non-treated rabbit is mixed with the serum of the blood of the foreign species.

In 1897 Kraus discovered that if the filtrates from cultures of bacteria (e. g., typhoid bacillus) are mixed with the serum of an animal immunized with the same serum (e. g., typhoid serum) it causes a precipitate; and that this precipitin reaction is specific. This fact was confirmed and has been extended by the work of many authors.

Tchistowitch in 1899 observed that the serum of rabbits which had received injections of horse or eel serum caused a precipitate when mixed with the serum of these latter animals.

Bordet found in 1899 that if milk is injected into a rabbit the serum of such a rabbit acquires the power of precipitating casein, and Fish found that this reaction is specific inasmuch as the lactoserum from cow's milk can precipitate only the casein of cow's milk but not that of human or goat milk. Wassermann and Schütze reached the same result independently of each other.

Myers and later Uhlenhuth showed that if white of egg from a hen's egg is injected into a rabbit, precipitins for white of egg are found in the serum of the latter, and Uhlenhuth' found, by trying the white of egg of different species of birds, that the precipitin reaction

Uhlenhuth, P., and Steffenhagen, K., Kolle-Wassermann, Handb. d. pathol. Mikroorg., 2nd Ed., 1913, iii., 257.

called forth by the blood of the immunized animal is specific, inasmuch as the proteins from a hen's egg will call forth the formation of precipitins in the blood of the rabbit which will precipitate only the white of egg of the hen or of closely related birds.

To Nuttall' belongs the credit of having worked out a quantitative method for measuring the amount of precipitate formed, and in this way he made it possible to draw more valid conclusions concerning the degree of specificity of the precipitin reaction. He found by this method that when the immune serum is mixed with the serum or the protein solution used for the immunization a maximum precipitate is formed, but if it is mixed with the serum of related forms a quantitatively smaller precipitate is produced. In this way the degree of blood relationship could be ascertained. He thus was able to show that when the blood of one species, e. g., the human, was injected into the blood of a rabbit, after some time the serum of the rabbit was able to cause a precipitate not only with the serum of man, or chimpanzee, but also of some lower monkeys; with this difference, however, that the precipitate was much heavier when the immune serum was added to the serum of man. The method thus shows the existence of not an absolute but of a strong quantitative specificity of blood serum. This statement may

Nuttall, George H. F., Blood Immunity and Blood Relationship, Cambridge Univ. Press, 1904.

be illustrated by the following table from Nuttall. The antiserum used for the precipitin reaction was obtained by treating a rabbit with human blood serum. The forty-five bloods tested had been preserved for various lengths of time in the refrigerator with the addition of a small amount of chloroform.

TABLE II

QUANTITATIVE TESTS WITH ANTI-PRIMATE SERA
Tests with Antihuman Serum

Among the Primate bloods that of the Chimpanzee gave too high a figure, owing to the precipitum being flocculent and not settling well, for some reason which could not be determined. The figure given by the Ourang is somewhat too low, and the difference between Cynocephalus sphinx and Ateles is not as marked as might have been expected in view of the qualitative tests and the series following. The possibilities of error must be taken into account in judging of these figures; repeated tests should be made to obtain something like a constant. Other bloods than those of Primates give small reactions or no reactions at all. The high figures (10%) obtained with two Carnivore bloods can be explained by the fact that one gave a loose precipitum, and the other was a somewhat concentrated serum.1

We have mentioned that even the proteins of the egg are specific according to Uhlenhuth. Graham Smith, one of Nuttall's collaborators, applied the lat

Nuttall, Blood Immunity and Blood Relationship, pp. 319 and 320.

ter's quantitative method to this problem and confirmed the results of Nuttall. A few examples may serve as an illustration.

Frog, Amphiuma, and Dogfish sera, as well as Tortoise and Dogfish egg-albumins, were also tested, with negative results.