Q2 TABLE 40 {continued) 



PROBABLE VALUES OF THE GENERAL PHYSICAL CONSTANTS 



Backlin's results lead to 4.794 ±0.015. 



The investigation by Backlin is a pioneer piece of work, and it is quite 

 likely, as such, to contain unsuspected systematic errors. If the three values 

 of e (4.768 from Millikan's oil-drop work, 4.776 by Wadlund, and 4.794 by 

 Backlin) are weighted according to the apparent probable error of each, the 

 result is still suspiciously high. The thorough examination made of the actual 

 value of e and its probable error, from the oil-drop work, was carried out 

 because of this inconsistency. It seems best to reject the Backlin value, and 

 to use the weighted mean of the remaining two values, viz. 4.768 ±0.005 an ^ 

 4.776 ±0.008, or 4.770; as usual adopt as its probable error the smaller of the 

 two individual errors, rather than that given by least squares; the latter is 

 meaningless when only two observations are concerned. The finally adopted 

 value is then 



e= (4.770 ±0.005) x 10 10 abs, es units. 



The specific charge of the electron (e/m). — A very complete and critical 

 account of all work on the measurement of e/m, up to 1919, has been given by 

 Bestelmeyer. 1 His final conclusion is that e/m = (1.76 ±0.02) x io 7 em units. 

 A more recent discussion is that by Gerlach, 2 who concludes that e/m= 1.766 

 X io 7 em units. The question is discussed very briefly by Henning and Jaeger, 8 

 who however adopt Gerlach's value. The I.C.T. adopts 1.769 ±0.003. 



The latest work greatly exceeds in accuracy all the preceding; it seems 

 legitimate to confine the discussion to these new results. The value of e/m 

 has been obtained with considerable accuracy by three distinct methods, 

 (a) deflection of electrons in electric and magnetic fields, (b) Zeeman effect, 

 (c) fine structure and relative wave lengths of H and He + spectral lines. It 

 may be obtained also from Bohr's theoretical expression for the Rydberg 

 constant, R^, provided one assumes the value of e and of h. This last method 

 is not as accurate as the preceding. A fifth involves the Compton shift. This 

 also is as yet a relatively inaccurate method. 



The latest and most accurate work with method (a), that by Wolf, 4 is 

 carried out with every possible refinement. The essential point is the employ- 

 ment of a longitudinal magnetic field. The electron velocity is calculated from 

 the potential fall. He concludes that e/m— (1.7679 ±0.0018) X io 7 em units. 

 T.7679 should be corrected for the difference between the int. and abs. units. 

 It then becomes (1.7689 ±0.0018) X io 7 abs. em units. 



1 Marx, Handb. Radiologic, 5, 1, 1919. 2 H.P., 22, 41. 'H.P., 2, 504. *Ann. Phys., 83, 

 849, 1927. 



Smithsonian Tables 



