ATOMIC WEIGHTS 51 



0.9867, whence NHgrr 17.0323. Jaquerod and Scheiier/ by a different 

 formula, and using only the density determinations of Guye and Davila, 

 find NH3 = 17.0148. If H= 1.0078, N = 14.0089, Perman and Davies' 

 method, or 13.9914 by Jaquerod and Scheuer. The first value is ap- 

 parently the best and will be adopted here. 



There are now four independent values for N, as follows: 



From No N = 14.0073, ± .00020 



From N„0 " = 14.0037, ± .00095 



From NO " = 14.0083, it .00055 



From NH3 " = 14.0089, ± .00065 



General mean N = 14.0074, ± .00018 



From compressibility data Eayleigh '^ found from Ng? N = 14.008, and 

 from ^2^} N = 13.998. His low pressure value, as previously cited, was 

 N"= 14.009, and Jaquerod and Perrot, at high temperatures, found N = 

 14.0077. To include these values in the general mean would change 

 the final result inappreciably, if at all, and they may therefore be dis- 

 regarded. They have, however, confirmatory significance. 



Some of the determinations utilized in the foregoing combination are 

 evidently overvalued, especially the figure derived from atmospheric 

 nitrogen. The " probable errors," scrutinized in detail, merely show that 

 the density measurements are much more concordant than the gravi- 

 metric analyses. Moreover, the errors of the critical constants have not 

 been taken into account, for they can hardly be estimated correctly. Al- 

 lowances for these uncertainties might be made, but their effect upon the 

 final combination would be trifling. The " probable error " here assigned 

 to N, simply indicates the weight which it should receive in calculating 

 other atomic ratios.' 



As a check upon the other determinations of the atomic weight of 

 nitrogen, Guye and Pintza * have determined the composition of ammonia 

 by volume. The gas was decomposed by a spiral of platinum wire heated 

 to redness, and from the density of the mixed gases, Nj + SH^, compared 

 with the known densities of nitrogen and hydrogen, the required datum 

 was calculated. For the weight, in grammes, of a normal litre of the 

 gaseous mixture, the following figures were obtained : 



' Compt. Rend., 140, 1384. 1905. From NO Jaquerod and Scheuer found N = 14.005. 



= Proc. Roy. Soc, 74, 446. 1904. 



' For a general discussion of the atomic weight of N, see Guye's lecture delivered before the 

 Cliemieal Society of Paris, June 10, 1905. Also Compt. Rend., 140, 1386, and 144, 1360; and Ber., 

 39, 1470. Two papers by Berthelot, of a controversial nature, are in Compt. Rend., 144, 76 and 

 269, and one by Leduc in Compt. Rend., 146, 399. 



* Compt. Rend., 147, 925. 1908. 



