144 NATURAL HISTORY OF AQUATIC INSFXTS CH. 



the appendage itself, 15 the sheath of the appendage. 

 Now it becomes necessary to remark that Di will 

 give rise to the pupal respiratory organ, whether 

 tubular or tufted, further that Di may be in an early 

 stage somewhat wing-like, forming a thin, flat sheath. 

 What was its original purpose we cannot even con- 

 jecture. No Insect is known with a functional pro- 

 thoracic wing, and it is hard even to imagine an Insect 

 with three pairs of wings. The sheet-like prothoracic 

 appendage of the dorsal series may now curve round 

 until its edges approach and at last become united. 

 In this way the tubes or funnels which constitute the 

 pupal respiratory siphons of the Gnat,^ Corethra, &c. 

 are formed. Or the sheet may become supplied with 

 a network of large, branching tracheal tubes, which 

 repeatedly fork, as tracheal tubes in general do. If 

 we suppose that the intermediate substance, proving 

 superfluous, shrinks or is absorbed, the tubes will then 

 stand out as a tuft of forked filaments, exactly like 

 the pupal respiratory tufts of Chironomus plumosus, 

 Simulium, &c. I believe that this is the way in which 

 either a trumpet-shaped siphon or a tuft of branching 

 tubes is derived from a wing-like fold of skin. 



The fly of Chironomus is a common object upon 



^ Dr. C. H. Hurst, The Pupal Stage of Culex. Studies from 

 the Biological Laboratory of Owens College. This and the 

 same author's "Life-history of a Gnat" {Tratis. Manchester 

 Micro. Soc, 1890) contain much interesting information. 



I have adopted Dr. Hurst's view that " the first pair [of dorsal 

 appendages] become rolled up to form tubes, the respiratory 

 siphons, while the other two remain flat plates," as agreeing 

 with the facts hitherto observed. I do not, however, consider 

 that it has been definitely proved. 



