Studies of Fe absorption in pregnancy often make unfounded assumptions of erythrocyte incorporation. Therefore, we measured the absorption and utilisation of Fe during early and late pregnancy by the erythrocyte incorporation of two stable isotopes. 8·5 mg 57Fe (oral) and 0·5 mg 58Fe (intravenous) were given to five non-pregnant women, to five women in early gestation (12 weeks) and five women in late gestation (36 weeks). The stable isotope ratios in whole blood 14 d later were measured by MS. Together with estimation of body Fe mass, this enabled the calculation of Fe absorption and erythrocyte incorporation. In non-pregnant women, Fe absorption averaged 20·3 (range 10·2–34·3) %. It was not significantly different in early pregnancy (11·8 (range, 4·4–24·8) %), but during late pregnancy Fe absorption increased to 59·0 (range 38·2–77·2) %. All non-pregnant and early-pregnancy subjects had normal Fe status, but two women in late pregnancy had evidence of Fe insufficiency. During early and late pregnancy, mean erythrocyte incorporation was 63·4 (SD 12·1) % AND 71·0 (sd 10·4) % respectively, significantly reduced (P=0·003) compared with non-pregnant subjects (90·1 (sd 6·0) %). Decreased erythrocyte incorporation of absorbed Fe in early pregnancy is compatible with reduced Fe demand and low oral absorption. However, during late pregnancy decreased erythrocyte incorporation associated with high absorption and Fe insufficiency is different from the high erythrocyte incorporation which occurs in non-pregnant Fe-deficient women. This suggests that part of the aetiology of Fe deficiency during pregnancy may be the reduction of Fe utilisation.