( 411 ) 



figs. 47 — 72), and this constancy indicates that the character was at h'ast iacipiont 

 in the ancestral antenna of the resjiective gronp. 



As we liave thns seen that the antennae with a reduced number of bristles 

 represent a more specialised state than those with a less, or not, reduced series, and 

 that the forms with the series irregular are less generalised than those which have 

 a regular belt, the development of the bristles admits the following conclusions to 

 be drawn as to the phylogen}' of the families of Butterflies : — 



The preservation of the most ancestral form of arrangement of the bristles in 

 Lycaenidae mates it evident that the Butterflies taken as a whole are not a further 

 development of any recent family of Moths (apart from Jagaluc'), and that they are 

 closer connected in the characters of the bristles with the ancestral Lepidopteron 

 than any Frenate Moths are. The Liicui'iudne being the only family in which the 

 ancestral character is preserved, all the other Butterflies being more or less 

 specialised, it is further clear that the Li)c<irniilae are not a derivation from any 

 other recent Butterfly family. 



The Hi'sjieriiiliie, having in many instances an ancestral, ventral, belt of bristles, 

 can but be derived from a Lepidopteron with a regular postmedian series, and they 

 may, therefore, have developed from the Lycaenidae. But the agreement of the two 

 families in the preservation of tlie ventral generalised series of bristles is not a con- 

 clusive argument for their being very close allies ; for the Hesju'riidKc may just as 

 well have originated direct from the ancestral Butterfly, or even from the ancestor 

 of all Lepidoptera. The apical, specialised, position of the row of bristles iu a few 

 Hesperiidae and Lycaenidae (figs. 2, 7) would at first thought seem to suggest, that 

 we had here to do with a specialisation which was an expression of relationship of 

 the two families ; bnt the antennae in which that specialisation occurs are such 

 as have the base of the joints depressed, and hence it is possible that the apical 

 position of the bristles is merely a consequence of this configuration of the surface 

 of the joints. 



The nearly complete belt of liristles on the apical joints in some En/riiiidai', and 

 the constant position of the mesial and submesial bristles near the ape.x of the joints 

 on the club, show that this family is a derivation from a form which agreed with 

 tiie early Lycaenidae in the iiossession of a belt of bristles, and deviated from the 

 ancestral stock by the apicad movement of the mesial bristles. The more highly 

 specialised Krycinidae agree closely with the Fieri da e in the possession of one 

 or more lateral bristles and an apical submesial pair ; and as this character is 

 constantly found on the club in those Pieridae in which the submesial bristles are 

 not obliterated, and in no other specialised family, we must conclude that there 

 is a closer phyletic connection between the Pieridae aud Krycinidae than between 

 the Pieridae and any other family. The Pieridae certaiidy cannot be a derivation 

 from the Parnassiinae, on account of the absence or very different jiosition of the 

 always much reduced bristles o{ Parnassiinae; while, on the other hand, the P/Vrnfee 

 also cannot have given origin to any other recent Butterfly family, as the Pierid 

 specialisation is very difl^'erent from the Nymplialid specialisation, and as in the 

 other families wo find antennae of a much more ancestral type. 



In Pajiilionidae we meet again with a subaucestral development of the bristles 

 in some members (f. 39, Leptocircus) : the family canuot, therefore, be a develop- 

 ment from the always specialised Nymphalidae or Pieridae, but can be derived, as 

 far as the bristles are concerned, from the Lycaenidae or Hesperiidae, or may have 

 originated direct from the ancestral stock common to all Butterflies ; all throe 



