102 



Froceedings of the Royal Irish Academy. 



Beside's these analyses there were several which yielded a far less 

 quantity of nickel. The same method was then applied to cobalt^ 

 with this difference, that the solution was boiled only for a few 

 minutes, as I found that length of time sufficient for the decomposition 

 of the last traces of hypochlorite. The amounts of iodine liberated 

 were much greater than would be due to the oxide Coo O3, while they 

 agreed perfectly with an oxide C03 O5, thus : — 



Cobalt used, 

 grams. 



Iodine liberated. 

 grams. 



Theory of Iodine 

 for C03O5. 

 grams. 



•186-5 



•5338 



•5343 



•1865 



•5380 



•5343 



• •1865 



•5328 



•5343 



I now repeated the experiments with nickel, taking care to boil 

 the liquid only a minute or two. In one or two instances it was not 

 boiled, but the precipitate filtered off and washed. The results were 

 as follows : — 



Mctel used, 

 grams. 



Iodine liberated, 

 grams. 



Theory of Iodine 

 for NisOs. 



gTams. 



•1570 



•4428 



•4521 



•0785 



•2263 



•2260 



•1835 



•5318 



•5284 



•1570 



•4490 



•4521 



"With a mixture of -1863 gram, cobalt, and ^1835 gram, nickel : — 

 Iodine foimd. 



gi-ams. 

 1-0532 



Theory for Iodine due 

 to Ms O5 & C03 Oa- 

 grams. 



1^0627 



In the last case the oxides were not boiled, but the solution was 

 allowed to stand over the steam bath for a few hours. 



"With solutions of known quantities of nickel, I now made the 

 following experiments. The solution with the suspended oxide was 

 boiled for some hours. 



