The iron status of a population of 1564 subjects living in the northwestern United States was evaluated by measurements of transferrin saturation, red cell protoporphyrin, and serum ferritin. The frequency distribution of these parameters showed no distinct separation between normal and iron-deficient subjects. When only one of these three parameters was abnormal (transferrin saturation below 15%, red cell protoporphyrin above 100 mug/ml packed red blood cells, serum ferritin below 12 ng/ml), the prevalence of anemia was only slightly greater (10.9%) than in the entire sample (8.3%). The prevalence of anemia was increased to 28% in individuals with two or more abnormal parameters, and to 63% when all three parameters were abnormal. As defined by the presence of at least two abnormal parameters, the prevalence of iron deficiency in various populations separated on the basis of age and sex ranged from 3% in adolescent and adult males to 20% in menstruating women. It is concluded that the accuracy of detecting iron deficiency in population surveys can be substantially improved by employing a battery of laboratory measurements of the iron status.