PNEUMOCOCCUS INFECTIONS IN MAN 463 



concentration of 1 per cent and 5 e.c. of defibrinated rabbit's blood. 

 These substances, sterile, must be mixed in tubes containing about 

 5 c.c. each and must not be resterilized after mixture. 



The sputum is very carefully washed in sterile salt solution. 

 We prefer to do the washing for this particular technique in succes- 

 sive test tubes of sterile broth. After about three washings, the 

 sputum is gently ground in the last broth tube with a glass rod, 

 and about 0.5 to 1 c.c. of this inoculated into tubes of the medium 

 given above. Growth is allowed to take place for from five to eight 

 hours, and when pneumococci have appeared in large numbers in 

 the supernatant fluid, the red cells are thrown down by centrifuga- 

 tion, and agglutinations done on the supernatant fluid as above. 



In all such typings it must not be forgotten that agglutination 

 reactions must be accompanied by a bile test on the materials used in 

 order to make sure that the organism is a pneumococcus. 



Typing' by Protection Experiment. The protection experiment 

 is not very practical for clinical use where speed is required. It 

 is, however, important that the technique of specific mouse protec- 

 tion should be well known to the bacteriologist. This technique will 

 be given in greater detail in connection with the determination of 

 the potency of immune serum in a subsequent section. 



Typing by Precipitins. In broth cultures the pneumococcus 

 gives up an antigenic substance which is soluble to the broth which 

 has been made use of by Blake 70 for typing purposes. The method 

 is useful when cultures or peritoneal exudates from mice are too 

 heavily contaminated for clear agglutination tests. The fluid to be 

 examined, either culture or peritoneal exudate, is centrifuged at 

 very high speed to throw down all the organisms, and the clear 

 supernatant fluid taken off for precipitation experiments. It is then 

 mixed in quantities of 0.5 c.c. with equal quantities of serum specific 

 for the three main types of pneumococci. Type I serum is used in a 

 1-10 dilution, Type II, undiluted, 1-10, and Type III, 1-5. In this 

 connection Dochez and Avery 71 have made an interesting observation 

 which is of great theoretical as well as practical importance. They 

 have shown that a soluble antigenic substance which gives specific 

 precipitin reactions with anti-pneumococcus serum, appears in the 

 urine of over 60 per cent of pneumonia cases. The substance may 



70 Klake, Jour. Exper. Med., 26, 1917, 67. 



71 Inches and Avery, Jour. Exper. Med., 26, 1917, 477. 



