1. Small quantities of antiserum bring about instantaneous agglutination of pneumococci in the circulation of the rabbit; the reaction is specific and occurs in every case in which sufficient serum is given to influence the course of the septicemia or to prolong the life of the animal. 2. The agglutinating titer of antipneurnococcus serum can be made considerably higher by adding only a small quantity of culture to the tests, thus making the test a finer differential. 3. Typhoid bacilli agglutinate spontaneously in the circulation of the normal rabbit; the reaction is positive in vivo even in cases in which undiluted serum gives a negative result in vitro; heating the bacilli to 80°C. for thirty minutes renders them more agglutinable in vivo. 4. Dysentery bacilli of the Shiga type do not agglutinate in the blood stream of the normal rabbit, but a small quantity of antiserum injected into the circulation causes immediate agglutination; while all strains of the Flexner group undergo spontaneous agglutination. 5. Non-virulent influenza bacilli agglutinate spontaneously in the circulation of the normal rabbit; virulent strains remain in the blood unclumped. 6. In all instances so far investigated of both passive and natural immunity, agglutination of the bacteria within the blood of the infected animal was followed by a rapid removal of the bacteria from the circulation, and by phagocytosis and destruction of the agglutinated bacteria in the capillary systems of the viscera; while those bacteria which are not agglutinated remain in the circulation and produce a progressive septicemia. 7. Hence the agglutinins seem to play the decisive part in at least certain instances of bacterial infections.