Normal dogs and two Eck fistula dogs, receiving a daily diet containing an average of 1 gm. of vegetable protein per kilo of body weight, showed after average intervals of 7 to 9 weeks, slight decreases in amounts of circulating plasma protein (Table 21). A third Eck fistula dog under similar circumstances was unable to maintain its plasma protein concentration above the edema level. This dog by biopsy was shown to have an abnormal liver and the evidence indicated that the other organs were normal. The animal showed active thirst and diuresis as compared with controls (Table 25). This Eck fistula dog had less than one-tenth the capacity of the normal dog to form new plasma protein when various food proteins were added to the basal diet, and no significant quantitative differences in the relative potency of these foods (liver, kidney, heart muscle, soy bean, salmon) could be distinguished (Table 22). It appears that the liver abnormality is responsible for this abnormal reaction. This observation gives strong support to the thesis that the liver is actively concerned with fabrication of new plasma protein.

No abstract

Any experimental technique which involves disturbance of liver function is of particular interest in regard to plasma protein studies, for there is evidence that the liver is associated in the production of plasma protein. When a solution of gum acacia is injected intravenously in the dog, there is a marked decrease in the blood plasma protein concentration (5). That the liver is concerned with this process is suggested by the observation that injected gum acacia is rapidly removed from the blood and deposited to a large extent in the liver, whence it is slowly eliminated in the bile (4).In any attempt to evaluate this phenomenon following gum injection, it is important to ascertain, first, whether decrease in plasma protein amcentratio~ is associated with a decrease in total circulating protein, and second, the degree of impairment of liver function under these conditions. The data given here represent such an attempt: By means of plasma volume determinations it was possible to show a decrease in the total circulating protein. Determinations of plasma fibrinogen were made as a means of estimating the relative degree of hepatic function. Marked disturbance in fibrinogen concentration was demonstrated both in dogs receiving single injections of the gum, and in those in which repeated doses were given. It was possible by means of repeated weekly injections of acacia to maintain dogs at low plasma protein and fibrinogen levels for several weeks. Following 1 We are grateful to Eli Lilly and Company for furnishing valuable material used in these experiments.

This paper deals with the factors responsible for the pigmentary disturbances previously described (8) in the splenectomized bile fistula dog. In the absence of demonstrable infection in such animals it had been assumed that the periods of bile pigment surplus and anemia were of physiological origin related to a lack of spleen and bile constituents. The observations given below point to "Bartonella canis" as the probable cause. The bile fistula dog alone shows no such changes, and the splenectomized dogs in our kennels have not spontaneously developed an anemia. Dogs having a combination of splenectomy and a bile fistula, sooner or later develop the picture of anemia and pigment excess described below. This condition was first described by Hooper and Whipple (2), working with open bile fistula dogs subsequently splenectomized. These dogs always showed moderate infection of the bile fistula tract. The use of the closed sterile bile fistula (10) and the gall bladder-renal fistula (3) did away with bacterial infection of the biliary tree, yet such dogs when splenectomized showed the same periods of anemia and excess pigment ' output.

In an earlier paper from this laboratory (4), it was pointed out that the splenectomized bile fistula dog showed cycles of bile pigment overproduction and anemia. Because of the very great amounts of bile pigments produced under these conditions we studied this reaction in considerable detail with the hope of a better understanding of body pigment metabolism. Our first paper (4) came to the conclusion that some obscure intrinsic factor related to the spleen was responsible for this reaction. The accompanying paper gives strong evidence that an extrinsic factor (Bartonella canis) is responsible. Splenectomy prepares the way for the Bartonella infection.Granting that Bartonella infection explains the blood destruction with resulting surplus bile pigment production, the mechanism of hemoglobin production under these conditions is of peculiar interest to students of anemia. These dogs must produce enormous amounts of new hemoglobin and red ceils during an active cycle on a diet which permits of but little new hemoglobin formation in simple anemic dogs. We have suggested that from the released hemoglobin the pyrrol aggregate is split off to form bile pigment while the globin fraction is turned over to form new hemoglobin. This proposal assumes that the body can readily produce a large excess of the pyrrol aggregate (four pyrr01 rings) and there is much evidence (1) to support this argument.To gain further information about blood destruction in bile fistula dogs we carried out a prolonged experiment with acetyl phenylhydrazine (Table 23) which gives similar results and supports the Bartonella observations.