452 BRITISH BUTTERFLIES. 



before reaching it. Examination of rl. xi., fig. 2, shows the trumpet- 

 hairs as manifested by this pupa (x 200 diameters). If it be compared 

 with the similar one of R. phlaeas (pi. x., figs. 1-2), the resemblances 

 will be seen to be very close, and differing a good deal from the 

 somewhat allied ones of Heodes virgaureae (pi. xi., fig. 1), and very 

 much from the somewhat remarkable form they assume in Loiceia amphi- 

 damas (pi. xiv). It will be noted that the pupal surface is, in both 

 cases (0. dispar and R. phlaeas), and, indeed, in the others, marked out 

 into small polygonal areas by raised ribs, which have, at their junctions, 

 rounded tubercles, of which the darker interior shows some indication 

 of radial division into sections. This arrangement is probably identical 

 with that in Thestor hallux (pi. xv., figs. 1-2), where the cells are 

 much smaller and the ribs and tubercles larger, and the latter with 

 more obvious detailed structure. One observes here also that the ribs 

 joining the tubercles, in some cases, do not, in fact, always do so, but 

 lose themselves by spreading out on the flat areas, often fairly close 

 together, but with the aspect of preferring to take a slightly different 

 direction and finish rather than meet their neighbours. This phase is 

 well illustrated in T. balliis (pi. xv). In some areas, all the tubercles are 

 linked up to their neighbours and there are no loose ends ; in others, the ribs 

 merely continue the hexagonal structure of the tubercles and alternate 

 with neighbouring ones instead of meeting them. These two phases 

 are most pronounced on areas somewhat apart, nevertheless they are 

 shown in pi. xv., fig. 2. The upper and right hand sides show the 

 tubercles linked together, whilst the alternate arrangement is seen in 

 several tubercles towards the middle of the lower margin. It has 

 been suggested (Ent. Rec, xviii., p. 145) that these tubercles represent 

 skin-hairs, but I believe, now, that this is a very doubtful suggestion, 

 since, as a fact, they never by any chance carry hairs. The few very 

 minute hairs that occur on the pupa of T. hallux always occupy the 

 clear interspaces, and are, therefore, the representatives of the trumpet- 

 hairs of the Chrysophanids. It is especially to be observed in both 

 R. phlaeas and C. dispar that the trumpet-hairs arise from bases in 

 the interspaces, and never from the ribs or their associated tubercles. 

 Plate xi., fig. 2, shows the ribs and tubercles strong and clear in places, 

 fading out in others ; the trumpet-hairs are, perhaps, a shade smaller 

 than those of R. phlaeas, 0*05 mm. in height, 0-Olmin. wide at base, and 

 0-03mm. to 0'06mm. wide at top in different specimens. They appear 

 to be colourless or glassy in material, any dark lines they show being due 

 to refraction. At the top and the bottom of pi. xi., fig. 2, are two circles 

 that may be "lenticles," but are more probably the bases of trumpet- 

 hairs that have been broken off. One is rather annoyed in a field of view 

 to find one or more hairs obviously broken away, and sometimes the 

 whole field cleared, but, looking at their frail and evanescent appearance, 

 one after all wonders how, in the accidents that befall pupae (alive or 

 dead), at the hands of the collector even, any proportion at all of the hairs 

 happens to remain. In the photograph (pi. xii., fig. 1) the spiracle of 

 the left side of the 6th abdominal segment of the pupa of G. dispar is 

 shown, with its surrounding territory ( x 100). We have here a very 

 definite difference between the sculpture and hairs, otherwise so 

 much alike of Rumicia phlaeas and Chrysqphanus dispar. In R. phlaeas 

 there appear to be no hairs except the trumpet-hairs, whilst in ('. 

 dispar we have long hairs (0'08mm. to 0'17mm.) of more ordinary 



