Many stratagems exist that bring about tolerance to endogenous body components to allow self-nonself discrimination and avoid potentially embarrassing autoimmune responses. At the turn of the century, the great Erlich recognized the necessity to establish the mechanisms that prevented self-reactivity and suggested that this could be mediated by the lymphoid system. However, these protective mechanisms against autoreactivity were not always successful; for example, Donath and Landsteiner’ described an antibody in paroxysmal cold haemoglobinuria that bound to red cells in the cold and then in the presence of complement would cause lysis as the temperature was raised. Interestingly, evidence for thyroid autoimmunity in human disease had been obtained as long ago as 19 1 1, when a young Romanian lady named Alexandrina Papazolu found that a third of the serum samples from patients with thyrotoxicosis gave positive complement fixation tests (CFT) with thyroid extracts. The audience, of course, was quite unmoved by these revelations because at that time they did not have the background knowledge to appreciate their significance. Also, in the aftermath of the discovery of the Wasserman reaction for syphilis, the studies were carried out as part of a popular vogue, looking at CFT with various organs rather than deliberately attempting to examine the hypothesis that autoimmunity was responsible for the disease. In fact, direct links between autoimmunity and disease were not really recognized until well after the Coombs’ test had been introduced in 1946 when the role of autoantibodies to red cells in the pathogenesis of acquired haemolytic anaemia was proposed. That deliberate induction of autoimmunity could cause tissue damage was well known from animal studies. Autolysed preparations of monkey brain, when injected back into the same species, were known to produce encephalomyelitis’. The introduction of the water in oil emulsion with killed TB by Dr Freund provided immunologists with a powerful way to provoke autoimmune lesions by incorporating autoantigens in an emulsion with that adjuvant. Freund himself showed quite unequivocally that sperm in Freund adjuvant could readily induce aspermat~genesis~. many other examples of autoantigen-induced tissue lesions were described, but particularly relevant in the present context were the experiments of Rose and Witebsky4 described elsewhere in this issue, who showed that extracts of an animal’s own thyroid incorporated in Freund adjuvant would induce both thyroid antibodies and a damaging invasive lymphocytic infiltration of the residual thyroid gland. In other words, the induction of autoimmunity to components of the thyroid gland resulted in an attack by the immunologically sensitized cells on the thyroid, with the autoantigen itself providing the target.