628 PROFESSOR DITTMAR AND MR JOHN M'ARTHUR ON 



two aliquot parts, both gravimetrically (with a known weight of silver used as 

 nitrate) and titrimetrically, i.e., by determining the silver left unprecipitated, 

 by means of our form of Volhard's method. 



Analysis, I. II. 



(1) Sal-ammoniac used, -268735 -537277 grm. 



(2) Chlorine calculated from the weight of the AgCl obtained, -178059 -356143 „ 



(3) From the weight of the silver precipitated, . . . -178028 -355885 „ 



(4) Mean of (2) and (3), 



(5) Substance analysed per 35 - 456 parts of chlorine, 



Mean, . 

 Stas' value is, 



. -178044 -356014 

 53-515 53-507 



53-511 



53506 



This sal-ammonic was used in the following experiments. 



I. Series. 

 ■ The general mode of operating was as follows : — 



A known weight of the salt (about "54 grm.) weighed out as a standard 

 solution, was mixed with a slight excess of chloroplatinic acid, the mixture 

 evaporated to almost dryness over a water-bath, the residue treated with a 

 mixture of equal volumes of absolute alcohol and absolute ether (both specially 

 distilled with a little syrupy phosphoric acid, to eliminate any ammonia), and 

 washed with such ether-alcohol until free of soluble chlorine. Only a small 

 portion of the precipitate was allowed to get on the filter. To recover this 

 small portion, it was allowed to dry in the air, then dried further at 100°, and 

 what could not be removed mechanically, dissolved off with hot water ; the 

 solution was evaporated to dryness over a water-bath, the bulk of the pre- 

 cipitate added, and the whole kept in a drying-chamber at 100° to 110° until 

 constant in weight. The chloroplatinate was then transferred to an Erlen- 

 meyer flask, with some 300 c.c. of water, the platinum reduced out with 

 hydrogen, filtered off, and weighed. The filtrate was diluted to a known 

 weight, and a small portion utilised for a preliminary determination of the 

 chlorine. A larger portion then served for the exact determination of the 

 chlorine, by means of that combination of the gravimetric and Volhard's 

 method, which had served for the analysis of the sal-ammoniac. 



The ether-alcohol washings were collected and utilised for the determina- 

 tion of the small amount of ammonia which had escaped precipitation. These 

 determinations, however, proved valueless, because we found out (only when 

 all the work had been done) that the chloroplatinic acid used was contaminated 

 appreciably with the nitroso-compound, PtCl G (NO) 2 . 



We therefore here satisfy ourselves with stating that the weights of unpre- 

 cipitated sal-ammoniac appeared to amount to from 06 to 0*9 mg. per 054 grm. 

 of sal-ammoniac operated upon. 



Five experiments were made in this manner. The results are given in the 



