]\gt ; * The Effect of Junctions on the of Blectric Waves along Conductors .
By LORD RAYLEIGH , O.M. , F.R.S. ( Received December 2 , 1912 , \mdash ; Read January 16 , 1913 .
) Some interesting problems in electric wave propagation are suggested by an experiment of Hertz .
* In its original form waves of the simplest kind travel in the positive direction ( fig. 1 ) , outside an infinitely thin conducting cylindrical shell , AA , which comes to an end , say , at the plane Co-axial with the cylinder a rod or wire BB ( of less diameter ) extends to infinity in both directions .
The conductors being supposed perfect , it is required to determine the waves propagated onwards beyond the cylinder on the positive side of , as well as those reflected back outside the cylinder and in the annular space between the cylinder and the rod .
: FIG. 1 .
So stated , the problem , even if mathematically definite , is probably intractable ; but if we modify it by introducing an external co-axial conducting sheath CC ( fig. 2 ) , extending to infinity in both directions , and if we further suppose that the diameter of this sheath is small in comparison with the wave-length ( ) of the vibrations , we shall bring it within the scope of approximate methods .
It is under this limitation that I propose here to consider the present and a few analogous problems .
Some considerations of a more general character are prefixed .
If , be components of electromotive intensity , , those of magnetisation , Maxwell 's general circuital for the dielectric give , ( 1 ) " " Ueber die Fortleitung electrischer Wellen durch Drffite 'Wied .
Ann 1889 , , p. 395 .
' Phil. Trans 1868 ; 'Maxwell 's Scientific Papers , ' vol. 2 , p. 128 .
and two similar equations , and , also wimilar equations , eing telocity oropagation .
Froml 0Lord Eayleigh.unctions o ( 1 ) and ( 2 ) we may derive ; and , further , that , ( 4 ) where .
( 5 ) At any point upon the surface of a conductor , regarded as perfect , the condition to be satisfied is that the vector be there normal .
In what follows we shall have to deal only with simple vibrations in which all the quantities are proportional to , so that may be replaced by It may be convenient to commence with some cases where the waves are in two dimensions only , supposing that , vanish , while , are independent of .
From ( 1 ) and ( 2 ) we have At the surface of a conductorP : : : , , are proportional to the direction cosines of the normal ; so that the surface condition may be expressed simply by , ( 6 ) which , with suffices to determine .
In ( 7 ) .
It will be seen that equations ( 6 ) , ( 7 ) are identical with those which apply in two dimensions to aerial vibrations executed in bounded by walls , then denoting velocity potential .
When is known , the remaining functions follow at once .
It may be remarked by the way that the above analogy throws light upon the question under what circumstances electric waves are guided by conductors .
Some high authorities , it would seem , regard such guidance as uing in all cases as a consequence of the boundary condition fixing the direction of the electric force .
But in Acoustics , though a similar condition holds good , there is no guidance of rial waves round convex surfaces , and it follows that there is none in the two-dimensional electric vibrations under consideration .
Near the concave surface of walls there is in both cases 1912 .
] Propagation of Electric Waves along Conductors .
a whispering gallery effect .
* The peculiar guidance of electric waves by wires depends upon the conductor being encircled by the magnetic force .
No such circulation , for example , could ensue from the incidence of plane waves upon a wire which lies entirely in the plane containing the direction of propagation and that of the magnetic force .
Our first special application is to the extreme form of Hertz 's problem ( as modified ) which occurs when all the radii of the cylindrical surfaces concerned become infinite , while the differences CA , AB remain finite and indeed small in comparison with .
In fig. 2 , then represent ..\mdash ; - . . .
FIG. 2 .
planes perpendicular to the plane of the paper and the problem is in two dimensions .
The two halves , corresponding to plus and minus values of are isolated , and we need only consider one of them .
Availing ourselves of the analogy , we may at once transfer the solution given ( after Poisson ) in ' Theory of Sound , ' S264 .
If the incident wave in CA be represented by and that therein reflected by , while the waves propagated along , AB be denoted by , we have ( 8 ) AB and .
( 9 ) The wave in AB is to be regarded as propagated onwards round the corner at A rather than as reflected .
As was to be anticipated , the reflected wave is smaller , the smaller is AB .
It be understood that the validity of these results depends upon the assumption that the region round A through which the waves are irregular has dimensions which are negligible in comparison with 'Phil .
Mag 1910 , vol. 20 , p. 1001 ; ' Scientific Papers , ' vol. 6 , p. 617 .
Lord Rayleigh.unctions oimpler example iketched i where f : FIG. 3 .
various lines represent planes or cylindrical surfaces perpendicular to the paper .
One bounding plane is unbroken .
The other boundary consists mainly of two planes with a transition at AB , which , however , may be of any form so long as it is effected within a distance much less than .
With a notation similar to that used before , may denote the incident positive wave and the reflected wave , while that propagated onwards in is .
We obtain in like ma1mer , ( 10 ) When AB vanishes we have , of course , .
A little later we shall consider the problem of fig. 3 when the various surfaces are of revolution round the axis of Leaving the two-dimensional examples , we find that the same general method is applicable , always under the condition that the region occupied by irregular waves has dimensions which are small in comparison with Within this region a simplified form of the general equations avails , and thus the difficulty is turned .
An increase in means a decrease in .
When this goes far enough , it justifies the omission of in equations ( 1 ) , ( 2 ) , ( 3 ) ) .
Thus become the derivatives of a simple potential function , which itself satisfies ; that is , the electric forces obey the laws of electrostatics .
Similarly are derivatives of another function satisfying the same equation .
The only difference is that may be multivalued .
The magnetism is that due to steady electric currents .
If several wires meet in a poin , the total current is zero .
This expresses itself in terms 1912 .
] Propagation of Electric Waves along Conductors .
107 : 3 virtue oanother bonditions which.argins oegion iorming tutions which aarts adistance oation between tations.\ldquo ; ethod tonsistsides fregion oarity , ccommodating t ection)onsidered iformer pndent vressi tectric conditionsi oethod iemconductors aution.round zthough titation iication t ; ( 12 ) also and vanish .
In the present case we have for the negative side , where there is both a direct and a reflected wave , , ( 13 ) where is the distance of any point from the axis of symmetry , and are arbitrary constants .
Corresponding to ( 13 ) , .
( 14 ) In the region of regular waves on the positive side there is supposed to be no wave propagated in the negative direction .
Here accordingly , ( 15 ) V , ( 16 ) being another constant .
We have now to determine the relations between the constants , hitherto arbitrary , in terms of the remaining data .
For this purpose consider cross-sections on the two sides both near the origin and yet within the regions of regular waves .
The electric force as expressed in ( 13 ) , ( 15 ) is purely radial .
On the positive side its integral between the radius of the inner and that of the outer conductor is , with omission of having the value proper to the section .
On the negative side the corresponding integral is ' Phil. Mag 1897 , vol. 44 , p. 199 ; 'Scientific Papem , ' vol. 4 , p. 327 .
dentified wnity Airst ration isrecognise these turther theconsider tntermediate rwhere ebeing tadius onner conductor aLord Reigh . .
( 17 ) In like manner the magnetic force in ( 14 ) , ( 16 ) is purely circumferential .
And the circulations at the two sections are as and H2 .
But since these circulations , representing electric currents which may be treated as steady , are equal , we have as the second relation\mdash ; .
( 18 ) The two relations ( 17 ) , ( 18 ) give the wave propagated onwards and that reflected in terms of the incident wave .
If , we have of course , If we suppose all great and nearly equal and expand the logarithms , we fall back on the solution for the two-dimensional case already given .
In the above the radius of the outer sheath is supposed uniform throughout .
If in the neighbourhood of the origin the radius of the sheath changes from to , while ( as before ) that of the inner conductor changes from to , we have instead of ( 17 ) , , ( 19 ) while ( 18 ) remains undisturbed .
In the logarithmic functions are proportional to the reciprocals of the electric capacities of the system on the two sides , reckoned in each case per unit of length .
From the general theory given in the paper referred to we may that this substitution suffices to liberate us from the restriction to symmetry round the axis hitherto .
The more general functions which then replace on the two sides must be chosen with such coefficients as make the circulations of magnetic force equal .
The generalisation here indicated applies equally in the other problems of this paper .
In Hertz 's problem , fig. 2 , .
the method is similar .
In the region of regular waves on the left in CA we may retain ( 13 ) , ( 14 ) , and for the regular waves on the right in CB we retain ( 15 ) , ( 16 ) .
But now in addition for the regular waves on the left in AB , we have , ( 20 ) : : 1912 .
] Propagation of Electric Waves along Conductors .
109 Three conditions are now required to determine in terms of We shall denote the radii taken in order , viz. , , SAA , , by respectively .
As in ( 17 ) , the electric forces give .
( 22 ) The magnetic forces yield two equations , which may be regarded as expressing that the currents are the same on the two sides along BB , and that , since the section is at a negligible distance from the insulated end , there is no current in AA .
Thus .
( 23 ) From ( 22 ) and ( 23 ) , ( 24 ) .
( 25 ) If exceeds but little , tends to vanish , while and approach unity .
Again , if the radii are all great , ( 24 ) , ( 25 ) reduce to , ( 26 ) as already found in ( 8 ) , ( 9 ) .
The same method applies with but little variation to the more general problem where waves between one wire and sheath divide so as to pass along several wires and sheaths ( etc. , always under the condition that the whole region of irregularity is negligible in comparison with the wave-length .
* The various wires and sheaths are , of course , supposed to be continuous .
With a similar notation the direct and reflected waves along the first wire are denoted by , and those propagated onwards along the second , third , and other wires by , etc. The equations are\mdash ; , ( 27 ) It is hardly necessary to detail obvious particular cases .
The success of the method used in these problems depends upon the assumption of a great wave-length .
This , of course , constitutes a limitation , but it has the advantage of eliminating the irregular motion at the junctions .
* This condition will usually suffice .
But extreme cases may be proposed where , in spite of the smallness of the intermediate region , its shape is such as to entail natural resonances of frequency agreeing with that of the principal waves .
The method would then fail .
110 Hon. R. J. Strutt .
Duration of Luminosity of [ Dec. 19 , In the two-dimensional examples it might be possible to pursue the approximation by determining the character of the irregular waves , ab least to a certain extent , somewhat as in the question of the correction for the open end of an organ pipe .
Duration of Luminosity of Electric Discharge in Gases Vapours .
By the Hon. R. J. STRUTT , F.R.S. , Professor of Physics , Imperial College of Science , South Kensington .
( Received December 19 , 1912 , \mdash ; Read January 16 , 1913 .
) S1 .
Introduction .
electric discharge produces luminosity in any gas or vapour through which it passes .
The question presents itself , Does the luminosity persist after the current has ceased , or does it stop immediately ?
A full answer is likely to be of great importance in unravelling the cause and mechanism of the luminosity .
There are existing observations bearing on the subject , but these are somewhat scattered in the literature , and , so far as I am aware , their mutual relations have not been pointed out .
It is hoped in this paper to do something towards systematising and extending them .
The most conspicuous phenomena in this connection are the various forms of afterglow which have been discussed in previous papers .
* But these are not really relevant to the present subject , for they are due to secondary causes of a chemical nature .
Some of them , produced in gaseous mixtures containing oxygen , are due to the interaction of ozone with other substances present .
Others , again , are connected with the formation of an active modification of nitrogen .
In none of these cases can after-luminosity be considered continuous with the luminosity of the discharge which produced it .
For it is always much less brilliant , even at first , and always has a quite different spectrum .
Several experimenters have found , by examining the leyden spark with a rotating mirror , that its luminosity may persist much longer than the 'Phys .
Soc. Proc 1910 , vol. 23 , pp. 66 , 147 ; 1911 , vol. 24 , p. 1 ; 'Roy .
Soc. Proc 1911 , , vol. 85 , pp. 219 , 377 , 633 ; 1912 , vol. 86 , pp. , 106 , 262 , ; 1912 , vol. 87 , pp. 179 , 302 .

