On Certain Grass-Eating Insects. 8i 



Before attempting to seek the cause of specialization, it may be 

 well to ascertain just what has occurred within generic lines. In 

 this study, based as it is on less than half of the North American 

 species, one can hardly expect to find unbroken lines of develop- 

 ment. We should expect to find gaps in the living forms and 

 much larger ones when only a part of these are represented. 

 Again, the larger the gaps, the greater the liability to error. In 

 this study we shall assume that the better an organ is fitted to 

 perform its functions the more highly specialized it is and, con- 

 versel5^ there may be a specialization by reduction as when an 

 organ becomes useless and there is a reduction of the superfluous 

 part. 



The most striking and at the same time most superficial specific 

 characteristics are the markings. As in many other groups of 

 moths, we find that the most generalized species have a rather 

 somber gray or yellow color with indistinct markings. The 

 species studied ma}^ be divided into two very natural divisions on 

 color alone. The grayish or dark colored species form a very 

 natural division, the Caliginosellus Division. Later we shall see 

 this division is united by strong structual afiinities. The rest of 

 the species are yellowish in color or of a color that may be derived 

 from the yellow, and these represent an entirely different line of 

 development, the Luteolellus Division. The lowest species of 

 this division is C. luteolellus, which stands just below C. calig- 

 inosellus and is closely related to this species : the two probably 

 having sprung from a common ancestor. This division may be 

 divided into groups on color and no violence be done to structual 

 affinities. The first group, the Decorellus Group, consists of yel- 

 lowish moths with few or no markings. Next we might separate 

 the Unistriatellus Group by the broad white stripe on a rufous 

 fore wing. But as we attempt to classify the more specialized 

 forms on maculation, we find the variation in color has increased 

 more than structual variations and therefore our conclusions are 

 at fault. In the more specialized forms the tendency to acquire 

 adaptional colors has been greatly increased ; con.sequently color 

 may be a quite reliable systematic character in the more gen- 

 eralized species, but it is not in the more specialized forms. 



In the venation there are considerable variations. These varia- 

 tions are useful in indicating the relative degree of specialization 



