456 



LORANDE LOSS WOODRUFF. 



division on the average about three quarters of division per 

 day lower than in the cultures subjected to the fresh culture 

 medium, etc. This is without doubt due, as already discussed, 

 in large part at least to the accumulated excretion products in 

 the s series. But whatever the cause, the experiment affords 

 an opportunity, to study the effect of naturally changed condi- 

 tions, involving a lowered fission rate, on the periodicity of 

 rhythms and endomixis. 



Now since the s subcultures divided at a much lower rate 

 than the parent cultures, and since endomixis appeared fairly 

 synchronously in parent and 5 sets, it is obvious that endomixis 

 consistently appeared in the s subcultures within a smaller num- 

 ber of generations. In other words, the treatment of the 5 



series apparently resulted merely in reducing the number of cell 

 divisions in a given time and had practically no effect (except 

 in the first period) on the occurrence of endomixis. On the 

 basis of this set of cultures, then, endomixis is to a certain 

 extent independent of the number of generations and more 

 closely .related to a time factor, if such an expression may be 

 employed. 



The B pair of cultures affords a fairly typical example (Fig. 

 n). In B endomixis occurred at periods 59, 63 and 70; while 

 in Bs it occurred at periods 59, 64 and 70. Thus the length of 



