BY R. J. TILLVARD. 641 



disappeared from the membrane, while tliose left upon the main 

 veins have become short, stont spines, resembling those found in 

 many Diptera. 



As the true cross-veins are always weakly chitinised, and 

 entirely free from macrotrichia in this Order, we have, in the 

 macrotrichia, a valuable aid in the tracing out of the courses of 

 the main veins. In many genera, the extreme base of Cuj is 

 sharply bent up to join M not far from its origin, and so takes 

 on the appearance of a cross-vein. But the true nature of this 

 vein is still proclaimed by the pi'esence of well-developed macro- 

 trichia upon it. 



To sum up, then, we may characterise the Archetype of this 

 Order as differing from that of the Protomecoptera only in the 

 complete loss of the archedictyon. While the microtrichia remain 

 constant throughout the course of evolution, the macrotrichia 

 undergo a gradual reduction in size and number. 



Order DIPTERA. 



(Plate Ixviii., fig. 12; Plate Ixix., fig. 14, and Text-figs. 26-29). 



We take this Order next, since it appears to be, on the evidence 

 of the wing-trichiation, a direct evolutionary derivative from the 

 base of the Order Mecoptera. 



Throughout the Order, microtrichia are present upon the wing. 

 Macrotrichia are to be found upon the main veins and their 

 branches; they tend to become reduced both in size and number. 

 In only one family that I have examined are macrotrichia to be 

 found upon the membrane, viz., in the Rhyphidcp (Plate Ixviii., 

 fig. 12, and Text-fig. 26). In this family, the structure and size 

 of the macrotrichia are closely similar to those of the older 

 Mecoptera. The arrangement of the macrotrichia upon the wing- 

 membrane, however, shows the pattern of the archedictyon much 

 more definitely than in any existing Mecopteron; so that, on this 

 character, the Rhyphidce can only be derived from a very early 

 form of Mecopteron, if they are to be derived from that Order 

 at all. In this connection, it is interesting to note that the 

 family RhypJiidct combines within itself characters of both the 

 Nemocera and Brachycera. This accords well with the condition 



