The origin and development of the wings of Coleoi^tera. 531 



are also numerous. Where the wing fundament has reached its 

 maximum invagination, which rarely exceeds two or three times the 

 thickness of the hypodermis, the dorsal portion of the invaginated 

 tissue thickens slightly and is evaginated downwards between the 

 hypodermis and the cuticula, thus forming the beginning of the larval 

 wing (PI. 17, Figs. 28—31). 



This development takes place entirely in the last larval stage 

 and produces a larval wing lying between the somatic hypodermis and 

 the cuticula without a trace of a wing sac (peripodal sac) and only 

 enough invagination to allow the wing to be formed beneath the cuti- 

 cula. The wing is thus from the start outside of the body and differs 

 in that respect from the usual recorded type of Holometabolic insect 

 wing development. The evagination of the wing is evidently a rapid 

 process, for in no case have I been able to obtain preparations in this stage, 

 although stages just before or just after evagination are easily obtained. 



With the evagination of the wing the embryonic cells increase 

 in number and become aggregated into large groups or masses 

 irregularly arranged near the opening of the wing cavity. There, 

 is, however, a considerable variation in the number of these cells 

 in different individuals (PI. 17, Fig. 30 ce). In the corresponding 

 stage of Chrysohothris femorata (PI. 18, Fig. 45) and in Buprestis 

 rufipes there is a dense mass of cells surrounding the tracheal trunk 

 and closing the opening to the wing cavity. Some of these may arise 

 from the "embryonic cells", but the larger part are surely due to the 

 proliferation of the hypotrichal membrane of the tracheal trunk. 



The further development of the wing of Phymatodes in the larva 

 consists in the continued downward growth of the wing fold between 

 the body and the cuticula. In Figs. 7 — 10 (PI. 14) are shown sur- 

 face views of the developing wing of P. variabilis made by graphic 

 reconstruction from serial sections. Fig. 7 (PI. 14) shows the first 

 invagination in the same stage as is shown in Fig. 26 (PI. 17) in 

 transverse section. In Fig. 8 (PI. 14) the evagination of the wing 

 has already produced a well marked wing fold. This same stage is 

 also shown in Fig. 30 (PI. 17) in transverse section. In Figs. 9 and 

 10 (PI. 14) and Figs. 27—29 (PI. 17) are shown later stages in sur- 

 face view and in transverse section. The development of the wings 

 in Phymatodes is interesting and shows the simplest condition of wing 

 development yet described. 



In addition to Phymatodes other Cerambycidae have the wings 

 developing in the same way: Prionus, Calladium^ and several larvae 



34* 



