286 CARNEGIE INSTITUTION OF WASHINGTON. 



place only in the female, and not in the male. Such is the case. It would 

 be expected that this reversion would occur only late in the history of the 

 germ-cells, since crossing-over is known to occur late. This also is found to 

 be the case. An extreme allelomorph of bar, called by Zeleny ultra-bar, has 

 been obtained from bar, and it has now been shown that in this event also 

 crossing-over occurs within three units of the bar locus. These results can 

 be explained if we represent the bar gene by B and two adjacent loci by M and 



M B N 



N respectively. The homozygous bar female will be ^, ^ ^ ,- If we sup- 

 pose that an unequal cross-over takes place in such a way as to produce 

 M BB N' and M'N, the former will represent ultra-bar, and the latter the 

 reversion to normal. A confirmation of this assumption is found in Zeleny's 



data (collected for another purpose), namely, that z^ (homozygous bar) has 



B 



very nearly, if not quite, the same number of ommatidia as ultra-bar heterozy- 

 gous for normal. These should be alike, by hypothesis, since each has two B's. 

 Further tests of the hypothesis stated above are now under way. 



In cooperation with E. G. Anderson, a study is being made of the crossing- 

 over in the X-chromosome of the triploid (3n) female. Each of the three 

 X-chromosoraes is followed by means of five mutant loci. For example, one 

 of the five points was marked by yellow in one X; by its wild-type allelomorph 

 in a second X; and by scute, which gives no crossing-over with yellow, in the 

 third. In addition, each X carried a recessive mutant gene crossveinless. 

 Such a 3n female was outcrossed to a not-crossveinless male. All the cross- 

 veinless daughters of this outcross have received from the mother two X- 

 chromosomes that have emerged from the same reduction division. A study 

 of the characters representing the five marked points present in such females 

 shows the origin of the different parts of each of the two X-chromosomes. It 

 has been found that crossing-over may occur between all threeX-chromosomes 

 simultaneously, since chromosomes emerge that are composed of parts taken 

 from each of the three chromosomes that entered. The most frequent type 

 of female is one that carries one chromosome that has not undergone crossing- 

 over (and is a duplicate of one of the entering chromosomes) and another X 

 that is composed of parts of the other two. A fairly frequent result is the 

 emergence of two X's that are duplicates of each other for part of their 

 length but not for the remaining part. This result indicates that each of 

 the original chromosomes was in a two-strand stage when crossing-over 

 occurred. In these 3n females crossing-over is relatively less frequent in the 

 right end (from lozenge to bar) than it is in the normal (2n) female, while 

 in the left it is relatively much more frequent. Thus, about 15 per cent of the 

 emerging chromosomes have come from crossing-over between yellow and 

 ruby, while the normal percentage is about 7. In the extreme left end cross- 

 ing-over has been observed between yellow and scute, which had not been 

 certainly detected in the diploid form. A more extended analysis of the data 

 is expected to throw light on the details of the process of crossing-over. 



In addition to the sex-linked mosaics, due to elimination of one of the 

 two X-chromosomes of an egg at an early cleavage, a number of mosaics have 

 appeared in which the two parts differ in autosomal characters. Most of 

 these are interpreted as due to two separate nuclei in the original egg, each 



