There is much evidence that the development of allergic disorders may be related to early exposure of allergens, including those in breastmilk. We have tried to find out whether avoidance of food and inhaled allergens in infancy protects against the development of allergic disorders in high-risk infants. In a prenatally randomised, controlled study 120 infants with family history of atopy and high (greater than 0.5 kU/l) cord-blood concentrations of total IgE were allocated randomly to prophylactic and control groups. In the prophylactic group (n = 58), lactating mothers avoided allergenic foods (milk, egg, fish, and nuts) and avoided feeding their infants these foods and soya, wheat, and orange up to the age of 12 months; the infants' bedrooms and living rooms were treated with an acaricidal powder and foam every 3 months, and concentrations of Dermatophagoides pteronyssinus antigen(Der p l) in dust samples were measured by enzyme-linked immunosorbent assay. In the control group (n = 62), the diet of mothers and infants was unrestricted; no acaricidal treatment was done and Der p l concentrations were measured at birth and at 9 months. A paediatric allergy specialist unaware of group assignment examined the infants for allergic disorders at 10-12 months. Odds ratios were calculated by logistic regression analysis for various factors with control for other confounding variables. At 12 months, allergic disorders had developed in 25 (40%) control infants and in 8 (13%) of the prophylactic group (odds ratio 6.34, 95% confidence intervals 2.0-20.1). The prevalences at 12 months of asthma (4.13, 1.1-15.5) and eczema (3.6, 1.0-12.5) were also significantly greater in the control group. Parental smoking was a significant risk factor for total allergy at 12 months whether only one parent smoked (3.97, 1.2-13.6) or both parents smoked (4.72, 1.2-18.2).