CLASSIFICATION OF BUTTERFLIES BY THEIR ANTENNAE. 125 



specialisation. This may perhaps render it easier for either of them 

 afterwards to encroach on and evict the other, if any advantage accrues 

 from such a movement. 



In the classification that Dr. Jordan arrives at, there is a largo 

 majority of items with Avhich one agrees both on general grounds, and 

 because the antennal structures described justify the conclusions. 

 Some of these may be briefly noted. 1. — The ancestral position of 

 the HcspcriiiJat', though not as Dr. Jordan concludes, specialised in the 

 matter of scaling, and in this regard the highest of the butterflies. 

 Dr. Jordan's strongest ground appears to be in the arrangement of the 

 bristles. 2. — A similarly early origin for Li/ccwnidac, largely based on 

 the bristles. Although the conclusions, founded on the absence of the 

 dorsal scaling apically, cannot be accepted, there is much to be said 

 on antennal and general grounds for deriving butterflies directly from 

 the Jiu/atac, since their relationship to the earlier Fimatae is quite 

 vague, and no families can with any certainty be referred to the in- 

 termediate line. It is certain that one must descend to the lower 

 Invdmpletae, if not to the Jtii/atac, to pass from butterflies to obtect 

 J'halacnac. 3. — The divergence of Lijcacnuhir on one side, and 

 rajiiliniiular on the other, neither being very closely related, and 

 certainly not derived from the other. 4. — The association of Erycinids 

 with Lycaenids, and of Nymphalids with Papilionids, as very close in 

 origin, if not actual derivations. 5. — The homogeneity of Nymphalids 

 as evidenced by their tricarinate antennae, extending throughout the 

 Avhole family, notwithstanding its great extent, and the possibility of 

 dividing it into subfamilies, a homogeneity that does not exist in any 

 other families except, perhaps, the Hesperids and Lycaenids. 



The point on which it is necessary to part company with the 

 author, is in his placing the Pierids with the Erycinids. This seems 

 quite inadmissible on many grounds, and a close scrutiny of the 

 antennal facts, shows that they do not by any means justify this con- 

 clusion, much less enable it to stand against the evidence to the 

 contrary from other organs. 



It is necessary in the first place to define a little clearly what is 

 meant by such relationship. Dr. Jordan does not do so, but his 

 argument occurs most frequently in accordance with the hypothesis 

 that Xyinphalidac are developed from Pajiilioiiidac, and that Jj/ramidac 

 give origin to Krijcinidac, and these to I'icriilae. There can be little 

 doubt that this way of stating the matter suggests an erroneous view. 

 I should say that Xi/mphalular and Pajnlionidae had a common origin, 

 the common ancestor being neither a Papilin nor a Xi/injJialis, but if 

 one were bound to call it one or other, one would rather call it a 

 I'apilio, if for no other reason than that the pupa was no doubt sup- 

 ported by a girth. Similarly, Li/cacnidac and I'! ri/rinidar had a common 

 ancestor, which was neither Lycaenid nor Erycinid, but probably more 

 Lycaenid of the two. The common ancestor of I'ajiili) and Ij/carna, 

 unquestionably had Hesperid affinities. In rejecting, then, the derivation 

 of Pierids from Erycinids, I do not assert that they were derived from 

 Papilios, but that they were derived from the common ancestor of 

 Papilionids and Nymphalids, and adhered for a time to the Nymphalids 

 after the division from the Papilionids. Though Dr. Jordan's views as 

 to the evolution of scales on the antennae tinge his whole argument, 

 he does not specifically state that the scaling of I'ininar gives any 



