On the other hand, the admirable researches of Kyes and Sachs, and subsequently those of Morgenroth, pursued under the direction of Ehrlich at the Laboratory of Experimental Therapy at Frankfort, have proved the readiness of venom to enter into chemical combination with certain elements of normal serums, in particular with _lecithin_, a combination which results in the formation of _hæmolysing_ and non-toxic _lecithides_, the _neurotoxin_ being left free. It therefore seemed impossible to deny the existence of a chemical reaction between the venom and the serum, which was until quite recently considered as proved. We shall see presently that this is not the case. But let us first endeavour to determine the laws that govern the neutralisation of variable quantities of venom by antivenomous serum. If, in a series of test-tubes, we bring the same quantity of cobra-venom (_e.g._, 0·00005 gramme, a dose which is invariably lethal to the mouse in two hours) into contact with progressively increasing quantities of an antivenomous serum (_e.g._, 0·01 c.c., 0·02 c.c., &c., up to 0·1 c.c.), and, after thirty minutes of contact, inject these different mixtures subcutaneously into a series of mice, we find that all those that have received the mixtures containing less than 0·05 c.c. of serum succumb after variable intervals, while all the rest survive. It is evident that, under these conditions, the serum experimented upon has shown itself capable of neutralising _in vitro_, in a dose of 0·05 c.c., 5 centimilligrammes of venom. The same serum should therefore neutralise 1 milligramme of venom in a dose of 1 c.c., that is to say, that this mixture injected into a mouse ought to be entirely innocuous. Experiments show, however, that in reality it is necessary to mix 1·2 c.c. of serum with 1 milligramme of venom in order that the inoculated mouse may not succumb. This proves that, in the initial mixture of 0·00005 gramme of venom + 0·05 c.c. of serum, there remained an exceedingly small quantity of non-neutralised venom, and that this quantity of venom in a free state was insufficient to cause the death of the animal, or even any apparent malaise. When multiplied by twenty, however, it becomes capable of producing toxic effects; it is for this reason that, when it is desired to inoculate a mouse with twenty times the lethal dose of 0·00005 gramme _neutralised_, it is necessary to mix with this twenty times lethal dose a dose of serum _a little larger_ than twenty times that which renders 0·00005 gramme of venom innocuous to the mouse, that is to say, 1·2 c.c.