PERIODICITY IN THE PRODUCTION OF MALES. I89 



generations, in a period of many males, the males are produced 

 mostly by the daughters at the end of the family in the first of 

 these generations, by the daughters in the middle of the family 

 in the second generation, and by the daughters in the first part 

 of the family in the third generation. In the manner in which 

 these results have been published, the males of the first and third 

 generation (appearing under date of the beginning of their 

 respective families) seem to be four or five days apart, whereas 

 they may be only one day apart, or even hatched on the same day. 

 It seemed advisable to obviate as many of these difficulties as 

 possible, and to obtain evidence of periodicity less open to 

 objection. The method adopted was as follows: First, a line 

 was selected which was producing only a moderate number of 

 males, with the expectation that the periods of male production 

 would be completely separated by periods in which there were 

 no males at all. Second, instead of isolating all the daughters of 

 a single family in each genertion, parts of a number of families 

 were isolated. This second precaution was taken to smooth out 

 irregularities and to reduce the apparent length (and hence the 

 overlapping) of the periods, both of many and of few males. 



Lines Exhibiting Periodicity. 



The line selected to meet the above conditions was one of the 

 F;^ lines from a cross between an English and a Nebraska line, 

 described in an earlier paper (Shull, '15). Several females of 

 this line, all of approximately the same age, were placed together 

 in a dish. When they reached maturity, the daughters of the 

 early part of their families were rejected; but the daughters 

 produced after about 24 hours of egg laying were preserved. 

 From these daughters the "sex ratio" (ratio of male-producing 

 to female-producing females) was determined, and from them 

 also the next generation was reared. The first daughters were 

 in each case rejected because, as I have shown before (Shull, '10), 

 the first daughters are less commonly male-producers than are 

 daughters in the middle of the family. 



The daughters selected for rearing the next generation were iri 

 like manner kept together in one dish until about 24 hours after 

 their first daughters appeared. At this time a second lot, all of 



