122 THE CANADIAN ENTOMOLOGIST. 



where they are not functional for feeding, are not phylogenetic characters, 

 what explanation is there of their structure? This is manifestly adaptive, 

 and this view is supported by the following arguments : 



i. There is an obvious advantage in the flattened and projecting 

 labium as an aid in separating the epidermis combined with the undoubted 

 specialization of the mandibles, which, however, is not as extreme as in a 

 true Gracilarian instar. 



2. This specialization is correlated with the mode of life, and ceases 

 when the mode of feeding which produced it comes to an end. The 

 animal kingdom affords us abundant instances of species which undoubt- 

 edly are closely related within the same genus, one of which possesses a 

 specialized larval form not present in the other, which is correlated with 

 different larval habits. 



3. The period of wrinkling the loosened epidermis by means of the 

 silk spun across it, which takes place in the cylindrical group in the fourth 

 instar, takes place in this group in the sixth instar, after the period of 

 feeding has passed, which may account for. the partial degeneration of the 

 mouth-parts in the sixth and seventh instars. Why there should be two 

 instars of the character of the sixth, remains to be explained, but it may 

 be suggested that these may be periods of development, in which the larva 

 approaches the normal form. 



More important than thepe, perhaps, are the structural gradations 

 which are disclosed by the examination of larvae of species of the cylin- 

 drical-larval group, which in larval habits and type of markings approach 

 the flat-larval group. This feature will be discussed in a future paper. 



Are the differences sufficient to constitute a division of generic rank, 

 which has originated early from Gracilariad stock ? Have we not here to 

 deal with a group in the process of differentiation, rather than a genus 

 already formed and isolated, since it is by adaptation and the resulting 

 change of structure that the group will finally be isolated ? 



That the flat-larval group is, it seems to me, a more recent develop- 

 ment than the typical Lithocolletis, is shown by its restricted geographical 

 range (compared with that of the typical Lithocolletis), and the compara- 

 tively slight variation in the marking, together with their remarkable 

 resemblance to the type of markings characteristic of Cremastobombycia. 

 In determining the descent, two possibilities present themselves. The 

 first is that Lithocolletis (typical) is an earlier offshoot from Cremasto- 

 bombycia, either from stock somewhat different from the modern 



