1890.] on dilute solutions of Nitrates and Nitrites. 53 



formed in order if possible to get some kind of regularity in the 

 numbers found, but without success. 



The method employed, is that one described in text-books 

 on quantitative analysis. It consists in placing a quantity of 

 well washed CuZn couple in a dilute solution of the salt (KN0 3 ) 

 and boiling. The steam is conducted through a condenser into 

 a measured quantity of standard acid, so arranged that all the 

 ammonia is absorbed. 



It is usually stated that the reaction is at an end in one 

 hour, and that afterwards no further evolution of ammonia takes 

 place. As this did not appear to be the case in the first set 

 of observations, the boiling was continued in the succeeding ones 

 until no more ammonia was given off. In doing this it was 

 found necessary to evaporate to dryness, then add more water 

 and evaporate again. In some cases this process was repeated 

 six times before every trace of the alkaline gas was removed. 



Sometimes the reaction stopped long before the whole of the 

 water passed over, and yet on the addition of more water a 

 further evolution of gas took place. This was caused by the 

 oxide of zinc, produced during the progress of the reduction, 

 forming a covering to the copper zinc couple and thus preventing 

 the access of the liquid. When more water was added, some of 

 this oxide was washed off and the surface thus exposed con- 

 tinued the reduction of the nitrate. 



The experiments gave numbers varying from 29 per cent, 

 of the calculated quantity up to 67 per cent, when the ammonia 

 had been produced slowly, while it rose from 70 to 80 per cent, 

 when the boiling was urged. 



Careful search was made for insoluble or soluble nitrogen 

 compounds, which might resist the action of the hydrogen, such 

 as basic nitrates, but no such bodies could be detected. 



Nitrites were however found to be produced in all cases. I 

 give the numbers for these experiments as an example. 



On the 1st boiling A gave 55 per cent., B gave 50, G gave 55 



2nd „ 3 „ „ 20 „ 8 



3rd „ 10 „ „ „ 9 



4th „ 7 „ „ - „ 6 



5th „ „ ZL » _°_ 



75 • 70 78 



Thus the highest number was a little more than 20 per cent, 

 short of the calculated quantity. As the highest numbers were 

 obtained from those solutions which had been made to boil most 

 strongly, experiments were designed to try if the variation of 

 the temperature caused any alteration in the chemical reaction. 



5-2 



