BRISTLE INHERITANCE IN DROSOPHILA 453 



render the race susceptible to selection. This seems to be an 

 unusual case ; the general experimental result is to find some mod- 

 ification in the means following selection, The greater the 

 environmental control, and accordingly the less accurately the 

 character represents the germinal condition, the more ineffec- 

 tive becomes selection. In the case of the vestigial wings, it is 

 quite possible that the strong temperature factor sufficiently 

 upset the connection between the germinal and somatic condi- 

 tions to make selection ineffective. 



An interesting race of flies in which one extra dorsocentral 

 bristle frequently appeared has been reported by Reeves ('16). 

 All flies in the race, both the normals and those bearing extra 

 bristles produced an excess of normal offspring as well as some 

 with the extra bristle. The ratios range from 90:1 to 3.3:1. 

 There appeared more extra flies from matings between extras 

 than from matings between normals and extras, which, in turn, 

 gave fewer extras than the matings between two normals. The 

 sums of all the normal and extra flies from each type of mating 

 in this race give the following ratios: four generations extra by 

 extra, 603 normals 90 extras 6.7 : 1 ; five generations extra by 

 normal, 1684 normals, 87 extras, 19.3:1; one generation normal 

 by normal 335 normals, 8 extras, 41.8: 1. The race was started 

 by five flies with extra bristles being crossed with normals from 

 the same stock; after this brother by sister matings of the above 

 three types were variously made during five generations. The 

 ratios in individual families were found to show a certain amount 

 of grouping about the ratios 3:1, 15:1, 63:1. On the basis of 

 the grouping of these ratios, the conclusion was drawn that a 

 Mendelian explanation was possible; the higher ratios indicating 

 that respectively two and three multiple or duplicate genes were 

 involved. The author is quite right in saying that there is no 

 dominance in the ordinary sense, and in so saying proves the 

 utter impossibility of explaining the ratios in the above simple 

 Mendelian sense. The ratios 3:1, 15:1, and 63:1 are F2 ratios 

 indicating the proportion of full recessives that appear after an 

 Fi in which all the individuals are dominants. To assume that 

 the ratios obtained may possibly be Mendelian supposes that 



