Résumé : The advantages of breast-feeding practice, both in industrialized and developing countries, are now well recognized. Recently, this optimistic view has been challenged by the potential risk of transmission of the Human Immunodeficiency Virus type I (HIV 1) through breast milk. The aim of this paper is to review virological aspects of milk transmission of animal and human retroviruses and to summarize the virological and epidemiologic data available on the potential role of breast-feeding in mother-to-child transmission of HIV 1. Colostrum and breast milk are considered as a major mode of transmission of many animal retroviruses. Evidence is strong to implicate transmission through breast milk of Human T-cell Leukemia Virus I (HTLV 1). Furthermore, bottle-feeding is able to prevent mother-to-child transmission of HTLV-I. Mother-to-child transmission of HIV seems to be mainly the consequence of transplacental infection and/or contamination of the newborn in the birth canal during delivery. Transmission of HIV 1 by breast milk is considered as biologically plausible. To date, eight cases strongly suggesting transmission of HIV 1 by breast-feeding have been reported in the litterature. In addition, two recent cohort studies carried out in Africa showed that post-partum HIV I seroconversion in mothers was associated with a high risk of postnatal transmission to their offspring, most probably via breast milk. The risk of postnatal transmission of HIV 1 from mothers who are already infected during gestation or at delivery remains unknown. It must be remembered however that transmission rates estimated from studies performed in Africa where breast-feeding is commonly practiced are generally higher than those estimated in european or american studies. Based on all the available information, the World Health Organization concluded in 1987 that the risk of transmission of HIV 1 by breast milk was small and that the promotion of breast-feeding should be continued in non-industrialized countries, even by HIV-infected mothers. However, many crucial questions regarding this mode of transmission remain unresolved and must lead to the following proposals: 1) careful, large scale, prospective and controlled cohort studies carried out under optimal ethical conditions are of utmost priority to quantify the risk associated with this mode of transmission and to appreciate it public health impact; 2) the present recommendations on breast-feeding in the context of HIV 1 infection should be modulated and perhaps reformulated.