550 BIOLOGY OF PNEUMOCOCCUS 



fining process, in filtration for sterilization, and in obtaining lots 

 consistently free from chill-producing elements. 



Dialysis : The removal of crystalloids by the dialysis of serum in 

 water was studied by Avery 32 as a means of concentrating anti- 

 pneumococcic serum. The procedure resulted in an incomplete pre- 

 cipitation of antibody. Dialysis, however, may be utilized to ad- 

 vantage at some stage in processing serum by salting-out methods. 

 Banzhaf 70 found that serum precipitated by sodium chloride or 

 ammonium sulfate contained about 90 per cent of the immune 

 bodies following dialysis. Later Banzhaf and Klein 72 recommended 

 dialyzing serum until practically free from salts as the initial step 

 in eliminating inert protein and chill-producing substances. Good- 

 ner 528 reported that it was unnecessary in dialysis to remove elec- 

 trolytes completely in order to obtain a precipitate containing the 

 majority of the agglutinins. The method of dialysis has certain 

 advantages not possessed by other processes of concentration. It is 

 inexpensive and frequently results in a volume concentration not 

 readily obtained by other procedures, and it entails less harsh 

 phvsical treatment of the antibody than the addition of chemicals. 

 Some unpublished data, however, show that solutions of antibody 

 made by the method are not stable. 



Dilution with Water: Felton (1924) 396 observed that a heavy 

 precipitate formed when antipneumococcic serum was diluted in a 

 proportion of 1 to 10 with water, and believed that the amount of 

 water-insoluble precipitate so obtained gave an indication of the 

 protective strength of the serum. A higher yield of precipitate was 

 recovered when small amounts of tartaric, citric, oxalic, or acetic 

 acids were added to the water. Experiments on the influence of 

 hydrogen ion concentration on the precipitation of immune sub- 

 stances from specific serum suggested that the isoelectric zone of 

 the protective antibody lies between pH 6.6 and 7.5, possibly at 

 about pH 6.8. 



In later papers, Felton 400 and Felton and Bailey 421 reported ad- 

 ditional experiments dealing with the water precipitation of pneu- 



