272 STUDIES IN AUSTRALIAN NEUROPTKRA, IV., 



larval habit. Thus we have, as semi-aquatic offshoots, the 

 Liassic and Jurassic Proheynerobiidce, the recent Osmylidce. ( some 

 semi-aquatic, some terrestrial), and the highly-reduced Suyridce, 

 whose larvae dwell in freshwater sponges. From the first of 

 these, our Australian Psycliopsidce undoubtedly arose, by a 

 unique specialisation of the wing-venation, and by the adoption 

 of a larval habit closely resembling that of the Raphidians in 

 the Northern Hemisphere. From the Osmylidce, a small, terres- 

 trial side-branch, the Nymphidce, favoured by an exceptionally 

 fortunate larval development, started out on the upward patVi 

 which led to the great dominant groups of the Myrmeleontidce 

 and Ascalaphidce. A small side-branch of the Prohemer-obiidce, 

 the Jurassic Mesochrysopidce, probably gave rise to the modern 

 Apoclirysidce and Chrysopidce. The highly-reduced remnant of 

 the main stem, after throwing off the aquatic Sisyridce, reached 

 the extreme limits of reduction in the Coniopterygidce. The 

 positions of the Nemopteridce and Mantispidce are uncertain, but 

 the former are probably allied to the Nymjjhidce, the latter to the 

 Chrysopidce. 



We see, then, that of the five families dealt with in this paper, 

 only two are of the true Hemerobioid stock, while the other three 

 possess Osmyloid affinities. The resemblance between Hemero- 

 biidce and Sisyridce is purely due to convergent reduction, both 

 these families being specialised in comparison with the other 

 three, though generalised enough when contrasted with the more 

 dominant groups, such as the Clirysopidce. Of the three most 

 ancient families, the Ithonidce appear to possess traces of Sialoid 

 affinities, while the Bervth^dce and Trichomatidai must not only 

 stand very close to the base of the Neuropteroid stem, but may 

 even lie not far off from the more specialised line which led to 

 the Micropterygidce in the Lepidoptera. Lepidopterists cannot, 

 indeed, afford to ignore the growing importance of the Neurop- 

 tera in helping them to a true view of the phylogeny of their 

 Order. In this connection, the scales on the wings of Berothidce, 

 the fringe of long hairs in this family and in the TnchovicUidce, 

 the very obvious resemblance between the venation of the 

 Mici opteryy'dct and the most highly reduced Neuroptera {Coniop- 



