( 397 ) 



la Papilionidae the number of the ventral bristles is very variable and their 

 position very irreguLar ; besides Leptocircus with an irregular transverse series on 

 a number of joints, we find sometimes an indication of a snbajncal or postmedian 

 series in some Pajjilios (for instance in the fciwde of P. uqamemnon^ f. 3(5) ; otlier 

 species have a number of lateral bristles without any regular arrangement, and the 

 Papilios with distinct lateral grooves {Pharmacophagus) possess often some bristles 

 between the grooves (f. 40). These latter bristles are perhaps homologous to a 

 pair of bristles that appear constantly in Ni/mphalidac. The ventral bristles of 

 Ni/mplmlidac (inclusive of Calinaga, which was described as a Pajnlionid) arc 

 surprisingly regular and constant in number and position, the mode of distribution 

 being quite unique among all Lepidoptera (figs. 47 — 72). There are in all forms 

 of this vast family, if the bristles are not obliterated, two pairs of bristles, one pair 

 at the base of the joint near the mesial carina, the other lateral at the dorsal side of 

 the lateral carina. The second pair varies in position in the different forms of 

 Nym])halid Butterflies in so far as it is sometimes more basal, sometimes more apical. 

 Both pairs are in a number of cases much reduced in length, but seldom not 

 visible (except in a transverse section of the joint). 



Among Moths we meet forms which resemble certain Butterflies very much in 

 the mode of distribution of the bristles. There is in a number of Hepialidae 

 {Hepialus humuli, for instance) a belt of bristles all round the joints -as in some 

 Lycaeiiidae {Tltecla, Arhopala). Other Hepialids, such as have the dorsal side 

 of the antennae scaled, have a ventral transverse series of bristles, and in 

 many other Moths with not-pectinated antennae an irregular series is found. 

 Coeytia has a transverse series of bristles arranged as in certain Hesperiidae, while 

 of Castniidae only the distal joints of the small Australian species show distinctly 

 a transverse row of short bristles, the large American forms possessing only one or 

 two bristles at each side standing close to the dorsal covering of scales, except the 

 very thin apical joints which bear a greater number of irregularly arranged long 

 bristles. In many Moths, for instance in Agarmidae (Nov. Zool. 1896. t. 4. f. 1, 10), 

 with not-pectinated antennae there are two pairs of bristles on the not-scaled ventral 

 area, one submcsial and apical, the other lateral and basal, recalling to a certain 

 degree the two jjairs of bristles of Ntjmphalidae and especially of Pieridae. 



B. A Classification of Butterflies. 



We have seen on the preceding pages that the deviation in the antennae of 

 the various forms of Butterflies concerns a number of special organs and structures. 

 A classification of these insects based upon the development of the antennae must 

 take into account all those parts which we have found to offer distinguishing 

 characters. As we have further seen that some of the antennal organs and 

 structures undergo modifications quite independently from others, it is clear that 

 the observations at these independently mutating parts of the antennae furnish us 

 with facts which, if corroborating eacli other, admit conclusions to be drawn with a 

 high degree of correctness as to the phyletic connection between the families of 

 Butterflies. To find out this connection, the blood-relationshij), is the ultimate aim 

 of the classifier. The descriptive ]iart of classificatory work provides the necessary 

 facts ; the classification is a summary of tlie conclusions the classifier bases upon 

 these facts. The correctness of the classification, therefore, depends firstly on the 



