Women (n = 45; 18-48 years) with serum ferritin less than or equal to 20 micrograms/l and haemoglobin greater than 120 g/l participated in an iron supplement bioavailability study. They were randomized to one of three groups and given one of three different low-dose supplements (18-20 mg iron/d) for 6 months. One of the supplements contained haem iron and non-haem iron, the others contained non-haem iron only. Serum ferritin, haemoglobin, serum iron and total iron binding capacity (TIBC) were determined at start and after 1, 3 and 6 months. Dietary intakes were recorded during 4 and 3 consecutive days after 1 and 3 months, respectively, by the aid of household measures and a set of food models. The increase in mean serum ferritin was significant (P less than 0.01) for two of the supplements, the one containing haem iron giving the best result. All the supplements resulted in a significant (P less than 0.011) mean decrease in TIBC. No significant differences between the groups were found in daily intakes of food factors known to influence iron absorption. The improvement in iron status was therefore attributed to the supplements. None of the supplements caused iron intolerance. Thus low-dose iron supplements given to iron depleted non-anaemic women may be an alternative to high-dose therapy.