522 W. L. TOWER, 



upon each segment and contains from three to ten large gland 

 cells. 



According to the observations of Wheeler (1889) upon L. decem- 

 lineata^ (and my own substantiate his), every segment of the thorax 

 and abdomen develops in embryonic life a well marked spiracular in- 

 vagination. This, I find, arises as a thickened area of the hypodermis 

 (ectoderm) which later becomes invaginated to form the tracheal tubes. 

 According to Wheeler (1889) the prothoracic spiracle degenerates, the 

 mesothoracic spiracle migrates anteriorly and becomes the functional 

 prothoracic spiracle and the metathoracic spiracle also degenerates. 

 My observations also confirm this, but in the migration of the spiracle 

 from the side of the mesothorax to its larval position, the spiracle 

 alone migrates and the thickened area of hypodermis about it remains 

 and probably becomes the fundament of the elytron. 



The tracheal invaginations ramify anteriorly and posteriorly and 

 the main branches of these fuse to form the longitudinal tracheal 

 trunks, after which the openings in the prothorax and metathorax 

 are rapidly cut off, leaving a disc shaped mass of cells which have a 

 somewhat concentric arrangement. The further stages in the degener- 

 ation of these spiracles I have not been able to observe. That this 

 rudiment of the spiracle is converted into the imaginai disc of the 

 wing seems probable, however, for the following reasons: 1) The disc 

 of the wing always appears in exactly the same area as that in which 

 the spiracle arose and de<^enerated. 2) The wing disc frequently 

 shows a concentric arrangement of the cells in early stages, but loses 

 this before invagination to form the wing begins. 3) If the wing fun- 

 dament is not derived from the remains of the spiracle, then, since 

 the wing disc has the exact position occupied by the spiracle, the 

 latter must entirely degenerate and be replaced by new hypodermis, 

 and from this the wing disc must arise. There is, however, absolutely 

 no ground for belief in such a process and the only conclusion that 

 seems at all tenable, is that the wing fundament is derived directly 

 from the remains of the spiracle. This agrees with "Verson's (1890) 

 conclusions. 



The discs are as a rule better developed in the mature embryo 

 than in the young larva, and it seems certain that they must be 

 formed in all of the embryos of the same species at the same stage 

 of development. The existence of the discs well developed in the 

 embryo and less so in the larva indicates either a degeneration or a 

 subsidence into a quiescent period. Furthermore, the fact that this 



