160 DK. R. J. TILLYARD ON THE 
In interpreting the above scheme, we must compare the imaginal venation 
shown in text-fig. 10 with the tracheafion shown in text-fig. 9. There can 
then be no doubt as to the correct homologies of the main veins. Two points 
only are open to doubt. Firstly, what is the correct notation for the middle 
branch of the triad o£ E.2+3? This appears to be a branch of K3. But in the 
penultimate instar' o£ Oniscigaster distans, I find the trachea Rj supplying 
this branch. I therefore conclude that the tracheation'is variable, and give 
preference to the notation R^b for this vein. This is done because, in the 
fore-wings of recent Mayflies, and also in both wings of the fossil Protereisma, 
R3 is an unbranched vein. Secondly, what is the correct notation for the 
middle branch of the triad of M ? In this case, both in the larva of 
Oniscigaster distans and in the one figured, the trachea supplying it comes 
from M3_^4. In addition to this, the vein itself, in the imaginal wing, arises 
from M3_,_4 as a definite branch. All the evidence, then, points to the correct 
notation for this vein being M3, the other members of the triad being Mi_^2 
and M4. In this respect, we must reckon that the hind-wings of Ameletus 
and Oniscigaster differ' from their fore-wings, and also from both wings of 
Protereisma. The explanation probably lies in a shifting, at some fairlj' early 
point in the evolution of the hind-wing, of the point of attachment of the 
middle member of the triad, followed at a slightly later time by the 
development of a corresponding new tracheal supply. 
Alternative Interpretations. 
Having now considered all the evidence available under the three headings 
already indicated, we may ask whether any of them offer any basis for an 
alternative interpretation. The answer to this must be, that, in the case of 
the evidence from Convex and Concave Veins, and from the Fossil Record, 
there can be no doubt whatever as to the true homologies of the veins, and 
no alternatives to the solution here given can be offered. In the case of the 
larval tracheation, we have noted a considerable amount of variation both in 
the branch tracheal supply and in the mode of origin of the main tracheae 
from the alar trunk. To anyone, therefore, who pins his faith upon this kind 
of evidence alonej and who refuses to admit that offered under the two other 
headings, alternative solutions are possible. The two chief alternatives, of 
course, are those given in the table on p. 145 as the Comstock-Needham and 
Morgan notations respectively. We may ask whether the evidence from 
the tracheae alone does really su23port either of these two schemes rather than 
the one we have given in this paper. 
With regard to the original interpretation given by Comstockand Needham, 
this clearly assumed that there had been a basal fusion between Rs, M, and- 
Cu. The evidence of the tracheation certainly supports the idea of a basal 
fusion between Rs and M ; but it shows clearly enough that Cu remains 
quite distinct. It does not seem necessary to say more about this alternative 
