THE COLOUR OF THE KINGFISHER. 469 



above and not below the structural layer. Viewed from the 

 inside, the barb of the green Eelectus is exactly the colour of 

 that of the Kingfisher, i. e. a fine azure blue. Nor is this all, 

 for the blue of the female is produced in the same way, and the 

 purple is due to the barbules and part of the barb being tinted 

 with araroth ; here the red does not overlie the cells, but shows 

 on each side of the layer. The above facts simplify the question 

 of the colouring of the sexes in this interesting Parrot, but 

 cannot be further discussed here. 



In several other green and blue birds these granulated cell- 

 caps are present. The common Amazon Parrot (Chrysotis) is an 

 accessible example. I have examined thoroughly the feathers 

 of but a small number of birds ; but I may as well say that I 

 have never examined a blue, green, or purple feather (I do not 

 now refer to "metallic" feathers) without finding the colour to 

 Le produced by the granular tops of the special cells I have 

 described, but always as blue, modified by red pigment in the 

 purple feathers, and by yellow in the green ones. It must not 

 be forgotten that many apparently blue feathers are really grey. 

 For instance, the blue feathers on the wing of the Shoveller are 

 black, with a fine line of white along the centre of each barb. 

 This is the method of blue production in many birds, but such 

 feathers are never " enamelled." In a similar manner, green is 

 formed by a combination of zoofulvin and melanin, as in the 

 Green Woodpecker. 



What do these facts suggest ? I read it that the bird can 

 only produce blue or blue compounds in this single manner. If 

 blue is required, the protecting outer sheath is colourless ; if 

 green, it is tinted with yellow ; and if purple, a line of blue 

 modifies the colour of a feather pigmented with red. Dark blue, 

 as in the tail of the Kingfisher, or in the feathers of many other 

 birds, is caused by narrow lines of blue-producing cells with wide 

 and black interspaces. I need hardly say that were the colours 

 due to prisms, or to thin plates, the overlying pigment would 

 be unnecessary. Yet, as we see, the base of these colours 

 is always the sky-blue reflected from the caps of elaborately 

 fashioned cells. 



Perhaps I may summarize, then, that the blue colour of the 

 Kingfisher is not due to pigments ; nor, as Dr. Gadow suggests, 



