The objectives of the present study were to study the correlation between the change in serum Fe and Fe absorption when administering 100 mg Fe (as FeSO4) orally, and to study the correlation between the absorption from a 3 mg and a 100 mg Fe (as FeSO4) dose. The study was conducted in a group of eleven male blood donors, without any evident infection, who had given blood 8 weeks before the study. On three consecutive mornings the subjects were served a wheat roll fortified with Fe. On the first 2 d the roll was fortified with 3 mg Fe labelled with 59Fe; on day 3 the roll was fortified with 100 mg Fe labelled with 55Fe. The serum Fe response to the 100 mg dose was followed for 6 h. Fe absorption was measured by whole-body counting. High correlations were seen between the absorption of Fe and the change in serum Fe after 100 mg Fe (r2 0·94, P>0·001), between the absorption from 3 mg and 100 mg Fe (r2 0·88, P>0·001), and between the absorption from 3 mg Fe and change in serum Fe after 100 mg Fe (r2 0·90, P>0·001). This strengthens the evidence that it is possible to use the change in serum Fe as a measure of Fe absorption, e.g. when establishing the relative bioavailability for Fe powders. The results also imply that the induced serum Fe increase following 100 mg Fe added to a food could predict the Fe absorption of a small dose of Fe added to the same meal.