Iron supplementation of infant formulas is recommended by most national and international organizations, but the optimal form of supplementation has not been determined. We have compared the bioavailability and tissue distribution of iron from four iron chelates and two commonly used iron salts. Weanling C57BL/6J mice were fed for 1 week an evaporated cow's milk diet supplemented with vitamins and minerals (except for iron). Following the adjustment period, mice were divided into 12 groups of 20 each. Six groups continued to receive the cow's milk diet for 18 hours, while the other six groups were fed a similar diet based on human milk. Individual groups received a single dose of milk radioactively labeled with Fe(II)Cl2, Fe(II)SO4, Fe(III)NTA, Fe(III)EDTA, Fe(III)citrate or Fe(III)lactobionate. Wholebody retention was measured after 4 days; animals were then killed and individual tissues were counted for radioactivity. Iron from FeCl2, FeSO4 and FeNTA were the best retained from both milk diets. Fe citrate had a significantly lower iron retention than all other groups in either diet and is probably not an effective chelate for delivering iron to milk diets. Iron bioavailability was higher from the human milk diets than from the cow's milk diets from all vehicles used except citrate and lactobionate. Absorption of Fe citrate was similar from the two milk diets, while percent retention from Fe lactobionate was higher from cow's milk than from human milk. Tissue distribution of retained iron was similar for the milk diets and among the groups, indicating that, once absorbed, iron from the different vehicles is metabolized in a similar manner.