540 THE PANORPOID COMPLEX, iii., 



Veins. Text-fig. 35 shows, diagrammatic-ally, the condition of 

 the veins when viewed by a cut taken somewhat obliquely across 

 the wing, (a) not far from the base, so as to include only the 

 main stems of the main veins, and (b) beyond the first forkings 

 of R and M 1 _ 4 .. Rs, like R 1( is convex near its origin. Fur- 

 ther distad, there is a tendency for the branches of M and R to 

 arrange themselves at very much the same levels on a plane 

 surface, so that the part of the wing lying between the two very 

 strongly convex veins R x and Cu x is often nearly flat. However, 

 in many types, it can be seen that there are actual differences 

 of level, as in Text-fig. 35 b, right to the apex of the wing. 



The general rule for the insect wing is that concave and convex 

 veins follow one another alternately. Consequently, if two con- 

 vex veins are seen next to one another, a strong presumption 

 arises that a concave vein has become suppressed between them. 

 Thus, for example, in the higher Lepidoptera, two very strong 

 convex veins (Rs and Cuj) form the anterior and posterior 

 borders of the closed basal celh This fact should at once sug- 

 gest that the concave vein M has been suppressed between them; 

 and an examination of the pupal tracheation proves this to be 

 the case. 



A more difficult problem, of which no solution has been at- 

 tempted before, lies in the fact that, in the Panorpoid type of 

 wing, veins M 3+4 and Ci^ are both convex, yet lie next to one 

 another. There is thus a strong antecedent probability that a 

 missing branch of M, viz. M 5 has been suppressed between 

 them. The working out of this problem is shown in this paper, 

 and offers a very interesting example of how the four types of 

 evidence here arranged under the headings A to D may be used 

 together, to furnish a satisfactory proof, which could not be 

 supplied by anything less than all four of them. 



There is also the question of the anal veins, which, as shown 

 in Text-fig. 35 a, are all convex, though sometimes separated by 

 depressions or furrows. The suggestion that they may be, after 

 all, only so many branches of a single convex vein, as in the 

 Odonata, is worth consideration, and has been studied in this 

 paper, in the light of the evidence afforded by the pupal trachea- 

 tion. 



(E) Atavisms, or Beversions to Type. It sometimes happens 

 that a single individual of a species, here and there, reproduces 

 a more archaic condition in its venation than is usual for the 



