A/J mice are proposed as a model of the human lupus diathesis since we previously determined that they express a slow acetyltor phenotype while others showed them to have a predisposition to develop spontaneous and drug-induced antinuclear antibodies. A/J mice were mated with C57BL/6J mice, a rapid acetylator phenotype which is relatively resistant to spontaneous and drug-induced antinuclear antibodies, to assess the importance of slow acetylator status as a component of the lupus diathesis. Procainamide, a potent inducer of antinuclear antibodies, was acetylated to a lesser degree by A/J mice than by C57BL/6J mice. This difference, detectable by in vitro assay but not by urinary levels of acetylated drug, represents a genetic polymorphism which can be detected in F2 and backcross progency of these two strains. We confirmed that A/J mice have a higher incidence of spontaneous antinuclear antibodies than C57BL/6J mice and that in A/J mice these antibodies can be induced by oral procainamide (6 g/l of drinking water for 37 weeks); procainamide tended to suppress antinuclear antibody formation in C57BL/6J mice, however. Rapid and slow acetylators among F2 and backcross populations were identified by a test for N-acetyltransferase activity in blood hemolysates. These two groups together with their respective rapid and slow acetylator parents were compared in respect to incidence of antinuclear antibodies. Slow acetylator phenotypes among F2 and backcross mice were predisposed to high titers of antinuclear antibodies like the slow acetyltor A/J strain. However, long-term exposure to procainamide suppressed antinuclear antibody formation as was found in the rapid acetylator C57BL/6J strain. Thus, the ability to N-acetylate procainamide is not the sole factor controlling the ability of this drug to induce antinuclear antibodies.