of fluorescin is much faster when the intramuscular path is used than when given subcutaneously, but the writers state that the rate falls far behind that of the intravenous administration. The difference in toxicity we observed between feeding by mouth and subcutaneous injection, although distinct, was not very great. It was much less than Maurel(55) obtained with the hydrobromid of caffein in the rabbit. Whether this difference between his results and ours is due to the use of the pure alkaloid in our experiments and the hydrobromid employed by Maurel can not be stated at present with any degree of accuracy. It is hoped that the work in progress in the laboratory will throw some light on the subject in the near future. But Maurel's(56) experiments show that various substances behave differently in this regard. Thus the toxicity of strychnin, he states, is three times as great when given subcutaneously as when given by mouth and six times that of the minimum fatal dose by vein. It may be remarked, however, that examination of his data shows that his doses are much too large for the rabbit. In experiments with other drugs little or no difference between the two modes of administration was noticed. Thus, digitalin was but slightly more active when given subcutaneously than by mouth, while the toxicity of emetin hydrochlorid was just the same, whichever one of these methods of introducing the substance was used. Differences in the toxicity of substances have also been observed between subcutaneous and intravenous modes of administration, but here, too, the differences for various substances were unequal. EXPERIMENTS ON GUINEA PIGS. The toxicity of caffein was studied in a large number of individuals. The experiments were conducted on full-grown animals and were carried out at different seasons of the year in a variety of ways. Special attention was given to diet as a possible factor influencing resistance to caffein, and the effect of different modes of administration on toxicity. Some animals were therefore fed oats, some carrots, others received both hay and oats. Caffein was introduced subcutaneously, intraperitoneally, and by mouth. SUBCUTANEOUS INJECTION. SERIES A. Preliminary experiments carried out on three guinea pigs, which received 360, 300, and 290 mg of caffein per kilo subcutaneously have shown that such doses were rapidly fatal. Two of the animals were seized with convulsions half an hour after the introduction of caffein and died during the attack. The other had tetanus two minutes after the injection of caffein. Repeated attacks followed, which terminated in the death of the animal two and a half hours later. The fatal and toxic doses must therefore be considerably under 0.3 gram of caffein per kilo when introduced by this path and smaller doses were therefore injected. The results are shown in the experiments of the next series. SERIES B. Experiments with 2 decigrams per kilo constituted this series.