OF ARTS AND SCIENCES. 39 



0-5479 gr. gave 0-0980 gr. SO, Co = 17-87% cobalt. 

 0-8183 gr. gave 0-1459 gr. SO, Co = 17-83% cobalt. 

 0-4640 gr. gave 0-2170 gr. water = 5-19% hydrogen. 

 1-0060 gr. gave 0-4605 gr. water = 5-09% hydrogen. 

 0.4091 gr. gave 118-5 c.c. nitrogen = 2.S-97% nitrogen. 

 0-3635 gr. gave 109-4 c.c. niti-ogeu = 29-93% nitrogen. 



These analyses correspond tolerably well with the empirical for- 

 mula, — 



Co(NH3)P,(N03),+OH„ 

 which requires, — 



Cobalt, 2 



Hydrogen, 34 



Nitrogen, 14 



Oxygen, 18 



In judging the analyses it must be remembered that the salt cannot be 

 recrystallized, as it is instantly decomposed by water. By standing 

 for a long time over sulphuric acid, the salt loses water and some am- 

 monia. In a salt so dried, — 



0-5088 gr. gave 0-2690 gr. SO.Co = 20-12% cobalt. 



0-3290 gr. gave 0-0859 gr. NH3 =26-11% ammonia (by boiling 



with KHO solution and titrition). 

 0-4308 gr. gave 0-1124 gr. NIL =26-09% ammonia (by boiling 



with KHO solution and titrition). 



The undecomposed anhydrous salt would contain 19*66% cobalt, and 

 would give off by boiling with potash 28-33% ammonia. I consider 

 the true formula of this salt * to be twice as high as that given above, 

 so that it becomes, — 



Co,(NH3),o.O,.(N03),+20H,. 



Fremy found very nearly the same percentages of nitrogen and hydro- 

 gen, but a much higher percentage of cobalt, 20 77%. He also found 

 that the salt is readily decomposed by water with evolution of oxygen. 

 In a single analysis to determine the quantity of oxygen evolved, I 

 obtained the following results : — 



* In a letter to the German Chemical Society, I have represented the niti-ate 

 of oxycobaltiaque as contaming four atoms of ammonia to one of cobalt ; but a 

 careful revision of my analyses leads to the formula given above. Deutsche 

 Chem. Gesell. Berichte, iv. 790. 



