Female SJL (H-2s) mice developed serum IgG anti-nucleolar antibodies (ANoA) after 5 weeks treatment with 0.05% or 0.01% silver nitrate (AgNO3) in drinking water. Five more weeks of treatment increased the ANoA titre to 3410 +/- 853 and 640 +/- 175 (reciprocal mean +/- s.e.m.), respectively. Controls receiving ordinary tap water and mice given 0.002% AgNO3 showed no antinucleolar antibodies. The high-titre ANoA targeted a 34-kD nucleolar protein identified as fibrillarin, the major autoantigen in murine mercury-induced autoimmunity and in a fraction of patients with systemic scleroderma. Serum autoantibodies to chromatin or histones, kidney, spleen, stomach, thyroid, or skin antigens (except the nucleolus) were not found in any of the mice. There was no consistent significant increase of serum IgG1, IgG2a, IgG2b, or IgG3 concentrations after AgNO3 treatment compared with controls. Mice treated with 0.05% AgNO3 for 10 weeks showed a slight decrease in serum IgG1, IgG2b and IgG3 concentrations. These mice also showed a small but statistically significant increase in renal, mesangial IgM deposits, which was not accompanied by any increase in C3c deposits, whereas mice given lower doses of silver nitrate showed no significant increase in mesangial immunoglobulin immune deposits. Systemic vessel wall immune deposits were not found in any of the mice. In mice given 0.05% silver nitrate, the kidney showed the highest concentration of silver (12.2 +/- 0.09 micrograms Ag/g wet weight; mean +/- s.e.m.), followed by the spleen (8.7 +/- 1.3), and the liver (3.9 +/- 0.4). Treatment with 0.01% silver nitrate caused a different distribution of silver, with the highest concentration in the spleen (2.1 +/- 0.16 micrograms Ag/g), followed by the kidney (0.63 +/- 0.037), and the liver (< 0.29 micrograms Ag/g; mean). Silver seems to be a more specific inducer of antinucleolar/anti-fibrillarin autoantibodies than mercury and gold, lacks the general immune stimulating potential of mercury, and has only a weak tendency to induce renal immune deposits. These observations suggest that the autoimmune sequelae induced in mice by metals is dependent, not only upon the genetic haplotype of the murine strain, but also on the metal under investigation.