656 



NATURE 



[November 3, 1923 



attack would be succcMful in completing and classify- 

 ing these spectra. 



These are the broad outlines ; let us now turn to 

 finer details. It is well known that the theory, though 



It ^ ; .1 Information about all optical spectra, 



g(, : (lid in all its finer details the spectrum 



due U> ilic binding of the first electron. The only 

 spectra of this type yet experimentally realised are 

 what we may now call HI and Hell, that is, the 

 spectrum of atomic hydrogen of which the most con- 

 spicuous feature is the well-known Balmer Series, 

 and the spectrum of ionised helium. Now the pre- 

 dictions of the theory not only give the exact position 

 of each line, but, as is well known, also assign to each 

 line a definite complex structure. Under very high 

 dispersion and first-class conditions this structure can 

 be observed. In the case of Hell, where the separa- 

 tions are greater and the conditions less severe, the 

 confirmation of the theory was completed some 

 years ago by the photographs of Prof. Paschen and 

 others. Until recently, however, the similar more 

 difficult experiments for the Balmer Series have been 

 inconclusive and discordant. For this series the 

 theory demands that each line should split into two 

 close lines of the same frequency difference, which 

 should themselves have a still finer detailed structure. 

 Into this we need not enter, beyond saying that this 

 ultimate structure should slightly reduce the apparent 

 separation of the lines of longest wave-length, parti- 

 cularly Ha. Now the last lacuna has been filled by a 

 brilliant piece of work in Prof. McLennan's laboratory, 

 for good photographs have been obtained showing 

 clearly the main separations of the five lines of longest 

 wave-length. The agreement with the theory is com- 

 plete. To illustrate the fineness of the detail it may 

 be mentioned for example that for the fourth line H«, 

 wave-length 4101-73x10"* cm., the theoretical main 

 separation is only 6-i x lo"^** cm. 



It has been known for some time that the energy 

 required to remove one electron from neutral helium 

 was (in the usual terminology) about 25 volts. We 

 thus express the energy acquired by an electron in 

 falling freely through such a potential difference. 

 Thus expressed the energy required to remove the 

 most lightly bound electron is known as the ionisation 

 potential. Until recently the known part of the 

 spectrum He I made no allowance for a normal atom 

 in which the electrons were so firmly held. It ap>- 

 peared that these ought to be a series of lines in the 

 far ultra-violet, not hitherto observed, associated with 

 the reconstruction of the normal atom. Four such 

 lines have now been observed by Lyman. The wave- 

 lengths are very short, from 500 to 6ooxio~* cm., 

 and indicate an ionisation potential of 24-5 volts, 

 in good agreement with direct observation. Our ex- 

 perimental knowledge of He I is thus properly rounded 

 oflf. Much valuable work on the theon,- of this spectrum 

 has also been completed, but the results are negative. 

 It is now certain that none of the models so far pro- 

 posed possess the proper permissible orbits, computed 

 according to the rules of the present quantum theory, 

 to account for the spectrum He I and the ionisation 

 potential. It is an advance to be sure of this. The 

 interaction of the two electrons in helium (and a 

 fortiori the Z electrons in the general atom) must be 



NO. 2818, VOL. 1 12] 



even more subtle, and the detaiUd theory' of their orbits 

 must lie even deeper, than has been hoped hitherto. 



Recent work has shown the very great value of the 

 study of the absorption spectra of atomic vapours in 

 the coldest state in which the>' can be procured at 

 reasonable densities. Such vapours, as is well kno\Mi, 

 absorb selectively a number of sharp lines which are 

 a selection of the lines of the first optical (emission) 

 spectrum. But since the atoms of the vapour must 

 in general be in their normal state, only thosr lines 

 can appear which belong to atomic recf ns 



ending in this normal state. We can ' • • i 



from the whole mass of lines just those associate d 

 with one particular state of the atom, and that th( 

 most important. In this way certain difficulties hav< 

 been cleared up in connexion with the spectrum Al I 

 and its analogues. It had been believed that th( 

 normal orbit of the most lightly bound electron was 

 of the same type for all atoms — that is, specified by 

 a certain value (unity) of one of its quantum numbers. 

 This is the theoretical interpretation of the empirical 

 belief that the absorption spectrum would always 

 consist of the same type of series. But the known 

 facts about this group of spectra did not fit in with 

 this belief, and it is now definitely established by the 

 study of absorption spectra that this belief is fal?>( . 

 The normal orbit in question may have at any rai< 

 one or two for the value of this quantum number, 

 and has the value two for aluminium and its analogues. 

 Thanks to this we now know that our account of these 

 spectra is reasonably complete. The study of absorp- 

 tion spectra will doubtless prove of great value in 

 disentangling the difficult spectra of the lead-tin 

 group. A good start has recently been made in their 

 classification. 



Let us with Prof. McLennan conclude by referring 

 to the effect of an applied magnetic field on the atcmic 

 orbits, with which is bound up the question of the 

 way in which the atom orientates itself in space under 

 such an influence. The effect on the spectrum is 

 known as the Zeeman effect, and its study is proving 

 of the utmost importance to the theory' of atomic 

 structure. It is here that we shall probably win the 

 next advance. We can scarcely expound these ques- 

 tions shortly and cannot enter into details here. But 

 it may be said that the proper classification of the 

 empirical facts, largely the work of Prof. Lande, seems 

 already fairly complete, and that their theoretical 

 interpretation has been begun on a sure basis. We 

 must not, however, omit to mention the cognate 

 beautiful experiment of Stem and Gerlach, which 

 consists in directing atoms of silver of known velocity 

 through a strong non-uniform magnetic field. If the 

 atom possesses a magnetic moment it must be deflected, 

 unless its axis is always perpendicular to the field. 

 Such deflexions were observed, and appear to prove, 

 simply and directly, that the normal free atom of silver 

 possesses a definite magnetic moment and always 

 sets itself with its magnetic axis parallel or anti-parallel 

 to the field. Experiments such as these are of the 

 greatest importance. They admit of unambiguous 

 interpretations and pro^^de the necessant- strong points 

 from which the attack on the complicated Zeeman 

 effect and related phenomena can be securely launched. 



R. H. Fowler. 



