Rats and guinea pigs developed pronounced erythrocytosis at one to four months after unilateral intrarenal (ir) injection of nickel subsulfide (Ni3S2). For example, at two months after ir administration of Ni3S2 (5 mg) to rats, blood hematocrit values averaged 70 +/- 3 percent (p less than 0.001 vs. 48 4/- 2 in controls); at two months after ir administration of Ni3S2 (20 mg) to guniea pigs, blood hematocrit values averaged 67 +/- 6 percent (p less than 0.001 vs. 49 +/- 1 percent in controls). Hamsters and gerbils did not develop erythrocytosis after ir injection of Ni3S2 (5 mg/animal). Administration of Ni3S2 to rats by intrasplenic injection did not increase blood hematocrit; splenectomy did not prevent erythrocytosis in rats that received ir injection of Ni3S2. Erythrocytosis in rats was completely blocked by excision of the Ni3S2-injected kidney but was unaffected by excision of the non-injected kidney. Partial inhibition of Ni3S2-induced erythrocytosis in rats occurred after simultaneous ir injection of Mn, Cu, or Al dusts, benzo(a)pyrene, or subcutaneous (sc) infusion of sodium diethyldithiocarbamate. Erythrocytosis induced by ir injection of Ni3S2 was augmented by ir injection of Cr dust or intramuscular (im) administration of iron-dextran. Erythrocytosis occurred in rats after ir implantation of Ni3S2 within semi-permeable cellulose tubules, indicating that phagocytosis of Ni3S2 particles is unnecessary for erythropoietic stimulation. Erythropoietin (Ep) activity in rat serum increased sixfold at two weeks after ir injection of Ni3S2 (p less than 0.001 vs. controls), but Ep activity in pooled extracts of Ni3S2-treated rat kidneys did not increase significantly. This study identifies several factors that influence erythropoietic stimulation by Ni3S2, and furnishes salient information concerning the pathogenesis of Ni3S2-induced erythrocytosis.