82 Table 40 (continued) 



PROBABLE VALUES OF THE GENERAL PHYSICAL CONSTANTS 

 H = 1.00777 ±0.00002. 



Aston, from positive ray analysis, obtains H= 1.00778 with a limit of error 

 0.00015. The recent discovery of an isotope of oxygen makes it permissible 

 to use Aston's value only as an indication of the relative abundance of O i8 and 

 Oi6, not as an atomic weight determination. It is in perfect agreement with the 

 chemical value, which indicates a very low abundance of O i8 . 



Helium. — The true atomic weight of helium must be close to Aston's value. 

 The chemical value is at present slightly less accurate than Aston's, and Doctor 

 Birge accordingly adopts his value but his assumed error as the probable error, 

 although he considers such a procedure may be open to criticism, in view of the 

 situation regarding the oxygen isotopes. He accordingly writes 



He = 4.0022 ±0.0004. 



Nitrogen. — The error in the atomic weight of nitrogen produces practically 

 the entire error in the atomic weight of silver. Since the great majority of 

 the accepted atomic weights are derived more directly from silver than from 

 oxygen, that of silver is of the highest importance. 



The atomic weight of nitrogen can be obtained by direct comparison with 

 oxygen, and also from density measurements, using the adopted value of R. 

 According to Clarke, 1 the final average of these two methods gives N = 14.0076. 

 The atomic weight can be obtained indirectly in many ways. The results of all 

 methods, including the two just mentioned, are summarized by Clarke 1 and 

 give N= 14.0081, presumably the best value in 1920. Now it is generally 

 agreed that, as in the case of helium, the atomic weight of nitrogen can be 

 determined most accurately from its density and deviation from a perfect gas, 

 by the use of R = M(i—a)/L where R is 22.41 46 ±0.0008 (see p. yy), 

 (1— a) = 1.00043 ±0.00002/ and L= 1.2 5046 ±0.000045 3 whence 



N = 14.0083 ±0.008. 



Aston 4 obtains N= 14.008, but his assumed accuracy is only one part in 

 5000. Aston gives always the limit of error, and his probable error should be 

 much smaller. His values all agree beautifully with the chemical values; the 

 decision as to his actual probable error may be left open. 



Silver. — The best atomic weight of silver is at present directly dependent 

 on that of nitrogen. A summary is given by Moles and Clavera. 2 Of the many 

 methods for obtaining the value of Ag, the most accurate is based on the reduc- 



1 Mem. Nat. Acad. Sci., 16, 1920. 2 Z. anorg. Chem., 167, 49, 1927. 3 Ibid., 167, 40, 

 1927. * Proc. Roy. Soc, 115 A, 487, 1927. 



Smithsonian Tables 



