THE COLOUR OF THE KINGFISHER. 465 



that the yellow pigment in green feathers always underlies the 

 structural layer. The inner surface of the sheath is often 

 (perhaps invariably) marked out in hollows corresponding ex- 

 actly in shape with the caps of the cells below. There is a strong 

 probability that these serve as tiny lenses, and increase the 

 brilliancy of the parts below. 



A single example of the polyhedral cells may now be examined 

 in detail. It is about one-thousandth of an inch in breadth, 

 and rather more in height. Its inferior wall or floor is formed 

 by the upper part of the pigmented base of the barb, and the 

 vertical walls are perfectly smooth, transparent, and colourless. 

 The cell is empty, and the coloured nucleus figured by Fatio 

 does not exist in fact. 



None of the feather elements so far considered have any part 

 in the production of the blue colour. The pigmented parts of 

 the barb, the transparent sheath, and the vertical cell- walls may 

 all be removed without destroying the blue. This is only pro- 

 duced by the tops or caps of the polyhedral cells. In spite of 

 the fact that these objects rarely exceed "025 mm. in length, it 

 is comparatively easy to separate one of the caps, and to arrange 

 it for examination under the high powers of a microscope. 



Even by means of a good hand-lens this tiny plate of ceratin 

 is seen to be brilliantly blue. Under the compound microscope, 

 using the highest power possible with incident light, the surface 

 is seen to be finely granular, without the slightest trace of the 

 ridges seen by Dr. Gadow in Pitta. I am unable to see the 

 shape of these grains or wrinkles, but in the aggregate they look 

 like the roe of a fish. The largest of them are plainly less than 

 the one-fifty-thousandth of an inch apart, for I have counted 

 forty separate cells in the length of a millimetre of barb, and 

 estimated far more than this number of grains across a single 

 cell top. 



This plate of ceratin is not flat, but slightly conical (or 

 patelloid), especially in the centre of the barb. In manipulation 

 the apex may be pushed down into the middle of the ceil, and 

 so be rather misleading. Under a proper light it is all over 

 brilliantly blue, and the colour and structure are the same on both 

 surfaces. This highly important fact can be seen by mounting 

 a single cap very firmly between two thin cover-glasses and 



