026 PROFESSOR DITTMAR AND MR JOHN M'ARTHUR ON 



Platinum, 194*35 



2RC1, by calculation, 24379 



Loose chlorine, 139-81 



Hydroxyl (0-0566 x OH), 0-96 



578-91 

 M found, . . =581-37 



Water (?), . = 246 



The 2RC1, however, may have included NaCl, which would make the 

 deficit still greater, or it may have contained relatively more chloride of caesium 

 than the original " chloride of rubidium," which would have the opposite effect. 

 Taking the experiment as it stands, it would appear that Pt - 194*35, i.e., a 

 little less than Seubert's value ; but it would be absurd to draw this conclusion 

 from this one isolated experiment. A more plausible hypothesis is that the 

 chloroplatinate contained surplus alkyl- chloride, which of course depresses the 

 value Pt as calculated from the fixed chlorine. Assuming that the chloro- 

 platinate contained its full complement of chlorine, and reducing the results to 

 x CI parts of total chlorine, we have M = 586*90; and for the composition 

 of those M parts — 



Platinum, 196-20 



1-00952 x RjClj, 246-11 



4xCl, 141*82 



584*13 

 M found, . . =586*90 



Deficit, . . = 2*77 



If we reduce to Pt == 195*50, we have M = 584*82; and for the composition 



of M parts, 



Platinum, 195*50 



1*00593 x R 2 C1 2 , 245*23 



3*9668 x CI, . . ' 140-64 



0-0332 xOH= 0-56 



581-93 

 M found, . . =58482 



Deficit, . . = 2*89 



That the chloroplatinate analysed contained 0*00593 x R 2 C1 2 of free alkyl 

 chloride is no improbable assumption. 



None of our rubidium experiments afford the data for calculating even a 

 limit value for the atomic weight of platinum; yet they are interesting, as 

 showing that the tendency of chloroplatinate of potassium to carry down 

 surplus platinum, in the form of chiefly hydroxide Pt(OH) 6 H 2 , when pro- 

 duced by precipitation of chloroplatinic acid with alkyl chloride, is greatly 



