312 BRITISH LEPIDOPTERA. 



but, later, they enlarge and fuse, producing the appearance of a white 

 band, upon which is a single row of tubercles, each emitting a 

 minute hair. These tubercles are the original shagreen-dots, and 

 each is placed in the centre of a white area, which has spread from 

 the base of the former into the ground-colour. The areas form the 

 white stripe, and they can be readily distinguished, as each is 

 situated upon one of the secondary rings into which each segment 

 is divided. The secondary annulation began in the first stage and 

 is present throughout larval life. There are eight of these annuli 

 on each of the segments that bear the oblique stripes, except the 

 8th abdominal. Where this latter segment is crossed by the upper 

 part of the 7th stripe, entering the base of the horn, the annulation 

 is not present and the adjacent areas are not separated by furrows 

 (between the annuli on other segments), and, therefore, fuse at an 

 early date. The 7th stripe is also much whiter and more conspicuous 

 than the others. As the larva advances in this stage, the subdorsal 

 and the stripe above it become indistinct, but the 8th stripe becomes 

 more prominent and is especially well seen as a V when the larva 

 is looked at from above. The head is shagreened as in previous 

 stages*. The most inteiesting fact about this stage, however, is the 

 appearance of the purple borders to the white stripes. These were 

 never present at ecdysis, and, in some instances, they did not appear 

 at all in this stage (in the case of very light varieties). So, also, 

 the time at which they made their appearance varied greatly, and 

 the extent to which they were developed. The stripes are linear and 

 very narrow ; they first appear as a brownish rather than purple edge 

 to the central part of the 1st and 7th stripes. Then they appear in 

 front of the others nearly at the same time and without any definite 

 order. It seemed that the first stripe gained a border rather before 

 the 7th. The purple edge is not a modification of the white stripe, 

 but is distinctly due to a darkening of the ground-colour. So far 

 from the shagreen-dots having any relation to the border, they are 

 either absent from it or very small (which is also true of the ground- 

 colour anterior to the whole length of the stripe). There is nothing 

 spot-like in the first appearance of the border, it is always very 

 narrow and linear. Its first appearance confirms the view that I 

 expressed last year (Trans. Ent. Soc. Loud., pt. 1, 1884), i.e., that 

 the border is a modification of the ground-colour in front of the white 

 stripe, and is not due to the drawing out of patches of colour that 

 appear in this position, in fact that the border is linear primarily 

 and not secondarily. Kleemann states that the larvae acquire the 

 purple borders in the 4th stage, and Weismann says that he has 

 observed the same thing. Nearly all my larvae acquired the borders 

 in the 3rd stage as above. At the conclusion of this stage the larva 

 is about 25mm. long, when comfortably at rest. This stage only 

 lasted about 6 days. A larva at the end of this stage, comfortably 

 extended at rest, is shown at fig. 4X2. Fourth instar: The larva 

 at the beginning of this stage is exactly similar to one in the last 

 except for the greater size of the purple borders. The subdorsal 

 and line above it are present at first, but subsequently disappear, 

 while the 8th stripe remains. The pair of distinct dots are at first 

 visible on each thoracic segment but they afterwards cease to be 

 * Should be " stage." The head is smooth in the 1st instar, shagreened in the 2nd. 



