JJSr THE FEATHERS OF BIRDS. 223 



periods of the incident light, and that the colours which they 

 reflect or transmit are due to resonance. 



The actual dimensions, forms, and rigidities of molecules are 

 unknown, but their diameters are apparently of the order of a 

 ten-millionth of a centimetre; and one may inquire what the 

 longest natural period of a sphei'e 1/10,000,000 cm. diameter 

 would be if it were as rigid as steel, or, which comes to the same 

 thing, wba,t is its least natural frequency. Without going into 

 the details of the calculation, it may be stated that this least 

 natural frequency is somewhere about 5 x 10^^ vibrations per 

 second. All the other natural modes of vibration would have 

 higher frequencies, and there is good I'eason to suppose that the 

 rigidity of molecules far exceeds the rigidity of the matter formed 

 by their aggregation. The frequency of yellow light is 5x10^* 

 vibrations per second. Thus it seems that the natural frequencies 

 of molecules and of visible light-waves ai'e at any I'ate of the 

 same order. 



All tliese pigmentary and dispersion colours depend on the 

 constitution of the molecule itself. In the second class it is the 

 relation of the size of particles, or on their disposition in space 

 as compared with the wave-length which determines the selective 

 influence of the matter on white or composite light. 



The origin of the colouring presented to view as the result of 

 selective action of the structure on wave-length can be fairly well 

 determined by the following tests : — 



(a) Mechanical compression or extension, 



(&) Immersion in various fluids. 



(c) Change of colour with the angle of incidence of the light. 



Of these, the compression test is the most decisive ; for, if the 

 mechanical distortion of the structui'e changes or obliterates the 

 colour, it may be assumed that the colour itself depends on some 

 special arrangement of the parts, and not on the molecular pro- 

 perties of the material of which it is built up. 



There may be some apparent exceptions, as for instance when 

 a material transmits one colour with less loss than another, so 

 that the predominating colour is dependent on the thickness of 

 the layer through which the light travels (e. gr. manganese glass 

 or a solution of chlorophyll). 



In the circumstances, however, in which this test is applied 

 to organic structures, such as feathers, these exceptions will 

 hardly operate. 



The greatei' part of the colours of feathers have their oi'igin in 

 pigments of the nature of which little is known. Except in 

 one instance, no solvent has been found for them, and the 

 pigments themselves vaiy much in physical properties. Some 

 are neaily opaque, while others are transparent and transmit the 

 complementary colour. Many, again, polarize the incident light, 

 and this is especially noticeable with transparent yellows. These 



