Marked discrepancies (values up to four times higher in on assay than in the other) were observed when the plasma concentration of immunoreactive human calcitonin (iCT) was measured by two radioimm8noassays in 18 patients with medullary thyroid carcinoma. The two antisera used had different binding affinities for the NH2- and COOH-terminal regions of synthetic calcitonin monomer (CT-1-32). Except for this difference, the assays were identical and reacted equally with CT 1-32. Plasma samples from patients with medullary thyroid carcinoma were gel filtered on columns of Bio-Gel P-150, and the immunoreactivity in column effuent fractions was measured with both assays. The one utilizing the antiserum with prominent NH2-terminal binding affinity (and giving higher iCT values) recognized at least five molecular species that eluted with or before CT 1-32. The other assay, utilizing the antiserum with a COOH-terminal binding affinity, recognized two fo these molecular species-one eluting with CT 1-32 and the other in a position consistent with a dimer. A mixture of athreotic asthma and added CT 1-32 contained a single immunologic species that was recoqnized equally by both antisera. No forms smaller than CT 1-32 were detected in any study. The results suggest that iCT circulating in the plasma of patients with medullary thryoid carcinoma is hetergeneous. The absolute iCT concentration measured by radioimmunoassays depends on recognition of these distinct molecular species as well as on the specific binding affinities of the antiserum used to detect them. These observations may partially explain the variations among iCT values reported by different laboratories.