GENERAL TECHNIC 437 



until coagulation has occurred and the serum has separated; then place 

 the whole in the ice-chest until needed. Blood may be collected in 

 centrifuge tubes and allowed to coagulate; it is then broken up with a 

 glass rod, centrifuged, and the serum secured at once; such complement, 

 however, is likely to be unduly hypersensitive to the anticomplement 

 action of organic extract and of mixtures of extract with normal serums. 

 As a general rule, therefore, it is good practice to bleed the animal late in 

 the afternoon preceding the day on which the experiment is to be made, or 

 at least some hours before the regular work of the day begins. The serum 

 should be clear and contain no corpuscles. 



Preservation of Complement. Various methods have been advo- 

 cated from time to time for the preservation of complement serum. Se- 

 rum kept in a frozen state will preserve its complement properties over 

 a period of several weeks. In my own experience, complement serum is 

 best preserved with chemically pure sodium chlorid. The sera of 

 several guinea-pigs are secured and mixed, and 0.425 gram sodium 

 chlorid dissolved in each 10 c.c. This serum is best preserved in the 

 refrigerator, sealed in amounts of 1 c.c. in ampules; before use each 

 cubic centimeter is diluted with 19 c.c. of distilled water, making a 

 1 : 20 or 5 per cent, dilution ( = 0.05 c.c. undiluted serum) in 0.85 per 

 cent, salt solution. If a 1 : 10 dilution (= 0.1 c.c. undiluted serum) is 

 desired, 0.85 gram sodium chlorid should be added to each 10 c.c. 

 of serum and each cubic centimeter diluted with 9 c.c. of distilled 

 water. 



Complement serum preserved in this manner will maintain its 

 hemolytic and fixing properties for several weeks; loss in fixability or 

 delicacy in the complement-fixation test is usually apparent before an 

 appreciable deterioration in hemolytic activity. 



And now we come to a very important question, namely, the amount 

 of complement that is to be used in conducting complement-fixation tests. 

 Many present-day observers use exactly one unit of complement and 

 one unit of amboceptor. This is permissible, providing the complement 

 is titrated in the presence of a constant dose of antigen and a constant 

 dose of serum, in order that due allowance for the anticomplementary 

 action of these may be made in the titration. Under these circum- 

 stances, however, it is necessary to titrate each patient's serum with the 

 complement, because one serum or even the pooled serums of different 

 persons should not be taken as a standard in the titration, for two impor- 

 tant reasons: (1) The patient's serum which we are about to test may 



