252 THE entomologist's record. 



hairs, placed one in front of the other (not side by side). 

 There are three circumspiracular tubercles, of which the post- 

 spiracular is small on 3rd and following segments, a marginal 

 tubercle to the 7th, but not after. The central tubercles of 13 

 are conjoined, and the anal plate has eight hairs. When full- 

 grown in this skin, the larva is greenish-grey, with yellow 

 dorsal and lateral lines, and is in fact already a miniature of 

 the full-grown larva. 



The second skin differs from the first, in the yellow bands 

 being free from the minute black points or bristles, which give 

 a smoky look to the white (green ?) portions. The head has 

 sundry pale markings. The plate of the 2nd segment is 

 divided into two portions, each with four hairs. The 3rd seg- 

 ment is decidedly the largest, then the 4th. The yellow band 

 broadens on 3, so as to include the dorsal tubercles, and is 

 nearly evanescent on 4. 



In the third skin the larva is at first deep blue-black and 

 yellow, but as it grows, the blue becomes a pale violet blue, 

 and the yellow, clearer and purer. The marginal tubercles 

 now present a complete series, and there is a pre-spiracular 

 tubercle just to be detected. Each lateral plate on 2nd seg- 

 ment has now broken up into four tubercles, of which the two 

 posterior are still conjoined. The yellow forms a broad 

 transverse dorsal stripe on the 3rd, including the dorsal 

 tubercles, and forming a hump. The 12th is also raised ; in 

 these and other respects the larva is almost identical with the 

 ftill-grown larva. 



The cocoon is 22 mm. by 8 mm,, fairly cylindrical, and 

 tapering at each end, but varying a good deal in size and out- 

 line, according to its position. In captivity, it is formed in an 

 angle of the top or bottom of the cage, or more frequently on 

 a stem of the foodplant, especially at a fork. In such a situa- 

 tion it pulls a leaf or two together in which to spin, but instead 

 of curling them round the cocoon, it chops them up into frag- 

 ments which it incorporates with the silk of the cocoon. It 

 will also accept a surface of sawdust and cover the cocoon 

 densely with this. The material is a white silk, afterwards 

 becoming rather darker, apt in a cocoon formed without much 

 extraneous material to be rather brittle. On several occasions 

 I have found the cocoon on the trunk of an apple tree, and 

 except that it was more rounded and prominent and, of course, 

 softer, it much resembled a Ceriira cocoon, the lichen of the 

 surface of the back being most cleverly worked into the cocoon 



