268 Qenetical Studies in Moths 



to enter the soil, they died. Luckily, I had taken the precaution of 

 isolating one specimen to determine exactly what course it would pursue. 

 For five weeks it remained motionless, and then it assumed a verj'^ 

 unhealthy colour suggesting that it was moribund. Clearly, however, 

 this colour change was only preparatory to exuviation, for when I looked 

 at it later I discovered that it had moulted and had yielded, not a fifth 

 stage larva, but a strange monstrosity, larval in colouration and semi- 

 imaginal in structure. At first sight its anomalies seemed to exist chiefly 

 in the pairs of unusual appendages borne by the meso- and metathorax, 

 these evidently being external wing buds exactly comparable with those 

 seen in heterometabolous forms like the Orthoptera. More minute 

 inspection showed many further approximations to imaginal conditions 

 which were more particularly visible on the structures appertaining to 

 the head. 



In size the head was much inferior to that of the last stage larva, 

 and in this and in shape it agreed more nearly with the imago. Nor 

 was the resemblance confined to size and shape, for the antennae were 

 much longer than the ordinary larval type and possessed numerous 

 joints, it being impossible to determine the exact number owing to their 

 being fused ; in shape they closely resembled an inverted Indian club. 

 The eyes, moreover, partook of the characteristics of both larva and 

 imago, for the imaginal compound eye was represented by rounded 

 projections of normal size, more or less regularly reticulated but not 

 developing regular facets ; on the right member of these eye masses 

 appeared two ocelli with rudiments of others and on the other one 

 ocellus with rudiments. 



The larval jaws were present and, as a matter of fact, apparently 

 duplicated, but this I suspect to be due to imperfect moulting of the 

 earlier pair. 



The external wing rudiments were not all of equal length, that on the 

 right of the mesothorax being 3 mm. long and the other three roughly 

 1'75 mm. ; all possessed vestiges of the usual imaginal neuration. 



The legs, whilst in the main larval, are difficult to describe owing to 

 the presence of obvious traces of imaginal structures. Although perhaps 

 a little flattened, the prolegs do not call for special treatment. 



The advent of external wing buds in a holometabolous insect is 

 capable of two explanations; either we are dealing with atavism, in 

 which case we are receiving glimpses of the. course matters took long 

 ago in the evolution of winged imagines from original larval forms, or 

 we are concerned with, on a large scale, the phenomenon of anticipation 



