This study was undertaken to investigate body iron stores in so-called remunerated blood donors as well as to evaluate the sensitivity of hemoglobin determination in detecting iron deficiency in two populations of blood donors. The authors studied 522 male donors who were divided into three groups: Group I, first-time volunteer donors with hemoglobin levels greater than or equal to 13 g per dl; Group II, remunerated donors with hemoglobin levels greater than or equal to 13 g per dl; and Group III, remunerated donors rejected because their hemoglobin levels were less than 13 g per dl. Iron stores were evaluated with an enzyme-linked immunosorbent assay for plasma ferritin. In Group I, 4.5 percent were iron-deficient with a mean ferritin value of 55.3 ng per ml; in Group II, 59.7 percent were iron deficient with a mean ferritin level of 17.4 ng per ml, and in Group III, 82.5 percent were iron-deficient and the mean ferritin level was 8.4 ng per ml. The last values represent the highest percentage of iron deficiency and the lowest mean ferritin value thus far reported. In Group I, hemoglobin determination had a sensitivity of 95 percent in detecting iron deficiency, but in Group II had only 40 percent sensitivity. These results indicate that a more accurate and reliable test, such as a plasma or serum ferritin determination, may be necessary to detect iron deficiency in blood donors when they donate more than five times per year, particularly those who are remunerated.