The precipitin in the antiserum is specific, and produces the effect only with its own antigen, not with that derived from another species of animal, but will act with one derived from a closely allied species, _e.g._ man and the higher apes, sheep, and goat. In order to eliminate this the antigen should be diluted to 1:1000. The blood used as antigen may be dried and sterilised, thus it may be kept for a long time, and when required for use dissolved in normal saline. The most convenient human antigen for general use is ascitic or pleuritic fluid, and may be kept in good condition when mixed with a small quantity of chloroform. To produce the antiserum 3 to 10 c.c. of the antigen serum is injected into a vein or the peritoneal cavity of a rabbit, and repeated at an interval of four or five days until 25 c.c. have been injected. The serum of the rabbit is then tested with the antigen diluted 1:100. The injections are then continued until 70 to 80 c.c. have been administered. The animal is then killed, the blood collected, and stood in a refrigerator for the serum to separate. This is then drawn into sterile pipettes, sealed, and kept in the cold. For purposes of preservation one-tenth its volume of 5 per cent. phenol may be added before pipetting. The antiserum may be dried on slips of black paper for future use. For medico-legal purposes, the antiserum should always be tested with its antigen to be sure of its efficacy. The antigen to be tested should be diluted 1:1000 at least, as potent antisera may give reactions with strong solutions of antigens derived from other species of animals. Various antisera may be made by inoculating rabbits with the serum of different animals. Metchnikoff has shown that intraperitoneal injection is not absolutely necessary in order to produce an antiserum, feeding a rabbit on the blood will act in the same way. The procedure for testing a stain is as follows. First of all the stain must be proved to be blood by the usual methods; this because the antiserum will give reactions with other albuminous substances, _e.g._ mucus, pus, semen, milk, or albuminous urine derived from the animal providing the antigen. Having proved the stain to be blood, a solution of it should be made in normal saline, sufficiently strong to give the HNO₃ reaction for albumin, or to foam when shaken up. If the amount of solution be small, the tests can be carried out in capillary tubes or pipettes. The solution of the stain must be cleared by filtration or the centrifuge. The tests must be controlled by comparison with known human blood, and blood from several domestic animals. For this purpose the various antisera should be kept in stock. Two sets, A and B, of six small tubes are used, into each of A is placed 0.05 c.c. of human antiserum from a sensitised rabbit. Into each tube of Set A is then added double the amount of diluted antigen, as follows: