( cliii ) 
obtains up to the postmedian area, where a longer narrower 
pattern begins to obtain with similar irregular apices; an 
occasional very blue and somewhat different scale is also 
present, having a very smooth surface, similar in shape, but 
tapering out more widely to the apex which is crenate not 
serrate. In direct light on a dark background these are all 
very blue, whilst the ordinary scale has a paler and a mauve 
tint. I have been quite unable to trace any difference in 
structure between these and the ordinary predominant 
pattern, the ribbing is similar yet it looks smoother and is 
decidedly different in colour. The brown neural scales are of 
two patterns, a somewhat bell-shaped one expanding rather 
rapidly and with an irregular much serrated apex, and a 
narrower and longer one also serrated but less irregularly ; 
this latter is less abundant. The “ blasenschuppen ” are very 
numerous and are not unlike those obtaining in Celastrina 
argiolus, but not quite so large, and the apex is not truncate ; 
it is therefore nearly globular but a little too long to be quite 
so, they have as a rule fourteen rows of reticulations on one 
side, but in certain cases I have counted fifteen, though 
fourteen appears to be the normal number. In the secondaries 
the scales above the costal vein are brownish, of unusual width 
and of moderate length, tapering to the apex which is rounded 
off unevenly, the outer apex being generally the longer and 
descending irregularly to the inner edge, the base of these 
scales is deeply cleft at the attachment stalk and very deeply 
and heavily lobed on the costal side, this lobing which obtains 
in certain species in this particular area is very interesting 
and is evidently an accommodation to the shape of the wing, 
I am not aware of its occurring where the scales have an 
entirely free course; the basal brown scales are short, broad, 
of even width, with a regularly rounded base with but little 
incision for the stalk, and with an unevenly dentate apex, 
the blue median and postmedian scales being a modification 
of this pattern, as might be expected in this case where they 
freely intermingle; they are rather longer, tapering slightly 
towards the tri-scalloped apex, the central division of which 
is the widest ; those on the abdominal fold are long and narrower, 
even in width with a truncate apex, these are brownish at the 
