368 PROCEEDINGS OF THE AMERICAN ACADEMY 



chloride was obtained, except that it was prepared with " very pure 

 antimony and dry chlorine, and distilled several times over finely 

 powdered antimony." We have, consequently, no guaranty for its 

 purity beyond the assertion of the author. But this is a point upon 

 which we are entitled to demand the most precise information in a 

 work which claims not only to be an exact determination of a chemical 

 constant, but also to decide the question whether the equivalents are 

 or are not multiples of that of hydrogen. It is well known that upon 

 this very point, the purity of the antimony obtained by a certain pro- 

 cess, there was for a long time a disagreement between chemists as 

 eminent as Berzelius and Liebig ; and that it was finally conceded 

 that the metal so prepared does contain an appreciable quantity of 

 arsenic. Now what proof have we that the antimony used in these 

 determinations may not have been prepared by the process of Liebig ? 

 and that the trifling differences in the results of M. Dumas and myself 

 may not have been caused by the presence of a minute quantity of 

 arsenic, a substance difficult of detection when mingled with a large 

 amount of antimony, and whose atomic Aveight is lower than that of 

 antimony in the proportion nearly of 3 to 5 ? If it be considered 

 unfair to M. Dumas to assume that his material was impure, it is 

 certainly fair to demand that he should give the proof that it was not. 



The properties of the chloride of antimony, also, are such as to 

 render extreme accuracy difficult of attainment. A body which at- 

 tracts moisture with the utmost avidity, and which, if every trace of 

 moisture be not excluded, gives on distillation a product containing 

 free hydrochloric acid, is not one which would willingly be selected 

 for the purpose of a normal determination. M. Dumas says that he 

 tried without success to effect the determination by the reduction of 

 the oxide and of the sulphide, and by oxidation of the metal with 

 nitric acid.* 



But the last-named process gave Berzelius agreeing results, and I 

 have shown that the accuracy of those obtained by myself in the same 

 way is about twice that of the determinations made upon the chloride. 

 It is not easy to believe that a process which has succeeded in the 

 hands of others would offer any difficulty to so consummate a manipu- 

 lator as M. Dumas. 



* " J'ai essaye sans succes de determiner I'equivalent soit en reduisant les fleurs 

 argentines d'antimoine, soit en reduisant le sulfure d'antimoine, soit en acidifiant 

 I'antimoine par I'acide nitrique. Je n'ai obtenu par ces divers proce'des que des 

 resultats tres-discordants." 



