By a method of permanent intubation the entire output of bile from dogs has been obtained in a sterile state over long periods of time, and studied quantitatively. The secretion of the first few days after the initial operation is scanty, contains more pigment than that secreted later, and is sometimes so viscid as to cause temporary obstruction within the collecting tube. The small amount and thick consistency of the fluid are referable to direct injury of the liver and ducts, while the abundant bilirubin is derived, in part at least, from the hemoglobin of extravasated blood. In some animals in which there was accidental mercuric chloride poisoning at the time of operation a suppression of the bile followed which, in one instance, was complete during the 48 hours before death. In a dog developing mixed infection of the biliary tract, an almost colorless fluid, glairy as white of egg, gradually took the place of the bile. Not until a week or 10 days after operation does the bile acquire the character which it maintains later. The quantity of this later bile, as measured in successive 24 hour specimens, is greatly less than that recorded by previous investigators, a difference attributable to disturbing influences inevitable to the method of collection they employed. By our method, it averaged from 3.5 to 9.5 cc. per kilo of dog in 24 hours, though transient variations from 1 cc. up to 14 cc. were encountered. Some dogs give consistently far more bile than others. The frequently recorded effects of fasting to lessen the rate of secretion, and of a meat diet to increase it more than a carbohydrate one, were noted often. Contrary to expectation, vigorous exercise does not act to increase the quantity of the bile. During hot weather this may sink greatly, although the animal remains in good condition; and during intercurrent diseases unassociated with jaundice, the flow may almost cease. One of the best of cholagogues, bile by mouth, fails of effect under such circumstances. It acts best when the animal is healthy, the weather not oppressive, and the food intake abundant. The bilirubin output, after the immediate effects of the operation have worn off, remains nearly constant from day to day, though often exhibiting slow, wave-like variations, each extending over a week or more. These slow changes are synchronous with similar alterations in the hemoglobin percentage of the circulating blood. A recognition of them is important to studies of the bilirubin yield. A mild anemia, absent in controls maintained under like conditions, regularly develops in unexercised dogs completely deprived of bile, despite their apparent health, and as it does so the bilirubin output falls off. Prior to the development of the anemia the output averages about 7.5 mg. per kilo in 24 hours, a yield which closely corresponds with that recorded by previous workers. Vigorous exercise of animals previously sedentary causes an increased pigment output presumably as result of increased blood destruction. Since the daily pigment output is approximately constant, whereas the fluid quantity undergoes frequent and great changes, it follows that the pigment concentration must vary inversely as the fluid quantity, and vary greatly. It does both. No matter how large or small the 24 hour specimen of bile may be within "normal" limits, therein will be found the customary quota of pigment. When, under pathological conditions, the bile flow almost ceases, the pigment concentration becomes extremely great. A similar reciprocal relationship between concentration and fluid quantity would seem to hold for the mucinous element of the bile. Scanty biles are ropy as a rule, and copious ones are watery. Temporary obstruction was produced by clamping the outlet tube from the common duct, and the bile yield following upon its relief was studied. Secretion was for a time far more copious than usual, with a low pigment content per cubic centimeter, and tended to remain so until the accumulated pigment had been voided. As much extra bilirubin is put forth after 24 hours of obstruction as if the liver continued to manufacture the pigment without interruption during it. Obstruction for this length of time appears to cause no subjective disturbance in most instances, though bilirubinuria develops. Even this sign of bile retention may be lacking when obstruction endures but 12 hours. The physiological and clinical significance of these facts is briefly discussed.