130 Mr. E. C. Kemble on the Probable 



at 585 A. U., due to atoms returning to the normal state from 

 (2, S) (crossed orbit system). In the early experiments o£ 

 Fricke and Lyman with slightly impure helium this line 

 was not observed. Later, the line was found under different 

 conditions with much purer gas. The experimenters are 

 inclined to attribute the discovery of the. line in part to an 

 increased intensity of emission in the purer helium. If such 

 a variation in the emission intensity with the purity of the 

 gas does occur, it may be accounted for by hypothesis (c). 

 Other things being equal, the emission intensity of the line 

 in question will be proportional to the number of atoms in 

 the state (2, !S). Let us assume that the atoms can pass from 

 (2, S) to the normal state much more readily through the 

 intermediate state (i) than by the direct path. Then when 

 there is considerable impurity in the gas, the atoms knocked- 

 into the state (2, S) will immediately pass through (i) back 

 to the normal state, and will not accumulate to any marked 

 extent. On the other hand, when the helium is very pure,, 

 the atoms can return to the normal state only by the difficult 

 direct route, and will therefore accumulate markedly in the 

 the state (2, S), and o cause a greatly increased emission of 

 the wave-length 585 A.U. 



Hypothesis (b) is essential to the validity of the above 

 argument. Paschen's experimental investigation of the ab- 

 sorption and fluorescence of the lines at 10830 A.U. and 

 20582 A.U. gives additional evidence for this hypothesis.. 

 Paschen (I. c.) passed a weak discharge through a vacuum- 

 tube containing helium, and found that the lines at 10830 A.U. 



O ~ ' 



and 20582 A.U. became strongly absorbing when the current 

 density was only one ten-thousandth part of that required to 

 produce an appreciable absorption of other lines of the 

 spectrum. The absorption of the wave-lengths in question 

 is due to the transfer of atoms from the state (2, S) to the 

 state (2, P), and from (2, s) to (2, p). (See fig. 1.) Hence 

 the experiment proves that even with a very low current 

 density there is a large accumulation of atoms in the states 

 (2, Si and (2, s). This can occur only if the atoms which 

 are knocked from the normal state into either (2, S) or (2, s) 

 have difficulty in radiating their energy and reverting to 

 their original condition. 



It should be observed that since our first hypothesis (a) 

 rules out direct jumps from the principal series stationary 

 states of both systems to the normal, it requires that both 

 the line at 20582 A.U. and that at 10830 A.U. shall be 

 " resonance " lines in the narrow sense of the word : that is, 

 all the radiation of either wave-length which is absorbed 



