VIII 



(The values for t == 99 and 125 have been omitted 

 as quite uncertain). 



By looking at the above calculations it is evident, 

 how satisfactory the concordance is, especially if the 

 unavoidable error of experiment as regards the deter- 

 mination of x is taken in consideration. Also t 

 might not always be exactly determined, the time for 

 the bleeding sometimes being allowed to vary within 

 6 hours. 



Considering this regularity in passive immunisation, 

 it was of interest to examine if the same regularity 

 should be found in the case of active immunisation. In 

 the following tables, which refer to the experiments men- 

 tioned in the preceding paper, the day, on which the 

 maximum of an agglutinin curve occurs, is taken as 0, 

 and the other values for t are counted from this point. 



- 9 



