BRISTLE INHERITANCE IN DROSOPHILA 67 



both types of crosses the children showed dominance of the 

 normal, and all grandchildren gave the same general ratios of 

 normal to extra. IMoreover, the distributions of these F2 extras 

 were similar in the two types of crosses (table 9 and figure 4). 

 This is enough to indicate that there is no sex linkage involved. 



3. MODIFICATION OF THE DISTRIBUTION OF EXTRACTED EXTRAS 



In comparing the distribution of the extra bristles in the F2 

 of the above crosses with that of the extra bristles in the selected 

 race of the corresponding generations, a marked difference is 

 observed. This is shown in figure 2. In these curves the males 

 and females have been put together and the normals occurring 

 in the inbred have been omitted, since corresponding normal 

 appearing flies in the F2 of the crosses could not be separated 

 from the genetically normal flies. The data have been plotted 

 on the basis of 500 in each curve, to facilitate comparison. In 

 general the distribution of the extras appearing in the F2 of a 

 cross is lower than the distribution of the corresponding gener- 

 ation of uncrossed extras. The mode of the extracted extra 

 bristles from the first cross is at +1, although the frequency of 

 + 2 is nearly as great; the mode of the second selected gener- 

 ation is at + 2. In the later crosses, the mode of the extracted 

 extra bristles is at -|- 1, and the mode of the tenth selected 

 generation is at + 3. 



With these curves in mind it will be well to compare the ex- 

 tracted extras from flies crossed before continued selection, with 

 the extracted extras from flies crossed after selection. It is 

 readily seen that the extracted curve from the later crosses is 

 strikingly higher than that from the first crosses. The range 

 is much higher; but the mode is still at -|- 1, even more promi- 

 nently so than in the first case. The standard deviation in the 

 first set of extracted extras is .695 ±.017, that of the correspond- 

 ing inbred generation is .863 ± .005. In the later crosses standard 

 deviations are as follows: extracted extra, 1.329 ±.018, inbred, 

 1.402 ±.021 (table 6). It appears to be a very definite fact 

 that the modification of the distribution of extra bristles by 



