r 408 ) 



of Popi/ionichf and Xymphnlidae. Thnt tliis is so, is not so very singular as it 

 wonld at first appear. The bilateral development in the antennae of insects is very 

 often disturbed in so far as the inner side is less developed than the outer side; 

 for instance, we see that in Heterocera with pectinated antennae the inner branches 

 are often shorter than the outer ones, and sometimes entirely obliterated, in which 

 case the antennae apjioar nnipectinate. Bnt we need not go to the Heterocera to 

 find an analogous case of development: tlie .\</mjilifilitl<i(' themselves show a 

 similar development in many forms. In a number of species of this family, 

 especially often among Sati/rinae, as well as in Libi/tkeinae, the inner grooves 

 (figs. 51, 5G) are smaller than the outer ones: in the species with a large develop- 

 ment of ventral sealing the inner grooves become narrower and narrower the 

 farther down the stalk we go, until they disap])ear entirely (f. u(j). Tliis develop- 

 ment leads, therefore, in consequence of the disappearance of the inner grooves, 

 also to one-grooved joints superficially similar to those of Pierinae. It is certainly 

 very remarkable, that we have both in Papilioniduc and Ntjmphalidae such 

 asymmetrical development of the ventral surface of the joints. 



As we have thus shown that the aberrant Parnassiine antenna belongs, like 

 the aberrant antennae of some Sai/trinac (f. 50), to the pair-grooved type, the above 

 objection to a closer phyletic connection between Piipilionidnp and Xi/mphaUdae 

 becomes invalid, and it must be accepted as established that the present state 

 of development of the fine sense-hairs and grooves points in every respect to the 

 Papilionids and Nymphalids belonging together to a i)liylnni of Butterflies with 

 pair-grooved antennae. 



The second form of antennae, the odd-grooved type, is fonnd only in Enjcinidae 

 (figs. 10 — 14) and P/c;vV/rte (figs. 15— 31), the former and the /-"/(V/wrtc possessing one 

 mesial impression (often shallow) or groove on a joint, while the Dismorphiiri'ir 

 have three apical grooves, one being mesial and two lateral. The not-grooved 

 antenna of Ltfcnerndde (figs. 0—9), which has the sense-hairs uniformly distributed 

 over the not-scaled area, is, for this character, certainly of a more generalised form 

 than the grooved Erycinid antenna, and the latter, which has the sense-hairs not 

 confined to the grooves, which are moreover often scarcely indicated, is again less 

 specialised than the antenna of Pierinae, with the fine sense-hairs restricted to 

 the grooves. As the gradations from tlic simply convex joint of the Lycaenid 

 antenna, often scarcely distinguishalde from a joint of a Hesperid antenna, to the 

 deeply grooved joint of En/cinirlae and of Pierinae, is complete, there is no doubt 

 that the odd-grooved antenna develojied from a not-grooved one similar to that of 

 l.iicncriidae, and further that the Pierinc antenna is derived from a less specialised 

 odd-grooved form of the tyjjc fonnd among Enjciniduc. Though in Kn/cinidae the 

 fine hairs arc generally not restricted to the grooves, as they are in Pierinae, it 

 is in some cases, where the scaling reaches close to the edge of the grooves, 

 difficult to say whether the joint is of a Pierid or an Erycinid type. There is, 

 moreover, no conspicuous line of demarcation between joints witliout restriction of 

 sense-hairs {Lycaenidae), io\ni& with slight concentration of the hairs {Erycinidae), 

 and joints witli obvious restriction of the sense-hairs to the grooves {Pierinae). 

 While, however, in Erycinidae the jiroximal joints are always without grooves, 

 being of a generalised character, these joints, if ventrally not scaled, are provided 

 with more or less obvious grooves in Pierinae. 



The close, agreement between the one-grooved antennae of Pierinae and 

 Erycinidae, and between the Erycinid antennae with rndimcntary grooves and the 



