Dec, 1917-] Crampton : Abdominal Segments and Appendages. 231 



Some of the representatives of each superorder have retained 

 characters common to certain other superorders, and may therefore 

 be spoken of as annectent between these groups. Thus, the Mantids 

 and Isoptera resemble the Grylloblattids and Phasmids in some re- 

 spects, while in others, they resemble the Embiids and Dermaptera. 

 Similarly, the Grylloblattids and Phasmids resemble the Mantids and 

 the Isoptera as well as the Embiids and Dermaptera : and, in the 

 same way, the Embiids and Dermaptera resemble the Mantids and 

 Isoptera as well as the Grylloblattids and Phasmids. Of these an- 

 nectent forms, the most important are the Mantids, Grylloblattids and 

 Embiids. 



The older method of representing the relationships of the orders 

 of insects by means of a dichotomously branching treee, drawn in one 

 plane, is very unsatisfactory, since it does not allow for the fact that 

 several lines of descent may approach one another from different 

 directions (i. e., the branches of the tree should be represented in 

 three planes) and that one group of insects may be intermediate be- 

 tween two other groups, being hardly more closely related to one 

 than to the other. If the relationships of the three superorders here 

 discussed were to be representd graphically, it would be more exact 

 to represent these superorders as forming the three apices of a tri- 

 angle, each apex of which is connected with the other two by mutual 

 bonds of relationships — or better yet, to represent them as three over- 

 lapping circles, each of which intersects the other two, and all having 

 a certain amount of territory in common, although each circle forms 

 a distinctly demarked group, when considered separately. In this area 

 common to the three circles, the Embiids, Grylloblattids and Mantids 

 would be placed, since these three members of the different super- 

 orders have a surprisingly large number of features in common, and 

 are the most important "annectant" insects of the groups under dis- 

 cussion. Of slightly less phylogenetic importance are the Dermaptera, 

 Phasmids and Isoptera, although they too furnish many valuable 

 clues as to the relationships of the groups, so that if the interrela- 

 tionships of the three superorders were expressed briefly in a formula, 

 the annectent insects of less importance would be placed in paren- 

 theses as follows. "Mantids (Isoptera) — Grylloblattids (Phasmids) 

 — Embiids (Dermaptera)." 



In opposition to Dr. Walker's view as to the close relationship of 



