We have examined the fine specificity of a panel of cloned T cell hybridomas generated from Lewis rats immunized with guinea pig (GP) or Lewis rat myelin basic protein (MBP) to determine the autoimmune T cell repertoire that develops in experimental allergic encephalomyelitis (EAE). This analysis has demonstrated that GP MBP, which was approximately 10-fold more potent for EAE induction than the autologous rat MBP, produced a population in which almost one-fourth of the members responded to GP-unique determinants and displayed no crossreactivity on the self antigen. The remaining majority of GP MBP-induced clones were specific for the 68-88 encephalitogenic determinant and could be subdivided into three groups based on their varying responses to the 68-88 peptide and rat and rabbit MBPs. Surprisingly, one of these groups showed equal reactivity to GP and rat MBPs. In contrast, the clonotype composition of the T cell population induced by rat MBP was quite different. One-half of these clones comprised a single group responding to the 68-88 determinant, reacting equally with GP and rat MBP. All of these responded to the same range of antigen concentrations as their GP-induced counterparts. The remaining half of that population contained a collection of clones that was nearly as encephalitogenic as the 68-88 population after propagation as a short-term T cell line. These clones were specific for at least three distinct antigenic determinants, all displaying extensive cross-species reactivity, and required as little or less rat MBP for maximal stimulation as did the 68-88-reactive clones. We therefore conclude that the T cell repertoire for MBP does include clones with reactivity to both 68-88 and non-68-88 determinants of GP and rat MBPs, and that both MBPs appear to be equally capable of stimulating these clones in vitro. However, the differences in the clonotype composition of the populations induced by immunization with these two antigens suggest that rat and GP MBPs are subject to different immunoregulatory constraints in the animal and may account for the difference in the encephalitogenic potential of these two antigens.