1. The turnover rate of urinary Tamm-Horsfall glycoprotein in rabbits was determined by two different methods. The first involved measurement of the pool size of the glycoprotein in rabbit kidney and the daily urinary excretion rate by a radioimmunoassay from which the turnover rate was calculated. 2. The second method made use of the incorporation in vivo of Na(2) (14)CO(3) and sodium [(14)C]acetate. After a single intramuscular injection of one of these compounds, urine collections were made every 24h and the glycoprotein was isolated and its specific radioactivity was determined. 3. Incorporation of the label into urinary HCO(3) (-), urea and plasma fibrinogen was also examined. The specific radio-activities of the O-acetyl, sialic acid, aspartic acid and glutamic acid residues isolated from the Tamm-Horsfall glycoprotein were compared and their half-lives were compared with that of the intact glycoprotein. The two methods gave results in quite close agreement and indicated a half-life for the glycoprotein of approx. 9h. 4. An attempt was made to localize the glycoprotein within the kidney and within the cell. It is present throughout the kidney, but was not detected in the brush-border fraction isolated from the proximal tubules. From differential cell-centrifugation studies, the glycoprotein seemed to be predominantly present in the soluble fraction (100000g supernatant). This suggests that it is either largely a soluble cytoplasmic component or is very loosely bound to a membrane, being readily released under the gentlest homogenization procedure. 5. The half-life of Tamm-Horsfall glycoprotein in human kidney was found by the radioimmunoassay method to be approx. 16h. The similarity between the composition of Tamm-Horsfall glycoprotein and human erythropoietin is discussed.