I. TRÄGÅRDH, LEPIDOPTEROUS LEAF-MINERS. 15 



outline and the larva has changed from the sap-feeding type 

 into a more normal one it proceeds, in the well known 

 manner, to raise the »roof» of the blotch. It commences 

 then to feed on the tissue left on the »floor», the only place 

 where there is any food left, the rest of the walls being 

 composed of the thin epidermis of the leaf. The tissue is 

 eaten from the surface patch after patch and it would seem 

 that the larva only needs to be able to discern anything on 

 the floor; hence the reduction of the two dorsal eyes. 



By this discussion of the adaptations of the ocelli of the 

 leaf-miners, it will be noticed that I have taken for granted 

 that the primitive number of ocelli in the lepidopterous larve 

 here treated is six. 



If, however, we take into consideration that two of the 

 more primitive genera, Eriocrania and Nepticula only possess 

 one pair of ocelli, it would seem that this is to take too 

 much for granted, and this the more so as, according to 

 Chapman [11. pl. 6, figs. 1 — 3.], the larva of Micropteryx 

 calthella, an external feeder belonging to the most primitive 

 genus known also has only one pair of ocelli. 



Against this suggestion the argument cannot be used 

 that in Lithocolletis, for instance, there can be no doubt about 

 the one pair of ocelli being a recent modification, as this 

 characteristic can of course in one genus be a primitive, in 

 another a recently acquired feature. 



5. Antennae. 



As might be expected, the antennse of the leaf-miners are 

 reduced and it is easy to follow this reduction in the diffe- 

 rent genera. 



The antennse are essentially olfactory organs and even 

 in the external feeders amongst the lepidopterous larvse com- 

 paratively small. 



As an example of a typical antenna I have chosen that of 

 a fullgrown Gracilaria syringella (fig. 19). We notice 3 distinct 

 joints, the terminal one of which is much narrower than the 

 other two. On the basal joint there are no hairs nor any sen- 

 sillse or pores; the second joint, on the other hand, is plenti- 

 fully provided with such structures, having distally two conical 



