Folate assays by use of radiolabeled folate provide obvious practical advantages over the standard microbiological assay, but remain incompletely tested. We therefore compared results for 415 sera with a kit involving 3H-labeled folate and the Lactobacillus casei microbiological method. We examined the patients' data when there were discrepancies between the two methods. Although the correlation overall was satisfactory, results were discrepant in 25% of cases. In 74% of the latter, the radioassay result appeared to be the correct one, primarily because L. casei results were suppressed by antibiotics being taken by the patient. The radioassay occasionally gave falsely high values for patients with liver disease and falsely low ones for patients who had received isotopes for scanning purposes. Several assay kits that make use of 125I- or 75Se-labeled folate were also tested. Although these results correlated with the results of 3H-labeled folate assay, various problems appeared, including the possible need for serum-supernate control tubes in one kit. Answers to these and other questions and careful clinical correlation of results are needed for any folate radioassays before their adoption for routine clinical use.