Seven hundred seventeen healthy male blood donors regularly donating four or more units a year were surveyed for haemoglobin and serum ferritin levels. One hundred fifty-one (21%) had a haemoglobin less than 13.5 g/dl and were therefore disqualified from further blood donation, having a mean serum ferritin of 28 micrograms/liter. Of the remaining 566 donors with haemoglobin levels equal to or greater than 13.5 g/dl, the mean serum ferritin was 33 micrograms/liter, although in 299 (53%) the value was less than 28 micrograms/liter. To document response to iron therapy 46 donors with haemoglobin levels equal to or greater than 13.5 g/dl were stratified into those with the lowest iron stores (group 1; n = 23), defined as a serum ferritin less than 20 micrograms/liter, and controls (group 2; n = 23), with serum ferritin between 50 and 150 micrograms/liter. Within each stratum donors randomly received ferric polymaltose at a dose of 100 mg elemental iron twice daily for 56 days (groups 1a and 2a) or an identical iron-free placebo tablet administered on the same schedule (groups 1b and 2b). Iron therapy in the iron-deficient group (group 1a:n = 11) resulted in a significant rise in haemoglobin (p = .03) and iron stores reflected in serum ferritin (p = .002) compared to those receiving placebo (group 1b). In the control group iron therapy or placebo was without significant effect. Thus, ferric polymaltose preparation is bioavailable and is notable for the virtual absence of gastrointestinal tract side effects.