113 The Mechanism of Evolution in Leptinotaksa 



agents shows them to have. At present they are but graphic representations of 

 the associations and interaction of these gametic agents in the production of 

 specific end-results. 



The results shown in plate 9, while suggestive of the crossing of one hetero- 

 zygous and one homozygous parent, are in reality not that at all, but are dif- 

 ferent, as shown by the test reaction applied to the F^ fraternities of any of 

 these arrays, which uniformly produces in the following F, fraternity, after 

 the test, pure-breeding diversa types along with several others, showing that 

 the pure-breeding F^ signaticollis race has in reality in it the Y determiner, 

 which could not possibly have come from any other source, and that the F^ 

 signaticollis type, while it bred true in aspect, is in reality a masked hetero- 

 zygote with the V determiner inactive, as far as the production of stripes is 

 concerned. This is further made certain if an F2 fraternity of the questionable 

 signaticollis type be separated into three groups by measurements of the form- 

 index and then applying the test to the modal classes of each group. 



The test, when it is applied to the portion of the trimodal F, polygon having 

 the highest or signaticollis index, gives in all following fraternities nothing but 

 signaticollis, regardless of what is done to it. The lowest mode, having the 

 diversa index, always gives in F2 pure-breeding extractive types that are diversa, 

 while the middle mode also gives in F, following the test diversa types and 

 others of complex nature. Everything shows clearly that the race of F^ ex- 

 tracted signaticollis is a masked heterozygous strain, a " fixed hybrid," in which 

 the agents that are commonly employed to differentiate the species are inactive. 

 However, biometric testing uniformly shows that Pj in any such race is tri- 

 modal, the extreme modes being that of the parent species form, proving clearly 

 that the F^ is really heterozygous. The further fact that the race from the 

 lower mode, pure-breeding, and in appearance signaticollis, will give the diversa 

 type as pure diversa following the test reaction; that the other race derived 

 from the higher mode can not be made to do so, shows that V is constantly in 

 association with the complex that segregates to produce the lower mode, and is 

 not present in the upper modal group of the F, array. 



The further observation that in the race derived from the lower mode of the 

 F2 array, while it is constantly signaticollis in its aspect, shows the development 

 of pattern in connection with the punctation system in a way not known any- 

 where else in all of my materials or in nature, led to the hypothesis that Y had 

 changed in its relations in the gametic system, such that it was entirely asso- 

 ciated with the punctation-determiner, the two acting as a single group. This 

 was shown by the ability to develop, out of the race (plate 8, fig. 11) derived 

 from the lower mode, a derivitive race that showed in all of the members of the 

 fraternities irregular pigment-pattern developing in association with the punc- 

 tation system where none was known before, and the ability to transfer this 

 condition entire to other series of combinations as in crossing with L. unde- 

 cimlineata or L. panamensis. From any such race, either originally or sec- 

 ondarily placed by crossing, the test reaction uniformly gave in the Fj following 

 the test the pure-breeding extracted diversa type. 



All of these results show that the real explanation of the apparent inhar- 

 monious array shown in plate 9 is in reality due to the forming of new arrange- 

 ments in the gametic agents. This change consists, as far as determined, in 



