October 11, 1912] 



SCIENCE 



485 



electrical charge to the mass of the simple 

 electron. This has been interpreted to mean 

 that the rotating (vibrating) electron does 

 not produce an appreciable magnetic field but 

 that the complexity arises from " linkage " of 

 the electrons in different ways. I will illus- 

 trate this simplicity of step arrangement by 

 citing from Mr. King's work three iron lines, 

 viz., 3722.7 a, 3872.6 a and 5447.1 A. The 

 spacings are all " normal " and the steps are 

 (1, 2, 3, 4) for one kind of vibration and 1, 2, 

 for the other kind of vibration so that King 

 writes the line ± (1, 1, 2, 2, 3, 4) a. From his 

 tabulated measurements I have computed the 

 value a, and find it for the three lines to be 

 respectively .730 ± .008, .752 ± .006 and 

 .758 ± .006, whereas his field strength would 

 give .753. One notices but a slight deviation 

 in the first, which arose from a probable error 

 in one measurement. These little details in 

 computation may add much to the conclusive- 

 ness of a statement. Neither such simplicity 

 nor so conclusive a relationship is present in 

 the twelve component thorium line 4086.7 A.° 

 Por most substances the majority of lines 

 are triplets, which corresponds to the simplest 

 form of the theory developed by and pre- 

 dicted by Lorentz. However, these triplets 

 show great variety in magnitude and appear- 

 ance." Mr. King thinks there is some tendency 

 for these lines to group about magnitudes re- 

 lated to the "normal." For example, he finds 

 that thirty-five sharp lines (Table IV.) can be 

 grouped under the magnitude 3a(=2.26). I 

 find twenty of these lines to lie between 2.21 

 and 2.32 with an average of 2.275. This is a 

 diilerence of five per cent, in the extremes. 

 For a considerable distance upon either side 

 of this space (2.21 to 2.32) there are but few 

 lines of corresponding sharpness. This fact 

 seems favorable to a single group in this list. 

 But the real necessity here is the same as I 

 pointed out in case of thorium, viz., a little 

 greater accuracy (or greater resolving power 

 of the spectroscope). For example, reduce 

 this extreme error in variation from five to 



^ Astrophysical Journal, XXX., p. 151, 1909. 

 ^ See King 'a references to Purvis, Moore, Jack, 

 Cotton and Babeoek. 



one per cent., possibly even two per cent., then 

 we shall know whether there are steps in the 

 magnitude of the triplet separations or 

 whether there are a great variety of separa- 

 tions differing by small increments of, value. 

 The latter result implies that the electrons 

 are vibrating in a self induced, as well as the 

 superimposed magnetic field. 



I fail to see the validity of Mr. King's law 

 (p. 54) : " Since -DA/A^ {i. e., change in fre- 

 quency of vibration per sec.) is shown to he 

 nearly a constant, . . . the mean separation 

 of the n-Components (i. e., components which 

 vibrate perpendicular to the lines of force) 

 varies as the square of the wave length." The 

 notes in parentheses are added by way of ex- 

 planation. One thing made certain by ex- 

 periment and theory is that different types of 

 separation correspond to different physical 

 aggregates, unions, or " linkages " of elec- 

 trons ; and Preston's law shows that any single 

 type may repeat itself in lines throughout the 

 spectrum, according to this very law that King 

 enunciates. That the mean of all types should 

 give a uniform value throughout the spec- 

 trum, implies only that all types are fairly 

 well distributed throughout the spectrum. 

 Such a statement has no particular value. 

 For in a chance distribution, by the law of 

 probabilities, such uniformity must increase 

 as the number of lines increase. Mr. King's 

 Table VII. shows this. 



The most important place in Mr. King's 

 paper is given to a discussion of the relation 

 of the magnetic separation to a displacement 

 in the position of the spectral lines which 

 arises when the radiating particles are sub- 

 jected to an external pressure. The latter is 

 a well-known phenomenon. However, no ex- 

 planation of it has found extended acceptance. 

 Humphrey's theory involves a strong mag- 

 netic field induced by the rotating electrons. 

 As observed above, such a field is by no means 

 certain. But this theory and the prominence 

 of both phenomena in solar lines form the 

 principal reasons for an examination of their 

 possible relationship. Tables XL and XII. 

 show the ratio of the two quantities for dif- 

 ferent lines to vary twenty-five fold. Then at 



