44 



stage do not correspond to these outlined veins, and that the latter 

 correspond more nearly to the veins of the mature wing. This is 

 so. The tracheae at first have a very generalised disposition, and 

 extend beyond Poulton's line to the wing limit. As development 

 goes on, the tissues between them diminish or increase in relative 

 extent, so that the tracheae, from being nearly straight and parallel, 

 become bent or curved, approach one another or fall more widely 

 apart, and thus take up gradually the positions of the longitudinal 

 veins of the wing. This is a fair description of what occurs in all 

 cases, though to be complete it would have to refer to veins forming 

 where there are no trachecX ; to the coalescence or actual disappear- 

 ance of the latter in many cases. Whilst the wing is still in a 

 somewhat early stage, and before scales are fully formed, each 

 persistent trachea becomes furnished near its extremity with two 

 lateral branches that meet those of the next trachea, and form a 

 continuous line of tracheal tube completely round the wing. These 

 are formed by the expansion of some of the finer tracheae with which 

 the wing is largely permeated. At the same time the wing, beyond 

 the line of marginal tracheae, which is to the developing wing in 

 precisely the same relation as Poulton's line to the pupal surface, 

 contracts just as the wing tissue does between two tracheae that are 

 approaching each other, the tracheae it contains disappear, and it 

 becomes represented by a line of wing tissue outside the marginal 

 trachea. This portion of the wing is much narrower proportionally, 

 than the area outside Poulton's line in those cases in which, from 

 exigencies of pupal form or other causes, it is widely expanded. It 

 is, however, fairly related as regards extent to that area in the larger 

 proportion of species, those, viz., in which Poulton's line occurs close 

 to the margin of the pupal wing. There can be little doubt that the 

 hypodermis covering this wing margin is that which lined the 

 chitinous covering of the wing in the pupa, which has no doubt 

 undergone many changes, but has preserved throughout structural 

 continuity. The only developmental interpretation that I can give 

 to the picture suggested by Professor Poulton would be that at 

 Poulton's line the upper and lower hypodermis met by a kind of 

 cicatrising process, cutting off the tissues beyond Poulton's line, and 

 that these were afterwards eaten or digested by the developing cells 

 of the cilia. Besides its inherent, might I say, absurdity, this would 

 not explain the fact that when the wing has developed, this space 

 contains only fluid, even in those cases in which the wing retreats 

 basally from Poulton's line instead of going beyond it, both these 

 cases occurring in different instances. 



That a marginal nervure is a normal constituent of all Lepidopterous 

 wings is proved, in the first place, by the marginal tracheae I have 

 already mentioned as occurring during the development of the wing, 

 and often to be discovered in the perfect wing, by the structure in 

 many imagines, when it exists as a hollow tube, and further, by the 



