47< 



NATURE 



[September 12, 1895 



tivity of a substance to march with its thermal conductivity. 

 Again, on this theor)' the resistance clearly increases with 

 increase of distance between atoms — i.e. with increase of 

 temperature. On the contrar)', in electrolytic conduction the 

 same junction of filament ends is brought about, not by oscilla- 

 tions of molecule to molecule, but by disruption of the molecule 

 and passiige of atom to atom In this case conduction is easier the 

 more easily a molecule is split up, and thus resistance decreases 

 with increase of temperature. To explain the laws of electro- 

 lysis it is only necessar)- to assume that the strengths of all 

 filaments are the same. -•V similar hypothesis, as we have seen, 

 lies at the basis of J. J. Thomson's explanation of chemical com- 

 bination, although it is not necessarily the case that we are dealing 

 with the same kind of fil.-iments. It is evident that the theor)- 

 easily lends itself to his views as to the mechanism of the electric 

 discharge through gases. The modus operandi of the production 

 of the mechanical foreive on a conductor carrying a current in a 

 magnetic field and of eleclrodynamic induction is not clear. 

 Prolably the full explanation is io be found in the stresses pro- 

 duced in the ether owing to the deformation of the cells by the 

 passage of the filaments through them. The fluid moves accord- 

 ing to the equation of continuity without slip, and subject to the 

 surface conditiiins at the conductors. This motion, however, 

 distorts the cells, and stresses are called into play. .-Vny theory 

 which can explain the mechanical forcives and also Ohm's law, 

 must, on the principles of the conservation of energ)-, also explain 

 the induction of currents. 



The magnetic rotation of the plane of polarisation of light 

 does not depend on the structure of the ether, or on the magnetic 

 field itself, but is a result of the atomic configuration of the 

 matter in the field modified by the magnetism. It is generally 

 recognised as caused by something in the field rotating round 

 the direction of the magnetic lines of force. Now the vortex 

 atom, as usually pictured, is incapable of exhibiting this ])roperty. 

 It is, however, an interesting fact, and one which I hope to 

 demonstrate to this Section during the meeting, that a vortex 

 ring can have two simultaneous and independent cyclic motions 

 — one the ordinary- one, and another which is capable of pro- 

 ducing just the action on light which shows itself as a rotation of 

 the plane of polarisation. The motion is rather a complicated [ 

 one to describe without a diagram, but an idea of its nature may 

 be obtained by considering the case of a straight cylindrical 

 vortex. The ordinary straight vortex consists, .as every one 

 knows, of a cylinder of fluid revolving like a solid, .-ind sur- 

 rounded by a fluid in irrotational motion. In the core the 

 velocity increases from zero at the axis to a maximum at its sur- 

 face. Thence it continuously decreases in the outer fluid as the 

 distance increases. Ever)'where the motion is in a plane 

 perpendicular to the axis. Let us now consider a quite different 

 kind of vortical motion. Suppose the fluid is flowing along the 

 core like a \-i.scous fluid through a pii>e ; the velocity is zero at the 

 surface and a maximum at the axis. Everywhere it is par.-illel to 

 the axis, the vortex lines are circles in planes perpendicular to 

 the axis, and concentric with it. Since the velocity at the surface 

 of the core is zero, the surrounding fluid is also at rest. Now 

 superpose this motion on the previous one, and it will be found 

 to Ik steady. If a short length of this vortex be supposed cut 

 off, bent into the shajM; of a circle and the ends joined, we shall 

 have very a rough idea of the c.mipound vortex ring of which I 

 speak. I say a very rough idea, Iiecause the actual state of motion 

 in a ring vortex or a Hill's vortex is not quite so simple as the 

 analogy might lead one to think. 



Now a compound vortex atom of this kind is just what we 

 want to pr<xluce rotation of the plane of polarisaticm of light. 

 The light pa-ssing through such a vortex has the direction of 

 %-ibration twisted in the wave front. In ordinar)- matter no such 

 rotation is produced, because the various atoms are indifferently 

 directed, and Ihey neutralise each other's effects. Let, however, 

 a m.agnetic field lie produced, and they will range themselves 

 .wi that, on the aver.agc, the primary' circulations through the 

 apertures will (loinl in the direction of the fickl. Consequently 

 the average ilireclion rif the seconilary spin will \k in planes 

 |)cr|>inilir ular to this, anil will rotate the plane of |Milari.«ition of 

 any lit;tit whose wave front |)asses them. The rotation is pro- 

 duced iinly on the light which is transmitted Ihronsh the vortex. 

 The rotation •.liMTvcil i> a resultant effect. In fact it is clear 

 that in the case of rL-fr.i<-iiim the optical media belong to the type 

 in which every (Mjrlinn transmits the light, and not to the ty|)e in 



< '* Primary " reftr* in itir mniion m usually uniier^lood ; " Mrcondar>* " to 

 ihc aupcTpmcd, a* rt[>lainrfl A\*t\c. 



which refraction is produced by opaque bodies embedded in the 

 ether. The atoms are only opaque if they contain vacuous 

 cores. The question of the grip of the particles on the ether 

 does not enter, but difference of quality — showing itself ii> 

 refraction and dispersion — is due to difference in average 

 rotational quasi-elasticity produced by the atomic circulations, 

 and possibly absorption is due to precessional and nutational 

 motion set up by the secondary spins. These, however, are 

 perhaps rather vague speculations. 



Instead of attempting to invent ethers, to deduce their pro- 

 perties from their specifications, and then seeing whether they 

 fit in with experience, we may begin half-way. We may assume 

 diflTerent forms for the function giving the energy of the medium 

 when disturbed, apply general dynamical methods, and dis- 

 tinguish between those which are capable of exi>laining the 

 phenomena we are investigating and those which are not. 

 Invention is then called upon to devise a medium for which the 

 desired energy-function is appropriate. This was the method 

 applied by MacCull.agh to the luminiferous ether. He obtained! 

 an algebraical form of the energy function which completely 

 satisfied the conditions for a luminiferous ether : its essential pro- 

 perty being that the energ)' depended only on the rotational dis- 

 placements of its small parts. He was unable, however, to 

 picture a stable material medium which would possess this 

 property. We recognise now that such a medium is possible if 

 the rotational rigidity is produced by intrinsic motions in the 

 small jiarts of the medium of a gyrostatic nature. In a most 

 masterly manner Lannor(" .\ Dynamical Theory of the Electric 

 and Luminiferous Medium, " /'/(//. Trans., 1S94) has recently 

 investigated by general dynamical methods the possibility of 

 explaining electric and magnetic phenomena by means of the 

 same energ)' function. Electric lines of force are rotational 

 filaments in the ether,' similar in fact to those I suggested at 

 Bath, while a magnetic field consists of a flow of the ether. 

 The same dilViculty in accounting for electro-dynamic induction 

 arises, but the general form of the equations for the electro- 

 dynamic and magnetic fields are the same as those generally 

 received. 



Towards the end of this paper he is led to jwstulatc a theory 

 ' of electrons whose convection through the ether constitutes an 

 electric current. Two rotating round each other are supposed to 

 produce the same effect as a vortex ring. The mass of ordinary 

 I matter is attributed to the electric inertia of these electrons. The 

 electron itself is a centre or nucleus of rotational strain. If I 

 express a doubt as to the possibility of the existence of these 

 nuclei as specified, I do so with great diffidence.-' Whether they 

 I can or cannot exist, however, the general results of the investi- 

 gation are not afiected. 



Since this paper was published Larmor has read a second one 



' on the same subject before the Royal Society, developing further 



} his theory of the electron. The pulilieation of this will be 



1 awaited with interest. It is impossible in an address such as 



' this to go .ffr/(i//«/ into the numerous points which he takes up- 



and illuminates, because the malhematical treatment of the 



general question docs not lend itself ea,sily to oral exposition even 



to an audience composed of i')rofessed mathematicians. There is 



no doubt but that this paper has put the theory of a rotationally 



elastic ether — and with it that i>f a fluid vortex ether — on a 



sounder b,asis, and will lead to its discussion and elucidation by 



a wider circle of investigators. 



f)ne further cl.ass of physical phenomena yet remains, viz. 

 those of gravitation. The ether must be capable of transmitting 

 gravitational forces as well as electric and optical effects. Does 

 the rotational ether give any promise of domg this ? No satis- 

 factory explanation of gravitation on any theory has yet been 

 offered. Perhaps the least uns.atisfactory is that depending on 

 the vortex atom theory of matter (" On the Problem of Two 

 PuRiting Spheres in a Eluid," Proc. Canih. J'/iii. S<v., iii. p. 

 283), which attributes it to pulsations of hollow vortex atoms. 

 Hut this necessitates that Ihey .shouUI all pulsate with the same 

 jx-riod and in the same phase. It is very difficult to conceive how 

 this can happen, unless, as I^armor suggests, all matter is built 



I Tfic necewily tlinl ifir fil.imcni'i shall be in p.liM doe* not seem to he 

 rccoKniwd. Thi^ i«, however, essential Moreover, if the complementary 

 circulations of the rilamcitis liclwcen (say) a plate condenser be placed other- 

 where than in the same rejiion, the lilaments l>elweei) the plates must rotate 

 as a whole ; that is, an electric field would always be combined with a 

 magnetic one. . 



•J It would appear that the same results would flow il two particles 

 oppositely electrilied— 1>. joined by two complementary fib-uncnts, as already 

 described— were to rotate round each other. 



NO. 1350, VOL, 52] 



