OTHER THEORIES OF SEXUAL COLOURING 327 



the patterns of so many birds and mammals. The 

 whiteness of a hair or feather is produced, just as the 

 whiteness of snow is produced, by the presence of gas 

 entangled in the loose meshes between the component 

 parts of their substance (see pp. 3-6). We cannot 

 suppose that the surplus vitality which is believed to 

 be efficient in producing some new or especially bright 

 colouring matter on one part, will on another part be 

 equally efficient in withholding it * and in causing the 

 substitution of bubbles of gas. 



But white is not the only difficulty; the most 

 beautiful of all colours in nature, the iridescent tints 

 of many animals, are not due to pigment at all, but 

 frequently to interference of light, the cause which 

 produces the colours of a soap-bubble or that of 

 mother-of-pearl (see pp. 6-10). 



The interference colours of animals are similarly 

 due to fine lines on the surface of structures, or more 

 frequently to excessively thin sheets of air or occa- 

 sionally of fluid, enclosed between layers of denser 

 substance. The varying tints are caused by excessively 

 minute differences in the width of the chinks in which 

 the air is contained. But it would be a very rash 

 hypothesis which suggests that a surplus of vitality 

 regulates the width of these chinks to the production 

 of this or that colour. There is absolutely no reason 



' When a permanent white patch appears upon a mammal, the 

 pigment is withheld ; it is only retained, and masked by the forma- 

 tion of gas-bubbles, in the whitening of existing dark hairs (see 

 pp. 98, 99). 



