332 



HERMANN J. MULLER 



The offspring were as seen in table 3. 



TABLE 3 



Numbers of females expected = numbers of males expected. 



A considerable differential viability came into play in these 

 bottles, as is proved by that fact that certain 'contrary' or oppo- 

 site classes which were necessarily produced in equal numbers 

 at fertilization (no matter whether linkage was involved or not), 

 gave rather different counts of adult flies (e.g., 206 females; 233 

 males; 97 pink straight, 62 red bent; which were the worst discrep- 

 ancies). On allowing for these differences obviously due to 

 viability, we find that there are no discrepancies due to linkage of 

 bent with black or pink, for the average between black straights 

 and gray bents (pink plus red), is to the black bents as 3:1, and the 

 average between pink straights and red bents (gray plus black) is 

 to the pink bents as 3:1, as expected on independence. These 

 results therefore confirm those previously obtained in regard to 

 the independence of black (Chromosome II) and bent; they show 

 in addition, however, that bent is independent of pink (Chromo- 

 some III). 



The distribution of bent with regard to black and pink was next 

 determined separately for each sex, by means of back-crosses. 

 A normal male was mated to black pink bent females. The Fi 

 males, which were heterozygous for all three factors, as well as for 

 sex, were then mated to triply recessive black pink bent females. 

 If the factors were independent, all classes in the next generation 

 should be equal in number. The actual result was as shown in 

 table 4. 



