44 SECTIONAL ADDRESSES. 



' red ' and ' blue ' spectra of argon. By using a cylindrical tube for this 

 form of discharge, L. and E. Blocb^® were able to distinguish the lines of 

 mercury representing the first and second stages of ionisation by observing 

 the extension of the lines from the edges towards the centre ; and in the 

 same way, Esclangon** has sorted out lines attributed to four successive 

 spectra of cadmium. 



Some of the newer methods have definitely brought additional spectra 

 within the range of laboratory experience. Thus, by electric bombard- 

 ment of lithium at a temperature of 1,000 deg. C, Werner^" succeeded in 

 producing the spectrum of ionised lithium, which had resisted all attempts 

 to obtain it by ordinary spark discharges. In accordance with theoretical 

 expectation, the new spectrum was resolvable into series having 4R for 

 the series constant, and was found to correspond closely with the spectrum 

 of neutral helium. A similar, but less complete, spectrum of ionised lithium 

 was also obtained by Schuler.^^ who made use of the ' hollow cathode ' 

 method, and by Morand*^ with an apparatus in which the metal was 

 excited by anode rays. 



One of the few sources which can at present be employed for observa- 

 tions in the extreme ultra-violet is that known as the 'vacuum spark,' 

 which has been extensively utilised by Millikan^^ and his colleagues. In 

 this method the spectrograph and spark chamber are highly evacuated, 

 and a powerful spark is made to pass between electrodes separated by 

 one or two millimetres or less. The spectra include lines representing 

 various stages of ionisation in a single photograph, but their disentangle- 

 ment can be effected with the help derived from the analysis of spectra 

 obtained under better controlled conditions in more accessible regions. 

 The interpretation of such spectra, however, has been greatly simplified by 

 the recent remarkable work of Bowen and Millikan,^* who, by utilising high 

 orders of the grating, have obtained a high degree of resolution of complex 

 groups, and wave-lengths of a degree of accuracy approaching that 

 obtainable in ordinary parts of the spectrum. 



Another class of ' experiments,' as I have previously mentioned, is 

 provided by the heavenly bodies. Saha's theory of high-temperature 

 ionisation, further developed by Fowler and Milne^^ and by Miss C. H. 

 Payne, 8« has already been utilised in the prediction of the ionisation 

 potentials of certain multiply-ionised atoms for which the structures of 

 the corresponding laboratory spectra have not yet been sufficiently 

 determined to indicate the energies of the normal states. In this way it 

 is conceivable that we may obtain approximate values of the actual 

 energy levels in some of the complex atoms for which only relative values 

 can at present be directly determined from the spectra. 



" Jour, de Phys., vol. 4, p. 333 (1923). 

 «» Dissert., Paris, 1926. 



50 Nature, vol. 115, p. 191 (1925). 



51 Die Naturwiss., July 11, 1924. 



52 Comptes Rendus, vol. 178, p. 1528 (1924). 

 " Astrophys. Jour., vol. 52, p. 47 (1920), etc. 

 51 Phys. Rev., vol. 26, p. 150 (1925). 



55 Mon. Net. R.A.S., vol. 83, p. 403 (1923) ; vol. 84, p. 499 (1924). 

 5^ ' Stellar Atmospheres,' Harvard Obs. Monographs, No. 1 (1925). 



