RUMICIA PHLJEAS. 



393 



line of the ridges and the end with red hooks. The 9th abdominal seg- 

 ment with two raised points, one on each side of a median longitudinal 

 fissure. Surface of abdominal segments bearing short trumpet-hairs 

 (Tutt, June 13th, 1893). The pupa is 11mm. long, thick and rounded, 

 2mm. from front, it is 3mm. high at mesothorax, and 4mm. at 3rd 

 abdominal segment after a slight waist ; it is still 3mm. high, 2mm. 

 from the posterior end, i.e., at the 6th abdominal segment. In occasional 

 specimens the wings encroach on, and are attached to, the 5th abdominal 

 segment. Generally they do not pass the usual position on the 4th. The 

 widths are almost identical with the heights, but the venter, being fairly 

 straight, the dorsal curves are about double the depths (or curvatures) of 

 the lateral ones. The trumpet-hairs are easily made out (with a lens), 

 but are inconspicuous, except when the pupa becomes very dark before 

 emergence, when they stand out brilliantly as white points. The 

 spiracles are pale, but not conspicuous ; the cover of the prothoracic one 

 is a bright little yellow line. The projection of metathorax repre- 

 senting -the hindwing ends just before the spiracle of the 2nd ab- 

 dominal segment. The last six abdominal segments ventrally measure 

 altogether only about 1mm., so much are they narrowed; a further 

 lmm. of the ventral aspect of the pupa consists of the dorsal surfaces 

 of the 9th and 10th segments ; so that, though, ventrally, the pupa 

 extends 2mm. beyond the ends of the wings and antennae, only half 

 of this is the true ventral surface. The pupa seems entirely incapable 

 of any movement, yet, on dehiscence, there is usually some opening 

 dorsally of the 5th to 6th abdominal incision, and sometimes of those 

 immediately before and behind it, 4th-5th and 6th-7th. In mounting 

 a specimen other incisions may open a little without actual fracture. 

 The cremastral region is a flat round area, about l-4mm. in diameter, 

 consisting of the 9th and lObh abdominal segments. The suture 

 between the 9th and 10th segments is lost, except for a small medio- 

 dorsal portion. Across the middle of the 10th segment is a transverse 

 ridge or suture, hardly raised at all, with a mediodorsal branch. This 

 ridge, which represents, no doubt, the transverse end of the cremastral 

 spine (as seen in Pieris and Vanessa) is just over lmm. in length, and 

 carries on either side about 36 brown hooks (probably brighter in 

 colour in the living pupa). They are about 0*05mm. long, pro- 

 portionally very thick, slightly curved, and end in anchor-shaped 

 or double hooks. Some traces of the suture between the 9th and 10th 

 abdominals may be seen or imagined ventrally; on the ventral aspect of 

 the 9th abdominal, is a central eminence, not very clearly double, with 

 (in different specimens) from one to ten similar hooks on either side, 

 outside and beyond, not on, the eminence. There is, indeed, much 

 variation in the distribution of the hooks. In one specimen, the 

 hooks of the 10th abdominal segment are dorsal to the terminal suture, 

 and this series is continuous laterally with the first series, forming a 

 circle, most dense laterally, round the anal scar, broken only in front 

 by the eminence of the 9th abdominal segment. In a specimen with 

 the anal scar well marked, it is a smooth area with a small point on 

 either side and radiating lines especially dorsal and ventral to it. The 

 8th abdominal segment in front is distinct in ^ specimens, obscure in 

 its demarcations from the 9th abdominal in $ s. The eminence on 

 the 9th abdominal is finely spiculated, as are also, but less markedly, 

 the margins of several hind segments, the general surface being 



