40 



It is well developed in Antherea pernyi and in Saturnia pyri, less 

 so in our carpi?u, specimens of which I have selected to illustrate 

 this structure. In none of these is it useful as a cutter, though it 

 comes to a rather sharp edge in pyri, at least, no observations have 

 been recorded, and the empty cocoons \n pyri and carpini show that 

 the provision made by the caterpillar is sufficient for the emergence 

 of the moth, as well also as in pernyi, though of a different character, 

 the softening of the cocoon by a fluid allowing the threads to be 

 pushed aside in such a way as to show that they are spun each to 

 its own side without crossing, just as by carpini or pyri, although 

 the full cocoon looks precisely the same all over, just like that of 

 Lasiocampa qjiercfis, or Limacodes. 



In these cases, then, the "cocoon-opener" does not open the 

 cocoon, but it may very well act as an efficient lever to assist the 

 moth in climbing out of the cocoon. I have noticed in pyri, pernyi, 

 and lima that there is on the side of the thorax against which the 

 "cocoon-opener" impinges when the wings are closed, a grove to 

 receive the opener, with an overhanging flange. The groove is 

 obviously a sheath to receive the rather formidable weapon, and 

 prevent the thorax being injured, but I have also a suspicion that 

 the flange locks on the opener in the closed position of the wings, 

 and helps to hold them in the closed position. Experiments with a 

 living specimen are necessary to determine whether this is so, but 

 such as were possible vc\ pyri by relaxing the specimen were rather 

 confirmatory. 



Not only the " cocoon-opener " but the hollow and ridge on the 

 thorax that correspond to it exist in many other species if not in 

 all, but have no special development, showing that the "cocoon- 

 opener" and its sheath are not new structures, but parts of the 

 ordinary arrangements specially elaborated and developed. 



In the "Transactions of the Linnean Society" (May, 1891) Professor 

 Poulton called attention to a pupal structure, which for that reason 

 I have called " Poulton's line." 



The actual matter that Professor Poulton concerned himself with 

 in that paper in regard to this line seems to me to be treated with 

 perfect correctness, and in fact his idea that the pupal traces of the 

 imaginal wing are less specialised than the wing of the imago itself, 

 if not fully proved, is at least fully and soundly supported by his 

 facts. 



It is nevertheless the case that his starting-point has always 

 seemed to me to be m some way erroneous, and I have neglected no 

 opportunity of observing any facts bearing on the subject. 



Poulton's line is a line that may be observed on almost any pupa, 

 though much more easily on some than others. It runs round the 

 inner and hind margin of the pupal wing, at some little distance 

 from its margin. "What is it ? Why has the pupal wing a double 

 margin, and what do they severally represent in the imaginal wing, 

 in its development and evolution ? Professor Poulton in the paper 



