distribution to hairs (that is "sense hairs") that is too wide owing to 

 its including areas with "bristles," and no "sense-hairs." I have 

 avoided this as far as may be by seeing that usually the hairs are of 

 the same characters over all the area noted as clothed with hairs, 

 *. e. that usually none of the hair-covered surface is devoted to one 

 kind of hair to the exclusion of others. 



I ought, perhaps, earlier to have given some account of a lepi- 

 dopterous antenna when fairly evolved, to enable any member who 

 has not studied antenna to see better whither evolution is guiding the 

 primitive antenna we have just considered. 



There is a curious parallelism between the pupal structure and the 

 antennal structure in this respect : that throughout the generalised 

 moths that constitute the Incompletae there is great variation in 

 antennal as in pupal structure, but all show some tendency — one 

 might almost say desire — to acquire the structure characteristic of the 

 Obtectae both in pupae and antennae ; whilst again, both in pupa? and 

 antennae a somewhat different departure is taken by butterflies. 



The obtect antenna obtains throughout nearly all the Phalenae 

 Obtectae and the Pyrales. It also occurs in some Tortrices and else- 

 where. It is, therefore, of- greater range than the obtect pupa. It 

 also presents a much greater variety of forms than is to be found in 

 the obtect pupa. 



Reduced, however, to its essentials, and omitting consideration of 

 its numerous varieties, it may be very simply described. 



It has an upper outer surface covered with scales, to the exclusion 

 of hairs, and an inner lower surface covered with hairs, to the exclu- 

 sion of scales. 



I ought also to say that I am considering throughout only the 

 clavola or flagellum of the antenna, omitting the two proximal joints, 

 the scape and pedicel. 



Hairs and scales are, no doubt, homologous structures ; and it is, 

 no doubt, very difficult to say what is the precise function of any par- 

 ticular hair in an insect. 



Nor do we know how the scales of Lepidoptera arose — whether 

 they arose de novo from epidermal cells that were not previously 

 hairs, or whether they were preceded by, that is arose, by a change in 

 the development of pre-existing hairs. 



Let us suppose the latter to be the case, since I think it is the 

 usually accepted theory. Why it should be I am not quite clear, 

 since the preceding forms, Panorpidae and Hemerobiidae, are not 

 remarkable for abundant hairs ; whilst scales and equivalent hairs are 

 very abundant in Lepidoptera, Trichoptera, and lower Diptera that 

 take a common origin from somewhere amongst them. However, 

 let us suppose scales are developed from preceding hairs : are they 

 developed from any hairs — from mere protective hairs, from tactile 

 hairs, or also from special sensory hairs ? A priori, it seems that 

 mere protective hairs, if there are such apart from tactile hairs, are 

 the most likely sources, and that tactile hairs also might originate 



