To investigate the pathophysiologic mechanisms of nicekl-induced erythrocytosis, serum erythropoietin activities were measured in (a) pooled serum from rats at 2 wk after intrarenal injection of alpha Ni3S2 (5 mg/rat), and (b) pooled serum from control rats at 2 wk after intrarenal injection of sterile NaCl vehicle (0.4 ml/rat). A sensitive erythropoietin bioassay was employed, which entailed repetitive administration of test serums to post-hypoxic polycythemic mice in divided doses (12 s.c. injections of 0.5 ml of serum at 6 hr intervals for 3 da; total dose = 6 ml of serum/mouse). The erythropoietin detection limit was congruent to 20 I.U./liter of serum. In mice which received pooled serum from alpha Ni3S2-treated rats, erythrocyte 59Fe-uptake averaged 28% (S.D. +/- 5) (vs 3.7 +/- 1.1% in control rats; P less than 0.001). Based upon a 7-point calibration plot, the erythropoietin activity in pooled serum from alpha Ni3S2-treated rats averaged 130 I.U./liter (S.D. +/- 18) (vs 27 +/- 6 I.U./liter in control rats; P less than 0.001). In vitro addition of Ni(II) to rat serum (100 microgram/liter) had no effect upon serum erythropoietin activity. This study demonstrates that induction of erythrocytosis in rats by intrarenal injection of alpha Ni3S2 is mediated by increased serum erythropoietin activity.