MRL/Mp-lpr/lpr autoimmune mice consistently show an approximately 25% incidence of the systemic lupus erythematosus marker autoantibody anti-Sm. In the present report, we show that the failure to find anti-Sm antibodies in three-quarters of 5-mo-old MRL/lpr mice was not an artifact of an insensitive assay, but rather that the mice fell into two populations as regards their anti-Sm positivity. Based on an extensive analysis of the incidence of anti-Sm positivity in 5-mo-old mice according to their cage of residence, we found no evidence for genetic, environmental, or parental influences on the propensity of an individual animal to become anti-Sm positive. Also, the gender of the mouse, its Sm antigen level, or its length of survival were not related to anti-Sm antibody, nor was the anti-Sm antibody status of either parent. Some animals became anti-Sm positive after 5 mo of age, but this was less likely than becoming positive before 5 mo of age. Finally, a survey of 205 autoimmune C57BL/6-lpr/lpr mice confirmed the uniqueness of the MRL background for this autoantibody response. These results together indicate that the possibility of making anti-Sm antibodies is under genetic control, but that the expression of this capability in an individual animal is governed by stochastic events. We hypothesize further that such random processes may involve the expression of particular immunoglobulin variable-region genes combined with mechanisms of extensive somatic mutation or positive feedback amplification, which would transmute an initial monoclonal response into an eventual polyclonal one.