654 Professor A. Fowler [March 31, 



was seemingly given by the subsequent observations of Evans and 

 Stark, who found, as I have since found myself, that the line 4686 

 can be obtained from helium tubes which show no trace of hydrogen. 

 But it is always difficult to be certain that no hydrogen remains in 

 the tubes ; it is apt to hide itself in the glass, or in the electrodes, 

 and to come out when not expected. 



Other evidence in favour of Bohr's view, however, was found in 

 a more general investigation of series of enhanced lines — that is, of 

 lines which are intensified in passing from the arc to the spark 

 spectrum. In -the first instance, some additional lines of magnesium 

 associated with the well-known spark line 448 1 were discovered, and 

 it was found that they could be arranged in a series of the new 

 variety indicated by Bohr in connexion with the 4686 series of the 

 helium tubes. Further experiments revealed many other enhanced 

 lines of magnesium which had previously escaped observation, and it 

 was found that they formed a large family of related series, generally 

 similar to the ordinary series, but differing from them in requiring a 

 " series constant " having four times the value of that derived from 

 hydrogen. The same was found to be true of the enhanced lines of 

 calcium, strontium and barium. 



We have therefore now to recognize two distinct classes of series : 

 (1) series of the arc type, having Eydberg's X for the series constant ; 

 and (2) series of the spark type, having 4 X for the series constant. 

 No relation has been found between the arc and spark series occurring 

 in the same spectrum, and they seem to be as distinct as if they 

 belonged to different elements. A logical extension of Bohr's theory 

 leads to the supposition that series of the usual type are produced 

 when only one electron is detached from each atom by the exciting 

 agency, and those of the spark type when two electrons have been 

 removed and one is regained by the nucleus. 



Xow the 4686 series is clearly one of the spark type, both as 

 regards its mode of production and the formula which represents the 

 lines.. It can therefore no longer be supposed to be associated with 

 the hydrogen series, which is of the arc type. We now see that the 

 numerical relations between them are incomplete, and only approxi- 

 mate, and that they furnish no evidence of a common origin. Taking 

 account also of the experimental results, we are therefore forced to 

 the conclusion that the 4086 series is not due to hydrogen but to 

 helium, as required by Bohr's theory. If it was one theory that led 

 us astray in the first instance, it was another that helped to set us to 

 rights. Nevertheless, the assignment of the 4686 series to a form 

 of helium may now be considered to be independent of any theory. 



The chief outstanding point needing confirmation by observation 

 was that the Pickering series— which should run parallel to the 4686 

 series and also form a series of the spark type — should include inter- 

 mediate lines very close to the hues of the hydrogen series and on 

 their violet sides, as indicated in Ym. 6. The existence of these 



