1882. ] DR. H. GADOW ON THE COLOUR OF FEATHERS, 419 
that through the combination of two different colours, as, for instance, 
yellow and blue, there would appear a third one, in this instance 
green. Thus, spectrum a, produced by the prism a, is partly 
covered by the spectrum £, produced by prism 4; the blue of 8 
covers the yellow of a. Now as red and green give white, and blue 
and yellow give green, we should in this case probably see only the 
colours red and orange, produced by prism a. 
Another circumstance, which might make this process very com- 
plicated, is implied in the consideration that the surface of the pris- 
matic sheath is frequently uneven. How many different systems of 
prisms result from this arrangement, and how in the purple feathers 
of an Aithopyga the rays of light become broken by the surface- 
ridges into blue and violet, we are unable to explain. 
A third phenomenon, which needs explanation, is that some of 
the most gorgeously metallic feathers cannot be made to look black 
in position A or C. An example of this is the beautiful coppery-red 
to deep blue of aJacamar. Under the microscope the compartments 
of the radii of such a feather are extremely convex, as in fig. 5, 
Plate XXVIII. ; consequently there will be always some part of such 
a compartment which presents a vertical plane to the eye, and which 
therefore is always more or less in position B. 
Now to sum up. We have to distinguish between several cate- 
gories of colours in feathers. 
1. Objective chemical colours directly produced by pigment. To 
these belong black, brown, red, orange, and yellow. 
2. Objective structural colours. The feather may contain no pig- 
ment at all, and the colour be produced solely by a special structural 
arrangement of the feather-substance, for instance white, and fre- 
quently yellow; the latter if the surface is composed of very fine 
and narrow longitudinal ridges, Or the feather contains a yellow to 
brownish-black pigment, and the colour actually observed, as green, 
blue, and violet, is produced by a specially produced and particularly 
constructed transparent layer between the pigment and the surface. 
Of non-changing colours 4/we and violet are always structurally 
objective. Green seems to be only in a few cases the result of 
yellow pigment combined with blue surface-structure. In most 
cases it seems to be not a mixture of two colours, but due to yellow- 
pigment light being broken into green. A green pigment seems to 
be very exceptional. 
3. Colours which change and which entirely depend on the position 
of the light and eye. They are produced by a transparent sheath, 
which acts like a prism. Any changing colour represented in the 
solar spectrum may be thus produced in feathers. 
The facts which I have laid down in this communication indicate 
the desirability, nay even the absolute necessity, of a standard 
method, not only of describing but also of drawing birds and other 
animals which show metallic colours. These three standard positions, 
I venture to submit, should be the following :— 
Position A, in which the eye is placed between the bird and the 
light, the eye and light almost in a level with the planes to be examined. 
