82 BULI.ETIN 64. 



of the different forms, but it is more in degree than in kind. In 

 the antennae there are some specializations and in two species this 

 has been carried to a remarkable extent. In the labial palpi and 

 the other mouth parts there are some modifications but they are 

 of no great systematic value. The relative size of the species is 

 of some value. The more generalized species seem to have a cer- 

 tain medium size from which the more specialized forms have 

 varied each way. In the exterior genital organs there are great 

 variations in form and these are of considerable value in tracing 

 the phylogeny of the species. With the genitalia as the main 

 witnesses and the other organs as subordinate witnesses, we will 

 attempt to trace the afiinities of the species. For the general plan 

 of classification see the synoptical table. 



1. The most generalized type is seen in C. luteolellus. Plates I. , 

 v., XIII., Fig. I. The valve is elongate, oval in cross section and 

 thickly covered with rather short hairs. The harpe is more 

 chitinous and bears a dense row of rather short hairs. In the 

 primitive Crambus the clasps were probably similar independent 

 organs. In this species the uncus is longer than the lower limb, 

 which is more slender. The scaphium is long and keel-shaped. 

 From this simple type of genitalia there have been wide devia- 

 tions along several lines. This species has the somber yellow 

 color and the medium size of a generalized Crambid. The wings 

 are rather broad and the veins branch nearer the base of the 

 wings than in most species. The antennae and palps exhibit 

 little specialization. This insect is more closel}^ related to the 

 progenitors of the many species now living than any other form 

 studied. Structurally, this insect is closely related to the 

 Caliginosellus Division, but its color connects it with the lighter 

 colored moths. As color in these more generalized species seems 

 to be a quite reliable character, it may be allowed that C. luteo- 

 lellus is not the progenitor of the Caliginosellus Division but that 

 the separation occurred in a more primitive form. Therefore 

 C. luteolellus, while being the most generalized species studied, is 

 most closely related to the insect through which the more special- 

 ized forms have come. 



Tlic Caliginosellus Division. — Plates I., V., XII., XIII., Fig- 

 ures 2-5. 



2. Crambus caliginosellus. Structurally this insect is most 



