308 H. S. JENNINGS 



isolated certain split pairs, in which conjugation had not been 

 consummated; one of these gives the long hne forming the branch 

 A of our diagram. 



This branch A was then propagated without conjugation, 

 from some period before March 4, 1910, while the other two lines 

 B and C conjugated repeatedly; the experiments consisted mainly 

 in comparison of the progeny from these conjugations with the 

 progeny of the non-conjugating branch A. Finally, on June 4, 

 a part of A was allowed to conjugate, the rest not, and these 

 two parts compared. Again, on August 10, a part was allowed 

 to conjugate, and compared with the part that had propagated 

 from the beginning of the experiment without conjugation. 

 In the following an account is given of the results of these various 

 comparisons. 



Experiment 5: January 29 to February Jf., 1910: Paramecium 



aurelia 



Experiment 5 was a brief one, dealing with 10 pairs (20 indi- 

 viduals) and 10 split pairs (20 individuals), belonging to the 

 branch B of race k (fig. 1). These 40 hues were cultivated side 

 by side under the same conditions for six days. Of the descend- 

 ants of the pairs, three were accidentally lost during the course 

 of the experiment, so that we have finally but 17 lines descended 

 from conjugants; 20 descended from non-conjugants. 



Of the 17 Unes of conjugants, 8 died out during the experi- 

 ments, a mortality of 47.06 per cent. Of the 20 Unes of non- 

 conjugants, 9 died out, a mortality of 45 per cent. 



The number of fissions for the six days of the experiment is 

 given for the surviving conjugant and non-conjugant lines in 

 table 11, while the results are summarized in table 12. 



Thus, as the tables show, this experiment, so far as it goes, 

 illustrates the usual conditions: 



1. Those that have' conjugated multiply less rapidly than those 

 that have not. 



2. The rate of fission is much more variable among those that 

 have conjugated — both the standard deviation and the coeffi- 



