Serum vitamin B12 levels of 2523 individuals were measured using the hemoglobin-coated charcoal assay. In one hospital, the test was performed on 1698 patients under hematologic evaluation for erythrocytic abnormalities of all types, while 825 patients from five other hospitals were studied because of suspected B12 deficiency. The incidence of low serum vitamin B12 was surprisingly low for both groups (2.3% and 3.5%, respectively). The value of the test as a diagnostic tool was further reduced by two characteristics of physician performance. First, with those patients who demonstrated a clear-cut macrocytic anemia and were suspects for intrinsic factor deficiency (pernicious anemia), extensive evaluations for B12 malabsorption were usually carried out prior to the receipt of the low B12 level. Second, when a deficiency state was not suspected, the return of a low value did not reliably stimulate physicians to begin evaluation or institute therapy. This was true despite the presence of clinical evidence of a possible gastrointestinal abnormality in many of the patients.