XV111 PREFACE. 



therefore, may be used without the risk of any great 

 error, except those of the arseniates and phosphates, 

 where the error is considerable ; but they consist of much 

 more complex numbers than mine, and the atomic weights 

 are all kept out of view by his method of giving the 

 constituents in 100 parts of the compound bodies. On 

 the score of facility and simplicity, I consider my numbers 

 as much preferable : indeed, the theory of Berzelius is so 

 complicated and intricate, that it would be surprising 

 if it were a true representation of what takes place in 

 nature. 



A few of my determinations of the atomic weights of 

 various bodies, were published in the Annals of Philo- 

 sophy, during the years 1820 and 1821. But on per- 

 ceiving that I had occasionally fallen into mistakes by 

 stating results deduced from particular views of a subject, 

 before I had been able to consider it in all its bearings, 

 I thought it safer to refrain, till I had brought the in- 

 vestigation to a conclusion, and then to lay the whole 

 at once before the public. This was the reason that 

 induced me to discontinue my original plan of publishing 

 my labours in periodical essays, in proportion as the ato- 

 mic numbers were determined. 



