110 



The Mechanism of Evolution in Leptinotaksa 



idea was supported by the ability to separate the F^ signaticollis race into the 

 three groups in Fj, having the modal conditions of form-index respectively of 

 the two parent species, and the known heterozygote, and the presence in the lower 

 portion of this trimodal polygon of the development of dark pigment in an 

 association not known in the original signaticollis materials. If the V deter- 

 miner * was present, but in new combinations in the gametic complex, it was at 

 least worth the effort to try to recover V from this race F^ of signaticollis in 

 appearance, but trimodal in form-indexes in Fj, and derive the diversa form 

 from this race as an extracted product by bringing about in some manner the 

 recombination in the gametes of the proper arrangement of gametic agents, 

 which were possibly present but in new arrangements. 



rt (^ ^ ^_ IT) lO vD lO l<_ r\ 03 CO <T>_ 05 O o_ -_ — tv; i\ o ro V. 

 w ru" (\r oi oi" N~ (\j fvj" c\i" f\j" cj" cy oj cm' CO ro co oi ro oo" co fO en 



oooooooooooooooooooo OOO 



m o ^ ■* iS ui ID CD r^. !<■. CO oq <J) <n o q -. - fy t^_ to n ■^^ 

 pj" (vT cu <\J' tu" rj rvj cj" (y oj (\l cm' oj tu (^ n' n co ro c'i' « co co 



Pure §tockof 

 L. diversa. 



Pure stock of 

 L. signaticollis 



Fig. 7. 



To test this point I used only strains of signaticollis, so that there might be 

 introduced no traces of the V determiner other than that already there. In 

 1908 and 1909 I made many crosses of this peculiar derivative from the tri- 

 modal signaticollis race (index mode 2.600-2.649), in which both had the Ac 

 determiner with a value of Ac'", with the uniform result that in no fraternity 

 did any trace of the V determiner producing the diversa striped elytra appear. 

 All of the progeny were signaticollis in aspect, but two series of crosses that 

 I had measured showed F^ heterozygous for form-index, monomodal, and as of 

 the usual F^ type, and trimodal conditions in Fg. In all 22 crosses were made, 

 with the constant results described, and from these it is certain that the S'igna- 

 ticollis pure species does not have any agent that matches the V, or otherwise 



^ Determiner for elypitcal stripe see tables 2 and 3, pp. 86, 87. 



