Iron bioavailability from species-specific milk is assumed to be high for the offspring, possibly due to species-specific iron-binding proteins in the milk. To assess this bioavailability using radioisotopes, the validity of extrinsic labeling technique needs to be proven. Using the suckling piglet as an animal model, we have compared iron bioavailability from sow's milk labeled extrinsically and intrinsically. During intrinsic labeling transfer into milk of 59Fe given intramuscularly was slow and was found to be at maximum 14 h post-injection. Recovery of isotope in the milk was only 0.00014%. Extrinsic and intrinsic labels were distributed differently among milk fractions; intrinsic iron bound primarily to the fat fraction but the extrinsic iron bound primarily to the casein fraction. Iron retention from intrinsically labeled milk was considerably higher than from extrinsically labeled milk. These results show that the extrinsic tag method is not valid for studies on iron absorption from sow's milk and suggest that the situation may be the same for human milk.