8 SOUTH-AFRICAN BUTTERFLIES. 



Tliorax generally ratlier slender or moderately stout, sometimes 

 robust. IVinr/s large, variable in outline ; forc-ivings usually ratlier 

 truncate, but occasionally produced, — hind-margin entire, rarely elbowed 

 superiorly ; subcostal nervure three- to five-branched, usually four- 

 branched ; discoidal cell closed ; hind-ivings rounded, with entire or very 

 slightly denticulated hind-margin, or produced in anal-angular portion, 

 which often bears from one to three longer or shorter tails ; inner 

 margins often forming an incomplete groove about the abdomen ; dis- 

 coidal cell closed by very slender nervule. Legs rather short, often 

 thick, scaly, rarely hirsute ; tibial terminal spurs usually small, some- 

 times minute, rarely wanting ; fore-legs of $ (with rare exceptions) ivith 

 tarsus not articulated, but consisting of a single long Joint ending in one 

 slightly curved claw; those of $ with the ordinary articulations and 

 terminal hooked claws. 



Ahdonien usually slender and rather short ; rarely thickened or 

 elongate. 



Larva. — More or less onisciforra, broadest and thickest about 

 middle, often with dorsal humps or corrugated ; sometimes downy or 

 with fascicles of hairs ; head and feet very small, inferior, hidden from 

 view above. 



Pupa. — Short, thick, usually much rounded ; blunt at extremities. 

 Attached by the tail, and (usually) by a girth of silk round the middle ; 

 rarely unattached, and lying in the earth or under stones. 



This family is a very distinct, compact, and natural one, the char- 

 acter of the unarticulated tarsi of the first pair of legs in the male sex 

 being all but universal, and the principal other points of structure 

 presenting but little variation. This sameness throughout so very 

 numerous an assemblage of species renders the task of classification 

 exceedingly difficult ; and no lepidopterist has hitherto found charac- 

 ters adequate to warrant the establishment of divisions or sub-families. 

 To discover natural limitations to the genera is a matter of scarcely 

 less difficu.lty; and all who have studied the family will admit that, 

 notwithstanding the labours of Westwood, Hewitson, Felder, Moore, 

 and other entomologists, the existing definitions of many of the accepted 

 genera are anything but satisfactory. The work of discriminating spe- 

 cies is an arduous one in all large genera, but it becomes specially so 

 in such immense groups of closely-related forms as Lyeccna, Thccla, and 

 Amhlgjyodia. 



Between fifteen and sixteen hundred species have been recorded, 

 and about fortj— seven genera created for their reception, — the three 

 genera just mentioned by themselves including fully half of the entire 

 number of known species. 



Among the genera which show more divergence from the common 

 type are specially noticeable the Oriental Lijjhgra and the Ah-ican Lij^tcna, 

 Pentila, D' Urhania, Pseuderesia, Alcena, Miinaera:a, Dcloneura, Arrugia, 

 and Lachnocncma^ all of which bear some resemblance to butterflies of 



