MODELS OF POSITIVE ION 323 



in the same manner by the formation of large 

 molecules. 



The effect is proportional to 



3 : 



1/ atomic volume x (linear expansion), when the 



. . atomic weight 

 (atomic volume) = 3 r; 



^ ' density 



Further, from a consideration of the spectra of 

 impurities, the amount of material present does 

 not appear to be the important factor in pro- 

 ducing the shift in the spectral lines, but rather 

 the total pressure on the size of the molecular 

 groups and not on the actual density of the 

 vapour under consideration. An explanation of 

 this nature has been offered by FitzGerald 

 (see Larmor, Physical Review, May, 1901, 

 p. 311). The inevitable outcome of this theory 

 is that with sufficient pressure the spectrum 

 of one substance may be so displaced as to 

 correspond approximately with that of another, 

 that is, the next in the series of elements to 

 which it belongs, if the discharge could be taken 

 at a sufficiently high pressure.^ Thus the spectrum 

 of mercury may be converted into that of cad- 

 mium, that of cadmium into zinc, zinc into 

 magnesium, and so on. The obstacles, however, 

 in the way of accomplishing this would be very 

 great. The effect is one which the acute observer 



1 With our present appliances it does not seem to be 

 possible to accomplish this ; but it is conceivable that such 

 events may have taken place in Nature when the earth's crust 

 was in a semi-molten state. 



y 2 



