THE ENTOMOLOGIST S RECORD. 



Avbich, tliG sense hairs are restricted ; the other is the distribution of 

 the bristles, two being placed basally close to the median carina, and 

 a lateral pair outside the grooves, almost always median, but some- 

 times more or less apical or basal. The scaling has a wide range of 

 variation, from a complete dorsal scaling and a scaling encroaching 

 beneath on the grooves, or restricting the groove to patches, as in 

 L. proma, to a nearly complete absence of scaling in Danauiae. 



Dr. Jordan gives us at considerable length the principles which 

 should underlie the use of these different characters in determining 

 classiiication. One must agree with all he says under this head. The 

 difficulty lies in the application in any concrete instance. In his 

 analysis there arc, indeed, one or two errors of fact that are not really 

 of much importance, such as the statement that in the Trichi>])b'ra the 

 antennae are furnished all over with a covering of sensory hairs and are 

 taken as representing the not-scaled hairy ancestral type of the 

 antennae of LejndniiU'ia. Such antennae do occur amongst the Trkhtt- 

 ptera, but also they have well-scaled antennae as in (Kcetis. He also 

 says that no lepidoptera have wholly scaled antennae. These errors 

 would have been avoided had Dr. Jordan noted in Bodine's paper that 

 both these points are correctly stated, as Bodine refers to the scaled 

 antennae of a Trichopteron, Mi/stacith'n ninra, and mentions pointedly 

 that the antennae of Eriooania ( jnir/no-eUa j are scaled all over. The 

 result of these mistakes is that a completely scaled antenna is taken 

 to bo the highest, instead of, as it really is, the lowest type of antenna ; 

 and that the idea that hairs, even special sensory hairs, are still being 

 constantly evolved into scales throughout the LrjiiiJujitcra obtains Dr. 

 Jordan's adherence. 



In discussing the scaling of antennae, the assumption is made that 

 in tolerably specialised lepidoptera sense hairs in the antennae may be 

 developed into scales, and that, as a matter of fact, when the scaling 

 covers a more extended surface, it does so by the sense hairs on the 

 invaded area undergoing an evolution into scales. This seems to be 

 erroneous, if not absolutely impossible. Scales originally developed 

 from sensory hairs or some similar cutaneous organs, and if so may do 

 so again ; but that they do so frequently and constantly, as must be 

 the case to account for the infinite gradations of scaling in the antennae 

 of butterflies (and elsewhere) is inconceivable. To restrict ourselves 

 to antennae, a typically evolved antenna possesses a series of joints, 

 which are to a great extent homologous with each other, each joint is 

 scaled dorsally, and has sensory appendages ventrally. The scaling 

 is more inveterate as a character on the basal joints, the sensory struc- 

 tures towards the apex, at least in butterflies. So homologous are the 

 joints that they might almost be called interchangeable, and an increase 

 or decrease of their numbers by meristic variation readily occurs. So 

 long as one antennal joint was left in the clavola, one can conceive a 

 complete antenna being regained in the descendants, Avithout the prin- 

 ciple of " lost structures never leiiig regained " being transgressed. The 

 same homology of parts, moreover, governs the relations of scaled to 

 unsealed areas. The dorsal scaled area is ready to become wider and 

 wider by encroaching on the sensory area. The sensory area is ready 

 to encroach on the dorsal scaled area according as natural selection may 

 determine. The ids or biophors (or whatever be the proper word) of 

 scales are ready to replace those of sensory hairs at the margin of the 



