46 SECTIONAL ADDRESSES. 



This rule permits only the pair of green nebular lines in doubly ionised 

 oxygen which we have discussed ; and in deference to it only two are 

 observed, instead of the three which apart from this might have been 

 expected from the triplet ground state. 



Yet we find the blue singlet line X4363 of this ion violating the same rule, 

 and the same applies to the analogous case of the aurora line, if we adopt 

 McLennan's view of its position in the scheme of the arc spectrum. 



In the case of the mercury spectrum, which lends itself well to experi- 

 mental observation and of which much detail is known, we have laboratory 

 examples of the violation of this rule, as originally shown by experiments 

 of Takamine, Fukuda and other Japanese physicists. The lines were 

 originally obtained under conditions where a strong electric field was acting, 

 and this was sometimes urged in mitigation for breaking the rule. Again, 

 the lines were of low intensity, and this too was thought to be a partial 

 excuse. 



Whatever might have been thought of these apologies originally 

 their irrelevance was, I think, clearly shown in some experiments of my 

 own, in which one of the ' forbidden ' mercury lines was obtained as the 

 second strongest line in the entire mercury emission spectrum, in the 

 vapour passing through a discharge, but altogether away from the region 

 in which the discharge itself was taking place, and consequently in the 

 absence of an extraneous electric field. 



In another experiment I was able to obtain the other forbidden line 

 as an absorption line in unexcited mercury vapour, and thus apparently 

 in the absence of any disturbing conditions. In this experiment the 

 quantity of vapour used was very large, about ten million times the amount 

 required to bring out the resonance line of mercury in absorption. The 

 probability of the transition thus indicated is very low, and for the other 

 forbidden line it is apparently still lower. But for all that, as we have 

 seen, this forbidden line can be got in considerable intensity in emission. 

 The necessary condition in the mercury experiments appears to be a 

 large accumulation of mercury atoms in the relevant metastable state, 

 so that even with a low probability of transition for the individual 

 excited atom a considerable number of transitions occur. 



It has even been proposed to define a metastable state as one with a 

 low probability of transition. This takes us far from the original concep- 

 tion, and makes ' metastability ' merely a question of degree. Some 

 recent results which I hope to bring before the section at a later stage in 

 our proceedings seem to indicate that even the normal excited state may 

 possibly persist for a much longer time than has hitherto been supposed. 

 If this conclusion is accepted, a far-reaching revision of our present 

 notions may become necessary. The general softening of outline in our 

 picture of atomic events resulting from the substitution of wave groups 

 for particles seems likely to afford what is required, and to allow the 

 occasional transition downwards from a metastable state. 



Ozone. 



The case of the nebular spectrum affords an illustration of how spectro- 

 scopic theory, working on laboratory data gained in the remote ultraviolet 

 region, enables us to some extent to turn the difficulties which arise from 



