A method has been devised to produce eosinophilia in birds experimentally. Multiple intraperitoneal injections of bovine serum albumin (BSA) produced a greater response than horse serum in young adult fowls, and injections of 2 ml BSA were more effective than injections of 0.2 ml. The highest eosinophilic responses were obtained with injections of BSA together with aluminium hydroxide. Thymectomy had little effect on eosinophil numbers following horse serum injections but irradiation produced a depressed response. Horse serum that had been heated to 56 degrees C for 30 minutes to inactivate complement and to denature proteins, including possible horse IgE, induced an eosinophilic response in 50 per cent of birds compared with only 14 per cent with unheated horse serum. Furthermore, when the former birds were reinjected after a rest period of 13 weeks, the eosinophilic response followed a similar curve as before but a higher proportion of birds responded. Birds aged over two years did not give a good eosinophilic response after horse serum stimulation. There was no consistent correlation between raised IgE levels and eosinophil numbers. Tolerance to horse serum was detected in some birds using a radial immunodiffusion technique. Analyses of blood plasma from birds with eosinophilia showed a significant increase in total amino acids compared with normal birds.