( 404 ) 



veutrul siirfticc and, besides, the dorsal surface of at least the apical luilf nf the cliili 

 covered with sense-hairs. As there is in all tlirec families not one siieeies wiiich is 

 without dorsal scaling on all the joints of tiic stalk and proximal joints of the club 

 — in Mesapia, a Pierid, with sparse scaling, more recent obliteration of the scales 

 has taken jilace — it is obvious that the ancestral form of each of the tlirec families 

 most probably was already in an advanced stage of development of the dorsal 

 scaling. Si/miihididnc and Ftj/iHionidac, tliongh very often without antennal scaling 

 in consecpience of obliteration of the scaling, are to be derived from a form with the 

 entire ventral surface and the dorsal side of at least the club not-scaled. 



From this review the one point is most obviously clear, that the Ucxpcriiilae 

 cannot have been the basis from which any of the other Butterlly families develoj)ed. 



The modification of the fine sense-hairs into scales is, as we have seen, accom- 

 panied in many instances Ijy a modilication of a portion of the remaining hairs into 

 another kind of sense-organ, the setifcrous jiunctnres. We have, tlierefore, lo 

 distinguish between Lepidoptcra which have preserved the ancestral cliaracter, a 

 covering of fine sense-hairs, on the not-scaled area of the antenna, and such in which 

 that sj)ccialisation obtains. The generalised character is met with on the dorsal 

 side only among Lyraeiiidae, in no other Lcpidopterous family apart from .hifiatac, a 

 fact of high significanee : while the ventral not-scaled area of the antennae is of the 

 same generalised character not only in all Lyeaenidac, but also in all Ilcsperiidae, in 

 thu females of certain Pajnlionidue, and in the Heterocera ; but among the last the 

 scnse-liairs show often specialisation in so far as they are often longer at certain 

 points than at others, esijecially in the nvilcs. 



The modification which the underside has undergone is of particular interest, us 

 the preservation of different steps of the progressing specialisations reveals to us 

 divergent lines of development of the Bnttertly antennae. When speaking here and 

 in the following pages of the underside of the antenna, I mean that ))ortion of the 

 under surface that is not covered with sealing ; it is the interno-ventral side. 



The setiferons punctures appear first on the distal joints, as may be seen in 

 Erycinidae and Papilionidae, and the specialisation proceeds in an aj)ici-basal 

 direction. The fine sense-hairs are not entirely replaced by setiferons punctures on 

 any joint, there remaining part of the surface of the joint covered with hairs. Tlie 

 apices of the joints or the sides are the first to acquire the punctures, and this 

 corresponds exactly with what we know of the development of the ventral scaling, 

 which begins in very many instances also at the apex and sides of the joint, leaving 

 a mesial or a basi-mesial s])ace free (figs. (!, :{8). The fine sense-hairs are, therefore, 

 first restricted to the basal and mesial portion of the joint. Now, we have seen that 

 in the greater proportion of Butterflies with restricted sense-hairs the ])ortion of the 

 underside that bears the patch (or patches) of fine sense-hairs is more or less 

 impressed ; hence it is obvious that the apjiearance of one, two, or three grooves 

 stands in close connection with the restriction of sense-hairs. As in I'apilionidae, 

 for instance in Pajnlio machaoti and allies, the division of the basal area of sense- 

 hairs of the distal joints sets in without tlie previous deveIoi)ment of grooves, it 

 is evident that the grooves reiiresent a cliaracter wliicli developed pliyletieally sub- 

 sequently to the restriction of the fine sense-hairs. If the restriction of these 

 sensory organs, however, preceded the development of the grooves, then it is con- 

 ceivable that the further restriction of the sense-hairs has not in every case been 

 followed immediately liy a corresponding dimiinitioii in the size of the grooves. 



