1 34 Roberts — Reduction of Nitric Acid by Ferrous Salts. 



24*8 to 22'6 cubic centimeters. In experiment (3), after the 

 lapse of an hour with similar treatment, the volume of gas 

 had decreased four cubic centimeters, and one cubic centimeter 

 was lost by the gas simply standing over night over caustic 

 soda. 



To determine whether the nitric oxide went into solution as 

 such or in the form of nitrite, the following experiments were 

 performed. About 30 cubic centimeters of the liquid were 

 drawn off by stopcock at bottom of Hempel's burette into an 

 acidified solution of potassium iodide, freely exposed to air, 

 and the amount of iodine set free determined by titration with 

 sodium thiosulphate. Again, 30 cubic centimeters more of the 

 same solution were treated in exactly the same way except 

 that the flask containing the potassium iodide was first filled 

 up with carbon dioxide and air carefully excluded. On two 

 trials it was found that the amount of iodine set free under 

 the latter circumstances was very small, and less than one-half 

 as much as that set free when air was admitted freely, showing 

 very plainly that some, and perhaps most, of the nitric oxide 

 was dissolved as such, without chemical action of the caustic 

 soda. 



The principal conclusions to which this study of the estima- 

 tion of nitrates, by measurement of the volume of nitric oxide 

 swept along by carbon dioxide and collected over caustic soda, 

 has led me, are as follows : 



1. The best analytical results are obtained according to this 

 method by passing the gas through potassium iodide before 

 collection, and estimating from the total volume of gas col- 

 lected.* 



2. If air be present in the apparatus at the outset, it will 

 not affect the total volume of gas if potassium iodide be not 

 used, or if the nitric oxide meet the air beyond the Will and 

 Varrentrapp bulb. When potassium iodide is used, air in the 

 apparatus would make the total volume too large, but with 

 small amounts of air, this error is counterbalanced by the solu- 

 bility of nitric oxide in caustic soda. 



3. Since the nitric oxide is perceptibly soluble in caustic 

 soda upon shaking, the gas should not be left long in contact 

 with the liquid before the volume is measured. 



4. If the hydrochloric acid used is very dilute, the reaction 

 proceeds very slowly, but leads finally to the same results. 



5. If the nitrate and ferrous salts are brought together hot, 

 some higher oxides of nitrogen may be formed, but the error 

 is corrected by the use of potassium iodide. 



6. If potassium iodide is used the results are accurate even 

 when the ferrous chloride is present only in very slight excess. 



*This result accords with Glaser's experience in collecting the gases evolved 

 unmixed with C0 2 , over KI — [loc. cit.J 



