A total of 888 randomly selected dairy farmers participated in an epidemiological study to evaluate the prevalence of precipitins to farmer's lung antigens, and the socioeconomic factors associated with their presence. Precipitins were present in 75 farmers (8.4%) (65 to Micropolyspora faeni, seven to Aspergillus spp, two to both Aspergillus and Micropolyspora faeni, and one to Aspergillus and Thermoactinomyces vulgaris). The titres ranged from a dilution of 1/32 to a concentration of X 2 (Ouchterlony's double diffusion method). In the study population there were 544 who had never smoked, 146 ex-smokers, and 198 smokers. Sixty nine precipitin positive subjects were either never smokers or ex-smokers; only six were smokers. The negative relationship between cigarette smoking and precipitins was highly significant (p = 0.004). Factors positively associated with positive precipitin reactions were: size of farm, time spent in the barn, and the presence of a family member previously diagnosed as having farmer's lung disease. Interestingly, positive precipitin reactions were not associated with any of the following: use of silos, hay conditioners, or hay dryers; the presence or quantity of mouldy hay; or the presence of respiratory symptoms. It is concluded that precipitin analysis is not useful as a screening method for farmer's lung, though it can be of diagnostic value in acute farmer's lung disease.