AN ANALYSIS OF THE EFFECT OF SELECTION. 



27 



both the speck and the not-speck offspring, and in identical propor- 

 tions. Therefore this factor should not influence the end-result. 



Table 20 shows the results obtained from such experiment The 

 arrangement is the same as in table 19. 



As was expected, the differences here are less than in the correspond- 

 ing tests, but are still present and in the same direction when significant . 

 This result proves that one or more of the second-chromosome modi- 

 fiers cross over from speck in the female. 



Table 20. — 9 Tests, Chromosome II. 



THIRD-CHROMOSOME MODIFIER. 



If we cross two races, one of which is Dichset rough, the other wild- 

 type, the Fi female will have the constitution D' r . If such a female 

 be mated to a not-Dichffit rough male, there will be two types of 

 Dichset offspring— the non-crossovers will be rough, the crossovers 

 not rough. 1 If the two original chromosomes differed in modifying 

 factors somewhere near rough, these two types of offspring will differ 

 in their bristle number. 



Such tests have been carried out, with the results shown in table 

 21. 2 In only one case (the third) was a significant difference obtained; 

 but that case proves that there was a dominant plus modifier in the 

 1002 line, located somewhere near rough, or a dominant minus modifier 

 in the speck stock, but not in the 1331 line and in the same region. 

 Since the 1331 line was derived from a cross involving the speck Btock, 

 and had been minus selected ever since that cross, it is probable that 

 dominant minus modifiers present in the speck stock would have been 

 preserved in the 1331 line. It is, therefore, almost certain that the 



'There will be some double crossovers, but these will be rare. There will, be 0OUTM, also be 

 two classes of not-Dichsets. 



2 In the case recorded in the second row, the Dichyet and rough ra.no from different parent*, 

 so that the non-crossover and crossover classes are reversed. The experiment is the HUM id 

 principle as that outlined above. 



