422 XXXIV.-A New Jlethod of Estimating Nitrites either crlone 01' in presence of Nitrates and Chlorides. By T. CUTHBEKT DAY. JN most text-books on chemistry a method is given for the prepara-tion of nitrogen by heating a mixture of potassium nitrite and ammonium chloride the change being represented by the equation-K02K + NH4C1 = KC1 + 20Hz + NZ, which niay be simplified by saying that ammonium nitrite when heated splits u p into nitrogen and water thus :-NO,*NH = N2 + OH, the quantity of nitrogen produced being exactly double of that present as nitrous acid in the nitrite. The method here given consists in applying this well-known reac-tion quantitatively. I have never seen any mention of the use of this reaction for quantitative purposes and I scarcely think that a method, based on its use would have been omitted in any work on analytical chemistry if accurate results had been obtained by its means.I n order t o test the accuracy of the method I have devised I pre-pared a quantity of pure silrer nitrite NO,Ag and determined the amonnt of nitrogen it contained comparing the results obtained with that required by the known composition of the salt. The process is conducted in the following way :-The FIG. 1. solution containing the nitrite is introduced into the flask A (Fig. 1) (about 75 C.C. capacity) and an excess of solid ammonium chloride is then added till the solution is thoroughly saturated. It is important that the solution should be concentrated otherwise considerable di€Ecultr will be found in driving off the last traces of nitrogen.The flask is closed by a cork carrying the capillary tube DAY A NEW METHOD OF ESTIMATING NITRITES. 423 and the delivery tuk)e C of somewhat wider bore 1.5 t'o 2.0 mm. Carbonic anhydride generated from white marble and hydrochloric acid in the apparatus shown in the figure is then allowed to flow through the flask until all the air is removed. The small U-tube n, is filled with glass beads moistened with a little weak potash solution, and is intended to intercept any acid spray which may be carried over by the current of gas. As soon as it is considered that the whole of the air in the flask is displaced by carbonic anhydride the screw-tap E is closed and the flask disconnected from the carbonic anhydride apparatus. The end of the delivery tube C is now brought under the graduated absorp-tion-tube F which contains mercury and a few C.C.of strong potash solution to absorb the carbonic anhydride. The solution in the flask is now heated till it boils. As long as any nitrogen is given off the boiling liquid effervesces freely but as soon as the last traces have been expelled the boiling assumes the peculiar bumping character noticed in water free from dissolved gas. The boiling is continued till all the nitrogen produced in the reaction has been carried over by the steam and collected in the absorption-tube. The screw-cock E may now be opened while the liquid is still boiling and the flask removed. The absorption-tube containing the collected gas is removed to aJ vessel full of water the mercury is allowed t o fall out.and the volume of nitrogen obtained is measured after standing a few hours. The observed volume of nitrogen when due allowance has been made for temperature pressure and tension of aqueous vapour is halved its weight calculat'ed and we have then all the data necessary for calcu-lating the amount present in the nitrite under analysis. The following experiments will show the degree of accuracy attain-able by this method :-The weighed quantity of silver nitrite was in each case dissolved in hot water and decomposed by a solution of sodium chloride. The solution was filtered through a small filter into the flask A and the precipitated silver chloride washed repeatedly with small quantities of hot water adding the washings to the solution in the flask :-Gram N. P. c. N. N02Ag. Gram N02Ag. Half vol. P. c. N. Theory. P. c. Expt. 1. 0.1364 gave 0.012136 = 8.90 9.02 Loss 0.12 , 2. 0.1700 , 0.0155'79 = 9.12 9 9 Gain 0.10 , 3. 0.1712 , 0.015452 = 9.02 9 9 Exmt. I repeated the determinations in presence of excess of potassium About 0.5 gram of this salt was added to the solution i n nitrate. each experiment : 424 RUREMANN AND CARNEGIE THE ACTION OF ACETONE Gram N. P. c. N. N0,Ag. Gram N0,Ag. Half vol. P. c . N. Theory. P. c. Expt. 1. 0.1716 gave 0.015455 = 9.12 9-02 Gain 0.10 , 2. 0.1808 , 0.016556 = 9.16 7 7 Gain 0.14 The results of these two experiments show that the presence of the nitrate does not affect the accuracy of the determinations