EPIDEMIC INFLUENZA in 



tests. On the other hand, the serum specimens from 

 seventeen of the nineteen persons who had influenza 

 during the recurrence of 1922 reacted positively and 

 thus afforded presumptive eivdence that they had 

 recently been infected with Bacterium pneumosintes. 

 One of the persons chosen as a control subsequently had 

 influenza. It was interesting to find that his blood 

 serum, previously negative, reacted positively with 

 Bacterium pneumosintes when tested on the tenth and 

 eighty-ninth days after recovery. In other instances 

 demonstrable antibodies persisted in the blood for at 

 least five months following an attack of influenza. 



The second series of experiments that were made possi- 

 ble by the acquisition of new and pathogenic strains of 

 Bacterium pneumosintes deak with the immunizing 

 effects in rabbits of subcutaneous injections of appro- 

 priate doses of the heat-killed organisms. When a 

 number of rabbits had been prepared by three injections 

 of the killed bacteria the protective effects of the vac- 

 cination were demonstrated in two ways. By examina- 

 tion of the blood serum it was found that eleven among 

 fifteen vaccinated animals had developed specific 

 antibodies against Bacterium pneumosintes. Their re- 

 sistance was then tested to doses of the living organisms 

 which were pathogenic for normal, unvaccinated ani- 

 mals. In all but two instances the protection was com- 

 plete. Not only did the vaccinated rabbits fail to show 

 the characteristic signs of infection with Bacterium 

 pneumosintes but, with the two exceptions noted above, 

 they were normally resistant to secondary infection 

 with ordinary bacteria. Incidentally, it was observed 



