90 University of California Publications. [Entomologt 



nothing in the other parts of the body to suggest such a rela- 

 tionship. 



Some members of the family possess a venation that could 

 be compared rather closely with that of Cicada, but here, like- 

 wise, there is the entire absence of all the characteristics of a 

 wing of that group, and the body characters show that here 

 there is still less ground for comparison. 



There is a very evident tendency to specialize one or two 

 rows of cross veins parallel with the hind margin of the wing. 

 The same tendency is seen to some extent in the Sialidte and 

 very distinctly in the Hemerobida^. The Sialidse is a compar- 

 atively undifferentiated group, so that its venation is not very 

 distinct from that of the Megascopterida", but there are some 

 characters which will at once separate them. The Megascop- 

 teridje have a rather more consistent venation than the Sialidee, 

 and in the specializations of the cross veins are also dis- 

 tinctly in advance of that group. As to the Hemerobidse, 

 it is much easier to homologize their venation with the Sia]id£e 

 than with these insects. We must therefore concur in the 

 separation of the Megascopteriche from all the existing groups. 



Their real relationship I conceive to be with the Odonata, 

 Protodonata, and Ephemeridte. All the fossil remains we have 

 of these insects show them with their wings spread, and the 

 appearance of the base of the wings would indicate that they 

 were not provided with means of folding. The structure of 

 the notum of the thorax, moreover, as far as it can be made 

 out, lends further support to this idea. 



If this supposition is correct, the Megascopteridae and Pro- 

 todonata represent two extremes in the development of this 

 primitive type, for there is little similarity between the two 

 groups. A group, doubtless derived from the latter family, 

 has in later time gone through a process of reduction, in which 

 the cross veins suffered to quite as great an extent as occurred 

 in the Megascopterida?, and produced the ^f//-io» of the present 

 day. In all respects ^(/rio?i is much in advance of the ancient 

 Megascopteridae. 



PROTODONATA. 



There can be no doubt of the correct identification of the 

 insect remains that have been placed in this group as being 

 allies of the Odonata. The wings possess neither stigma, 

 nodus, nodal sector, arculus, nor triangle. But this list of 



