356 BKITISH BUTTEKFLIES. 



the posterior border of the segment, is a similar, but small, scar of the 

 eversible gland occupying that position in the larva ; it is a little dark 

 puckered line (slightly oblique), about O05mm. long, and with little or 

 no altered surrounding area. The 9th and 10th abdominal segments 

 are difficult to distinguish, except as one mass, with certain scars — 

 anal and genital, that may belong to one or other. What is very 

 notable is that there is no trace of any cremastral armature, and this 

 notwithstanding the fact that the insect forms at least an apology for 

 a girth for pupation. A few other points deserve notice. On the 3rd 

 abdominal segment the netting in the dark patch above the spiracle 

 often has the appearance of radiating from a centre ; it is all 

 thoroughly linked up with no loose ends, as is rather the rule on the 

 general surface, and the meshes (as seen elsewhere) are filled in with 

 smaller cells about a fourth the diameter of the others; these become 

 fainter marginally, but may be detected on other parts of the surface, 

 and are, therefore, probably everywhere present, but too faint to be 

 seen, their evanescence explaining the frequent loose ends in the 

 network as generally visible. The subdorsal mark of this segment is 

 similarly provided with abundant small cells, but has a remarkable 

 central structure, like a scar, forming an irregular centre as of several 

 ill-formed lenticles surrounded by radiating lines, with an area about 

 0'08mm. in diameter ; on the 2nd abdominal segment, this is repre- 

 sented by a slightly puckered fulness and prominence of the ordinary 

 lines; it is repeated, well-developed in the 4th, 5th, and 6th abdominal 

 segments, but is wanting on the 7th. The skin-points, or rosettes, 

 which seems a tempting name for them in this pupa, that occur at the 

 intersections of the network are comparatively few, perhaps a tenth of 

 the number that a similar area would show on the pupa of Thestor 

 ballus. In those of T. ballus is a small central rosette surrounded by 

 a very wide border continuous with the broad lines of the network, 

 the points being nearly Olmm. wide ; here we have only the central 

 rosette, not at all unlike in size or structure that in T. ballus, but still 

 large enough to overlap the lines that meet against it. It consists of 

 a minute central ring with from three to six branches. It is, indeed, 

 very rose-like, with central ring (stamens), and the margin divided into 

 (generally) five portions, much like rose-petals in form. Beneath these 

 the lines of network enter and disappear; the width of the lines being 

 perhaps 0003mm. in L. boeticus, and about ten times that, 0-03mm., in 

 T. ballus. The spiracles are about 0'12mm. across their long diameter 

 (transverse to segment), and have a very elaborate grid- work. The 

 hairs are more numerous near the spiracles and also longer there, 

 about 003mm. to 0-04mm., half that or less on the thorax ; they are 

 of nearly uniform width from end to end, or even a little thicker 

 terminally, and end in a sharp spicule with several others towards the 

 end, sometimes long and spreading. The lenticles are most numerous, 

 and largest, near the spiracles, where they are about O'Olmm. to 

 0-01 5mm. in diameter, or even a little larger sometimes ; the margins 

 are wide, only about the central third being occupied by the dotted 

 membrane. In certain regions are areas of minute spiculated skin- 

 points, which, in places, run along the lines of network, suggesting 

 that the ribs of the network are formed by skin-points run together ; 

 these skin-points occur along the abdominal incisions, beginning on 

 that of the 2nd-3rd incision, and are seen on the intersegmental mem- 



