10 PROCEEDINGS OF THE AMERICAN ACADEMY 



small excess of nitric acid. The chloro-flnosilicate appears to combine 

 with auric and platinic chlorides to form crystalline salts containing 

 silicic fluoride. Nitrate of purpiireocobalt gives with a solution of 

 fluosilicic acid, SiF,jII., a beautiful violet-red granular crystalline pre- 

 cipitate, but slightly soluble in cold water. I have not examined this 

 salt; it is probably the normal fluosilicate Co2(NH3),Q(SiF,5)3. 



It occurred to me that the fluorine compounds of the cobaltamines 

 might offer means of separating certain metallic elements which have 

 hitherto proved intractable by ordinary methods. The numerous ex- 

 periments made with this end in view have not, however, led to really 

 valuable results. I shall therefore content myself by briefly describ- 

 ing in this place a few reactions which will serve as starting-points to 

 those who may be disposed to enter upon "this field of investigation. 



Nitrate of croceocobalt gives a fine granular crystalline salt with 

 fluosilicic acid, and a beautiful salt in large granular crystals with 

 potassic fluo-titanate, TiF^.lC. Both salts are soluble in much hot 

 water, the solutions yielding large and perhaps measurable crystals. 



^Nitrate of xanthocobalt gives very fine granular crystalline salts 

 with fluosilicic acid and jJOtassic fluotitanate. Both salts may be 

 dissolved in a large quantity of hot water, and recrystallized without 

 decomposition. 



Sulphate of roseocobalt gave no precipitate with fluosilicic acid or 

 pot;issic fluosilicate even after long standing; but a solution of the 

 iodo-sulpbate of roseocobalt, Co^(NIL)j||(SO^)2T2» gave with fluosilicic 

 acid a dull violet-red crystalline precipitate very slightly soluble in 

 water. 



Nitrate of luteocobalt gives beautiful granular orange-yellow crys- 

 talline precipitates with fluosilicic acid and solution of potassic fluo- 

 titanate. The fluotitanate is somewhat lighter in color than the 

 fluosilicate. Both salts are almost insoluble even in boiling water. 

 Krok * observed that iodo-sulphate of luteocobalt gives a very in- 

 soluble crystalline precipitate with fluosilicic acid. I find that this 

 salt contains iodine, and it will probably prove to have the formula 

 Co2(NII.j),2(SiF,,).J2. Iodide of luteocobalt gives a beautiful orange 

 crystalline preci|)itate with potassic fluozirconate, which requires a 

 large (]uantity of boiling water for solution, and may be recrystallized 

 without d«'com[)()sition. Iodide of luteocobalt gives also crystalline 

 very slightly soluble precipitates with potassic fluotantalate, fluo- 

 uiobate, and oxyfluoniobate, not differing from each other, so far as I 



* Acta Universit. Lund. 1870. 



