116 RICHARDS — THE ATOMIC WEIGHT OF NITROGEN. [April 7, 



SOURCES OF ERROR IN THE DETERMINATION OF 

 THE ATOMIC WEIGHT OF NITROGEN. 



(Contribution from the Chemical Laboratory of Harvard College.) 

 BY THEODORE WILLIAM RICHARDS. 



{Read April 7, WO^.) 



The combining weight of nitrogen presents a problem of un- 

 usual interest, because of the uncertainty which still clings to it, in 

 spite of the careful work of some of the most accurate of chemical 

 experimenters. Uncertainty of this kind implies a lack of com- 

 prehension of some unknown variable or variables, and it is always 

 possible that the determination of these variables may lead to the 

 discovery of- some new important fact or principle. Thus the 

 accurate work of Lord Rayleigh in demonstrating that the less 

 active gases of the atmosphere are somewhat heavier than pure 

 nitrogen, led to the discovery of argon and the other inert gases. 



The data for computing the atomic weight of nitrogen are mani- 

 fold, because nitrogen enters into many well-defined compounds. 

 Unfortunately, however, it is always necessary to find the weight of 

 the nitrogen indirectly. The most extended series of experiments 

 wa,s instituted by the great Belgian chemist Stas, who attacked the 

 problem in various ways, converting silver into the nitrate, con- 

 verting this nitrate into chloride, converting the nitrates of potas- 

 sium and sodium into chlorides, and comparing amnionic chloride 

 and bromide with pure silver. The average results of these experi- 

 ments have been variously computed, the extreme estimates of the 

 atomic weight of nitrogen ranging between 14 039 and 14.058 if 

 the atomic oxygen is taken as 16.000.1 



^ The early work of Stas involving argentic chloride must all be rejected, 

 because insufficient precautions were taken concerning its solubility. Among 

 the other pertinent data obtained by him the following, easily traced in Clarke's 

 convenient Recalculation of the Atomic Weights (1897), seem to me the most 

 important. 



(10 = 16.000; Ag =r 107.930; CI = 35.455. H = 1.0076, Br = 79.95s) 



100.000 parts of silver gave 157.478 of its nitrate N = 14.036 



100.000 parts of silver correspond to 49.599 of ammonic chloride. .N = 14.047 

 100.000 parts of silver correspond to 90.830 of ammonic bromide. .N = 14.048 

 Difference between molecular weights of alkaline nitrates and 



chlorides = 26.589 N = 14.043 



Marignac's work on argentic chloride and nitrate leads to a much lower value 



