ARE RECESSIVES LOSSES OF GENES? 89 



taken place in the mother of the male between forked and 

 fused. The total cross-overs between forked and fused 

 are less than 3 per cent and yet they include all the rever- 

 sions to full eye. 



Only the reverted sons have been spoken of above in 

 order to simplify the situation, but of course the reverted 

 chromosome might have passed into an egg that de- 

 velops into a female. The experiment can be so planned 

 that evidence of crossing-over will also be detected in the 

 reverted daughter. The ordinary daughters will be homo- 

 zygous bar (see Fig. 49, 1). The reverted daughters will 

 be heterozygous for bar eyes, and either forked or fused. 

 None of them are both forked and fused and none of them 

 are not-forked not-fused. 



The crossing-over, that brings about reversion to 

 round eye, must not only have left one X-chromosome 

 without a bar gene, but must have put the other bar gene 

 into the bar chromosome (Fig. 51a). In appearance a 

 male with two bar genes (double bar) is similar to a 

 male with one bar gene, but its eyes are smaller and 

 the number of its facets is reduced. It has been named 

 double bar (Fig. 49,6). The presence of two allelomorphic 

 genes in the same linear series is an exceptional occur- 

 rence that has as yet not been observed in any other 

 mutation. It can be pictured only by supposing that the 

 bar genes lying opposite to each other before crossing- 

 over are shifted a little when the crossing-over occurs. 

 The result is that the double bar chromosome is length- 

 ened by one bar gene at least, and conversely that the 

 other chromosome has been shortened by the "loss" of 

 the bar gene. 



Sturtevant has put the theory of reversion of bar to a 

 number of critical tests. There is an allelomorph of bar 

 (that arose as a mutation of bar) called infra-bar (Fig. 

 49, 7 and 8), whose eyes are somewhat different in size 



