1032 workers exposed to organic dusts in 9 different industries were surveyed using an administered questionnaire and exposure sampling. Information concerning; demographic features, occupational histories, respiratory symptoms (including temporal patterns) and smoking habits were collected. In addition total dust exposure was measured in a representative sample in each industry using personal sampling techniques. Exposure to endotoxin was then measured using a simple water extraction of collected dust and assay with an automated quantitative turbidimetric assay (LAL 5000). The highest prevalence of respiratory symptoms were found in animal handlers (swine confinement and poultry) and the lowest in wool workers. Logistic regression analysis was used to determine predictive factors of respiratory symptoms entering parameters of demographic details, smoking habits and measured exposures in a forward stepwise model (SPSS). Smoking habits, increasing dust/ endotoxin exposures and pre-existing pulmonary disease were the significant predictor variables.