much the largest antennae. This is the same also in the pupae ; 

 but in these, though the pupal antenna of the female is smaller 

 than that of the male, it is larger than the imaginal antenna it 

 contains. 



Professor Poulton was of opinion that this indicated that the 

 female antenna was smaller than it had been in its ancestors, their 

 larger antennae being recorded in the large pupal antenna. The 

 difficulty of accepting this explanation is that these female Saturnians 

 have very widely pectinated antennae, not so widely pectinated as 

 the males, but for female moths perhaps the most pectinated 

 antennae we find anywhere, and it seems very problematical that 

 they could be derived from female ancestors with wider antennae. 

 It is also observed that in all moths whose males have well-pec- 

 tinated antennae the female pupa has antennae cases decidedly in 

 excess of the requirements of the imaginal antenna. 



The explanation seems rather to be that the male and female 

 antenna tend to be alike, especially in the pupa. Wherever we 

 have sexual dimorphism (or, inutatis imitandis, any dimorphism) 

 it is as if we had two species mingled, which are separately 

 responding to the effects of natural selection, and each tending 

 to vary in a different direction, and yet unable to avoid a certain 

 amount of crossing, making each to a certain extent follow the 

 other. So if the male moth gets a large antenna by selection, 

 the female tends to do so by inheritance from the male. Selection, 

 however, keeps action on the female antenna to prevent its further 

 development ; but does not act so strongly on the pupal antenna, 

 which is a quiescent and unimportant structure, and so the pupal 

 antenna outstrips the imaginal one. 



It is difficult to handle any subject without a sound beginning, 

 and so in considering the evolution of lepidopterous antennae it 

 is necessary, if it can possibly be done, to ascertain, and if not, to 

 build up, some representative of the actual ancestor of the antenna 

 of the order. 



Bodine gives us the " antenna of Micropteryx {Eriocrania) 

 se7nipurpurella as the ancestral type, and describes its arrangement 

 of hairs, but in dealing with the evolution says nothing about scales ; 

 so that, though he elsewhere describes this antenna as scaled all over, 

 there is no doubt he regards the primitive ancestor of the lepidop- 

 terous antenna as covered all over with sensory hairs. Dr. Jordan 

 gives us the same definition, that the primitive lepidopterous antenna 

 was covered all over with sensory hairs, but adds that it was without 

 scales, and gives us Hepiahis as a representative. 



Stated in this way the two views appear to be fairly in accord, 

 and yet they might be more correctly described as exactly opposite 

 to each other. This is much obscured by the fact that Bodine pro- 

 ceeds to deal with the further evolution of the antennae with refer- 

 ence to the further development and arrangement of sense hairs, 

 and says little or nothing of scales. Jordan, on the other hand, 



