Examination of the protocols shows that the absorption of caffein from the gastro-intestinal canal was quite rapid, symptoms having been observed as early as 20 minutes after its introduction. The duration of life, it will be remarked, varied with the size of the dose. When approximately 3 decigrams per kilo were fed, all the animals died in the night. They lived, therefore, less than 18 hours. Two decigrams per kilo were likewise fatal, but the duration of life was longer. To decide whether or not this is the smallest fatal dose, smaller amounts were fed. It seemed at first that about 150 mg per kilo was the smallest toxic dose, and about 200 mg per kilo the minimum fatal dose. Macroscopic examination of the organs, however, threw some doubt on this supposition, for well-marked lesions were noticed in all of the guinea pigs which received 0.2 gram per kilo. It is quite possible, therefore, that the minimum fatal dose may be somewhat higher, as we have reason to believe that, at least in some pathologic conditions, the susceptibility to caffein is increased. The presence of fatty changes in the liver of No. 128 and the rapid death in this case lends especial support to this view. Hence, the minimum fatal dose is probably greater than 0.2 gram per kilo for the normal guinea pig. The doses employed for the tests on guinea pigs Nos. 129, 130, and 131 may be considered therefore the minimum fatal dose for these animals. It will be also remarked that macroscopical examination of the organs of these animals failed to reveal the presence of severe lesions. That the minimum toxic dose is probably much smaller than 0.28 gram per kilo is indicated by the experiments on guinea pigs Nos. 135 and 134, in which 0.15 gram caffein per kilo induced mild symptoms in from two to three hours. Both of these, however, and also No. 137 died four to six days after the drug was fed. As already pointed out, caffein may be a factor in the delayed death of guinea pigs which received moderate doses of it. That this supposition may also be true for guinea pigs Nos. 134, 135, and 137 is indeed made probable by the observation that after moderate amounts of caffein symptoms may persist in the guinea pig for about 24 hours, and also by the fact that the secretion of urine in these animals was very scanty, as shown in the preceding record of the experiments; this means slow elimination of caffein and its products of decomposition. It is conceivable that the presence of toxic amounts of caffein in the body for a considerable length of time would induce changes that ultimately lead to the death of the animal or that morbid processes are set up by the combined action of caffein and some preexisting poison. Since some guinea pigs, however, survived the doses indicated, it is more probable that such changes would be brought about by caffein in the presence of a preexisting poison. The death of these pigs, and also of No. 137 several days later, is difficult to