ARTICLES 



609 



specimens produced by fission by a single parent differed in 

 their rate of division as indicated in Fig. 1. ' 



Specimens were isolated on hollow-ground slides and 

 carefully maintained under uniform conditions. An attempt 

 was made to obtain a " high line," in which the fission rate 

 was rapid, by selecting generation after generation the offspring 

 of specimens that divided most quickly; and a " low line," 

 in which the fission rate was slow, by selecting for further 

 reproduction the offspring of specimens that divided slowly. 

 Selections were thus carried on for 1 30 days, during which the 

 differences in the fission rate between the two lines gradually 

 increased until it averaged 21-2 per cent. During this period 



Fig. i. 



the number of generations produced in the high line averaged 

 183, and in the low line 122 (Fig. 2). The next problem was 

 whether these high and low lines were heritably different, or 

 would undergo complete reversion, as had been shown to occur 

 in other organisms by many investigators. Accordingly, 

 some of the members of the high line and some of the low line 

 were, from time to time, cultivated without selection. The 

 difference in the two lines persisted as long as they were main- 

 tained without selection (102 days), and the diversity in 

 fission rate thus seemed to be inherited and to have been built 

 up by means of small heritable variations. At the conclusion 

 of the series of experiments the high and low lines were again 

 subjected to selection, but in the reverse direction ; slowly- 



1 For an explanation of these figures see the end of the article. 



