184 THE entomologist's record. 



a thread of silk as it crawls. If annoyed it wriggles. When the larva 

 is full fed, it remains for a certain time in its mine until it has lost the 

 broad green dorsal stripe. It then cuts a slit in the upper cuticle and 

 leaves its mine for ever. I was fortunate one morning in seeing a 

 larva, which had left its mine, crawl down the leaf stalk on to the twig 

 and from the twig to the branch, from the branch to the main stem 

 and so down the main stem to the ground. This was in September 

 and most of the. larva? of the second brood which came under ray 

 notice did likewise, or attempted to do so. On the other hand most 

 of the larvje of the early brood appeared to spin up on the leaflets. 



The cocoon and the method of its preparation are both highly 

 worthy of attention. It is one of those which have sometimes been 

 called hammock cocoons, as the real cocoon is not fastened to the 

 foundation on which it is built (as, for instance, that of Dicrunnra 

 viniila) but is swung between strands of silk. When the larva has 

 selected the situation in which to form its cocoon, it commences by 

 spinning a number of strands of silk, not around itself as most larvte 

 do, but only on two sides of itself. After a short time, the larva is 

 seen in the centre between two parallel walls of silk. It remains in 

 the centre while spinning, but faces one wall at a time, and 

 stretches its head to the right and left as it passes the strand 

 of silk from one to the other end of the wall. After a time 

 it will turn round and face the opposite wall, adding silk to 

 that in the same way. This action, with frequent rests, is 

 continued for five or six hours, when the walls are nearly completed. 

 In a well made cocoon the walls will be about 12mm. long and 2mm. 

 high and 1mm. apart at the top. These walls are not perpendicular. 

 One approaches that position and the other leans greatly towards the 

 opening between thera. The leaning wall is necessarily the larger and 

 on the top of this wall the larva spins a number of loose fluffy threads 

 which partly hang over the top of the wall. No such threads are spun 

 on the other wall. At the bottom between the walls the cocoon proper 

 is spun. This is spindle shaped and not much more than half the 

 length of the walls. It is frequently spun almost into the base of one 

 of the walls. Though the whereabouts of the pupa can be seen through 

 the cocoon, the latter is rather closely woven and moderately tough, 

 though soft. When these larvte spin up on the leaflets of the 

 laburnum, the walls of silk draw the leaflet into a curve, like the string 

 does a bow, so that a convenient space is formed in which the cocoon 

 proper can be spun. I have noticed that the cocoons spun on the 

 leaves in June are often rather poor structures compared with the 

 elaborate cocoons spun in Autumn. Pupation takes place three or four 

 days after the completion of the cocoon and the larval head-shell and 

 skin are usually thrust out of the cocoon proper. 



The pupa is a very singular one, very flat and covered with strise 

 or minute ridges, running over the limb-cases. The limbs come down 

 almost to the end of the body. The eye- caps at the bases of the 

 antennae are very conspicuous. There appears to be no cremaster for 

 anchoring the pupa to the cocoon ; I presume the transverse ridges on 

 the limb-cases help to retain the pupal shell in the cocoon when the 

 moth escapes, as the pupa-case is left in the cocoon when the moth 

 emerges. 



( 



