September 8, 1923] 



NATURE 



359 



Letters to the Editor. 



[ The Editor does not hold himself responsible for 

 opinio7ts expressed by his correspondents . Neither 

 can he undertake to return^ nor to correspond with 

 the writers of, rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 takeii of anonymous comm.unications?\ 



On the Regularities of the Spectral Lines of Iron 

 and the Atomic Magnetic Field. 



Since our short account of the method of observing 

 the Stark effect with a stabiUsed arc was written (see 

 Nature, March 31, p. 431), we have made experiments 

 on about twenty different metals. With elements 

 having a simple spectrum, as silver, copper, zinc, and 

 others, the separation of the lines into different series 

 is facilitated from the similarity in the winged 

 appearance of the lines in the strong heterogeneous 

 electric field near the anode, though the broadening 

 is generally asymmetric and there is some difference 

 among the polarised components. The examination 

 of many thousands of iron lines is not yet completed, 

 but choosing lines between \X 2400 to 3000 A showing 

 the simplest type of the effect, in which they are 

 enhanced and slightly shifted towards shorter wave- 

 lengths, we have found that a few lines can be 

 arranged in regular triplets, quartets, and sextets. 

 These mostly belong to spark lines. In addition to 

 these regularities, we can arrange the enhanced lines 

 into a large number of quadruplets as shown below : 



12 34 



The frequency diiference, Ai'(i,2), is equal to Aj'(3,4) to 

 a fraction of the wave-number per cm. The relations 

 between t^v[\,i) and A>'(2,3) are various, but the values 

 of Ai'(2,3) and Ai'(i,4), and especially those of Aj'(2,4) 

 and Aj'(i,3), are common to many of the quadruplets. 

 The remarkable numerical relation between Ai'(i,2)'s 

 is that they come out in groups as given in the 

 subjoined table : 



Values outside the ranges above cited do not appear. 

 Counting from group (a), the mean Av(i,2)'s, except- 

 ing the second, are almost exactly in the ratio 



1:2:3:4:6:7:8. 



It is singular that 5 does not enter in the above 

 r.itif) ; 111-' absence of this number will probably 

 uuiKilir t!ic principle of choice. Perhaps the above 

 ratio has sonic l)farin,L( on llie quantum theory, and is 

 connected wiili ilu: iiuur (|u;uitum number (" innere 

 Quantenzahl "). if we interpret the existence of 

 regular separations as due to the action of an atomic 

 magnetic field, the above relation seems to be one 

 aspect of Runge's rule in the Zeeman effect. Taking 

 363 as the standard separation, the above ratio can 

 be written as representing 1/6, 1/3, 1/2, 2/3, i, 7/6, 

 4/3. The intervals of quadruplets in group (g), for 



NO. 2810, VOL. I 12] 



which Aj'(i,2) =485, frequently occurs and is closely 

 related with the separations of numerous quadruplets, 

 so that it seems to have some important signification. 

 The same number occurs in two regular triplets. 



In forming these quadruplets, there is no criterion 

 but that of taking the interval Ai'(i,2) =Aj'(3,4), with 

 corresponding symmetry in the intensity of lines. 

 Analysing the distribution of lines, it is found that 

 the same line can be looked upon as belonging to more 

 than one quadruplet. Most of them are perhaps not 

 real quadruplets, but belong to portions of more 

 complex separations, the true nature of which is 

 difficult at present to unravel. Dealing with many 

 hundreds of Unes, chance coincidences may frequently 

 occur, yet the probability of the existence of regular- 

 ities can scarcely be doubted. Eventually, we shall 

 be able to arrange the iron lines in spectral series by 

 utilising the Stark effect, if such really exist. 



If we assume that the separations are due to the 

 Zeeman effect of the atomic magnetic field, they will 

 probably amount to aliquot parts of a normal triplet, 

 if they follow Runge's rule. This is not usually 

 obeyed in iron lines by applying an external field, but 

 if we roughly assume that the triplets (Av(i,2) =485) 

 are normal, the field must amount to 10' gauss, which, 

 will approximately give the order of magnitude of 

 magnetic force acting on the light-emitting electrons. 

 As the above value of A»' corresponds to the widest 

 separation observed, the field will be generally 

 smaller ; by choosing Ai/=354, which is found in one 

 of the triplets and a number of quadruplets, the 

 atomic field is found to be 6.6 x 10* gauss, coinciding 

 with the value found by Weiss from experiments on 

 magnetisation. This gives strong support to the 

 magneton theory, and though the problem of atomic 

 field is still in a hypothetical stage, the close agree- 

 ment of the results obtained from measurements 

 made on different phenomena is worthy of further 

 consideration. 



In Bohr's equation for calculating the frequency of 

 light, the change of electric energy is taken into account 

 only when an electron passes from one orbit to 

 another during the emission of light. If we assume 

 that, in the interior of an iron atom, a strong magnetic 

 field as given above is prevalent, we must also 

 examine the change of magnetic energy during the 

 emission. This adds a further complication to the 

 discussion, especially when the orbits are not coplanar. 

 The question is, where does the magnetic field come 

 from ; does the seat lie in the nucleus or in the 

 orbital motions of electrons ? The intricate nature 

 of the spectral lines in ferromagnetic metals may 

 ultimately be traced to the existence of an inner 

 atomic field. 



The list of lines and different separations will be 

 published shortly in the Japanese Journal of Physics, 

 vol. 2. * H. Nagaoka. 



Y. Sugiura. 



Institute of Physical and Chemical Research, 

 Hongo, Tokyo, July 20. 



Embryology and Use -Inheritance. 



Having read with great interest in the supplement 

 to Nature of August 18 the Huxley lecture of my 

 friend Sir Arthur Keith, and the comments upon it 

 in " Current Topics and Events " of the same issue, I 

 should like as an embryologist to make some remarks 

 on the subject. Sir Arthur, in his fascinating style, 

 describes the manner in which during development 

 indifferent embr^'onic cells are marshalled so as to 

 build up structures of functional and adaptational 

 He arrives, however, at the surprising con- 



clusion that " functional adaptation 



is a property 



