30 PROC. ENT. SOC. WASH., VOL. 20, NO. 2, FEB., 1918 



tain Gracilariidae, but still distinctly belong to the tissue feeding 

 type of larva. 



On account of the crowding forward of the triangular plates of 

 the hypostoma the labial and maxillary parts are greatly reduced 

 in size. The hypopharynx and maxillae project well forward of 

 the labrum, but are so small that they are almost entirely cov- 

 ered by the mandibles, only the galea, lacinia, and palps of the 

 maxillae and the anterior third of the lobes of the maxillulae 

 projecting beyond. 3 The labial palpi are very short. Stipes 

 labialis (Stl) is faintly indicated. Cardo (C) and stipes maxillaris 

 (St) are proportionately large for the reduced size of the mouth 

 parts and clearly indicated; but palpiger is not discernible and 

 mentum (M) and submentum (SM) are fused. On the lacinia 

 of the maxillae there is a spreading brush (Br) of moderately 

 long, hair-like spines. The maxillulae (fig. 2) consist of two 

 fleshy lobes densely spined, attached to the long rod-like arms of 

 the mentum; the upper moves freely, is moderately wide, and 

 extends the entire length of the hypopharynx. Its inner margin 

 is fringed with fine, long hairs. The lower lobe is much shorter 

 and about half covered by the upper one; there are no blades. 



To compare such a highly specialized type with any of the 

 more generalized forms is obviously unprofitable. It is neces- 

 sary to seek among the leaf and bark miners where similar biolog- 

 ical factors have influenced larval development, and there to 

 find, if possible, the forms approaching it. If these show tenden- 

 cies to develop certain parts of the head at the expense of others 

 and if such tendencies persisted in, would result in a head^ struc- 

 turally like that of Opostega we may consider them as probable 

 relatives. If, on the other hand, we see tendencies of development 

 from the normal in opposite or other directions we can not justly 

 connect the forms in any immediate relationship. By such a 

 test we can at once dismiss 'such groups as Brenthia, Bedelia, the 

 Gracilariidae, 4 or the Bucculatrigmae, Nepticula, Ectoedemia, 



3 In Opostega nonstrigella Chamb. (fig. 6), the specimens also show a 

 large well developed spinneret of the broad flattened type, slightly con- 

 st ricted at the tip, projecting well forward. In the species from Hawaii 

 the spinneret is much reduced. This, as well as the differences in the 

 antennae (figs. 3, 5) would indicate that the Hawaiian specimens are either 

 an earlier feeding stage than those of nonstrnjcll/i which are presumably 

 last instar larvae, or that the two species represent distinct but very closely 

 allied genera. 



4 The Gracilariidae represent a distinct line of larval development 

 different from other lepidopterous families and quite opposite in its ten- 

 dencies to Opostega or its allies. In Gracilariidae the anterior regions of 

 the head develop at the expense of the posterior, (he triangular plates of 

 hypostoma are being continually thrust further and further back arid 

 more and more reduced in size, while t he maxillary st ines and sub-mentum 

 become longer and longer until in some forms they occupy most of the 



