1263 
teristic genus Phassus, the three types of colour-pattern described 
above, can again be distinguished, and the connection between 
them is seen to be the same. For instance Ph. giganteus corresponds 
to Charagia mirabilis & in the regular alternation of biconcave and 
biconvex markings over almost the whole space of the wing-surface. 
Ph. chamyl is the counterpart of Ch. ramsayi 4; in the occurrence ot 
silver markings of different shape and size spread over the wingsurtace 
in an apparently irregular manner, and in the great contrast between 
these conspicuous blotches and the rest of the design, which for the 
greater part is indistinct, the markings showing a tendency to coalesce. 
Finally several species of Phassidae, e. g. purpurascens, herzi 
signifer, and also the just mentioned chamyl, possess the Hepialid 
triangle, in the same position as Ch. splendens, lignivorus and lewini. 
The same is the case in many species of the genus Hepialus, 
where it is often combined with transverse rows of secondary and 
primary markings. Nevertheless in- H. bebrensi, the primary markings 
have completely vanished, and nothing is left of the pattern but 
two rows of secondary silver-markings, strongly contrasting with 
the uniform dark ground-colour (ramsayi © type). H. heetus tends 
to the same type, but less perfectly, and within the specifie limits 
of H. fuseonebulosus a variety gallicus oceurs, in which the redue- 
tion of the markings has led to almost complete unicolourism, 
though the type possesses a very elaborate pattern. Again in the 
male of H. humuli all traces of pattern have given way to a uni- 
form coloration of silvery shade, the female at the same time 
showing red markings on a yellow ground, which by careful ana- 
lysis can be deduced from the general Hepialid plan. 
Though a well known fact, it nevertheless is worth remembering, 
that in the Shetland-islands a race of humuli is found, in which 
the males agree in colour-pattern with the females. 
Pielus and Trictena are two Hepialid genera, whose colour-pattern 
is modified in the same remarkable way, leading to a close simila- 
rity between the two groups, so much so that I still doubt the 
validity of their separation. In all cases it is easily shown that also 
this highly complicated colour-pattern may be deduced from the 
same groundform as that of other genera, and has developed from 
it along similar lines. 
The characteristic longitudinal white streak, running from wing- 
root to tip, with its fantastically crooked and vacillating contours, 
is evidently the result of the deformation, coalescence and partial 
obliteration of spots, originally arranged in transverse rows and 
similar to each other in form and size. 
