TABLE 40 (continued) 81 



PROBABLE VALUES OF THE GENERAL PHYSICAL CONSTANTS 



Henning and Jaeger a consider that the variation from unity of either r/p or 

 q is less than the experimental error, and think it more practical to assume 

 r/p = q = i .0000. It seems best to accept the mean value of r/p = 0.99995, 

 determined in four different laboratories, with the probable error as ±0.00005. 

 Assuming then no distinction between int. volt (v), and int. volt (a), we have 



p = 1.00051 ± 0.00002 pq = 1.00046 ± 0.00005 



q = 0.99995 ± 0.00005 pq 2 = 1 .00041 ± 0.00010 



The atomic weights of certain elements. — In evaluating some of our con- 

 stants, it is necessary to use the atomic weights of various elements. In the 

 ultimate analysis, only ratios of atomic weights enter our formulas for the 

 general constants. All atomic weights are determined from ratios, but in 

 general not directly from the particular ratios we need. Hence it is necessary 

 to consider individual atomic weights. 



The present atomic weights are based on the arbitrary assumption that the 

 weight of oxygen is 16 exactly. In choosing oxygen as a basis, it is assumed 

 that it has always the same atomic weight ; i. e., it has no isotopes. Giauque 

 and Johnston 2 have very recently found an isotope of atomic weight 18, from 

 an analysis of the atmospheric absorption bands of oxygen. H. D. Babcock 

 states that experiments performed on absorption coefficients in these bands 

 indicate that Oi 8 has an abundance of only one part in 1250 (probable error 

 some 25 per cent). Aston's atomic weights should be greater than the chemical 

 values by about one part in 10,000. Babcock's determination of relative abun- 

 dance, involves the assumption that the absorption coefficient is the same, per 

 molecule, for each species of molecule (Oi 6 — O a6 and O x6 — Oi 8 ), and this 

 may not be true. The atomic weights determined by Aston, 3 from the mass 

 spectrograph, need not be identical with those determined by chemical means, 

 since Aston's atomic weights are based on the mass 16 isotope of oxygen con- 

 sidered as exactly 16, while the chemical atomic weights are based on the 

 ordinary mixture of the two isotopes considered as exactly 16. We shall see 

 that Aston's atomic weights of hydrogen, helium, nitrogen and iodine seem 

 to agree with the chemical values within his limit of error (one part in ten 

 thousand to one part in five thousand). 



Hydrogen. — Moles 4 lists nine results lying in the narrow range 1.00766 to 

 1.00783, with a mean value of 1.00777 ±0.00002, or a rounded figure of 1.0078. 

 The final average represents the result of 223 different measurements by five 

 different investigators, using four different methods, and seems to be the most 

 reliable now available. Doctor Birge accordingly adopts 



1 H.P., 2, 501. 2 Journ. Amer. Chem. Soc, 51, 1436, 1929. 3 Proc. Roy. Soc, 115 A, 487, 

 1927. "Berichte, 61 B, 1, 1928; 59, SII, (A) 1926. 



Smithsonian Tables 



