396 



NA TURE 



[September 15, 1923 



give their assistance in any scientific capacity in the 

 event of war coming dcspltd their efforts to prevent 

 it. Without trained technical assistance the warfare 

 of the future will be impossible. If they wish to 

 carry a rifle, by all means let them : they will not do 

 much harm with a rifle. But a refusal to give their 

 technical .issistance would not only bring any war to 

 a standstill, but would also be the strongest possible 

 guarantee against it breaking out. If this measure 

 is not taken, and promptly, we may well fear that the 

 new order that rises from the rums of the old will 

 persecute science as whole-heartedly as ever did the 

 rulers of the Middle Ages, and with better reason. 



W. D. Evans. 

 68 Argyle Road, West Ealing, 

 London, W.13. 



The Heisenberg Theory of the Anomalous 

 Zeeman Effect. 



In his theory for doublets Heisenberg {Zeit. f. 

 Physik, 8, 273, 1922) assumes that the atom may be 

 looked at as made of two parts : (i) the shell and (2) 

 the valence electron. Expressing angular momenta 

 in multiples of A/2t and choosing the direction of 

 the angular momentum of the shell as positive, the 

 electron is allowed to have angular momenta 

 I = i. ± *. ±li • • • in t\ies,p,d, . . . states respectively, 

 and the shell has in all of the states the angular 

 momentum J. The observed Zeeman patterns show 

 that I = ^ in 2/>, and I = - f in 2/)j. The observed 

 energy levels show that the energy in 2^1 is higher 

 than in 2p2. The writer experienced the following 

 difficulty in accounting for this relative pKJsition of 

 energy levels. 



Various hypotheses can be made as to the nature of 

 the interaction between the shell and the electron. 

 We may suppose, for example, that the magnetic 

 field of the electron induces a precession in the shell 

 in a manner analogous to that in which an external 

 magnetic field induces a precession in the electronic 

 orbit. We then suppose, too, that the field due to the 

 shell produces a precession of the electron. The 

 contribution to the kinetic energy of each of thes^ 

 precessions is - mH cos ^ , where m, H, ^ are respectively 

 the magnetic moment of the shell, the field at the 

 shell due to the electron, and the angle between the 

 positive directions of /x and H. The contribution of 

 both is - 2mH cos ^. The mutual energy of the 

 magnetic fields is + mH cos ^. There is no contribu- 

 tion to the energy of the electric field because the 

 radius of the orbits is unchanged (Sommerfeld, 

 " Atombau und Spektrallinien," third edition, p. 380). 

 The energy to be added to that coming from other 

 sources is then -^Hcos^. On this hypothesis, 

 therefore, the 2pi state has the lower energy, while 

 the reverse is actually the case. 



If there were no induced precession in the shell, but 

 if the electronic precession should be still hypothesised, 

 the 2pi and the 2p2 levels would coincide. 



If the shell and the electron should be supposed to 

 have no induced precession, the energy of the mag- 

 netic field becomes the only source for the energy of 

 separation of the 2^2 levels. This energy is + mH cos ^ 

 and thus makes the 2/>, level the state of higher 

 energy, as it is actually observed to be. 



It seems questionable, however, whether the 

 hypothesis just made can be maintained, for it pre- 

 supposes that the dimensions of the orbits of the 

 valence and the shell electrons are the same in the 

 2pi and the 2p2 state. This may be contrary to 

 quantum conditions if the energy of the magnetic 

 field is considered as kinetic energy. If two electrons 

 should be constrained to move on the opposite ends 



NO. 281 I, VOL. I 12] 



of a diameter of a circle of variable radius (as in 

 Bohr's first helium model), the kinetic energy becomes 

 of the form : 



where m,, m,, v,, r, are respectively the masses and 

 velocities of the electrons and M^v,t/, ia the mutual 

 energy of their magnetic fields. The case ly 



analogous to the hydrogen atom, and a s. .^n 



in well-known formulas shows that the total energy 

 becomes decreased if M,, is increased. The reason 

 for this is traceable to a decrease in the radius of 

 the orbit. Thus again the effect on the 2/>, level 

 is opposite to that observed. 



1 he matter of the sign of the energy in the doubl 

 terms thus does not appear to the writer to be suii 

 ciently clear. 



The same question of sign is present in the case oi 

 triplet terms. In addition to this the 2/>, term of 

 triplets does not seem to be accounted for properly 

 by Heisenberg. His arrangement of angular momenta 

 accounts for the energy level of the 2/), state. I 

 obtain, however, a different result for the Zeeman 

 resolution. On going through Heisenberg's calcula- 

 tion his lines 5, 6, counted from the bottom of pat 

 292 and leading to the equation cose = m//),, do ni 

 appear obvious. His pi^ is the projection of a vector 

 in the direction J, m is the projection of the 

 same vector in the direction H, and e is the ang!' 

 between J and H. The above equation is tht . 

 cos (JH) = cos(AH)/cos(AJ), which does not appear 

 to be generally valid. It becomes correct, however, if 

 A and J are the same. They arc the same for doublets 

 and for the 2/),, ip^ terms of triplets, but not for the 

 2pj. term. G. Breit 



(National Research Fellow 

 The University of Minnesota, 

 Minneapolis, U.S.A. 



Thunderstorms and Ozone. 



The question — What chemical changes, if any, ar< 

 associated with atmospheric electrical discharges ? — 

 does not appear hitherto to have received a definite 

 answer. Nitrogen peroxide and ozone are both 

 referred to in scientific literature, although neither 

 appears to have been satisfactorily identified, and 

 their presence has been perhaps inferred from th- 

 phenomena observed while " sparking " air b\ 

 artificial means. 



No reliance can be placed upon observations made- 

 with guaiacum or starch-potassium iodide paj)ers, 

 and the work of the more serious investigators on 

 ozone in the air (Pring, Proc. Roy. Soc., 1914, 90a. 

 204; Hayhurst and IMng, Jour. Chem. Soc., 1910. 

 868; Kaiser and McMaster, Am. Chem., July i. 

 1908, 39, 96; Henriet and Bonissy, Comp. rend., 

 1908, 146, 977 ; and the older work of Houzeau. 

 Schone, H. de Varigny, Hached and Amy, and 

 Thierry) has thrown no light on this subject. 



Unexpectedly clear evidence on the above p>oint 

 was obtained by me in connexion with the severe 

 thunderstorm which passed over the metropolis from 

 south to north, during the early hours of July 10 

 last. The lightning on this occasion was generally 

 described in the London press as the most \'i\-id and 

 prolonged display in living memory {vide Nature, 

 July 21, p. 113). 



I have for some time been measuring the prop>ortion 

 of certain variable gaseous constituents in London 

 and country air, and succeeded last spring in working 

 out an improved method of estimating ozone, in 



