340 



NA TURE 



[March i6, 1922 



Letters to the Editor. 



\The Editor does not hold himself responsible for 

 opinions expressed by his correspondents. Neither 

 can he undertake to return, or to correspond with 

 the writers of rejected manuscripts intended for 

 this, or any other part of Nature. No notice is 

 taken of anonymous communications^ 



A Magnetic Model of Atomic Constitution. 



The following preliminary description of a new 

 magnetic model atom is given on account of the 

 remarkable coincidence between the results we have 

 obtained experimentally and the views which Dr. 

 Aston has put forward with regard to the atomic 

 constitution of certain of the lower elements. 



An attempt has already been made by Sir J. J. 

 Thomson to draw a parallel between atomic structure 

 and the arrangements into which Mayer's magnets 

 group themselves ; but this parallelism suffered from 

 the fundamental flaw that there was no numerical 

 relationship between the strength of the central pole 

 and the united strengths of the floating magnets. In 

 order to have a true parallel, it is essential that the 

 strength of the central pole should increase -pari passu 

 with the number of magnets afloat at one time ; for 

 only in this way can a " magnetic neutrality " be 

 obtained which will represent the electrical neutrality 

 of the atom. 



In our preUminary work we achieved a close 

 approximation to this state of affairs by using a 

 series of equal magnets. A single magnet was fixed 

 vertically at the bottom of a lead tank filled with 

 water, above the surface of which a second vertical 

 magnet was held in a glass tube. On the water 

 surface, a third vertical magnet was floated by means 

 of a cork disk. All the magnets had their north 

 poles upward. Since both poles of a floating magnet 

 repel those of another floating magnet, it is necessary 

 to have two fixed magnets for each floating one, in 

 order to establish neutrality. When two floats are 

 in the dish, two magnets are placed at the bottom 

 and two m.agnets in the glass tube, and so on. In 

 this way, the strength of the central pole is always 

 exactly equal and opposite to the combined powers 

 of the floating magnets. 



When this arrangement is tested with numbers of 

 floating magnets corresponding to units of atomic 

 weight in the various elements, the following results 

 are obtained. The magnets arrange themselves into 

 two sharply defined and clearly separated groups, the 

 central one of which may for convenience be termed 

 the nucleus, whilst the second group may be named 

 the ring. In the table below are the experimental 

 results showing the distribution of the magnets into 

 the two groups : — 



Total minus 



Nucleus. 



O 

 2 



3 

 3 

 4 

 5 

 5 

 6 



7. 

 8 



9 

 10 



Corresponds to 



Hydrogen. 



Helium. 



Lithiuiii-6. 



Lithium-7. 



Beryllium. 



Boron- 10. 



Boron-ii. 



Carbon. 



Nitrogen. 



Oxygen. 



Fluorine. 



Neon. 



Inspection will show that this series of groupings 

 corresponds exactly to the suggestions put forward 

 by Dr. Aston. He regards each unit of atomic weight 



NO. 2733, VOL. 109] 



as a proton carrying one positive charge ; and the 

 atomic number of the element is, as Prof. Soddy 

 suggested, the algebraic sum of the positive and 

 negative charges in the atomic nvicleus. In the new 

 atomic model, it will be seen, the systems arrange 

 themselves spontaneously so as to form a perfect 

 analogy to the hypothetical atomic structures. Thus 

 in the case of Boron- lo the central pole contains 

 ten magnets ; five of these are " neutralised " by 

 the five floating magnets of the nucleus, leaving a 

 net " charge " of five, which is the atomic number 

 of boron. In the case of Boron-ii there are eleven 

 magnets in the central pole ; six floating magnets 

 " neutralise " six of them, leaving free five — the 

 atomic number of both isotopic forms of boron. 



All the known elements and isotopes in the series 

 find their exact parallel in these magnetic models ; 

 and the coincidence, if it be merely coincidence, is 

 certainly surprising. 



It should be pointed out that hydrogen is anoma- 

 lous, since obviously a single magnet cannot simul- 

 taneously form part of two groupings, nucleus and 

 ring. If the floating magnet in this case be reckoned 

 as a ring magnet instead of a nuclear one, then 

 hydrogen would have the atomic number i. 



It seems worth while to point out that the number 

 of magnets in the ring is always equal to the maximum 

 valency of the element plus two units ; and the 

 occurrence of two magnets in the ring of the system 

 corresponding to helium suggests that all the eight 

 elements up to fluorine are built up with two residual 

 non-detachable electrons in the ring. In the case of 

 neon, the non-detachables evidently number lo ; and 

 this suggests a connection with the Rydberg series 

 2(1^+2^+2^+3^ . . .), which evidently has some con- 

 nection with atomic structure. 



It appeared of interest to examine the cases of the 

 atomic weights 2, 3, 5, and 8, which have at present 

 no corresponding elements. With two floating mag- 

 nets, one forms the nucleus and the other the ring 

 group, which gives an atomic number i. This sub- 

 stance would therefore by analogy be an isotope of 

 hydrogen with an atomic weight 2. With three 

 magnets, one again forms the nucleus, with two 

 others in the ring : this corresponds to the atomic 

 number 2, so that the element, if it exists at all, 

 may be an isotope of helium. Five magnets give a 

 system of two in the nucleus and three in the ring — 

 an isotope of Hthium. In the case of eight magnets 

 there are two possible groupings, almost equally 

 stable. Four in the nucleus and four in the ring 

 would correspond to an isotope of beryllium. Five 

 in the nucleus and three outside is the analogue of 

 a fourth Hthium isotope. These two groupings would 

 be isobaric systems. 



It would occupy too much space to discuss the 

 regularities of the magnetic groupings within the 

 nucleus, but one point of interest may be mentioned. 

 In the case of Boron-ii and carbon, both nuclei 

 contain six magnets arranged in two pairs of con- 

 centric triangles. In view of the general resemblance 

 in physical characteristics between boron and carbon, 

 and especially of the fact that in its hydrides boron 

 is quadrivalent like carbon (yielding BgHj like CgHg 

 and not BH3 as might be expected from its position 

 in Group III.), this pecuharity seems not without 

 meaning. 



We propose to extend this investigation imme- 

 diately with improved apparatus which we hope will 

 surmount some of the obvious experimental difficulties 

 in the case of more complex systems. 



J. K. Marsh. 

 A. W. Stewart. 



The Sir Donald Currie Laboratories, 

 Queen's University, Belfast. 



