]\gt ; Discussion on the Theory of tivity .
( February 5 , 1920 .
) J. H. JEANS : During the last century , two great dominating principles of physics emerged\mdash ; the Conservation of Energy and the Second Law of Thermodynamics .
The present century has already added a third member to this list , the principle of Relativity , which we are to discuss to-day .
The three pl'inciples have in common that they do not explain how or why events happen ; they merely limit the types of events which can happen .
Thus the principle of Conservation of shows that water cannot flow uphill ; the Second Law of Thermodynamics shows that heat cannot flow from a cold body to a hot ; the principle of elativity shows that a planet cannot describe a perfect ellipse about the sun as focus .
But it would be as unreasonable to expect the principle of Relativity to explain why a planet describes an orbit or how a ray of light is propagated as it would to propound the same questions to the principle of Conservation of Energy or the Second Law of Thermodynamics .
All three principles deal with events , and not with the mechanism of events .
The main interest of the new theory , however , is not merely that it discloses a new universal principle ; it is rather that it discloses a new universe .
Our former belief that the foundations of science had been laid for all time has been shattered ; we now find that the land on which we had built was largely a mirage .
New and mysterious continents appear for science to explore , but it is not for the theory of Belativity to explore them .
The methods of that theory are destructive rather than constructive , when the theory predicts a positive result , it is invariably for the same reason , namely , that a process of exhaustion shows that any other result would be impossible .
Historically , the theory had its origin in the famous experiment of Michelson and Morley .
Light sent from a source to a mirror waa reflected back to , and it was found that the average to-and-fro velocity was the same for all directions of OP in space .
To those who believed in a fixed ether through which light propagation took place , the result was surprising , for , on the ether-hypothesis , the average to-and-fro velocity would depend on the orientation of OP relative to the earth 's motion through the ether . .
As Ritz and others have pointed out , the result is intelligible enough if we adopt an emissive or corpuscular theory of light , but such a theory is difficult .
to reconcile with the majority of electromagnetic phenomena , and seems Discussion on the Theory of Relativity .
be definitely negatived by recent experiments of Majorana .
Majorana finds that the velocity 5f light is not altered when it is reflected by a moving mirror , and neither is it altered when emitted by a moving source .
Thus we have almost conclusive proof of the truth of a hypothesis , first put forward by Einstein in 1905 , which has formed the foundation of the whole theory of reIativity .
The hypothesis is that light from any source appears to any observer to travel always with the same velocity , or , more concisely , the wave-surface is a sphere surrounding the observer .
There is a much better justification for this hypothesis than is to be found in the experiments of Michelson-Morley and bIajorana ; it is that the hypothesis explains a vast number of physical phenomena , and has , so far , not been foumd inconsistent with a single one .
Suppose that two observers , , move with different velocities , but happen to coincide at the origin at an instant , at which a light signalis sent out from the origin .
A wave-surface will be formed , and each observer will assert that it is spherical , and that he is at its centre .
Since a sphere cannot have two different centres at once , we can only conclude that the two qervers mean different things by their statements , or , in other words , they are using different co-ordinates .
The -Lorentz contraction hypothesis had prepared the ground for this conception .
If appears both to and to travel in straight lines with uniform velocity , then the co-ordinates used by and used by must be connected by a linear transformation , and must transform into .
It follows that , in a four-dimensional continuum having and as rectangular co-ordinates , the axes of must be derived from those of by a pure rigid-body rotation .
Corresponding to a in the plane , the equations of transforma- tion are readily seen to be , where and is the relative velocity of the two observers .
These are precisely the equations of the famous Lorentztransformation .
Before the appearance of Einstein 's theory , Lorentz had already discovered these equations as the condition that all optical and electromagnetic phenomena should be the same for all observers , regardless of their relative velocities .
These equations explain and predict a great numbel of physical phenomena , , the variation of mass with velocity , electromagnetic forces in ponderable media , Fizeau 's water-tube experiment .
Nothing in physics has yet been found to be inconsistent with them .
The theory of Relativity insists upon the truth of the equations , at any Discussion on the rate in the ) sense of a gravitational field , but it tells us nothing as to their physical interpretation .
The earliest attempt at an explanation was the erald contraction hypothesis , but its inadequacy is easily discovered .
Following Ehrenfest , let us imagine a cart wheel revolving about an axle at rest in the ether which is postulated by the contraction theory .
Each element of the rim moves with the same velocity , and so is contracted in a uniform ratio , but the spokes , moving only at right to their length , undergo no contraction at all .
If the contraction-theory were true , a rotating wheel of radius would have a rim of length The apparent paradox resolved when we think in terms of ths fourdimensional continuum , and only then .
A point in the continuum is an event To each observer separately an " " event\ldquo ; is a point at an instant ( t ) , but we must remember that different observers divide the continuum into space and time in different ways .
To one observer , the interval between two events is so much time and so much space , but another observer estimates both the time and space differently .
The interval between a man 's birth and death may be estimated by as miles and 75 years , but may call it millions of miles and 76 years .
The only final objective measurement is that of a single range in the continuum ; its square is one million square miles minus the square of 75 light-years .
The contraction hypothesis expressed in terms of the continuum imme- diately becomes consistent and complete .
What calls a length , is resolved by into or along axis of length , and along 's axis of , representing a time interval .
The contraction in length is compensated by an expansion in time .
The successive positions of a particle or a ray of light will be represented by a continuous series of points in the four-dimensional continuum , what Minkowski calls the " " world-line\ldquo ; of the particle or ray .
The analytical equation of this " " world-line\ldquo ; is an integral of the equations of motion of the particle or light-ray .
The Newtonian law of motion is that the world-line of a particle acted on by no forces is a straight line .
The equation of this straight line may be put in the invariant form , this being of course invariant as regards all rotations of the axes and so being the same for all observers .
The general relativity-condition is that all the laws of nature shall be invariant as regards diff'erent sets of orthogonal axes in the continuum .
If any single law of nature shows a special connection with any special set of axes , we shall have an objective means of identifying one special set of axes , and this leads at once to a possibility of determining absolute velocity in Theory of Relativity .
space .
As a matter of experiment , the laws of light and electromagnetism are found to conform to the relativity-condition , with the consequence that no absolute velocity can be revealed by optical or electromagnetic means .
But it is easily seen that the Newtonian law of gravitation does not conform to this condition , so that if the Newtonian law were true , it would be possible to determine absolute motion in space .
Sir O. Lodge , assuming the Newtonian Law to be true , determined what velocity of the Sun in space would be required to account for the observed motion of the planet Mercury , but Prof. Eddington showed that quite different velocities would be determined from the motions of the other planets .
Einstein 's first theory of gravitation proceeded by simply modifying the Newtonian law so that it conformed to the condition of invariance , but this modified law predicted a motion of the perihelion of Mercury equal to only one-sixth of the amount observed , and a gravitational deflection of light equal only to about half of that discovered at the recent eclipse .
Einstein 's second theory is of a much more radical nature .
He does not merely modify the expression of the laws of nature in the four-dimensional continuum ; he modifies the continuum itself .
The continuum is supposed to be affected with twists and kinks centering round those values of and which represent positions of matter at different instants .
Gravitation is no longer a physical force ; it is merely our interpretation of the fact that the continuum is not Euclidean ; parallel lines meet or diverge , so that particles projected along parallel lines may in time collide with one another .
Their collision is a consequence of the special metrical properties of the space , but we have up to now ascribed it to a supposed attractive force which we have called gravitation .
According to Einstein 's theory it is no more accurate to say that the earth a fallin body than to say that the kets of an uneven billiard table repel the balls .
A somewhat imperfect will perhaps explain the fundamental conceptions of Einstein 's theory .
Let us imagine a race of cave-dwellers shut away for ever from the earth 's surface , and therefore having no reason to suspect either the roundness or rotation of the earth .
To them , gravitation would appear a simple matter ; they would notice that bodies fell to the floor of their cave with a constant uniform acceleration , and they might ascribe this to a " " force\ldquo ; of gravity , acting always in one direction , which they would call the vertical .
To the cave-dwellers a system of mechanical laws identical with our Newtonian laws would seem quite consistent and plausible .
The expression ' at rest\ldquo ; would of course mean " " keeping the same distance from the walls of the cave and it would probably never occur to them that any more precise analysis or definition was needed .
Discussion on the Some day a cave-dwelling Foucault might try the experiment of suspending a weight from the roof of the cave and examining what orbit it would describe , when projected in any way .
He would expect it to be an ellipse indefinitely repeated , but would find that it was an ellipse which rotated slowly in its own plane .
The cave-dwellers might further mark out straight lines which satisfied all geometrical tests perfectly , but ( as a consequence of the earth 's rotation ) , refined experiments would show that light did not follow these straight lines .
The cave-dwellers would now have two reasons for doubting their simple laws .
Light would appear to be deflected from straight lines , and orbits which ought to be recurrent ellipses would be found to be revolving ellipses .
In the results of the recent Eclipse expedition and in the observed motion of Mercury , we have precisely similar reasons for doubting our simple laws , and the } ) tion of their failure which is suggested to us by Einstein is precisely of the same nature as that which might be suggested to the cavedwellers by a odyte scientist .
We have tacitly assumed fixed axes where nothing is fixed : we have formed wrong ideas of the nature gravitation : and our definition of a straight line is interwoven with the ideas of an untrue system of geometry .
We are familiar with the conception of the gravity-acceleration on the earth 's surface as being due partly to centrifugal force and partly to the earth 's attraction .
Our level surfaces are equipotentials under a potential and we were taught in our youth that the first term represents real forces and the second term only fictitious forces .
The nition by the cavedwellers that the second term represented only fictitious forces would remove their troubles with the Foucault experiment .
The suggestion of generalised relativity is that the further niCion by us that both terms represent fictitious forces will remove the troubles which have so far centred round our Newtonian law of gravitation .
In the four-dimensional continuum already considered , the interval between two adjacent events is ( )2 , ( A ) this being of course invariant for all observers .
According to Einstein 's generalised theory this expression is only true at an infinite distance from all gravitating matter .
In the neighbourhood of matter it must be replaced by ( )2 , ( B ) in which the coefficients , specify the gravitational field .
Theory of Relativity .
To see the meaning of this , imagine that our cave-dwellers had suspected their cave to be in rotation and so had transformed expression ( A ) to rotating axes .
It would have to be replaced by Their old laws of motion , to which particles and light rays described straight lines with uniform velocity , could be put in the form with given by ( A ) .
They would find that the same equation with given by ( C ) would predict exactly the observed deflection of their light rays and would enable them to deduce the orbit of their Foucault experiment .
With other values for , etc. , in ( B ) Einstein shows that the same law of motion will remove our difficulties about the perihelion of Mercury and the dellection of At first sight it might be thought that with so many adjustable constants , , etc. , this would be a foregone conclusion .
But the values of , etc. , which conform to the relativity condition are practically limited to a single set , in which is iven by ( )2 , ( D ) where stands for .
There are in reality no adjustable constants , and it was this circumstance which enabled Einstein to predict without any uncertainty the deflection of light to be expected .
We notice that in both ( C ) and ( D ) the coefficient of is ( potential of apparent field force ) , and when the coefficient of is fixed , the others follow almost uniquely from the conditions of invariance .
A particle which is not acted upon by netic forces of any kind will have a world-line which is a geodesic ( defined by ) in the twisted space defined by ( .
This world-line is the straightest line ( if I may be permitted this use of words ) which can be drawn in the space in question , but it will not be straight in the Euclidean sense .
An observer has been taught to think in terms of the conceptions of Enclid and Newton will expect the world-line to be a Euclidean straight line , and will ascribe its deviations from his ideas of straightness to the action of gravitational force .
Mathematicians are accustomed to be asked why they cannot express the new law of gravitational force in simple form .
The answer is easy : there is no force , and therefore no law of force to be expressed .
The laws of can be expressed in the simplest conceivable form , namely What , however , is meant by ?
The simplest interpretation is to regard it merely as a conventional algebraic symbol defined by ( D ) .
Then the law Discussion on the of motion predicts Mercury 's orbit and light-deflection , but does not , I think , make any prediction about shift in the Fraunhofer fines .
Einstein gives to a special interpretation in terms of time and space ; for him is a lineelement in a distorted space-time continuum .
The hypothesis from which this special interpretation is derived requires inevitably , Einstein considers , shift of the Fraunhofer lines towards the red , and I do not think we can dispute that he is right .
It also requires us to believe in an objective curved four-dimensional space , and grave difficulties are disclosed by a consideration of the limits of this space .
Indeed , the difficulties are so grave that , as Prof. Eddington has pointed out , Einstein is prepared to abandon the whole central principle of relativity in an attempt to remove them .
For myself , I have never felt convinced that any sufficient reason compeIs us to adopt Einstein 's interpretation of .
Fermat 's law giving the path of a ray of light in an anisotropic refracting medium can be put in the form , where ( ds)2 but we should be led into terrible difficulties if we tried to interpret as a line element in a twisted three-dimensional space .
I will hasten to add that I am aware that the cases are not parallel ; the most I suggest is that the analogy of Fermat 's law may justify us in hesitating before accepting Einstein 's physical interpretation of .
If we do not insist on Einstein 's interpretation of , the theory becomes less rich and less inclusive , but at the .
same time less beset by difliculties .
The reality of the four-dimensional continuum is , I think , beyond dispute , but the reality of the twists and kinks in it do not appear to be .
Decisive knowledge as to the shift of the Fraunhofer lines would go far towards settling this question .
Prof. A. S. EDDINGTON : The sun 's gravitational field has been summed up in the rather mysterious expression I would like first to explain in a general way how the ordinary Newtonian attraction of the sun is here represented .
is the distance from the sun and the time ( both being reckoned rather conventionally , so far as exact definition is concerned ) .
If we want to make a graph of the motion of a particle , we take and as horizontal and vertical co-ordinates .
I call such a representation a picture .
I want to distinguish between a picture and a map .
If we are given the latitudes and itudes of places on the earth , we can make a picture by taking latitude and longitude as vertical and horizontal co-ordinates ; but of that does not give us a map .
In any ordinary map of Europe the lines of longitude run obliquely , and the lines of latitude are curved .
Why is this Because the map aims at representing all the distances of places in their proper proportions .
Distance is the important thing , which we want to represent correctly .
Now , in the space-time round the sun , the interval is the important \mdash ; the thing with an absolute nificance in nature .
We want to make a map , so that all the intervals are represented in their proper proportions ; and the above formula gives us the rules for drawing the map .
If I had time , I could show you in detail how to derive the shape of the -lines and -lines on the map , which may be compared with the longitude and latitude lines on a terrestrial map .
Now , the law of motion , whether gravitation is acting or not , is extremely simple\mdash ; every particle describes a straight line on the map .
Lt us draw such a straight line .
Now go back to the picture ; we distort the whole figure till the -lines and -lines become vertical and horizontal , and the graph of the particle 's track is then distorted into a concave to the .
We say the particle is attracted .
This curving comes from the difference of the map and the picture , and it is the same phenomenon that 'is usually described as the force of the sun 's gravitation .
There are two methods of exploring space and time .
One is with a scale and clock ; the other with moving particles and light-pulses .
In each way we can find out the geometry of space-time .
It is conceivable that the two methods might give different geometries , each true in its own sphere .
If we had to make a choice , I think we should prefer the geometry of moving particles and light-pulses , because these are very simple things , whereas scales and clocks even of the simplest form have elaborate structure ; another reason is that , in practice , moving particles and light-pulses a very much more sensitive method .
The astronomer uses this method ; by his moving particles , the planets , and especially the planet Mercury , and by the light-pulses from stars during the solar eclipse , he has established that the eometry is summed up in the expression for which I have already written .
If you like to reject Einstein 's theory altogether , and disagree with every idea in Mr. Jeans ' exposition , that will not alter the fact that the formula is true experimentally to a certain high degree of approximation for the geometry of moving particles and -pulses .
Is the geometry of scales and clocks the same ?
There are certain theoretical reasons for thinking it is\mdash ; at least , to a degree of approximation sufficient for our purposes .
But scales and clocks are so insensitive that only one rather crude test seems to be possible .
That is to have a clock on the sun and a similar clock on the earth , and to compare them .
Discussion on the If the clocks are similar , they will each beat out similar values of the absolute , the interval ; what they beat in-time depends on how you measure time , because time is not absolute .
If we measure time in the usual way for the solar system by , and if our clocks are at rest so that do not change during the vibration , then by the formula so that .
Thus , if the two clocks measure equal values of , they will not measure equal values of .
The solar clock , for which is small , will have the longer period , In the actual experiment , two atoms are used for the clocks .
We cannot , of course , determine the period of the solar atom directly ; we have to measure the period of the waves emitted by it when they reach the earth .
It turns out , however , that the period is carried to us unaltered\mdash ; a result due to our judicious choice of the co-ordinate which we consider to be the time .
I expect later speakers will deal with the results of this test .
Sir F. W. DYSON : Leverrier found that the perihelion of Mercury advances a century in excess of the amount produced by the action of the planets .
The following hypotheses have been advanced in explanation:\mdash ; ( 1 ) Oblateness of the sun .
( 2 ) An intra-Mercurial planet .
( 3 ) A of planets between Venus and Mercury , or between Mercury and the sun .
( 4 ) Extended matter whose presence is indicated by the Zodiacal light .
( 5 ) Alteration of the law of attraction from to where Taking these in order , ( 1 ) would require.the equatorial diameter to be greater than its Polar diameter .
Comparison with Jupiter , which rotates in 10 hours , while the sun rotates in 25 days , indicates an oblateness of less than one-tenth of this amount .
Further , measures of the sun 's diameter entirely negative this supposition .
( 2 ) An intra-Mercurial planet has been carefully looked for in Eclipse Expeditions .
Although stars down to the 9th magnitude have been photo- graphed , no unknown body has been found .
( 3 ) Such a ring would need to lie nearly in the plane of Meroury , or it Theory of tivity .
would affect the movement of the node as well as the apse .
Mercury is highly inclined to the invariable plane of the planetary system .
It is very unlikely that a ring of small bodies would be in the plane of Mercury , and it is not certain that a ring would be stable .
( 4 ) The Zodiacal light extends to at least from the sun .
It is attributed to the reflection of sunlight from diffused matter extending in the form of a flattened disc beyond the earth 's orbit .
Prof. attempted to explain the movement of the apse of Mercury by the attraction of this matter , assumed that nine-tenths of its mass was concentrated in a flat spheroid inside Mercury 's orbit .
Further , that the equatorial plane of this spheroid was nearly in the plane of Mercury .
This is necessary in order to secure other elements of the orbits of the inner planets , such as the node of Mercury , from being disturbed .
The arbitrariness of the assumption is evident .
( 5 ) The slight alteration of the inverse square law will the required movement to the apse of Mercury , and will not disturb any other planetary elements .
It is a quite arbitrary assumption .
( Further , it should be stated that the assumption is negatived by the movement of the moon 's apse .
Here a shift of 13 would be produced .
E. W. Brown 's discussion shows such excess of motion .
) Prof. A FOWLEIi : It is enerally known that Einstein 's theory predicts a shift of the Fraunhofer lines towards the red by amounts equal to those which would be given by a radial velocity of kilom .
per second .
In order to eliminate displacements which might arise from possible difference of pressure between the sun 's reversing layer and the arc comparison , observers have given special attention to the lines the band of cyanogen 3883 , which is free from pressure effect .
In this region the theoretical displacement is A. Since radial convection currents would produce no displacemenb at the sun 's limb , observations have been made at the limb as well as at the centre of the disc , the sun 's rotation being eliminated by restricting observations to the vicinity of the sun 's poles .
Account has to be taken also of the slight general increase in the wave-lengths of the Fraunhofer lines in passing from centre to limb , which is called the " " limb effect The actual observations may be summarised as follows:\mdash ; Discussion on the Evershed 's first value has been omitted in forming the general mean , and in the case of Schwarzschild the limb displacements have been taken as the means of his values at distances and of the sun 's radius from the centre of the disc .
The numbers in brackets indicate the numbers of lines measured .
If we could accept St. John 's view that the greatest weight should be given to the weaker lines , and that there is practically no displacement either at the edge or at the centre , I think we should conclude , without hesitation , that the predicted Relativity effect does not exist .
It would be too much to suppose that there could be two equal compensating effects , one at the centre and one at the limb , of totally different origin .
If we take the general mean , we have a displacement at the centre of in place of and if the relativity effect , exists we have to account for a compensating effect of to tlJe violet .
This might conceivably be explained by an ascending motion of the solar gases with a corresponding velocity ; but there must be descending as well as ascending gas , and it seems more likely that the descending gas would contribute to the absorption in greater degree than the ascending gas , so that the requisite compensation does not seem probableAt the limb , a compensating displacement of towards the violet would be required ; that is , " " limb effect\ldquo ; opposite in direction to that found for the metallic lines .
If the Einstein effect be supposed absent , a more probable explanation of the mean results can be given .
At the centre we should merely have to suppose that the predominant absorption was given by the descending gas , giving a red displacement of ; and at the limb we should have assume alimb effect of to the red , in accordance with the effect observed Theory of Relativity .
in other cases .
Diflerent interpretations may possibly be put on the observations , but my own opinion is that they are decidedly unfavourable to the existence of the Einstein effect .
Mr. E. CUNNINGHAM : In Einstein 's prevision of an observable shift of the spectrum two distinct hypotheses are involved : the hypothesis of Relativity , and that of Equivalence .
If we take the illustration of two identical atoms , , at a distance apart in a field free from gravitation , the arrival at of the waves from A must synchronize with the vibrations of B. If a gravitational field be simulated by viewing these atoms in a frame of reference in whioh they have an accelerated motion , the principle of elativity requires that this synchronism not be broken , since a coincidence of two events in time and space may never be resolved into non-coincidence .
But the principle of Equivalence has nothing to say in this case .
It only comes into play when we compare the action of an atom in a simulated gravitational field with its behaviour in a real gravitational field which , not being susceptible of being transformed right away , is associated with the presence of matter .
Then , inasmuch as a real gravitational field may be simulated as far as the values of .
and their first differential coefficients are concerned , if the behaviour of the atom depends on these only we may , I think , admit Einstein 's hypothesis that as the measure of the " " interval\ldquo ; between the emission of two successive pulses is the same for a given atom in all fields .
But in the case of the real field it appears that the mode of transmission through the ether will not be such as to disguise the consequent change of period .
Here then appears to be a real difference between a real gravitational field and a simulated field .
The early restricted principle of Relativity was blamed by some physicists for taking away their ether .
Does the generalisation make the situation worse , or does it clear the way for an attempt at reconstruction ?
The criterion for any conception of an objective ether is that the specification of it must be by means of invariant relations .
A definite velocity must be assigned to any element of it which transforms in any change of the spacetime framework in the same way that the velocity of any material element transforms .
It is certain that a revision and completion of the customary specification of stress , momentum , energy , and flow of energy requires a velocity to be attributed to the medium in which these exist , in order that the stress may account exactly not only for the flux of momentum but for the flux of energy as suggested by Pointing .
shows that along these lines an objective moving ether may be conceived which is consistent with the requirements of the Relativity theory .
Discussion on the Prof. H. F. NEWALL : Einstein has said that the merit of theories of principle is their logical perfection and the sec\amp ; ubar ; rity of their foundations .
We are here dealing with a theory of principle .
We have to be very sure of the foundations before we accept it .
When one thinks of Fitzgerald 's guess .
about the celebrated interference experiments , it is difficult to conceive that two different substances , like Michelson and MorJey 's sandstone slab and Morley and Miller 's pinewood planks , should make such a beautiful adjustment of a contraction in length to a compensating expansion in time .
I wish to put a point to Mr. Jeans , relating to the central which he has referred to as difficult to grasp .
Observers and are in stationary ether at the origin when a flash of light is emitted from it ; remains there , moves along a radius .
White diffusers are distributed all over a sphere , of radius say 100 light-seconds , round the origin .
Then after 200 seconds will see all the diffusers flash simultaneously .
My question is : Is it now maintained that will not see the fore diffusers flash first , the lateral diffusers take it up next , and the aft diffusers flash last ?
If we superpose a uniform ether-wind in any direction , we get back to the general statement of the -Morley experiment in terms of an instantaneous .
flash instead of continuous illumination .
I have some remarks to make about refraction\mdash ; a subject which will take me seven minutes to set forth .
Perhaps it will be better to refrain at this .
late hour .
Mr. JEANS : As regards Prof. Newall 's problem , it is not maintained that the would be seen simultaneously by S'\mdash ; he would not expect them to be , because the material surface which appears to be a sphere at to would appear to be quite different to Prof. F. A. LINDEMANN : The assumption that an atom is not changed by being brought into a gravitational field seems doubtful .
Its frequency times some pure number .
increases on account of the radial contraction exactly enough to balance the Einstein effect , unless this increase in its turn is compensated by a change in and .
If so , the spectral shift would be required by the definition of .
If not , its absence , if it really is absent , would be explained .
Whether the fact that Einstein 's theory contains the principles of the conservation of energy and momentum is in its favour , seems doubtful , since they are derived by using Hamilton 's principle which notoriously fails in the ultimate processes of atomic interaction .
Theory of tivity .
Relativity , however , appears to favour the undulatory view of the tion of radiation as opposed to light-quanta .
If these existed they would have a finite mass due to their electric field for an observer moving them .
This mass should vary with the velocity as that of an electron , an become infinite for any other observer .
Incidentally , Relativity deprives us of evidence for the assumption that an electron 's mass is electromagnetic .
Any mass would vary in exactly the same way .
I regret the tendency to emphasize the metaphysical rather than the physical interpretation of Einstein 's equations .
Thus the statement that a sphere of water of radius cm .
would occupy all space appears to me wrong .
That an observer on such a sphere would believe it to occupy all space I agree , but we are not bound to accept his conclusions .
Dr. L. SILBERSTEIN : In view of the late hour but two brief remarks : the attempt at connecting the spectrum-shift question with the quantum theory cannot help very much .
As I have shown at the last meeting of the Royal Astronomical Society , the predicted shift cannot thus be compensated without artificiality .
And as to the effect itself , Einstein himself considers it as theoretically unavoidable , and so do the leading Continental physicists , among whom W. Wien , who agrees with me that it is premature to see the Eclipse results a confirmation of Einstein 's theory .
Secondly , and at the risk of committing what nowadays may seem a heresy , I should like to draw attention to the possibilities offered by Stokes-Planck 's ether which seems to acquire vitality exactly due to the change of light-velocity around the sun observed by the astronomers .
Details are given in my paper in the ' Philosophical Magazine ' for February .
In this connection also a certain crucial first order experiment suggests itself , which I hope to elucidate at a later opportunity .
VOL. XCVII.\mdash ; A.

