114 The Mechanism of Evolution in Leptinotaesa 



the lower or diversa mode produces an F5 zygote that is heterozygous, but 

 appears as signaticollis in type. 



The Ac values of all the lines, and indeed after F^ of the original cross, are 

 Ac^°, all trace of the value J-C*" having gone from the lines. When the cross is 

 made between the test signaticollis at the value of Ac^^ or Ac*°, under the gen- 

 eral conditions of the experiment as far as temperature and relative humidity 

 are concerned, there always appear, following the test, diversa types in the Fg 

 fraternities. The numbers present are not constant, and it has been found that 

 the highest number appears when there is little or no wind-movement, hence 

 low evaporation ; that the least frequent comes with wind-movement producing 

 rapid evaporation. Plate 11 shows only the fact that the diversa type comes out 

 as the result of the test reaction. 



In plate 12 is shown the full array that comes in the test reaction when made 

 between pure signaticollis with J.c*° and the tested form from the lower mode 

 with Ac^°, under the conditions given. The full F2 array shows the following 

 types : 



{a) diversa, a type in all respects diversa that breeds true. 

 (&) mid-type a, that is, in all respects a heterozygote in appearance and in 

 reaction between the two species. 



(c) mid-type h, of ligliter appearance than the preceding, heterozygous, and 



giving in F2 : 1 diversa, 3 si^na/icoZHs-appearing types within close 

 limits to the expected ratios. 



(d) mid-type c, different from the other two, and in Fg giving 1 diversa, 2 mid- 



type c, 1 new signaticollis type, also in close realizations to the ex- 

 pected frequency. 



(e) signaticollis type, that is, the extracted lower mode race, breeding true. 

 (/) signaticollis type, in appearance pure signaticollis, in reality heterozygous, 



giving in Fj all signaticollis, but arrayed in trimodal polygons, the 



modes being on the index-modes of the parent species. 

 (g) signaticollis type, that is, pure, true-breeding signaticollis. 

 (h) signaticollis type, that is, heterozygous, for a new arrangement that comes 



out in the test. 

 (i) signaticollis type, that appears as signaticollis, but is really heterozygous 



for the two arrangements Pu-V and Elp-V. 

 (;') signaticollis type, that breeds true, is homozygous, for the association of the 



V determiner with the Elp factor producing general pigmentation in 



the elytron, and by inbreeding almost black individuals can be 



obtained. 



As far as experience goes, the F^ test hybrid produces two main classes of 

 gametes, 50 per cent that are signaticollis and 50 per cent diversa. The signa- 

 ticollis gametes in all tests are apparently pure and unaltered, but the diversa 

 gametes show with changing conditions recombinations between the three 

 essential agents involved. The condition shown in plate 12 is that realized most 

 frequently under the conditions of experiment, with three classes of diversa 

 gametes present. The relative number of these is apparently entirely the 

 combined product of the external conditions, the Ac values in the two stocks 

 interacting in the gametogenesis or at some other time in the life of F^. 



This in general is the result of the test reaction, revealing as it does the 

 existence of the V determiner and its restoration in some of the gametes to its 

 normal relations, resulting in the recovery of the lost diversa type. Of interest 



