428 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 



of diffraction bands. Insufficient thinness, however, is by no means 

 the only obstacle to resolving the grain in the structure. 



Thin sections are hi general very transparent, and the only source 

 of variation of intensity in the image formed by the microscope de- 

 pends on the varying amount of retardation affecting the waves which 

 traverse their different parts, that is (since the section is of uniform 

 thickness) on differences of refractive index; but, in order to view such 

 sections at all, it is necessary to mount them in some refractive 

 medium, and this greatly reduces the chance of detecting a fine- 

 grained structure. 



I have tried washing out the bedding material and examining the 

 sections when dry, but, although great care was taken in keeping the 

 cutting edge of the knife smooth and sharp, stria? always appeared in 

 the direction of the cut, which quite obscured the real structure. The 

 fact is, that there are a very few objects on which the highest micro- 

 scopic powers can be used with advantage. Even the test diatom, 

 A. pellucida, which, of course, has to be mounted dry, or in a medium 

 whose refractive index greatly exceeds that of silica, is too thick to 

 give a satisfactory image, and small solids, whose dimensions are less 

 than a wave length, give images which are not their enlarged geo- 

 metrical outlines, but phenomena depending on the wave length. 



Although the microscope, in my hands, at any rate, has failed to give 

 direct evidence of a "periodic" structure, other tests point strongly to 

 "interference" as the origin of the colors. 



In some cases the color-producing film is backed by an extremely 

 opaque layer, and in others the whole of the structure is transparent, 

 and transmits the complementary color with nearly the same intensitj^ 

 as the color reflected. Even where there is an opaque backing, this is 

 often thin enough to allow of examination by strong transmitted 

 light, and the prevailing color is a brown, tinged with the unabsorbed 

 complementary to the color reflected. These opaque backings are 

 present in most feathers and in some insect scales, but in the majority 

 of cases the scales of insects are transparent. 



Both theory and observation show that, when the reflected color 

 depends on interference the tint will change toward the blue as the 

 angle of incidence increases, so that reds become yellows, yellows 

 change to greens, and greens to blues, and also that if the color- 

 producing structure is immersed in a refractive fluid instead of air the 

 reflected color will change toward red and have its intensity reduced. 

 Two causes are operative in producing this change : In the first place, 

 if the color-producing film is protected from the fluid by an imperme- 

 able outer layer with which it is in optical contact the only effect of 

 the fluid is to diminish the angle of incidence of a ray of given 

 obliquity in air, so that the color reflected is that due to the smaller 

 angle of incidence. Secondly, if the fluid penetrates the layers in 



