To determine whether the plasma ferritin assay reliably reflects body iron stores, as determined by the presence of stainable iron in the marrow, 63 iron-deficient patients were serially studied after receiving a total dose infusion of iron dextran (Imferon). In 58% (Group I) anemia did not recur. In 19% (Group II) classical iron deficiency was present with low plasma ferritin levels, reduced transferrin saturation and absence of stainable iron in the marrow. In 13% (Group III) evolving iron deficiency was seen with low plasma ferritin level but normal hemoglobin and transferrin saturation. The development of overt iron deficiency was associated with a fall in the transferrin saturation and absence of stainable iron in the marrow. In the remaining 10% (Group IV) the initially raised plasma ferritin fell in parallel with the developing anemia, reduction in transferrin saturation and absence of reticuloendothelial marrow iron. These studies demonstrated that in the development of iron deficiency anemia following total dose infusion of iron dextran (Imferon) there was a relationship (r = 0.9625; p less than 0.05) between decreasing body iron stores and falling plasma ferritin levels. Accordingly, the latter measurement should replace bone marrow studies carried out for this purpose.