Previous investigators have found that complete occlusion of the ureter may lead either to hydronephrosis or to atrophy. In Lindemann's series of six dogs;, for instance: Two animals showed simple hydronephrosis, three animals showed simple atrophy, and in one animal the kidney was slightly enlarged and the ureter and pelvis dilated, but fluid was absent. In his series of four rabbits, all showed hydronephrosis. The result, whether hydronephrosis or atrophy, is evidently not determined by the time elapsing after the operation. Lindemann found that the intrapelvic pressure resulting from the ligation obliterates the lumen of the vessels, first of the veins and subsequently of the arteries; but that this is compensated by an increase of the collateral blood supply through the capsule, the degree of this compensation determining the presence or absence of hydronephrosis. If the blood supply is free, the fluid after tapping will accumulate again and again. It is somewhat remarkable that all of our dogs showed hydronephrosis after the first operation. The results of establishing a urinary fistula differed in the two cases in which it was tried: The fluid did not re-form in Dog 3 even when sodium sulphate was injected: whereas in Dog 4, a very abundant quantity of fluid reaccumulated spontaneously; but it differed notably in composition from the original fluid, having more the character of a purulent inflammatory exudate. The histological changes consist in necrosis of the renal cells, obliteration of the glomeruli, increase of connective tissue, and endarteritis and periarteritis. Different areas in the same kidney are affected in very different degree, some areas appearing almost normal. The glomeruli are generally less altered than the tubules. The collecting tubules are generally displaced so as to run parallel to the surface; many are dilated. The changes correspond closely to those described by Lindemann. The sound kidneys showed slight hyperaemia and hypertrophy, but no necrosis. This corresponds with the findings of Pearce and of Ames. The uniformity in chemical composition of the fluid obtained, after the first operation, from the four dogs, as shown by Column III of Table I, is very striking, and points to a uniform origin by a process which is but little affected by the interval elapsing after the operation. The specific gravity, total solids and proteids correspond to those of a very dilute lymph, being but a trifle above those of cerebro-spinal fluid and aqueous humor, and much lower than those of serum, lymph and most cystic fluids (the proteid content of the latter being generally from 2 to 6.5 per cent.). The human fluid (Column II) which had remained in the kidney for a very long time had a particularly low proteid percentage; while that of the second fluid of Dog 4 (Column VIII) was very much higher; this last fluid having a pronounced inflammatory character and being of recent formation. The absence of notable amounts of the specific urinary constituents is particularly important.(4) Odorous principles are entirely absent. Urinary pigments appear to be present in the four dogs' urines, but absent from the human case,(5) and after the second operation in case of the dogs. It seems fair to assume that the pigments were secreted shortly after the ligation, when the kidneys were still functional, and that they were reabsorbed with extreme slowness. Urea was present in all the fluids, but its quantity was very small in the dogs, and probably in the human case. It is on the whole somewhat greater than in the serum (0.103 per cent., in place of 0.05 per cent.), but the difference may be within the analytical error. The same applies to the ammonia, phosphates and sulphates. An important difference between these fluids on the one hand and blood serum, lymph and ordinary exudate on the other, lies in the higher contents of chlorid, and the consequently greater molecular concentration. The ordinary chlorid content of body fluid varies between 0.55 and 0.70 per cent., mean about 0.6 per cent (as NaCl), while that of the first kidney fluid, in the dogs, varied between 0.68 and 0.75 per cent., mean 0.725 per cent.; that of the human fluid was 0.83 per cent., that of the second fluid of Dog 4 only 0.52 per cent. The depression of the freezing point in the human case was 0.715 degrees C., as against the normal value, for human serum, of 0.491 to 0.562. (Possibly the blood of this patient had a higher concentration than normal, since uraemia existed.) The high chlorid percentage has evidently no relation to the length of time during which the fluid sojourned in the kidney. It is probably to be explained by the relatively slow absorption of this ion from the kidney pelvis. It is also to be remarked that cerebrospinal fluids generally have a somewhat high chlorid content (0.573 and 0.6 per cent.), but this never reaches the height of these ureteral fluids.