APPENDIX II 411 



estimated. It is important to remember that in striking these 

 averages only an approximate estimate can be obtained, and that 

 therefore the greater the number of tubes made the better, as the 

 error decreases with the greater number of leucocytes counted. In 

 the experiments which I am about to record I have counted about 

 500 leucocytes in each case by making five films from each of five 

 tubes, and counting about 20 leucocytes in each film. Since it is 

 obvious that the greatest error may occur when the number of 

 living approximates the number of dead cells in a tube, the follow- 

 ing experiments would appear not to be very erroneous, judging by 

 the application of Poisson's formula, which shows that supposing 

 there are half a million leucocytes in the five tubes, which is an 

 excessive estimate, a count of 500 cells would give a possible error 

 of not more than about 6 per cent, even when the numbers 

 approximate. 



Before enumerating the actual measurements there is yet 

 another question to be considered, a point upon which I wish 

 to lay great emphasis namely, that all measurements of the 

 lives of leucocytes should necessarily be comparative. For instance, 

 it would be fallacious to say that a typhoid plasma killed a person's 

 leucocytes more rapidly than a septicsemic patient's plasma, when 

 the typhoid measurement was made to-day and the septicsemic 

 measurement made three days ago; for even if there was a great 

 difference in the length of the lives and the same person's leucocytes 

 were used, one cannot say that that person's leucocytes were in the 

 same state to-day as they were three days ago, although the person 

 is apparently in the same healthy condition. 



Again, I have shown (4) that the factor Jieat in accelerating 

 the diffusion of substances into cells also materially affects the 

 lives of the leucocytes, since the cells are necessarily resting in a 

 citrate solution which is itself poisonous to some extent, and even 

 the temperature of incubators is variable. It is thus of the utmost 

 importance that when the lives of a person's leucocytes, which have 

 been placed in the plasma of a person suffering from an infective 

 disease, are measured, a simultaneous measurement of the same 

 leucocytes shed at the same time must be made in the plasma of 

 a healthy person. And it is only by the difference between the 



