1. Stable isotope methods are being used to investigate the absorption of dietary iron. In order to be certain that this new methodology is accurate, we have compared results obtained using stable isotopes and inductively coupled plasma mass spectrometry with those determined using a radioisotope and whole body counting. 2. The stable isotope 54Fe (2.8 mg) was given to 10 healthy non-pregnant women. Six women received the isotope in aqueous form, and four took it with a meat meal. The 54Fe served as a carrier for 10 ng of the radioisotope 59Fe. An ampoule (200 μg) of the isotope 57Fe or 58Fe was then given intravenously, and in serum samples taken over the next 10 h the ratios of the stable iron isotopes were measured by inductively coupled plasma mass spectrometry and the oral iron absorption was calculated. This was then compared with the results obtained by using a whole body counter to measure (on day 0 and day 14) the γ-activity emitted by the radioisotope. 3. The mean iron absorption measured by both methods ranged from 8% to 45%. Measurement of the post-absorptive serum enrichment of the stable isotopes provided estimates of absorption from both aqueous and food iron which were similar to that yielded by whole body counting, the mean difference being −1.5% (95% confidence interval −5.2 to 2.1%). Absorption estimated by stable isotopes exhibited the same inverse relationship with the serum ferritin level (body iron stores) to that known to exist with whole body counting. Similar estimates of food iron absorption were obtained irrespective of the type of isotope used as an extrinsic label, implying that stable isotopes are as valid as radioisotopes in reflecting intrinsic food iron absorption. 4. This study validates the use of stable isotopes and post-absorption curves as a new and accurate technique in the measurement of iron absorption.