Absorption of iron from seven nonprescription dietary iron supplements was measured in both iron-replete and iron-depleted rats by use of the extrinsic label technique. Both ferrous and ferric forms of the radioactive extrinsic tracer iron were used; the iron was in the form of 59FeS04, 59FeC13, or f e r r~u s -~~F e gluconate. Of six products tested with both ferrous and ferric labels, significant differences in absorption measured by the two different labels were found for three products. This suggests that there is incomplete exchange between the intrinsic and extrinsic iron pools. Iron-deficient animals absorbed significantly more iron than iron replete animals when the ferrous label was used in both experiments. Although iron-replete animals discriminate among iron sources, absorption is always less than in iron-deficient animals. The results with iron deficient animals show the true bioavailability of the iron. The use of the extrinsic label may not always be valid when the metal in question can exist in more than one oxidation state.