1 lO 



NATURE 



[March 24, 192 1 



in which the electrons may move, and each group 

 corresponds to one of the original circular orbits. 

 Some of the orbits in each group are elliptical. 

 It appears that the energy of the electron would 

 be the same in all the orbits of any one group 

 were it not that when an electron moves in an 

 ellipse its velocity is not always the same. Now 

 a fast-moving electron shows a variation in mass 

 when its speed alters, and this does affect slightly 

 the energy of the orbit. Consequently, the elec- 

 tron that steps from an orbit belonging to one 

 group to an orbit belonging to another group 

 may part with an amount of energy which is not 

 always exactly the same. The frequency of the 

 consequent radiation may, therefore, have two 

 or more values differing slightly from each other : 

 the single spectrum line is doubled or trebled. 

 This is what Sommerfeld calls the " fine structure " 

 of the lines. 



Now there is far more than mere speculation 

 in this. The formula which Sommerfeld gives as 



KiG. 2. — Model of the arrangement of carbon atoms in the diamond. All 

 the atoms are alike, but those represented by light spheres differ in 

 orientation from those represented by dark spheres. 



the result of an analysis which is as reasonable as 

 can be expected does more than account for 

 known effects ; it has predicted the existence of 

 numerous lines, and even their intensities, and 

 the predictions have been verified by experiment 

 in the most remarkable way. The story is told 

 in Sommerfeld 's work on "Atom-building": a 

 story of the work of himself, Bohr, and others 

 during the last six years or so. 



We see that in this fundamental inquiry into 

 the nature and properties of radiation the electron 

 plays a very direct and important part. Our eyes 

 are designed to detect waves, not electrons, and 

 so our first attention is directed to radiation in 

 wave form. But we now find that radiation 

 energy may alternatively be carried by electrons, 

 and that many things become clearer when we 

 appreciate this fact. We can make further pro- 

 gress in our understanding of radiation, and in- 

 deed in our understanding of the electron, only by 

 getting to know more about the reciprocal con- 

 NO. 2682, VOL. 107] 



version of one form of energy into the other, 

 since evidently it is one of the most frequent and 

 most fundamental operations in Nature. 



So far our conception of the structure of an 

 atom would consist of a positive nucleus and of 

 electrons attached thereto in some way, with 

 the further idea that the energy attached to these 

 electrons can have only certain definite values. 

 Bohr assumes that they have these values because 

 they can move round the core in certain orbits 

 only, and Sommerfeld enlarges this idea, as 

 already explained. But, of course, this can be no 

 more than a partial picture of the whole atomic 

 structure. The atom so conceived cannot fill the 

 part required of it in the building of molecules 

 and crystals. 



When we come to examine these structures we 

 find atoms attaching themselves to each other 

 through the action of forces which cannot always 

 be considered as acting from centre to centre. 

 For instance, the arrangement of the carbon atoms 



Fig. 



-An extra ball can be in.serled in the model without disturbing, 

 other ball.s. 



in a diamond, as recently determined by X-ray 

 methods, is such that every atom is at the centre 

 of gravity of four others, arranged round it in 

 tetrahedral fashion, as shown in the model. The 

 representation of an atom by a smooth sphere 

 and nothing more would be in agreement with 

 the idea that the properties of the atom in any 

 one radial direction are exactly the same as in 

 any other radial direction, and that any forces 

 between two atoms are between centre and centre. 

 But if that were the case, the carbon atoms would 

 pack themselves together more closely than they 

 do. As a matter of fact, if the top of this model 

 is lifted, another carbon atom can be inserted and 

 the top replaced exactly as it was. If a 

 more extensive model were employed, it would 

 be seen that just twice as many atoms could be 

 packed into any volume as are already there. 

 We must conclude that there are definite sub- 

 centres of force on the outskirts of the atom, 

 and that in the carbon atoms of which the dia- 



