A RECESSIVE CHARACTER AND SELECTION 159 



One male and one female were taken at random as the begin- 

 ning of all the series. From the offspring of these two individuals, 

 the male and female with the longest vestigial wings were used to 

 begin selected series A. The brothers and sisters of these bred 

 inter se started control series B to be used as a check. The 

 third series, which will be referred to as 'crossed-in-and-selected' 

 series C, had its origin as follows: ten selected males from the six- 

 teenth generation of selected series A were crossed to long- 

 winged females of wild Drosophila obtained from bananas in a 

 local grocery. The first generation was long-winged, while in the 

 second, long-winged and vestigial-winged appeared. Selection 

 was begun among the vestigial-winged segregates of the second 

 generation, those individuals with the longest vestigial wings 

 being used as parents of the next generation, and so on. Some 

 of the vestigial-winged males of the second generation of the 

 cross were mated again to long-winged females. This process of 

 'crossing-in' was repeated eight times in some of the series, but 

 in this paper a series which was crossed-in only once is reported 

 for the reason that it consists of a larger number of generations 

 than any of the others. Duplicate series were also run, formed 

 by crossing vestigial-winged females to long-winged males, the 

 reciprocal of the above cross. 



The length of the vestigial wing is the measurement used to 

 determine the effect of selection and of 'crossing-in.' The ideal 

 method would have been to measure the surface of the wing and 

 note the venation, but this was prohibitive on account of the labor 

 involved. However, the length is a good character to use as an 

 indication of any changes in size. It is better than width, for 

 in large thin vestigial wings, curling sometimes renders the 

 measurement of the width impossible or extremely difficult. The 

 generations were preserved in 75 per cent alcohol and measure- 

 ments made later. 



In all cases selections were made by simple inspection with a 

 hand lens, the effectiveness of which can be seen from tables 1, 

 3, 5, 6, 7 and 9. 



Measurements were made by the use of a camera lucida, 

 tracing the length of the wing and measuring the line thus pro- 



