1815.] An Address to Chemists. 12S» 



of oxides with each other, that we can hope to penetrate into the 

 scLTfts ol' ofiranic coinposition. When that law is demonstrated, 

 the theory of chemical proportions ma}' be considered as esta- 

 blished. 



It is absolutely necessary, that every jierson who repeats an 

 analysis in order to ascertain its accuracy, should be acquainted 

 with all the requisites to make an exact analysis, in the same 

 decree as the auiiior of tlie work which lie proposes to examine. 

 Otherwise, he takes upon him to judge his master. Accurate 

 ehemical analysis is half an art, half a science. He who does not 

 unite both in one, will never be in a condition to make an exact 

 analysis. VVe must always chu*e that method in which the pre- 

 cision of result depends the least on tlie dexterity of the operator. 

 We must likewise be able, from long experience, to avoid all the 

 circumstances by which the result would be rendered inaccurate, 

 that neither from forgetfulness, carelessness, or ignorance, we rriay 

 neglect to observe and remove them. These circumstances are 

 generally of such a nature that chemists usually pay no attention to 

 them; yet they are of coiisideral)le importance when our object is 

 to obtain the utmost possible accuracy. The most skilful chemist 

 will continue long an apprentice in the art of analysis, if he has 

 not been in the habit, from the commencement of his studies, to 

 attend to these circumstances. Long experience is even often 

 necessary before we discover their existence, unless we have the 

 advantage of being informed by others. 



When we have to judge i)efvvccn two different analytical methods, 

 or between two different chemists, we must be well acquainted 

 with the metliods employed ; and boili tlie method and the repu- 

 tation of the author ought to enter for something in our judgment. 

 If, for example, we have to comp^ire two experiments oii the com- 

 position of a m^'tallic oxide, one of which was made by dissolving 

 the metal in nitric acid in a phial, from which the acid was then 

 evaporated, and the residue expo«;ed to a strong heat : if the other 

 was made by precipitating the solution in nitric acid by means of 

 an alkali, filtrating, &c., it is presumable, that supposing the same 

 dexterity in the operators, the first experiment is more accuratt* 

 than the second, because it has been less exposed to the influence 

 of circumstances. It is necessary, however, that ho who judges of 

 two results, should know the circumstances that may reiider the 

 one or the other inexact. We must know, for example, that the 

 first experiment can scarcely have any otlicr error, than giving too 

 great a quantity of o.xygcn to the oxide, 'i'his may be either owing 

 to the application of too little heat, or (which is more usual) to 

 the glass not having resisted the action of the nitric acid during the 

 evaporation, s-o that a little nitre is formed, the acid of which in- 

 creases the weight of the oxygen that we think wc have found. 

 (This circumstance long deceived me in my first experinu nts before 

 1 perceived it.) The oilier method of operating can have no other 

 error than indicating loo little oxvgcn, if aiuinonb has beeu «m- 



VoL. V. N^ II. I 



