BY R. J. TILLYARD. 277 



it is clear that Comstock and Needhain's proposition(3), unsup- 

 ported, as far as I can see, by any evidence, cannot be accepted. 

 This proposition was to the effect that additional sectors of R 

 had been added, one after another, from the distal end, and had 

 thus progressed gradually basad. Now the whole study of Neu- 

 ropterous venation emphasises the fact that the tendency of 

 branches of R (and of M and Cu likewise) is to move progressively 

 distad, as we pass fi'om the older to the newer forms. Again, if 

 Comstock and Needham were correct, the old original Es, with 

 its many branches, would have to be the most basally placed, 

 whereas it is actually the most distally placed in all three 

 families where more than one sector exists. Thirdly, Comstock 

 and Needham's proposition would necessitate a recognition of 

 the smallest simplified Hemerobiida', with only two sectors, as 

 archaic types, from which, by progressive elaboration of the 

 venation, the forms with many sectors (such as Drepanepteryx, ' 

 Megalomus) have been built up; whereas it must be obvious, to 

 the most superficial student of the Order, that the very opposite 

 is the case. The only argument in favour of Comstock and 

 Needham's proposition is the fact that fossil Neuroptera, so far 

 as they are known, all have a single Rs in the forewing. But 

 the only fossils known, other than those of Tertiary age, are a 

 a small group of forms from the Upper Lias and Upper Jurassic 

 (the Prohemerohiidce of Handlirsch, together with one or two 

 other forms) which are clearly allied to our Osmylidce and 

 Psychopsidcf, and show already, in the Lias, a degree of special- 

 isation which places them very far from the beginning of the 

 Neuropteroid stock. Nobody would claim, I suppose, that such 

 an admittedly archaic group as the Neuroptera arose in the Lias, 

 or even in the Trias. It must have been already in existence 

 alongside the Carboniferous Protodonata, these latter being, in 

 fact, a very vigorous side-branch of the main stem, specialising in 

 the assumption of an aquatic larval life-history. Why, then, have 

 we so small a recoi'd of fossil Neuroptera ? The answer is obvious, 

 viz , that they have all along been essentially a non-aquatic 

 group, with a preference for dry climates. We cannot hope, 

 then, to find their record written completely in freshwater beds, 



22 



