OF MUTANT CHARACTERS. 



179 



one, before one becomes " sensitized" to that particular mutation. 

 Some of our mutant characters had long been present in stocks or 

 experiments, so that many flies showing the character must have been 

 seen, before attention become sharply focused upon the differences 

 shown. Contamination and errors of one sort or another have also 

 added to the number of apparent reoccurrences of mutations. It is 

 therefore to be doubted if more then two of the apparent reoccur- 

 rences of purple were genuine remutations. 



REPETITION OF THE PURPLE VESTIGIAL BACK-CROSS TESTS. 



Because of the number of disturbing conditions that had been met 

 with in the first set of tests of the linkage of purple and vestigial, a 

 second and more extensive set was started. These second experi- 

 ments were carefully planned, and in the results obtained approach 

 present standards of uniformity and reliability. The viability of 

 vestigial was excellent, and the equality of contrary classes throughout 

 the experiments speaks for the favorable culture conditions. The new 

 experiments were conducted with a purple vestigial stock descended 

 from that used in the experiments of table 28, but cleared of mutations 

 and perhaps other disturbing factors by out-crossing to wild and by 

 selection, started among the F 2 progeny and maintained for several 

 generations until it seemed probable that the stock was clean. Also, 

 from the progeny of table 28 some purple (not-vestigial) cross-overs 

 were selected and from them was secured in a few generations a simple 

 purple stock free from vestigial and from the other mutant characters 

 known to be present. 



TABLE 30. F 2 offspring from cross of a purple male to a wild female. 



A preliminary test of the qualities of this purple stock was made by 

 out-crossing a male to a wild female and carefully examining all F 2 

 flies (table 30). The F 2 showed only purple (150) and wild- type 

 flies (300) as expected, but the ratio was 1 : 2 instead of 1 : 3. While 

 this deviation was significant (4.1 times the probable error), it indicated 

 a peculiarity of the wild parent rather than of the purple, and was not 

 further regarded. The vestigial stock used was that from which pur- 

 ple itself was derived. It had been examined frequently and seemed 

 to be clean. 



The question of crossing-over in the male was the first point attacked. 

 Complementary P! matings were made (June 13, 1913) by crossing 

 purple vestigial to wild (" coupling") and by crossing purple by vesti- 



