CRITICAL EXPERIMENTS FOR DETERMINATION OF POTASSIUM, ETC. 629 



following table, in which P stands for the weight of platinum used per 

 2NH 4 C1 parts of sal-ammoniac ; A signifies the weight of sal-ammoniac used : 

 C, that of the chloroplatinate obtained ; p, that of the platinum obtained from 

 the chloroplatinate ; Pt' the weight of platinum present per 6 x CI parts of 

 total chlorine. 



P, 



A 

 A 

 Pt', . 



C, 



Experiment, I. II. III. IV. V. 



224 very nearly throughout. 



. -23873 -23970 (-24139) -23948 -23984 



. -54464 -54658 lost -54616 -54743 



. 196-90 196-59 ... 196-73 19648 



. -43833 -43854 ... -43847 -43812 



Mean Values contrasted with those calculated from Seubert's Pt = 194-8. 



Mean. 



•23944 

 •54620 

 196-68 

 •43837 



Calculated from 

 Pt = 194-8. 



•24121 

 •54934 

 1948 

 •43909 



A : C (omitting III.), .... 



A : p, 



Pt', 



p':0, 



As the values Pt', and more still those of p : C, agree fairly ; while those for 

 A : C and A : p vary to an unpleasant extent, we thought that a variable part 

 of the ammonia might have been lost in the evaporations, — the presence of 

 nitroso -compound in the reagent had not been noticed yet, — and therefore 

 tried the following experiment : — The same quantities of standard sal- 

 ammoniac and chloroplatinic acid solutions as has been used in Experiment V. 

 were mixed in a retort connected with a Liebig's condenser, and distilled down 

 in a current of air, while immersed in a steam-bath, and thus kept near 100° C. 

 This somewhat tedious operation was continued until only about 1 c.c. of 

 liquid was left in the retort. In the distillate, after neutralisation with 

 ammonia-free caustic soda, the ammonia was determined by Nessler's 

 colorimetric method. It amounted to only 0*01 mg. This tends to show that 

 under the circumstances chloroplatinate of ammonium is not liable to disso- 

 ciation (into 2NH 3 and PtCl 6 H 2 ) ; but then the ammonia, which otherwise 

 would have escaped as such, may have been destroyed by the nitroso- 

 compound. 



II. Series. 

 The chloroplatinates obtained in the above experiments were obviously not 

 of a constant composition, reducible to the general formula PtCl 6 (NH 4 ) 2 . 

 Thinking that the irregularities were perhaps caused by the process of 

 evaporation, we prepared two quantities of chloroplatinate of ammonium by 

 mere precipitation in the cold. In one case (I.) the chloroplatinic acid was in 

 excess, and the sal-ammoniac poured into it ; in the other (II.), the sal- 



