founds his lines of evolution very much on the gradual spread of 

 scales, which he conceives as wanting in the primitive lepidopteron, 

 but developing from time to time out of the sense hairs, and so 

 securing che distribution over the antenna that they have in each 

 form.* 



Jordan's picture is of a primitive Lepidopteron fully scaled except 

 as to the antenna, which has a covering of only sensory hairs. 

 Then the sensory hairs on the dorsal aspect of the antenna begin at 

 the base to change into scales, and this process extends thence till it 

 reaches the apex dorsally, and also extends, from any point of the 

 dorsum that it has reached, downwards, till it meets beneath the 

 segment, covering it entirely with scales. This hypothesis has 

 only one fact to support it, that is that the distribution of scales 

 and hairs of the "macro" Heterocera and of the butterflies would in 

 each instance, in almost all cases, correspond to some stage in 

 the process so postulated. There are, however, at least two over- 

 whelmingly fatal objections to it. The first is that it does not 

 correspond to the relative distributions of hairs and scales in the 

 lowest Lepidoptera. The second, that it implies an almost mira- 

 culous succession of discontinuous variations, by which elaborate 

 (probably olfactory, certainly special) sense-organs are per saltum 

 converted into scales, not now and then, but in the case of almost 

 every species. A third and equally great (from Dr. Jordan's stand- 

 point probably a greater) difficulty is that the progress is not always 

 forwards ; sometimes there is retreat, as in no other way can 

 the great variety of scaling in related forms be explained, and a 

 per saltum (or in any other way) change of scales into sense-organs 

 (no intermediate forms exist) seems to be impossible. 



Both Bodine's and Jordan's hypothesis, that the antenna from 

 which that of the Lepidoptera was derived was without scales, 

 and covered all over with hairs, is one that may readily be accepted 

 as true ; but they both err in looking for this form within the order. 

 Bodine finds it in Eriocrania, quite ignoring the fact that he fully 

 recognises elsewhere, that it has a complete covering of scales. This 

 error is from Bodine's standpoint trivial, as he devotes little atten- 

 tion to the evolution of the scaled areas, and discusses fully the 

 evolution of sensory surfaces ; in fact, he rather ignores scales 

 than, strictly speaking, makes any error in regard to them. 



Jordan finds the desired primary type in Hepinlus, a more serious 

 error, since the fully haired and scaleless antenna of Hepiaius is not 

 a primary form, but secondary to the loss of scales. Hepiaius is 

 often spoken of as a very primitive and generalised form, and in 

 many respects it is so. In many of its structures, especially its 

 venation, it belongs to the Jugatae, in association with the Microp- 

 teryges and Eriocraniae, which are the most primitive Lepidoptera ; 

 and so, regarded from a standpoint of Phalen?e Obtectse, fit seems 



* Page 400 of his paper. 



