10 PROCEEDINGS OF THE AMERICAN ACADEMY 



scribed his method of procedure, that others might correct his 

 results if they were found to be based upon an incorrect assump- 

 tion. He added an excess of silver to the chloride to be investi- 

 gated, and then added the standard solution of a chloride until no 

 more cloudiness was observable. Such a method under ordinar}'- 

 circumstances requires from two to eight milligrammes less of sil- 

 ver to correspond with a given weight of chloride than would be 

 required if the solutions were added in the inverse order. 



A number of years afterward * Stas changed his method of pro- 

 cedure, and selected the point half-way between the two extremes 

 as the true end point of the silver reaction. He gave reasons for 

 this change of view, but wholly ignored his previous results. Com- 

 mentators have laid hardly enough stress upon this important 

 difference between the two series of determinations, although it 

 necessarily involves an error in one series or the other. 



Working before even the earliest date of Stas's publication upon 

 this subject, the experimenters upon the atomic weight of barium 

 naturally overlooked the whole question. As nearly as may be 

 guessed from their incomplete accounts, they usually selected the 

 end point obtained by gradually adding argentic nitrate to baric 

 chloride; hence their results cannot be compared with either of 

 Stas's series. 



Much time during the past eighteen months has been spent upon 

 this question. The investigation of baric chloride showed that re- 

 sults for the atomic weight of barium varying from 137.35 to 137.50 

 might easily be obtained from the purest possible salt, according to 

 the interpretation of the data. At last a definite conclusion was 

 reached, and the work is now nearly ready for publication. 



The necessity for some other basis for the atomic weight of 

 barium early led to the search for a new starting point. In the 

 course of this search, most of the available baric salts were inves- 

 tigated with regard to their adaptability for the present purpose. 



Baric nitrate holds water with great obstinacy, and no certain 

 point of constant weight could be reached by gradually heating it. 

 Besides, the only two methods available for its analysis are ex- 

 tremely unsatisfactory. The conversion into the chloride is ren- 

 dered very difficult because of the insolubility of both the nitrate 

 and chloride in strong acids. The complete conversion of the 



* The Memoir was presented in 1876, according to the title page. Mem. de 

 I'Acad. de Belg., Nouv. Ser., XLIII. See also Van der Plaats, Chem. News, 

 LIV. 52, 88. 



