304 WILLIAM LAWRENCE TOWER. 



pretatlon would perhaps be perfectly valid were it not for the 

 fact that identical materials placed under dissimilarc onditions, 

 as in Exp. No. H 410, give such a result that one would not be 

 warranted in assuming that either of the parents was hetero- 

 zygous, but that both parents were homozygous in the strictest 

 sense. Even this might possibly be accounted for by assuming 

 that chance had brought about the selection of homozygous 

 individuals for one series of experiments and heterozygous in- 

 dividuals for the other. This interpretation, however, is invali- 

 dated by Exp. No. H 409/411, in which a male and a female 

 of material identical with that used in Exp. No. 409 and Exp. 

 No. 410 were allowed to reproduce under the conditions of Exp. 

 No. 409 for part of their reproductive period, and under the 

 conditions of Exp. No. 410 for the remainder. From the eggs 

 developed and fertilized during the first period came results in 

 every way comparable to those obtained in Exp. 409, and from 

 the eggs developed and fertilized during the second period came 

 results in every way the duplicate of those obtained in Exp. No. 

 410. 



These experiments are crucial as far as this case is concerned, 

 for the simple reason that it is inconceivable that during one 

 period all the heterozygous germ cells would be developed, and 

 during another period all the homozygous cells. Any such as- 

 sumption is preposterous, and the only interpretation that one 

 can legitimately use is that conditions surrounding and incident 

 upon the germ cells at the time of and immediately following 

 fertilization were in some way productive of the differences in 

 behavior found. 



In the series of experiments between L. signaticollis and 

 L. undecimlineata, a more complicated behavior is found, due 

 to the fact that more characters are involved. As in the 

 first series, similar materials reared under dissimilar condi- 

 tions gave dissimilar behaviors, but without the production of 

 any new types or characters in the materials; that is, there are 

 apparently no strange attributes present, but simply a varia- 

 bility in the behavior of attributes already prssent, and this 

 variability in behavior may be of profound importance in evolu- 

 tionary processes. 



