158 ELMER ROBERTS 



maintained that in asexually reproduced organisms and in indi- 

 viduals arising by self-fertilization selection is futile except for 

 the piu-pose of isolating pure lines. The original interpretation 

 of a pure line was that it consisted of the offspring of a self-fer- 

 tilized individual. Later, the term, pure line, was extended to 

 homozygous characters in individuals reproduced dioeciously or 

 biparentally. It has been the belief of the majority of workers 

 in genetics that selection only separates pure lines and is ineffec- 

 tive when applied to a unit character. Such an hypothesis must 

 also assume a gametic or factorial purity which is unchanging 

 except through mutation. 



The questions wdth which the present investigation deals are: 



1. Does long-continued selection have an effect upon a Men- 

 delian character? 



2. Can there be a gametic or factorial contamination of a char- 

 cter, and, if so, is selection then efficacious? 



II. MATERIAL AND METHODS 



In Drosophila ampelophila, the fruit fly, there is a wingless 

 or vestigial winged variety, the origin of which was described by 

 Morgan and Lynch ('12). This vestigial wing behaves as a re- 

 cessive character when crossed to individuals with normal long 

 wings. The offspring in the first generation have long wings 

 while in the second generation both long-winged and vestigial- 

 winged are produced. However, the ratio of long-winged to 

 vestigial-winged is not 3:1. In this experiment, 9248 F2 indi- 

 viduals were produced, 7381 of which were long-winged and 1867 

 vestigial, or a ratio 3.95: 1. With this number of individuals 

 such a departure from a 3 : 1 ratio cannot be attributed to chance. 

 Morgan explains the departure from the theoretical ratio as 

 being due to the low viability'' of the vestigial-winged variety. 

 A vestigial and a normal wing are shown in plate 1, figures 1 

 and 14. This character, vestigial wing, is well adapted to such 

 an investigation as outlined, because it is a well-defined Mende- 

 Uan character and shows some variability in size and venation. 

 The stock was originally obtained through the courtesy of Prof. 

 T. H. Morgan, of Columbia University. 



