90 Berzelius's Experiments to determine the Composition [Feb. 



can go, and to give an equal degree of precision to the analyses 

 of the most important bodies. If by these efforts the inevitable 

 errors in the results become proportional, even though they are 

 not absolutely correct, they will be of the same utihty to us as if 

 they were so, because the mistakes are all within the limit of our 

 observation, and as nearly as possible proportional for all combi- 

 nations. 



In endeavouring to determine to what degree of accuracy it 

 was possible to arrive, I have found that by very simple methods 

 it was possible to come within one-thousandth part of the weight 

 employed ; so that variations in the results of experiments made 

 by the same process difi'er only from each other in the ten-thou- 

 sandth parts ; but even this degree of accuracy requires much 

 care and attention in all the circumstances which may contribute 

 to render the results inaccurate, and which usually vary with the 

 manner of operating. I have never been able, except accident- 

 ally, to make the results coincide beyond the ten-thousandth 

 part, and indeed very frequently, notwithstanding all my care, 

 they have not agreed perfectly beyond the hundredth part. 



Some of the experiments, which I intend to relate, were made 

 with the views just stated. Others are of an older date, and 

 have been already related in the third volume of the Annals of 

 Philosophy by Dr. Thomson. But among these last several have 

 been repeated, and some mistakes in my original memoir have 

 been corrected. I shall first give a description of the experi- 

 ments in which I endeavoured to obtain the greatest possible 

 precision ; I shall then state those whose accuracy I cannot gua- 

 rantee to the same degree. 



Experiments to determine the exact Composition of the Bodies, 

 which in a great Nnmber of Analyses serve as Bases to the 

 Calculations of Chemical Proportions. 



Muriate of Potash and Muriate of Silver. — Of all the methods 

 to determine the exact quantity in potash and in the oxide of 

 silver, none seems to me more likely to be exact than to deter- 

 mine in the first place the quantity of oxygen in the oxymuriate 

 of potash, and afterwards to analyze the common muriate. By 

 experiments already known, it has been established that oxymu- 

 riate of potash exposed to a red heat loses ten times the quantity 

 of oxygen contained in the potash of the residual muriate. If we 

 determine accurately the quantity of oxygen disengaged from the 

 oxymuriate, and then calculate, by the way just stated, the quan- 

 tity of oxygen which the potash should retain, the error resulting 

 will amount only to ^th of the error of observation in the analysis 

 of the oxymuriate. 



It is long since I published an analysis of the oxymuriate of 

 potash by which I had found that 100 parts of that salt give out 

 38-845 parts of oxygen ; but to reach a greater degree of preci- 

 sion I considered it as requisite to repeat the experiment, la 



