(AZO NEW YORK STATE MUSEUM 
Odonata,’ I came to the conclusion that our hypothetical type 
wing did not represent all the venation of the primitive insect 
wing, but only the main skeleton of it. That the longitudinal veins 
of that type were possessed by the primitive insect I do not doubt: 
they represent the main lines of chitin deposition along primeval 
tracheae; but the interspaces between these veins were occupied, 
I believe, by a more or less irregular meshwork of cross veins, 
which disappear with the progressive differentiation between strong 
veins and thin membrane. Redundant cross veins are still char- 
acteristic of many generalized insects, and were so of most of the 
older fossils known. I have given in the paper just cited [p. 725- 
28] a theory as to the mode of differentiation of strong cross veins’ 
in the dragon flies. 
ditere= iss inuchesless 
evidence as to how 
the reduction may 
have occurred in the 
Diptera; but I have 
no doubt that the 
supernumerary cross veins and spurs of veins, so common in 
Tipulidae, indicate the location of some few remnants of the large 
numbers that were probably possessed by the early neurop- 
teroid ancestors of the Diptera. It may be assumed that in any 
process of reduction cross veins favorably situated, joining the 
principal veins advantageously, would tend to grow stronger, while 
others, less favorably situated in intervening spaces, would tend 
to weaken and disappear. 
I have drawn and present in figure 12 a typical Tipulid wing 
in which the principal veins with their full complement of branches 
are represented in solid black, and the typical cross veins are 
represented in double contours. ‘This wing is based on a tracing 
of the wing of Macrochile [pl. 14, fig. 1] and differs very little 
therefrom. Then, in order to see what sort of wing it would be if 
all the supernumeraries occurring anywhere in any crane fly should 
appear together, I located these supernumeraries, all in their: 
proper places, one by one, and I represent them then in dotted 
lines in this figure. How like a Panorpid wing is the result! If 
one compares it with the wing of Bittacus, for example, He will 
see that the differences are very slight, and are confined chiefly 
to the anal area. ‘There is the same type of branching of all the 
1U. S. Nat. Mus. Proc. 1903. 26:703-64. A Genealogic Study of Dragon 
Fly Wing Venation. 
