Autoimmune inflammatory diseases, including many human arthritides, cause significant morbidity and mortality. The control of their immunological aspects is central to management of the diseases and usually involves drugs that are, in the immunological sense, non-specific in their effects. The efficacy of such drugs may be limited, and they may have side-effects so serious that if immunologically specific means were available to control these diseases, there would be significant benefits. Immunological specificity is carried only by antibodies, T cell receptors, MHC molecules and, of course, antigen. Based upon the experimental outcome of inducing systemic specific immunological unresponsiveness-the so-called oral tolerance effect-by feeding antigens, several groups have investigated the effects on experimental autoimmune diseases of delivering autoantigens across gastric and respiratory mucosal surfaces. Three forms of arthritis, those induced by type II collagen, adjuvant and oil, respectively, have been examined in this way, and the results from these studies show that disease can be specifically prevented or ameliorated. In parallel, examination of experimental encephalitis and uveitis, as examples of neurological diseases, and diabetes in the NOD mouse, have produced the same results. This review discusses the results of these experimental studies and draws out their common features. The uniform finding is that T cells are made hyporesponsive, and that the most likely mechanism is one of active suppression mediated through the selective activation of T cells, of either CD4+ or CD8+ phenotype, that make cytokines which in turn can suppress pathological T cells responsible for the disease lesions. There are many unanswered questions concerning optimal dosage regimes, routes and vehicles for antigen delivery and antigen pharmacokinetics that need to be answered if the promising results of early human trials are to be exploited with benefit. At the fundamental level, the full identity of the cell type, or types, responsible for the tolerance, most likely to be active peripheral suppression, is still elusive. Given the complexity of disease processes in the different situations that have been examined, it is likely that no one mechanism applies in all, and that therefore different therapeutic approaches will need to be tailored accordingly.