CLASSIFICATION OF PALEOZOIC INSECTS. 321 



have no fact beyond the subsequent development of biting types into groups ordinally dis- 

 tinct from sucking types (a fact paralleled in wing structure), to show that from the struc- 

 ture of the mouth parts Eugereon should be ordinally separated from Protophasina. 



Whether the paleozoic relics which have been referred to Coleoptera should also be 

 grouped with the Palaeodictyoptera is another question. That coleopteriform insects then 

 existed is I think probable, both from the traces which are reasonably referred to borings 

 similar to those made by existing types, and by the present structural relationship of 

 Coleoptera to types whose predecessors are most plainly recognized among paleozoic 

 forms, i.e. other Heterometabola. Troxites — the single relic from the paleozoic referred 

 to Coleoptera — is an obscure object, and may, as Brongniart has suggested, be the fruit of 

 a, plant. It seems to me most probable, all things considered, that Coleoptera sprang from 

 such Palaeodictyoptera as were wood-borers throughout life, and which in paleozoic times 

 had no greater differentiation of structure between the front and hind wings than exist in 

 other Palaeodictyoptera. Such differentiation would be likely to arise from the preserva- 

 tion of favored races with such a habit ; while the inherent probability that all the 

 heterometabolous types had their already diverging stems in paleozoic times, coupled with 

 the entire absence from these rocks of any shards of beetles, which in later rocks are the 

 most readily and frequently preserved of all insect remains, renders the supposition the 

 more acceptable. 



If then, Troxites be a fruit, and the above hypothesis account for what are apparently 

 beetle borings in the older deposits, we have left one insect only, Phthanocoris, claimed to 

 come from paleozoic rocks, which shows any considerable sign of such differentiation in 

 structure as led to the existing distinction between the front and hind wings of heterome- 

 tabolous types, as we now know then). 



Another reason for the claim here urged, viz., that all paleozoic insects should be grouped 

 in one order, Palaeodictyoptera, is to be found in the fact that whenever any of the 

 special groups which it includes, whose distinct affinities to special modern types are easily 

 recognized, are compared with these types, they are found to possess characters which 

 distinguish them as a whole from them. My meaning here will be clear by reference to 

 my paper on paleozoic cockroaches ; these insects, though plainly cockroaches or the 

 ancestors of existing cockroaches, are nevertheless structurally distinct from the latter to 

 such a degree that it was necessary to recognize them as a separate group, Palaeoblattariae, 

 taxonomically equivalent to the entire modern group Blattariae. The passage from one 

 group to the other took place in early mesozoic times. 



The above view of Palaeodictyoptera then reduces itself to simply this : that hexapodous 

 insects were not ordinally differentiated until post-paleozoic time. The example we have 

 given above, however, sufficiently indicates the next step we must take, and that is to dis- 

 tinguish between groups which the historic development of insects shows were the 

 precursors of types ordinally distinct. This it is difficult to do on any other basis than that 

 of family-continuity. It is comparatively easy to see that the Palaeoblattariae were the 

 probable ancestors of Blattariae, Protophasmida the precursors of Phasmida, Palephemeri- 

 dae of Ephemeridae. and Ileineristina perhaps of Sialina; but from wing structure alone, 

 Palaeoblattariae (ancient Orthoptera) are as nearly allied to Palaeopterina (ancient Neurop- 

 tera) as they are to Protophasmida (other ancient Orthoptera). Our clew is through the 



