Commercial oral iron preparations (drops) containing trivalent iron either as a citrate (C) or hydroxide-polymaltose complex (HP) were labelled with 59Fe in the thermal neutron flux of a research reactor. No measurable differences were observed between the original commercial preparation and the neutron-activated samples. In an intraindividual comparison oral doses of 100 mg 59Fe were administered as an aqueous 59Fe(II)-ascorbate solution (= reference), C (-59Fe) and HP (-59Fe) to starved subjects with normal and depleted iron stores. Two weeks later the whole body retention of absorbed 59Fe was measured and used for the calculation of 59Fe-absorption. Subjects with normal Fe-stores absorbed means a +/- SEM +/- SD = 8.53 +/- 0.29 +/- 1.2% of the 59Fe from the aqueous 59Fe(II)-ascorbate solution, 1.58 +/- 0.12 +/- 0.49 of the C-59Fe and 0.81 +/- 0.06 +/- 0.27% of the HP-59Fe. Subjects with depleted Fe-stores absorbed means a +/- SD = 17 +/- 3% of the 59Fe from the aqueous 59Fe(II)-ascorbate solution and only 2.4 +/- 1% from the HP-59Fe. The relative bioavailabilities were reduced from 100% (reference = Fe(II)-ascorbate) to 19% for the C-Fe and to 9.5% for the HP-Fe in subjects with normal Fe-stores. In subjects with depleted iron stores the relative bioavailability was decreased from 100 to 14% for the HP-Fe. Oral iron preparations with a relative bioavailability less than 10% or less than 30% are considered to be therapeutically ineffective or insufficient, respectively.