494 



SCIENCE 



[N. S. Vol. XL. No. 1031 



From an examination of the data it seems 

 probable to me tbat the effect is simpler than 

 Stark implies, and that such a relation as he 

 suggests does exist. The relative change of 

 frequency An/n is equal to AX/A, the separa- 

 tion of corresponding, symmetrically placed 

 components divided by the wave-length of the 

 particiilar line. If we plot this AA/A as a 

 function of the term number, smooth curves 

 drawn through the points are found to agree 

 closely in slope, etc., differing only in the 

 number of the terrrn at which they start. 

 Figs. 1 and 2 show the results for hydrogen 

 and helium I. The numerical data are given 

 in the following table. The numbers in 

 brackets are the term numbers. 



Hydrogen differs from the other elements in 

 that the components polarized parallel to the 

 field, and those polarized perpendicular to it 

 lie on different curves. 



The smooth curves shown in the figures all 

 correspond to the equation 



(1) 



lX10' = 0.89 (n — p) +0.01 (w — j 



where p is the number of the term where the 

 curve begins in each case. 



These constants refer to a field of 28,500 

 volts per cm. Assuming the separation to be 

 proportional to the field, which Stark proved 

 to be true for the hydrogen series, the equa- 

 tion may be written 



(2) ^Xl0' = 3.l(n — p) +.035 {n — py, 



where E is expressed in volts per cm. 



The reduced separations, AA/A, of the Hel 

 lines agree quite closely with this equation; 



the deviations in the case of the hydrogen 

 lines are larger, but both positive and negative. 

 Curves given by Stark for the variation of the 

 separation with the field strength^ seem to 

 show a greater accuracy for his measurements 

 than these deviations would imply. Yet a 

 simple curve can not be drawn through the 

 points representing the outer components of 

 the hydrogen lines so as to fit them better 

 than that corresponding to the above equation. 

 And in the reproductions given of the original 

 photographs, the lines are heavy and not paral- 

 lel and do not seem capable of more accurate 

 measurement. 



In the ease of helium, the separations for 

 lines polarized parallel to the field were con- 

 sistently found to be greater than for the corre- 

 sponding lines polarized perpendicularly. If 

 the effect is real, as it seems to be, the con- 

 stants of the equation would have to be slightly 

 different for the two sets of components. 



In the case of lithium, the reduced separa- 

 tions for 38,000 volts agree with the separa- 

 tions for the corresponding lines of helium. 

 When reduced to 28,500 volts they are a 

 quarter less. The measurements are stated by 

 Stark to be less accurate than for helium, so 

 whether the difference is real remains to be 

 proved. At least the relative separations of the 

 different terms of the series are the same. 



It should be stated that as regards asymmetry 

 of position and intensity, the corresponding 

 lines of hydrogen and helium apparently do 

 not behave alike. In fact. Stark reports differ- 

 ences in the behavior of lines of the same 

 series. As he suggests, the very complexity 

 of the phenomenon makes it a most promising 

 field in which to search for clues to a knowl- 

 edge of atomic structure. It may be too soon 

 to find regularities and the agreements noted 

 above may be accidental. But I think not. 

 If not, they suggest that there is something 

 which is common to the atoms of hydrogen and 

 helium, in addition to the presence of electrons 

 in both. 



Gordon S. Fuloher 



Wisconsin University, 

 July 11, 1914 



3 L. c, p. 997. 



