1

introduction

the general principles of dynamics, which have, since newton, served as foundation for physical science, and which appeared immovable, are they on the point of being abandoned or at least profoundly modified? this is what many people have been asking themselves for some years. according to them, the discovery of radium has overturned the scientific dogmas we believed the most solid: on the one hand, the impossibility of the transmutation of metals; on the other hand, the fundamental postulates of mechanics.

perhaps one is too hasty in considering these novelties as finally established, and breaking our idols of yesterday; perhaps it would be proper, before taking sides, to await experiments more numerous and more convincing. none the less is it necessary, from to-day, to know the new doctrines and the arguments, already very weighty, upon which they rest.

in few words let us first recall in what those principles consist:

a. the motion of a material point isolated and apart from all exterior force is straight and uniform; this is the principle of inertia: without force no acceleration;

b. the acceleration of a moving point has the same direction as the resultant of all the forces to which it is subjected; it is equal to the quotient of this resultant by a coefficient called mass of the moving point.

the mass of a moving point, so defined, is a constant; it does not depend upon the velocity acquired by this point; it is the same whether the force, being parallel to this velocity, tends only to accelerate or to retard the motion of the point, or whether, on the contrary, being perpendicular to this velocity, it tends to make this motion deviate toward the right, or the left, that is to say to curve the trajectory;

c. all the forces affecting a material point come from the action of other material points; they depend only upon the relative positions and velocities of these different material points.

combining the two principles b and c, we reach the principle of relative motion, in virtue of which the laws of the motion of a system are the same whether we refer this system to fixed axes, or to moving axes animated by a straight and uniform motion of translation, so that it is impossible to distinguish absolute motion from a relative motion with reference to such moving axes;

d. if a material point a acts upon another material point b, the body b reacts upon a, and these two actions are two equal and directly opposite forces. this is the principle of the equality of action and reaction, or, more briefly, the principle of reaction.

astronomic observations and the most ordinary physical phenomena seem to have given of these principles a confirmation complete, constant and very precise. this is true, it is now said, but it is because we have never operated with any but very small velocities; mercury, for example, the fastest of the planets, goes scarcely 100 kilometers a second. would this planet act the same if it went a thousand times faster? we see there is yet no need to worry; whatever may be the progress of automobilism, it will be long before we must give up applying to our machines the classic principles of dynamics.

how then have we come to make actual speeds a thousand times greater than that of mercury, equal, for instance, to a tenth or a third of the velocity of light, or approaching still more closely to that velocity? it is by aid of the cathode rays and the rays from radium.

we know that radium emits three kinds of rays, designated by the three greek letters α, β, γ; in what follows, unless the contrary be expressly stated, it will always be a question of the β rays, which are analogous to the cathode rays.

after the discovery of the cathode rays two theories appeared. crookes attributed the phenomena to a veritable molecular bombardment; hertz, to special undulations of the ether. this was a renewal of the debate which divided physicists a century ago about light; crookes took up the emission theory, abandoned for light; hertz held to the undulatory theory. the facts seem to decide in favor of crookes.

it has been recognized, in the first place, that the cathode rays carry with them a negative electric charge; they are deviated by a magnetic field and by an electric field; and these deviations are precisely such as these same fields would produce upon projectiles animated by a very high velocity and strongly charged with electricity. these two deviations depend upon two quantities: one the velocity, the other the relation of the electric charge of the projectile to its mass; we cannot know the absolute value of this mass, nor that of the charge, but only their relation; in fact, it is clear that if we double at the same time the charge and the mass, without changing the velocity, we shall double the force which tends to deviate the projectile, but, as its mass is also doubled, the acceleration and deviation observable will not be changed. the observation of the two deviations will give us therefore two equations to determine these two unknowns. we find a velocity of from 10,000 to 30,000 kilometers a second; as to the ratio of the charge to the mass, it is very great. we may compare it to the corresponding ratio in regard to the hydrogen ion in electrolysis; we then find that a cathodic projectile carries about a thousand times more electricity than an equal mass of hydrogen would carry in an electrolyte.

to confirm these views, we need a direct measurement of this velocity to compare with the velocity so calculated. old experiments of j. j. thomson had given results more than a hundred times too small; but they were exposed to certain causes of error. the question was taken up again by wiechert in an arrangement where the hertzian oscillations were utilized; results were found agreeing with the theory, at least as to order of magnitude; it would be of great interest to repeat these experiments. however that may be, the theory of undulations appears powerless to account for this complex of facts.

the same calculations made with reference to the β rays of radium have given velocities still greater: 100,000 or 200,000 kilometers or more yet. these velocities greatly surpass all those we know. it is true that light has long been known to go 300,000 kilometers a second; but it is not a carrying of matter, while, if we adopt the emission theory for the cathode rays, there would be material molecules really impelled at the velocities in question, and it is proper to investigate whether the ordinary laws of mechanics are still applicable to them.

2

mass longitudinal and mass transversal

we know that electric currents produce the phenomena of induction, in particular self-induction. when a current increases, there develops an electromotive force of self-induction which tends to oppose the current; on the contrary, when the current decreases, the electromotive force of self-induction tends to maintain the current. the self-induction therefore opposes every variation of the intensity of the current, just as in mechanics the inertia of a body opposes every variation of its velocity.

self-induction is a veritable inertia. everything happens as if the current could not establish itself without putting in motion the surrounding ether and as if the inertia of this ether tended, in consequence, to keep constant the intensity of this current. it would be requisite to overcome this inertia to establish the current, it would be necessary to overcome it again to make the current cease.

a cathode ray, which is a rain of projectiles charged with negative electricity, may be likened to a current; doubtless this current differs, at first sight at least, from the currents of ordinary conduction, where the matter does not move and where the electricity circulates through the matter. this is a current of convection, where the electricity, attached to a material vehicle, is carried along by the motion of this vehicle. but rowland has proved that currents of convection produce the same magnetic effects as currents of conduction; they should produce also the same effects of induction. first, if this were not so, the principle of the conservation of energy would be violated; besides, crémieu and pender have employed a method putting in evidence directly these effects of induction.

if the velocity of a cathode corpuscle varies, the intensity of the corresponding current will likewise vary; and there will develop effects of self-induction which will tend to oppose this variation. these corpuscles should therefore possess a double inertia: first their own proper inertia, and then the apparent inertia, due to self-induction, which produces the same effects. they will therefore have a total apparent mass, composed of their real mass and of a fictitious mass of electromagnetic origin. calculation shows that this fictitious mass varies with the velocity, and that the force of inertia of self-induction is not the same when the velocity of the projectile accelerates or slackens, or when it is deviated; therefore so it is with the force of the total apparent inertia.

the total apparent mass is therefore not the same when the real force applied to the corpuscle is parallel to its velocity and tends to accelerate the motion as when it is perpendicular to this velocity and tends to make the direction vary. it is necessary therefore to distinguish the total longitudinal mass from the total transversal mass. these two total masses depend, moreover, upon the velocity. this follows from the theoretical work of abraham.

in the measurements of which we speak in the preceding section, what is it we determine in measuring the two deviations? it is the velocity on the one hand, and on the other hand the ratio of the charge to the total transversal mass. how, under these conditions, can we make out in this total mass the part of the real mass and that of the fictitious electromagnetic mass? if we had only the cathode rays properly so called, it could not be dreamed of; but happily we have the rays of radium which, as we have seen, are notably swifter. these rays are not all identical and do not behave in the same way under the action of an electric field and a magnetic field. it is found that the electric deviation is a function of the magnetic deviation, and we are able, by receiving on a sensitive plate radium rays which have been subjected to the action of the two fields, to photograph the curve which represents the relation between these two deviations. this is what kaufmann has done, deducing from it the relation between the velocity and the ratio of the charge to the total apparent mass, a ratio we shall call ε.

one might suppose there are several species of rays, each characterized by a fixed velocity, by a fixed charge and by a fixed mass. but this hypothesis is improbable; why, in fact, would all the corpuscles of the same mass take always the same velocity? it is more natural to suppose that the charge as well as the real mass are the same for all the projectiles, and that these differ only by their velocity. if the ratio ε is a function of the velocity, this is not because the real mass varies with this velocity; but, since the fictitious electromagnetic mass depends upon this velocity, the total apparent mass, alone observable, must depend upon it, though the real mass does not depend upon it and may be constant.

the calculations of abraham let us know the law according to which the fictitious mass varies as a function of the velocity; kaufmann’s experiment lets us know the law of variation of the total mass.

the comparison of these two laws will enable us therefore to determine the ratio of the real mass to the total mass.

such is the method kaufmann used to determine this ratio. the result is highly surprising: the real mass is naught.

this has led to conceptions wholly unexpected. what had only been proved for cathode corpuscles was extended to all bodies. what we call mass would be only semblance; all inertia would be of electromagnetic origin. but then mass would no longer be constant, it would augment with the velocity; sensibly constant for velocities up to 1,000 kilometers a second, it then would increase and would become infinite for the velocity of light. the transversal mass would no longer be equal to the longitudinal: they would only be nearly equal if the velocity is not too great. the principle b of mechanics would no longer be true.

3

the canal rays

at the point where we now are, this conclusion might seem premature. can one apply to all matter what has been proved only for such light corpuscles, which are a mere emanation of matter and perhaps not true matter? but before entering upon this question, a word must be said of another sort of rays. i refer to the canal rays, the kanalstrahlen of goldstein.

the cathode, together with the cathode rays charged with negative electricity, emits canal rays charged with positive electricity. in general, these canal rays not being repelled by the cathode, are confined to the immediate neighborhood of this cathode, where they constitute the ‘chamois cushion,’ not very easy to perceive; but, if the cathode is pierced with holes and if it almost completely blocks up the tube, the canal rays spread back of the cathode, in the direction opposite to that of the cathode rays, and it becomes possible to study them. it is thus that it has been possible to show their positive charge and to show that the magnetic and electric deviations still exist, as for the cathode rays, but are much feebler.

radium likewise emits rays analogous to the canal rays, and relatively very absorbable, called α rays.

we can, as for the cathode rays, measure the two deviations and thence deduce the velocity and the ratio ε. the results are less constant than for the cathode rays, but the velocity is less, as well as the ratio ε; the positive corpuscles are less charged than the negative; or if, which is more natural, we suppose the charges equal and of opposite sign, the positive corpuscles are much the larger. these corpuscles, charged the ones positively, the others negatively, have been called electrons.

4

the theory of lorentz

but the electrons do not merely show us their existence in these rays where they are endowed with enormous velocities. we shall see them in very different r?les, and it is they that account for the principal phenomena of optics and electricity. the brilliant synthesis about to be noticed is due to lorentz.

matter is formed solely of electrons carrying enormous charges, and, if it seems to us neutral, this is because the charges of opposite sign of these electrons compensate each other. we may imagine, for example, a sort of solar system formed of a great positive electron, around which gravitate numerous little planets, the negative electrons, attracted by the electricity of opposite name which charges the central electron. the negative charges of these planets would balance the positive charge of this sun, so that the algebraic sum of all these charges would be naught.

all these electrons swim in the ether. the ether is everywhere identically the same, and perturbations in it are propagated according to the same laws as light or the hertzian oscillations in vacuo. there is nothing but electrons and ether. when a luminous wave enters a part of the ether where electrons are numerous, these electrons are put in motion under the influence of the perturbation of the ether, and they then react upon the ether. so would be explained refraction, dispersion, double refraction and absorption. just so, if for any cause an electron be put in motion, it would trouble the ether around it and would give rise to luminous waves, and this would explain the emission of light by incandescent bodies.

in certain bodies, the metals for example, we should have fixed electrons, between which would circulate moving electrons enjoying perfect liberty, save that of going out from the metallic body and breaking the surface which separates it from the exterior void or from the air, or from any other non-metallic body.

these movable electrons behave then, within the metallic body, as do, according to the kinetic theory of gases, the molecules of a gas within the vase where this gas is confined. but, under the influence of a difference of potential, the negative movable electrons would tend to go all to one side, and the positive movable electrons to the other. this is what would produce electric currents, and this is why these bodies would be conductors. on the other hand, the velocities of our electrons would be the greater the higher the temperature, if we accept the assimilation with the kinetic theory of gases. when one of these movable electrons encounters the surface of the metallic body, whose boundary it can not pass, it is reflected like a billiard ball which has hit the cushion, and its velocity undergoes a sudden change of direction. but when an electron changes direction, as we shall see further on, it becomes the source of a luminous wave, and this is why hot metals are incandescent.

in other bodies, the dielectrics and the transparent bodies, the movable electrons enjoy much less freedom. they remain as if attached to fixed electrons which attract them. the farther they go away from them the greater becomes this attraction and tends to pull them back. they therefore can make only small excursions; they can no longer circulate, but only oscillate about their mean position. this is why these bodies would not be conductors; moreover they would most often be transparent, and they would be refractive, since the luminous vibrations would be communicated to the movable electrons, susceptible of oscillation, and thence a perturbation would result.

i can not here give the details of the calculations; i confine myself to saying that this theory accounts for all the known facts, and has predicted new ones, such as the zeeman effect.

5

mechanical consequences

we now may face two hypotheses:

1o the positive electrons have a real mass, much greater than their fictitious electromagnetic mass; the negative electrons alone lack real mass. we might even suppose that apart from electrons of the two signs, there are neutral atoms which have only their real mass. in this case, mechanics is not affected; there is no need of touching its laws; the real mass is constant; simply, motions are deranged by the effects of self-induction, as has always been known; moreover, these perturbations are almost negligible, except for the negative electrons which, not having real mass, are not true matter.

2o but there is another point of view; we may suppose there are no neutral atoms, and the positive electrons lack real mass just as the negative electrons. but then, real mass vanishing, either the word mass will no longer have any meaning, or else it must designate the fictitious electromagnetic mass; in this case, mass will no longer be constant, the transversal mass will no longer be equal to the longitudinal, the principles of mechanics will be overthrown.

first a word of explanation. we have said that, for the same charge, the total mass of a positive electron is much greater than that of a negative. and then it is natural to think that this difference is explained by the positive electron having, besides its fictitious mass, a considerable real mass; which takes us back to the first hypothesis. but we may just as well suppose that the real mass is null for these as for the others, but that the fictitious mass of the positive electron is much the greater since this electron is much the smaller. i say advisedly: much the smaller. and, in fact, in this hypothesis inertia is exclusively electromagnetic in origin; it reduces itself to the inertia of the ether; the electrons are no longer anything by themselves; they are solely holes in the ether and around which the ether moves; the smaller these holes are, the more will there be of ether, the greater, consequently, will be the inertia of the ether.

how shall we decide between these two hypotheses? by operating upon the canal rays as kaufmann did upon the β rays? this is impossible; the velocity of these rays is much too slight. should each therefore decide according to his temperament, the conservatives going to one side and the lovers of the new to the other? perhaps, but, to fully understand the arguments of the innovators, other considerations must come in.