Plasma retinol and indices of iron status were measured in 148 school children (6-12 years) receiving a soup fortified with iron and vitamin C for a period of 15 weeks. The most significant change in serum iron (P = 0.0005) and transferrin saturation (P = 0.0002) was seen in subjects with plasma retinol > or = 40 micrograms/dl, while subjects with plasma retinol < 20 micrograms/dl showed no response. Serum ferritin improved most in the retinol categories < 40 micrograms/dl, suggesting that the absorption of iron was not impaired by marginal vitamin A status, but that it was rather the mobilisation of iron from stores that was affected. Changes in vitamin A status correlated positively and significantly with changes in serum iron (r = 0.37; P = 0.0001) transferrin saturation (r = 0.27; P = 0.004) and haemoglobin (r = 0.21; P = 0.03), but negatively with serum ferritin (r = -0.28; P = 0.003). The presence of marginal vitamin A deficiency in a community may limit the effectiveness of an iron intervention programme and vitamin A status should therefore also be considered when such programmes are planned.