Serum erythropoietin (EPO) and its relationship to hemoglobin (Hb), iron status markers and iron supplementation during normal pregnancy was assessed in a longitudinal, placebo-controlled study on 118 women, 61 took daily tablets containing 66 mg ferrous iron from the second trimester until delivery and 57 took placebo. Blood samples were obtained at 4-week intervals until delivery as well as post-partum. In the placebo-treated women, median serum EPO rose from 22.5 U/l at inclusion to 35.0 U/l at delivery (P = 0.0001). In the iron-treated women, median serum EPO rose from 23.9 to 29.9 U/l (P = 0.0001). Serum EPO showed a steeper increase in the placebo-treated women than in the iron-treated women (P < 0.05). After delivery, serum EPO became normal in both groups (P = 0.0001). Median Hb was lower in placebo-treated (iron depleted) than in iron-treated (iron repleted) women (P < 0.05). In the placebo-treated women there was a negative correlation and in the iron-treated women a positive correlation between serum EPO and Hb. In the placebo-treated women, inverse correlations existed between serum EPO and serum transferrin saturation and serum ferritin, reflecting the consequences of iron deficiency, whereas the iron-treated women displayed no correlation. A physiological, nonhypoxia-induced increase in EPO production accounts for the basic expansion of the red cell mass during pregnancy. In placebo-treated women, iron deficient erythropoiesis constitutes an additional hypoxic stimulus, which induces a further increase in serum EPO.