34 



Taking next the bristles or sensory hairs, we find these in many 

 Hemerobiidce. confined to the nervures and wing margins ; in Panorpa 

 they are found also on the general area of the wing. It is generally 

 assumed that the scales on Lepidoptera arose by modification of 

 these hairs or bristles, and there is much to be said for this hypo- 

 thesis. For example, in Culicida, Diptera that usually possess scales 

 like those of Lepidoptera^ there are many species that resemble the 

 Hemerobiidce. in possessing spicules all over the wing surfaces and, on 

 the nervures — not bristles but scales ; in other species of Diptera, 

 especially Cecidomyiads, scales of a rather hair-like structure, but 

 certainly not bristles, occur all over the wing amongst the spicules. 

 In one Culicid, at any rate, I have noticed the marginal scales be- 

 come more and more bristle-like towards the base of the wing. The 

 evidence from the Diptera is therefore very strong that scales arose 

 by modification of bristles. When we come to Lepidoptera, it is 

 necessary to note a very common misapprehension. Many Lepidop- 

 tera possess abundantly, a majority possess some, so-called hair- 

 scales, and it is implied, often distinctly stated, that these hair- scales 

 are an intermediate form between hairs and scales. That some 

 so-called hair-scales are in an intermediate state between hairs and 

 scales seems nearly certain, but the immense majority of " hair- 

 scales " are really much modified scales — are, that is, less, instead 

 of more, allied to hairs than scales are ; they are scales that are 

 developed, or degenerated, if you prefer it, by becoming very long 

 and very slender. Such are the hairs on the wings of the higher 

 FsychidcB, where typical scales are rare, whilst the usual positions 

 of scales over the wing surfaces are occupied by hair-like struc- 

 tures that are evidently, however, scales, and not hairs. 



How, then, shall we distinguish a bristle from a scale, that we 

 may be able to judge whether a given hair-scale is a bristle modi- 

 fied in the direction of a scale, or a scale still further modified ? 



A typical bristle is easily distinguishable from a typical scale. A 

 bristle is circular in cross-section, tubular, conical, tapering regu- 

 larly from base to apex, usually sharp-pointed, has no pigment, but 

 consists of a brown chitin, not bleached so readily as are scales 

 coloured with pigment. Its attachment is by a delicate membrane 

 to a superficial circle on the chitinous surface, and thus it has con- 

 siderable freedom of movement. 



A scale is flattened, often striated, colourless apart from pig- 

 ment, attached usually by a specially narrow process, which is 

 enclosed in an hour-glass shaped cup, through the open top of 

 which it proceeds downwards, then through the central constriction, 

 and terminates some way below the surface in the lower globe, 

 and so it possesses very little or probably no mobiHty. 



If w^e examine a wing of Micropteryx, as the earliest stage of a 

 Lepidopterous w'ing that is available, we find unmistakable scales, 

 broad, flat, pigmented, and typically attached to the wing, over 

 the whole of the wing surface and forming the ciliae ; if we ask 



