Elemental iron powders are widely used to fortify staple foods. Experimental evidence indicates that there is considerable variation in the bioavailability of different products. For some powders, it may be too low to permit a significant impact on iron status. This study was designed to evaluate possible approaches to screening commercial iron powders for predicted bioavailability, to identify products that have the potential to improve iron status, and to ascertain whether bioavailability is related to the method of manufacture. Nine commercial iron powders were allocated to one of five types based on the production process; carbonyl, electrolytic, hydrogen-reduced (H-reduced), carbon monoxide-reduced (CO-reduced), and other reduced. Structure by scanning electron microscopy and physical properties (pycnometric and apparent density, particle size distribution, Fisher subsieve size, and surface area) were determined on all samples. Selected samples (one or more of each type depending on the cost of the assay) were then subjected to five screening procedures that have previously been advocated for predicting bioavailability in humans--issolution rate in 0.1 mol/L HCl, dialyzability and Caco-2 cell iron uptake, both after simulated in vitro gastrointestinal digestion, relative bioavailability (RBV) with respect to ferrous sulfate by the AOAC rat hemoglobin repletion method, and plasma iron tolerance tests in human volunteers. The results for particle size distribution, surface area, Fisher subsieve size, dissolution rate in 0.1 mol/L HCl, and RBV in rats were significantly correlated and consistent for powders of the same type. However, values for different powder types were significantly different. There was no correlation between either dialyzability or Caco-2 cell uptake and the predicted bioavailability estimates based on the physical properties, dissolution rates, RBV in rats, or human efficacy data. Although human plasma iron tolerance tests were in general agreement with the other measures of predicted bioavailability, they did not provide information that would have improved the precision of bioavailability estimates based on physical properties, dissolution in HCl and/or RBV in rats. Our observations indicate that the dissolution rate in 0.1 mol/L HCI under standardized conditions is highly predictive of potential bioavailability and that it would be the most practical approach to developing a reliable and sensitive screening procedure for predicting and monitoring the bioavailability of commercial elemental iron powder products. Some, but not all, of the carbonyl and electrolytic iron powders had the highest predicted bioavailability values. The predicted bioavailability for the reduced iron products was lower and variable, with the lowest values being recorded for the carbon monoxide and other reduced iron products. Two powder types were selected for a human efficacy trial, electrolytic (because it is the iron powder type recommended by WHO) and hydrogen-reduced (because of its widespread use). Electrolytic/A131 and H-reduced/AC-325 had relative efficacies compared with ferrous sulfate monohydrate of 77% and 49%, respectively, based on the change in body iron stores in Thai women with low iron stores, who received an additional 12 mg iron per day, six days per week for 35 weeks in wheat-based snacks. We conclude that there is significant variability in the bioavailability of the commercial iron powders that we evaluated (those used for food fortification at the time that our studies were initiated), and that bioavailability is related in part to production method. The bioavailability of some carbonyl and electrolytic iron powders may be adequate for effective food fortification. The reduced iron powders that we tested are unlikely to have an adequate impact on iron nutrition at the fortification levels currently employed, although preliminary analysis of a new H-reduced product indicates that it may be possible to improve the bioavailability of individual powders of this type of product. We did find significant differences among products in both the electrolytic and carbonyl categories. Therefore, all products should be screened rigorously.