100 TABLE 40 (continued) 



PROBABLE VALUES OF THE GENERAL PHYSICAL CONSTANTS 



erg • sec • esr 1 . Substituting, one finds c/m= (1.772 ±0.006) x io 7 abs. em 

 units, the final error being due almost entirely to the error in AA. It seems pos- 

 sibly significant that this value agrees better with the deflection than with the 

 spectroscopic value of e/m, for the theory used in the derivation of the equa- 

 tion is essentially the collision theory of classical dynamics for free electrons. 



e, e/m, and h 1 appear in many important constants, h depends for its value 

 on e, therefore the e appears implicitly, if not explicitly, in every quantum 

 relation. The outstanding discrepancy was between the work of Wagner and 

 of Duane and co-workers, on the value of h from the X-ray continuous 

 spectrum. The recent work of Feder, using this method, gives h in exact 

 agreement with the value adopted, and explains Wagner's low value. Doctor 

 Birge now feels that the value of h/e listed in Table 43 can be assumed with 

 some confidence. The real problem concerns the values of e and of e/m. 



The need of two values of e/m is" very annoying, and fundamentally unsatis- 

 factory. The same situation seems to be arising in regard to e. Millikan's 

 value has been accepted ; it was the only one available. The new work on 

 X rays opened another possibility. The value of Backlin is one-half per cent 

 higher than Doctor Birge's adopted value. As a final result Doctor Bearden 

 obtains for the absolute wave length of the (unresolved) Cu Ka line, 1.5439 

 ±o.ooo2A, and for the Cu K/3 line, 1 .3940 ± 0.0002 A. These results are ob- 

 tained under many varied conditions. The first is 0.345 per cent higher than 

 Siegbahn's value, the second 0.336 per cent. The relative wave lengths are 

 in agreement with Siegbahn, but the absolute wave lengths lead to a value 

 for calcite of rf' 20 = 3-O398A, and £ = 4.825 x io~ 10 abs. es units, 1.15 per cent 

 above Doctor Birge's adopted value of c. It is desirable to consider the various 

 relations that have been suggested between these constants. The most famous 

 connects e, e/m, h, and c in Bohr's formula for the Rydberg constant. This 

 was used to evaluate h, and the value (6.54713) is identical to four digits with 

 that adopted. Hence, the indirectly calculated value of e/m is also practically 

 identical with that adopted. Thus the adopted values of e, e/m, h and c form a 

 self -consistent system, as judged by the Bohr formula for R^. 



Lewis and Adams 2 (theory of ultimate rational units), have obtained, with 

 the aid of Planck's radiation law, the relation: //c/27r<r = 877- (87^/15)3. The 

 right side equals 137.348; the left side, with the constants here adopted, equals 

 137.294 ±0.1 1. The left side equals the reciprocal of the fine structure con- 

 stant a, and the value quoted is taken directly from Table 43. The numerical 

 agreement is very striking. The present agreement shows that this method 

 yields a value of h almost identical with that adopted. 



a is a dimensionless constant involving fundamental general constants ; 

 it should be remembered that to make a dimensionless, we must include with 

 the factor he the unknown dimensions of specific inductive capacity. 



^ote added by Birge April, 1929 (abbreviated). 2 Phys. Rev., 3, 92, 1914. 

 Smithsonian Tables 



