224 MR. A, MALLOCK ON COLOUR-PRODUCTION 



form rather brilliant objects when viewed between crossed 

 nicols. 



But by far the most brilliant colouring of birds has its origin 

 in interference, that is to some periodic structure in the sub- 

 stance of the feather, vv'here the spacing of the parts is a multiple 

 of the half wave-lengths of the light they reflect. Such is the 

 case among humming-birds, sun-birds, peacocks, birds of paradise, 

 and ducks, to mention only a few instances of what are spoken 

 of as "metallic" colouring. 



All these colours disappear when subjected to pressure, and in 

 all cases the colour-producing substance is confined to a very thin 

 layer overlying an intensely opaque black or brown substratum. 



The general effect when viewed from a distance depends to a 

 great extent on the form of the surfaces on which the colouring 

 layer is disposed. If these surfaces are planes, the relative 

 positions of the eye and source of light with regard to the feather 

 has to be rather carefully adjusted, in order that any colour may 

 be visible. 



When, however, the surfaces are rounded, the range of incidence 

 is mnch extended, and from almost any point of view some 

 colour appears, although the intensity is lessened, just as a tray 

 of small glass beads will scatter sunlight in all directions 

 although the intensity in any one direction is much less than 

 what would be produced by a plane mirror adjusted to reflect 

 lights in that direction only. 



The examples chosen all show distinctive structure peculiar to 

 the orders to which the examples belong. 



A feather may be described as consisting of a stem, branches, 

 and leaves (text-fig. 1) (named by zoologists respectively rachis, 

 rami or barbs, and barbules, the latter sometimes as carrying 

 barbicels). It is in the modifications of the leaves that the 

 distinctive features are found. 

 I will notice these in order. 



In all the ornamental feathers of Humming-Birds the branches, 

 but especially those of the gorget, are so bent that their ends 

 are parallel and the colour-bearing surface is on the leaves. A 

 cross-section of the branches and leaves forms a succession of 

 hollows in text-fig. 2 e. 



The section of the leaves themselves is something like the 

 numeral 7 (text-fig. 4 c), and the colour-producing material lies 

 in a thin layer on the upper surface of the leaves, and is nearly a 

 plane, so inclined that the normal makes an angle of about 60^ 

 with the branch, but lying in a plane parallel to the latter. (This 

 angle varies in diff'erent species.) 



To keep the leaves in this position, each barb terminates in a 

 curved plate (text-fig. 4(i), which interlocks with the similar plates 

 of tlie two or three leaves in advance. 



Each branch ends in a long bristle (fig. 2 6), which, when the 

 plumage is in good order, lies in the trough of the valley formed 



