A SYSTEM OF PARTICLES. 251 



18. This argument, undoubtedly, does not carry with it much 

 weight, although it appears to a certain extent plausible. It will be 

 extremely difficult to attempt a formal proof of the law after which 

 we are seeking. Experimenters have had their attention drawn to any 

 thing rather tlian this; for it must be remembered that it is not by 

 the examination of broad facts and familiar jihenomena that we can 

 generally construct a Theory, but by the intricate pursuit of the slight 

 deviations from, and apparent exceptions to, the general rule. 



Supposing such a law established as we have given above, it is 

 manifest, that the velocity must be decreased in the higher regions of 

 the atmosphere, whilst the variation of the velocity, due to an altera- 

 tion of the length of the wave, is increased in the same ratio. 



We might then hope to arrive at satisfactory conclusions, and indeed 

 completely establish a law of repulsion, could we make observations in 

 the higher regions of the atmosphere, as for instance, on or near the 

 summits of considerable mountains. 



At the same time, supposing such experiments have been made, 

 and that the apparent result of them is contrary to that obtained, on' 

 the above supposition, I should still hesitate before I reject a law of 

 force which is, as far as we at present know, the universal law for 

 all particles not in actual contact. 



The case here is widely different irom that of light. In the latter 

 the length of a wave is only about the ^^ part of an inch; in the 

 former, it is several inches; calling to mind then that even for lio-ht 

 the difference of velocity bears but a small ratio to the whole we 

 should not expect, unless some of the other quantities proportionally 

 mcrease, that a difference so minute would be sensible to the ear 

 Should this be the case it is unfortunate, as it deprives us of a ready 

 mode of ascertaining the fundamental property which pervades the 

 whole. 



19. But in the absence of the observations requisite for obtaining 



accurate results, we may still discover, by some indirect process 



Vol. Vr. Paut II. ]f ,5 r j 



