( 405 ) 



which explains the stage of mutation i'ouuil in many SijinphaUdnc (for instance in 

 DanuiiMc) which have small patches of sense-hairs and large grooves (f. 71); while, 

 on the other hand, the close connection between the development of the patches of 

 sense-hairs and the grooves is an explanation of the small size of both patches of 

 sense-hairs and grooves in many other Butterflies (figs. 1.3, 30, 58). Among allied 

 forms the antennae with restricted seuse-hairs are more specialised than those which 

 have the patches of sense-hairs extended, and the former again are less specialised 

 than the antennae in which_ the patches of sense-hairs and the grooves correspond 

 in shape and size. The highest degree in this direction of development wonld be 

 total modification of the sense-hairs and subsequent disappearance of the grooves, a 

 stage of development nearly reached in some Nijmphalidae {Sutyrinae, f. 59). 



The families in which we observe the beginning of tiie modification of the 

 fine sense-hairs (ventrally) and the development of grooves, namely in Kn/cinidac 

 and PajiiUonidae, show already divergency in the first steps towards specialisation, 

 as we learn from a comi)arison of the Lycaenid with the Erycinid antennae, or of 

 the generalised antennae of certain females of Papilios with the antennae of the 

 respective males, or of the more generalised proximal with the specialised distal 

 joints of a Papilionid or Erycinid antenna. In Enjcimdue the appearance of 

 setiferons punctures begins laterally, leaving a mesial sjiace entirely covered with 

 tine sense-hairs, while in Fajnlionklae the modification goes on ajjically, restricting 

 the fine sense-hairs to a basal patch of variable size that extends laterad. A next 

 step in the development of the hairy areas in Fajiilionidae is that the area becomes 

 sinuate distally in the mesial line (f. 42), and subsequently, the modification 

 proceeding basad, divided into two patches, the development resulting finally in 

 two well-defined, relatively small, sublateral (in Papilio pr/anius and allies) or 

 submesial (P. leosthenes) patches. On the other hand, the incipient specialisation 

 observed in certain Enjcinidae is carried to a higher degree in other forms of this 

 family {Semeobiiis, f. 11), and to a still higher degree in Pierinae, the antennae of 

 all these Butterflies having the sense-hairs, at least on the distal portion, restricted 

 to one mesial patch. 



However, among Papilionidae there is another kind of specialisation obscry- 

 able. In certain species of Paniassiinac the sense-hairs are similarly restricted 

 basad as in Papilio machaon and allies, without the development of impres- 

 sions, while in other species a very irregular impression appears, that in 

 others again, at least on the stalk, becomes more rounded and more regular in 

 shape (tigs. 32 — 34), and we note that the more regular grooves occur in the 

 male sex. If we now remember that in certain species of Papilioninae {P. 

 priamus, P. ambrax, pohjtes, memnon, etc.), the males are in advance of their 

 females in the specialisation of the sense-hairs respecting setiferons punctures, it 

 is evident that here again the mule antenna with the more regular grooves is 

 higher in specialisation than the antenna of the female with irregular and more 

 shallow grooves. That means that the not-grooved Parnassune antenna is closer 

 related to the little specialised Papilionine antenna as found in P. podalirius, 

 nuichaon, etc., than is the grooved Parnassiine antenna. The Papilionidae 

 represent, therefore, two lines of develojiment leading from tlie most generalised 

 form, as found in the females of Papilio jioli/tes, etc. (f. 41), to one extreme with 

 two grooves in /'. priamus and allies (f. 40), and the other extreme with one groove 

 in Paniassiinae (f. 34). As the Xymphalidae have always two-grooved and the 

 Fierinae always one-grooved antennae, it would seem very natural to conclude 



