x Boussinesq's Mathematical Argument. 179 



fectly complete, yet every increase in the accuracy 

 and perfection of astronomical knowledge brings us 

 nearer to such absolute proof ; and it seems extremely 

 improbable that they should be subject to any limit 

 whatever. The Brownian motions, the motions of 

 the molecules of gases, the undulatory motion which 

 constitutes light all these, however minute, are 

 motions, and we cannot doubt that they are rigidly 

 subject to the laws of motion. It is uncertain how 

 far chemical actions can be resolved into the motions 

 of atoms ; but the law of the absolute invariability of 

 chemical properties and actions the proof of which, 

 it is true, can never be complete, though every in- 

 crease of chemical knowledge strengthens it makes it 

 probable, with a probability approaching indefinitely 

 near to certainty, that the laws of chemical action 

 admit of no more limitation or exception than the 

 laws of motion. We must consequently hold with 

 scientific men generally, that all motions, whether on 

 a planetary or an atomic scale of magnitude, are 

 determined by the laws of motion, with a certainty 

 which, though not mathematical in its nature, is 

 practically equal to mathematical certainty. 



But do the laws of motion ensure absolute de- 

 terminism 1 An attempt has been made by Professor 

 Boussinesq, of Lille, to show that absolute determinism, 

 though generally true in mathematics, is not always 

 so, and therefore is not necessarily always true in 

 mechanics. 1 He chiefly makes use in his argument 

 1 See Paul Janet's article in the Contemporary Revieiv, June 1878. 



