418 DR. H. GADOW ON THE COLOUR OF FEATHERS. [May 2, 



red lying at r (i. e. towards the upright edge of the prism), and with 

 the violet at v. 



No rays, or at least no visible ones, are thrown out into the space 

 between! and 2 ; consequently with our eye at 1, corresponding to 

 position A, we shall perceive black. At 3 the first red rays will 

 become visible, at 4 the blue ones, and so forth, till at 6 we come 

 across the ultra-violet rays, where we see again black, corresponding 

 with position C. Between 1 and 6 will be a place from where we 

 can look at the object under full light (position B) ; and this, of 

 course, is the way in which we generally describe an object. 



As this agrees with observation, i. e. as every metallic feather (if 

 examined in the way explained above) shows precisely the same 

 phenomena as a prism under similar circumstances would show, we 

 have every right to consider the explanation of " metallic varying 

 colours" as proved. There are, however, several observed facts 



Fig. 2. 



z^- 



Diagrammatic section through the barb of a " metallic " feather. 



which need an explanation, since they seem rather to upset this 

 theory. 



First, why does not every metallic feather display all the colours 

 of the spectrum ? and why do they generally range not over more 

 than a few neighbouring colours? Of course any prism, however 

 small it be, displays all the colours of the spectrum ; but this does 

 not mean that all of them reach our eye ! Part of the spectrum 

 might be hidden by some other object standing between it and our 

 eye ; for instance we can easily cut off either end of a spectrum by 

 a screen. In the feathers the screen would be represented by a 

 neighbouring radius, so that, as, for instance, in fig. 2, the lower half 

 of the spectrum j3 may be concealed, in which case that feather 

 would only vary between red and greenish. Or two neighbouring 

 prisms, even if they belong to the same radius or barbule, may be so 

 situated that their spectra partly overlap one another. This would 

 have a double result: first, that where two complimentary colours fall 

 upon each other they would simply produce white light; secondly. 



