scales, since a scaled surface is less in need of tactile sense than are 

 unprotected ones. But it seems very unlikely that special sensory 

 hairs should undergo such an evolution. 



Leaving, however, the ground of probability, and looking for 

 facts, when we find the primitive lepidopterous antenna of Eriocrania 

 possessing both scales and hairs equally distributed, we find the con- 

 clusion almost irresistible that the hairs are the sense-hairs, special 

 sensory, and at least some tactile also, of the pre-lepidopterous 

 antenna remaining unchanged, and that amongst these — in hypo- 

 dermic cells, in protective hairs, or in tactile hairs — the proper 

 structure to evolve into scales has been found on the antennae, as on 

 other parts of the surface. 



If the antennae are really the special sensory organs we believe 

 them to be, then it seems as unlikely that the structures — hairs, or 

 whatever else they may appear — that receive the special sensory 

 nerves should be transformed into scales, as that, say, the facets of 

 the compound eyes should do so. 



It would almost appear that the antennae became scaled unwill- 

 ingly, but were obliged to do so because they possessed some of the 

 same structures of the chitinous envelope that developed into scales 

 elsewhere. This appears from the fact that in Eriocrania and some 

 Incurvarias the scales are, as it were, buried amongst the hairs, and 

 can be of little use either for protection or coloration. 



In nearly all the orders of insects there appear to be instances of 

 antennae without any definite orientation, but also in all, or nearly all, 

 there is a strong tendency to specialisation of surfaces, one side 

 having a different character of development to the other. This 

 tendency soon took effect in the Lepidoptera ; and if we go to the 

 immediate descendants of the Eriocraniads, we find a greater variety 

 in the results of this tendency than in any other family. The 

 antennas have not yet taken any determined line of evolution, and 

 are, therefore, decidedly more plastic than they afterwards become. 



The lower Adelidae, and especially the genus Incurvaria {muscu- 

 lelld), are closer to the Eriocraniadae than any other Frenate is to any 

 Jugate family. 



The male of Incurvaria tnusculella has unipectinate antennae. The 

 pectinations are flat spathulate projections from the middle of the 

 lower surface of each joint. This is a strong determination of growth 

 in one direction. The most remarkable feature of the antenna, how- 

 ever, is the presence of both hairs and scales over its whole surface. 

 The scales are not confined to the dorsum of the antenna, and the 

 margins of the pectinations as in the pectinate antennae of many 

 Obtectae, but are tolerably uniformly scattered over the faces of and 

 between the pectinations. They are more abundant than elsewhere 

 along the dorsum, and also at the opposite surface, viz. the lower or 

 proximal surface of the pectinations. At precisely the same points 

 the hairs are most abundant also, and overtop the scales. 



In Incurvaria (xhlmaiiniella, which has non-pe^^tinate antennae, the 



