BRISTLE INHERITANCE IN DROSOPHILA 93 



where the inbred and extracted distributions are compared. 

 The first set of curves show this fairly well, but in the second 

 set, in which the extra parents had been selected much longer, 

 the phenomenon is very clear. On the other hypothesis, if the 

 factor for extra bristles were modified in Fi, one would expect 

 to find the whole curve of the Fo extras lowered, and another 

 supposition would have to be made to explain the occurrence of 

 high grades. Such an added supposition would involve the 

 modification of the factor, sometimes in various degrees,, and 

 other times not at all — a supposition easy to make (Castle '14) 

 but difficult to explain. 



That there is a greater modification in the F2 extracted extras 

 when the extra parents had been selected for several generations, 

 than when they had not been selected, is the result expected 

 if the selection had accomplished no more than to drop out 

 certain accessory restricting factors, which being present in the 

 wild parent would produce their effect in the second generation 

 at the normal end of the distribution just as strongly as they did 

 in the crosses before selection. To interpret this without using 

 accessory factors one would have to suppose that a factor for 

 extra bristles, that had been made more extra by selection, 

 was more susceptible to contamination, just as squeezing a sponge 

 will make it take up more water; but as has been shown, this 

 kind of a mechanism will not explain the occurrence of the high 

 extremes in this modified distribution. When one considers 

 that the extra bristled condition is due, not to a single factor 

 of which anything is known, but to the nucleus of determiners 

 that carry the main heritage, such a modification hypothesis 

 becomes vague and confused. 



The occurrence of extras in the Fi of the crosses may be ex- 

 plained by incomplete dominance, in which case the proportion 

 of F2 extracted extras should be too high instead of too low. 

 Their occurrence may be explained by a heterozygous condition 

 of the restricting factors in the wild New York 1912 race, which 

 had been escaped being weeded out by selection. The occurrence 

 of extras in the New York 1912 race suggests such a heterozyg- 

 ous condition. It seems probable that the slightly low propor- 



