Iodine by Multiplex Excitation, 683 



these companions for the present, we will consider the series 

 of doublets. The extent of this series has been somewhat 

 extended since the earlier investigation, and it now consists 

 of twenty-five members (all double), two of which are missing. 

 Some of the others are very faint. Certain lines were found 

 to be missing in the case of the resonance spectra of sodium 

 vapour, and this peculiarity is probably characteristic of all 

 resonance spectra. Dunoyer has made the suggestion that 

 each line in the series originates from a special configuration 

 of an aggregate of two or more sodium atoms or aggregates 

 of sodium and the residual gas in the tube, and that the 

 missing lines correspond to aggregates which are unstable or 

 which cannot exist. This is an interesting hypothesis, and 

 we are at once reminded of the equilibrium figures formed 

 by floating magnets, some of which are very unstable, while 

 others cannot exist at all. 



My first idea regarding the series of doublets was that a 

 close pair of absorption-lines were excited by the green line, 

 and that the spacing of the lines in the two series thus 

 originated was slightly different, the two series starting at 

 the same point (the green mercury line) . This would, how- 

 ei^er, give us a series of doublets with constantly increasing 

 separation, whereas the observed separation is very nearly 

 constant (about 2 Angstrom units). The law governing the 

 spacing of the doublets is not the same as that which obtains 

 in the case of the resonance spectra of sodium, in which case 

 I have found that the lines are spaced at very nearly equal 

 distances along a normal spectrum. This makes the wave- 

 length differences constant, but the law is by no means 

 strictly followed, and matters become only worse if we take 

 frequency differences. In the case of iodine the distance 

 between the lines increases steadily as we pass towards the 

 region of longer wave-lengths, but if we take the frequencies, 

 or what amounts to the same thing, the reciprocals of the 

 wave-lengths, we find that the frequency differences are 

 : approximately constant. They are not strictly so however, 

 and it seems possible that the departures may be the result 

 of slight perturbations which modify the periods. 



The wave-lengths were determined from photographs made 

 on Wratten and Wainwright films, with a concave grating 

 of 15,000 lines to the inch and a radius of curvature of six 

 feet. The wave-lengths to which they are referred are the 

 most recent values given for the neon lines by Priest, of 

 the Bureau of Standards (Interferometer determinations). 

 In the following table I have given the wave-length dif- 

 ferences, the reciprocals of the wave-lengths (frequencies), 

 and the frequency differences. The first six lines (starred) 



