( 406 ) 



that the Payilionulae, consisting of forms with one, twn, or no grooves, had been 

 the basis from which the X'jtnjthalidce as well as the Vieridae-Krycinidae developed, 

 the former appearing as a further specialisation of the two-grooved branch of 

 Papilios, and the latter as a derivation from the Farnassiinae with one-grooved 

 antennae. Does a closer examination of the facts warrant such conclusions ? 



Before entering upon the discussion of these points, it is jjcrhaps necessary to 

 mention that the two-grooved antenna cannot be derived from the one- or from 

 the three-grooved antenna, and that the inverse is also not possible. For in both 

 cases we should have to assume that the ventral surface, which in the two-grooved 

 antenna is liighest where it is most deeplv imjiressed in the one- and three-grooved 

 Erycinid-l'ierid antenna, had developed in exactly the opposite direction to its 

 former line of development — an assumption wliich is not admissible, (1) because all 

 the forms wliich stand at the top of the one line of development are so much 

 sjiecialised in many respects that they cannot have been the basis from which the 

 forms descended that show oj)])osite direction of development, and are in many otiier 

 characters less specialised than the former, and (2) because there are no intergrada- 

 tions between the pair- and odd-grooved antennae, the assumption being, therefore, 

 entirely unsupported by facts. 



Let us now examine first the two-grooved antennae. The gronp of I'api- 

 Uonidae in which the fine sense-hairs are most restricted, and in which the 

 restriction is most constant in all the species belonging to the gronp and hence 

 represents a high development of the grooves, must have arrived at this stage by a 

 laterad restriction of the fine sense-hairs, since in the more generalised Painlioiddae 

 there is a covering of fine sense-hairs in and near the mesial line. In Nymphalidae 

 the small patches of sense-hairs as found in Danainae, in certain Satyrinae, etc., 

 within the larger grooves, as well as the small grooves of Iphthima asterojie (f 58), 

 and the larger, but nevertheless much restricted, grooves of Calinaga (f. 47), stand 

 apart from each other, while in the more generalised Nymphalids with extended 

 covering of sense-hairs the two patches fill nj) nearly the whole grooves, being 

 separated from each other by the not-hairy mesial carina ; the restriction is, 

 therefore, also laterad, and, of course, basad as well as apicad, resulting always 

 in a sublateral rounded patch. However, if in both families specialisation proceeds 

 in the same direction, the results mnst be essentially similar. The agreement in the 

 diminution and position of the patches of sense-hairs as well as in the shape of 

 the grooves in members of different subfamilies of Nymphalidae, and the resem- 

 blance (not identity) of the antennae of certain Acraeinae (f. .52) with those of 

 ( 'uliiiaya (f. 47), and of the antennae of this genus with the antennae of certain 

 Fapilionidae {P. priaiuus and allies, f. 40), are, therefore, not necessarily arguments 

 for these forms being phyletically closer related to each other than to forms which 

 have not reached that degree of specialisation, the resemblance being exjilaiued by 

 the agreement in the direction of develoinnent. The fact that the same direction 

 of development obtains in both the Symphalidaa and I'apilionidae is, however, 

 weighty evidence for the close phyletic connection between the two families. Tlie 

 less specialised antennae of certain Xympkalidar-, with extended grooves and large 

 patches of sense-hairs, show further that the antenna of the Nymphalids cannot be 

 a direct development from the higher specialised Papilio antenna, nor is it possible 

 to derive the antenna of the Papilionids, on account of the very generalised forms 

 of antennae that occur in this family, from the Nymphalidae, all the species of this 

 latter family being specialised; and as we have to infer from the presence of sucli 



