292 BELL SYSTEM TECIIXIC-IL JOURNAL 



ments of sliallow-hing electrons; and fnially, that the wa\'c-lcngths of 

 the excited rays lie below 13A, many of them in the inaccessible range, 

 some in the range newly opened to exploration. This is a most un- 

 fortunate coincidence, for instead of being able to apply laws which 

 prevail in other ranges to compensate for our inability to measure 

 wave-lengths in this range, we have to expect distinct laws within 

 it. Must shallow lying electrons be extracted altogether from the 

 atom if they are to be displaced at all or ha\e they certain transient 

 sojourn orbits to some or all of which they may be raised by electron- 

 impacts? Do the emitted rays result from a step-by-step return 

 of the displaced electron? or from a return in a single leap? or from 

 a rearrangement of the remaining electrons? or from a compounding 

 of changes of the two latter types? So long as the emitted wa\e- 

 lengths are not measured, these cjuestions cannot be answered with 

 confidence. 



Some little can be inferred fnmi numerical relations among excita- 

 tion-potentials. McLennan and Clark, for example, observed three 

 excitation-potentials of lithium, at 37.0, 31.8 and 12.0 volts. The 

 first two of these voltages stand nearly in the ratio of the first two 

 frequencies of the Lyman series in the h\drogen spectrum, which 

 suggested to the discoverers that the processes involved in the excita- 

 tions were the raising of a X-electron to the first and second of a 

 pair of transient-sojourn orbits, standing in the same relation to the 

 normal orbit of the iC-electron as the orbits of energy-values —Rh/A 

 and —Rfi/9 stand to the normal orbit of energy-value —Rh in the 

 hydrogen atom. Tliat is to say, they conceive these excitation- 

 potentials to be ((imij.iiable to resonance-potentials, and the K- 

 electron of litliiuni to beiiaxe liku a \alence-electron. They also 

 found excilalion-polenlials of ber\llium at 20.3 and 10.0, and of 

 boron at 27.92 and 23.45. The ratio of each pair of numbers is about 

 equal to the ratio of the first two frequencies of the Balmer-series, 

 suggesting that these are resonance-potentials of an L-electron; the 

 details of the analogy may be left to the reader to work out. Each 

 of the latter elements disiilayed additional higher potentials, to be 

 associated with the X-electrons. Rollefson lately disco\ered seven 

 excitation-potentials of iron in the range between 160 and 264 volts, 

 expressible by a formula {a — b «-') if the integer \alues 5, 6, 7, 8, 9, 10 

 and 12 are successively gi\eii to ;;. If the.sc se\en potentials corre- 

 spond to ele\ations of a certain shallow-King electron to seven tran- 

 sient-sojourn orbits, the extraction-iiotcntial for this electron can be 

 calculated b\- an extrajiolation (so also in the cases cited from Mc- 

 Lennan and Clark). Rollefson interprets certain other excitation- 



