

THE LIMIT OF KNOWLEDGE. 261 



we know what we have observed, and not a jot 

 more ; and if we think that we do, we are in error. 



Now, when we carefully, attentively, and without 

 any visionary theory or notion formed previous to 

 knowledge, and therefore groundless and delusive- 

 look at nature around us, we find two great classes 

 of natural productions. The one class perfectly 

 passive to the operation of the laws of matter, hav- 

 ing in themselves no principle of change, suffering 

 no alteration though ever so long kept apart from 

 other substances, and altering only when they are 

 affected by something external of themselves. Those 

 substances we can, in many instances, resolve into 

 their elements, or constituent parts ; and we also 

 can, although not in so many instances, reproduce 

 them back again out of the very elements into which 

 they were previously resolved. If we cannot do 

 that, we always can account for all the parts, and 

 say into what other substances they have been 

 compounded ; and scatter them as we may, through 

 any number of combinations, not one of them is lost 

 either in its quantity of matter or in any of its quali- 

 ties ; but in all cases in which we can bring the in- 

 gredients together under the proper circumstances, 

 and these all observable circumstances, we get the 

 original compound, unaltered and undiminished in 

 any one of its qualities. 



These are the substances of which, it has already 

 been mentioned, no part, mechanically considered, 

 is necessary to the existence^ and perfection of 

 another. If we cut them with a sharp instrument, 

 break themi>y a blow, or otherwise divide them by 

 any mechanical operation, all the parts are, size and 

 weight exeepted, just the very same substance that 

 the larger mass was before the mechanical division. 



And as we cannot make them smaller, except by 

 taking away a part, and the part and what is left 

 still make up the whole, so we cannot add to their 

 quantity in any other way than by adding matter of 



