It is shown that marine teleosts swallow relatively large quantities of sea water which subsequently undergoes absorption. Most of the ingested water, Na, K and Cl are absorbed from the intestine, leaving a residue rich in Mg and SO4 and approximately isotonic with the blood. The salts of the urine consist primarily of Mg and SO4, with only traces of K and small quantities of Na and Cl, the solution being isotonic or hypertonic to the blood. It is shown that the intestinal Mg and SO4 come from ingested sea water, and that the urinary Mg and SO4 come from alimentary absorption of these substances. Quantitative analysis of the composition of the ingested sea water, the intestinal residue and the urine lead to the following conclusions: 1) The os-motically dilute nature of the intestinal residue and urine elaborated from sea water shows that the marine fish is doing osmotic work at some point other than in the kidneys and gastro-intestinal tract. 2) The fact that Mg and SO4 occur in greater concentrations in both the intestinal residue and the urine than in the ingested sea water from which these fluids are derived shows that part of the water absorbed from the alimentary tract is excreted by some route other than the kidneys. The fact that relatively more Na, K and Cl are absorbed from the alimentary tract than Mg and SO4, while less Na, K and Cl are excreted in the urine than Mg and SO4 shows that a large part of the absorbed NaCl and KC1 is being excreted by some route other than the kidneys. By considering these 3 processes as identified with a single process, it is inferred that the marine fish excretes by some extrarenal route a solution of NaCl and KC1 which is hypertonic to the ingested sea water and thus leaves part of the absorbed water free for the formation of the osmotically dilute urine and intestinal residue. The extrarenal excretion of water is demonstrated in Anguilla and Myoxocephalus by the addition of phenol red to sea water and the quantitative, simultaneous determination of the quantity of water ingested and the quantity of urine excreted in a given time. The extrarenal excretion of salts is demonstrated in the catadromous eel, Anguilla, when kept in fresh water. The greater part of the NaCl and KCl absorbed from the alimentary tract is excreted by the extrarenal route, while essentially all the Ca, Mg, SO4 and PO4 are excreted by the kidneys. In the absence of other salts, SO4 takes Na into the urine with it, and Mg and Ca take Cl. Reasons are given for believing that the extrarenal excretion of salts and water and the extrarenal performance of osmotic work are attributable to the gills. The water for the formation of urine in fresh water fish is not absorbed from the stomach or intestine, but probably from the oral membranes. The water cycle in marine fish is of such a nature as to reduce the quantity of water excreted as urine to a small fraction of that excreted in fresh water fish. It is suggested that this relative anuria is the principal causative factor in an evident tendency for the glomeruli to disappear in the kidneys of strictly marine teleosts.