1887.] 



Dispersion Equivalents. 



409 





Eefraction Equivalent. 



Dispersion Equivalent. 



Potassium. 



Sodium. 



Difference. 



Potassium. 



Sodium. 



Difference. 



Chloride . . . 



18-83 



15-40 



3-43 



1-27 



1-18 



0-09 



Bromide. . . 



9K 'OK 



91 -SA 

 £iL oU 







9 -Oft 





Iodide 



35-78 



32-52 



3 26 



4-42 



4-33 



0-09 



Hydrate . . . 



12 -60 



9-26 



3-34 



0-79 



0-72 



0-07 



Formate . . . 



20-01 



16-60 



3-41 



1-07 



0-96 



0-11 



Acetate. . . . 



27-52 



24-34 



3-18 



1-32 



1-23 



0-09 



Carbonate. . 



28-63 



22-17 



2(3 -23) 



1-40 



1-34 



2(0-03) 



Oxalate. . . . 



37-55 



.. 





2-18 





Nitrite 



18 -99 



15-65 



3 34 



1-30* 



1-17 



0-13 



Cyanide . . . 



17-18 







0-94 







regard to the refraction equivalent, my original determination was 

 8*1; but Kanonnikoff gives only 7*75, which led me, three years ago, 

 to recalculate the observations, taking Briihl's values for oxygen, and 

 to reduce my previous estimate to 7*85. This is determined mainly 

 from the organic salts, and the nitrate, and cyanide. I did not draw 

 any conclusion from the haloid salts, as the chlorine, bromine, and 

 iodine in them appear to have somewhat higher values than what they 

 have in organic compounds. 



How are we to determine the corresponding equivalent of disper- 

 sion ? From the haloid salts it would seem to be about 0'8, but it 

 seems likely that the disturbing influence, whatever it be, which 

 increases the refraction of the haloid salts, would affect the- 

 dispersion. The formate and acetate, KCH0 3 and KC 2 H 3 2 , pro- 

 mise more trustworthy results, as we can subtract from their 

 dispersion equivalents the numbers already determined for carbon, 

 hydrogen, and oxygen. This will give respectively 0*53 and 0*44 for 

 the dispersion equivalent of K. 



If we view potassium hydrate, KHO, as water in which one 

 hydrogen atom is replaced by potassium, water being - 265, we obtain 

 the value of 0'565 for K. 



From the nitrite, KN0 2 , by subtracting 0*82 for N0 2 , we obtain 

 0-48 for K. 



In like manner from the cyanide, KCN, by deducting 0"36 for 

 cyanogen, we get 0*58 for K. 



From the carbonate, K 2 C0 3 , by taking the probable value of the 

 C0 3 at 0*60, we get 0'40 for each K. 



From the oxalate, K 2 C 2 4 , by deducting 1*00 we get 0*59 for 

 each K. 



These figures, varying from 0'40 to 0'59, are too uncertain, and too 



* This is estimated from measurements of A, F, and Gr, and is somewhat open to* 

 doubt, as there seems to be something abnormal in the spectrum. 



