We followed 183 infants for two years, 31 of whom were breast fed less than three and a half months (median 70 days; short breast feeding group) and a further 31 of whom were exclusively breast fed for more than nine months (long breast feeding group). We assessed heredity for atopy, number of infections, and duration of breast feeding as determinants of atopy. During the first year of life 14 infants has signs of atopy. During the second year parents reported signs of atopy in a further 31. Heredity was the only significant predictor of atopy. Atopy was seen in 33% of infants with a positive heredity and in 16% without family history for atopy. The duration of breast feeding affected the incidence of atopy only among the infants without family history for atopy: fewer in the short breast feeding group (1/18) had atopy than in the long breast feeding group (5/13). Duration of breast feeding did not associate with incidence of respiratory infections. Diarrhoea was more common in the short breast feeding group than in the long breast feeding group during the first year of life. We conclude that prolonging exclusive breast feeding from the median of 70 days to nine months did not contribute to the prevention of infantile atopy and respiratory tract infections.