DIRECTION AND FREQUENCY OF MUTATION 227 



either male or female sex-cells. The conclusion is based on the 

 assumption that approximately equal numbers of males and 

 females were observed. In most of the present counts males 

 and females were not listed separately, but other counts indicate 

 that while there may be a wide departure from equality in indi- 

 vidual bottles the departures occur in both directions, and in a 

 large series of observations the number of males and females 

 approaches equality. In one part of the present observations, 

 the selection lines, accurate records were kept of the sex of all 

 individuals. Here there are 7189 females and 7138 males, a close 

 approach to equality. Of the eight mutants to full in these 

 selection lines seven are females and only one a male. This ratio 

 distinctly favors the view that mutations may arise in male sex- 

 cells as well as in the female. 



The other mutations in the bar series do not throw any light 

 on the present question because of the probability of missing 

 heterozygous females. In fact, they give fewer females in pro- 

 portion than is to be expected on the hypothesis that the muta- 

 tions occur in the female sex-cells alone. The only recorded 

 mutation from full to bar appeared in a son. Of the five muta- 

 tions from ultra-bar to full only one is a daughter, while four are 

 sons. Of the three mutations from ultra-bar to bar all are sons. 

 Finally, of the three mutations from bar to ultra-bar one is a 

 daughter and two are sons. 



It is interesting to speculate further on the degree of inde- 

 pendence of the changes in the bar gene. For instance, is the bar 

 gene of one chromosome in the female independent in this 

 respect of the bar gene in the other or do they change 

 simultaneously? Unfortunately, no satisfactory test of this 

 proposition has been devised for the bar series. Muller ('20) 

 has, however, pointed out that mosaic mutants involving re- 

 cessive sex-linked genes are always males and that this indicates 

 that mutations occur in only one member of a pair of chromosomes 

 at a tune. 



There is another question of equal interest which may perhaps 

 be capable of solution. Does the change occur with equal readi- 

 ness in all individual animals? This cannot be tested in females. 



