CLASSIFICATION OF BUTTERFLIES BY THEIR ANTENNAE. 127 



as at the middle of the segment. In some Erycinids there are patches 

 of unsealed surface in the middle of the ventral surface of various 

 segments of the shaft. Dr. Jordan describes these as grooves, but I 

 cannot determine that they are hollowed in any way, beyond the 

 depressed appearance due to the scales rising round them. 



Now in Picruiac, the grooves begin as very definite circular or oval 

 pits, on the middle of the ventral surface of the segments not extending 

 to their extremities (Aporia crataci/i, many Delias, Kcarsayi, J'^iirJiarix, 

 A;/()stina, and Catasticta), and this is more or less the case throughout 

 the greater part of the Pierinae, only in some of the more evolved 

 forms such as Picris alcesta, Anthocharis hclia, do the grooves extend 

 to the ends of the segments, forming a continuous groove along the 

 antenna, but in such species there always remains a certain narrowing 

 of the groove at the margins of the antennal segments, showing their 

 development from a central point. Curiously enough. Dr. Jordan 

 describes Stalachtin as belonging to this group. It certainly has a very 

 straight groove, with margins quite parallel from end to end, in thorough 

 Erycinid fashion. Its antennae, therefore, show it to be an Erycinid, 

 as we always used to suppose it to be. Why Dr. Jordan calls it a 

 Pierid remains unexplained, I can only imagine that Dr. Jordan, by some 

 accident which escaped his notice, got his notes and figures of this 

 genus inadvertently transferred from his Erycinid to the Pierid portfolio 

 (ligs. 25 and 27 are apparently transposed by some similar oversight). 

 The restricted patch of hairs is always within the depression, no hairs 

 occurring outside the groove as in Erycinids. 



Dr. Jordan makes a very strong point of the fact that Papilionids 

 show a decided tendency to restrict the hairs to two patches, and that 

 the Papilionids are, therefore, a two-grooved family, and that conse- 

 quently a one-grooved family like Picridae cannot be derived from it. 

 Though perhaps hardly consciously doing so, he here again assumes 

 that if Pierids are close to Papilionids, it must be by derivation from an 

 already well-characterised or even highly specialised Papilionid, instead 

 of from an early form, now lost, that one would only call Popilio, 

 because it is more Pap'dio than Picris. This early form had already 

 the tendency to restriction of the area of sense hairs, a tendency which 

 was wanting in the Lycaenid branch, and which only appeared there 

 in the Erycinid division after its separation from Li/rama, the 

 Lycaenid and lower Erycinid being without grooves. In Painlin 

 (archican) also this was a tendency only, but it early took form, and 

 did so in all the branches of the stem, resulting in Papilionidac 

 merely in restriction, and that in the higher branches only, so that 

 the Nymphalid grooving could not be derived from these. The 

 Nymphalid grooving very early acquired great fixity. 



The Parnassian branch showed depressions of a character not unlike 

 those of Picris, and quite unlike anything in the main Papilionid stem. 

 I have not been able to satisfy myself that these depressions are not 

 central but lateral, as Dr. Jordan believes he can demonstrate. The 

 point is one of considerable interest, but of no importance as regards 

 the phylogencticfquestion before us, as in any case no one, I think, 

 has any idea that I'icris descends from Parnassins, but if it did, since 

 it is difficult to see the unilatoralness of I'aruaxsius, the same 

 unilateralncss may quite conceivably underlie the Pierid form, though 

 now impossible to detect. This much is certain, that the Pierine and 



