E. H. Sellards — Types of Permian Insects. 257 



features, can scarcely be said to exist. To this extent they 

 approach the Permian type.* 



The bridge. — To the supplementary longitudinal vein inter- 

 polated between M^ and M 3 and connecting with the radial 

 sector, Needham has applied the term bridge. The bridge 

 appears as a normal feature in Tupus. A part of the bridge 

 is seen also in Paralogus. Unfortunately the tip of the wing 

 of that genus is not preserved. The crossing of the radial sec- 

 tor has doubtless promoted the early development of this 

 structure. 



Classification. 



In the scheme of classification of Paleozoic insects proposed 

 by Professor Samuel H. Scudder, the Paleozoic dragon-flies 

 were disassociated from their modern descendants and placed 

 along with all other Paleozoic insects in the Paleodictyoptera. 

 Brongniart (Recherches des Insectes fossiles) places the Proto- 

 donata as a family group under the large order of Neuroptera. 

 Professor Anton Handlirsch, in his recent publication (Revision 

 of American Paleozoic Insects) has advanced the Protodonata 

 to ordinal rank to stand as a Paleozoic order coordinate with 

 the Odonata of Mesozoic and modern times. 



It will be remembered that Professor Scudder, in defense of 

 his general plan of classification, has urged repeatedly that the 

 insects of the Paleozoic were more closely related inter se than 

 to their descendants of the Mesozoic and later times. A classi- 

 fication based upon Scudder's principle must, in the nature of 

 the case, break down with the advance of knowledge of extinct 

 forms, for as the life history of the various lines of descent is 

 continuous, it necessarily follows that the artificially placed 

 dividing line between the Mesozoic and Paleozoic can remain 

 a distinct break only so long as there remains time for a con- 

 siderable development between the earliest known Mesozoic 

 and the latest known Paleozoic representative of that particu- 

 lar line. In Scudder's classification emphasis is thrown on the 

 interrelation of associated but diverging groups, or what may 

 be called the lateral relation of organisms, rather than on the 

 lineal or phylogenetic relations. In this connection I have 

 urged a principle of classification* by no means new in its 

 application, as follows : — " Any natural group of organisms 

 should be recognized as extending back in time until a point 

 is reached at which that group coalesces with a group or 

 groups of coordinate rank, or unites with the parent stock." 

 The principle proposed by Scudder necessitates breaking phyla, 



* As a genus with cubitus unexpectedly similar to that of Tupus, compare 

 Pseudophcea. 



f This Journal, vol. xviii, p. 121, 1904. 



