of Edinburgh, Session 1885 - 86 . 
649 
convert the quantity of ammonio-ferrous solution used into its equiva- 
lent in c.c. of permanganate, and in the ordinary way calculate the 
percentage of nitrite the sample contains. 
To test the method, nitrite of silver was precipitated from a 
moderately strong solution of nitrite of potassium by nitrate of 
silver, and two grammes of the recrystallised nitrite dissolved in 
lukewarm water were made up to one litre. As this nitrite is not 
very stable, the amount of silver was estimated by chloride of 
sodium, and calculating from this basis it was computed that every 
50 c.c. of the solution contained ’02465 gram of anhydrous nitrous 
acid. The following six consecutive estimations show the quantities 
actually found : — 
1. -02464 
2. -02465 
3. -02463 
4. -02464 
5. *02465 
6. -02466. 
The more commonly employed inorganic nitrites being that of those 
alkalies, I may here note that I find commercial specimens of 
nitrite of sodium, either in rods or crystals, usually contain nitrous 
acid equal to 94-95 per cent, of nitrite ; specimens of nitrite of 
potassium an amount equal to from 86-88 per cent, of the salt ; 
whilst nitrite of ammonium, being very deliquescent, is consequently 
sold in solution which generally contains, when not old, about 12 
per cent, of actual nitrite. 
In this paper, therefore, I have endeavoured ta show — 
1 . That the permanganate process as ordinarily employed for the 
estimation of inorganic nitrites, viz., by gradually adding perman- 
ganate to the acidified solution of the nitrite, is unsatisfactory. 
2. That it is desirable to use a large excess of permanganate, and 
titrate back with a reducing agent. 
3. That very accurate results are obtained by using comparatively 
weak solutions of permanganate and of the reducing agent. 
4. That the most suitable reducing agent is ammonio-ferrous sul- 
phate, and that both it and the permanganate must not be in strong 
solution. 
