classificaTi'on' of buttekflies by theik antennae. 87 



may be five apical joints dorsally, and seven ventrally (in one form 

 only) the dorsal area shortening to two apical joints, and the ventral 

 area extends jxn-i }>assii, even to the base, often however divided up by 

 transverse bands of scaling. Setiferous punctures cover the dorsal bare 

 area, but hairs maintain themselves over the whole ventral area. The 

 arrangement of bristles on the whole resembles that in the Li/cacnlilac. 

 IV. Picn'ihh-. The antennae clearly divide this into two subfamilies, 

 Vierinac and JUsinorpliiiiiac {si)iaj)is is our only British representative of 

 the latter). IVa. ricrinac : These are characterised (tliere is only one 

 exception) by having the sense hairs restricted to a longitudinal 

 depression on the ventral aspect of the antenna, the depression may be 

 continuous from one end of the segment to the other, or restricted to 

 a central round or oval patch, the scaling covers the dorsum or all 

 except the last one to four segments, and varies ventrally from absence, 

 to complete scaling, except of the club, often leaving the depressions of 

 stalk as unsealed islets. The bristles are usually one on each side of 

 the middle of the joint, and a ventral pair distad. The antennae of 

 P>it'ii<Ii)j)()ntia are shown to be undoubtedly Pierid. IVb. BUmoyiihiinae : 

 The last two joints are always united. The not scaled area is restricted 

 to the club, and is rather more extended ventrally than dorsally. The 

 depressions here are a central and two lateral ones, always at the 

 apical end of the segment. The bristles are as in Picrinac. V. 

 PapiUoniihu' : The antennae differ somewhat in (a) Pajnlioninac (b) 

 Thai ili)taL' and {c)Parua>isiina('. The hairs are ahvays restricted to ventral 

 surface. There is often some lateral compression, making in some 

 species a ventral ridge, with a groove on either side of it, the ridge 

 never forms a carina as in Nymplialids. There is great variation in 

 the manner and extent of the restriction of the hairs, in some Parnas- 

 aiinac they are restricted to a groove or series of depressions, difficult 

 to distinguish from those of PU'viiJac, more usually they suggest a 

 Nymphalid distribution. The scaling is very various in extent also, 

 it may cover the whole dorsal surface except the last joint, or may be 

 nearly wanting ; the punctures usually cover any unsealed dorsal area. 

 The bristles are not unlike those of Hisjieriidae, and Pijrafni'lac, but are 

 more irregular in distriljution, and may be nearly evanescent ventrally. 

 VI. NjiwpJudiilar : Though these present great solidarity as against the 

 other families, in antennal, as in many other, characters. Dr. Jordan 

 does not yet determine that he Avill not finally give family rank to 

 some of the following sub-families which he recognises — (JalUnaiiiiuu., 

 Psriidcri/olijiai', Libi/theinac, Arrarinac, HcUconinac, Sati/riiuic, Pras- 

 sdlinac, Morpliiinae, Xi/nijihaliiuic, Ncotrojt'uiac, })anai)uu\ In so 

 immense and dominant a family, a differentiation of groups has often 

 gone further than in smaller and more stagnant families, and so family 

 rank may appear to be deserved by some groups ; the temptation to 

 recognise these, however, owing to its convenience in handling large 

 groups, ought to be resisted. On the other hand, if a group be well 

 defined, it is often a verbal, rather than a real, question, whether it is to 

 be called a family or sub-family. The Nymphalid antenna presents 

 two very constant features throughout the whole family that distinguish 

 it at once from any other antennae, and these should go a long wny in 

 opposing the sub-division of the group. These are the presence of 

 three raised lines or carime on the ventral aspect of the antenuii. 

 having between them two grooves, to which, or to patches within 



