]\gt ; I31$ .
] Electrolytes rapidly lating C III .
' The Resistance of Electrolytes to the Passage of very rapidly alternating Currents , with some Investigations on the Times of Vibration of Electrical ystems By J. J. THOMSON , , F.R.S. .
dish Professor of Experimental Physics , CaInbridge .
Received January 9 , 1889 .
The netic effect of the currents in a conducting plate by alternations iu a primary system in its neighbourhood , is , at a point on the side of the plnte opposite to the primary system , in the contrary ection to netic effect of the primary .
Such a plate , therefore , tends to shield oif from a secondary the induction due to the , the diminution it produces in the cnrrent induced in the secondary upon the ctivity and thickness of the plate and the rate of reversal of the primary .
If the rate of reversal is infinitely rapid , a thin plate of very badly ) substance will be to screen from the cuit all the induction arising from the prim , while , if the rate is very slow , a thick plate of the best conducting metal will hardly be suflicient to do this .
When the current in the jmary is a few hundred times per second , a metal plate of very thickness will completely shield off all induction .
If the thickness of the plate excceds this limit , the currents induced in the layers next the primary will sbicld off all electromotive force from those layers which are more remote , so that in these layers no currents will be formed , the induced currents will thus be confined to the skin of the conductor , the thickness of the skin varying inversely as the conductivity of the plate and the rate of of the current .
In Hughes ' induction ance this screening effect of metal plates 18 made use of to compare the resistances of metals , but with } it is hardly possible to make the iently rapid to produce appreciable eflects with substances conduct so badly as electrolytes ; we can , however , b ] tbo vibrations of electrical systems such those used by Hertz in recent periments on the rate of propagation of elecrodyna]nic action get oscillations sufficiently rapid to make shielding effect of moderately thin plates of electrolytes quite appreciable .
Before the experiments made on this point , we shall consider the theory of the scree1ling effect of a slab of a lctor bounded by two planes .
Let us that these planes are rePresented by the equations ; let represent the components parallel to the axes of respectively of the vecror potential on the side of the slab on which the primary system Let be the ] ectrostatic potential , and let us suppose that all lantities vary as , then where is tant w depends on the theory of electricity we adopt .
If we assume Maxwell 's theory , , and as we shall see reason latcr 011 for believing that has always this value , we shall henceforlh in this investigation assume this value for it .
represent transverse disturbances propagated in a dielectric with velocity of propagation of electrodynamic action .
Let where the terms of the type )resent the disturbance proceedin from tlje primary , and those of the type the bance reflected from the plate .
Let Electrolytes to alter wting .
271 The boundary conditions are that continuous as we cross from one medium to another , that the induction at right angles to the bounding surface is also continuous , and that the magnetic force parallel to the surface , the components of which along the axes of and are respectively where is the magnetic permeability , should also be continuous .
Let us first the special case where the electromotive forcQ is everywhere to the conducting plate , as this is case hich is most tant for the of our experiments .
In this case , , and , and we have , since is continuous at the surface since it is continuous since is continuous , we have if is the etic permeability of the plate , .
Solving these equations we get . . . . .
( 2 ) , . . . . .
( 3 ) , . . . . .
( 4 ) , .
Equation ( 2 ) may be written and if the plate is so thin that is small , this may be written . . . . . . . . . .
( 6 ) .
Now the transverse disturbances satisfy in the dielectric equation ' of the form where is the velocity of propagation of the electrodynamic action ; in the plate they satisfy equations of the form where is the specific resistance of the substance of which the plats is Dlade .
From these equations we see that and Now if the primary system is a circular coil whose plane is parallel to the plane of the plate , and will be of the order , where is the radius of the coil ; hence if as in our experiments is large compared with , we may put Since was small compared with for the vibrations used , we have approximately , , therefore , is small compared with ; hence from lation ( we get Etectrolytes to very rapidly alternating .
273 or .
( 7 ) .
But is the proportion in which the electromotive force is reduced by the conducting plate ; hence we see that if this is considerable must be large , and in this case the reduction is bional to the thickness of the plate , the number of reversals in the ection of the current second , and the specilic resistance .
The term will not if the primary remains undisturbed .
We see from the above vestigation that if with the same rate of two different plates produce the same effect upon the induced rent , their thicknesses must be proportional to their specific resistances , or , in other , the resistance of slabs of the same area to currents parallel to bounding surfaces must be the same .
The above case is the one that is most ally useful ; there is no difficulty , however , in writing dowIl the solution of the most general case when the vector potential is noti assnmed to be parallel to plate .
Using the same notation as before we have .
J. J. Thomson : TJoe Resistance of The solutions of these equations are From these equations we can at once find and , and hence screening effect of the plate ; xactly the same conclusions hold for this as for the special case previously considered ; if the screening effectof plates is the same their thicknesses must be proportional to their specific resistance .
The rapidly alternating rents , which in the experiments wore screened by the plates , were those resulting from the electrical vibra .
tions which are set up when the electrical equilibrium of a system is .
We shall now proceed to give a somewhat detailed in vestigation of the periods of such vibrations , as the ordinary expression for the time of vibration of a condenser , whose plates are connected by an induction coil , is not applicable to this case , and , in addition , I think the result of these ations taken in conjunction with some expel.iments by Hertz , will enable us to decide the vexed question as to whether the currents flow like an incompressible fluid , and to show that Maxwell 's hypothesis on this point is correct .
The case we shall investigate is that of a straight wire connectlng two spherical balls .
Let us take the axis of the wire as the axis of 4 and let be the components of the vector potential , the electrostatic potential .
Then whers and where is a constant .
According to Maxwell 's theory according to .
Helmholtz 's more general theory , where is the velocity of propagation of the electrostatic potential , and a quantity which may be determined by the equation Electrolytes to very rapidly alternating Currents .
On this theory is also equal to where is a quantity such that the effect of the polarisation produced in a parallelipipedal element of dielectric by an electromotive force X may be represented by distributions of electricity of surface plus and minus over the faces of he parallelipipedon at right angles to X. Let all the variable quantities be proportional to .
Then in the conductor since where is the magnetic permeability and the specific resistance of the conductor , we have or , since the axis of he wire is an axis of , if be the discance of a in the wire from this axis ; and if the solution of bhe cquation is where repref cnts the Bessel 's function of zero which is finite when In the diclcctxic surrounding the wire satisfies the differential equation where is the velocity of propagation of electrodynamic action the dielectric .
Transforming this as before , this may be written where Prof. J. homson .
The Resistance of The solution of this is where is the Bessel 's function of zero order which when is infinite .
We may by symmetry , since there is no current in a plane at right angles to the wire , put\mdash ; where since and all satisfy differential equations of the same form have in wire , and in the dielectric .
if ' and are the ties of propagation of the electro .
static poteutial in the wiro and respectively , we in the wire , where ; and in the dielectric , where Since is continuotls as we cross from the wire to the , we have if be the radius of the wire , and since is continuous , we have here and are the values of in the wire and dielectric tively .
Since and are continuous , we have Electrol.ytes to very rapidly alternating Currents .
or Since the fnetic force parallel to the surface of the wire is continuous , is continuous , and therefore cm , or ( 11 ) .
From equations 9 ) and ( 11 ) we get . .
( 12 ) , and from ( 10 ) and ( 11 ) we get ( 13 ) .
HeIIce , eliminating A and from these equations , we get .
In the cases dealt with in these experiments the rate of vibrat:on was so rapid that na was very large .
In this case .
If Maxwell 's theory is true , the right hand side vanishes , fiince , and we have : or . . . . . . . . . . . . . . . . . .
( 15 ) .
The right hand side of this equation is very small , so must be very small .
In this case approximately , Prof. J. homson .
The Resistunce of where so equation ( 14 ) becomes solution of which ( see ' London Math. Soc. Proc vol. 17 , p. 316 ) is and erefore in this , since the second term on the right-hand side is small compared with the first , the distul.bance is propagated along the wire with the same velocity as that of electrodynamic action through the dielectric .
The amplitude of the vibration will sink to of original after traversin , a distance If , however , does not vanish , and if we snppose which will be the case unless the velocity of opagation of ostatic potential is exceedingly small compared with that of electrodynamic action , since in this case since is very large compared with u1lity , equntion becomes , remembering that is large , ; unless be very small , the right hand side in this very large compared with the first on the left , and the tion becomes 1@89 .
] Electrolytes to very rapidly atter lating Cnrrents .
279 The right hand side is small so that will be small , and the equation to determine it or The solution of this equation is approximately or say ; but therefore and the velocity of of the disturbance through the is or .
Since the imaginary part of ?
does not involve and occurs Tlllder logarithm , the rate at which the vibrations die away will in this be ctically i dependent of the resistance and size of the wire .
Thus , unless Maxwell 's is truc , the rate of propagation of a very rapidly alternatin disturbance throug.h a conductor is not the same as that of electrodyuamic action the su dielectric ; if is positive it goes faster through the wire than ough the dielectric , while if is it goes more slowly .
The rate of propagation through the wire is most t not quite of size and conductivity of the vire and of the rapidity of the ations .
Thus , if it could be ed that the velocity of a disturb through a wire appreciably from the velocity of action , and the rate at which the vibrations die away did not depend upon the sistance , it be to sbow that Maxwcll 's assumption is tble .
would seem to show thab the rate of propttgation through a metallic wire is less that of odynamic action through the dielectric ; but I believe he has lately follnd that the increases rapidly with pidi t of tbc tions , which is luconsistenb with the above result , if and are independeut of No experiments seem to have been made on late at which the vibrations die away , though this would be one of the best ways of dist. lnguisbing between the theories .
same as that of electrodynamic action through air .
We shall now investigate the time of vibration of a system sisting of a straight wire connecting two spherical balls .
Let us the middle of the wire as the origin , and suppose that the flow of electricity is symmetrical about this point ; at points equidistanb from the origin the electrostatic potential will be equal and opposite .
Using the same notation as before , let in the wire , in the dielectric , in the wire , in the dielectric .
If is the intonsity of the current parallel to the axis of .
The quantity of electricity which has passed across any section at right angles to the axis is given by since , , we see that If ends of tho wire are given by , the rate at which electrioity flows across the end is given by , or by equation ( 13 ) ml ; if , however , is the capacity ( electromagnetic measure ) of the condenser at the end so that ml .
Equating for we get .
Since it is very large compared with u1lity , and if and are small , ' The form of the solution will depend upon the nitude A .
If this is small then will be small , and ue have or since , if Maxwell 's theory be true , This result , however , is only true when is llot large compffi with , in this case ml will be large , and therefore will be approximately , and so on .
Thus in this case the ends of wire are nodes of the electrical vibrations , and the gravest modo of vibration is that in which the wave-length is the length of the wire ; here the wave-length , and therefore the rapidity of will be independent of the capacities of the condensers at the ends .
If is finite , since the second term on the right hand equation ( 17 ) will in this case be large compared with the first , is large , the equation reduces to\mdash ; ; Electrolytes to very rapidly alternating Cu.rents .
283 the capacity of the condenser were exceedingly large , much greater than that requisite for the same purpose in the preceding case , the time of vibration would be independent of the capacities of the ends ; and conversely , if we could prove that the time of vibration depends upon the capacity , we should prove .
Now Hertz in his experiments seems to have been able to bring two circuits into resonance by altering the capacity of the ends , though these capaoities were exceedingly small compared with 1 , / 10 of a microfarad .
This , therefore , is exceedingly testimony in favour of the truth of Maxwell 's theory , at any rate for conductors .
[ Note added February 15 , 1889.\mdash ; We can find the ratio of to the values of for a dielectric and conductor respectively , by considering the reflection of an disturbance at a metallic surface .
Using the }ation of the ) of the paper , let the iucident waves of the vector potential be expressed by the reflected waves by .
Then assuming that are with we find\mdash ; Thus the electromotive force parallel to the surface of the reflector does not vanish at the surface unless .
Hertz ( ' Wied .
Ann 34 , 616 ) found that when the plalle of the secondary cuit was Parallel to reflecting surface , the sparks vanished at the refiecting surface , thus showing that is at a , ny small .
The method founded on the law of decay of the vibrations is more delicate , as it whether or not is small and , is a large quantity .
] In the above work we have assumed that is small , but if be If for all substances , theu since the electromotive to the axis of is , and since the component of the electromotive force is and components are .
Thus the force is Propagated with the velocity of the transverse vibrations on Electrica heories , \ldquo ; ' Brit. Assoc. Report , ' Aberdeen , 1885 , p. 138 ) , and since satisfy the solenoidal condition , there is uo condensation .
The rate of propagation of a disturbance through a conductor only equal to that of the electrodynamic sction through a dielectric when .
is small , and though this will be so for the rapid vibrations we are dealing with when the conductor is metallic , it would not be so if the conductor were a dilute electrolyte or a rarefied gas .
In this case the rate of propagation of the disturbanoa through the conductor would not be the same as that through the dielectric .
In this case action gated along the conductor , and that opagated through bhe dielectric , would when they interfere and set up vibrations , so that along the conductur there would be a series of stationary nodes at which the current vanished , in other words , the current the conductor would sbriated .
In the discharge of electricity through rarefied gases have the ent p through a conducor of high , and it seems possible that the striations ich are observed in the may be du to the interference of the disturbance propagated the gas and that passing tlJrougb the dielectric .
of the striae on rarefaction , and on the dia1nekr of charge tube , are with this view .
The resistances of the electrolytes to the very rapidly alternaMg cnrrents were compared in the following way:\mdash ; are three coils .
two of which and are of the same dimensions , and are nearly but not quite closed .
rical balls fastened to the ends of these coils .
The two screening effect of conducting plates , for we saw by equation ( 7 ) when a screen of was interposed the electromotive is only ' when is the conductivity of the metal .
Since the electromotive forco in the plane of the screen , which fr taken as the plane of , is of the form by topping and starting the coils several times ; if , during this \mdash ; ro parks passed in the secondary , the sparks were considered gped .
A variation of 5 per cent. in the quantity of electrolyte present would cause the system to pass this point one way or another in 1narked way .
balls at the extremity of the secondary were adjnsted so that @parks passed freely before the electrolyte was put in , after each eximent the electrolyte was removed , and care was taken to ascertain that sparks still passed as freely as before so as .
to guard aga.lnst any accidental disarrangement of the secondary durlng the experiment .
Three sets of coils were used which we shall describe by I , II , III .
Set I consisted of two circular brass coils , .
in circumference .
The diameter of the brass rod of which they were made was about 0 . .
; the balls at the extremities were 2 .
in diametel .
The time of vibration of this coil , calculated by equation ( 18 ) , is about seconds .
Set II consisted of two circular copper coils .
in circumference , the rod of which they were made being about .
in diameter ; the balls at the extremities were 1 .
in diameter .
The time of vibration is about seconds .
With these small coils balls of the secondary had to be exceedingly close togethcr ) order to get sparks , but when the micrometer screw was properly adjusted the sparks were very bright and the indications quibe deiinite .
The coils we ] about .
apart .
Set consisted of two rectangular coils made of sheet lead , one side was , the other 40 , the breadth of sides was 5 , and the diameter of the balls at the extromity 2 .
The distance between the coils was .
The time of vibration about seconds .
The olytes used were solutions of\mdash ; Sulphuric acid , specific gravity of solution . . . . . . .
Ammonium chloride , , , , . . . . . . .
Sodium Potassium . . . . . . .
Ammonium nitrate , , , , . . . . . . .
Potassium carbonate , , , , . . . . . . .
In the following table the relative thickness of the films of these rukIauces required to stop the spark given , each number bein the mean of several observations .
The thickness of the film was taken as unity .
An observation with sulphuric acid was made before and the observation with any other electrolyte .
The thicknesses of the films are by equation ( 7 ) proportional to the ecific resistances , so that the numbers in the fourth line give the elative resistance of the electrolytes to currents whose directions ue reversed from to times per second .
In order to see whether these resistances are the same as those with an almol infinitely slower revcrsat , I determined the resistance of the lytes by a commutator which reversed the current through the electrolyte a ) times a second , and kept the direction of current through the galvanometer constant .
The electrodes .
and no polarisation could be detected .
The numbers given in the last line of the above table , and agree sufficiently well to enable us to say that the ce of olytes is the saw when the current is reversed a hundred million times a second as Electrolytes to very alternating Currents .
289 bhe resistance of the graphite film was 67 obms , and that of the huric acid 7.2 obms ; thus we may say that the ratio of the specific stances of grap 1 an and sul huric acid is the same when the nurrents are steady as when they are reversed times per second .
Since the ratio of the resistances of such things as graphite ffi electrolytes remaius the same , we may conclude that the res stances hemselves remain unaltered .
The method described above for comparing the resistances of electrolytes is one that can be very easily and quickly applied , and only requlres the mplcst apparatus : an induction coil , or , if that is not available , au eleetrical machine , being all that is required .
The method has the advantagc of avoiding the use of electrodes , as all the circuits in the electrolyte are closed .
Since electrolytes are transparent , they must , if the electromagnetic theory of is true , act as lators when the currents are revel.sed as often as light vibrations , or about per second .
We have seen , however , that they conduct as well when the currents are reversed times a second as when they are steady ; thus the molecular processes which cause electroiytic conduction must oCCtlpy a time be ween 1 and seconds .
Another point which can be ] by this method is , whether a vacuum is a conductor or sulator .
According to one view the great resistance which a highly exhausted offers to the passage of electricity is due to the difficulty .
the current from the electrode into the refied gas : when once the current has got there , there is , according to this theory , no resistance to its passage : if this theory is corroct , a highly exhausted receiver placed between the primary and secondary circuits ought to stop the sparks the latter , as incc all the circuits are closed to be no obstacle to the passage of the induced .
In order to test this I took a box , .
by50 .
by 4 , the top and of sheets of plate-glass fitting into vooden sides ; the shects of glass were also Upported by five ebonite pillars placed at equal intervals .
the snrface .
The box was repeatedly pped into a bath filled with melted paraffin until it was surrounded by a coating of paraffin about 2 .
thick .
The paraffin was then smoothed over with a hot soldering iron , and then eovered with a layer of shellac varnish .
The box was then exhausted by a mercury-pump , and it was found that the prescould be reduced to about 1 mm. of mercury , but no further .
When this vacuum was placed ) etweeu the primary and secondaxy coils it did not )roduce the slightest effect upon the sparks , so that must be very small indeed compared with that of the ectrolytes used in the preceding experiments .
I am having an earthenware vessel made with which I hope to repeat the at much higher exhaustions .
I tried whether the conductivity of the eloctrolyte was altered latter ; a current of about 2 amperes , which was reversed about times a second , was then sent through the sulphuric acid , but th passage of the current did not seem to produce any effect upon the sparks in the secondary .
I conclude , therefore that resistance of an electrolyte is not aflected by the passage of a currert .
I wish to express my thanks to my assistant , Mr. E. Everett , for the zeal and skill he has displayed in these experiments .
[ Note added February 15 , 1889 .
have recently tried the effect of very high vacuum in stopping the sparks .
The primary circuit sisted of two straight wires with spheres fastened to one end of each ; these wires were connected with the poles of an induction coil , and the sparks passed between the spheres .
The secondary consisted of two similar wires , with smaller balls at the ends , the distance between the balls being very small .
The length of the wires of the secondary wafi aItered utltil it was in } ance with the The secondary was placed in a hollow cylinder orrned of two coaxial glass tubes , sealed on to a mercury pump , by means of which a very high vacuum was obtained in the space between them , which sur rounded the secondary .
This vacuum , however , did not produce the slightest effect on the sparks .
] Presents , January 1 , 1889 .
Transactions .
Breslau :Schlcsische Gesellschaft fur vaterlandiscbe Cultur .
Jahresbericht .
1887 . .
Breslau 1888 .
The Society .
Brisbane :Royal Society of Queensland .
Proceedings .
Vol. V. risbane 1888 .
The Society .
Dublin :Royal Irish Academy .
Proceedings .
Ser. 3 .
Vol. I. No. 1 .
Dublin 1888 ; Transactions .
Vol. XXIX .
3-4 . .
Dublin 1888 .
The Academy .
Edinbnrgh:\mdash ; Geological Society .
Transactions .
Vol. V. Part .
Edinburgh 1888 .
The Society .
Falmouth :Royal Cornwall Polytechnic Society .
Annual Report 1887 . .
Falmouth [ 1888 ] .
The Society .
Florence :Biblioteca Nazionale Centrale .
Bollettino dell blicazioni Italiane .
Giugno\mdash ; Dicembre , 1888 .
The Libraq$ Hermannstadt:\mdash ; Siebenburgischer Verein fur Naturwissenschafton

