When this blood is heated, after having been suitably diluted with three or four parts of distilled water, in order to prevent it from coagulating, it is found that a temperature of 70° C. maintained for fifteen minutes is sufficient to cause it to lose all toxic effect. The same applies to the blood of the other poisonous or non-poisonous snakes, and to that of the _Murænidæ_. Now, since the majority of venoms resist even prolonged heating at this temperature, it cannot be supposed that the toxicity of the blood is due to its containing venom derived from the internal secretion of the poison-glands, as was thought by Phisalix and Bertrand. On the contrary, it is probable that the toxicity results from the fact that the blood contains diastasic substances of cellular origin, which themselves represent certain of the constituent elements of venoms. These substances, moreover, possess some of the properties of venoms, as, for instance, the faculty of producing hæmorrhages and of being influenced by antivenomous serum, which causes them to lose a large portion of their toxic qualities. I have found that they can even be utilised to vaccinate animals against venom; by injecting weak, non-lethal, and repeated doses of dilute _Cobra_-blood into guinea-pigs and rabbits, I have succeeded in rendering them immune to doses of _Cobra_-venom several times greater than the lethal dose. There is no doubt that it is to these substances that the poisonous and non-poisonous snakes owe the partial immunity that they themselves enjoy with respect to venoms. We know, in fact, that common snakes suffer without danger many bites from vipers (Phisalix and Bertrand[78]), and that the _Cobra_ is relatively little affected by inoculation with its own venom or with that of other COLUBRIDÆ, such as _Bungarus_, or even of VIPERIDÆ, such as _Vipera russellii_. This immunity, however, is far from being absolute; I have killed common snakes (_Tropidonotus natrix_) with doses of _viper_-venom ten times greater than the lethal one for the rabbit, and _Lachesis lanceolatus_ (from Martinique) with 0·02 gramme of the venom of _Naja tripudians_. Phisalix,[79] on his part, has shown that, while it was necessary to inject from 100 to 200 milligrammes of _viper_-venom into other vipers or common snakes, beneath the skin or into the peritoneum, in order to cause death, the introduction of only 2 to 4 milligrammes of this venom into the brains of these reptiles was sufficient to kill them with the same symptoms of intoxication. This dose, however, is only twenty-five to thirty times greater than the lethal one for the guinea-pig. The practical lesson to be learnt from the establishment of the foregoing facts is that poisonous snakes of different species must never be placed in the same cage, for these animals sometimes bite each other, and may thus kill one another.