Causes of the errors above described readily suggested themselves because the patient from whom the uriñe was obtained was in diabetic coma and secreting many substances either totally foreign to the urine or abnormal quantities of normal constituents. Of the latter, ammonia was the most striking. On Oct. 18, this amounted to 3.07 grains and on the next day to 4.39 grains in contrast to the normal daily excretions of 1 gram. This led me to employ a solution of ammonia instead of urine in the

more »

... aratus for the urea test. The result was not unexpected. Water was displaced in quite the same way as if urine liad been added. Eijkmann 4 several years ago showed that this reaction took place. The quantity of acetone excreted by this same patient was the largest I have found recorded, -Oct. 18, 32.3 grains, Oct. 19, 26.7 grams. The test was, therefore, repeated, using a solution of acetone instead of urine. Water was freely displaced. Boxybutyric acid was also similarly tested, and with like result. On Oct. 18 this acid reached 96.3 grams and on the following day 69.8 grams. It is evident from the above that the hypobromite method is not safe clinically when ammonia, acetone and b-oxybutyric acid are in the urine. Other bodies like creatinine, uric acid and urates also interfere with the test, but to a less extent. Since acetone, b-oxybutyric acid and ammonia are chiefly found in diabetic patients, it is a good rule to scrutinize carefully reports of large quantities of urea in this disease. If these bodies are present, and this is readily determined, such reports are untrustworthy if they are based on the hypobromite method. The simplest procedure to enable us to judge of the presence of these substances is the ferric chloride test for diacetic acid. When this is positive, acetone and b-oxybutyric acid are always present and an excess of ammonia is probable.