4 Dr. Thomson on the true Atomic Weight of [Jan. 



those precautions which a knowledge of the existence of error 

 would lead him to look for. 



The mode of experimenting which I have employed has the 

 great advantage of informing the experimenter whether or not 

 nis results be accurate. When I mix together a given weight 

 of sulphate of potash and nitrate of lead, it is easy, by examin- 

 ing the clear liquid which remains after the precipitate has 

 subsided, to ascertain whether that liquid holds in solution any 

 sulphuric acid or any oxide ; thus showing whether either of 

 the salts has been employed in excess, and which of the two. 

 We are thereby enabled to vary the weight of each salt till at 

 last we hit upon proportions which exactly decompose each 

 other without leaving any excess whatever. The knowledge of 

 these proportions enables us, by the methods explained in the 

 papers above alluded to, to determine the real weight of the 

 atom of the bodies which are the subject of experiment. 



This method is tedious, and requires no great share of dexte- 

 rity in experimenting. We must take care that our salts are 

 pure and dry ; that they are accurately weighed out, completely 

 dissolved in distilled water, and the solutions well mixed without 

 any loss whatever. Whoever attends to these precautions will 

 be enabled to come to results as near absolute accuracy as is 

 consistent with experimenting. I have made considerable pro- 

 gress in determining the weights of the atoms of the acids and 

 bases by this method ; but at present 1 mean to confine myself 

 to the experiments made to determine the atomic weights of the 

 five substances whose names are placed at the head of this 

 essay. The following table exhibits the numbers assigned to 

 the atom of these bodies by Dalton, WoUaston, Berzelius, Prout, 

 and myself: 



To enable the reader to compare Berzelius's atoms with the 

 others contained in the table, it will be proper to bring them 

 down to smaller numbers. This may be done by dividing each 

 by 2. They will then be reduced to what they are in the 

 foUowing table : 



• In the fifth edition of my System of Chemistry, I made it 4*5, deceived hy the expe- 

 xiinents of Berzelius and Dalton. But I have long ago corrected this mistake. 



