Our knowledge of erythropoiesis and iron in renal disease is limited. The accepted view of the control of erythropoiesis was founded on observations made in a variety of disorders, but the control mechanism in healthy individuals may not be quite the same. Evidence suggests that mechanisms other than erythropoietic stimulation may play a role in increased red blood cell production. Measuring erythropoiesis is complex. The quantitative reticulocyte count is probably the closest practical assessment of erythropoietic activity we can achieve, yet there is very little correlation between circulating erythropoietin level and reticulocyte count in normal and near normal subjects. Oxygen transport in humans depends entirely upon iron. In renal disease, the failure of the erythropoietin positive feedback mechanism can be readily and directly remedied; recombinant human erythropoietin therapy can replace the missing erythropoietin, but this will be negated if iron supply to the erythroid marrow falls short of demand. Measurement of iron stores is also complex. The use of serum ferritin concentration as a direct quantitative estimate of iron in the stores is not advisable, and in practice we have not found the transferrin receptor assay to be useful in identifying patients who require iron therapy. Use of percentage hypochromia as a measure of iron deficiency is complicated by the fact that hypochromic cells are not exclusively a consequence of functional iron deficiency. There are clearly lessons still to be learned in this field and there is much that we do not yet understand about the control of erythropoiesis and iron metabolism in humans.