180 CARL DOWNEY LA RUE 



isolated by selection during six periods.' The difference between the means 

 is 0.51 0.109 spines or less than five times its probable error. 



During the six selection periods following the change in the method of 

 selection (periods 11 to 16) the divergence between the two groups did 

 not constantly increase but fluctuated from period to period. In the 

 eleventh period it was .52 spines, in the twelfth it fell to .49 spines, in the 

 thirteenth it rose to .84, the highest point reached. In the fourteenth 

 period the difference fell to .28 spines, in the fifteenth to .23 spines, which 

 is probably not significant, and in the sixteenth or last period it again 

 rose to .66 spines. If the divergence in period thirteen is due to selection, 

 it is difficult to see why further selection for the same number of periods 

 that caused this divergence, should cause a decrease in the difference of 

 the two groups. 



After selection was discontinued the two groups were kept under obser- 

 vation during five different periods which varied from eight to twenty- two 

 days in length. The mean spine number of the individuals in the low- 

 selected group was 5.19 while that of the high-selected group was 5.58. 

 The error of the difference between the means of the two lines is 0.178 

 spines while the difference is only 0.39 spines, or slightly more than twice its 

 probable error. The difference would not be considered significant if it 

 were not for the fact that the variations in the various lines, though great, 

 are insufficient to bridge the gap between oppositely selected lines. Is it 

 a fair criticism that Difflugia is insufficiently plastic to show rapid changes 

 from a certain condition of the shell after that condition is once attained? 

 The results suggest that the shell itself plays a part in determining the 

 character of the new individual formed by division of the old one. This shell 

 effect may be regarded as superposed upon the other factors determining 

 the form of the new shell. Such an effect might be expected upon physical 

 grounds. Naked protoplasm protruded from a large orifice is under dif- 

 ferent surface conditions than if the orifice were smaller; the extruded 

 protoplasm would be expected to assume a different form and size in the 

 two cases, and the newly secreted shell would therefore be different. In 

 other words, any accidental size modification might be maintained through 

 several generations, not through any protoplasmic modification, but 

 merely because the nature of the shell imposes certain physical conditions 

 upon the development of the new individuals. A modification having 

 once occurred in the shell through any cause, becomes comparable to an 

 environmental factor which is maintained for several successive genera- 

 tions. The nature of new individuals is largely determined by the environ- 

 ment, and of the environmental factors the shell is one of the chief. If this 



