10. When ready to give the third injection prepare a fresh blood agar subculture from another O.C. tube and after twenty-four hours incubation prepare a minimal lethal dose (as determined in 5) and inject it subcutaneously into the rabbit's abdominal wall. A slight local reaction will probably be observed as well as the weight and temperature reactions. 11. A week to ten days later inject a similar minimal lethal dose into the peritoneal cavity. 12. Observe the weight and temperature of the rabbit very carefully, and regulating the dates of inoculation by the animal's general condition, continue to inject living cultivations of the pneumococcus into the peritoneal cavity, gradually increasing the dose by multiples of ten. 13. At intervals of two months samples of blood may be collected from the posterior auricular vein and the serum tested for specific antibodies. 14. Under favourable conditions it will be found after some six months steady work that the rabbit may be injected intraperitoneally with an entire blood agar cultivation without any ill effects being apparent; and this characteristic--resistance to the lethal effects of large doses of the virus--is the sole criterion of _immunity_. Further, the serum separated from blood withdrawn from the animal about a week after an injection, if used in doses of .01 c.c., will protect a mouse against the lethal effects of at least ten minimal lethal doses of living pneumococci. In the foregoing illustration it has been assumed that complete acquired active immunity has been conferred upon the experimental rabbit in consequence of the formation of antibody, specific to the diplococcus pneumoniac, sufficient in amount to ensure the destruction of enormous doses of the living cocci--the _antigen_ (that is the substance injected in response to which _antibody_ has been elaborated) in this particular case being the bacterial protoplasm of the pneumococcus with its endo-toxins. But provided death does not immediately follow the injection of the antigen, specific antibody is always formed in greater or lesser amount; and in experimental work a sufficient amount of any required antibody can often be obtained without carrying the process of immunisation to its logical termination. For instance, if the immunisation of a rabbit toward Bacillus typhosus is commenced on the lines already set out it will often be found, after a few injections of "killed" cultivation that the blood serum of the animal (even when diluted with several hundred times its volume of normal saline) contains specific agglutinin for B. typhosus--and if the sole object of the experiment has been the preparation of agglutinin the inoculations may well be stopped at this point, although the animal is not yet immune in the strict meaning of the word.