In Different Species. 61 



forked. Hairs on costa stand out at an angle of about 60°. Females fertile; eggs 

 normal. Resembles most closely forked virilis. Heterozygous females not examined. 



Relationships. 



One of the first points revealed by this comparison is that a 

 similarity in name does not necessarily indicate a close similarity in 

 the appearance of the characters in different species. For instance, 

 forked-1 and forked-2 in willistoni suggest singed in mrilis instead 

 of forked, which is more like stubby in willistoni. 



Our search for some feature upon which to base a dichotomous 

 separation of the characters in the whole series has been largely 

 unsuccessful. The range of modifications shown by the forked 

 allelomorphs in melanogaster, for instance, covers most of the condi- 

 tions found in any of the ''forked," "singed," or ''stubby" characters 

 in the other species. No single criterion, therefore, can be used 

 exclusively for determining homology or lack of homology. The 

 best we can do at present is to relate the characters on the basis of 

 the degree of resemblance. 



The two singed allelomorphs in melanogaster, the two forked 

 allelomorphs in willistoni, and singed in obscura agree in regard to 

 female sterility and presumably in regard to egg abnormality (see 

 descriptions). This gives good ground for considering them to be 

 parallels. It is supported also by the general appearance of the 

 characters, which are of the t}n?^ having twisted or curled rather 

 than jagged and sharply bent bristles. Correlated with this are the 

 depressed bristles and curled and depressed hairs on the thorax. 

 We may group these characters together, then, in the "singed" 

 series. 



The other character in willistoni (stubby) differs in all of these 

 respects and bears a closer resemblance to the forked allelomorphs 

 of melanogaster. So we may group these together in the "forked" 

 series. In obscura the character stubby represents such a slight 

 modification that it is hard to compare with the others, especially 

 since no specimens are available for examination. Since singed 

 seems to belong to the singed series, however, stubby may tentatively 

 be correlated with the forked series. 



These relations are shown schematically in figure 14, in which the 

 chromosome maps are oriented so that the loci of parallels^ cor- 

 respond. 



In simulans forked has been shown by Sturtevant (1. c.) to be 

 allelomorphic to that in melanogaster, and consequently the simulans 

 map would correspond to that of melanogaster, and iwojy be omitted 

 here. 



* The term parallel is used to indicate a resemblance which suggests homology. 



