Having gained some information respecting the effects of repeated doses of caffein on rabbits, it was of interest to find out how carnivora reacted to the drug when similarly administered. A number of dogs were used for the purpose. Considerable variation in the mode of experimentation, as will appear later, was allowed. Since the condition of the animal, its age, environment, or diet might be factors influencing toxicity, tests were made on full-grown and on young growing dogs whose food was varied. The subjects of the experiment were kept under observation for a few days to several weeks before the administration of caffein was begun, in order to determine whether or not any morbid condition existed, as well as to ascertain whether the new environment had any effect on these animals. Caffein was given chiefly by mouth, but the subcutaneous method was also employed during a portion of the experimental period in some dogs. The initial dose, which varied for different individuals, was maintained for a variable length of time. It was then progressively increased, in most cases until the death of the animal. With larger doses the intervals between successive injections were also increased. SERIES A. Six dogs were used in this series. Caffein was administered by mouth for periods of six days to five weeks. It was given daily or at intervals of two, and sometimes of three, days. In a few instances the drug was withheld for four or even for seven days, and its administration was resumed at the end of this time. The initial dose in these experiments varied approximately between 40 and 140 mg per kilo. The doses were then increased gradually, and thus the maximum resistance of the subject to caffein was tested. The diet consisted either exclusively of meat or largely of carbohydrates with a minimum amount of meat to give flavor to the food. _Dog 11. Female._ Diet consisted of rice, 250 grams; cane sugar, 250 grams; meat, 50 grams; cracker meal, about 100 grams. Caffein was given by mouth daily or at intervals of one day, when the dose did not exceed 1.5 grams. Before the dose was increased to 2 grams, or approximately 0.213 gram caffein per kilo, an interval of two days was allowed. Symptoms were noticed the next day. An interval of two days was therefore allowed again at the end of which the same dose was repeated. It will be remarked that there were no symptoms this time, and the general condition of the dog seemed to be good. Two grams of caffein were, therefore, given daily during the next two days without any untoward effects; the dose was then increased to 2.5 grams. Even after this enormous quantity no symptoms were observed except slight tremors. When this dose was repeated 26 hours later, it proved fatal. No albumin or sugar was found in the urine, although the dog was fed on a very liberal carbohydrate diet. The following is a complete record of the experiment. April 20: Urine acid, no albumin, no sugar.