsequent to that recommendation however suggested that breast feeding may increase the rate of perinatal HIV transmission by 14% and may be the cause of higher perinatal HIV transmission rates seen in some developing countries. Even so WHO/UNICEF still recommend that in a setting of high infant mortality associated with infectious disease or malnutrition all women should breast feed their infants. Breast feeding is not recommended in more affluent settings. This recommendation is made and maintained as the less of 2 evils; bottle feeding infants in some settings may threaten child survival more than breast feeding them with potentially infected milk. The paper describes the attempt of HU et al to compare the hazards of bottle feeding and the risks of perinatal acquisition of HIV through a decision analysis model. It remains to be seen which approach should be adopted to truly maximize the long-term potential survival of these infants.