4 PROCEEDINGS OF THE AMERICAN ACADEMY 



Timgstate.. — When chloride of purpureocobalt is boiled with sodic 

 tunc^tate, WO^Na.,, it is quickly converted into a violet granular 

 crystalline mass, which, after washing with cold water, is perfectly free 

 from chlorine. When dried in vacuo over sulphuric acid, the crystals 

 have a fine deep violet color. The salt is but slightly soluble in either 

 cold or boiling water, even in presence of free nitric acid. It dissolves 

 readily in a solution of sodic or ammonic carbonate; the solutions have 

 a iine violet color. Of this salt, 



1-0294 gr. gave 0-o418 gr. WO3 (by mercurous nitrate) = 5G'26% 



WO,. 

 2-3912 gr. gave 1-71G3 gr. WO.Co (by careful ignition) = 75-9G%. 



The formula Co,(NIl3)i„.0.(WOJ2 requires 55-ll%^VO, and 7G-7o% 

 WO.Co. 



The salt cannot be recrystallized, and was therefore probably not 

 absolutely pure. 



Oxalo-chloride. — W^hen a solution of ammonic oxalate is added to 

 one of chloride of purpureocobalt, violet needles are soon deposited, 

 which Genth and I considered as the normal oxalate of this series, and 

 to which we gave the formula, as we should now write it : — 



Co,(NH3),o.O.(CA).2+30II,. 



Two determinations of cobalt and one of oxalic acid agreed very 

 closely with this formula. Krok * subsequently discovered that this salt 

 contains chlorine, and he assigns to it the formula : — 



Co,(NiL),,a(CA).. 



On carefully re-examining this salt, I find that the percentage of 

 chlorine varies considerably in different preparations. Thus : — 



J 0-7108 gr. gave 0-2512 gr. silver = 11-G1% chlorine. 

 (0-3G27 gr. gave 0-218G gr. SO, Co = 22-95% cobalt. 

 ( 0-5093 gr. gave 0-2930 gr. SO, Co = 21-9(1% cobalt. 

 I 0-57G8 gr. gave 0-22-41 gr. silver == 12-78% chlorine. 



It is therefore probable that the chloro-oxalate contains as an admix- 

 ture a greater or less percentage of another oxalate, possibly COjj(NII.j)jp. 

 0.(C^,0,)2. I did nut succeed in obtaining this salt from nitrate of 

 purpureocobalt and ammonic oxalate, the reaction residting only in 



* Acta Univ. Lund. 1840. I believe that this is the only error which has 

 boon detected in the first part of this paper. 



