OF MUTANT CHARACTERS. 139 



limited at the two values, 67 and 90 per cent, the accumulation of the 

 modifiers continued to make the wings shorter, and the shortest indi- 

 viduals gave the highest percentage of truncates, since they were on 

 the average the ones carrying most modifiers. 



The increase in the number of completely sterile individuals was 

 independent of the other changes and due to a mutation which occurred 

 early in the selection. Hyde (1914) was able to eliminate this sterility 

 from the truncate stock by breeding for some generations from those 

 families which showed least sterility. But there was no rise in pro- 

 ductivity parallel to the elimination of the sterility. This sterility 

 behaved as a recessive in the FI out-crosses, and in F 2 reappeared, but 

 in such proportion and distribution as to suggest that it also was 

 complex. 



The fact that the main gene for truncate is in the second chromosome 

 was established by Muller and Altenburg through the non-occurrence 

 of cross-overs in back-cross tests of males heterozygous for truncate 

 and black. In back-cross tests of similar females there was somewhat 

 under 25 per cent of crossing-over. A back-cross test showed that 

 there is somewhat over 25 per cent of crossing-over between star and 

 truncate, and this information, in connection with the known distance 

 of about 50 between star and black, showed that truncate is located 

 about midway between star and black and slightly nearer to black. 

 Since the amount of data secured in these tests is not large and because 

 truncate is so elusive a character, the location is not precise, though 

 the locus is probably not far from 28 (see Snub). 



TRUNCATE REOCCURRENCES. 



Many of our mutant characters have made their appearance more 

 than once, and occasionally under circumstances which make it certain 

 that there has been a new occurrence of the same mutative process 

 that was responsible for the original appearance of the character. In 

 other cases it is probable that the character is not reappearing because 

 of a fresh mutation, but that the original mutant gene had been intro- 

 duced in some previous cross and has been unsuspectedly present for 

 several generations but has been unable to appear because of the way 

 in which the crossing has been carried on. In still other cases a char- 

 acter appears which resembles very closely another already known 

 character, but the two are the result of mutations in entirely different 

 loci which are often in different chromosomes. Thus, we have at least 

 half a dozen mutant characters of the "genus" pink which are so 

 similar as to be practically indistinguishable in mixtures, yet which are 

 dependent on entirely distinct genes. Characters of the genus truncate 

 are the most frequently recurring of any, with the possible exception 

 of the beadeds. These two kinds of character have both come up in 

 the breeding work three or four times each year. 



