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



analyzed, and thus may have led to an underestimate of Its atomic 

 weight. Nevertheless, the case is one of those exceptional ones 

 which needs further investigation before it can be cast wholly aside. 



At that time, by request of Sir William Ramsay, I made some 

 preliminary experiments upon entirely a new method, namely, the 

 conversion of sodic carbonate into nitrate. These led to a value at 

 least 14.02 for the quantity in question, thus concurring rather 

 with the usually accepted value than with the widely deviant results 

 of the English experimenters. This work has not been completed, 

 and therefore need not receive further discussion. 



Still more recently, in connection with Dr. E. H. Archibald, 

 still another method was tried with success.^ The nitrates of 

 potassium and caesium were decomposed by finely divided pure 

 silica, the nitric acid being completely expelled. If the atomic 

 weights of these two metals are assumed to be respectively 39-139 

 and 132.879 (values calculated from other accurate data), that of 

 nitrogen is found to be in the two cases respectively 14.037 and 

 14.040, in good confirmation of the work of Stas upon the nitrates. 

 Viewed as a means of determining the atomic weight of nitrogen, 

 these analyses must nevertheless be regarded as preliminary, since 

 hardly large enough quantities of material were taken to attain the 

 best results, although indeed the average amount of nitrogen 

 weighed was ten times as great as that weighed by some of the 

 previously mentioned experimenters. The real purpose in this case 

 was to determine the atomic weight of caesium by a wholly new 

 method, assuming nitrogen to be known. 



On considering all these data and their possible errors discussed 

 above it seems probable that the atomic weight of nitrogen is not 

 less than 14.02 and not over 14.04, probably being nearer to the 

 latter value than to the former. The occasional wide deviations are 

 not certain enough to demand the assumption of inconstancy in the 

 atomic weights, or the necessity of disbelieving in the law of the 

 conservation of weight ; but the irregularities which exist are 

 enough to point to the desirability of further investigation. Such 

 investigation could hardly fail to yield an interesting outcome, 

 since any uncomprehended relation in nature must be due to some 

 fact or facts not hitherto recognized. 



^ Richards and Archibald, Proc. Am. Acad.^ 38, 458 (1903). 



