Notwithstanding the diversity in the method of experimentation, there was a striking uniformity in some of the results obtained. All the experiments of the series showed absence of cumulative action of caffein. The experimental evidence presented indicates that moderately large doses may be given at intervals of about 24 hours without inducing any symptoms of nervous or any other disturbance. This is illustrated in the tests on dog 11, which were preliminary in character. In this subject 100 to 150 mg of caffein per kilo were ingested daily for several days without showing any changes. Later in the course of the experiment, after larger doses were given, mild symptoms only, such as tremors, were observed. Additional evidence of the absence of cumulative action of caffein was furnished by the results of the following experiments: Dog 23 received 142 mg of caffein per kilo on three successive days. His general condition indicated that these amounts of caffein were toxic, but he survived. In another series of tests, made after he was allowed to rest a few days, he again failed to show any cumulation of the drug, as he survived this time a series of tests of longer duration than the first. A much better illustration of the absence of cumulative action of the drug is furnished by the experiments on dog No. 20. In this case 100 to 125 mg of caffein per kilo, given on 10 consecutive days, did not cause any marked effects. Diarrhea and restlessness were the only symptoms observed. These experiments therefore show that the elimination and decomposition of caffein are apparently effected by the body within twenty-four hours or thereabouts. Experiments on dog 19, however, form an exception--the third dose of 125 mg caffein per kilo having proved fatal. The very low protein content of the diet of this dog suggests itself as a possible cause of the lower resistance to caffein of this subject. But it may be observed that the same diet was furnished to dog 20, which stood such amounts of caffein much longer. The presence of a trace of albumin in the urine of dog 19 is likewise inadmissible as a cause of the difference in the toxicity of caffein in this dog, for the urine of dog 20 likewise contained a trace of albumin. The alkaline reaction of the urine, together with the fact that the first dose of only 60 mg of caffein per kilo induced symptoms of toxicity, suggests the presence of an abnormal condition which in all probability was the cause of the death of this subject under the conditions indicated.