KELLOGG: TAXOXOMIC VALUE OF SCALES OF LEPIDOPTERA. 69 



ized by a more highly specialized condition of scale-covering within 

 its borders, than is shown on the rest of the wing. 



An exaaiple of the loss of sharpness in marking when the scales are 

 not specialized is presented in Citlwronia rt'ga//s. The covering of 

 the forewing is pretty generalized, consisting of long one- or two- 

 pointed scale-hairs and shorter, more specialized, three- to several- 

 pointed scales. In the yellow spots which occur on the wing there 

 is little or no difference of condition of the wing-covering, and the 

 limiting or bounding of these spots is much less sharp than in the 

 case of the color spots of butterflies. On the hindwings the covering 

 is much more generalized than on the forewings, consisting almost 

 entirely of long, soft scale-hairs. Here the yellow regions fade out 

 into the reddish-brown ground with little definiteness of outline. 



The colors of scales are produced by two causes: (i), the presence 

 of pigment; (2), the overlapping, lamination and striation of the 

 scales which produce those familiar but striking optical phenomena 

 due to the interference of the waves of light. These two sorts of 

 colors have been called by Hagen "natural colors," referring to the 

 colors due to absorption of waves of certain lengths and the reflec- 

 tion of the others by pigment, and " optical colors," due to wave 

 interference incident to reflection through the lainimt and striated 

 surfaces (diffraction gratings) of the scales. Combinations of these 

 causes are usually present so that the resulting color effects are prac- 

 tically incapable of analysis. 



As the pigments mostly transmit the same colors as they reflect 

 (the colors complementary to the colors absorbed) the colors of 

 scales which produce color by pigments are usually the same by 

 transmitted light as by reflected light. But in the case of "optical 

 colors" (colors by interference of reflected rays) this is not the case. 

 The scales producing these colors are often transparent, as with 

 Microptcrxx, and these when viewed by transmitted light are colorless, 

 or they contain pigment and when viewed by transmitted light a 

 color is seen which is due wholly to the pigment and not at all to the 

 structural features (lamination and striation) so that it may be entirely 

 different from the colors seen when the scales are viewed by reflected 

 light.* 



^Vhere the scales do not overlap they present two superposed lamellae 

 (the opposed cuticular sides of each scale); where they irregularly 

 overlap there are at some points two and at other points four super- 

 posed lamella; and where the scales are arranged in double-tiered 



*An interesting expariment is to take a wing of Lyaeaa or Morpho and examine it 

 under tne microscope by reflected li.ulit and by transmitted lisht. In examining by 

 transmitted light care .should tie taken to prevent any light falling on the wing from 

 above. The colors will be very different by tlie two lights, 



