T wo O-hydroxylysyl glycosides, Hyl-Gal-Glc and Hyl-Gal, have been isolated from normal human urine and shown to be identical to two glycosides isolated from alkaline hydrolysates of collagen. A relatively sample and reproducible analytical procedure has been devised to measure the levels of these glycosides in human urine. By the use of this procedure it was shown that a normal diet has only a small effect on 24-hr urinary excretion levels of these glycosides indicating an endogenous origin. Urinary glycoside levels appear to be highest in children, roughly paralleling collagen turnover as indicated by urinary hydroxyproline levels. Collagen turnover equivalents calculated from urinary hydroxylysyl glycoside levels were found to be significantly larger than collagen turnover equivalents calculated from urinary hydroxyproline levels. This suggests that urinary glycosides are more quantitative indicators of collagen metabolism than urinary hydroxyproline.The ratio of Hyl-Gal-Glc to Hyl-Gal was measured in urines of diseased as well as normal individuals and a bimodal distribution was found. Alkaline hydrolysates of different human connective tissue collagens showed that only bone collagen, of the collagens examined, had a low ratio of Hyl-Gal-Glc to Hyl-Gal compared to human urine. Other collagens examined had higher ratios than found in human urine. On the basis of these results it is postulated that the bimodal distribution of glycoside ratios represents two populations of collagen turnover, the lower ratio population having a high bone collagen turnover, the lower ratio population having a high bone collagen turnover relative to the second population. Examination of the types of subjects making up the two populations supports this hypothesis. These data suggest that urinary O-hydroxylysyl glycoside excretion, in addition to providing a more quantitative estimate of collagen turnover than urinary hydroxyproline, may prove to be of value as a specific means of studying the metabolism of bone collagen.