XIII 
THE COLOURS OF BIRDS 
271 
like the tiles of a roof. The ultimate causation 
of the physical colour Gadow ascribes to the trans¬ 
parent sheaths of keratin covering these compart¬ 
ments, which he thinks act like a series of prisms. 
An important point in connection with these metallic 
barbules is, that they are so modified that both 
hooklets and folds are completely lost, and therefore 
there is no connection between the barbules or the 
barbs. Metallic feathers of this type have therefore 
a peculiar looseness of texture which is, for example, 
very obvious in the ornamental feathers of the 
peacock ; the solidity of the flattened metallic 
barbules gives, however, to such feathers an appear¬ 
ance quite different from that of ordinary downy 
feathers, in which also the barbs are unconnected. 
The unconnected nature of the barbs is of especial 
interest, because it would render the feathers quite 
unfitted for purposes of flight if the variation were 
to occur in quill-feathers. In sun-birds it is usually 
the contour-feathers which are metallic, rarely the 
tail-quills, and apparently never the wing-quills. 
Development of Colours .—The types of coloration 
already described in the sun-birds are seen in the 
specialised feathers, especially of the male. In the 
unspecialised feathers, such as the general contour- 
feathers of the female, we find what may be regarded 
as the primitive condition. These feathers are of a 
dull olive colour, and are divided into three regions—a 
basal downy region usually of an ashy colour, a median 
slightly V-shaped region in which the barbules have 
a very close texture and are of a brown colour, an 
apical region in which the barbules are unconnected, 
slightly modified, and faintly pigmented with dark 
