Resonance Potentials for Electrons in Metallic Vapours. 43 



intensity of the lines remains the same when the voltage is 

 further increased. However, if the K group is to be looked 

 upon as a series of lines, the above comparison to the 

 behaviour of thallium is not justified. 



The single line spectrum of thallium has never been 

 observed. McLennan has suggested that it lies o far in the 

 ultra-violet. Our work indicates that \= 11515 A., lying in 

 the infra-red, is the true single line spectrum of thallium. 



If we continue the analogy of the behaviour of thallium to 

 that of sodium and potassium we shall expect to find the 

 ionization potential determined by the convergence of the 

 principal series, \ = 4389. This corresponds to an ionization 

 potential of 2*81 volts, a value which is clearly disproven by 

 the present work. Accordingly thallium does not behave 

 similar to sodium and potassium, and actually in the light 

 of the work with other metals we should expect a diHerent 

 behaviour, since it belongs to a new family in the periodic 

 table as discussed above. The observed ionization occurred 

 at 7*3 volts corresponding to a convergence frequency (wave 

 number) of about 59000. The highest convergence frequency 

 so far known in the thallium spectrum is that of the shorter 

 wave-length member of the 2nd subordinate series, v = 49263. 

 The other member of this series converges at 41470- 

 McLennan * has attempted to predict 1*5 S for thallium by 

 assuming, as in the case of zinc, mercury, etc., that the single 

 line spectrum is given by the notation v=15 S — 2p 2 where 

 v corresponds to \ = 5351, the line appearing prominently in 

 the flame-spectrum of thallium, and 2p 2 = 49263 the con- 

 vergence of one member of the first subordinate doublet 

 series. In this manner he obtained the value v=l'5 S = 67947. 

 If one employs the proper value for the single line spectrum, 

 one obtains v = 1*5 S = 57946. This corresponds to an 

 ionization potential of 7*2 volts in good agreement with the 

 experimentally determined value of 7*3 volts. However, the 

 above method of deducing the convergence frequency 1*5 S 

 is incorrect. The term 1*5 S belongs to the principal series- 

 of single lines, and 2p 2 properly belongs to the principal 

 series of triplets. Since neither single line series nor triplet 

 series are known in thallium there is no means of deter- 

 mining 1*5 S or 2p 2 . If 2p 2 is taken to denote the con- 

 vergence of one member of the first subordinate series,. 

 z> = 49263, then the only series deducible is v = l'5 s — mp 2y 

 which represents the principal series converging at v = l'5s 

 = 22786. Our work accordingly predicts for thallium a 

 new series of the form 1/5 S— t?iP, a single line series 

 * McLennan, J. Frank. Irist. clxxxi. p. 201 (1916). 



