144 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I37 



An exception to these generalizations may perhaps be exemplified 

 in the Dermaptera, which are often looked on as orthopteroid insects. 

 In fact, the Dermaptera could be derived directly from our orthop- 

 teroid-psocopteroid composite, with its 17 spinasternal muscles, did 

 they not possess a isps-2ils that has not been found in any of the 

 orthopteroid orders that we have considered. Among the latter, 

 isps-fuo usually stands in place of the isps-2ils of the Dermaptera, and 

 it may be that the two elements are really homologues in this instance. 

 But, whether one is inclined to equate the isps-2ih of Dermaptera 

 with the isps-fn2 of other insects, or to accept its apparent equivalence 

 to the isps-2ils of Thysanura, Megaloptera, Coleoptera, and Neurop- 

 tera, in some of which both muscles occur together, it becomes obvi- 

 ous that the Dermaptera have struck out on a morphological path 

 that has been followed by no other orthopteroid group. At least 

 to this extent then, the Dermaptera have set themselves somewhat 

 aside from the main line of orthopteroid descent. 



Among the orders yet to be considered are those that are usually 

 referred to as neuropteroid, and of them we have for the Megaloptera, 

 Coleoptera, Neuroptera (s.s.), and Mecoptera sufficient data to war- 

 rant discussion. There has been some question as to whether the 

 Coleoptera should be included with these other groups, but the simi- 

 larities in the spinasternal musculature of larval dytiscids and Meg- 

 aloptera are decisive on this point. Adult Coleoptera, however, have 

 progressed to such a degree of thoracic specialization that at most 

 two spinasternal muscles are left, and the majority of beetles are thus 

 without diagnostic value in the present context. 



As already indicated, the neuropteroid orders, as I shall continue 

 to call them, make up a somewhat looser aggregate than we have seen 

 in the orthopteroid groups. The Megaloptera, Coleoptera, and Neu- 

 roptera, which are the more primitive among them, agree in lacking 

 the cruciate cervicocoxal muscle, icv-cxiX, which is characteristically 

 present in equally primitive orthopteroid orders ; but this same muscle 

 comes to light again in the Mecoptera, which have but seven spina- 

 sternal muscles in all. Thus we must choose either to exclude the 

 Mecoptera from the neuropteroid series, or else to postulate that 

 icv-cxiX was present at the base of the stem from which the Mecop- 

 tera were derived after the branches leading to the Megaloptera, 

 etc., had been given off. The second of these alternatives seems to 

 accord better with other data and is therefore the one adopted here, 

 although it implies that the stock that produced the Mecoptera has 

 been distinct from the other neuropteroid lines for a very long time. 



