VII. POTENCY TESTS 657 



number of mice to prove death due to uncontaminated Pneumococcus 

 infection. Streaking a sample of the heart's blood from a sterile capil- 

 lary pipette on a blood-agar plate is adequate. 



d. Interpretation of the test. In interpreting the test, no attempt is 

 made to identify an end-point dilution, because such a procedure is sta- 

 tistically unsound inasmuch as it considers the fate of too small a group 

 (10 mice). As the same number of mice is used for each serum and as 

 the doses used for each serum are in the same ratio to each other, it is 

 permissible to take into account the whole number (30) injected with 

 each serum. The method of computation of potency values is, therefore, 

 based on the equation : 



x / a (s — s')\ 



Units per cc. in unknown serum = — a -I } 



v \ n / 



a (s — s')^ 



y 



x = largest amount of control serum injected 

 y = largest amount of unknown serum injected 

 a = units per cc. in control serum 

 s = number of mice surviving on unknown serum 

 s' = number of mice surviving on control serum, and 



n = total number of mice injected with all dilutions of any one serum (see 

 above). 



If s is less than s', then the part of the formula within the inner pa- 

 rentheses has a minus value and is subtracted from a, not added to it. 

 The value of n will be 30, 60, 90, etc., depending upon the number of 

 tests from which calculations are made. 



The form of protocol used for reporting to the National Institute of 

 Health does not show the value for x and y. The formula can be applied 

 to these protocols if the following definitions are used for these values: 



x = the smallest number of units tested for with unknown serum, and 

 y = smallest number of units tested for with control serum. 



The application of the formula to the interpretation of the CP 77 B 

 Types I and II protocols (pages 659 and 660 respectively) summarized 

 on the attached National Institute of Health protocol (page 661), using 

 the second definitions for x and y, is: x = 1500, y = 150, a = 300, 

 s = 14, s' = 15, n = 30; so 



x/ a(s-8')\ 1500/ 300(14-15)\ 



- ( a + — J = 1 300 + I = 2900 Type I units per cc, and 



750/ 150 (15- 9) \ 

 = — ( 150 + J 1 = 1800 Type II units per cc. 



