GENETIC EVIDENCE FOR DU'LOID MALES. 447 



4. When three of these factors affect one structure (the wing 

 in this case) if one recessive and two dominants are contributed 

 by one parent, their allelomorphs by the other, this structure in 

 biparental males shows all the dominant characters. 



5. From these results it is concluded that biparental males are 

 diploid (Whiting, P. W. and Whiting, Anna R., 1925) at least 

 for the four chromosomes that can be identified genetically. 



6. Biparental males and their daughters are often abnormal in 

 appearance and usually sterile or nearly so. When fertile they 

 breed as dominants (with one, and possibly two, exceptions noted 

 above) . 



DISCUSSION. 



When different types of crosses are made between inbred re- 

 lated stocks and results summarized it has been found that definite 

 relationships exist between various percentages derived from these 

 summaries. 



Percentage of males among biparental offspring, (previously 



biparental <$<$ X 100 . 

 called percentage of patroclmy), biparental ^ + 33 > ' s 



99 X 100 



tively correlated with percentage of females, - - , and with 



total, 



biparental rft? + 9$ X 100 

 percentage of total biparentals, - ^ . - . In 



other words when a type of cross results in a high percentage of 

 males among biparentals there is a low percentage of females and 

 of total offspring from fertilized eggs. The female percentage is 

 lower than it would be if the decrease were due only to the trans- 

 formation of some fertilized eggs into biparental males. This 

 indicates that there is a mortality of fertilized eggs in these crosses 

 directly correlated with number of biparentals which are males. 

 Types of crosses that produce no biparental males have the high- 

 est percentage of females. 



The question as to why these individuals are males cannot be 

 answered at this point. Dr. Castle has suggested that they may 

 correspond to the intersexes which Goldschmidt gets in Lymantria 

 (Goldschmidt, Richard, 1927). Some of the facts support this 

 and it may be that the presence of certain genetic factors in some 

 individuals causes sex reversal. No gradations have been ob- 



