FAM. GRACILARIAD/E 3 



but no near connection can be made out. The characteristic porrected three-jointed maxillary p;ilpi 

 probably indicate real affinity to the true PluteUidae, and so also does the normal absence of long fine 

 hairs (not bristles) on the posterior tibiae; on the other hand the basal obsolescence of the upper margin 

 of cell of forewings, and the costal termination of vein 7, are notable Tineid characters. The fourteen - 

 legged larva, and the peculiar attitude of the imago are unique features. 



Due attention to the family characters as denned above will show that Tischeria and Bedellia, 

 which have often been classed with Lithocolletis, do not belong here at all. Other genera wrongly referred 

 here by their authors are Metriochroa, Busck, which is allied to Tischeria, and Philodoria, Walsingham, 

 which belongs to the Glyphipierygidae y.hrt.0^.djb^Q y ^ula/~flijt -nl4" Af-i^t^v ftrtfcl *£+ "Aft 



The internal prrylogeny of the family is not difficult. The most primitive genus is undoubtedly 

 Gracilaria : from this comes Parectopa, giving rise to two main branches represented bv Acrocercops and ^ N- ** 



Lithocolletis respectively. The two first-named genera are of Indian origin, and the two branches repre- 

 sent tropical and northern temperate lines of development respectively. These conclusions, coupled 

 with the association of the species with Dicotyledonous plants of advanced development, indicate an 

 origin in comparatively recent times, probably the Eocene period. 



In delimiting the genera the characters of the legs are in this family of chief importance. The 

 purpose of the curious and characteristic posture, in which the two prominently ornamented anterior 

 pairs of legs are conspicuously displayed, is not known ; it looks as though it were sexual, but I am not 

 aware that there is any direct evidence in that sense. The neuration varies to some extent within most 

 of the genera, though there is a progressive diminution of veins culminating in Lithocolletis; in the higher 

 forms some of the veins are often partially obsolescent or faintl}' developed, and are liable to be over- 

 looked without careful observation. The scaling of the labial palpi, formerly employed to distinguish 

 genera, is in this family quite unreliable. I characterise 17 genera, with 583 species. 



KEY TO THE GENERA 



1. Middle tibiae clothed with rough projecting bristles 2. 



— Middle tibiae not clothed zcith rough projecting bristles 3. 



2. Forewings with 11 absent 5. Genus Spanioptila, Walsingham. 



— Forewings with 11 present 6. Genus Phrixoscei.es. Meyrick. 



3. Posterior tibiae with series of bristly scales 4. 



— Posterior tibiae without series of bristly scales 8. 



4. Middle tibiae thickened with dense scales 5. 



— Middle tibiae not thickened with dense scales 6. 



5. Basal joint of antennae with dense flap of scales 1 5. Genus Chilocampyla, Busck. 



— Basal joint of antennae without dense flap of scales n. Genus Cyphosticha, Meyrick. 



6. Forewings with 8 absent 9. Genus Stomphastis, Meyrick. 



— Forewings iviih 8 present . . 7. 



7. Head with oppressed scales, labial palpi curved 8. Genus Acrocercops, Wallengren. 



— I lead more or less rough, labial palpi straight 7. Genus Epicephala, Meyrick. 



v Head rough on crown 9. 



— Head with oppressed scales i3. 



'.. Forewings with 11 absent 10. 



— Forewings with 11 present _i 1 . 



; . Forewings with 4. present 2. Genus Cremastobombycia, Braun. 



— Forewings with 4 absent . . .' 1. Genus Lithocolletis, Hiibner. 



