NEURATION IN THE LEPIDOPTERA. 155 



the ancestral median 1 of the fore-wing ; not the specialist in 

 Sphingides and one or two other near super-families. 



In the super-family Micropterygides, the wings are of similar 

 shape, are narrow, and taper to a blunt tip, having, therefore, only 

 two margins (upper and lower), and the jugum (Plate I., fig. 1, J). 

 The neuration of all the wings is similar. It exhibits no discoidal cell 

 as such, and the hind-wings have a greater number of nervules than 

 any other lepidopterous group, except the Hepialides, showing a close 

 affinity between this group and the Trichoptera (see also Dr. Chapman's 

 " Evolution of the Lepidopterous Pupa," Eut. Ecc, vol. v., p. 25). 



The members of the super-family Hepialides have a jugum. The 

 wings are all of similar shape, but are broader at the extremity than 

 those of the Micropterygides, forming another margin (the outer). The 

 neuration is modified by the formation of the discoidal cell (Plate I., 

 Fig. 1, Hi'pialm hioiinli). 



Some individual amongst the earlier Hepialides — which probably 

 more closely resembled the Micropterygides — developed the frenulum 

 in place of the jugum (Plate I., fig. 5, F). As all the other groups 

 possess the frenulum — except some of the higher, which develop an 

 extension of the hind-wing base, and the Rhopalocera — they seem to 

 have evolved from a common ancestor. 



Those families which, in shape of wing, neuration, etc., most widely 

 differ from the primitive ancestor (which, as already inferred, had 

 narrow wings, and all of similar shape) logically seem to be the higher 

 families. 



Thus of the groups possessing the frenulum, the super-family 

 TiNEiDES (genus Tinea and its allies) has probably the most primitive 

 wing-shape and neuration ; how unnatural the old group Tineina was, 

 neuration — on the argument that modification tends to reduce the 

 number of nervules — clearly shows. Lyonetidae (fig. 2) is very highly 

 specialised, whilst Elachistidcu' also takes a high position. 



The super-family Pterophorides has also a highly specialised wing- 

 shape. 



In the super-family Tortricides (including Tortrix, Cossiis, etc.) the 

 family Cossid.e has a high wing-shape, but the neuration places it as 

 a low group, whilst Zeuzera pi/rina (Plate I., fig. 8) belongs to an 

 earlier super-family (Zeuzerides) than that to which Cossus lif/niperda 

 (fig. 4) belongs. This is evident from the alteration of position of 

 the median nervules of the fore and hind-wings, and other details, as 

 well as the somewhat broader wings of the latter. The super-families, 

 which are more highly specialised than the ( 'ossidac, are the Sphingides, 

 BoMBYCiDEs, NocTuiDEs, Geometrides and intermediate families. Thus 

 Stauropu^ fa;/i (Plate I., fig. 6) represents the Pseudo-Bombycides, 

 and is intermediate — although highly specialised in larval characters 

 — between Tortricides {Cosmlac) and Sphingides, the members of the 

 latter, by the way, appear to have two clearly defined sub-families, 

 Spliimiidi and Snierinthidi — illustrated by Sphinx Uijmtri (Plate I., 

 fig. 7) and Swerinthm jxijudi (Plate I., fig. 8). 



In removing the wing scales from species of Rhopalocera, I was 

 struck by the fact that they (the scales) appear to be more firmly con- 

 nected to the membrane than are the scales of any other group. And 

 although the neuration superficially — in fact, closely — resembles that 

 of the BoMBYciDES, near which the Rhopalocera is placed by Corastock, 



