The proximate cause of autoantibodies characteristic of systemic autoimmune diseases has been controversial. One hypothesis is that autoantibodies are the result of polyclonal nonspecific B cell activation. Alternatively, autoantibodies could be the result of antigen-driven B cell activation, as observed in secondary immune responses. We have approached this question by studying monoclonal anti-DNA autoantibodies derived from unmanipulated spleen cells of the autoimmune MRL/lpr mouse strain. This analysis shows that anti-DNAs, like rheumatoid factors (19), are the result of specific antigen-driven stimulation. In addition, correlation of sequences with fine specificity shows that: (a) somatic mutations can cause specificity for dsDNA and that such mutations are selected for; (b) arginine residues play an important role in determining specificity; and (c) anti-idiotypes that recognize the majority of anti-DNA are probably not specific for any one family of V regions.