CLASSIFICATION OF GRACILARIA AND ALLIED GENERA. 8? 



larva has no primitive condition, in any of its aspects that are 

 so modified. 



This picture of larval changes in the Lepidoptera is probably, 

 to a great extent, a true one. Nevertheless, it is probably much 

 less frequently so than we imagine. 



Let us try to apply it to the Gracilarians. An ancestral 

 larva has lived some sort of life as a leaf-miner, like a Nepticula, 

 or a Tischeria, but in its last moult takes on the special structure 

 of Gracilaria, and feeds in the Gracilarian manner, and then 

 passes the change back to all the earlier instars. It is just 

 conceivable ; but to follow this life in its last skin, a Gracilaria 

 would need a very large and very succulent leaf. It may perhaps 

 be said that Phyllocnistis practically does this, though its last 

 instar shows that even here this is not so ; and we may derive 

 the group from Phyllocnistis, 



But how are we to get back in the later stages to the ordinary 

 form of larva. The embryonal centres have lost the power to 

 , develop the ordinary trophi ; they can develop Gracilarian 

 trophi, and afterwards the imaginal ones. But the ordinary 

 ones have been eliminated, and no suitable imaginal discs to 

 give rise to them remain. It is not possible to picture a 

 Phyllocnistis giving rise to a form with a larva possessed of 

 ordinary mouth-parts. They are gone and cannot return. Any 

 modification of the mouth-parts of Phyllocnistis larvse that are 

 possible would probably be less like the ordinary form than they 

 are at present, though there is no reason why a modification 

 might not occur fulfilling very similar functions to those of 

 the ordinary trophi, but structurally they would be decidedly 

 different. No such forms appear to exist. 



When we remember that it is the first stage that is always 

 Gracilarian, and that it persists into the second or some further 

 stage, and that it is useful in very small larv£e only, and there- 

 fore especially in the first stage. That later there is always an 

 ordinary stage, though not completely so in Phyllocnistis, since 

 in it — this also is modified so as to possess no jaws, and only a 

 spinneret as an actually functional organ — the conclusion is 

 inevitable that the Gracilarian form arose by modification in the 

 first instar, and thence moved forward into the second, and in 

 other cases further. 



This modification in the early instars of Gracilaria is by no 

 means an isolated instance of such an occurrence, but it is 

 probably the most pronounced and the most unmistakable case 

 in which an early larval instar undergoes modification, indepen- 

 dently of any change in the later ones. 



We are familiar with the four stages of egg, larva, pupa, and 

 imago, and that modifications may take place in any one of 

 these, without any corresponding change necessarily occurring 

 in any of the others. And we are tolerably prepared to find 



