The results of the experiments on rabbits show considerable variation in the toxicity of the single dose. Individuals differed so widely in their resistance to this drug that the same experiments had to be repeated many times with each method of administration before satisfactory conclusions could be drawn. This is strikingly illustrated in the experiments by intravenous injection in which a dose of nearly 0.2 gram per kilo was not fatal. Similar instances of exceptional resistance or of sensitiveness to caffein were observed when it was given in other ways. A comparison of the toxicity of caffein administered by different methods in this investigation shows well-marked differences in its activity, although they are not quite so striking as similar experiments with other alkaloids reported by several observers. The toxicity of caffein in these experiments on the rabbit indicates that it is greatest when given by vein and least when given by mouth. The ratio of the minimum toxic doses by these two methods of introduction of caffein was about 7.1; the relation of the minimum fatal dose was about 3.1. The toxicity when given subcutaneously is about 15 to 20 per cent greater than when given by mouth. The difference between the intramuscular and subcutaneous injection is even more marked. The toxicity of caffein when injected into the muscles is about midway between that administered by the subcutaneous and intraperitoneal routes, and is about half that injected intravenously. Meltzer and Auer,(58) who experimented with a number of drugs found that the intramuscular method of administration is as effective as the intravenous, fluorescin forming the only exception according to their observations. In the experiments of Sollman and Brown(81) with ergot, the effect was quite different from those obtained by Meltzer and Auer(58) with the drugs they used. It is quite possible that the result obtained with ergot is merely illustrative of a difference in the behavior of various substances in this regard. This appears probable on account of the difference in the rate of absorption for various substances. Thus, according to Achard, Gaillard, and Ribot (Compt. rend. Soc. biol., 1907, _62_: 90), absorption from the peritoneal cavity varies with the concentration of the solution and the size of the molecule. The smaller the molecule and the greater the concentration the more rapid the absorption. That the rate of absorption from the intramuscular tissues is unequal and varies for different substances appears from the experiments of Meltzer and Auer.(58) The difference was very striking between intramuscular and subcutaneous administration of curara or adrenalin; the results were somewhat different with morphin and with fluorescin. As shown in their protocols, the onset of the symptoms after the intramuscular injection of morphin was sooner than after subcutaneous injection, but in time the difference diminishes and disappears altogether. The absorption