OP ARTS AND SCIENCES. 365 



mony (SbCl.s) was dissolved in a solution of tartaric acid, and the 

 quantity of silver required for the exact precipitation of the chlorine 

 determined by volumetric process. Four determinations, said to have 



cess has the advantages of simplicity in execution, and of giving the atomic weight 

 directly dependent iipon that of oxygen ; but in consequence of the high equiva- 

 lent of antimony the result is very much affected by the unavoidable errors of the 

 operation. This will appear when we consider that the atomic weight is connected 

 with the observed values by the equation 



32 n 



Sb = ^. 



q—p 



in which p denotes the weight of metallic antimony taken, q that of the SbO^ ob- 

 tained, and the atomic weight of hydrogen is taken as unity. 



If we suppose that the whole error of the operation lay in the determination of q, 

 and differentiate the above equation in respect to this value, we have 



c? Sb = — ?! — dq = — dq : 



(q-pf 32p 



and putting for Sb the value deduced from the determinations, viz. 122.34, and for 



p the number belonging to any one of these determinations, for example 3, we find 



that 



d Sh = —15& dq ; 



or, that in this case a small error committed in the determination of q is multiplied 

 about 1 56 times in its effect upon the atomic weight, and that this eflFect increases 

 as the square of the atomic weight to be determined. 



These considerations led to many trials of the possibility of deducing the atomic 

 weight from the quantity of gold which a known quantity of antimony could precipi- 

 tate from an acid solution of the double chloride of gold and potassium (AuCls, KC'l). 

 These attempts proved unsuccessful ; the quantity of gold precipitated seemed to 

 be dependent upon the amount of acid in the gold solution, being greatest when 

 the solution contained the least free acid. They showed that this process, which 

 has been recommended for the quantitative analysis of mixtures of SbOo and SbOg 

 is entirely unreliable. It was noticed that a solution of SbCla in hydrochloric acid 

 which had been digested with metallic antimony until it no longer gave a blue color 

 with starch and iodide of potassium, acquired this property after having been ex- 

 posed for some hours to the action of the air. It would seem that absorption of 

 oxygen takes place with formation of antimonic acid, or the corresponding perchlo- 

 ride of antimony. A similar absorption of oxygen by the salts of arsenious acid 

 has been observed by Fresenius. * 



The nitric acid employed for the oxidation of the antimony was most carefully 

 prepared from distilled sulphuric acid, and the purest nitre of commerce from which 

 every trace of chlorine was removed by means of nitrate of silver. It was still fur- 

 ther purified by a second distillation from the same nitre, conducted at a tempera- 

 ture below the boiling-point of the acid, in order to prevent the possibility of any 

 solid matter being carried over mechanically with the vapor. The acid thus ob- 



* Ann. d. Ch. u. Pharm., Bd. XCIII. S. 384. 



