G30 



PROFESSOR DITTMAR AND MR JOHN M'ARTHUR ON 



ammoniac solution was in excess, and the chloroplatinic acid poured into it. 

 In either case, 3*00 grins, of platinum (as PtCl G H 2 ) were operated upon, and the 

 two solutions, before being mixed, attenuated by addition of quantities of water 

 so adjusted that the total volume on both sides was 150 c.c. In both cases the 

 mixture was allowed to stand over night. Next morning the clear liquid was 

 decanted through a small filter; the precipitate was then washed by decanting 

 filtration, first with small instalments of water until the soluble chlorine was 

 reduced to a fraction of milligramme (by calculation), then with absolute 

 alcohol until all the soluble chlorine was proved to be away (by testing with 

 nitrate of silver). The precipitates were dried at 105° to 110° until constant in 

 weight. Each of the two chloroplatinates was divided into two approximately 

 equal parts, and each part analysed by itself; the platinum was reduced out, 

 filtered off, and weighed ; the filtrate divided into two aliquot parts, and in 

 each the chlorine determined gravimetrically. Yolhard's method was not 

 used in this case, as our stock of standard silver had become exhausted. All 

 duplicate determinations made agreed very well with each other. In the 

 following table we give the values, C : p ; Pt' = weight of platinum per 

 6 x CI parts of total chlorine ; Pt" = weight of platinum per C1 6 (NH 4 ) 2 

 = 248-84 parts of non-platinum in the precipitate, also the approximate 

 weight P of platinum employed per 2NH 4 C1 parts of sal-ammoniac. 



Experiment, I. II. c , ., 



1 ' Seubert s 



Part, A. B. A. B. Values. 



p : C, . . . -43838 43837 "43836 -43834 -43909 



Pt', .... 197-04 196-97 197-13 197-14 \ ig . g 



Pt",. . . . 194-19 194-23 194-22 19420 J 



P, . . . . 200 192 



The most remarkable feature in these results is that the values p : C, and 

 P', were the same in Experiment I. as in Experiment II. Clearly neither 

 the mean of the values Pt' nor that of the values Pt" can pretend to be a close 

 approximation to the true Pt. A reasonable hypothesis to account for the 

 difference between Pt' and Pt" is to assume that the chloroplatinate was a 

 mixture of the composition PtCl 6 (NH 4 ) 2 + <#Pt(OH) 6 H 2 -f ?/H 2 0. By cal- 

 culating the four analyses as one, we found per 6 x CI parts of chlorine — 



Chloroplatinate, . . . . . = M = 449 - 53 

 Platinum, = Pt' = 19706 



Assuming 6 x CI to be associated with 2NH 4 C1 = 107 01, we have for the 

 composition of the precipitate — 



Platinum, = 197-07 



2NH 4 C1, = 107-01 



4xCl, = 141-82 



Unaccounted for, = 3 - 63 



M, . . . = 449-53 



