GENETIC EVIDENCE FOR DIPLOID MALES. 439 



type is more easily made in females than in males and is espe- 

 cially clear in light females heterozygous for orange or ivory. 



Three pairs of allelomorphs affecting the wings have also been 

 used. Wrinkled (w), recessive to type (W), prevents complete 

 expansion of the wings and often the normal development of legs 

 and antennas. It is almost completely recessive and somewhat 

 lethal in the homozygous and azygous conditions. Reduced (r), 

 completely recessive to type, reduces the wings, especially the 

 primaries, in size and venation. No overlapping with type occurs. 

 These mutations are discussed at length elsewhere (Whiting, P. 

 W., 1926). Defective (d) (referred to in previous publication 

 as duos. Whiting, P. W., 1924), like wrinkled, is almost com- 

 pletely recessive but shows some overlapping with type (D). It 

 reduces the length or causes disappearance of the fourth branch of 

 the radius vein (r 4 ) in 90-95 per cent, of pure stock under stand- 

 ard conditions. Unfortunately reduced and defective cannot al- 

 ways be recognized with certainty in wrinkled, while defective 

 cannot be identified in reduced where the veins are so generally 

 disarranged. 



There is no linkage between any of these factors. 



CROSSES PRODUCING BIPARENTAL MALES. 



Table I. gives summaries of crosses which produce biparental 

 males. Whenever significant differences involving locus D occur 

 they are indicated in the formulae. The impaternate haploid males 

 arise from unfertilized eggs and resemble their mothers. Where 

 the mothers are heterozygous these males fall into two classes as 

 expected. Females show all the dominant traits of both parents 

 as do the biparental males. 



The first cross in the table will serve to illustrate this. Mothers 

 are light, fathers black. All daughters are black, most sons light, 

 the few biparental sons black. 



In classes c and e, section I., mothers are defective (d), fathers 

 normal (D). If numbers in these classes be combined it is found 

 that 6 of the 58 biparental males are defective, 10.34 2.70 per 

 cent.; 753 of the 820 impaternate males, 91.83 0.64 per cent.; 

 46 of the 569 females, 8.08 0.77 per cent. Similarity of per- 

 centages in biparental males and females is apparent. The 13 



