The Evolution of the Sciences 



separate, or rather, since we have no means of 

 separating them, they act, so far as we are 

 concerned, as elements, and we must only 

 assume them to be compound when experiment 

 and observation have supplied the necessary 

 proof." 



We are thus led to consider compounds as 

 formed by the association of a certain number of 

 elements which they contain. We are here 

 leaving the domain of facts for a hypothesis 

 which has been disputed by prominent chemists, 

 and among others by Ostwald: that oxide of iron 

 can always be decomposed into oxygen and iron, 

 and can be obtained by the combination of these 

 elements, is an incontrovertible fact, but that 

 the oxide of iron actually contains oxygen and 

 iron is not at all certain. A circular motion may 

 be decomposed into two oscillating motions along 

 two diameters of the circumference, and may be 

 produced by the superposition of these two move- 

 ments, and yet not actually contain either. 



As a matter of fact all chemists implicitly 

 admit this hypothesis, that the millions of bodies 

 now known are derived from the association of 



seventy-two different elements, which cannot be 



6 4 



