452 Dr. R. J. TiUyard on 



is more higlily specialised than the typical Mecopterous 

 larva in all points, except only in the form of the antennae. 

 The Mecopterous larvae of the genera CJiorista and Panorpa 

 have a more generalised form, and chaetotaxy, a head 

 which, while definitely retractile, is not so reduced or so 

 regularly concealed as in Sabatinca, compound eyes with 

 from forty to more than fifty facets, and much more 

 general ground-feeding habits. Also these larvae do not 

 form cocoons, but pupate free in the earth, or in a hardened 

 earthen cell, the result being that the pupa is more normally 

 elongate, and not so much curved round upon itself. The 

 general chara.cters of the papa are, apart from this, almost 

 entirely archaic, and indicate that it has persisted un- 

 changed, except for inclusion within the cocoon, from the 

 earliest times in M^hicli Holometabola existed. 



The above conclusions agree well with what we know at 

 the present time of the ancestry of the Lepidoptera from 

 palaeontology. The fossil Behnonlia, from the Upper 

 Permian of Belmont, N.S.AV., has been claimed by me to 

 represent the ancestral type from which both Trichoptera 

 and Lepidoptera were derived.* Recently a related fossil, 

 Parabelmontia, has been discovered in the same beds, and 

 differs chiefly from Belmontia in having the first cubitus of 

 the wing simple instead of forked. t This new fossil, 

 therefore, completes the record of the ancestry of the 

 Diptera, through the Triassic Paratrichoptera backwards, 

 imtil it merges into the same fossil Order Paramecoptera 

 in which the common ancestor of the Trichoptera and 

 Lepidoptera is to be found. In forming the new Order 

 Paramecoptera for the fossil Belmontia, I pointed out 

 how this type was closely similar to the older Mecopterous 

 type of wing, but differed from it in important respects, 

 which showed that it was not a direct oft'shoot of the Order 

 Mecoptera as it existed in the Upper Permian, and as it 

 still exists almost unchanged at the present day. Thus we 

 now see that the common stem of the three closely related 

 Orders Trichoptera, Lepidoptera and Diptera is to be found 

 in the Upper Permian within the Order Paramecoptera, 

 the Diptera lying closer to the Mecoptera than do the other 

 two. If these conclusions be correct, we should not be 

 surprised to meet with some existing archaic Lepidopteron 

 which showed distinct Mecopterous affinities in its larva and 



* Proc. Linn. Soc. N.S.W., 1919, xliv, pt. 2, pp. 231-256. 



t Ibid., 1922, xlvii, pt. 3, pp. 284-287, pi. xxxiii, f. 2, text f. 3. 



