OF WASHINGTON, VOLUME XVIII, 1916 165 



ulation" theories. For detailed study the reader is referred to 

 the able dissertations of the above mentioned authors. A synop- 

 sis of the most salient features in the development of these larvae 

 will suffice for our present purpose. 



To obtain their food, the sapfeeders burrow in leaves, stems, or 

 twigs. By means of a narrow slit they separate the epidermis 

 from the parenchyma and thus cause a flow of the plantsaps. 

 This habit necessarily involves profound structural modifications. 

 The head-capsule takes the form of a wedge. The mandibles 

 are flattened into thin blades with small teeth and a serrate cut- 

 ting edge. Becoming greatly enlarged, the hypopharnyx forms 

 an ideal receptacle for the flowing saps. The spinneret is either 

 absent or so greatly reduced as to be functionless. In some 

 cases, the labial palps appear to be missing whilst in others they 

 are merely rudimentary. The maxillar\ r palps are very incon- 

 spicuous. It is readily seen why all these appendages of the lower 

 lip should undergo such reduction. Their presence in the normal 

 form would indeed prove very cumbersome to the larva in the 

 making of its peculiar mine. The body also undergoes a general 

 flattening and becomes moniliform. Legs and prolegs are rudi- 

 mentary or absent. 



The sapfeeding habit is a very high specialization and must be 

 considered as a comparatively recent acquisition in the Lepi- 

 doptera. It has been found only in two families, the Phylloc- 

 nistidae and the Gracilariidae. There is, however, a great dif- 

 ference in this respect between the two families. As far as we 

 know, the Phyllocnistidae feed on plantsaps exclusively through- 

 out their whole larval existence. Such is not the case in the 

 Gracilariidae. Here we have every indication of a gradual evo- 

 lution tending to establish the sapfeeding habit as yet not fully 

 acquired in many genera. In their early instars all gracilariids 

 are sapfeeders and present as such the typical form induced by 

 this habit. The larvae of many genera become external feeders 

 or tissue feeding leaf miners in their later instars. On changing 

 their habit they change their form and return to the normal type 

 of lepidopterous larvae. This return occurs at different periods of 

 larval life. Thus, Gracilaria and Ornix have only two flat instars. 

 Acrocercops strigifcnitella has two and a partial one. Phyllono- 

 rycter has three. With Marmara and Cameraria, on the other 

 hand, the sapfeeding habit persists throughout the entire feeding 

 period. That the early stages should specialize rather than the 

 later ones is in itself a most remarkable feature. Chapman calls 

 especial attention to this fact and advances the theory that 

 "There is a tendency of a peculiarity acquired at any stage to be 

 passed to the preceding or following stage. Hence that the young 



