When immunoreactive human parathyroid hormone (hPTH), extracted by three different solvents (20% acetone in 1% acetic acid, 8 M urea, or normal saline) from parathyroid glandular tissue was subjected to Sephadex G-100 gel filtration and immunoassay using two different antisera (273 and C-329), four distinct fractions were observed. The first (I), a void volume peak, was detected by both antisera with similar immunoreactivity, as was a second (II), which had the elution and sedimentation properties of highly purified bovine parathyroid hormone (bPTH); a third (III) eluted between [(125)I]growth hormone and [(125)I]insulin, sedimented with the velocity of a molecule of approximately 6,000 mol wt, and was detected primarily by antiserum 273; a final fraction (IV), detected primarily by C-329, eluted just prior to [(125)I]insulin. The elution profiles of the acetone-acetic acid and 8 M urea extracts were similar and contained fraction II as their major component. In saline extracts, however, fraction III predominated. Three fractions, having gel filtration and immunologic characteristics similar to fractions II, III, and IV, respectively, of saline glandular extracts, were detected in the plasma of patients with both primary (adenomatous or carcinomatous) and secondary hyperparathyroidism. The predominant component in every plasma was the intermediate fraction that, like III, was detected primarily by antiserum 273, while the least abundant form was consistently the final fraction, detected primarily by antiserum C-329. The first fraction, like II, was detected with about equal potency by both antisera and had an elution volume on Sephadex corresponding to that of intact bPTH. It bore a reciprocal relationship to serum calcium and disappeared from the plasma of a uremic patient during calcium infusion or following parathyroidectomy with a half-time of no more than 20 min. This component therefore probably represents biologically active hormone. The intermediate and final fractions had turnover times in the plasma of a uremic patient more than 100 times greater than the active form, remained elevated even in the presence of post-parathyroidectomy hypoparathyroidism in this patient and were presumed to be biologically inactive. The ratio of biologically inactive fragments to the active form was greater in secondary hyperparathyroidism. The evidence presented favors a glandular origin for the fragments. Comparison of hormonal assays with the two antisera reveals a striking advantage in the preoperative diagnosis of primary hyperparathyroidism with antiserum 273 that is due to the enhanced sensitivity occasioned by its detection of a biologically inactive as well as the biologically active hormonal form.