We have been employing two mouse models for the study of human thyroid xenografts from patients with autoimmune thyroid disease (AITD). The first mouse strain is that of the athymic “nude” mouse which accepts human thyroid xenografts, but the passenger lymphocytes are lyzed in this model over several weeks; in the second model, the severe combined immunodeficient (SCID) mouse, both the xenograft and its lymphocytes survive. The AITD thyroid xenograft returns to normal function and morphology in the nude mouse over several weeks whereas the same AITD thyroid xenografts undergo aggravation of their lesions in the SCID mouse. Normalized (“cleansed”) thyroid tissue in the nude mouse, now bereft of the passenger lymphocytes, can be removed and then re-xenografted into SCID mice. There it will remain normal unless autologous peripheral blood mononuclear cells are added, whereupon the AITD lesion will be reproduced. Autologous “irrelevant” muscle tissue having undergone the same process will not show such lesions. There is thus no evidence for a primary thyroid cell disturbance in AITD, the abnormality appearing to be only in the immune system. Moreover peripheral blood mononuclear cells appear to contain sufficient memory cells to be able to mount an immune assault on the autologous normalized thyroid tissue (to which they had been previously sensitized) but not on irrelevant autologous (muscle) tissue.