with fragments of gnawed wood." He does not tell us where this 

 cocoon is placed, but the fact of it being made of " silk mixed with 

 fragments of gnawed wood " evidently implies that it is within the 

 tree. But in an editorial note published some ten years later 

 (" Entom.," vi, p. 487) he says, "I have found this larva under- 

 ground in a cocoon formed of silk and earth, without a particle of 

 its home being made of sawdust." Wilson ("The Larvae of the 

 British Lepidoptera," p. 45) says, "Pupa: generally within the 

 tree, but it has been found occasionally in a cocoon in the earth." 

 Barrett ("The Lepidoptera of the British Islands," vol. ii, p. 149) 

 describes the method of pupation, thus : — " In a tough, oval cocoon 

 of silk and raspings of wood or earthy particles, being usually the 

 hybernaculum wathin which the full-grown larva has passed the 

 winter, and which it does not voluntarily leave, though the pupa 

 state is not assumed until June or July." 



Meyrick (" Handbook of British Lepidoptera," p. 560) contents 

 himself with the simple statement, "Pupa subterranean;" while Tutt's 

 ("The Common British Moths of England," p. 339) account — "In 

 the autumn, when full-fed, it enters the ground and spins a cocoon in 

 which it winters ; this it leaves in the spring, spins another cocoon 

 almost directly, in which it changes to a pupa, and emerges in about 

 three or four weeks " — appears to be to the same effect. 



In reviewing these records in their chronological order one is 

 brought face to face with the fact that whereas the earlier authors 

 regarded the pupation as invariably taking place within the tree, and 

 are at some pains to point out that the structure of the pupa is 

 singularly well suited to such a situation, some of the most recent 

 are equally assured that pupation takes place only in the earth. 



It is difficult to reconcile two methods so diametrically opposed 

 to each other, and one is almost led to ask whether it is possible 

 that the habits of the species have changed during the period men- 

 tioned. I do not, however, think there is any evidence to support 

 such an hypothesis, and I am confirmed in this opinion by sundry 

 cases that have come under my personal observation. 



One of my earliest experiences was with a badly, indeed very 

 badly, infested willow tree that stood in what is now the Hilly Fields 

 Recreation Ground at Lewisham. In the spring or early summer of 

 a year now long passed, this tree was either felled or blown down, 

 and the rotten trunk afforded a rich harvest of Cossi/s pupas. They 

 were certainly in the wood of the tree, but whether in the tunnels in 

 which the larvae had fed or not I am unable to say with certainty, 

 but my impression is that they were not, as the portion containing 

 them was very rotten, — in fact, a mass of touchwood, a condition that 

 would be unsuitable for sustaining the life of the larvae. 



During recent years I have frequently picked up straying larvje 

 i/i. autumn, and holed them in a poplar tree that grows in my garden 

 by boring a deep hole with an auger, inserting the larvce and plugging 

 the entrance with a thin piece of cork. The first larva experimented 



