Osteocalcin is initially synthesized as an 11 kD molecule consisting of a 23‐residue translocation signal peptide that is cleaved during translation, a 26‐residue propeptide that targets the protein for gamma‐carboxylation, and the 49‐residue mature protein. Although the majority of newly synthesized osteocalcin is deposited into bone matrix, a small amount can be detected in blood, and it is this characteristic that has led to its current clinical use as a specific index of osteoblastic activity. Nothing is known, however, about the fate of the propeptide. If osteocalcin and the propeptide are cosecreted, then the concentration of the propeptide could also be useful as a marker of osteoblastic function and, further, may be superior to osteocalcin because it would be unaffected by binding to bone. To test this hypothesis, we synthesized a peptide corresponding to 21 residues of the osteocalcin propeptide from humans and produced a polyclonal antibody to this peptide. Human sera were screened for the presence of the propeptide, and the human osteosarcoma cell line MG‐63 was tested for secretion of the propeptide. We could not detect any osteocalcin propeptide in sera from normal adults or individuals with renal failure or primary hyperparathyroidism or those on long‐term coumadin therapy. Likewise there was no propeptide present in media from cells grown in the presence of vitamin K, l,25‐(OH)₂D₃, warfarin, or warfarin plus 1,25‐(OH)₂D₃. In contrast, the cell extract, characterized by high‐performance liquid chromatography, contained mature osteocalcin, free propeptide, and the proosteocalcin precursor when cells were grown in the presence of 1,25‐(OH)₂D₃ alone. When warfarin was added to the culture, only proosteocalcin precursor was detected in the cell extract. These data establish that the osteocalcin propeptide is not secreted in vitro or in vivo and verify that carboxylation is not required for the conversion of the precursor to osteocalcin.