78 ILLINOIS BIOLOGICAL MONOGRAPHS [78 



Cossula magnifica is so different from other cossid larvae that at 

 first it seems to deserve the rank of a separate subfamily. Certain 

 European larvae form a series connecting this type with the others, hence 

 such a division could not be defended. For example, in Phragmatoecia 

 castaneae Hiibner, the crochets are uniordinal but arranged in a com- 

 plete circle and the last pair of spiracles are enlarged and moved dorsad 

 as in Cossula; Dyspessa ulula Borkh. has rudimentary crochets in indefi- 

 nite transverse rows but all the spiracles are in line. 



Prionoxystus robiniae is our most common species. The larva, which 

 bores in locust and other trees, often reaches a length of three inches. 

 P. macmurtrei, the only other species, is smaller and bores in a number of 

 different trees. The crochets of the latter species are circular in arrange- 

 ment while those of the former are in a very narrow ellipse. 



Cossus centerensis has a larval structure nearly identical with that 

 of Prionoxystus. The two are hard to distinguish and the character 

 given in the table does not hold for European species of Cossus. There 

 is a Rocky Mountain species, the larva of which I have not seen. 



Zeuzera pyrina has a world-wide distribution. Like Cossula mag- 

 nifica, it is distinct from the other forms and is the only species of the 

 genus found in North America. 



Larvae of the western genus Hypopta and the subtropical Inguro- 

 morpha have not been seen. 



Family Psychidae 



The Bagworm moths are a peculiar group which Dyar at one time 

 considered as deserving the erection of a separate suborder. He seems 

 later to have abandoned this view, which was based on the fact that 

 alpha and beta are located on the same annulet of the abdominal seg- 

 ments instead of on successive ones. The homology seems clear enough 

 altho the position of the setae is, it is true, different from that found in 

 other larvae. This combined with the absolutely unique horizontal pro- 

 thoracic spiracle shows that the family is an ancient one. The different 

 genera are remarkably uniform. 



Head overarched by the prothorax above but large and not truly 

 retractile ; front and adfrontals various ; labrum notched to about half 

 its depth; ocelli normal. Body with primary setae only, these very 

 minute in later stages. Prothorax with all setae normal. Kappa group 

 trisetose, epsilon and rho distant from each other ; mesothorax and meta- 

 thorax normal except that the Pi group is bisetose. Abdomen with 

 alpha, beta, and rho in almost a straight line above the spiracle, but 

 alpha and beta close together on segment 9 ; kappa and eta adjacent ; 

 mu present; Pi group normal. Prolegs present as usual on segments 3, 



