362 PROCEEDINGS OF THE AMERICAN ACADEMY 



the equivalent obtained by Schneider, than as an independent research 

 upon the atomic weight. He gives as trustworthy the result of but 

 one analysis of the chloride, and that not made by himself. From this 

 he derived the number 120.7. 



Of the researches which place the atomic weight at about 122, my 

 own * was the first in point of time, and was commenced before the publi- 

 cation of that of Schneider. After attempting without success to deduce 

 the atomic weight from the quantity of gold which a known weight of 

 antimony could precipitate, with the view of rendering as small as pos- 

 sible the influence of the errors of the operation upon the result, the 

 determination was finally effected by oxidizing the pure metal by 

 nitric acid, and ascertaining the weight of the resulting antimoniate of 

 oxide of antimony. That this body has really the composition Sb04 

 was also proved by a separate experiment. The process, besides its 

 simplicity and accuracy, has the advantage of giving the atomic weight 

 directly dependent upon that of oxygen. In ten determinations the 

 equivalent was found from 122.24 to 122.48. The most probable value 

 is 122.33 ; and the probable error of the individual determinations 

 is 0.05. * 



In the last edition of the Handwbrterhuch der Chemie t this determi- 

 nation is spoken of as being not so accurate {weniger genau), by which 

 is probably meant that it is not so correct, as that of Schneider. As no 

 reason whatever is given for this opinion, I do not think it calls for 

 any other remark than that the writer, 11. Bolley, apparently had not 

 read the paper which he thus criticises. For in the very next article,! 

 by the same writer, the determination of chloride of antimony by means 

 of the double chloride of gold and potassium is recommended ; although 

 this process had been thoroughly tried by me, in the hope of making it 

 available in the determination of the atomic weight, and stated to lead 

 to completely erroneous results. That this process had given Pro- 

 fessor Rose apparently satisfactory results when the equivalent of an- 

 timony was taken at 129, would not genercdly be thought a good reason 

 for supposing that it would give equally good ones when the equivalent 

 was found to be only 120. In fact, Rose has since admitted in a 

 similar case, and referring to my experiments, that the process must 

 be abandoned. § 



* Pogg. Ann., Bd. C S. 563. § Pogg. Ann., Bd. CX. S. 541. 



t Braunschweig, 1858. Art. Antimon., Bd. II. S. 43. 



X Art. Antimon., Bestimmung und Trennung, S. 56. 



