414 DR. H. GADOW ON THE COLOUR OF FEATHERS. [May 2, 
not applicable to them. Now our surface-coating of 0°004 is about 
ten times thicker than the thickest of colour-producing plates ; 
consequently this surface-coating cannot be the cause of the blue 
colour. Moreover, we find quite a similar sheath surrounding red 
and black feathers; and therefore the function of this sheath will 
be, besides merely protecting the feather, to give the blue colour, 
produced by other parts, a glossy brilliant appearance. This, how- 
ever, does not mean that this transparent sheath is superfluous or 
unimportant to the appearance of blue; since, supposing the blue 
is produced by the underlying cone-stratum, there must be some 
material to reflect this blue to our eye, in a similar way that a piece 
of wood shows its colour up much better when polished or varnished. 
We must therefore look for other reasons for the appearance of 
blue. It is true that all blue feathers contain a yellowish to brown 
pigment ; but the same is the case with many others, like yellow 
and green feathers, and, besides this, a yellow pigment alone can 
never produce blue. 
The most essential part of blue feathers is the layer of prismatic 
columns; but as these vary considerably in size, from 0°011 to 0-003, 
they alone cannot be the essential part, nor can it be the thickness 
of the transparent coat of the little columns themselves, since even 
this extremely thin coating is sometimes too thick to allow the 
application of ‘ thin-plate colours.” 
As the primary cause, we have to consider the fine ridges which 
we observe on the outer surface of all these prismatic columns. 
Many of them are so fine that they are even narrower than the 
length of one wave of light (the length of one wave of red light 
being 0:0007, that of violet being 0:0004 mm.). As they form a 
system of ridges, I am inclined to apply to them the theory of 
‘*Gitterfarben,’’ and I explain the blue colour of feathers by this 
theory. 
The colour produced by thin plates depends entirely on the pro- 
portion of the thickness of the plate to the length of the waves (w) ; 
consequently if this proportion is equal to #, and y is the thick- 
ness of the plate when first looking red, this same colour will appear 
every time when the thickness (y') of the plate has increased to an 
odd multiple of wx. But after a certain thickness is obtained, the 
plate loses the power of producing colour. Very similar conditions 
apply to the theory of colours produced by a system of narrow ridges. 
However as these ridges in the cones are so minute that a cipher 
generally does not appear before the fourth decimal, we are unable 
to measure them with exactness. Moreover these ridges do not 
appear always as straight lines, but seem to be waved; to measure 
the length and deviations of these waves would be mere guesswork. 
Whether this system of ridges is the only cause of the blue colour 
is doubtful ; very likely the transparent coating and the cones them- 
selves will considerably influence the light passing through them. 
The production of blue therefore in a feather would be the result 
of a very complicated process. 
Let us throw only a furtive glance at some of the changes which 
