Dr. E. J. Mills on certain Cabaltamines. 251 



taken in connexion with the fact that a lower cobaltic oxide than 

 the sesquioxide had been formed, naturally led me to look for 

 hydric nitrate in the filtrates. The sodic solution just mentioned 

 was evaporated to dryness, and hydric nitrate determined in it 

 by Stein's method*. The percentage number obtained (0*06) 

 is of sufficient value to show that the quantity of nitrate is ex- 

 tremely small, and adequate to account for only a minute frac- 

 tion of the missing substance. 



The presence of hydric nitrate in these liquids may also be 

 shown qualitatively by boiling them with pure sodic hydrate 

 until all free ammonia has been expelled, employing Har- 

 court's process, and applying Nessler's test to the distillate. 



II. Cobaltic fi-Pentammoniotrichloride. 



The " purpureochloride " of Gibbs and Genth, which is still 

 termed the "roseochloride" by French writers, may be conve- 

 niently designated by the above name, in accordance with its 

 dynamical history. It can be prepared, as is well known, by 

 adding potassic permanganate in slight excess to an ammoniacal 

 solution of cobaltic chloride, and then boiling with abundance of 

 strong aqueous hydric chloride. This process can be carried 

 out on small quantities in a few minutes, and is already recog- 

 nized as an excellent qualitative test for cobalt. It is, however, 

 quite unsuited for quantitative workf, as may be seen when the 

 method is employed on a large scale ; the filtrate contains co- 

 baltic chloride and cobaltic hexammoniotrichloride in appreciable 

 proportions and to a varying amount. Indeed no plan has yet 

 been proposed whereby cobaltic chloride can be at once and 

 entirely transformed into either of its poly ammoniacal deriva- 

 tives. Calcic oxychloride is an economical and convenient sub- 

 stitute for potassic permanganate in the preparation of the j3 

 compound. 5 parts of crystallized cobaltic chloride and 5 parts 

 of sal-ammoniac are dissolved in 100 of water ; the solution is 

 cooled, mixed with 22 parts of strong aqueous ammonia and 2 

 parts of "chloride of lime," and allowed to stand freely exposed 

 to air for twenty hours. At the end of that time the liquid is 

 boiled with a large excess of strong aqueous hydric chloride and 

 allowed to crystallize. The precipitate is the compound required. 

 The very acid mother-liquid is poured off and employed as 

 a precipitant in the next operation ; after repeated use, it will 

 contain a small quantity of the hexammoniotrichloride and a 

 more considerable proportion of cobaltic chloride. 



* Jahresbericht fur Chemie, 1851, p. 627. 



t The contrary is maintained by Terreil, Comptes Rendus, vol. Jxii. 

 p. 139. Braun (Zeit. analyt. Chem. vol. v. p. 114) supports the view here 

 adopted, but on other grounds. 



