Synthetic human C-peptide bearing a Tyrosine group at its amino end is labelled with 125iodine using chloramin T or hydrogen peroxide and lactoperoxidase. The results are compared applying both methods. Antiserum to synthetic human C-peptide (without Tyrosine) which was partially compared to rabbit albumin, is raised in guinea pigs and goats. Goats show to be superior to guinea pigs concerning antibody production. The so-called "hook effect" phenomenon is observed in setting up the standard curves for the radioimmunoassay. Monotonically decreasing standard curves are obtained on dilution of antiserum with a high antibody titer which was produced by repeated immunization in goats. Free C-peptide and C-peptide bound to antiserum are separated with the anxion exchange resin Amberlite. Using this separation technique we excluded unspecific binding of labelled C-peptide to protein fractions in serum of diabetics. The sensitivity of our radioimmunoassay is approx. 0.3 ng C-peptide/ml serum. Intra- and interassay variability are below 10%. Human proinsulin is the only substance found to crossreact with the antiserum.