Inheritance of Size 69 



obviously increased progressively with the genera- 

 tions. This was shown in particular by the close cor- 

 relation for size between parents and offspring, which 

 among ten families (Jennings's table 32) averaged 

 -j-0.37, and in one very large family was +0.67. Since 

 families gradually became diversified, it was possible 

 to select very large and very small individuals. Since 

 their progeny each tended to resemble the parents, the 

 diversity endured for some time. Selection, therefore, 

 increased the diversity. But it was apparent that the 

 progeny always varied back toward the mean. Un- 

 fortunately the permanence of the selected diversities 

 of size was not tested throughout many generations 

 after selection had ceased. 



It is possibly of significance that selection was able 

 to increase sizes to very great extents, while selection 

 was effective in decreasing the sizes only to small ex- 

 tents. The largest organisms showed a very high mor- 

 tality; a fact which constitutes a distinct limitation to 

 diversity of biotype. 



In several instances considerable variations of size 

 appeared suddenly, and were inherited for at least five 

 generations of Difflugia. But in the light of Jollos's 

 ('24) observations on Arcella this is not necessarily an 

 indication of genotypic change. 



Centropyxis. Another species in which the inheri- 

 tance of size has been studied is Centropyxis aculeata, 

 by Root ('18). Again clones proved to show perman- 

 ent differences in shell size, but data upon variability 

 within populations or clones were not published. In 

 two families there were distinct correlations between 

 shell sizes in parents and in progeny (of -f- 0.253 and 

 -f0.088) ; but these disappeared in the second and fur- 

 ther generations. In this species therefore the size 

 of the race was more important than the size of the 

 immediate parent; in Difflugia and Arcella the size 



