OF MUTANT CHARACTERS. 



157 



recessive. When none was found the experiment was discontinued, 

 since it was thought that the chances of finding a double were as good 

 in F2 as in any subsequent generation — an attitude excusable at I kit 

 time, before the fact of no-crossing-over in the male had yet been 

 discovered. 



Table 19. — Pi, blistered 9 X pink cf; Fi wild-type 9 and cT. 



The tests on the linkage of blistered were taken up again after the 

 discovery of no-crossing- over in the male had made it most probable that 

 the absence of the double recessive curved blistered was due to the 

 genes being in the same chromosome rather than to a masking effect of 

 curved. At this time (July 1913) it was important that any autosomal 

 mutant be tested as to its linkage with both the second and the third 

 chromosomes, and, indeed, it was by the uniform results of many such 

 tests that convincing proof was secured that the second and third 



