The mean feature of organ-specific autoimmune disease is that the individual produces an immune response against one or several localized autoantigens. In humans this immune response may be characterized by the description of B-and T-cell responses to particular autoantigens in individuals with autoimmune diseases. However, the pure descriptive character of these studies does not allow us to answer questions about how these autoimmune reactions are induced; whether the autoimmune reactions are the" driving force" of the disease or rather epiphenomena resulting from nonspecific inflammatory activity or an immune response against some uncharacterized infectious agent. To be able to study these questions experimentally, several autoimmune animal models have been developed. In this review we will discuss results produced with the type II collagen (Cll)-induced autoimmune arthritis model (CIA). CIA was first described in 1977 by David Trentham who showed that immunization with type II collagen (CII) led to the development of arthritis in certain rat strains (Trentham et al. 1977). The CIA model was also later estabhshed in mice (Courtenay et al. 1980) and primates (Cathcart et al. 1989, Yoo et al. 1988) whereas guinea pigs are insensitive (Hernandez et al. 1988). As will be discussed in this review, the CIA mode! provides the opportunity for studying autoimmunity to a defined autoantigen present in the healthy individual and how this autoimmunity may be converted into pathology as a consequence of certain provocations. However, to understand the picture that emerges from these studies we will first refer to some recent findings and theories about the physiological importance of autoimmunity and the mechanisms for self-nonself discrimination in the immune system.