32 ILLINOIS BIOLOGICAL MONOGRAPHS [274 



ludinally and transversely, furnishing good generic characters through- 

 out the family. It may be nearer to pi or to sigma and varies from a point 

 on the transverse line through nu to one just cephalad of the transverse 

 line through pi. 



Only one of the minute setae xa and xb is usually present on the abdomi- 

 nal segments, altho both are found in Chloridea arniigera and frequently 

 the persisting seta is associated with a minute spot, which evidently repre- 

 sents the vestige of the other. The transverse variation in the location of 

 xa with reference to alpha and beta affords some phylogenetic indications, 

 altho the minute size of xa precludes the use of its variations in tables. 

 Typically this seta is further dorsad of beta on segment 1 than on segments 

 2 to 7 inclusive, being most ventrad of all on segment 8. This variation 

 involves the migration of beta as much as that of xa. Two genera examined, 

 Catocala and Xylina, afford exceptions to the rule, xa being further dorsad 

 relative to beta on segment 8 than on 1. In all other groups investigated 

 xa is ventrad of the longitudinal line through beta, whereas in these two 

 genera it is much dorsad of this line, an instance of parallel development. 



The difference in the longitudinal position of the spiracle with reference 

 to the surrounding setae on successive segments in the individual follows a 

 certain definite plan throughout the family. On segment 1 the transverse 

 line through rho ranges according to the group from a position distinctly 

 cephalad to one a little caudad of the spiracle. Rho is clearly further 

 cephalad in segments 2 to 6 inclusive than in 1, its transverse line passing 

 cephalad of the spiracle or tangent to the cephalic margin. Segment 7 

 presents approximately the condition found in segment 1, the seta being 

 further caudad than in segments 2 to 6, and ranging from a situation 

 cephalad of the spiracle to one caudad of it. Two exceptions to this general 

 plan have been noted. In Sidemia devaslatrix the condition on segments 

 1 and 7 does not differ clearly from that in the intermediate segments and 

 in Papaipema nebris segment 7 exhibits the same location which it occupies 

 on segments 2 to 6, segment 1 differing from the rest as usual. The fact 

 that segments 1 and 7 show much greater variation in this respect than the 

 intermediate segments indicates that the condition found on segments 2 to 

 6 is the more primitive. Evidently in segments 1 and 7 the spiracle has 

 migrated cephalad of its primitive position. In some cases this process has 

 proceeded further on segment 1 and with other species on segment 7. The 

 condition on segment 8, where the position of rho varies according to the 

 group from a location cephalad to one caudad of the spiracle, reveals no 

 uniform relation to that on other segments. Altho these minor variations in 

 the longitudinal situation of the spiracle offer points of morphological interest 

 they usually do not lend themselves readily to taxonomic application. The 

 Catocalinae, however, apparently differ from all other groups within the 

 family in having rho distinctly cephalad of the spiracle in segment 1. 



