( 112 ) 



assumptions exjilain 0(|iially well the occurreuce of an irregular but complete 

 ventral belt of bristles among Papilionitlae, and hence are equally admissible. 

 The further develo])ment of the bristles leads in Papilioiiidae largely to obliteration; 

 but we observe in forms with two ventral grooves (f 40) a basaJ movement of the 

 mesial bristles. Such a basad movement of a snbmesial pair must have taken place 

 very early in the ancestral Xi/mphalidae, since this pair of bristles has in all 

 members of the very large family a basal position. On account of the great and 

 constant specialisation of the bristles the Xi/mphalukie are not the basis from which 

 any other family develojicd : but the similarity in the direction of the development 

 of the mesial bristles between Nf/mplialidae and some of those Papilioiiidae which 

 have, like the Nymphalids, pair-grooved antennae, is additional evidence of 

 relationshij) of the two families. 



To sum up we will briefly recapitulate our interpretation of the state of develop- 

 ment of the four ]iarts of the antennae which wo have especially been dealing with, 

 and state the principal conclusions arrived ut as to tlie jthyletic connection of the 

 various Butterfly families : — 



I. Hesperiidae. 



1 . Dorsal and ventral scaling specialised ; similar specialisation occurring in 



members of all other families except Papilionidne. Hesperiidae cannot 

 be the basis from which any other family developed. 



2. Ventral fine sense-hairs generalised. Hesperiidae cannot be derived from 



any other family but Lycaenidae, or the ancestor of Lyeaenidae. 



3. ( 'onfignration of ventral surface ancestral. Conclusion as to phylogeny as 



before. 



4. Sense-liristles ancestral, or specialised; specialisation as in certain Lycaenidae, 



or peculiar to Hesperiidae. Probably connection between Lycaenidae 

 and Hesperiidae. 



II. Lycaenidae. 



1. Scaling subancestral, or specialised as in otlier families. Dorsal side never 



without sealing. Indifierent. 



2. Fine sense-hairs distully of lowest type among all Lepidoptcra, apart from 



some .Tuyatae, or at least veutrally ancestral. Lycaenidae cannot be 

 derived from any other family of Butterflies. 



3. Configuration of ventral surface ancestral, but sometimes faint indication of 



Erycinid specialisation. 



4. Sense-bristles ancestral on distal joints dorsally and ventrally; lowest type 



among Lepidojjtera e.xcept certain Jugatae ; sometimes specialised as in 

 Erycinidae, seldom as in Hesperiidae. Lycaenidae cannot be derived 

 from any other Butterfly family, but stand in relationship with Erycinidae 

 and probably also Hesperiidae. 



III. Erycinidae. 



1. Scaling subancestral, or specialised as in other families. Dorsal side never 



without scaling. 



2. Fine seuse-hairs ventrally subancestral or distally specialised nearly as in 



Pierinae. Erycinidae closely connected with Pieridae, 



3. Confignration of the ventral surface subancestral, or specialised as in Pierinae: 



one-grooved antennae. Conclusion as before. 



