EMPIRICAL KNOWLEDGE, EXPLANATION, &c. 187 



in a hypothetical, but not therefore a less certain 

 manner. 



Though only two of the metals, gold and silver, have 

 ever been observed to be transparent, we know on the 

 grounds of theory that they are all more or less so ; we 

 can even estimate by theory their refractive indices, and 

 prove that they are exceedingly high. The phenomena 

 of elliptic polarization, and perhaps also the theory of 

 internal radiation 1 ', depend upon the refractive index, and 

 thus, even when we cannot observe any refracted rays, 

 we can indirectly learn how they would be refracted. 



In many cases large quantities of electricity must be 

 produced, which we cannot observe because it is instantly 

 discharged. In the common electric machine the cylinder 

 and rubber are made of non-conductors, so that we can 

 separate and accumulate the electricity. But even a little 

 damp, by serving as a conductor, prevents this separation 

 from enduring any sensible time. Hence there is little 

 or no doubt that when we rub two good conductors 

 against each other, for instance two pieces of metals, 

 much electricity is produced, but instantaneously con- 

 verted into some other form of energy. Dr. Joule, indeed, 

 believes that all the heat of friction is but transmuted 

 electricity. 



As regards phenomena of insensible amount, Nature is 

 absolutely full of them. We must, in fact, regard those 

 considerable changes which we can observe as the com- 

 paratively speaking infinitely rare aggregates of minuter 

 changes. On a little reflection we must allow that no 

 object known to us remains for two instants of exactly the 

 same temperature. If so, the dimensions of objects must 

 be in a perpetual state of variation. The minor planetary 

 and lunar perturbations are indefinitely, or rather in- 

 finitely numerous, but usually too small to be detected by 



h Balfour Stewart, 'Elementary Treatise on Heat/ ist edit. p. 198. 



