Urinary glycosaminoglycan excretion was studied in 24 cases of disseminated neoplasm, 12 of which had unequivocal evidence of skeletal involvement. Urinary hydroxyproline, cetylpyridinium chloride (CPC)-precipitable uronic acid, and CPC-precipitable hexosamine were expressed as a ratio to urinary creatinine. Glycosaminoglycans contained in urine concentrated x 1000 by vacuum-dialysis were separated by electrophoresis on cellulose acetate and stained with alcian blue. Of the 12 cases with clear evidence of skeletal involvement, eight (66%) showed elevation of serum alkaline phosphatase, five (42%) showed elevation of urinary hydroxyproline, and three (25%) showed elevation of urinary uronic acid. It is concluded that urinary uronic acid is not a sensitive index of skeletal involvement in disseminated neoplasm. The most striking feature of the study was the identification of a well-defined fraction indist inguishable from hyaluronic acid in seven (58%) of the cases with evidence of skeletal involvement. Hyaluronic acid is not normally identifiable in adult human urine. The hyaluronic acid excretors showed more consistent biochemical evidence of bone disease (elevation of serum alkaline phosphatase and urinary hydroxyproline) than the non-excretors. The possibility that the urinary hyaluronic acid is derived from degradation of skeletal hyaluronic acid is discussed. An alternative explanation is that the hyaluronic acid is derived from neoplastic cells as part of a reversion of glycosaminoglycan synthesis to a more ;fetal' state, a glycosaminoglycan counterpart of the production of oncofetal antigens by neoplastic cells.