The isotope dilution method has been useful in measuring either secretion rates or production rates of several important steroid hormones. This technique consists of measuring the dilution of a tracer amount of hormone into the plasma pool from the cumulative specific activity of a unique urinary metabolite of that pool. If the plasma pool is derived entirely from glandular secretion of the hormone, then a secretion rate is measured. If, however, both glandular secretion and peripheral conversion of one or more precursors contribute to this pool, then a production rate is measured. In either case the validity of the method depends strictly upon the presence of a urinary metabolite uniquely derived from the pool in question. Since it has generally been assumed that a steroid hormone conjugate would be such a unique metabolite, secretion rates of aldosterone and, in women, estradiol have been measured in this way.In discussing an isotope dilution method (1) for measuring the testosterone production rate (TPR), we considered the possibility that testosterone glucuronoside (TG) may not be a unique metabolite of the plasma testosterone (T) pool. This situation could arise as follows: It has been shown that T can be synthesized peripherally from both androstenedione (A) and dehydroepiandrosterone (D) (2-5). Presumably this transformation occurs in the liver, and indeed this has been demonstrated by perfusion studies with dog liver (6). When T is synthesized either in the liver or at any other peripheral site, if conjugation occurs before entry of this T into plasma some TG will be excreted that is not derived from plasma T. This concept is diagrammed in Figure 1. The portion within the * Submitted for publication April 15, 1964; accepted July 9, 1964. Presented in part at the Forty-sixth Meeting of the Endocrine Society, San Francisco, 1964. dotted line represents that hypothetical moiety of T derived from A, which is conjugated immediately. We have designated this route the A -> TG pathway.The following studies were undertaken to test for the existence of this pathway and evaluate its quantitative significance. Androstenedione-H3 and T-C14 were infused, and the H3/C14 ratios of free and conjugated T were compared. If all the T derived from A re-entered the plasma as free T, then the H3/C14 ratio of free and conjugated T should become equal as metabolism proceeds towards completion. Conversely, if all the T synthesized from A underwent immediate conjugation, plasma free T would contain no tritium. Our results indicate that the A -+ TG pathway exists, that it may contribute substantially to urinary TG, and therefore that urinary TG is not a unique metabolite of the plasma T pool. Thus the TPR must overestimate production of plasma T.