In subchronic nephritis with oedema this criterion is less reliable. In this disease periodic fluctuations in the excretion of albumin seem to occur spontaneously and seem to be altogether uninfluenced by the diet; they are in fact independent of the changes in the general disease picture. An attempt to estimate the effect of any given dietary on the course of the disease from the fluctuations in the amount of albumin excreted may lead to very serious errors.

This may be seen very clearly from the following case, that I had occasion to observe continuously during two months. The patient was a student of medicine, twenty-two years old, who was confined to bed on account of moderate oedema, and who consequently collected his urine very carefully. No drugs at all were administered. The patient was also forced to do without bathing and sweating, because he had only re cently recovered from an attack of pericarditis. Whenever the attempt was made to produce diaphoresis he developed violent palpitation.

First Period. - Pure milk diet. (Fourteen days.) Pure milk diet, i.e., daily 21/2 liters of milk and 3-10 of a liter of cream. The average amount of urine excreted was 1,950 c.cm., the average amount of albumin 8.7 g. (precipitation of the albumin by boiling with the addition of sodium chlorid and acetic acid. Determination of the nitrogen in the washed precipitate).

Second Period. - (Fourteen days.) Instead of the 2 liters of milk, the substitution of 200 g. of meat (weighed raw, usually beef, sometimes veal), 2 eggs, 100 g. of white bread and 50 g. of butter, 1,500 c.cm. of seltzer water, in addition 1/2 liter of milk and 3-10 of a liter of cream. The percentage of nitrogen and the calories of the diet were approximately the same as in the first period. The average amount of urine excreted was equal to 2,100 c.cm.; the average amount of albumin was 4.2 g. a day. During the first two days of this second period the excretion of albumin was slightly increased, but then was rapidly and continuously decreased (down to 2.6 g.) and then rose a little on the last three days (to 3.8 to 4.1 g.).

Third Period. - (Fourteen days.) Milk-cream diet exactly as in Period One. The average excretion of urine was 2,050 c.cm.; the average excretion of albumin 3.8 g. in the first week, and 7.2 g. in the second week. The excretion of albumin was still increasing when the fourth period was begun.

Fourth Period. - (Fourteen days.) The diet the same as in the second period. The average excretion of urine was 1,780 g. (The weather was very hot.) The average excretion of albumin was 8.1 g. in the first week, and then rapidly fell in the second week, namely, to 6.9 - 5.7 - 5.1 - 5.4 - 4.2 - 3.7 - 3.6 g., that is, an average of 4.9 g.

Six weeks later the patient was again placed under observation for eight days, and a mixed diet selected at random. The general condition of the patient had not under-gone any particular change in the meantime. The average excretion of urine was 1,680 c.cm. and the average excretion of albumin 6.1 g.

The patient died four weeks later from erysipelas.

We see distinctly from this series of observations how uncertain it is to draw any conclusions in regard to the value of a certain diet from a short series of determinations of the albumin excretion. If we were to limit ourselves to considering single short periods of these series, and to compare the different periods, we might interpret them to signify that in some cases the addition of meat and eggs was beneficial, in others again that it was detrimental. This case is not by any means isolated - in fact, I consider it typical, and the careful and protracted studies that I have carried out in different cases show that a periodic rise and fall of albuminuria is a common symptom. This should teach us that fluctuations in the excretion of albumin are valueless as an indicator of the advantages or disadvantages accruing from a certain diet.

In chronic contracted kidney albuminuria is still more unreliable as an index. The degree of albuminuria is, on the average, much smaller than in the subchronic form; even if the percentage of albumin excreted is doubled, the total excretion is usually only increased by a few tenths of a gram; if, for instance, a patient is voiding 2,500 c.cm. of urine and excreting a quarter pro mille of albumin on one day and a half pro mille on another, the difference in the daily excretions only amounts to 0.6 g. Acute exacerbations, that occur quite frequently in contracted kidney, particularly in young subjects, almost regularly produce an excretion of more albumin, and at the same time usually haematuria. The latter symptom slowly disappears; just as in acute nephritis. Aside from these fluctuations in the excretion of albumin, the total quantity of albumin passed in a day usually fluctuates within very narrow boundaries. I have frequently observed that any sudden change in the diet usually leads to a remarkable increase of the albuminuria, regardless of whether we pass from a milk diet to a meat diet, or inversely, from a meat diet to a milk diet, or to a vegetarian diet. At the end of a few days, or at most at the expiration of about a week, the old level is reached again.

Here, then, we see once more that we are by no means justified in drawing hasty conclusions from a comparison of short periods; even when longer periods are utilized for purposes of comparison, we should proceed with great care; whereas in acute nephritis in crease and decrease of albuminuria are apparently dependent on improvement or exacerbation of the primary disease; and whereas this also applies to a limited degree to subchronic forms of nephritis, we must concede that in chronic nephritis the degree of albuminuria is quite independent of the disease picture at any given moment. We may find, for instance, that the disease picture becomes more and more threatening, while at the same time the albuminuria decreases, and inversely we may find that the excretion of albuminuria increases (both procentically and absolutely) while at the same time certain threatening symptoms disappear. In cases of this kind, then, we witness a paradox phenomenon, namely, that the patient improves, and that the prognosis grows more favorable, while at the same time the albuminuria increases. This is a prolific source of error for those physicians who gauge the result of their therapeutic measures principally from the degree of albuminuria; it also constitutes a source of continuous and altogether unnecessary worry for those patients who are in the habit of frequently sending their urine to the apothecary for quantitative analysis, and who make their mood and their happiness dependent on the result of this analysis.