1883.] on the Size of Atoms. 197 



ferent in diflerent mediums, and should in most cases be smaller in 

 the denser than in the less dense medium, is quite what we should, 

 according to dynamical principles, expect from any conceivable consti- 

 tution of theluminiferous ether and of palpable transparent substance. 

 But that the velocity of propagation in any one transparent substance 

 should be different for light of different colours, that is to say, of 

 different periods of vibration, is not what we should expect, and could 

 not possibly be the fact if the medium is homogeneous, without any 

 limit as to the smallness of the parts of which the qualities are com- 

 pared. The fact that the velocity of propagation does depend on the 

 period, gives what I believe to be irrefragable proof that the substance 



Fig. 3. 



Diagram of Huyghen's construction for wave front of refracted light. 

 Drawn for light passing from air to flint glass. 



of palpable transparent matter, such as water, or glass, or the bisul- 

 phuret of carbon of this prism, whose spectrum is before you, is not 

 infinitely homogeneous ; but that, on the contrary, if contiguous por- 

 tions of any such medium, any medium in fact which can give the 

 prismatic colours, be examined at intervals not incomparably small in 

 comparison with the wave-lengths, utterly heterogeneous quality will 

 be discovered ; such heterogeneousness as that which we understand, 

 in palpable matter, as the difference between solid and fluid ; or 

 between substances differing enormously in density ; or such hetero- 

 geneousness as differences of velocity and direction of motion, in 

 different positions of a vortex ring in an homogeneous liquid ; or 

 such differences of material occupyiijg the space examined, as we 

 find in a great mass of brick building when we pass from brick to 

 brick through mortar (or through void^ as we too often find in Scotch- 

 built domestic brick chimneys). 



