= x ^ 



LIBRARY OF THE UNIVERSITY OF CALIFORNIA 




LIBRARY OF THE UNIVERSITY OF CALIFORNIA 

OV^TD 







LIBRARY OF THE UNIVERSITY OF CALIFORNIA 



OF CALIFORNIA LIBRARY OF THE UNIVERSITY OF CALIFORNI 




OF CALIFORNIA LIBRARY OF THE UNIVERSITY OF CALIFORNI 









OF CALIFORNIA 



LIBRARY OF THE UNIVERSITY OF CALIFORNh 



MODERN VIEWS 

OF 

ELECTRICITY 



MACMILLAN AND CO., LIMITED 

LONDON . BOMBAY . CALCUTTA 
MELBOURNE 

THE MACMILLAN COMPANY 

NEW YORK . BOSTON . CHICAGO 
ATLANTA . SAN FRANCISCO 

THE MACMILLAN CO. OF CANADA, LTD. 

TORONTO 



MODERN VIEWS 



OF 



ELECTRICITY 



BY 

SIR OLIVER LODGE 

D.Sc., LL.D., F.R.S. 

Principal of the University of Birmingham, 
Late Professor of Physics in University College Liverpool 



THIRD EDITION REVISED 



OF THE 

UNIVERSITY 




MACMILLAN AND CO., LIMITED 

ST. MARTIN'S STREET, LONDON 

1907 






P.ICHARD Cl.AY AND SONS, LIMITED, 

BREAD STKEET HILL, EC., AND 

BUNGAY, SUFFOLK. 

first Edition printed 1889. Second 1892. 
Third 1907. 



PREFACE 

TO THE THIRD EDITION 

IN these days it is customary and natural to attend 
closely to the doctrine of electrons, and to the recently 
discovered phenomena specially associated with that 
doctrine. But there is a mass of long-established 
electrical science before and beside all that ; and, so 
far from lessening their importance, the discovery of 
electrons brings the well-known facts of electrostatics 
into greater prominence. It is impossible rightly 
to follow the newest researches without acquaintance 
with the fundamental truths concerning electricity and 
magnetism in general, and yet there is some danger 
that these things may be temporarily overlooked and 
partially neglected ; consequently it seemed likely to 
be helpful, especially to self-taught students and 
general readers interested in science, if the book 
called " Modern Views of Electricity " were republished 
with a few additions arid such abbreviations as are 
permissible. 

In order to facilitate reference the numbering of 
sections has been retained the same as in previous 
editions, but no attempt has been made to indicate 
the little changes that have been freely introduced 
wherever desirable. It is noteworthy that few actual 

170101 



vi PREFACE 

corrections have had to be made ; showing that new 
discoveries are of a supplementary rather than of a 
revolutionary character. The chief difference is that 
whereas in previous editions the question of the 
inertia of electricity was constantly treated as an 
open question, within reach of being shortly answered, 
it may now be regarded as to a great extent settled, 
at least in a proximate manner. The atomic theory 
of electricity, begun by Maxwell and brought into 
prominence by Helmholtz, now holds the field ; and 
the experimental establishment of the separate exist- 
ence of electric charges has removed all rational 
doubt about the electromagnetic theory of their 
inertia. Ultimate explanation of inertia is an 
etherial, and perhaps insoluble, problem, but an 
electric treatment of the inertia of matter is legiti- 
mate as a working hypothesis ; and an estimate of the 
value even of ethereous inertia is not beyond us, so 
in the last chapter (XVII.) it is introduced. 

The subject of the discharge of Electricity through 
gases, however, has so greatly developed that it 
requires a separate treatise, and is accordingly but 
lightly touched on. Nor has it been thought appro- 
priate to attempt to refer to extremely recent work 
and theory on the nature of metallic conduction. 
These things are specially dealt with in the writings 
of Professor J. J. Thomson. 

To the general reader, as well as to the student, 
the lectures appended at the end of the book are 
especially commended : the old lectures remain quite 
useful, and some new lectures and articles have been 
added : one of them dealing in a semi-philosophic 



PREFACE vii 

way with the Interstellar Ether, and two others giving 
a popular account of recent discoveries as to the 
probable nature of matter, and some idea of the 
atomic theory of electricity, a subject which is 
dealt with in a more thorough fashion in my book 
called Electrons, or the Nature and Properties of 
Negative Electricity. 

For the rest we may say, much as was said in the 
preface to the first edition, that the doctrines ex- 
pounded in this book are : the electrical nature of 
light, a recent theory of matter, and an ethereous 1 view 
of electricity. Crudely one may say that as sound 
and heat are forms of energy, or modes of motion, so 
light and electricity and matter are all forms of ether, 
or modes of ethereous manifestation. 

Persons who are occupied with other branches of 
science or philosophy, or with literature, and who have 
therefore not kept quite abreast of physical science, 
may possibly be surprised to see the intimate way 
in which the ether is now spoken of by physicists, 
and the assuredness with which it is experimented 
on. They may be inclined to imagine that it is still 
a hypothetical medium, whose existence is a matter 
of opinion. Such is not the case. The existence of 
an ether can legitimately be denied only in the 
same fashion as the existence of matter can be denied : 
the evidence for its existence can be doubted 
or explained away, in the one case as in the other, 
but the evidence for ether is as strong and almost 
as direct, though not so plentiful, as the evidence for 
air. The eye may indeed be called an etherial 

1 See footnote on p. 329. 



viii PREP^ACE 

sense-organ, in the same sense as the ear can be 
called an aerial one, and somewhat in the same sense 
as the hand and muscles may be called a sense-organ 
for the appreciation of ordinary matter. 

A rough and crude statement, adapted for popular 
use, is that electricity and ether are identical. A better 
statement is that electricity is specifically modified 
ether ; though that is not all that has to be said, for 
there are two opposite kinds of electricity, and there 
are not two ethers. But there may be two aspects 
of one ether, just as there are two sides to a sheet 
of paper, or two aspects of a transparent clock face, 
or two constituents in common salt ; and similarly 
may positive and negative electricity be two aspects, 
or, as I have sometimes called them by chemical 
analogy, " components," of the ether. Anything 
which can be sheared and ether is incipiently 
sheared by every electromotive force applied to it 
must consist of two parts sufficiently different to 
travel or to be displaced in opposite directions : and, 
in the case of ether, these parts are only recognisable 
or manifested as two in the act of partial separation 
or shear which is called c electric displacement.' 

If this statement is vague, it is because our present 
knowledge of the structure of the ether is vague ; not 
because the relationship of electricity to ether is un- 
certain, or will be anything but definite so soon as we 
know the constitution of the ether more precisely. 
It is impossible to treat this part of the subject in a 
satisfactory manner at present, but an attempt to in- 
dicate the probable lines of development is made in a 
concluding chapter. 



PREFACE ix 

Explanation always progresses by stages ; no 
explanation is ultimate ; every explanation is a step 
up removal of something from a lower to a higher 
category. Thus comets, at one time, might have been 
anything ; they have been shown to be a form (or 
swarm) of meteorites. Meteorites, again, have been 
shown to be lumps of common matter usually iron 
or rock. There remains the question, What is iron or 
rock, or any form of matter ? Heat was once thought 
to be a form of matter ; it is now known to be a form of 
energy. There remains the question, What is energy ? 
Electricity may have been thought to be a form of 
energy ; it has been shown to be a form of ether. 
There remain the questions, What is ether? How 
does it oscillate to produce light ? how is it modified 
into an electric charge or electron ? and (we may 
now add) how do electrons build up matter ? 

These questions have not yet been answered. 

UNIVERSITY OF BIRMINGHAM, 
April, 1907. 



CONTENTS 



INTROD UCTION. 

PAGE 

THE ETHER . i 



PART I. 
ELECTROSTA TICS. 

CHAPTER I. 
FUNDAMENTAL NOTIONS ABOUT ELECTRICITY . 



CHAPTER II. 
THE DIELECTRIC 18 

CHAPTER III. 
CHARGE AND INDUCTION 32 



xii CONTENTS 

PART II. 

CONDUCTION. 

CHAPTER IV. 

PAGE 

METALLIC AND ELECTROLYTIC CONDUCTION 65 

CHAPTER V. 
CURRENT PHENOMENA 87 

CHAPTER VI. 

CHEMICAL AND THERMAL METHODS OF PRODUCING CURRENTS. 
CONDUCTION IN GASES 106 



PART III. 
MAGNETISM. 

CHAPTER VII. 
RELATION OF MAGNETISM TO ELECTRICITY 131 

CHAPTER VIII. 
NATURE OF MAGNETISM 146 

CHAPTER IX. 
STRUCTURE OF A MAGNETIC FIELD 166 



CONTENTS xiii 

CHAPTER X. 

PAGE 

MECHANICAL MODELS OF A MAGNETIC FIELD 172 

CHAPTER XI. 
MECHANICAL MODELS OF CURRENT INDUCTION 190 

PART IV. 
RADIATION. 

CHAPTER XII. 
RELATION OF ETHER TO ELECTRICITY .213 

CHAPTER XIII. . 

CONSTANTS OF THE ETHER 226 

CHAPTER XIV. 

ELECTRICAL RADIATION, OR LIGHT 238 

CHAPTER XV. 

ELECTRIC AND MAGNETIC INFLUENCES ON LIGHT 268 

CHAPTER XVI 
RECENT PROGRESS (1892) 290 

CHAPTER XVII 
MOST RECENT VIEWS CONCERNING THE ETHER (1907) . . . 318 



xiv CONTENTS 

APPENDED LECTURES 

LECTURE I. 

PAGE 

THE RELATION BETWEEN ELF:CTRICITY AND LIGHT 345 

LECTURE II. 
THE ETHER AND ITS FUNCTIONS . . 359 

LECTURE III. 
THE DISCHARGE OF A LEYDEN JAR 387 

LECTURE IV. 
MODERN VIEWS OF MATTER 409 

LECTURE V. 
THE INTERSTELLAR ETHER 450 

LECTURE VI. 
THE LESSONS OF RADIUM 463 

APPENDIX 479 

INDEX . , .. 503 



MODERN VIEWS 

OF 

ELECTRICITY 




MODERN VIEWS OF 
ELECTRICITY 

INTRODUCTION 

THE ETHER 

THE most interesting and important portion of 
physics at the present day is that constantly growing 
portion which is concerned with the properties and 
function of the universal connecting medium, the 
" ether." All the phenomena of light, of electricity, 
and of magnetism, as well as of what is spoken of 
as radiant heat, are intimately connected with, and 
indeed wholly dependent on, that medium ; it is 
through these agencies that we have gradually become 
aware of its existence and able in some degree to 
investigate it. Probably gravitation, cohesion, and 
chemical affinity are no less closely associated with 
the ether; and it is becoming doubtful whether matter 
itself could exist without it. 

The eye may be called an ethereal sense-organ, 
just as the ear is an aerial one. Both are adapted to 
appreciate ripples the one, large ripples or waves in 
air ; the other, minute ripples, or quivers, in ether ; 

B 



2 INTRODUCTION 

and the appreciation of those electrical ripples or 
quiverings, insignificant though they are from many 
points of view, is intensely important to us. Without 
the power of apprehending the ripples thrown off, 
and reflected from, and scattered by, the several 
objects in our neighbourhood, we should be both 
blind and deaf. It is only by inference from the way 
they deal with, and disperse, the ethereal quiverings, 
that we become aware of odourless objects beyond the 
range of our touch : We are said to " see " them. In 
a room where the ether is quiescent we can see 
nothing ; but if by any means, as by lighting a 
candle, we throw it into vibration, the vibration re- 
turns to our eyes from all the objects in the room ; and 
thereby, through habit and experience, we gain such 
quick intuitive information concerning those objects, 
that we are not conscious of the means of information, 
and express our perception as if it were direct and 
immediate, by saying that we see the objects them- 
selves. That is, in fact, what we mean by "seeing." 
We do not see the waves, we see the obstacles ; but 
it is entirely by means of the ethereal waves that the 
existence and nature of the objects are inferred. The 
extraordinary precision and rapidity of this mode of 
inference is perfectly marvellous, and could never 
have been believed if it had only theoretically been 
explained. 

Before beginning a book on electricity, where 
we are to come into contact with other properties 
and functions of this most interesting and omni- 
present medium, it is as well to exercise our imagina- 
tion concerning it, and try to realize the vitally 



INTRODUCTION 3 

important part it plays throughout the universe. 
Probably we are far from knowing all its functions 
as yet ; but knowing many of them we suspect more, 
and the reader is recommended to turn to a couple of 
lectures on the subject, which he will find printed on 
pages 359 and 450 further on, and read those, at 
intervals, before and during his perusal of the rest of 
this book ; for they will serve as a useful introduction. 
The book as a whole, beginning with a survey of 
well-known and elementary phenomena, leads up to 
the generalizations shadowed forth in Chapter XVII 
and Lectures IV and VI, which attempt to indicate 
how the idea of ether and the idea of electrons can be 
combined so as to supplement and illuminate each 
other. The hypothesis ultimately suggested is that 
excessively minute portions of ether have, by some 
unknown means, been dissociated here and there into 
electric charges, and that these immensely numerous 
mobile specks of electrified ether through the forces 
they exert and the disturbances they originate con- 
stitute the substratum of what appeals to our senses 
as matter. 



B 2 



PART I 

ELECTROSTA TICS 




CHAPTER I 

FUNDAMENTAL NOTIONS ABOUT ELECTRICITY 

I BEGIN by saying that the whole subject of elec- 
tricity is divisible for purposes of classification into 
four great branches. 

(1) Electricity at rest, or static electricity : wherein 
are studied all the phenomena belonging to stresses 
and strains in insulating or dielectric media, brought 
about by the neighbourhood of electric charges or 
electrified bodies at rest immersed therein ; together 
with the modes of exciting such electric charges-, and 
the laws of their interaction. 

(2) Electricity in locomotion, or current electricity : 
wherein are discussed all the phenomena set up in 
metallic conductors, in chemical compounds, and in 
dielectric media, by the passage of electricity through 
them ; together with the modes of setting electricity 
in continuous motion, and the laws of its flow. 

(3) Electricity in rotation, or magnetism : wherein 
are discussed the phenomena belonging to electricity 
in whirling or vortex motion, the modes of exciting 
such whirls, the stresses and strains produced by 
them, and the laws of their interaction. 



8 ELECTROSTATICS CHAP, i 

(4) Electricity in vibration or radiation : wherein 
are discussed the propagation of periodic or undu- 
latory disturbances through various kinds of media, 
the laws regulating wave velocity, wave-length, re-, 
flection, interference, dispersion, polarization, and a 
multitude of phenomena studied for a long time 
under the heading " Light." 

Although this last is the most abstruse and difficult 
portion of electrical science, a certain fraction of it has 
been known to us longer than any other branch, and 
has been studied under special advantages, because of 
our happening to possess a special sense-organ for its 
appreciation. 

Now in order to be able to survey these four great and 
comprehensive groups in moderate compass, it will be 
necessary for me to assume acquaintance with the ele- 
mentary facts and proceed at once to their elucidation. 

The views concerning electrification which I shall 
try to explain are in some sense a development of 
those originally propounded by that remarkable man, 
Benjamin Franklin. The accurate and acute experi- 
menting of Cavendish laid the foundation for the 
modern theory of electricity ; but, as he worked for 
himself rather than for the race, and as moreover he 
was in this matter far in advance of his time, Faraday 
had to go over the same ground again, with exten- 
sions and additions peculiar to himself and corre- 
sponding to the greater field of information at his 
disposal three-quarters of a century later. Both these 
men, and especially Faraday, so lived among phe- 
nomena that they yielded up their hidden secrets 



3 FUNDAMENTAL NOTIONS 9 

in a way unintelligible to ordinary workers ; but while 
these men themselves arrived at truth by processes 
that savour of intuition, they were unable always to 
express themselves intelligibly to their contemporaries, 
or to make the inner meaning of their facts and 
speculations understood. Then comes Maxwell, with 
his keen penetration and great grasp of thought, com- 
bined with mathematical subtlety and power of ex- 
pression ; he assimilates the facts, sympathizes with 
the philosophic but untutored modes of expression 
invented by Faraday, links the theorems of Green 
and Stokes and Thomson to the facts of Faraday, 
and from the union rears the young modern 
science of electricity, whose infancy a quarter of a 
century ago was so vigorous and so promising that 
we all looked forward to the near future in eager 
hope and expectation of some greater and still more 
magnificent generalization. 

3. Everyone knows that there have been fluid or 
material theories of electricity for the past century : 
all know, moreover, that there has been a reaction 
against them. There was even a tendency at one 
time to deny the material nature of electricity and 
assert its position as a form of energy. This was 
doubtless due to an analogical and natural, though 
unjustifiable, feeling, that just as sound and heat and 
light had shown themselves to be forms of energy, so 
in due time would electricity also. If such were the 
expectation, it has not been justified by the event. 
Electricity may possibly be a form of matter it is 
not a form of energy. It is quite true that electricity 
under pressure or in motion represents energy ; but the 



io ELECTROSTATICS CHAP, i 

same thing is true of water or air, and we do not 
therefore deny them to be forms of matter. Under- 
stand the sense in which I use the word electricity. 
Electrification is a result of work done, and is most 
certainly a form of energy ; it can be created and 
destroyed by an act of work. But electricity none 
is ever created or destroyed ; it is simply moved and 
strained, like matter. No one ever exhibited a trace 
of positive electricity without there being somewhere, 
connected with it by lines of force, an equal quantity 
of negative. 

The simplest proof of this statement consists in 
making experiments inside a closed conducting in- 
sulated room or shell ; it may be the size of a living- 
room, or the size of a beer-tankard, whichever is most 
convenient. All known electrical experiments being 
performed inside such a room, bodies electrified 
strongly, moved about, sparks taken, &c., &c., a 
sensitive electroscope connected to the room shall not 
show the slightest permanent effect. In other words 
the room will not become in the slightest degree 
charged. I say no permanent effect, because it is just 
possible that small transitory effects may occur during 
the rapid rearrangement of internal charges. I be- 
lieve that I have shown 1 that even transitory effects 
are not really possible ; but whether they are or not, 
they have nothing to do with our present argument. 
All the electrifications that have been going on will not 
have resulted in the creation of the minutest quantity 
of electricity ; the only way to charge the room is to 
convey a charge to it from some other body outside. 

1 Journal Inst. Elec. Engineers, vol. xix. pp. 371 379. 



4 FUNDAMENTAL NOTIONS 11 

This is the first great law, expressible in a variety 
of ways : as, for instance, by saying that total 
algebraic production of electricity is always zero ; 
that you cannot produce positive electrification with- 
out an equal quantity of negative also ; that what one 
body gains of electricity some other body must lose. 

Now, whenever we perceive that a thing is produced 
in precisely equal and opposite amounts, so that what 
one body gains another loses, it is convenient and 
most simple to consider the thing not as generated in 
the one body and destroyed in the other, but as simply 
transferred. Electricity in this respect behaves just like 
a substance. This is what Franklin perceived. 

4. The second great law is that electricity always, 
under all circumstances, flows in a closed circuit, the 
same quantity crossing every section of that circuit, so 
that it is not possible to exhaust it from one region of 
space and condense it in another. 

Another way of expressing this fact is to say that 
no charge resides in the interior of a hollow conductor, 
but that every trace of charge is on the outer surface 
and penetrates to no appreciable depth. 

Another is to say that total induced charge is 
always equal and opposite to inducing charge. 1 

This second law can also be illustrated by the in- 
sulated room or conducting cavity already mentioned. 
Having found that internal electrification produces no 
effect on an outside electroscope connected to the 
walls, proceed to pass a charge into the cavity through 

1 To explain how it is that these statements are equivalent, and 
generally to emphasize this fundamental part of the subject, is the 
object of 14 A at end of the next chapter. 



12 ELECTROSTATICS CHAP, i 

a temporarily opened window or lid. Instantly the 
chamber has become charged by a definite amount, viz. 
by the precise amount which has then been introduced 
into its cavity. The charge need not be in any way 
communicated to the chamber ; all that is necessary 
is that it shall be wholly inside. Moving the charge 
about, or letting it spark to the walls of the chamber, 
makes not the slightest difference to the electroscope 
outside. It may be watched through a microscope at 
the instant the spark occurs, and it will not show the 
slightest twinkle ; although it is connected to the 
enclosure by a metallic wire or rod. Both these 
experiments of the hollow chamber were made by 
Faraday, and the latter is well known and often 
quoted as his " ice-pail " experiment, because he hap- 
pened to use an ice-pail sometimes as his insulated 
chamber. In many modern experiments this principle 
is utilized, and the device is spoken of as Faraday's 
hollow or collecting vessel. 

Another mode of illustrating the same series of facts 
is afforded by an insulated parrot cage, with an electro- 
scope inside it connected by a wire to the bars of the 
cage. The cage may now be charged strongly : its 
potential may be changed from a million volts positive 
to a million volts negative : sparks of any length may 
be taken from it ; but, provided the meshes are close 
enough for it to be regarded as a really closed vessel, 
the electroscope inside is wholly unaffected. In mak- 
ing this experiment the electroscope must be connected 
with the bars of the cage ; for, if it be detached, it can 
be affected by charged air blown through the meshes. 
Directly a charge gets inside the cage it can affect the 



4 FUNDAMENTAL NOTIONS 13 

electroscope easily enough, unless there is good wall- 
connexion, which completely protects it. 

The open-work of a cage is objectionable on this 
account, that it allows the permeation of electrified air, 
just as it might allow an electrified pith ball to be 
thrown in. Making the meshes close is no guarantee 
against this source of disturbance : I have blown 
electrified air, from a point, through very fine earth- 
connected copper gauze, and affected strongly an 
electroscope on the other side. 

No doubt a solid metal-walled room is secure against 
this cause of disturbance, but then it is difficult to see 
the electroscope. Faraday however constructed a room 
for the purpose, big enough to get into himself, and 
thus performed the experiment quite satisfactorily. 

But perhaps the most rigorous mode of examining 
the precise truth of the property of electricity which 
lies at the bottom of all these experiments is that 
adopted in the famous Cavendish experiment : some- 
times referred to in French books by the name of Biot. 
This consists in charging strongly an insulated sphere 
provided with a couple of hemispherical caps which 
can afterwards be taken off mechanically, and connect- 
ing a delicate electroscope immediately afterwards to 
the disclosed ball. Not the faintest trace of charge is 
found upon it. This experiment was repeated by 
Clerk Maxv/ell and Dr. Donald McAlister in the 
Cavendish Laboratory, Cambridge, using a Thomson 
Quadrant Electrometer and all modern appliances, 
with absolutely negative result. 1 

1 See Maxwell's "Electrical Researches of Cavendish," pp. 104 
and 417; also Maxwell's "Scientific Papers," p. 612. An interesting 



14 ELECTROSTATICS CHAP, i 

This series of experiments is most vital, and gives 
us fundamental information regarding electricity ; let 
us consider how we can best express their teaching. 

5. When we thus find that it is impossible to charge 
a body absolutely with electricity, that though you can 
move it from place to place it always and instantly 
refills the body from which you take it, so that no 
portion of space can be more or less filled with it than 
it already is, that it is impossible by any rise of 
potential to squeeze a trace of electricity into the in- 
terior of a cavity, and that if a charge be introduced 
a precisely equal quantity at once passes through the 
walls to the outside ; it is natural to express the 
phenomenon by saying that electricity behaves like 
a perfectly incompressible substance or fluid, of which 
all space is completely full. That is to say, it behaves 
like a perfect and all-permeating liquid. ( I4A.) 

6. Imagine now that we live immersed in an infinite 
ocean of incompressible and inexpansible all-permeat- 
ing perfect liquid, as fish live in the sea ; how can we 
become cognizant of its existence ? Not very easily ! 
Not by its weight, for instance, for we can remove 
it from no portion of space in order to try whether 
it has weight. 

We can weigh air, truly, but that is simply because 
we can compress it and rarefy it. An exhausting or 
condensing pump of some kind was needed before 
even air could be weighed, or its pressure estimated. 

But if air had been incompressible and inexpansible, 

little experiment with soap bubbles, made by Mr. Vernon Boys, illus- 
trates the fact that the depth to which a charge penetrates is less than 
the diameter of a few molecules, for one soap bubble inside another is 
entirely screened from such electrostatic forces as can be applied. 



6 FUNDAMENTAL NOTIONS 15 

if it had been a vacuum-less perfect liquid, pumps 
would have been useless for the purpose, and we should 
necessarily be completely ignorant of the weight and 
pressure of the atmosphere. 

How then could we become cognizant of its ex- 
istence ? In four ways : 

(1) By being able to pump it out of one elastic 
bag into another : not out of one bucket into another, 
if you lived at the bottom of the sea you would 
never think about filling or emptying buckets, the 
idea would be absurd ; but you could fill or empty 
elastic bags, and could notice the strain phenomena 
exhibited by the bags, and their tendency to burst 
when over full. Water (or air) may be pumped out 
of one elastic bag into another ; and the analogy with 
an electrical machine charging two conductors oppo- 
sitely, i.e. pumping electricity from one into the other, 
may be perceived. 

(2) By winds or currents ; by watching the effect 
of moving masses of the fluid as it flows along pipes 
or through spongy bodies, and by the effects of its 
inertia and momentum. A hanging vane arranged 
in a tube, so as to be deflected by a stream of water, 
may be likened roughly to a galvanometer ; also the 
effect of suddenly stopping a stream of water, as in a 
water ram, is somewhat analogous to self-induction. 

(3) By making vortices and whirls in the fluid, and 
by observing the mutual action of these vortices their 
attractions and repulsions. Whirlwinds, sand-storms, 
waterspouts, cyclones, whirlpools, have some analogies 
with an electro-magnet. 

(4) By setting up undulations in the medium : 



16 ELECTROSTATICS CHAP, i 

i.e. by the phenomena which in ordinary media excite 
in us, through our ears, the sensation called " sound'' 

In all these ways we have become acquainted with 
electricity, and in no others that I am aware of. 
They correspond to the four great divisions of the 
subject which I have made above ( i). 

7. But there are differences, at first sight important 
differences, between the behaviour of a material liquid 
ocean, such as we have contemplated, and the be- 
haviour of electricity. 

The chief difference between the behaviour of elec- 
tricity and that of an ordinary incompressible fluid 
comes out in the fourth category that concerned 
with wave-motion. In an incompressible fluid the 
velocity and length of waves would both be infinite, 
and none of the phenomena connected with the 
gradual propagation of waves through it could exist. 
Such a medium therefore would be incapable of sound- 
vibrations in any ordinary sense. On the other hand, 
it is quite certain that the disturbances concerned in 
light-radiation take place at right angles to the direc- 
tion of propagation they are transverse disturbances, 
and such disturbances as these no body with the 
simple properties of an ordinary fluid can possibly 
transmit. Such vibrations can only be transmitted by 
a medium having something akin to the rigidity of a 
solid. We are bound to admit that ether, though 
fluid in the sense of enabling masses to move freely 
through it, has a certain amount of rigidity, at least 
for enormously rapid and minute oscillatory disturb- 
ances. Is there any way of conferring upon a fluid 
such rigidity, without solidifying it ? 



7 FUNDAMENTAL NOTIONS 17 

It can be shown that although a fluid at rest has 
no quasi-solid properties, a fluid in motion acquires 
them. Drive water rapidly through a flexible tangle 
of india-rubber pipe with both ends fixed, and it at 
once becomes semi-rigid. So, utilising that fact, 
imagine every particle of a mass of fluid as in a 
state of violent motion, not locomotion, but motion 
in minute closed curves vortex motion as it is called 
and it too becomes rigid or quasi-solid : it acquires 
the property of transmitting transverse waves. 

The gross analogy of a jelly is occasionally useful. 
A jelly is composed almost wholly of fluid, and yet it 
is rigid. It is rigid because of an elastic skin to each 
particle of water : it is like a multitude of little elastic 
bags of fluid. Distend a thin india-rubber balloon with 
water, and put it on a plate it looks and behaves very 
like a jelly, until collapsed by a needle-prick. Well, 
the effect of an elastic skin, which can thus in a 
manner rigidify stagnant water, can be imitated, much 
more perfectly though not so easily, by setting the 
water into a state of fine-grained vortex motion. All 
kinds of elasticity are to be explained, as Lord 
Kelvin has shown us, by simple motion. (See article 
" Elasticity," Ency. Brit.} Not a simple fluid, but a 
fluid in a violent state of minute spin a vortex- 
sponge, as it has been called is what the ether is 
going to turn out to be. But as this conception is 
difficult at present, we can vaguely say that ether con- 
tains electricity as a jelly contains water ; and that the 
rigidity concerned in its transverse vibrations belongs, 
not to the water in the jelly, but to the mode in which 
it is entangled in its meshes. 

C 



CHAPTER II 

THE DIELECTRIC 

9. RETURN now to the consideration of electro- 
statics. We are to regard ourselves as living im- 
mersed in an infinite all-permeating ocean of perfect 
incompressible fluid (or liquid), as fish live in the sea ; 
but this is not all, for if that were our actual state we 
should have no more notion of the existence of the 
liquid than deep-sea fish have of the medium they 
swim in. If matter were all perfectly conducting, 
such would be our state : in a perfectly free ocean 
there is no insulation no obstruction to flow of 
liquid : it is the fact of insulation that renders 
electrostatics possible. We could obstruct the flow, 
and store up definite quantities of a fluid in which we 
were totally submerged by the use of closed vessels, of 
course. But how could we pump liquid from one 
into another so as to charge one positively and 
another negatively? Only by having the walls 
elastic ; by the use of elastic bags, and elastic parti- 
tions across pipes. And so we can represent a con- 
tinuous insulating medium (like the atmosphere or 
space) by the analogy of a jelly, through which liquid 



9 THE DIELECTRIC 19 

can flow only by reason of cracks and channels and 
cavities. 

Modify the idea of an infinite ocean of liquid into 
that of an infinite jelly or elastic substance in which 
the liquid is entangled, and through which it cannot 
penetrate without violence and disruption ; and you 
have a crude model of the general insulating atmo- 
sphere. Our ocean of fluid is not free and mobile like 
water, it is stiff and entangled-like jelly. 

Nevertheless bodies can move through it freely. 
Yes, bodies can, it is the liquid itself only which is 
entangled. It is not easy to picture freely and natu- 
rally the motion of ordinary matter through the insu- 
lating medium of space, unless we step beyond our 
jelly analogy to the more refined and truer theory of 
the vortex-sponge of perfect liquid ; but it is charac- 
teristic of analogies to break down when pressed too 
far, though they are useful up to a certain point. 

Insulators being like elastic partitions, or impervious 
but yielding masses, conductors are like cavities ; or 
rather like porous or spongy bodies, perfectly pervious, 
though with more or less frictional resistance, to the 
flow of liquids through them. Thus, whereas bodies 
easily penetrable by matter are impervious to elec- 
tricity, bodies like metals, which resist entirely the 
passage of matter, are quite permeable to electricity. 
It is this inversion of ordinary ideas of penetrability 
that constitutes a small difficulty at the beginning of 
the subject. 

However, supposing it overcome, let us think of 
the ordinary lecture-table assemblage of insulated 
spheres and cylinders connected by copper wire, as so 

C 2 



20 ELECTROSTATICS CHAP, n 

many cavities and tubes in an otherwise continuous 
elastic impervious medium, which surrounds them and 
us, and extends throughout space wherever conductors 
are not ; all, however cavities as well as the rest 
of the medium being completely full of the universal 
fluid. 

The fluid which is entangled in insulators is free to 
move in conductors ; whence it follows that its pres- 
sure or potential is the same in every part of a 
conductor in which it is stationary. For if there 
were [any excess of pressure at any point, a flow 
would immediately occur until it was equalised. In 
an insulator there is a tendency to equalisation, but 
it cannot take effect : so differences of pressure are 
exceedingly common in insulators, and are naturally 
accompanied by a strain of the medium. 

It is instructive now to think over a number of 
ordinary electrical experiments from this point of 
view : to think of the fluid as flowing freely through 
conductors and settling down to a state of equilibrium 
or uniform pressure in them; but straining insulators, 
as high-pressure water might strain elastic walls or 
boundaries, straining them even to bursting if the 
partitions be made too thin. 

10. There have been, as you know, two ancient 
fluid theories of electricity the one-fluid theory of 
Franklin, and the two -fluid theory of Symmer and 
others. A great deal is to be said for both of them 
within a certain range. There are certainly points, 
many points, on which they are hopelessly wrong and 
misleading, but it is their foundation upon ideas of 
action at a distance that condemns them, it is not the 



io THE DIELECTRIC 21 

fluidity. They concentrate attention upon the con- 
ductors ; whereas Faraday taught us to concentrate 
attention on the insulating medium surrounding the 
conductors the " dielectric," as he termed it. This is 
the seat of all electrostatic phenomena : from the 
point of view of static electrification, conductors are 
mere breaks in it interrupters of its continuity. 

To Faraday the space round conductors was full of 
what he called lines of force ; and it is his main 
achievement in electrostatics to have diverted our 
attention from the obvious and apparent to the in- 
trinsic and essential phenomena. Let us try and 
seize his point of view before going further. It is 
certainly true as far as it goes, and is devoid of 
hypothesis. 

Take the old fundamental electric experiment of 
rubbing two bodies together, separating them, and 
exhibiting the attraction and repulsion of a pith ball, 
say ; and how should we now describe it ? Something 
like this : 

Take two insulated disks of different material, one 
metal, say, and one silk ; touch them together ; the 
contact effects a transfer of electricity from the metal 
to the silk ; rub slightly to assist the transfer, since silk 
is a non-conductor, then separate. As you separate 
the disks the medium between them is thrown into a 
state of strain, the direction of which is mapped out 
by drawing a set of lines, called lines of force, from 
one disk to the other, coincident with the direction of 
strain at every point. As Faraday remarked, the 
strain is as if these lines were stretched elastic threads 
endowed with the property of repelling each other as 



22 



ELECTROSTATICS 



CHAP. II 



well as of shortening themselves ; in other words, there 
is a tension along the lines of force and a pressure at 
right angles to them. When the disks are near, and 
the lines short, they are nearly straight (Fig. i), but as 



FIG. i. Rough diagram of the state of the medium near two oppositely charged 
disks when close together. 

the distance increases they become curved, bulging 
away from the common axis of the two disks and some 
even curling round to the back of the disk (Fig. 2), 
until, when the disks are infinitely distant, as many 




FIG. 2. Rough diagram of the state of the medium near two oppositely charged 
disks when separated. 

lines spring from the back of each as from its face ; 
and we have a charged body to all intents existing in 
space by itself. 



ii THE DIELECTRIC 23 

The state of tension existing in the medium between 
the disks results in a tendency to bring them together 
again, just as if they were connected by so many 
elastic threads of no length when unstretched. The 
ends of the lines are the so-called " electrifications " or 
" charges," and the lines perpetually try to shorten and 
shut up, so that their ends may coincide and the strain 
be relieved. If one of the disks touch another con- 
ducting body, some of its lines instantly leave it and 
go to the body ; in other words, the charge is capable 
of transference, and the new body is urged towards 
the other disk, just as the disk was from which it 
received the lines. If this new body completely sur- 
rounds the disk, it receives the whole of its lines, and 
the disk can be withdrawn perfectly free and inert. 
This is Faraday's " ice pail" experiment ( 4). 

n. Now take the two charged disks, facing one 
another, and let, say, a suspended gilt pith ball hang 
between them. Being a conductor there is no strain 
inside it, and so it acts partially as a bridge, and 
several of the lines pass through it or, rather, they 
end at one side of it and begin at the other : thus it 
has opposite charges on its two faces it is under 
induction (Fig. 3). Let it now be moved so as to 
touch one of the disks, the lines between it and the 
disk on that side have shut up, and it remains with 
those only which go to the other disk. Besides these, 
it receives, unbalanced, some of the lines which pre- 
viously belonged to the touched disk. Its line of force 
now will pull it over to the far disk and there shut 
themselves up. From that disk it receives more ; 
and it travels, with them, back to the first disk, 



ELECTROSTATICS 



CHAP. II 



and so on (Fig. 4), perpetually receiving lines and 
shutting them up, until they are all gone and the 
disks are discharged. The experiment is easily 
performed, and illustrates "discharge by alternate 
contact." 

This mode of stating the facts involves no hypo- 
thesis whatever it is the simple truth. But it may 
at once be admitted that " lines of force " have no 




FIG. 3. Rough diagram of the medium between two disks disturbed by the 
presence of an uncharged metal sphere. The two halves of the sphere are 
oppositely charged " by induction." 

more and no less existence than have " rays of light." 
Both are convenient modes of expression." 

12. But so long as we adhere to this mode of 
expression we cannot form a complete mental picture 
of the actually occurring operations. In optics it is 
usual to abandon " rays " at a certain stage and attend 
to " waves," which we know are of the essence of the 
phenomenon though we do not yet know very much 
about their true nature. 

Similarly in electricity, at a certain point we are 
led to abandon lines of force and potential theories, 



12 



THE DIELECTRIC 



and try to conceive the actual stuff undergoing its 
strains and motions. It is then that we get urged 
towards ideas similar to those which are useful in 
treating of the behaviour of an incompressible fluid. 
In a much-modified sense, we have still a fluid 




FIG. 4. Rough diagram of the medium near two oppositely charged disks between 
which a metal carrier ball is oscillating, having just touched the right-hand disk. 
(Discharge by "alternate contact.") 

theory of electricity, and a portion of the ideas of the 
old theories belong to the new theory also. 

Thus Franklin's view that positive charge was 
excess, and negative charge was a deficit, in a cer- 
tain standard quantity of the fluid which all bodies 
naturally possessed in their neutral state, remains 
practically true : except that we have reason to wish 



2 6 ELECTROSTATICS CHAP, n 

that he had happened to interchange the nomenclature ; 
since it is what he called negative that turns out to be 
the excess. But too much confusion would result from 
trying to change now. His view that the fluid was 
never manufactured, but was taken from one body to 
give to another, so that one gained what the other lost 
no more and no less remains practically true. Part 
also a less part of the two-fluid theory likewise 
remains true, in my present opinion ( 90) ; but this 
is not a branch of the subject on which I shall enter 
in the present part. It will suffice for the present to 
fix our attention on one fluid only. 

You are to think of an electric machine as a pump 
which, being attached to two bodies respectively, 
drives some electricity from the one into the other, 
conferring upon one a positive and upon the other a 
precisely equal negative charge. One of the two 
bodies may be the earth, in which case the charge 
makes little or no difference to it. 

13. But, as has been objected before, if electricity 
is like an incompressible and inextensible fluid, how 
is it possible to withdraw any of it from one body 
and give it to another ? With rigid bodies it is not 
possible, but with elastic bodies it is easy. 

The act of charging a sphere is therefore analogous 
to pumping water into an elastic bag ; or rather into 
a cavity in the midst of an elastic medium, whose 
thick walls, extending in all directions and needing a 
great pressure to strain them, better represent the true 
state of the case than does the thin boundary of a 
mere bag. 

Draw a couple of such cavities and consider fluid 



i 3 THE DIELECTRIC 27 

pumped from one into the other, and you will see that 
the charge (i.e. the excess or defect of fluid) resides on 
the outside. The interior was full and remains full. 

If the fluid is exactly incompressible, not the least 
extra quantity will be squeezed by the pressure into 
the space originally occupied by the cavity. This is 
the moral of the Cavendish experiment ( 4) : it 
proves that electricity is precisely incompressible ; a 
fact which is further explained in 14 A. 

You may also show that when both cavities are 
similarly charged, the medium is so strained that 
they tend to be forced apart ; whereas when one is 
distended and the other contracted they tend to 
approach. 

Further you may consider two cavities side by side ; 
pump fluid into (or out of) one only, and watch the 
effect on the other. You will thus see the phenomena of 
induction, the near side of the second cavity becoming 
oppositely charged (i.e. the walls encroaching on the 
cavity), the far side similarly charged (the cavity 
encroaching on the wall), and the pressure on the 
fluid in the cavity being increased or diminished in 
correspondence with the rise or fall of pressure in the 
charged or inducing cavity. In other words, con- 
ductors rise in potential when brought near a 
positively charged body ; and their charge, if any, 
though not altered in quantity becomes redistributed. 

The actual changes in volume necessary to the 
strain of these cavities are a defect in the analogy. 
To avoid this objection, one will have to accept a dual 
view of electricity a sort of two-fluid theory, which 
many phenomena urge one to contemplate ; for the 



28 ELECTROSTATICS CHAP, n 

two constituents of ether may then be sheared past 
each other one increasing as the other decreases in 
the charged cavity setting up a shearing strain or dis- 
tortion which involves no expansion or contraction in 
volume, and leaving the total quantity of ether every- 
where unchanged (cf. 18). 

14. Return Circuit. Sometimes a difficulty is felt 
about electricity flowing in a closed circuit as, for 
instance, in signalling to America and using the earth 
as a return circuit : the question arises, How does the 
electricity find its way back ? 

The difficulty is no more real than if a tube were 
laid to America with its two ends connected to the 
sea and already quite full. If now a little more sea- 
water were pumped in at one end, an equal quantity 
would leave the other end, and the disturbed level 
of the ocean would readjust itself. Not the same 
identical water would return, but an equal quantity 
would return. That is all one says of electricity. 
One cannot label and identify electricity. 

To imitate the inductive retardation of cables, the 
tube should have slightly elastic walls ; to imitate the 
speed of signalling, the water must be supposed quite 
incompressible, not elastic as it really is, or each 
pulse would take three-quarters of an hour to go : 
even through a perfectly rigid pipe. 

14 A. Incompressibility of Electricity. The equiva- 
lence of the different modes of expression at the 
beginning of 4 is, I find, not obvious. It seems 
worth while to explain these fundamental matters 
at greater length. One of the fundamental experi- 
mental facts concerning electricity is that it is impos- 



i 4 A THE DIELECTRIC 29 

sible to charge an insulated chamber by performing 
electrical experiments in its interior. Electrical 
machines inside a conducting chamber and electro- 
scope outside, or vice versa, are perfectly independent 
of each other. This may be expressed by saying that 
the field of force outside a conductor, and the field 
of force inside, are quite independent ; that there is 
entire dislocation of every line of force at the con- 
ductor ; that a closed conductor acts as a perfect 
electrostatic screen. A second fundamental experi- 
mental fact is that if a charge be in any way 
communicated to, or even passed into the interior of, 
a hollow conductor, the same precise quantity at once 
appears on the outside ; and a variant of this is that if 
the outside skin of a conductor be peeled off, all charge 
is removed with that outer skin : no trace remaining 
below it. 

These two facts lead us to say that electricity 
behaves like an incompressible fluid of which all 
space is precisely full ; a fluid moreover which cannot 
be manufactured but only moved about. For think of 
a closed chamber in the sea with pumps inside it. 
So long as there is no communication through the 
walls, all that the pumps could do would not result in a 
trace more liquid in the vessel than there was at first. 
The vessel cannot be thus charged. But now suppose 
some water to be pumped from the outside into the 
interior, through a suitable opening ; how could it be 
done ? The vessel being already by hypothesis quite 
full, either the water must be compressible, in which 
case the extra charge would exist uniformly through- 
out the enclosed space ; or the vessel must be expan- 



30 ELECTROSTATICS CHAP, il 

sible or porous, in which case the extra charge is not 
found in the original space at all, but outside it. 
Whatever extra quantity was pumped into the original 
space, that same amount precisely would instantly 
appear outside it ; and this is what happens with 
electricity '( 4). 

In order to examine whether water is incompressible 
or not we might perform this experiment : Inside a 
metal cavity provided with a pressure-gauge, introduce 
a strong elastic bag, and distend it by pumping water 
into it (Fig. 10). As the bag swells, the pressure in the 
outer vessel will rise, unless it is continually relieved 
(by connexion to " earth "). Having finished pumping 
into the bag, close any leak the outer vessel may have 
had, examine the gauge carefully, and then puncture 
the bag inside. The water in the bag suddenly escapes 
from pressure, and, if the slightest degree elastic, will 
increase the pressure in the outer rigid vessel and thus 
will cause a sudden small rise of the gauge. If it 
be incompressible the gauge will indicate nothing. 
Nothing will happen except the release of tension in 
the bag : the water will indicate nothing whatever, 
unless it had been actually compressed. This is 
precisely Faraday's ice-pail experiment ; and the 
result teaches us that electricity is perfectly incom- 
pressible. 

The Cavendish experiment where the skin is 
removed from a charged conductor, which is then 
tested and found exactly uncharged is another more 
stringent but more difficult mode of doing the same 
thing. If a real positive result could be got by either 
of these experiments, it would constitute a measure 



i4A THE DIELECTRIC 31 

of the compressibility or volume-elasticity of the 
ether. It would show that longitudinal thrusts (such 
as those of gravitation may be) do not exist instant- 
aneously throughout space, but are propagated in 
time at a certain velocity. 

It is easy to see that the facts related above carry 
with them such mere technical expressions as that 
total " induced charge " is always equal and opposite 
to "inducing charge "; because " induced charge" 
merely represents the condition of the interior surface 
of a chamber into which an inducing charge has 
been introduced : and, an equal quantity having been 
thereby extruded, the inner surface of the wall must 
have been left negatively charged by that precise 
amount. 

Furthermore, the incompressibility of electricity 
necessitates its flow always in a closed circuit. In 
most cases it is obvious ; but take a case where it is 
not obvious. When a charge of electricity is moved 
from place to place, as for instance from outside to 
inside a room, it is not thereby piled up in the room 
and withdrawn from elsewhere; for an equal quantity 
instantly goes through the walls and appears outside. 
It has not been exhausted from one part of space 
and condensed in another. So even in this case, it 
has moved in a closed circuit, and has behaved pre- 
cisely as an incompressible fluid filling all space must 
behave. That is what it does always. 

Thus the equivalence of the statements at beginning 
of 4 are clear. 



CHAPTER III 

CHARGE AND INDUCTION 

15. Condensers. Returning to the subject of charg- 
ing bodies electrically, how is one to consider the 
fact that bringing an earth-plate near a conductor 
increases its capacity so greatly, enabling the same 
pressure to force in a much larger quantity of fluid ? 
how is one to think of a condenser, or Leyden jar ? 

In the easiest possible way, by observing that the 
bringing near an earth-connected conductor is really 
thinning down the dielectric on all sides of the body. 

The thin-walled elastic medium of course takes 
less force, to distend it a given amount, than a thick 
mass of the same stuff took ; in other words a cavity 
enclosed by thin walls has much more capacity 
than if its walls were thick. Remember that capacity 
of elastic cavities cannot satisfactorily be measured as 
the capacity of buckets is measured, by the maximum 
quantity they will hold when full : elastic bags are 
never " full," till they burst ; and the amount required 
to burst them measures rather their strength than their 
capacity. The only reasonable definition of capacity 
in such cases is the ratio which any addition to their 



16 CHARGE AND INDUCTION 33 

contents bears to the extra pressure required to force 
it in : and this is exactly the way electrical " capacity " 
is, defined. A Leyden jar is like a cavity with quite 
thin walls in other words, it is like an elastic 
bag. 

But if you thin it too far, or strain it too much, the 
elastic membrane may burst ! Exactly, and this is the 
disruptive discharge of a jar, and is accompanied by 
a spark. Sometimes it is the solid dielectric which 
breaks down permanently. Ordinarily it is merely 
the air ; and, since a fluid insulator constitutes a 
self-mending partition, it is instantaneously as good 
as new again. 

There are many things of interest and importance 
to study about a Leyden jar. There is the fact that 
if insulated it will not charge : the potential of both 
inner and outer coatings rises equally. In order to 
charge it, for every positive spark given to the interior 
an equivalent positive spark must be taken from 
the exterior. There is the charging and the discharg- 
ing of it by alternate contacts, as by an oscillating 
ball ; and there are the phenomena of the spark- 
discharge itself, 

But, as a matter of fact, all charging is really re- 
presented by the case of a Leyden jar. The outer 
coat must always be somewhere the walls of the 
room, or the earth, or something you always have 
a layer of dielectric between two charges the so- 
called induced and the inducing charge. You cannot 
charge one body alone ( 5). 

16. To illustrate the phenomena of charge, I will 
now call your attention to these diagrams- which 

D 



ELECTROSTATICS 



CHAP. Ill 



less completely but more simply than hydraulic illus- 
trations, serve to make the nature of the phenomena 
manifest 

Fig. 5 shows an inextensible endless cord circu- 
lating over pulleys ; this is to represent electricity 
flowing in a closed circuit. Electromotive forces are 




FIG. 5. Mechanical analogy of a metallic circuit. The cord corresponds to 
electricity, the weight w to an electromotive force, and the perforated blocks 
to atoms of matter. 



forces capable of moving the cord, and you may 
consider them applied either by a winch, or by a 
weight on the hook W. A battery cell corresponds to 
a small weight ; an electric machine to a slow but 
powerful winch. Clamping the cord with the screw s 
corresponds to making the resistance of the circuit 
infinite. Instead of the cord, clamp, and driving 



i6 



CHARGE AND INDUCTION 



35 



pulley, one might consider an endless pipe full of 
liquid, with a stop-cock and a pump on it, but for 
many purposes the cord is sufficient and more simple. 
In Fig. 5, the only resistance to the motion is friction, 
and there is no tendency to spring back. The cord 
passes freely through fixed blocks or beads, which are 
to typify atoms of matter ; and they may be more 




FIG. 6. Mechanical analogy of a circuit partly dielectric : for instance, of a 
charged condenser. A is its positive coat, B its negative. 



or less rough, to represent different specific resist- 
ances. If the cord be moved, heat is the only result. 
Now pass to Fig. 6. Here the cord is the same as 
before, and still represents electricity in a closed cir- 
cuit, but the beads are now firmly attached to the cord, 
so that if it moves they must move with it. They repre- 
sent, therefore, the particles of an insulating substance. 
Nevertheless, their supports are not rigid they do 
not prevent the cord moving at all ; they allow what 

D 2 



36 ELECTROSTATICS CHAP, in 

is called electric " displacement? not conduction : 
they can be displaced a little from their natural 
position, but they spring back again when the 
disturbing E.M.F. is removed. The beads in this 
figure are supposed to be supported by elastic 
threads. If we used the analogy of a closed pipe 
full of water, the beads and threads would be replaced 
by elastic partitions. The specific inductive capacity 
of the dielectric is represented by the stretchability, or 
inverse elastic resilience, of the elastic threads. The 
stiffer the threads are to pull out, the less is the in- 
ductive capacity of the medium ; because evidently a 
greater E.M.F. is needed to cause a given displace- 
ment. 

Apply a given E.M.F. to this cord, as for instance 
the weight W, and a definite displacement is produced. 
One side, A, gets more cord than usual it is 
positively charged ; the other side, B, gets less it is 
negatively charged. If the applied E.M.F. exceeds a 
certain limit, the strain is too great; the elastics break, 
and you have disruptive discharge with a spark. But 
even when the strain is only moderate some of the 
supports may yield viscously, or be imperfectly elastic, 
and permit a gradual extra displacement of the cord, 
known to telegraphists as " soaking in " ; which is 
explained more fully in the next section. 

When, after this process, discharge is allowed, it is 
not at once complete ; a large portion of the displace- 
ment is at once recovered, but the rest gradually 
" soaks out " and causes residual discharges. 

If the dielectric is at all stratified in structure, so 
that some of the beads allow cord to slip through 



i; CHARGE AND INDUCTION 37 

them or yield more than others then this residual 
charge effect will become very prominent. 1 

1 7. These are matters which it is easy thoroughly to 
understand, and Fig. 7 illustrates different stages 
sufficiently. In Diagram I. are represented 8 strata, 
each displaced from its normal position by an amount 
3. The restoring force is proportional to the displace- 
ment, so the total restoring force can be called 24. 
The diagram represents, therefore, a Leyden jar or 
other dielectric strained by an applied E.M.F. of 24 
units. If every stratum insulates perfectly that is, if 
every bead is quite firmly attached to the cord nothing 
further happens, so long as the force is kept applied. 
This state of things may be maintained in two ways : 
either by keeping on the weight W that is, by keeping 
the condenser permanently connected to the battery ; 
or by clamping the cord and thus making the resistance 
infinite that is, by insulating the terminals of the 
condenser. 

But now suppose that some of the strata are not 
perfectly insulating ; let some of the beads slowly slide 
back along the cord towards their zero position. Then 
we shall witness different phenomena according to 
whether the weight has been left on or whether the 
cord has been clamped. 

Take first the case of leaving the weight on that 
is, keeping the battery connected. If every bead 
slides equally, all we get is a continuance of the state 
represented in Diagram I., combined with a slow oozing 

1 For the original details of these cord models as illustrative of 
Maxwell's theory of electrostatic induction, &c., see Lodge, Phil. Mag. 
November 1876. 



38 ELECTROSTATICS CHAP, in 

forward of the cord that is, a slow and steady leak 
through the condenser. But suppose every bead does 
not slide equally, suppose some do not slide at all ; 
then the slipping of some throws extra strain on the 
others, and the cord moves forward, but more and more 
slowly, until the insulating strata ultimately have to 
bear all the strain, and the cord asymptotically comes 
to rest. 

This process is observed in Atlantic cables and 
Leyden jars ; all condensers are liable to it except 
air condensers, and it is called " soaking in " ; it is 
accompanied by the development of internal charges, 
because plainly the original normal length of cord 
between the beads is no longer maintained ; some 
strata have acquired an extra length that is, are 
positively charged others are negatively charged. 

The strain is distributed very unequally, but its 
total amount, in this case, continues constant. 

Remove the electromotive force now, keeping the 
circuit still closed: i.e. short-circuit the* condenser. We 
get first a quick discharge ; and then a slow leaking 
out or reverse motion of the cord, as it is propelled by 
the still displaced insulating strata through the now 
oppositely displaced badly conducting ones ; the 
time of A soaking out" being comparable with that 
permitted to the soaking in. 

It is important to see how these phenomena are en- 
tirely reconcilable with the incompressible character of 
electricity that is, the unyielding character of the cord. 
So much for the case when the battery is left 
connected ; now attend to the case when the terminals 
are insulated. Refer to Fig. /. 



17 



CHARGE AND INDUCTION 



39 



Having got the dielectric into the state represented 
by I., screw down the clamp and wait. If some of 
the beads slide, while some do not, we shall shortly 




FIG. 7. Stages in the discharge of a stratified condenser, with some of its layers of 
imperfect insulating power ; showing one way in which the phenomena of 
"residual charge,"" internal charge," and '' soaking out " areproduced. Tension 
in the cord, measured by the right-hand numbers, represents electric potential. 

I. represents a recent charge, of E.M.F. 24. 

II. represents the same after lapse of time, reduced to 17 by partial internal 
leakage, and shows internal charge. The circuit itself is supposed to have 
been perfec;ly insulating all the time ; the charge on the plates therefore 
remains constant. 
Til. shows the fir^t discharge. 

IIII. shows the s'ate attained after again waiting, viz. a residual charge with an 
E.M.F. 3 in the old direction. 

arrive at the state represented by II. Beads Nos. 3, 
5, and 6 have slid partially back, and the total stress 
on the cord has been reduced to 17. The jar will 



40 ELECTROSTATICS CHAP, in 

appear to have partially discharged itself by internal 
leakage, and yet not the slightest motion of the cord 
has been permitted. Internal charges have appeared : 
positive between Nos. 3 and 4, and between 5 and 6, 
negative between 2 and 3, and between 4 and 5. The 
charges on the coatings at A and B have remained 
constant ; the jar has apparently increased in capacity, 
because the same charge is maintained by a less 
electromotive force. All these effects may present 
themselves at first sight as irreconcilable with the 
behaviour of an incompressible fluid, or cord of constant 
length ; but the diagram clearly says otherwise. 

Now unclamp the cord momentarily, i.e. discharge 
the condenser, and insulate again. At the instant of 
discharge a rush of electricity takes place, and the 
force falls to zero. The state of the discharged jar 
at the first instance after discharge is represented in 
III. fig. 7. The surface charges have not wholly 
disappeared ; the internal charges have been un- 
affected ; the displacement of none of the strata is 
zero. The insulating ones remain with some of their 
original displacement, the leaking ones have been 
forced into a position of inverse displacement, so as 
just to reduce the resultant force on the cord to zero. 
The most slippery, Nos. 3 and 5, have been most dis- 
placed in the inverse direction. But not long do they 
thus remain. They at once begin slowly to ooze 
back ; and before long they will have got into the 
state represented by II II, where the now almost un- 
balanced stress of the insulating strata exerts on 
the cord a force 3 in the original direction. This is 
known as the " residual charge," and on unclamping 



iS CHARGE AND INDUCTION 41 

the cord, the first residual discharge can be obtained. 
Not even yet, however, is the jar wholly discharged. 
Waiting again, another but feebler residual charge 
makes its appearance ; and so on, almost without limit, 
until the sum of all the residual discharges plus the 
original discharge make up exactly the charge origin- 
ally imparted to the jar make it up exactly if any one 
of the strata has declined to leak. If all leaked more 
or less, then of course there will be some deficiency. 

1 8. The only thing needful to guard against in 
following out this mechanical analogy is the idea 
that there need be any mechanical displacement of 




FIG. 7 A. Closer representation of electric displacement or shear. Cords alter- 
nately representing positive and negative electricity. Cf.aalso Fig. 46. 

the atoms of matter accompanying the electric dis- 
placement. Manifestly a model containing fixed 
beams for the attachment of the beads cannot really 
correspond very closely to electrical facts. To make 
the model more closely imitative of facts we should 
have to take a number of rows of beads, each row 
threaded on its cord and attached crosswise by elastic 
threads, as in Fig. /A. 

If these cords are simultaneously displaced alter- 
nately in opposite directions, and if they be con- 
sidered as representing positive and negative elec- 
tricity alternately, while the beads represent the 
electro-positive and the electro-negative elements of 



42 ELECTROSTATICS CHAP, in 

the material substance, then perhaps something 
more like the actual state of things may be imagined. 
There is here no displacement of the molecule as 
a whole, but there is a displacement of its con- 
stituent atoms ; there is a shearing stress applied to 
each molecule, which, if strong enough, may result in 
electrolytic disruption. I certainly regard disruptive 
discharge as being of this electrolytic character 
( 112 and 65 A ; see also 13). 

19. Return, however, to the simple discharge, and 
see how it occurs. Will it take place as a simple 
sliding back of the beads to their old position? Yes, 
if the resistance of the circuit is great ; but not other- 
wise. If the cord is fairly free the beads will fly 
past their mean position, overshooting their mark, 
then will rebound, and so, after many quick oscilla- 
tions, finally settle down in their natural position. 
Thus is represented the fact that the discharge of a 
Leyden jar is in general oscillatory. The apparently 
single and momentary spark, when analysed in a very 
rapidly rotating mirror, turns out really to consist 
of a series of alternating flashes rapidly succeeding 
one another, and all over in the hundred-thousandth 
of a second or thereabouts. These oscillatory cur- 
rents were predicted and calculated beforehand by 
Lord Kelvin ; they were first observed experi- 
mentally by Feddersen. 1 The oscillations continue 
until the energy stored up in the strained medium 
has all rubbed itself down into heat. The existence 
of these oscillations proves distinctly that electricity 

1 We now find that they were experimentally discovered with con- 
siderable clearness by Joseph Henry of Washington. (See Lect. III.) 



20 



CHARGE AND INDUCTION 



43 



in conjunction with matter possesses inertia. The 
rapidity of these oscillations is something tremendous : 
it may reach as high as a hundred million per second, 
or it may be as slow as ten thousand per second, 
according to the capacity and inertia of the circuit. 
With special arrangements, the above range of fre- 
quency for the alternations can be far exceeded both 
ways. 1 

The rapidity of oscillation, and its rate of dying 
away, as well as the circumstances which change the 




FIG. 8. Partial model of a dielectric. 

recovery from an oscillatory one into a simple one- 
directioned leak, are precisely analogous to those 
which regulate the recovery of a bent spring suddenly 
let go. If the spring is in a very viscous medium, or 
if it has but small inertia, it will not oscillate, but 
will merely return to its normal position. Under 
ordinary circumstances, however, it will make many 
oscillations before its energy is all rubbed into heat 



Fig. 8 shows part of a first actual model of the kind. 

20. To make the model represent charge by induc- 

tion, all that has to be done is to immerse a conductor 

1 Lodge, Proc. R. /. March, 1889, and Phil. Mag. April and July 



44 ELECTROSTATICS CHAP, in 

into the polarised dielectric ; in other words, to make 
one or more of the beads of the fixed and slippery 
conducting kind, the other beads on the cords being of 
the elastic and adhesive or insulating kind. Then, 
when displacement occurs, it is plain that a defi- 
ciency of cord will exist on one side of the metallic 
layer and a surplus on the other, as shown in Fig. 9. 
This state of things corresponds exactly to the equal 




FIG. 9. Metallic layer in the midst of a polarized dielectric, showing opposite 
charges " induced " on its surfaces. (Compare Fig. 3.) 

opposite induced charges on a conductor under in- 
duction, as in Fig. 3. 

An improved model of this kind is represented in 
Fig. 67. 

If the strain on one side be relieved by letting the 
beads on that side slip back on the cord, that corre- 
sponds to touching the conductor to earth, as in Fig. 
4. The other side has now to withstand the whole 
E.M.F., consequently the strain there and the charge 
there will have increased. Remove now the applied 
E.M.F., and the negative charge appears on both 
sides of the metal partition, either equally, or more 
markedly on that side which has fewest beads, i.e. 
which is nearest to other conductors. 

21. This being a matter which it is desirable 
thoroughly to understand, a series of figures illus- 



21 



CHARGE AND INDUCTION 



45 



trating the various stages of the process are appended 
in Fig. QA. 




B 



60 



II 



A 



\ I M N 



Till 



'-4 



O 



\ \ 



n v 



B 



60 









\ 




\ 


\ 




\ 




1 




















A 




\ 




\ 




\ 




\ 




I 












hB 










\ 




\ 




V 




\ 


V. 
















60 


m H 
3 


+ 


6 




I 




/ 




/ 




A 






<. 








( 


i 










/ 




/ 




/ 




/ 




























/ 




/ 




/ 




/ 

















































/ 




/ 








\ 




















( 






/ 




/ 


































/ 




/ 








4- 


2 
















-2 






-4 


\ 




\ 





4 








1 


























\ 


































\ 




\ 






















1 

























FIG. gA. Stages during the charge of a metal by induction and contact. C is a 
metallic stratum introduced into a dielectric between the opposing electrified 
plates, A and B, which are kept connected to a constant source, till one side 
has been discharged. Numbers preceded by + or - represent charges; 
numbers affixed to an arrow-head represent E.M.F. In this series the E.M.F. 
applied is supposed constant. To represent the other case, when the charge on 
the A plate is kept constant, the cord must not Joe allowed to move ; but the 
force on it will then decrease, from 60 in I. to 45 in II. and to 30 in III. 

I. represents an ordinary polarized dielectric, say 
air, between two oppositely charged bodies, A and B, 



46 ELECTROSTATICS CHAP, in 

maintained at constant difference of potential. For 
simplicity the field is taken of uniform strength, i.e. 
with its lines of force parallel straight lines, so that A 
may be considered as a large positively charged plate, 
and B an earth-plate facing it. The difference of 
potential between A and B is called 60, and is dis- 
tributed among 8 strata or units of thickness, each of 
which therefore bears the strain 7|, and is displaced 
f of the width of a square from its normal position. 
The charges on A and B we may call 3 re- 
spectively. 

An insulated metal plate, C, two units in thickness, 
is now introduced, replacing a couple of the dielectric 
strata. The remaining 6 have therefore more strain 
thrown upon them, viz. 10 on each, and accordingly 
each is now displaced a whole square-width from its 
normal position, the charge of the plates A and B has 
risen to + 4, and the effect is the same as if the thickness 
of dielectric had been reduced in the ratio of 8 to 6. 
The metal partition introduced has also a charge on 
the surface, viz., 4 on the side facing A, and +4 on 
the side facing B. See Diagram II. in Fig. QA. 

The next stage is to connect the metal momentarily 
to earth. The effect of this is entirely to relieve the 
strain on the B side, by replacing the dielectric with 
metal, which allows the cord to slip through freely. The 
cord makes another bound forward, and all the strain is 
now thrown upon 4 strata, which each have to bear 15, 
and are displaced i| from their natural position. Re- 
storing the dielectric (i.e. removing the temporary earth 
connexion) makes no further change, but leaves every- 
thing as shown in Diagram III. The charge on one side 



22 CHARGE AND INDUCTION 47 

of the metal partition is now 6, and on the other side 
is nothing. 

Finally remove the constant E.M.F. which has been 
acting all this time. The cord makes a bound back, 
the resultant force on it becomes nothing ; the 2 strata 
on the right have to balance the 4 strata on the left, 
and accordingly their displacements are I and | 
respectively. The charges on the faces of the partition 
are 2 and 4; both negative. The charges on A 
and B are +2 and +4 respectively, although they are 
at the same potential. The state of things is shown in 
IV., which exhibits the metal partition as completely 
charged negatively by means of induction. Of course 
it may have been charged equally on its two faces, 
but not necessarily ; that is a mere matter of the 
relative proximity of adjacent objects, A and B. 

If instead of maintaining A at constant potential, 
it possessed a constant charge, the series of operations 
would differ in a slight and easily appreciated manner. 
The resultant tension on the cord would then decrease, 
both when the thick metal layer was introduced, and 
when it was touched to earth ; but we may now regard 
the series of operations as practically the electrophorus 
series, such as go on rapidly and continuously in all 
inductive machines and replenishers. It will be worth 
while to sketch this electrophorus series more par- 
ticularly ; the process of working out what is happening 
in any given case will then be sufficiently illustrated. 

Electrophorus. 

22. Diagram I., Fig. 96, shows the cake excited 
negatively, resting on its sole. The negative charge 



ELECTROSTATICS 



CHAP. Ill 



on surface of cake is called 13 units ; of these, 12 
are what is sometimes called " bound " by the sole, 
and i is free. In other words, the strain due to 12 



(Ceiling 


) 




A- 











(( 


)) 








(- 


2) 








(- 


4) 






































































^^ 


_- - 1 




^ 




^ 


^ 


\ 


\ 


Atmosphere 




































+ 


4 




























- 


. ' 


-, 


& 






































+ 


2 


















































+ 


6 










-j 




























s 






/ 


Lowest position 
of Cover J 


( 
























+ 


8 








^ 


"P 




























\ 






/ 




\ 






/ 


Air film 
















+ 1 


O 








\ 


/ 








X, 


/ 












-i 


3 




\ 






/ 




\ 






/ 




\ 






/ 








/ 


\ 








A 


/ 








5 


/ 








\, 


/ 








/ 


/ 


\ 


. 






'v 


-I 










-i 


3 






/ 


\ 








/ 


/ 


\ 


\ 




^ 


,-^- 


*\. 


^ 




X 


^ 


^ 


N, 




/ 






\ 


Cake< 




/ 


/ 


\ 


N 


























/ 


\ 








/ 


/ 


\ 


\ 




^ 


^ 


'-^ 


^ 




X 


^ 


> 


\ 




/ 






\ 






t 






\ 


























/ 


\ 








/ 






\ 




^ 


^ 


*\. 


^ 




X 


^ 


x 


X 




/ 






^\ 








+i 


2 








-f- 


3 








+ 


5 








+ 


7 




Sole 




































































































I 


I 








i] 


[I 








I 


V 





FIG. QB. The electrophorus series of operations. Double lines mean rigid rods 
supporting smooth beads, and represent metal. Gripping beads supported by 
elastics represent dielectric. Numbers preceded by + or - represent charges. 
The charge excited on the surface of the electrophorus cake remains constant all 
the time. The sole is supposed connected with the floor ^and walls of the room. 
Numbers in parentheses at the top represent the charge induced on ceiling and 
walls. The whole thickness of the atmosphere does not pretend to be represented. 
It must be thick enough in I. precisely to balance the stress in the cake. The 
resultant force on the cord is in each case zero. 

I. shows the charged cake, with cover either off or in position but insulated. 
II. shows effect of connecting cover and sole. 
III. and IV. show effect of gradually raising cover. 

I. again shows effect of having removed cover completely and discharged it. 

units of charge is thrown on the layers of the cake, 
the remaining small strain is thrown on the atmo- 
sphere above. The strain in the atmosphere is small 



22 CHARGE AND INDUCTION 49 

because it is so much thicker than the cake there 
are so many layers in it that a very small displace- 
ment of each suffices to balance the stress in the 
cord. One unit of charge is induced on the ceiling 
and walls of the room by the electrified cake. 

We now bring down the insulated metal cover of 
the electrophorus. If of any appreciable thickness it 
displaces a few of the strained layers, and thus there 
is a little extra strain on the others ; but this effect is 
extremely small, and it is quite unessential. We 
may therefore take the cover as of no thickness, and 
bring it down into what is marked in the diagram as its 
lowest position ; the stress passes through it, and 
nothing is affected except the one layer whose place 
it takes. Diagram I. will serve to represent the cover 
thus put on, so long as it is insulated. The dotted 
lines show it in position. It does not make intimate 
contact with the cake ; a film, either of air or of the 
substance of the cake itself, intervenes between it and 
the negatively charged surface, and this is exhibited 
in the diagram. 

The next thing is to connect the cover and the 
sole together. This immediately brings about the 
state of things represented by Diagram II. 

A charge of 9 units has rushed round from sole to 
cover, making, with the charge I which previously 
existed on the walls of the room, a total of IO. 1 The 
strain above the cover is entirely relieved, and the 
whole excitement is now internal, between cover 

1 If the sole had been insulated, and connexion between it and cover 
also made in an insulated manner, then this unit on the walls of the 
room would stay there : the cover would only acquire a charge 9, and 
the slight strain above it shown in I. would continue to exist unaltered. 

E 



50 ELECTROSTATICS CHAP, in 

and sole. The strain in the cake is considerably 
relieved, but the work of balancing what remains 
is thrown on the very thin film between cover and 
top of cake. This, therefore, is highly strained. 

We now raise the again insulated cover. As it as- 
cends, fresh layers of dielectric intervene between it 
and the cake, and receive some of the strain. The 
effect of this is threefold. First they partially relieve 
the strain in the original very thin layer ; next, they 
increase the strain in the cake ; and thirdly, they 
put a little strain on the air above the raised cover. 
The sole therefore receives 5 units instead of 3 ; the 
cover retains its charge 10, but part of this is on its 
upper surface ; an induced charge 2 makes its 
appearance on the walls of the room. This state of 
things is shown in Diagram III. 

Diagram IV. continues the process of raising, 
until ultimately when the plate is removed to 
infinity ; its charge above and below is then equal, 
being 5 on each, and the cake and cover have 
returned to their original state I., ready to begin 
again. The cover having now a charge 10, the 
walls of the room, wherever the cover is, will have a 
charge 10; and it may be discharged to the walls, 
or to anything connected with them, whenever we 
please, without affecting the cake at all. Having dis- 
charged it, we can put it on, as in I., and perform the 
cycle again. 

If one chooses to put the cover on before dis- 
charging it, the cycle of operations is just reversed, 
from IV. to II. 

It is instructive to mount an electrophorus on an 



23 CHARGE AND INDUCTION 51 

insulating stand, and connect its sole to earth through 
a delicate galvanometer ; then the rush out of it when 
the cover is touched, and the flow back again as the 
cover is raised, can be easily watched. 

23. There is one more thing which is so important 
to see clearly that an illustration of it is desirable ; 
and that is the effect of inserting, not a metal, but 
a slice of some other perfectly insulating dielectric, 
with a different inductive capacity, in the midst of a 
polarized medium. Thus, for instance, between the 
plates of a charged condenser insert a thick slab of 
glass. The effects will differ according as the con- 
denser plates are charged each with a given quan- 
tity, or are maintained at a constant difference of 
potential. 

Refer to Fig. gc ; the 8 similar strata are supposed 
to be displaced with a total E.M.F. 24 ; the tension 
in the cord (negative electric potential) accordingly 
rises by a step 3 at each layer. Diagram I. shows 
this initial state. Clamp the cord, to represent 
a constant charge on the plates A and B, and now 
introduce a slab of glass that is, replace the 4 
middle layers by elastics only half as stiff (see II.). 
The stress in the cord steps up now by only ij at 
each of these layers, and the total difference of poten- 
tial, instead of being 24, is now only 18. Meanwhile 
the charges remain the same, and there is no charge 
on the surface of the glass ; the capacity of the 
whole condenser, which was -J, has now increased 
to J. 

There is no charge on the surface of the glass ; 
but the resultant effect is very much the same as if 

E 2 



ELECTROSTATICS 



CHAP. Ill 



there were. The effect on the cord will be precisely 
the same as if the replaced elastics were still of the 



T 


+ 


3 


\ 


\ 


\ 
\ 


\ 


\ 

\ 


\ 


\ 

\ 


\ 


\ 

\ 


\ 


\ 

\ 


\ 


\ 
x 


\ 


\ 

\ 


\ 


\ 


-3 






24 




i 


^ 


X 


X 


7 

X 


X 


/ 
/ 


/ 


X 

/ 


/ 


7 

X 


/ 


7 

X 


X 


2 


7 

/ 


X 


7 


B 





II 


+ 


3 


\ 


\ 


\ 
N, 


\ 


\ 

\ 


X x, 


\ 

x 


\ 


s 


\ 


N^ 


\ 


\ 


\ 


X 

\ 


\ 


\ 


-3 






18 






/ 


/ 


/ 
/ 


/ 


7 
/ 


/ 


/ 


^x' 


/ 


/ 


> 


/ 


X 


/ 


/ 


/ 


x 









in 


4- 


3 


\ 


\ 


\ 
\ 


\ 


\ 

\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 
\ 


\ 


\ 


-3 






18 






/ 


x 


7 

/ 


X 


z 


r 


/ 


/ 


/ 


/ 


/ 


A 


a 


7 

/ 


X 


7 









IV 



xxx 



xl 



24 





-t- 


4 


X 


^ 


^ 
^ 


^ 

\ 


\ 

\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\ 


\J 

x 


\ 
^ 




^ 


A 


l^. 


24 








^ 


/ 


/ 
/ 


/ 

/ 


s 

/ 


? 


/ 


/ 


/ 


/ 


/ 


/ 


+ 
^ 


2 

/ 


^ 
/ 


/ 

/ 


/ 


/ 







FIG. 90. Real and apparent effects of introducing a glass slab between the plates 
of an air-condenser. The numbers on the extreme right represent E.M.F.s. 

I. shows the original condenser, of capacity . 

II. shows the effect of inserting a slab of half the whole thickness, and of 
specific inductive capacity 2, the charge being kept constant. The 
capacity has risen to &. 

III. shows a spurious imitative mode of obtaining the same effect, without any 

change of inductive capacity, by help of boundary charges. 

I V. shows the effect of introducing the slab into the condenser I when it is 

supplied with a constant E.M.F. The capacity is still &. 
V. shows a spurious imitation of this effect by help of boundary charges. 



24 CHARGE AND INDUCTION 53 

same strength, but as if their beads had slid half-way 
back, into the positions shown in III., where surface 
charges exist as indicated by numbers. This, I 
repeat, is not the state of things caused by the glass, 
but it is so like it in effect as to be difficultly dis- 
tinguishable from it ; and we sometimes speak of 
the spurious or virtual charges set up on the glass 
surface meaning the charges in Diagram III., which 
so exactly imitates the resultant effect of II. 

So much for the effect of constant charge ; now take 
the case of constant potential. 

Diagram IV. shows the effect of replacing some of 
the elastics by weaker ones in this case. The E.M.F. 
is kept constant, so the strong elastics have more 
strain thrown on them than before ; no internal 
charge is possible so long as the substances insulate 
perfectly, so all the beads are pulled forward equally. 
The step of potential is now 4 at all the stiffer (or 
air) strata, and 2 at all the weaker (or glass) strata, 
making up the total E.M.F., 24. The charge on the 
plates A and B has increased from 3 to 4 
in accordance with the increase of capacity ; the 
ratio of increase of which is still 3 : 4. Here again 
the real effect shown in IV. may be simulated by 
spurious boundary charges, without any change of 
inductive capacity, as is sufficiently indicated by 
Diagram V., wherein all the elastics are supposed of 
the same strength. 

24. Hydraulic Model of a Leyden Jar. So much 
for the cord model, but I will now describe and ex- 
plain an hydraulic model which illustrates the same 
sort of facts; some of them more plainly and directly 



54 ELECTROSTATICS CHAP, m 

than the cord model. Moreover, since all charging is 
essentially analogous to that of a Leyden jar, let us 
take a Leyden. jar and make its hydrostatic analogue 
at once. 

The form of jar most convenient to think of is one 
supported horizontally on an insulating stand with 
pith ball electroscopes supplied to both inner and 
outer coatings. Or one may use, as I commonly do, 
in conjunction with the hydraulic model, a vertical 
coated pane, with a pith ball connected to each coat- 
ing; but if the electroscopes were of such a kind as 




FIG. 90. 

to show a difference between positive and negative 
potential, they would do better. 

To construct its hydraulic model, procure a thin 
india-rubber bag, such as are distended with gas at 
toy-shops ; tie it over the mouth of a tube with a 
stop-cock, A, and insert the tube by means of a cork 
into a three-necked globular glass vessel or " receiver " 
as shown in the diagram, Fig. 10. 

One of the other openings is to have another stop- 
cock tube, B ; and the third opening is to be plugged 
with a cork, as soon as the whole vessel, both inside 



24 



CHARGE AND INDUCTION 



55 






and outside the bag, is completely full of water 
without air-bubbles. 

This is the insulated Leyden jar : the bag repre- 
sents the dielectric, and its inner and outer coatings 
are the spaces full of water. 

Open gauge-tubes, a and , must be provided for 
tubes A and B, to correspond to the electroscopes 







FIG. io. Skeleton diagram of hydraulic model of a Leyden jar. (Compare 14.*. 

supplied to the jar ; and a third bent tube, C, con- 
necting the inner and outer coatings, will correspond 
to a discharger. Ordinarily, however, of course C will 
be shut. 

A water-pump screwed on to A will represent an 
electric machine connected to inner coating ; and the 
outer coating, B, should open into a tank, to represent 
the earth. The pump will naturally draw its supply 
of water from the same tank. 

The bag being undistended, and the whole filled 
with water free from air the level of the water in the 



56 ELECTROSTATICS CHAP, in 

two gauge-tubes will correspond with that in the 
tank ; and this means that everything is at zero 
potential, i.e. the potential of the earth. 

Now, C being shut, shut also B, open A, and work 
the pump. Instantly the level in the two gauges rises 
greatly and equally : you are trying to charge an 
insulated jar. Turn an electric machine connected to 
a real insulated jar, such as Fig. QD, and its two 
pith balls will similarly and equally rise. 

Now open B for an instant the pressure is relieved, 
and both gauges at once fall, apparently both to zero. 
Repeat the whole operation several times, however ; 
and it will be found that, whereas b always falls to 
zero, a falls short of zero each time by a larger amount ; 
and the bag is gradually becoming distended. This 
is charge by alternate contact. It may be repeated 
exactly with the real jar : a spark put into the inner 
coating, and an equal spark withdrawn from the 
outer coating each time ; and unless this outer spark 
is so withdrawn, the jar declines to charge : water 
(and electricity) being incompressible ( 14 A). 

If B is left permanently open, the pump can be 
steadily worked, so as to distend the bag and raise 
the gauge a to its full height, b remaining at zero all 
the time, save for oscillatory disturbances. 

Having got the jar charged, shut A, and remove 
the pump, connecting the end of A with the tank 
directly. 

Now of course by means of the discharger C the 
fluid can be transferred from inner to outer coat, the 
strain relieved, and the gauges equalized. But if this 
operation be performed while the jar is insulated, 



24 CHARGE AND INDUCTION 57 

i.e. while A and B are both shut, the common level of 
the gauges after discharge is not zero, but a half- 
way level ; and the effect of this is noticeable if 
you charge a Leyden jar ordinarily, then insulate 
it, discharge it, and touch it. 

Instead of using the discharger C, however, we can 
proceed to discharge by alternate contact ; and the 
operation is very instructive. 

Start with the gauge b at zero, and the gauge a 
at high pressure. Open stop-cock A ; some water is 
squeezed out of inner coating, and the a gauge falls to 
zero, but the suck of the contracting bag on the outer 
coat pulls down the gauge b below zero, the descent 
of the two gauges being nearly equal. 

Next shut A and open B ; a little water flows in 
from the tank still further to relieve the strain of the 
bag, and both gauges rise ; b to zero, a to something 
just short of its old position. 

Now shut B and open A again : again the two gauges 
descend. Reverse the taps, and again they both rise ; 
and so on until the bag has recovered its normal size. 
This is discharge by alternate contact, and exactly 
imitates the behaviour of an insulated charged Leyden 
jar whose inner and outer coats are alternately touched 
to earth. Its pith balls alternately rise, with positive 
and with negative electricity respectively, indicating 
potentials above and below zero. 

Figs. 1 1 and 1 2 are taken from photographs of 
apparatus I have made to use as just described. The 
glass globe with the partially distended bag inside it, 
the pump, the tank, the gauges a and #, the stop-cocks 
ABC, will be easily recognized. Two small stop-cocks 



ELECTROSTATICS 



CHAP. Ill 



A' and B', leading direct to tank, are extra, and are to 
save having to disconnect pump and connect A direct, 
when exhibiting the effect of " discharge by alternate 
contact." But the tank is not sufficiently tall in 




FIG. ii. First actually construcced mouel Leyden jar, with mercury gauges as 
electrometers ; the whole ringed up with things purchasable at a plumber's, 
except the pump. The glass globe contains an elastic bag, which swells as 
water is pumped into it. The tank is kept full of water, and its level represents 
the potential of the earth. Flexible tubes full of water effect the desired 
earth-connections when wished. The mercury gauges a and b represent electro- 
scopes connected to inner and outer coats of the jar respectively. 

Fig. 12 ; I have doubled its height since ; and the 
gauge tubes are really longer. 

In any form of apparatus it is essential to fill the 
whole with water pipes, globe, everything before 
commencing to draw any moral from its behaviour. 
It is rather difficult to get rid of a large bubble of air 
from the top of the globe of Fig. n, and though 



25 



CHARGE AND INDUCTION 



59 



it is not of very much consequence in this place, the 
stop-cock in Fig. 12 is added to make its removal 
easy. The gauges in. Fig. n may be replaced by 




FIG. 12. Latest form of hydraulic model of a Leyden jar with water gauges ; the 
whole arranged vertically to be more conspicuous. The pump is a force-pump 
with a communication between bottom of barrel and tank ; also one at the top, 
to get rid of stray water. The parts are labelled to correspond with the skeleton 
diagram, Fig. 10, as well as with Fig. n. 

others arranged as a lantern-slide, and connected by 
flexible tubing full of air. 

25. I have explained thus fully the hydraulic illus- 
tration of Leyden jar phenomena, because these 



60 ELECTROSTATICS CHAP, in 

constitute the key to a great part of electrostatics. 
The illustration is not indeed a complete one, but by 
combining with it a consideration of the endless cord 
models, and of what I have endeavoured to explain 
concerning conduction and insulation in general, a 
distinct step may be gained. 

Think of electrical phenomena as produced by an 
all-permeating liquid embedded in a jelly ; think of 
conductors as holes and pipes in this jelly, of an 
electrical machine as a pump, of charge as excess or 
defect, of attraction as due to strain, of discharge as 
bursting ; and think also of the discharge of a Leyden 
jar as a springing back or recoil, oscillating till its 
energy has gone : 

By thus thinking you will get a more real grasp of 
the subject, and more insight into the actual pro- 
cesses in Nature unknown though these may still 
strictly be than if you employed the old ideas of 
action at a distance, or contented yourselves with no 
theory at all on which to link the facts. You will 
have made a step in the direction of the truth, but 
I must beg you to understand that it is only a step ; 
that what modifications and additions will have to be 
made to it before it becomes a complete theory of 
electricity I am unable fully to tell you. I am con- 
vinced they will be many, but I am also convinced 
that it is unwise to drift along among a host of 
complicated phenomena without guide other than 
that afforded by hard and rigid mathematical 
equations. 

The mathematical theory of potential and the like 
has insured safe and certain progress, and enables 



25 CHARGE AND INDUCTION 61 

mathematicians to dispense for the time being with 
theories of electricity and with mental imagery. Few, 
however, are the minds strong enough thus to dispense 
with all but the most formal and severe of mental aids ; 
and none, I believe, to whom some physical repre- 
sentation or picture of the phenomena would not be a 
help if it were safely available. 



PART II 

CONDUCTION 



CHAPTER IV 

METALLIC AND ELECTROLYTIC CONDUCTION 

26. WE have now glanced through electrostatic 
phenomena, and seen that they could be all com- 
prehended and partially explained by supposing 
electricity to be a fluid of perfect incompressibility in 
other words, a perfect liquid extending everywhere 
and permeating everything ; and by further supposing 
that in conducting matter this liquid was capable of 
free locomotion, while in insulators and general space 
it was as it were entangled in some elastic medium or 
jelly to strains in which electrostatic actions are 
due. This medium might be burst, in a disruptive 
discharge, but easy flow could go on only through 
channels or holes in it, which therefore were taken 
to represent conductors ; and it was obvious that all 
flow, must take place in closed circuits. 

I now want to consider the circumstances of this 
flow more particularly : to study, in fact, the second 
division of our subject (see classification in Section i), 
viz., Electricity in locomotion. 

I use the term "locomotion " in order to discriminate 

F 



66 CONDUCTION CHAP, iv 

it from rotation and vibration : it is with translation 
only that we intend now to concern ourselves. 

Consider the modes in which water may be made to 
move from place to place. There are only two : it may 
be pumped along pipes, or it may be carried about in 
jugs. In other words, it may travel through matter, 
or it may travel with matter. Just so it is with heat 
also ; heat can travel in two ways ; it can flow through 
matter, by what is called " conduction," and it can 
travel with matter, by what is called " convection." 
There is no other mode of conveyance of heat. You 
frequently find it stated that there is a third method, 
viz., " radiation " ; but this is not truly a conveyance 
of heat at all. Heat generates radiation at one place, 
and radiation reproduces heat at another ; but it is 
radiation which travels, and not heat. Heat only 
naturally flows from hot bodies to cold, just as water 
only naturally flows down hill ; but radiation spreads 
in all directions, without the least attention to where 
it is going. Heat can only flow one way at any given 
point, but radiation can travel all ways at once. If water 
were dissociated on one planet into its constituent 
gases, and if these recombined on another planet, it 
would not be water which travelled from one to the 
other, neither would the substance obey the laws of 
motion of water water would be destroyed in one 
place, and reproduced in another ; just so is it with the 
relation between radiation and heat. 

Heat, then, like water, has but two direct modes of 
conveyance from place to place. For electricity the same 
is at first sight true. Electricity can travel with matter, 
or it can travel through matter ; by convection or by 



27 METALLIC CONDUCTION 67 

conduction. But it has recently been discovered that it 
can travel also without matter isolated, detached, 
in the form of electrons ; though into that process we 
will not now enter. 



Conduction in Metals. 

27. Consider, first, conduction. Connect the poles of 
a voltaic battery to the two ends of a copper wire, and 
think of what we call the " current." It is a true flow 
of electricity among the molecules of the wire. If 
electricity were a fluid, then it would be a transport 
of that fluid ; if electricityjs nothing material, then a 
current is no material transfer ; but it is certainly a 
transfer of electricity, whatever electricity may be. 
Permitting ourselves again the analogy of a liquid, we 
can picture it flowing through, or among, the mole- 
cules of the metal. Does it flow through or between 
them? Or does it get handed on from one to the 
next continually? The last supposition is believed 
most nearly to represent the probable truth. The 
flow may be thought of as a perpetual attempt to set 
up a strain like that in a dielectric, combined with an 
equally perpetual breaking down of every trace of that 
strain. If the atoms be conceived as little conductors 
vibrating to and fro and knocking each other, so as to 
be easily and completely able to pass on any electric 
charge they may possess, then, through a medium so 
constituted, electric conduction could go on much as 
it does go on in a metal. Each atom would receive a 
charge from those behind it, and hand it on to those 
in front of it ; and thus might electricity get conveyed 

F 2 



68 CONDUCTION CHAP, iv 

along the wire. Such a theory reduces conduction to 
a kind of electrostatics an interchange of electric 
charges among a set of conductors. If such a set of 
vibrating and colliding particles existed, then certainly 
a charge given to any point would rapidly distribute 
itself over the whole, and the potential would quickly 
become uniform ; and it is now thought that the actual 
process of conduction is something like this. This is 
not the simplest mode of picturing it for ordinary 
purposes. The easiest and crudest idea is to liken a 
wire conveying electricity to a pipe full of marbles or 
sand conveying water ; and for many purposes, though 
not for all, this crude idea suffices. 

Leaving the actual mode of conveyance for the 
present, let us review how much is certainly known of 
the process called conduction in homogeneous metals. 

This much is certainly known :- 

(1) That the wire gets heated by the passage of a 
current. 

(2) That no trace of a tendency to reverse discharge 
or spring back exists. 

(3) That the electricity meets with a certain amount 
of resistance or friction-like obstruction. 

(4) That this force of obstruction is accurately 
proportional to the speed with which the electricity 
travels through the metal that is, to the intensity of 
the current per unit area. 

28. About this last fact a word or two must be said. 
The amount of electricity conveyed per second across 
a unit area is called intensity of current ; 1 and 

1 Often called "density" of current, but "intensity" is perhaps a 
more proper expression for the purpose. 



29 METALLIC CONDUCTION 69 

experiment proves what Ohm originally guessed as 
probable from the analogy of heat conduction that 
this intensity is accurately proportional to the slope 
of potential which causes the flow ; or, in other words 
(since action and reaction are equal and opposite), 
that a current in a conductor meets with an obstruc- 
tive electromotive force exactly proportional to itself. 
Or, quite briefly, a current through a given conductor 
is proportional to the E.M.F. which drives it. The 
particular ratio between slope of potential and cor- 
responding intensity of current depends upon the 
particular material of which the conductor is composed 
and is one of the constants of the material, to be de 
termined by direct measurement. This precise ratio 
is called its " specific conductivity " or its " specific 
resistance," according to the way it is regarded. 

The law here stated is called Ohm's law, and is one 
of the most accurately known laws there are. Never- 
theless it is an empirical relation ; in other words, it 
has not yet been accounted for, it must be accepted as 
an experimental fact. Undoubtedly, it is one of vast 
and far-reaching importance : it asserts a connexion 
between electricity and ordinary matter of a definite 
and simple kind. Using the language of hydraulic 
analogy, it asserts that when electricity flows through 
matter the friction between them is accurately as the 
first power of the velocity, for all speeds. 

29. Now, if we think of this opposing electromotive 
force as analogous to friction, it is very easy to think 
of heat being generated by the passage of a current, 
and to suppose that the rate of heat-production will 
be directly proportional to the opposing force and 



70 CONDUCTION CHAP, iv 

to the current driven against it ; as, in fact, Joule 
experimentally proved it to be. 

But if we are not satisfied with this vague analogy, 
and wish to penetrate into the ultimate nature of heat 
and the mode in which it can be generated, then we can 
return to the consideration of a multitude of oscil- 
lating and colliding particles, moving with a certain 
average energy which determines what we call the 
" temperature " of the body. If now one or more of 
these bodies receives a knock, the energy of the blow 
is speedily shared among all the others, and they all 
begin to move rather more energetically than before : 
the body which the assemblage of particles constitutes 
is said to have " risen in temperature." This illustrates 
the production of heat by a blow or other mechanical 
means. But now, instead of striking one of the balls, 
give it an electric charge ; or, better still, put within its 
reach a constant reservoir of electricity from which it 
can receive a charge every time it strikes it, and at 
the same time put within the reach of some other of 
the assemblage of particles another reservoir of infi- 
nite capacity which shall be able to drain away all the 
electricity it may receive. In practice there is no 
need of infinite reservoirs : all that is wanted is to 
connect two finite reservoirs, or " electrodes " as one 
might now call them, with some constant means of 
propelling electricity from one to the other, i.e., with 
the poles of a voltaic battery or a Holtz machine. 

What will be the result of thus passing a series of 
electric charges through the assemblage of particles ? 
Plainly the act of receiving a charge, and passing it on, 
will tend to increase the original motion of each par- 



30 ELECTROLYTIC CONDUCTION 71 

ticle ; it will tend to raise the temperature of the body. 
In this way, therefore, it is possible to picture the mode 
in which an electric current generates heat. 

Heat, being some mode of motion, must also be 
handed on after some analogous fashion, so that when 
heat is supplied to one point of a mass it spreads or 
diffuses through it. It is difficult to suppose the con- . 
duction of heat to be other than the handing on of 
molecular quiverings from one particle to another ; and 
yet it takes place according to laws altogether different 
from those of the propagation of the gross disturbance 
called sound. The exact mode of conduction of heat 
is unknown, but, whatever it is, it can hardly be 
doubted that the conduction of electricity through 
metals is not very unlike it, for the two processes obey 
the same laws of propagation : they are both of the 
nature of a diffusion, they both obey Ohm's law, and 
a metal which conducts heat well conducts electricity 
well also. 

Conduction in Liquids. 

30. Leaving the subject of conduction in metals 
for the present, let us pass to the consideration of 
the way in which electricity flows through liquids. 
By " liquids," in the present connexion, one more par- 
ticularly means definite chemical compounds, such as 
acids, alkalies, salt and water, and saline solutions 
generally. Some liquids there are, lik'e alcohol, tur- 
pentine, bisulphide of carbon, and water, which, when 
quite pure, either wholly or very nearly decline to con- 
duct electricity at all. Such liquids as these may be 



72 CONDUCTION CHAP, iv 

classed along with air and gases as more or less perfect 
dielectrics. Other liquids there are, like mercury and 
molten metals generally, which conduct after pre- 
cisely the same fashion as they do when solid. These, 
therefore, are properly classed among metallic con- 
ductors. 

But most chemical compounds, when liquefied 
either by heat or by solution, conduct in a way 
peculiarly their own ; and these are called " electro- 
lytes." 

31. The present state of knowledge enables us to 
make the following assertions with considerable con- 
fidence of their truth : 

(1) Electrolytic conduction is invariably accom- 
panied by chemical decomposition, and in fact only 
occurs by means of it. 

(2) The electricity does not flow through^ but witli 
the atoms of matter, which travel along and convey 
their charges something after the manner of pith balls 
between two oppositely charged plates. 

(3) The electric charge belonging to each atom of 
matter is a simple multiple of a definite quantity of 
electricity, which quantity is an absolute constant 
quite independent of the nature of the particular sub- 
stance to which the atoms belong. 

(4) Positive electricity is conveyed through a liquid 
by something equivalent to a procession of the electro- 
positive atoms of the compound, in the direction called 
the direction of the current ; and at the same time 
negative electricity is conveyed in the opposite 
direction by a similar procession of the electro- 
negative atoms. 



32 ELECTROLYTIC CONDUCTION 73 

(5) On any atom reaching an electrode it may be 
forced to get rid of its electric charge, and, combining 
with others of the same kind, escape in the free state; 
in which case visible decomposition results. Or it may 
find something else handy with which to combine say 
on the electrode or in the solution ; and in that case 
the decomposition, though real, is masked, and not 
apparent. 

(6) But, on the other hand, the atom may cling to 
its electric charge with such tenacity as to stop the 
current : the opposition force exerted by these atoms 
upon the current being called polarization. 

(7) No such opposition force, or tendency to spring 
back, is experienced in the interior of a mass of fluid: 
it occurs only at the electrodes. 

32. The three first of these statements constitute a 
summary of Faraday's laws of electrolysis. These 
laws are of far-reaching importance, and appear to be 
accurately true. The first is called the " voltametric 
law," and asserts that the amount of chemical action 
electrolytically produced in any given substance is 
exactly proportional to the quantity of electricity that 
has passed through it. The vague phrase " chemical 
action " is purposely used here to include decomposi- 
tion or recomposition or liberation or deposition or 
dissolution, or any other effect that can be brought 
about, in either elements or compounds, by the pas- 
sage of an electric current. The weight of substance 
acted on measures the quantity of electricity which 
has passed ; hence a decomposition cell can act as a 
voltameter, and the law is called the voltametric law. 
Its truth enables us to make the first of the above 



74 CONDUCTION CHAP, iv 

statements ; which many qualitative facts concerning 
the details of electrolysis modify into statement 
No. 2. 

The second of Faraday's laws is called the law of 
electro-chemical equivalence. It asserts that, if the 
same current be passed through a series of voltameters 
for the same time, the amount of chemical action in 
each substance acted on is exactly proportional to its 
ordinary chemical equivalent ; not to its atomic 
weight merely, but to its atomic weight divided by 
what is called its valency, or atomicity, or quanti- 
valence ; this being its real chemical equivalent. Thus 
an atom of oxygen weighs sixteen times as much as 
an atom of hydrogen, and is equivalent to two such 
atoms in combining power ; hence the law asserts 
that 8 grammes of oxygen are liberated for every 
gramme of hydrogen. Again, an atom of silver is 108 
times as heavy as an atom of hydrogen, and is equal 
to it in combining power ; hence 108 grammes of 
silver are deposited in a silver voltameter while one 
gramme of hydrogen is being liberated by the same 
current in a gas voltameter. Once again, an atom of 
gold weighs as much as 197 atoms of hydrogen, and is 
able to replace three of them in combination : hence 
657 grammes of gold are deposited by the same 
current in the same time ; and so on. 

Now this law plainly means that the same number 
of monad atoms is liberated by the same quantity of 
electricity, no matter what their nature may be ; half 
that number of dyad atoms ; one third that number of 
triad atoms. Hence, assuming statement No. 2 above, 
that the current flows purely by convection each atom 



32 ELECTROLYTIC CONDUCTION 75 

conveying electricity it follows that every monad 
atom carries the same quantity, whether it be an atom 
of hydrogen or of silver, or of chlorine, or a complex 
radicle like NO 3 ; that each dyad atom carries twice 
as much, whether it be an atom of oxygen or of zinc 
or of copper, or a complex dyad radicle like SO 4 ; 
that each triad atom carries three times as much, and 
so on. And this is what is laid down in No. 3 of the 
above statements : which is virtually equivalent to 
Faraday's second law. 

True, it is possible that every atom may have a 
specific charge of its own with which it never parts ; 
but about such nothing is known ; we can only make 
experiments on the charge it is willing to part with 
at an electrode, and there is no doubt that this is 
accurately the same for all substances, up to a simple 
multiple. And this quantity, the charge of one monad 
atom, constitutes the smallest known portion of 
electricity, and is a real natural unit. Obviously this 
is a most vital fact. This unit below which nothing 
is known has been styled an "atom" of electricity ; 
and the phrase is now known to have a distinct mean- 
ing. I have ventured to suggest one or two effects 
which would result from the hypothesis that this unit 
quantity of electricity were really in fact an absolute 
minimum, and as indivisible as an atom of matter. 1 
This natural unit of electricity is exceedingly small, 
being about the ten-thousandth-millionth part of 
the ordinary electrostatic unit ; or in electromagnetic 
units almost exactly io~ 20 c.g.s. 

1 See paper on "Electrolysis" at Aberdeen (Reports of the British 
Association for 1885, p. 763). 



76 CONDUCTION CHAP, iv 

The charge 01 each atom being so small, its 
potential is not high. Something between i and 3 
volts is a likely difference of potential for two 
oppositely charged atoms. But they are so near 
together that even this small difference of potential 
causes a strong electrostatic attraction or "chemical 
affinity " between the oppositely charged atoms of a 
molecule. 

This electrical force between two atoms, at any 
distance, is ten thousand million billion billion times 
greater than their gravitative attraction at the same 
distance. The force has an intensity per unit mass, 
(= io 21 ) and therefore is able to produce an accelera- 
tion, nearly a trillion times greater than that of 
terrestrial gravity near the earth's surface. 

These are undoubtedly the forces with which 
chemists have to do, and which they have long called 
chemical affinity. 

33. But it may be asked, If the atoms in each 
molecule cling together by their electrostatic attrac- 
tions, and if there are an enormous number of atoms 
between two electrodes, how comes it that a feeble 
E.M.F. can pull them apart and effect decomposition ; 
moreover, how can the E.M.F. needed to effect de- 
composition help varying directly with the thick- 
ness of fluid between the plates? It does not depend 
on anything of the kind ; the length of liquid between 
the electrodes is absolutely immaterial. This fact 
proves that throughout the main thickness of liquid 
no atoms are torn asunder at all. Probably they 
frequently change partners, one pair of atoms not 
always remaining united but occasionally getting 



33 ELECTROLYTIC CONDUCTION 77 

separated and recombined with other individuals. 
During these interchanges there must be moments of 
semi-freedom during which the atoms are amenable 
to the slightest directive tendency, and it is probably 
these moments that the applied E.M.F. makes use of. 

The reality of such a state of continual interchange 
between molecules has been forced upon chemists by 
the facts of double-decomposition such facts as the 
interchange of atoms between strongly combined salts 
when their solutions are mixed so as to form very 
much weaker compounds ; the proof that such com- 
pounds are formed being very clear in the case when 
they happen to be insoluble. The fact that if a 
precipitate is insoluble enough it is bound to form, 
really proves that some small quantity of the corre- 
sponding compound is always formed in every case, 
whether it happens to be insoluble or not. 

The state of continual interchange results in a 
perfect sensibility to the migratory power of extremely 
weak forces, so 1 that even the faintest trace of an 
electromotive force is able to affect the charged 
atoms; on the average assisting the positive atoms 
down the slope of potential, and the negative atoms up 
the slope. The fact that the most infinitesimal force 
is sufficient to effect its due quota of decomposition 
was proved most clearly and decisively by the experi- 
ments of Helmholtz. (See his Faraday Lecture.) 

Sometimes the term "dissociation " is used to signify 
this practical freedom of atoms to locomotion ; and, 
as stated originally by Prof. Clausius, the idea of 
dissociation was certainly involved. It was thought 
that a certain percentage of atoms existed in the 



78 CONDUCTION CHAP, iv 

liquid in an uncombined state, wandering about 
seeking partners ; that it was these loose atoms on 
which the electromotive forces acted, and that the 
procession of these conveyed the current. It is likely 
that such atoms are not really and actually free ; they 
may be hangers on of molecular aggregates, but they 
are potentially free or dissociated in the sense that 
their bonds with their ordinary partners are broken 
by the solvent. But even this amount of freedom is 
not proven by the facts of electrolysis alone, for we 
now see that the addition of the idea of double- 
decomposition or interchange of partners to the 
original hypothesis of Grotthus explains all that is 
required by the facts, viz. a virtual or potential dis- 
sociation, a momentary state of hovering and inde- 
cision, without the need for any continuous and actual 
dissociation. Nevertheless something very like actual 
dissociation is now considered most likely ; the state 
being brought about by the action of the solvent on 
a portion of the dissolved compound. 

A certain average number of the atoms are virtually 
free, and amenable to any directive force, without 
necessarily remaining free for more than a brief though 
finite time. The more of these virtually dissociated 
atoms there are, the better is the liquid likely to 
conduct. The process of conduction is probably 
very much as Grotthus imagined it, a passing on of a 
charged ion through a chain of decomposing and 
recomposing molecules : but the molecular decom- 
position is automatic, i.e. it is not causeol, but only 
directed, by the applied electrical force. The result is 
as if a double procession of free charged atoms 



34 ELECTROLYTIC CONDUCTION 79 

migrated through the liquid between the electrodes, 
and it is simplest so to think of the process. 

34. I will now try and make the process of 
electrolytic conduction clearer by reverting to our 
mechanical analogies and models. 

Looking back to Figs. 5 and 6, we see illustrations 
of metallic conduction and of dielectric induction. In 
each case an applied electromotive force causes some 
movement of electricity ; but whereas in the first case 
it is a continuous almost unresisted movement or 
steady flow, through or among the atoms of matter, in 
the second case it is a momentary shift or displace- 
ment only, carrying the atoms of matter with it, and 
highly resisted in consequence : resisted, not with a 
mere frictional rub, which retards but does not check 
the motion, but by an active spring-back force, which 
immediately checks all further current, produces what 
we call " insulation," and ultimately, when the propel- 
ling force is removed, causes a quick reverse motion 
or discharge. But the model is plainly an incomplete 
one ; for what is it that the atoms are clinging to ? 
What is it ought to take the place of the beam in the 
crude mechanical contrivance? Obviously another 
set of atoms, which are either kept still or urged in the 
opposite direction by a simultaneous opposite displace- 
ment of negative electricity ; as in Fig. /A, 18. We 
are to picture two or any number of rows of beads, 
each row threaded on its appropriate cord ; the cords 
alternately representing positive and negative elec- 
tricity respectively, and being simultaneously displaced 
in opposite directions by any applied E.M.F. The 
beads threaded on any one cord have, in a dielectric, 



8o CONDUCTION CHAP, iv 

elastic attachments to those on some parallel and 
oppositely moving cord, and thus continuous motion 
of the cords in opposite directions is prevented : only 
a slight displacement is permitted, followed by a 
spring back and oscillation after the fashion already 
described. 

Very well ; now picture the elastic connections be- 
tween the beads all dissolved, and once more apply a 
force to each cord, moving half of them one way and the 
alternate half the other way, and you have a model 
illustrating an electrolyte > and the simplest view of 
electrolytic conduction. The atoms are no longer 
attached to each other, but they are attached to the 
cord. In the first respect, an electrolyte differs 
from a dielectric ; in the second, it differs from a 
metal. 

Moreover, electrolytic conduction is perceived to 
be scarcely of the nature of true conduction : the 
electricity does not slip through or among the 
molecules, it goes with them. The constituents of 
each molecule are free of each other ; and while one 
set of atoms conveys positive electricity, the other 
set carries negative electricity in the opposite direction; 
and so it is by a procession of free atoms that the 
current is transmitted. The process is of the nature 
of convection : the atoms act as carriers. Free loco- 
motion of charged atoms is essential to electrolysis. 

35. In order to compare with Figs. 5 and 6, so as to 
bring out the points of difference, Fig. 13 is drawn. 
The beads representing one set of atoms of matter 
are tightly attached to the cord, no trace of slip between 
beads and cords being permitted, but otherwise they 



35 



ELECTROLYTIC CONDUCTION 



81 



are free, and so are represented as supported merely 
by rings sliding freely on glass rods. The only 
resistance to the motion, beside the slight friction, 
is offered at the electrode, which is typified by the 
spring-backed knife-edge, z. This is supposed to be 
able to release the beads from the cord when they are 




FIG 13. Crude mechanical analogy, illustrating a few points in a circuit partly 
electrolytic. To compare with Figs. 5 and 6. 

pressed against it with sufficient force. The cling 
between the bead and cord (i.e. between each atom 
and its charge) is great enough to cause a perceptible 
compression of the springs, and accordingly to bring 
out a recoil force in imitation of polarization. 

The piece of cord accompanying each bead on its 
journey (i.e. the length between it and the next bead) 

G 



82 CONDUCTION CHAP, iv 

represents the atomic charge, and is a perfectly con- 
stant quantity : the only variation permissible in it is 
that some kinds of atoms have twice as much, or are 
twice as far apart on their cord, and these are called 
by chemists dyad atoms ; another kind has three 
times as much, another four, and so on ; these being 
called triad, tetrad, &c. 

If the cord be taken to represent positive electricity, 
the beads on it may represent atoms of hydrogen, or 
other monad cation, travelling down stream to the 
cathode. Another cord representing negative electri- 
city may be ranged alongside it, with its beads twice 
as far apart, to represent the atoms of a dyad anion, 
like oxygen. If the cords are so mechanically con- 
nected that they must move with equal pace in 
opposite directions, we have a model illustrating 
several important facts. The number of oxygen 
atoms liberated in a given time will then obviously 
be half the number of hydrogen atoms set free in the 
same time, and will therefore in the gaseous state 
occupy but half the volume. 

Moreover, for any element whatever, the number of 
atoms liberated in anytime is equal to the number of 
atoms of hydrogen liberated in the same time, divided 
by the "valency" of the element as compared with 
hydrogen. This is one of Faraday's laws, and appears 
to be precisely true; and inasmuch as the relative 
weight of every element is known with fair accuracy, it 
is easy to calculate what weight of substance any given 
current will deposit or set free in an hour, if we once 
determine it experimentally for any one substance. 



36 ELECTROLYTIC CONDUCTION 83 

Summary. 

We may summarise thus : 

If we apply E.M.F. to a metal we get a continuous 
flow, and the result is heat. 

If we apply it to a dielectric we get a momentary 
flow or displacement, and the result is the potential 
energy of " charge." 

If we apply it to an electrolyte we again get a 
continuous flow, and the result is chemical decom- 
position. 

36. There are a large number of important points, 
as to the mode by which an electric current is con- 
veyed through liquids ; and I will specially select 
one for further emphasis, viz., that it is effected by 
a procession of positively charged atoms travelling 
one way, and a corresponding procession of negatively 
charged atoms the other way. 

Whatever we understand by a positive charge and 
a negative charge, it is certain that the atoms of, say, a 
water molecule, are charged the hydrogen positively, 
the oxygen negatively ; and it is almost certain that 
they hang together by reason of the attraction between 
their opposite charges. It is also certain that when 
an electromotive force i.e. any force capable of 
propelling electricity is brought to bear on the 
liquid, the hydrogen atoms travel on the whole in one 
direction, viz., down stream, and the oxygen atoms 
travel in the other direction, viz., up stream ; using 
the idea of level as our analogue for electric potential 

G 2 



84 CONDUCTION CHAP. IV 

in this case. The atoms may be said to be driven 
along by their electric charges just as charged pith, 
balls would be driven along ; and they thus act as 
conveyors of electricity, which otherwise would be 
unable to move through the liquid. 

Each of this pair of opposite processions goes on 
until it meets with some discontinuity either some 
change of liquid, or some solid conductor. At a 
change of liquid another set of atoms continues 
the convection, and nothing very particular need be 
noticed at the junction ; but at a solid conductor the 
stream of atoms must stop: you cannot have loco- 
motion of the atoms of a solid. The obstruction so 
produced may stop the procession, and therefore the 
current, altogether ; or, on the other hand, the force 
driving the charges forward may be so great as to 
wrench them free : in which case the charges are given 
up to the electrode, which conveys it away by common 
conduction, while the atoms are crowded together in 
such a way that they are glad to combine with each 
other and escape. 

Speed of Travel. 

37. Now notice the fact of the two opposite proces- 
sions. It is impossible to get one kind of ion liberated 
at one electrode without having a precisely equivalent 
quantity of the opposite ion liberated, or deposited, 
or otherwise appearing, at the other electrode ; and 
this fact may be expressed by saying that it is im- 
possible to have a procession of positive atoms 
through a liquid without a corresponding procession 



37 ELECTROLYTIC CONDUCTION 85 

of negative ones. In other words, an electric current 
in a liquid necessarily consists of a flow of positive 
electricity in one direction, combined with a flow of 
negative electricity in the opposite direction. And if 
this is thus proved to occur in a liquid, why should it 
not occur everywhere? It is at least well to bear the 
possibility in mind. 

Another case is known where an electric current 
certainly consists of two opposite streams of electricity, 
viz., the case of the Holtz machine. While the 
machine is being turned, with its terminals somehow 
connected, the glass plate acts as a carrier conveying 
a charge from one collecting comb to the other at 
every half revolution ; but, whereas it carries positive 
electricity for one half of its rotation, it carries 
negative for the other half. The top of the Holtz 
disk is always, say, positively charged, and is 
travelling forward ; while the bottom half, which is 
travelling backward at an equal rate, is negatively 
charged. 

In the Holtz case the speeds are necessarily equal, 
but the charges are not. In the electrolytic case the 
charges are necessarily equal, but the speeds are not. 
Each atom has its own rate of motion in a given liquid, 
independently of what it may happen to have been 
combined with. This is a law discovered by Kohl- 
rausch. Hydrogen travels faster than any other kind 
of atom ; and on the sum of the speeds of the two 
opposite atoms in a compound the conductivity of the 
liquid depends. Acids, therefore, which contain replace- 
able hydrogen, in general conduct betterthan their salts. 

The following table gives the rates at which 



86 CONDUCTION CHAP, iv 

atoms of various kinds can make their way through 
nearly pure water, when urged by a slope of potential 
of i volt per linear centimetre : l 

H . . . ro8 centimetre per hour. 

K . . . 0-205 

Na . . . 0-126 

Li . . . 0*094 

Ag . . . 0-166 

Cl . .0-213 

I ... 0'2l6 

NO 3 . 0-174 

1 Brit. Assoc. Retort, 1886, p. 393. 



CHAPTER V 

CURRENT PHENOMENA 

Electrical Inertia. 

38. RETURNING now to the general case of conduc- 
tion, without regard to the special manner of it, we 
must notice that, if a current of electricity were any- 
thing of the nature of a material flow, there would 
probably be a certain amount of inertia connected 
with it ; so that to start a current with a finite force 
would take a little time ; and the stoppage of a 
current would also have either to be gradual or else 
violent. It is well known that if water is stagnant in 
a pipe it cannot be quite suddenly set in motion ; 
and again, if it be in motion, it can only be sud- 
denly stopped by the exercise of very consider- 
able force, which jars and sometimes bursts the 
pipe. The impetus of running water is utilized in 
the water-ram. It must naturally occur, therefore, to 
ask whether any analogous phenomena are experi- 
enced with electricity ; and the answer is that analo- 
gous phenomena are very conspicuous. A current 
does not start instantaneously : it takes a certain 



88 CONDUCTION CHAP, v 

time though usually a very short time to rise to its 
full strength ; and when started it tends to persist, so 
that if its circuit be suddenly broken, it refuses to 
stop quite suddenly, and bursts through the intro- 
duced insulating partition with violence and heat. It 
is this ram or impetus of the electric current which 
causes the spark seen on breaking a circuit ; and the 
more sudden the breakage the more violent is the 
spark apt to be. 

The two effects the delay at making circuit, and 
the momentum at breaking circuit used to be called 
" extra-current " effects, but they are now more 
commonly spoken of as manifestations of " self- 
induction." 

We shall understand them better directly ; mean- 
while they appear to be direct consequences of the 
inertia of electricity ; and certainly if electricity 
were a fluid possessing inertia it would behave to a 
superficial observer just in this way. 

39. But if an electric current really possessed inertia, 
as a stream of water does, it would exhibit itself not 
only by these effects, but also mechanically. A con- 
ducting coil delicately suspended might experience a 
rotary kick every time a current was started or stopped 
in it ; and a coil in which a steady current is main- 
tained should behave like a top or gyrostat, and 
resist any force tending to deflect its plane. 

Clerk Maxwell has carefully looked for this latter 
form of momentum effect, and found none. He took 
a bar electro-magnet, mounted it on gimbals so that it 
was free to rotate if it wished, and then spun it rapidly 
about an axis perpendicular to the magnetic axis. If 



39 CURRENT PHENOMENA 89 

there had been the slightest gyrostatic action, the 
magnet would have rotated about the third perpen- 
dicular axis. But it did nothing of the kind. One may 
say, in fact, that nothing like momentum has yet been 
observed in an electric current through solids or liquids, 
by any mechanical mode of examination. There is a 
now well-known exception in the case of gases ; but 
it is safe to say that a coil or whirl of electricity does 
not behave in the least like a top ( 185). 

I have looked for the effect in another way sug- 
gested by Maxwell, viz., by starting and stopping a 
current in a freely suspended coil [also in a suspended 
electrolytic tube, 1 and in a small mica condenser], and 
watching for recoil kicks at the instants of varying 
current strength. Terrestrial magnetism and the 
reaction between fixed and movable parts of the 
circuit caused spurious effects ; but when these were 
reduced to a minimum, by the thick soft-iron case of 
a "marine galvanometer" and other suitable pre- 
cautions, no certain residual effect due to change of 
momentum could be perceived. The experiments 
were by no means final, but they were sufficient to 
show that to detect any possibly existent effect of the 
kind considerable refinement must be employed. 

Suppose, however, that highly refined experiments 
directed to the same object still gave a negative 
result, would that prove that a current has no 
momentum of any kind ? Not necessarily. It might 
be taken as suggesting that an electric current con- 
sists really of two equal flows in contrary directions, 

1 For discussion of a probable electrolytic momentum, by reason of the 
different mass of the opposite ions, see Lodge, Phil, Mag. Nov. 1876, 



90 CONDUCTION CHAP, v 

so that mechanically they neutralize one another 
completely, while electrically i.e. in the phenomena 
of self-induction or extra-current they add their 
effects ( 89). Or it might mean merely that the 
momentum was too minute to be so observed. Or, 
again, the whole thing the appearance of inertia in 
some experiments and the absence of it in others 
may have to be explained in some altogether less 
simple manner, to which we will proceed to lead up. 

Condition of the Medium near a Current. 

40. So far we have considered the flow of electricity 
as a phenomenon occurring solely inside conductors, 
just as a flow of water through pipes is a phenomenon 
occurring solely inside them. But a number of re- 
markable facts are known which completely negative 
this view of the matter. Something is no doubt pass- 
ing along conductors when a current flows, but the 
disturbance is not confined to the conductor ; on the 
contrary, it spreads more or less throughout surround- 
ing space. 

The facts which prove this have necessarily no 
hydraulic analogue, but must be treated suorum 
generum, and they are as follows : 

(1) A compass needle anywhere near an electric 
current is permanently deflected so long as the 
current lasts. 

(2) Two electric currents attract or repel one 
another, according as they are in the same or 
opposite directions. 

(3) A circuit in which a current is flowing tends to 



41 CURRENT PHENOMENA 91- 

enlarge itself so as to inclose the greatest possible 
. area. 

(4) A circuit conveying a current in a 'magnetic 
field tends either to enlarge or to shrink or to turn 
part way round, according to the aspect it presents to 
the field. 

(5) Conductors in the neighbourhood of an electric 
circuit experience momentary electric disturbances 
every time a current in it is started or stopped or 
varied in strength. 

(6) The same thing happens even with a circuit 
conveying a steady current if the distance between it 
and a conductor is made to vary. 

(7) The inertia-like effects of self-induction, or 
extra-currents, can be almost abolished in a covered 
wire by doubling it closely on itself, or better by 
laying a direct and return ribbon face to face ; 
whereas they may be intensified by making the 
circuit inclose a large area, more by coiling it up 
tightly into a close coil, and still more by putting a 
piece of iron inside the coil so formed. 

Nothing like any of these effects is observable with 
currents of water ; and .they prove that the pheno- 
mena connected with the current, so far from being 
confined to the wire, spread out into space and affect 
bodies at a considerable distance. 

41. Nearly all this class of phenomena were dis- 
covered by Ampere and by Faraday, and were called 
by the latter " current-induction." According to his 
view, the dielectric medium round a conducting circuit 
is strained, and subject to~stresses, just as is the same 
medium round an electrically charged body. The 



92 CONDUCTION CHAP, v 

one is called an electrostatic strain, the other an 
electro-magnetic or electro-kinetic strain. 

But whereas electrostatic phenomena occur solely 
in the medium conductors being mere breaks in it 
interrupters of its continuity, at whose surface charge - 
effects occur, but whose substance is completely 
screened from disturbance, that is not the case with 
electro-kinetic phenomena. It would be just as 
erroneous to conceive electro-kinetic phenomena as 
occurring solely in the insulating medium as it would 
be to think of them as occurring solely in the conduct- 
ing wires. The fact is, they occur in both not only 
at the surface of the wires, like electrostatic effects, 
but all through their substance. This is proved by the 
fact that conductivity increases in simple proportion 
with sectional area ; it is also proved by every part of 
a conductor getting hot ; and it is further proved in 
the case of liquids by their decomposition. 

But the equally manifest facts of current attraction 
and current induction prove that the effect of the 
current is felt throughout the surrounding medium as 
well, and that its intensity depends on the nature of 
that medium ; we are thus wholly prevented from 
ascribing the phenomenon of self-induction or extra- 
current to simple and straightforward inertia of 
electricity in a wire, like that of water in a pipe. 

We are brought face to face with another suggestion 
to account for these effects, viz. this : Since the 
molecules of a dielectric are inseparably connected 
with electricity, and move with it, it is possible that 
electricity itself has no inertia at all, but that the 
inertia of the atoms of the displaced dielectric confer 



42 CURRENT PHENOMENA 93 

upon it the appearance of inertia. Certainly they do 
sometimes confer upon it this appearance, as we see 
in the oscillatory discharge of a Leyden jar. For a 
displaced thing to overshoot its mean position and 
oscillate till it has expended all its energy is a pro- 
ceeding eminently characteristic of inertia ; and so, 
perhaps, the phenomena of self-induction may be 
explicable ( 98). Anyhow the oscillatory discharge 
of a Leyden jar depends upon self-induction. 

Further consideration of this more difficult part of 
the subject is, however, best postponed to Part III. 
( 48 and 88). 



Energy of tJie Current. 

42. I have now called attention to the fact that the 
whole region surrounding a circuit is a field of force 
in which many of the most important properties of the 
current (the magnetic, to wit) manifest themselves. 
But directly we begin thus to attend to the whole 
space, and not only to the wires and battery, a very 
curious question arises. Are we to regard the current 
in a conductor as propelled by some sort of end-thrust, 
like water or air driven through a pipe by a piston or 
a fan ? Or are we to think of it as propelled by side 
forces, a sort of lateral drag, like water driven along a 
trough by a blast of air, or by the vanes of paddle- 
wheels dipping into it ? Or, again, referring to the 
cord models, Figs. 5, 6, and 13, were we right in 
picturing the driving force of the battery as applied 
at one place, as shown in the diagrams, or ought we 



94 CONDUCTION CHAP. V 

to have schemed some method for communicating 
the power of the battery by means of belts or other 
mechanism to a great number of points of the circuit ? 
Is a propelling force applied to electricity at one 
point, or is it transmitted through the dielectric to 
every part of the boundary of the conducting circuit ? 

Prof. Poynting has shown that, on the principles 
developed by Maxwell, the latter of these alternatives, 
though apparently the more complicated, is the true 
one ; and he has calculated the actual paths by which 
the energy is transmitted from the battery to the 
various points of a circuit, for certain cases. 

We must learn, then, to distinguish between the 
flow of electricity and the flow of electric energy : they 
do not occur along the same paths. Hydraulic ana- 
logies, at least hydraulic analogies of a simple kind, 
break down here. When hydraulic power or steam 
power is conveyed along pipes, the fluid and its 
energy travel together. Work is done at one end of 
the tube, in forcing in more water, and this is pro- 
pagated along the tube and reappears at the distant 
end, as the work of the piston. But in electricity it is 
not so. Electric energy is not to be regarded as 
pumped in at one end of a conducting wire, and as 
exuding in equal quantities at the other. The elec- 
tricity does indeed travel thus whatever the travel of 
electricity may ultimately be found to mean, but the 
energy does not. The battery emits its energy, not 
to the wire direct, but to the surrounding medium ; 
the surrounding medium is disturbed and strained, 
and propagates the strain on from point to point till 
it reaches the wire and is dissipated. This, Prof. 



42 CURRENT PHENOMENA 95 

Poynting would say, is the function of the wire : it is 
to dissipate the energy crowding into it from the 
medium, which else would take up a static state of 
strain and cease to transmit any more. The con- 
tinuous dissipation of the medium's energy into heat 
renders possible the continuous propagation of elec- 
tricity ( 107). 

The energy of a dynamo, for instance, does not 
travel to a distant motor through the wires, but 
through the air. The energy of an Atlantic-cable 
battery does not travel to America through the wire 
strands, but through the insulating sheath. This is a 
singular and apparently paradoxical view, yet it is 
well founded. 

Think of a tram-car drawn by an underground 
rope. A contact piece of iron protrudes from the 
bottom of the car and grips the moving rope, which is 
thus enabled to propel the car. How does the energy 
of the distant stationary engine reach the car ? Via 
the rope and the iron connector, undoubtedly. They 
both have to be strong, and are liable to be broken 
by the transmitted stress. 

Next, think of an electric tram-car driven by means 
of a current taken up from an underground conductor, 
as in several electric railways. A contact piece of 
wire rope protrudes from the bottom of the car and 
drags a little truck or other sliding contact piece 
along the conductor, which is thus enabled to supply 
electricity to the electro-magnetic motor geared to 
the wheels. How does the energy of the distant 
dynamo reach the car in this case ? Not via the wire 
connector ; not even via the underground conductor. 



96 CONDUCTION CHAP, v 

There is no strain on either. It travels from the 
distant dynamo through the general insulating 
medium between cable and earth, some little enters 
the conductor and is dissipated, but the great bulk 
flows on and converges upon the motor in the car, 
which is thus propelled. All the energy sent into the 
conducting wire is dissipated and lost as heat : it is 
the energy conveyed by the insulating medium which 
is really transmitted and utilized. 

When the attempt is made to transmit too much 
energy by a wire rope, the rope is liable to snap : 
showing what was the medium of transmission. On 
the other hand, when the attempt is made to transmit 
too much energy by an electric cable, it is not the 
copper wire that gives way, but the insulation which 
breaks down. It is the gutta-percha which is trans- 
mitting the energy, and so it is the gutta-percha 
which is liable to suffer from over stress. True, the 
wire might be melted by having to dissipate too 
much energy, just as the bearings of a rapidly 
travelling wire rope might be made red hot if in- 
sufficiently lubricated ; but the energy thus converted 
into heat is manifestly not transmitted : it is simply 
being wasted on the way, 

The paths by which energy travels in ordinary 
machinery are worth attention. In belting, energy 
travels against the moving matter, via the taut half of 
the belt ; in shafts, it travels along the axis of 
rotation, not in the direction of the material motion, 
but at right angles to it ; in mill-gearing or cog 
wheels, it travels neither axially nor tangentially but 
approximately radially, the exact paths being studied 



43 CURRENT PHENOMENA 97 

in treatises on the theory of machinery (such as that 
of Reuleaux). 

The manner in which the transmission of electric 
energy goes on we will attend to further in Part III. 
( 105 et 



Phenomena Peculiar to a Starting, or Stopping, or 
Varying Current. 

43. There is a remarkable fact concerning electric 
currents of varying strength, which was brought into 
prominence by the experimental skill of Prof. 
Hughes, viz., that a current does not start or stop 
equally and simultaneously at all points in the 
section of a conductor, but starts at the outside first. 
This fact is naturally more noticeable with thick wires 
than with thin, and it is especially marked with iron 
wires, for reasons which in Part III. will become 
apparent ; but the general cause of it in ordinary 
copper wires can very easily be perceived, in the light 
of the views of Prof. Poynting just mentioned. 

For, remember that a current in a wire is not 
pushed along by a force applied at its end, so as to be 
driven over obstacles by its own momentum combined 
with a vis a tergo ; but it is urged along at every point 
of its course by a force just sufficient to make it over- 
come the resistance there, and no more : the force 
being applied to it through the medium of the di- 
electric in which the wire is immersed. A lateral force 
it is which propels electricity ; and it naturally acts 
first on electricity in the outer layers of the wire or 

PI 



98 CONDUCTION CHAP, v 

rod, only acting on the interior portions through the 
medium of the outside ( 102). 

44. To illustrate this matter further, begin to 
rotate a common tumbler of liquid steadily, and 
watch the liquid ; dusting powder over it perhaps to 
make it more visible. You will see first the outer 
layer begin to participate in the motion, and then the 
next, and then the next, and so on, until at length the 
whole is in rotation. Stop the tumbler, and the liquid 
also begins gradually to stop by a converse process : 
the outside stopping first, and then gradually the 
central portions. 

If the liquid sticks together pretty well, like treacle, 
the motion spreads very rapidly : this corresponds to 
a poor conductor. If the liquid be very mobile, the 
propagation of motion inward is slow : this corre- 
sponds to a good conductor. If the liquid were 
perfectly non-viscous, it would correspond to a perfect 
conductor, and no motion would ever be communicated 
to it, deeper than its extreme outer skin. 

Think now of an endless tube full of water, say 
the hollow circumference of a wheel, or the rim of 
a top, and spin it : the liquid is soon set in rotation, 
especially if the tube be narrow or the liquid viscous ; 
but it is set in motion by a lateral not an end force, 
and its outer layers start first. 

Just so it is with a current starting in a metal 
wire. If the wire be fine, or its substance badly con- 
ducting, it all starts nearly together ; but if it be made 
pretty thick, and of well-conducting substance, its 
outer layers may start appreciably sooner than the 
interior. And if it were infinitely conducting, no more 



45 CURRENT PHENOMENA 99 

than the outer skin would ever start at all (see 
Chap. X. and 103). 

In actual practice the time taken for all the 
electricity in an ordinary wire to get into motion is 
excessively short something less than the thousandth 
of a second so that the only way to notice the effect 
is to start and reverse the current many times in 
succession. 

45. If the hollow-rimmed wheel above spoken of 
were made to oscillate rapidly, it is easy to see that only 
the outer layers of water in it would be moved to and 
fro ; the innermost water would remain stationary ; 
and accordingly it would appear as if the tube con- 
tained much less water than it really does. The virtual 
bore of the pipe would, in fact, for many purposes be 
diminished. So it is also with electricity ; the sectional 
area of a wire to a rapidly alternating current is 
virtually lessened so far as its conducting power is 
concerned ; and accordingly its apparent resistance 
is higher for alternating than for steady currents. 
The effect is, however, too small to notice in practice 
except with thick wires and very rapid alternations : it 
then becomes of commercial importance. 

Remember that a propelling force acts on electricity 
at the boundary between the dielectric and the con- 
ductor ; hence the more extensive this boundary, 
the more readily and quickly is the electricity got into 
motion ; and a conductor in that case need not have 
any inert central portion too distant from the propelling 
dielectric. In other words, by splitting up the con- 
ductor into a bundle of ^insulated wires, thus affording 
the dielectric access to a considerable surface of con- 

H 2 



loo CONDUCTION CHAP, v 

ductor, the force is applied much more thoroughly ; 
and so the throttling effect spoken of is greatly 
lessened. The same thing is achieved by rolling 
out the conducting rod into a flat thin bar. Making 
the conductor hollow instead of solid offers the ad- 
vantage of gain of surface per given weight ; but 
no other gain, because no energy travels via the 
hollow space, it still arrives only from the outside ; 
unless, indeed, the return part of the circuit is taken 
along the axis of the hollow, like a telegraph cable. 
In this last arrangement all the energy travels via 
the dielectric between the two conductors, and none 
travels outside at all. It will be perceived therefore 
that, as in static electricity, the term " outside " must 
be used with circumspection : it really means that 
side of a conductor which faces the opposite conductor 
across a certain thickness of dielectric. 

46. We learn from this that, whereas in the case 
of steady currents the sectional area and material of 
a conductor are all that need be attended to, the case 
is different when one has to deal with rapidly alternat- 
ing currents, such as occur in a telephone, or, again, 
such as are apt to occur in Leyden-jar discharges 
(see Part I., p. 46), or in wireless telegraphy, or in 
lightning. 

In all these cases it is well to make the conductor 
expose considerable surface to the propelling medium 
the dielectric else will great portions of it be 
useless. 

Hence, so far as electric facts are concerned, a 
lightning conductor should certainly not be a round 
rod, but a flat strip or a strand of wires, with the 






47 CURRENT PHENOMENA 101 

strands as well separated as convenient. Consider- 
ations of durability and chemical deterioration may 
modify the application of this statement to practice. 

47. I might go on to say here that iron makes an 
enormously worse conductor than copper for rapidly 
alternating currents. So it does for currents which 
alternate with moderate rapidity a few hundred or 
thousand a second like those from a dynamo or a 
telephone ; and so it does always as far as true con- 
ductivity is concerned, the throttling effect is always 
much more marked in iron than in copper ; but, 
singularly enough, when the rapidity of oscillation is 
immensely high, as it is in Leyden-jar discharges and 
lightning, iron acts just as well as copper does, because 
the current keeps to the extreme outer layer of the 
conductor in either case ; and though the conducting 
layer is decidedly thinner in iron than it is in copper, 
yet the total obstruction due to other causes is so 
enormous that this extra throttling is unimportant, 
and practically the material matters very [little as 
regards mere obstruction. But there is another effect 
that deserves special consideration, namely, the dis- 
sipation of energy by a conductor what is called its 
Ohmic resistance ; and for that the material makes 
all the difference. The magnetic quality of iron 
causes rapid alternations to keep to the surface, so 
as to avoid having to magnetise the interior, on the 
principle of choosing the easiest path. The extreme 
thinness of the superficial layer of iron which conducts 
rapid alternations causes it to rub out their energy 
with great rapidity, and accordingly iron acts as a 
most efficient dissipator of energy of this kind. A 



102 CONDUCTION CHAP, v 

Leyden jar discharge sent through thick copper is 
loud and violent. The same discharge taken through 
a good length of thin iron wire is soft and gentle ; 
most of its energy is expended in the wire instead of 
in the flash. 

The application of all this to the case of lightning 
conductors should be obvious. A lightning conductor 
is wanted, not only to carry the discharge current, 
but to enfeeble it also, to prevent it altogether if 
possible, by tapping a surplus charge, but if not then 
to carry it off quietly. Iron for this purpose is 
immensely superior to any other material ; and there 
is no need for it to be thick. A thin iron wire offers 
to most kinds of discharge not to all a path 
apparently as attractive as a thick copper rod does ; 
but whereas a flash to the rod is dangerously violent, 
a flash to a thin enough iron wire is comparatively 
harmless. It is likely indeed to fuse the wire : but by 
that time the danger is over. It will have protected 
the building, at the cost of its own life. The occur- 
rence of what are called B flashes, which discriminate 
between good and bad conductors so that for them 
a conductor must have some reasonable thickness, 
prevents this doctrine from being pressed to extremes : 
so also do chemical fumes in the atmosphere and 
other practical considerations. But undoubtedly 
galvanised or otherwise protected iron is the best 
material for lightning conductors ; and its cheapness 
enables many to be supplied, instead of only one or 
two. 



48 CURRENT PHENOMENA 103 

The Question of Electrical Momentum again. 

48. We are now able to return to the important 
question whether an electric current has any mo- 
mentum or not, as it would have if it were a flow of 
material liquid. Referring to Part I., 7, a hint will 
be found that the laws of flow of a current in con- 
ductors the shape of the stream-lines, in fact are 
such as indicate no inertia, or else no friction. Now 
Ohm's law shows that at any rate friction is not ab- 
sent from a current flowing through a metal ; hence 
it would appear at first sight as if inertia must be 
absent. 

The stream-lines bear upon the question in the 
following kind of way. If an obstacle is interposed 




FIG. 14. Stream-lines of water flowing through a pipe with an obstruction in it. 

in the path of a current of water, the motion of the 
water is unsymmetrical before and behind the obstacle. 
The stream-lines spread out as the water reaches the 
obstacle, and then curl round it, leaving a space full 
of eddies in its wake (Fig. 14). 

But if one puts an obstacle in the path of an electric 
current say by cutting a slit in a conducting strip 



104 CONDUCTION CHAP, v 

of tinfoil the stream-lines on either side of it are 
quite symmetrical, thus 




FIG. 15. Electrical stream-lines past an obstacle. 

And this is exactly what would be true for water 
also, if only it were devoid either of friction or of 
inertia, or of both ; or as Sir Geo. Stokes has now 
found, and Prof. Hele Shaw demonstrated, if the 
friction just compensated the inertia so as to make 
the motion dead-beat. 

49. Is not this fact conclusive, then ? Does it not 
prove the absence of momentum in electricity ? 

Plainly the answer must depend on whether there 
is any other possible mode of accounting for this kind 
of flow. And there is. 

For suppose that water, instead of being urged by 
something located at a distance from the obstacle 
instead of being left to its own impetus to curl 
round or shoot past as it pleases suppose it were 
propelled by a force acting at every point of its 
journey, a force just able to drive it at any point 
against the friction existing at that poinf and no more ; 
then the flow of water would take place according to 
the electrical stream-lines shown in Fig. 15. 

An illustration of such a case is ready to hand- 



49 CURRENT PHENOMENA 105 

Take a spade-shaped piece of copper wire or sheet, 
heat it a little, and fix it in quiescent smoky air ; 
looking along it through a magnifier, in a strong light, 
you will see the warmed air streaming up past the 
metal according to the stream-lines of Fig. 15 j 1 and 
this just because the moving force has its location at 
the metal surface, and not in some region below it. 
One cannot indeed say that it is propelled at every 
point of its course, but it is propelled at the critical 
points where the special friction occurs, and this 
comes to sufficiently the same thing. 

We learn, therefore, that stream-lines like Fig. 15 
prove one of three things, not one of two ; and the 
three things are: (i)that the fluid has no friction; 
or (2) that it has no inertia ; or (3) that it is propelled 
at every point of its course. 

If any one of these is true of electricity, there is no 
need to assume either of the others in order to explain 
the actual manner of its flow. Now we have just 
seen in 42 that, according to Prof. Poynting's inter- 
pretation of Maxwell's theory, the third of the above 
is true electricity is propelled at every point of its 
course ; consequently, as said in Part. I. 7, the 
question of its own intrinsic inertia so far remains 
completely open ( 88, 89, and 98). 

1 See Lord Rayleigh, Nature , vol. xxviii. p. 139. 



CHAPTER VI 

CHEMICAL AND THERMAL METHODS OF PRODUCING 
CURRENTS. CONDUCTION IN GASES 

Voltaic Battery. 

50. LEAVING the last-mentioned mode of regarding 
the subject for the present, to return to it in Chap. X. 
Part III., let us* proceed to ask how it comes about 
that a common battery or a thermopile is able to 
produce a current. (Read Chapter IV. again.) 

If we allow ourselves to assume the existence of an 
unexplained chemical attraction between the atoms 
of different substances, an explanation of the action 
of an ordinary battery-cell is easy. You have first 
the liquid, containing, let us say, hydrogen and oxygen 
atoms, free or potentially free that is, either actually 
dissociated, or so frequently interchanging at random 
from molecule to molecule that the direction of their 
motion may be guided by a feeble directive force 
( 33). Each of these atoms in the free state 
possesses a charge of electricity the hydrogen all a 
certain amount of positive electricity, the oxygen 
twice that amount of negative. Into this liquid you 



5i PRODUCTION OF A CURRENT 107 

then plunge a couple of metals which attract these 
atoms differently : for instance, zinc and copper, 
which both attract oxygen, but zinc more than 
copper ; or, better, zinc and platinum, the latter of 
which hardly attracts it at all ; or, better still, zinc 
and peroxide of lead, one of which attracts oxygen, 
the other hydrogen. 

Immediately, the free oxygen atoms begin moving 
up to the zinc, the free hydrogen atoms to the other 
plate. 

51. When one speaks of the plates attracting 
atoms, it is not necessary to think of their exerting a 
force on all those in the liquid, distant and near : all 
that it is reasonable to assume is a force acting on 
those which come within what is called " molecular 
range " of its surface a distance extremely minute, 
and believed, with excellent reason, to be about the 
ten-millionth part of a millimetre. If the zinc plate 
could remove and combine with all the oxygen atoms 
which come within this range, they would be speedily 
replaced by others from the next more distant layer 
by diffusion, and these again by others, and so on. 
And thus there would be a gradual procession of oxy- 
gen atoms all through the liquid towards the zinc, the 
rate of the procession being regulated by the rate 
of diffusion possible in the particular liquid used. 

The atoms which reach the zinc may be supposed 
to communicate to it the negative charge they carry, 
thus very soon making it negatively electrified 
enough to neutralize its attractive power on the 
similarly charged oxygen atoms ; and then everything 
would stop. But if a channel for the escape of its 



io8 CONDUCTION CHAP, vi 

electricity be provided, by leading a wire from it to a 
copper plate immersed in the same liquid the circuit 
is completed, the electricity streams back by the wire, 
and the procession may go steadily on. The negative 
electricity thus imparted to the copper, or platinum, 
neutralizes any attraction it may have exerted on the 
negatively charged oxygen atoms, and enables it even 
to assist the opposite procession of hydrogen atoms 
towards it ; which on their arrival deliver up their 
charges to it, combine with each other, and escape 
as gas. 

Without going into all the niceties possible, this 
mode of thinking of the matter at least calls attention 
to some of the more salient features of a battery. 

52. If, instead of two different plates, plates of 
the same metal be immersed, they will need to be 
oppositely electrified by some means, before they 
are able to cause the two opposite processions to 
so maintain a current in the liquid. This plainly 
corresponds to a voltameter. In a voltameter the 
maintained slope of potential between the plates must 
be supposed to act on all the ions in the liquid not 
only those within molecular range of the electrodes. 
Every ion is directly propelled. In that respect it 
differs from a voltaic cell, where the chief propelling 
force is the chemical attraction of one of thejplates for 
one of the constituents of the liquid through molecular 
range only. 

53. Taking advantage of the known fact that the 
atoms are charged, Helmholtz avoided the necessity 
for postulating any chemical (non-electrical) force 
between zinc and oxygen, by imagining that all 



54 PRODUCTION OF A CURRENT 109 

substances have a specific attraction for electricity 
itself, and that zinc exceeds copper and the other 
common metals in this respect. 

He would thus think of the zinc attracting, not the 
oxygen itself, but its electric charge ; and so would 
liken a battery cell rather more nearly to a volta- 
meter. The polarisation or opposition force acting at 
the hydrogen-evolving plate he would account for 
by the attraction of hydrogen for positive electricity, 
and the consequent repugnance of the hydrogen 
atoms to part with their charges. 

Voltcts so-called Contact Force. 

54. It maybe convenient to append to this account 
of the action of a battery a statement of the way in 
which the electric charges observed on plates of zinc 
and copper, which have been put into contact and 
separated, are brought about. It can be regarded 
in a simple fashion, though a great deal has been 
written about it. 

Plates of zinc and copper immersed in air are 
under precisely the same chemical conditions as if 
they were immersed in water. The only difference is 
that, whereas water is a conductor, air is an insulator. 
Until the plates of zinc and copper (or other pair of 
metals) are made to touch, nothing happens in either 
case, because the chemical tendency is uniform all 
over both plates ; and though the attraction of 
zinc for oxygen is pretty strong, it is impossible for 
charged atoms to move up to it equally on all sides, 
because by so doing they would be charging a con- 



i io CONDUCTION CHAP, vi 

ductor with one kind of electricity without any com- 
pensating opposite charge of some other conductor. 
(See 5, 3, 4, I4A.) A piece of pure zinc surrounded 
on all sides uniformly by charged atoms may be 
considered as in the same condition as the interior of 
a charged conductor ; and experiment shows that not 
a trace of charge can, in that case, make its appearance 
beneath the superficial layer. If the atoms were to 
move towards the zinc they would have to compress 
electricity into it, which is impossible. However much 
they are pulled towards it, therefore, the only effect is 
to increase the pressure, or slacken the tension, of the 
negative electricity in the zinc (speaking in language 
appropriate to the cord models, Figs. 5, 6, &c.) ; in 
other words, to lower the potential of the zinc below 
that of the air near it. The same thing, only to a less 
degree, occurs with the isolated piece of copper. 
Neither metal becomes the least charged so long as 
uniformity of conditions is maintained all round it. 

But directly metallic contact between the two 
metals is effected, all the oxygen atoms at this point 
are swept away, and an unobstructed communication 
or clear passage is opened from the zinc to the copper 
for the flow of electricity. That is what metallic 
contact achieves it disturbs the uniformity by 
removing the straining atoms at one spot. Unless, 
therefore, there is some E.M.F. at their junction 
which we have good reason for asserting there 'is not, 
of any magnitude worth speaking of an immediate 
rush of negative electricity from zinc to copper, or of 
positive the other way, occurs. The copper therefore 
becomes negatively charged, the zinc becomes positive; 



54 PRODUCTION OF A CURRENT in 

the charge being effected by all the oxygen atoms 
moving a little nearer to the zinc, a little farther from 
the copper. So far everything goes on just the same 
whether the plates are in acidulated water or in 
common air. 

What happens next depends upon the difference 
between water and air in conducting power. 
Acidulated water is able to conduct electricity ; air 
is not. Accordingly, when the plates are immersed 
in water, the negative charge is continually con- 
veyed back from the copper to the zinc through 
the liquid ; whereas when they are in the air nothing 
further happens, except the slight electrostatic strain 
into which the air is thrown by the quantity of 
electricity accumulated upon the metals, positive on 
the zinc, negative on the copper charges which have 
no vent or outlet. 

Unless care is taken to make the capacity of the 
free metallic surfaces considerable, by expanding 
them over a large surface or by bringing them very 
close together, these charges will be extremely small, 
the electromotive force producing them being rather 
under than over one volt ; and accordingly the electro- 
static strain in the air near a couple of zinc and 
copper rods in contact is extremely minute. By 
delicately suspending a highly-charged aluminium 
needle near such a junction, however, Lord Kelvin 
has been able to observe the state of strain ; the 
needle when positively charged moving perceptibly 
towards the copper. A more usual method of 
displaying the phenomenon, and the one originally 
used by Volta, is to increase the capacity of the 



ii2 CONDUCTION CHAP, vi 

arrangement by bringing two carefully ground plates 
very close together. Although the E.M.F, is small 
(just the same as with a mere point contact), yet now 
the capacity is so great that quite a reasonable 
quantity of electricity can be stored in the two 
opposing metals, opposing each other across a 
microscopic air film and only touching at a few 
points ; so that, when the plates are neatly separated, 
sufficient charge is found in them to affect sensibly 
even a common gold-leaf electroscope. 

5 5. The mistake which has been, and still frequently 
is, made, with regard to this simple and not very 
important experiment, has been to regard the charge 
as evidence of a peculiar E.M.F. at the point of con- 
tact ; an E.M.F. which causes a difference of potential 
between the two metals. And this fictitious contact 
E.M.F. has then been appealed to to explain the 
voltaic battery. 

The right way of regarding the matter is to con- 
sider the battery first, explaining its action chemically 
so far as it is possible to explain it at present ; and 
then to point out that similar things will occur in air 
(an air battery, in fact), with the slight difference that 
since air is a dielectric instead of an electrolyte no 
continuous current is possible, but merely a slight 
electric displacement, which is the Volta effect. 

56. The effective cause of the whole phenomenon 
in either case is the greater affinity of oxygen for zinc 
rather than copper. This by itself would cause a 
greater strain of negative electricity towards zinc a 
slackening of the negative cords in it, to speak in the 
language of the cord model and a consequent rise of 



56 PRODUCTION OF A CURRENT 113 

negative potential. A piece of isolated zinc is there- 
fore some r8 volts below the potential of the atmo- 
sphere, the difference of potential between metal and 
oxygen being calculable direct from their experi- 
mentally known heats of combination with oxygen. 1 
The same sort of thing is true for copper, except that 
the intensity of strain is less ; as evidenced by the 
less heat of formation of CuO compared with ZnO ; 
and accordingly a piece of isolated copper is about 
0*8 volt below the potential of the atmosphere. 

Directly the two metals touch they necessarily 
become of the same potential all parts of a conductor 
are at one potential unless there are disturbing internal 
forces, and the equalization of potential is effected 
by a rush of electricity across the junction, whereby 
the zinc receives a positive charge and the copper a 
negative charge, until their potential is equalised. In 
air the equalization is effected in an instant. In water 
it is a matter of eternity. That is all the difference. 
The thing observed in the Volta effect is not a differ- 
ence of potential between zinc and copper, but a 
difference of charge ; the two metals being charged 
so as to make their potentials the same, in spite of 
their unequal chemical affinities for oxygen. 

What is observed in the Lord Kelvin form of 
the experiment ( 54) is again, not a difference of 
potential between zinc arid copper, but a slope of 
potential in the air near them, from the zinc towards 
the copper. The metals when in contact are both 

1 See Lodge "On the Seat of E.M.F. in the Voltaic Cell," Brit.. 
Assoc. Report, 1884; also Journ. Soc. Tel. Eng. 1885, and Phil. Mag. 
vol. xix. 1885 and 1886. 

1 



CONDUCTION 



CHAP. VI 



at a common potential, 1*3 volts below the atmo- 
sphere, the mean of their original potentials ; but the 
original difference of potential between each and the 
air in contact with it remains unaltered ; hence there 
is a gradual slope of potential, of i volt, from the 
layer of air in contact with zinc to the layer in con- 
tact with copper ; and this slope of potential is what 







-4 i- 




J J - 




1 


i i i 




1 5 i 



II- 



i r 



/ / 1 



Zn. 



Cu. 



FIG. ISA. Diagrammatic representation of the Volta effect, on the plan of the cord 
models (Figs. 5, 6, 7, &c.). 

I. shows a piece of zinc and copper before contact, with a cord representing 

negative electricity passing through both, and beads representing oxygen 
atoms. The arrows indicate that the oxygen is being pulled by the 
zinc on all sides of it ; and that it is also pulled by the ix:opper, but with 
less force. The two metals differ in potential from each other and from 
the air near them, but this fact results in no sensible phenomenon so long 
as they are separate. 

II. shows the effects of sweeping away the oxygen atoms between the two 

metals, by establishing metallic contact, so that the greater atom-attracting 
force of zinc over copper can now produce an effect, until it is balanced 
by the elastic stress called out by an electric displacement. The surface 
of the zinc has now less than its normal share of negative cord it is 
positively charged the copper is negatively charged. The two metals 
are now at the same potential, though oppositely charged ; and a slope of 
potential is thrown upon the air in their neighbourhood. This is the. Volta 
effect. 



the electrometer needle feels. The diagram Fig. I5A 
may possibly help in making the whole thing 
clear. 

The matter is capable of being perceived quite 
distinctly with the expenditure of a little time and 
trouble, and it is worth an effort. 



58 PRODUCTION OF A CURRENT 115 

True Contact Force. 

57. So far we have assumed that there is actually 
no force at the contact of zinc with copper. There 
is indeed none of any appreciable magnitude, but 
the force is not absolutely zero. Between some 
metals, bismuth and antimony for example, the force 
is much larger, but it is still only a few hundredths 
of a volt. It is an important thing that there can be 
a true contact force at the junction of two metals, 
only it has nothing to do with the chemically pro- 
duced Volta effect. If the Volta effect be called a 
contact force at all, it is a contact force between 
metal and air ; any true contact force between metals 
acts as a slight and insignificant disturber of the 
simple Volta effect ; and what is really observed in 
electroscopic experiments is the sum of the two. 

58. That there is a true though weak contact force 
at a junction of metals is proved by the reversible 
heat effects which are found there when a current is 
passed across the junction : a current one way pro- 
duces more heat than a current the other way. In a 
simple homogeneous piece of metal the heat produced 
by a current is utterly independent of direction : it is 
called irreversible heat ; it is proportional to the 
square of the current strength, as Joule showed. But 
at a junction of different substances, or even at a 
junction of the same substance in two different states 
two different temperatures, for example, in addi- 
tion to the irreversible heat produced by mere resist- 
ance there is a reversible heat production, one which 
changes sign with the direction of the current, so 

I 2 



n6 CONDUCTION CHAP, vi 

that a current one way actually tends to cool the 
junction instead of heating it. With care this cooling 
tendency may be got to overpower and mask the 
irreversible heat ; and a junction may be positively 
cooled and water frozen by steadily passing a mode- 
rate current in the right direction across it. 

This curious fact was discovered by Peltier. 
It may be considered as the fundamental fact of 
thermo-electricity. Its meaning is that something in 
the metals, at the junction, is helping to propel the 
current along ; doing work, in fact, and consuming its 
own heat in the process. The vibratory motion of 
the molecules is getting used up in propelling elec- 
tricity. The contact force is acting in the direction 
of the current, and the junction is thereby cooled. 

If the current be reversed, it will be driven against 
the force exerted by the molecules, and an extra 
amount of heat will then be added to the irreversible 
or frictional generation of heat. 

59. This thermal evidence of contact force, though 
the most direct, was not the earliest discovered. The 
earliest known fact in thermo-electricity was that in a 
complete circuit of different metals a current could be 
excited by having the parts at different temperatures ; 
manifestly because these contact forces, of which we 
have been speaking, change with temperature some 
increasing, others decreasing, as the temperature is 
raised. They are accurately balanced in a circuit of 
uniform temperature, but they have a resultant when- 
ever the temperature is not uniform ; and this resultant 
E.M.F. is able to propel a current, as discovered by 
Seebeck. 



60 PRODUCTION OF A CURRENT 117 

Thermo-electric Pile. 

60. A thermopile may be thought of in the follow- 
ing way, but in trying to understand the nature of 
these actions, at present, one must admit that some 
speculation and vagueness exist. 

We have seen that when electricity is propelled 
through or among the molecules of a metal it experi- 
ences a certain resistance or opposition-force which 
is exactly proportional to the speed of its motion 
( 28). In other words, there is a connection between 
matter and electricity in many respects analogous to 
fluid friction, but varying accurately as the first power 
of the relative velocity. Hence, if an atom of matter 
be vibrating about a fixed point, it will tend to drive 
electricity to and fro with it ; but if it be only one of a 
multitude, all quivering in different phases, they will 
none of them achieve any propulsion. This may be 
considered the state of an ordinary warm solid. But 
if from any cause a set of atoms could be made to 
move faster in one direction than in the reverse 
direction to move forwards quickly and backwards 
slowly then such an unsymmetrically-rnoving set 
will exert a propulsive tendency, and tend to drive a 
current of electricity forwards, simply because the 
force exerted is proportional to the velocity, and so 
is greater on the forward journey than on the return. 
Referring back to the cord model, Fig. 5, Ohm's law 
requires that the friction between cord and beads 
should be directly as the velocity ; hence if a bead 
begins to oscillate unsymmetrically, travelling forward 
quickly and back slowly, it will propel the cord along 



ii8 CONDUCTION CHAP, vi 

in the direction in which it moves most quickly, some- 
what as a child can propel its chair by jogging it upon 
a rough floor. 

It might be thought that the weaker force of the 
slow return journey would be exactly compensated 
by the extra time it lasts. And so it would if there 
were only one propeller, or if they were all in the 
same phase ; but the quick moving ones would grip 
the cord and prevent its return, if there is a balance 
of motion being transmitted in one direction ; so that 
the result would be an effect analogous to the effect 
of barbs or valves : such an effect as is typified by 
the hammering in of a nail against a much weaker, 
though much more durable, extracting force. (For an 
incomplete account see Lodge, Philosophical Magazine ^ 
December 1876.) 

Wherever conduction of heat is going on along a 
substance the atoms are moving unsymmetncally. 
They are driven forward infinitesimally quicker by the 
more rapidly moving atoms at the hot end than they 
are driven back by the less rapidly moving atoms in 
front. And, whatever may be the precise way in 
which such dissymmetry acts, such a slope of tem- 
perature exerts a propulsive tendency: there is an 
electromotive force in a substance unequally heated. 

This fact was discovered theoretically and verified 
experimentally by Lord Kelvin. 

61. But not only is there such a force at a junction 
of a hot and cold substance ; there is also a force at 
the junction of two substances of different kinds, even 
though the temperature be uniform. It is not quite 
so easy to explain how it comes about that the atoms 



62 PRODUCTION OF A CURRENT 119 

at this kind of junction are moving faster one way 
than the other; nevertheless, such a thing is not 
unlikely, considering the state of constraint and 
accommodation which must necessarily exist at the 
boundary surface of two different media. However it 
be caused, there is certainly a weak E.M.F. at such a 
junction ; which is in general greater when the resist- 
ance, or grip between electricity and metal, is greater. 

Thus, then, in a simple circuit of two metals with 
their junctions at different temperatures, there are 
altogether four electromotive forces one in each 
metal, from hot to cold or vice versa, and one at each 
junction ; and the current which flows round such a 
circuit is propelled by the resultant of these four. 1 

These four forces, two Kelvin forces in the metals, 
and two Peltier forces at their junctions, may some 
of them help and some hinder the current. Wherever 
they help, the locality is to that extent cooled ; 
wherever they hinder, it is to that extent warmed. 

Frictional Electricity. 

62. But the contact force at a junction is by no 
means confined to metals. It occurs between insula- 
tors also, and it is to this fact that the striking effects 
produced by all frictional electric machines are due. 
The essential thing in the production of " frictional 
electricity " is the contact of dissimilar substances. 
It is by their contact force that electricity gets 
transferred from one to the other, so that one becomes 
positive and the other negative. A rapid succession 

1 See Lodge, Phil. Mag. Dec, 1876 ; also June 1885. 



120 CONDUCTION CHAP, vi 

of contacts, easily obtained by sliding one of the 
surfaces over the other, is usually necessary to aid the 
transfer, the substances being so badly conducting. 

By thus noticing that the connection between 
matter and electricity, known as resistance and defined 
by Ohm's law ( 28), is competent to produce contact 
electromotive forces, we may perceive how it comes 
to pass that in good conductors such forces are so 
weak, while in insulators they are so strong. 
Electricity slips through the fingers of a metal, as it 
were, and the driving force it can exert is very feeble ; 
while an insulator gets a good grip and thrusts it 
along with violence. A consequence of the above 
view is that in perfect conductors thermo-electric and 
all other contact forces would necessarily be zero. 

Specific Relation betiveen Matter and Electricity ; 
sometimes called " Specific Heat of Electricity!' 

63. The metals differ in their gripping power ; 
and, roughly speaking, the best conductor makes 
the worst thermo-electric substance. A bad conductor, 
like antimony, or, still better, galena, or selenium, 
or tellurium, makes a far more effective thermo- 
electric element than a well-conducting metal. Not 
that specific resistance is all that has to be con- 
sidered in the matter ; there is also a specific relation 
between each metal and the two kinds of electricity. 
Thus, iron is a metal whose atoms have a better 
grip of positive than of negative electricity, and so 
a positive current gets propelled in iron from hot to 
cold. Copper, on the other hand, acts similarly on 



63 PRODUCTION OF A CURRENT 121 

negative electricity, and it is a negative current 
which is driven from hot to cold in copper. And all 
the metals can be classed with one or other of these 
two ; except perhaps lead, which appears to grip 
both equally, and so to exert no differential effect 
upon either. 

How this relation can be likened to a " specific 
heat" may be thought out by attending to the last 
paragraph of 61, and by regarding electricity as a 
material fluid (see also 182). 

Pyro-electricity. 

Certain crystals, called by mineralogists hemihedral, 
having different forms at the two ends of their axis 
which may be called the A end and the B end respec- 
tively, exhibit some properties not quite the same in 
the direction A B as in the direction B A. They are said 
to be more easily scratched, for instance, in one sense 
than in the other. Such crystals, of which the class of 
tourmalines may be taken as the type, have other 
very singular properties. Such of them as are fairly 
clear are opaque to light in a singular fashion not 
opaque to light polarized in all planes, but selectively 
opaque. Vibrations occurring perpendicular to the 
axis are rapidly quenched, so that one cannot see at 
all through a slice cut perpendicular to the axis ; 
while vibrations occurring along the axis are trans- 
mitted with but moderate absorption. This opacity 
seems quite different from the conductive opacity of 
metals, about which we shall speak later, for, in 
the first place, the light stopped is not reflected, 



122 CONDUCTION CHAP, vi 

but absorbed ; and, in the second place, a crystal 
of tourmaline is not a conductor, but a very fair 
insulator. 

And yet there are some peculiarities about such 
conducting power as it has, which are very note- 
worthy, and which may be intimately connected with 
the selective opacity which fits a slice of crystal cut 
parallel to the axis for use as a " polarizer " in optics. 
One of these peculiarities was found by Dr. S. P. 
Thompson in conjunction with the present writer : viz. 
that while, like all other uniaxial crystals, the con- 
ductivities for heat along and across the axis are not 
the same (being, in the case of tourmaline, less good 
along the axis than across), yet, in addition to this, 
a warming crystal conducts heat better in the sense 
B A than in the sense A B, while a cooling crystal 
does the opposite. While the temperature is rising 
heat gets conveyed more easily towards A than 
towards B. 1 

Whether on account of inequalities of temperature 
thus set up, or for some more direct reason, electricity 
gets automatically transferred in one direction rather 
than in the other. And accordingly, while a crystal 
is rising in temperature, positive electricity accumu- 
lates at the A end, and negative electricity at the B 
end. So long as the temperature remains constant 
nothing further happens ; except ordinary leakage, 
principally no doubt ever the surface, which may in 
time completely mask the effect produced. On now 
cooling the crystal, an inverse electrification will be 
set up ; or, if no leakage has been permitted, the 

^ Phil. Mag. July 1879. 



64 CURRENT THROUGH GASES 123 

effect of cooling is simply to replace the electricity 
displaced by the warming. 

While the temperature of the crystal is steady, no 
difference in electric conductivity has been detected 
by the writer, between the sense A B and the sense 
B A. Neither is there any difference in the thermal 
conductivity when the temperature is steady. Both 
effects depend on a varying temperature. But the 
electrical phenomenon is more than a unilateral con- 
ductivity, it is a true axial electromotive force. 



Passage of Electricity through a . Gas. 

64. There remains to be said something about 
the way in which electricity can be conveyed by 
gases : a subject which has grown immensely of late 
years. But here we will be brief, and deal with the 
matter in another book. 

The first thing to notice is that gases and vapours 
can be divided into two classes : those which 
conduct electricity, and those which do not. 
Those which conduct, do so electrolytically that 
is to say, they conduct like liquids by the loco- 
motion of charged or carrier atoms, or ions ; and there 
is good evidence that such gases are in a state of 
dissociation so long as they possess any conducting 
power. A high temperature, or a recent electrical 
discharge, either of which is well known to be a 
dissociating agency, is found to confer upon some 
gases and vapours a conducting power which they do 
not possess in their ordinary state. A gas whose mole- 



124 CONDUCTION CHAP, vi 

cules are dissociated into charged ions becomes an 
electrolyte, and is said to be ionised. 

Undissociated or non-conducting gases and vapours 
appear not to conduct in the slightest degree ; in 
other words, a substance in this condition behaves 
as a perfect insulator perhaps the only perfect 
insulator there is. Not water vapour, not even 
mercury vapour, is found to conduct in the least ; 
except when ionised. An electrified liquid does not 
lose its charge by evaporation : steam carries off no 
electricity, though spray does. This shows that mere 
bombardment of molecules, such as is known to go 
on in gases, is not sufficient either to remove or to 
impart any electric charge. Schuster and J. J. 
Thomson have found, however, that some gases do 
conduct electrolytically ; and Schuster has described 
several curious facts concerning their conductivity 
during and shortly after an electric discharge. 

The commonest way in which electricity makes its 
way through a non-conducting gas, setting aside the 
mere mechanical conveyance by solid carrier, is that 
of disruptive discharge. 

When one says that a gas does not ordinarily act 
as a common electrolyte, the experimental grounds of 
the statement are that a finite electrostatic stress is 
certainly possible in its interior a stress of very con- 
siderable amount ; and when this stress does overstep 
the mark and cause the material to yield, the yielding 
is evidently not a quiet and steady glide or procession, 
but a violent breaking down and collapse, due to in- 
sufficient tenacity of something. One may therefore 
picture the molecules of a gas, between two opposite 



6s CURRENT THROUGH GASES 12$ 

electrodes or discharge terminals maintained at some 
great difference of potential, as arranged in a set of 
parallel chains from one to the other, and strained 
nearly up to the verge of being torn asunder. In 
making this picture one need not suppose any fixture 
of individual molecules : there may be a wind blowing 
between the plates ; but all molecules as they come 
into the field must experience the stress, and be 
relieved as they pass out. 

65. If the applied slope of potential overstep a 
certain limit, fixed by observation at something like 
33,000 volts per linear centimetre for common air, the 
molecules give way, the atoms with their charges rush 
across to the plates, and discharge has occurred. The 
number of atoms thus torn free and made able to 
convey a charge by locomotion is so great that there 
is ordinarily no difficulty in conveying any amount 
of electricity by their means. In other words, during 
discharge the gas becomes a conductor ; and, being 
a conductor by reason of locomotion of atoms it 
may be called an electrolyte. 

Schuster discovered that this conductivity of gases, 
or power of equalizing even very feeble potentials 
lasts for a little time and extends to some little 
distance from the region of a disruptive discharge ; 
as if a number of ions or charged carrier atoms 
had been liberated and diffused, for a time, until 
they recombined and neutralized each other once 
more. 

But whether the charge then possessed by each 
carrier atom intrinsically belonged to it all the time, 
or whether it was conferred upon the components of 



126 CONDUCTION CHAP, vi 

the molecules during the strain and the disruption, is 
a point not yet decided. 

What is called " the dielectric strength " of a gas 
that is, the strain it can bear without suffering dis- 
ruption and becoming for the instant a conductor 
depends partly on the nature of the gas, and very 
largely on its pressure. Roughly, one may say 
that a gas at high pressure is very strong, a gas 
at low pressure very weak. An ordinary electrolyte 
might be called a dielectric of zero strength. 

One reason why pressure affects the dielectric 
tenacity of a gas may be suggested : it is certainly 
not the only one, but it can hardly help being at 
least partially a vera causa ; and that is the fact 
that in a rare gas there are fewer molecules between 
the plates to share the strain between them. 



A Current regarded as a Moving Charge. 

66. To review the ground we have covered so far. 
We first tried to get some conception of the nature 
of electrostatic charge, and the function of a dielectric 
medium in static electricity. We next proceeded to 
see how far the phenomena of current electricity could 
be explained by reference to electrostatics. For a 
current, being merely electricity in locomotion, need 
consist of nothing but a charged body borne rapidly 
along. 

Charge a sphere with either positive or negative 
electricity, and throw it in some direction ; this con- 
stitutes a positive or a negative current in that 
direction. There is nothing necessarily more occult 



67 CURRENT THROUGH GASES 127 

about a current than that. And a continuous current 
between two bodies may be kept up by having a lot 
of pith balls, or dust particles, oscillating from one to 
the other, and so carrying positive electricity one way 
and negative the other way. But such carriers, as 
they pass each other with their opposite charges, 
would be very apt to cling together and combine. 
They might be torn asunder again electrically, or they 
might be knocked asunder by collision with others. 
Unless one or other of these things happened, the 
current would shortly have to cease, and nothing but 
a polarized medium would result. 

Instead of pith balls, picture charged atoms as so 
acting, and we have a rough image of what is going 
on, in an electrolyte on the one hand and in a di- 
electric on the other. The behaviour of metals and 
solid conductors is more obscure. Locomotive car- 
riage by progressing atoms, is not to be thought of in 
them ; but, inasmuch as no new phenomenon appears 
in their case, it is natural to try and picture the pro- 
cess as one not wholly dissimilar : to picture some- 
thing as being passed on from hand to hand, by a 
set of merely vibrating atoms, that something being 
electricity. And this is what in 27 we tried to do. 

67. I have said that an electric current need be 
nothing more occult than is a charged sphere moving 
rapidly ; and a good deal has been made out concern- 
ing currents by minutely discussing all that happens 
in such a case. But, even so, the problem is far from 
being a simple one. One has to consider not only the 
obviously moving charge, but also the opposite induced 
charge tied to it by lines of force (or tubes of indue- 



128 CONDUCTION CHAP, vi 

tion, as they are sometimes called), and we have this 
whole complicated system in motion. And the effect 
of this motion is to set up an altogether new pheno- 
menon in the medium a spinning kind of motion that 
would not naturally have been expected : whereby 
two similarly charged spheres in motion repel one 
another less than when stationary, and may even 
begin to attract, if moving fast enough ; whereby also 
a relation arises between electricity and magnetism, 
and the moving charged body begins slightly to 
deflect a compass needle ( 113 and 184). Of which 
more in the next Part. 



PART III 

MA G N E T I S M 



K 



CHAPTER VII 

RELATION OF MAGNETISM TO ELECTRICITY 

68. WE next proceed to consider electricity in a 
state of rotation. What happens if we make a whirl- 
pool of electricity ? Coil up a wire conveying a 
current, and try. The result is it behaves like a 
magnet : compass-needles near it are deflected, steel 
put near it gets magnetized, and iron nails or filings 
are attracted by it, sucked up into it if the current 
be strong enough. In short, it is a magnet. Not of 
course a permanent one, but a temporary one, lasting 
as long as the current flows. It is thus suggested that 
magnetism may perhaps be simply electricity in 
rotation. Let us work out this idea more fully. 

First of all, one may notice that everything that 
can be done with a permanent magnet can be imitated 
by a coiled wire conveying a current. (It would not 
do altogether to make the converse statement.) Float 
a coil attached to a battery vertically on water, and 
you have a compass-needle ; it sets itself with its axis 
north and south (Fig. 16). Suspend two coils, and 
they will attract or repel or turn each other round just 
like two magnets. 

K 2 



1 3 2 



MAGNETISM 



CHAP. VII 



69. As long as one only considers the action of a 
coil at some distance from itself, there is no need to 
trouble about the shape of the particular magnet 
which it most closely simulates ; but as soon as one 
begins to consider the action of a coil on things close 
to it, it is necessary to specify the shape of the 
corresponding magnet. 

If the coil be a long cylindrical helix like a close- 
spired corkscrew, as in Fig. 16, it behaves like a 




FIG. 16. Floating battery and helix acting as a compass-needle, 

cylindrical magnet filling the same space. If the coil 
be a short wide hank, like a curtain-ring, it behaves 
again like a cylindrical magnet, but one so short that 
it is more easily thought of as a disk. A disk or 
plate of steel magnetized with one face all north and 
the other face all south can be cut to imitate any thin 
hank of wire conveying a current. It will be round 
if the coil be round, square if it be square, and 
irregular in outline if the coil be irregular. It is called 
a magnetic shell. 

There is no need for the coil to have a great num- 
ber of turns of wire, except to increase its power : 
one is sufficient, and it may be of any shape or 



70 ELECTRO-MAGNETISM 133 

size. So when we come to remember that every 
current of electricity must necessarily flow in a 
closed circuit, one perceives that every current of 
electricity is virtually a coil of more or less fantastic 
shape, and accordingly imitates some magnet or other 
which can be specified. Thus we learn that every 
current of electricity must exhibit magnetic phe- 
nomena : the two are inseparable a very important 
truth. See Appendix (a). 

There is one detail in which the magnetized disk 
and the coil are not equivalent, and the advantage lies 
on the side of the coil : it has a property beyond that 
possessed by any ordinary magnet. It has a pene- 
trable interior, which the magnet has not. For space 
outside both, they simulate each other exactly ; for 
space inside either, they behave differently. The coil 
can be made to do all that the magnet can do ; but 
the magnet cannot in every respect imitate and 
replace the coil : else would perpetual motion be an 
every-day occurrence. 

70. Now I want to illustrate and bring home 
forcibly the fact that there is something rotatory 
about magnetism something in its nature which 
makes rotation an easy and natural effect to obtain, 
if one goes about it properly. One will not observe 
this by taking two magnets : one will see it better by 
taking a current and a magnet, and studying their 
mutual action. 

A magnet involves, as you know, two poles a 
north and a south pole of precisely opposite pro- 
perties : it may be considered as composed of these 
two poles, for many purposes ; and the action of a 



134 MAGNETISM CHAP, vn 

current on a magnet may be discussed as compounded 
of its action on each pole separately. Now how does 
a current act on a magnetic pole? Two currents at- 
tract or repel each other ; two poles attract or repel 
each other ; but a current and a pole exert a mutual 
force which is neither attraction nor repulsion : it is a 
rotatory force. They tend neither to approach nor to 
recede ; they tend to revolve round each other. A 
singular action this, and at first sight unique. All 
ordinary actions and reactions between two bodies 
take place in the line joining them : the force 
between a current and a pole acts exactly at right 
angles to the line joining them. 

Helmholtz long ago (in 1847) showed that the con- 
servation of energy could only be true if forces 
between bodies varied in some way with distance and 
acted in the line joining them. Now here is a case 
where the force is not in the line joining the bodies, 
and accordingly the conservation of energy is defied : 
the two things will revolve round each other for ever. 
This affords, and has afforded, a fine field for the 
perpetual motionist ; and if only the current would 
maintain itself without a sustaining power, perpetual 
motion would in fact be attained. But this after all 
is scarcely remarkable, for the same may be said of a 
sewing-machine or any other piece of mechanism : if 
only it would continue to go without sustaining power 
it would be a perpetual motion. Attend to pole and 
current only, and energy is not conserved, it is per- 
petually being wasted ; but include the battery as an 
essential part of the complete system, and the mystery 
disappears : everything is perfectly regular. 



7i 



ELECTRO-MAGNETISM 



135 



71. The easiest way perhaps of showing the rotation 
of a conductor conveying a current round a magnetic 
pole, is to take an 8-feet long piece of gold thread, such 
as is used for military uniforms, and, hanging it verti- 
cally, supply it with as strong a current as it will stand. 
Then bring near it a vertical bar-magnet, and in- 

ra 




Fig. 17. A long flexible conductor twisting' itself into a spiral round a powerful 
bar-magnet raised to meet it. 

stantly you will see the thread coil itself into a spiral : 
half of it twisting round the north end of the bar, 
and half twisting, as part of the same spiral, round 
the south end (Fig. 17). 

If the magnet were flexible and the conductor rigid, 
instead of vice versa, the magnet would in like manner 



136 



MAGNETISM 



CHAP, VII 



coil itself in a spiral round the current ; the force is 
strictly mutual. A rigid magnet, put near a stiff con- 
ductor, shows only the last remnants of this action : 
it sets itself at right angles to the wire, and ap- 
proaches its middle to touch it, but that is all it 
can do. 

The experiment with the flexible gold thread is 
simple, satisfactory, and striking, but the rotatory 




Fig. 18. Pivoted disk with radial current, revolving in a magnetic field and winding 
up a weight. The current is supplied to the axil by screw A, and leaves the rim 
by mercury trough M. The same apparatus obviously serves to demonstrate 
currents induced by motion ; both directly and by the damping effect. 

properties connected with a magnet may be illustrated 
in numbers of other ways. Thus, pivot a disk at its 
centre, and arrange some light contact to touch its 
edge, either at one point or all round, it matters not ; 
then supply a current to disk from centre to circum- 
ference, and bring a bar-magnet near it along its axis, 
or, better, two bar-magnets, with opposite poles one 



ELECTRO-MAGNETISM 



137 



on each side, near the contact place of the rim ; the 
disk at once begins to rotate (Figs. 18 and 19). 




Fig. 19. Another pivoted disk with flange to dip into liquid so as to make contact 
all round its rim. It rotates when a magnet is brought above or below ; or even 
in the field of the earth. 




Fig. 20. A couple of radii of the above disk, provided with points to dip into mercury 
and rotating constantly under the influence of the steel magnet A ; or under the 
influence of a current conveyed by a coil of wire round the circular trough. 

Instead of a disk one may use a single radius of it, 
viz. a pivoted arm (Fig. 20) dipping into a circular 
trough of mercury ; or we may use a light sphere 



138 



MAGNETISM 



CHAP. VII 



rolling on two concentric circular lines of railway 
(Gore's arrangement, Fig. 2I). 1 In every case 
rotation begins as soon as a magnet is brought near. 

72. Nor is the revolving action confined to metallic 
conductors and to true conduction. Liquids and 
gases, although they convey electricity by something 




Fig. 21. Gore's circular railway. The light spherical metal ball revolves round the 
two concentric metal hoops or rails whenever it is made to convey a current 
between them in a vertical magnetic field. 

of the nature of convection, are susceptible to rotation 
in a precisely similar manner. 

1 This is not what Gore's railway is commonly used to illustrate, nor 
is it the cause of the motion as observed by the inventor, or as described 
in Tyndall's Heat. Ordinarily the ball moves by reason of an irregular 
disturbance due to heat at its point of contact with the rails, and it is 
mere accident which way it goes. But, in so far as the earth's vertical 
magnetic field is strong enough, it should exhibit a preference for one 
direction over the other ; and if the field is strengthened by bringing the 
south pole of a bar-magnet below the apparatus, true magnetic rotation 
is bound to occur. It may, however, be convenient to state that the 
current's own lines of force are powerless to cause continuous motion in 
this case. An external field is essential in cases like this of fixed and 
finite area for the circuit. All that the current's own magnetism could do 
would be very slightly to urge the ball to the end of a diameter opposite 
to where the current enters. 



72 ELECTRO-MAGNETISM 139 

To show the rotation of liquid conductors under 
the influence of a magnet, take a circular shallow 
trough of liquid, supply it with stout sheet copper 
electrodes at centre and circumference, and put the 
pole of a magnet below it. The liquid at once begins 
to rotate, and by using a magnet and current of fair 
strength it can easily be made to whirl so fast as to fly 
over the edge of the trough (Fig. 22). 1 The experi- 
ment is plainly the same as Fig. 19, except that a 




Fig. 22. Rotation of a liquid disk conveying a radial current in a vertical 
magnetic field. 



liquid disk is used in place of a solid one. Or, again, 
it may be considered the same as Fig. 21. Reverse 
the magnet, and the rotation is rapidly reversed. 
Another method is to send a current along a jet 

1 In practice it is most convenient to split a battery current between 
magnet and liquid : i. e. to connect them in parallel instead of in series. 
It is also well to make the smaller surface of copper the cathode ; 
because with intense currents (say 3 amperes per square centimetre) 
a crust of oxide forms on the anode which almost entirely stops the 
current by its resistance. This alone is a curious experiment, worth 
trying with small copper electrodes. 



140 



MAGNETISM 



CHAP. VII 



of mercury near a magnet and note the behaviour of 
the jet. It twists itself into a flat spiral as shown in 
Fig. 23. 

The rotation of a gas discharge is most commonly 
illustrated by an arrangement like Fig. 24, where the 




u u 



/A 4\ It 

Fig. 23. A falling stream of liquid conveying a current between two magnetic poles, 
and being thereby twisted into a spiral. (Copied from a paper in Phil. Mag. b> 



Dr. Silvanus Thompson.) 



terminals of the induction coil are connected to the 
rarefied gas respectively above one pole and round the 
middle of a magnetized bar. If the discharge can be 
got to concentrate itself principally down one side, 



73 



ELECTRO-MAGNETISM 



141 



which it is not easy always to do (it seems to depend 
on the presence in the vacuum of some traces of 
foreign vapour, e.g. CS 2 vapour), the line of light so 
formed is seen to revolve. 




Fig. 24. Induction coil discharge from a to b through rarefied ga, rotating round 
a glass protected magnetized iron rod. 



Action between a Magnet and an Electric Charge in 
Relative Motion. 

73. Remembering all this, and the fact that a 
moving charge constitutes a current, it is not to be 
doubted that a charged pith ball moving in the 
neighbourhood of a magnet is subject to the same 
rotatory action. There is no known action between a 
magnet and a stationary charged body, but directly 



142 MAGNETISM CHAP, vi 

either begins to move there is an action between 
them, tending to cause one to rotate round the other. 
It is true that for ordinary speeds of motion this force 
is extremely small ; but still it is not to be doubted 
that if a shower of charged pith balls or Lycopodium 
granules are dropped on to a magnet pole, they will 
fall, not perfectly straight, but slightly corkscrew 
fashion. In this way electrified particles, shot out by 
the sun, are guided spirally along the earth's magnetic 
line of force ; and are thus led from the tropics to the 
poles, where they illuminate the rarefied atmosphere 
electrically. 

If a set of charged particles were projected horizont- 
ally and radially from the top of a magnet, their paths 
would revolve like the beams of a lighthouse. And if 
by any means their paths were kept straight, or deflected 
the other way, they would exert on the magnet an in- 
finitesimal "couple " tending to make it spin on its own 
axis. Conversely, if a magnet were spun on its axis 
rapidly by mechanical means, there is very little doubt 
but that it w r ould act on charged bodies in its neigh- 
bourhood, tending to make them move radially either 
to or from it. This, however, is an experiment that 
ought to be tried ; and the easiest way of trying it 
would be to suspend a sort of electrometer needle, 
electrified positive at one end and negative at the other, 
near the spinning magnet, and to look for a trace of 
deflection to be reversed when the spin is reversed, 
A magnet of varying strength might be easier to try 
than a spinning one. 1 (See 114 116.) 

1 For experiments directed to discover this effect and exhibiting 
a very minute trace of it, see Lodge, Phil. Mag. June 1889, p. 469. 



74 



ELECTRO-MAGNETISM 



143 



Rotation of a Magnet by a Current. 

74. The easiest way to show the actual rotation of 
a magnet is to send a current half-way along it and 
back outside. Thus take a small, round, polished 
steel bar-magnet with pointed ends, pivot it vertically, 
and touch it steadily with two flakes or light pads of 
tin-foil, one near either end and one near the middle ; 
supply a current by these contact pieces, and the 



=5 




Fig. 25. Round bright steel bar-magnet pivoted at its ends, spinning rapidly on its 
axis under the influence of a current supplied to either the bottom or top pivot, 
or both, and removed near the middle by a scrap of tin-foil lightly touching it. 

magnet spins with great rapidity. Reverse the current, 
and it rotates the other way. Conversely, by pro- 
ducing the rotation mechanically a current will be 
excited in a wire joining the two pieces of tin-foil 
(Figs. 25, 26, and 27). 

The two contacts may be made anywhere on the 



144 



MAGNETISM 



CHAP. VII 



magnet except symmetrically : if the two are equi- . 
distant from the middle, no effect will be produced. 
The nearer one is to the middle and the other to 




Fig. 26. Another mode of exhibiting the same thing as Fig. 25. The magnet is 
loaded so as to float upright in mercury. 




Fig. 27. The converse of Fig. 25. Spinning the magnet mechanically gives a current 
between two springs, one touching it near or beyond either end, the other 
touching it near the middle. 

the end the stronger the effect ; stronger still if 
one of the contacts is either at or beyond the end, 
as in Fig. 25, where the lower cup contains mercury. 



75 ELECTRO-MAGNETISM 145 

75. The customary or Faraday plan of exhibiting 
the effect depicted in Fig. 25, with a mercury ring 
trough round the magnet, into which a projecting wire 
carried by the magnet dips, is not quite so simple and 
obvious a method as Fig. 25 ; neither is it so effective 
unless the ring trough fits the magnet pretty closely. 
The arrangement in Fig. 25, where the contact is 
made actually on the surface of the magnet, gives the 
theoretically greatest force. 

Many more variations of the experiment could be 
shown, but these are typical ones, and will suffice. 
They all call attention to the fact that a magnet, 
considered electrically, is a rotary phenomenon. 



CHAPTER VIII 

NATURE OF MAGNETISM 

Ampere's Theory. 

76. THE idea that magnetism is nothing more nor 
less than a whirl of electricity is no new one it is 
as old as Ampere. Perceiving that a magnet could 
be imitated by an electric whirl, he made the hypo- 
thesis that an electric whirl existed in every magnet 
and was the cause of its properties. Not of course 
that a steel magnet has an electric current cir- 
culating round and round it, as an electro-magnet 
has ; nothing is more certain than the fact that a 
magnet is not magnetized as a whole, but that each 
particle of it is magnetized, and that the actual magnet 
is merely an assemblage of polarized particles. The 
old and familiar experiment of breaking a magnet 
into pieces proves this. Each particle or molecule of 
the bar must have its circulating electric current, and 
then the properties of the whole are explained. 

There is only one little difficulty which suggests 
itself in Ampere's theory How are these molecular 
currents maintained ? Long ago a similar difficulty 
was felt in astronomy What maintains the motions 



76 NATURE OF MAGNETISM 147 

of the planets ? Spirits, vortices, and other contriv- 
ances were invented to keep them going. 

But in the light of Galileo's mechanics the difficulty 
vanished. Things continue in motion of themselves 
until they are stopped. Postulate no resistance, and 
motion is essentially perpetual. 

What stops an ordinary current? Resistance. 
Start a current in a curtain-ring, by any means, and 
leave it alone. It will run its energy down into heat, 
in the space of half a second or so. But if the metal 
conducted infinitely well there would be no such 
dissipation of energy, and the current would be 
permanent. 

In a metal rod, electricity has to pass from atom to 
atom, and it meets with resistance in so doing ; but 
who is to say that the atoms themselves do not 
conduct perfectly ? They are known to have various 
infinite properties already ; they are infinitely elastic, 
for instance. Pack up a box of gas in an adiabatic 
enclosure for a century, and see whether it has got 
any cooler. The experiment, if practicable should be 
tried ; but our present experience warrants us in 
assuming no loss of motion among colliding atoms, 
until the contrary has been definitely proved by ex- 
periment. To all intents and purposes certainly atoms 
are infinitely elastic ; why should they not also be 
infinitely conducting? Why should dissipation of 
energy occur in respect of an electric current cir- 
culating wholly inside an atom ? Excepting for loss 
by radiation, there is no known reason why it should : 
there are many analogies against it. Magnetic proper- 
ties are probably due to such currents. 

L 2 



148 MAGNETISM CHAP, vm 

How did these currents originate ? We may as 
well ask, How did any atomic qualities originate ? 
How did their motion originate ? These questions are 
unanswerable. Suffice it for us, there they are. The 
atoms of a particular substance iron, for instance, 
or zinc have an electric whirl of certain strength 
circulating in them as one of their specific physical 
properties. 

This much is certain, that the Amperian currents 
are not producible by magnetic experiments. When 
a piece of steel or iron is magnetized, the act of 
magnetization is not an excitation of Amperian 
current in each molecule is not in any sense a 
magnetization of each molecule. The molecules were 
all fully magnetized to begin with ; the act of mag- 
netization consists merely in facing them round so as 
to look mainly one way in polarizing them, in fact. 
This was proved by Beetz long ago : I will not stop 
to explain it further, but will refer students to 
Maxwell (vol. ii. chap. vi.). 



Ampere's Theory extended by Weber to explain 
Diamagnetism also. 

77. Let us see how far we have got. We have made 
the following assertions : 

(1) That a magnet consists of an assemblage of 
polarized molecules. 

(2) That these molecules are each of them perman- 
ent magnets, whether the substance be in its ordinary 
or in its magnetized condition, and that the act of 



78 NATURE OF MAGNETISM 149 

magnetization consists in turning them round so as to 
face more or less one way. 

(3) That when all the molecules are faced in the 
same direction the substance is magnetically com- 
pletely saturated. 

(4) That if each molecule of a definite substance 
contains an electric current of definite strength, cir- 
culating in a channel of infinite conductivity, the 
magnetic behaviour of the substance is completely 
explained. 

But now, supposing all this granted, how comes it 
that the molecular currents are not capable of 
being generated by magnetic induction ? And if 
we cannot excite them, are we able to vary their 
strength ? 

78. The answer to these questions is included 
in the following propositions, which I will now for 
convenience state, and then proceed to explain and 
justify. 

(5) If a substance possessing these molecular 
currents be immersed in a magnetic field, all those 
molecules which are able to turn and look along the 
lines of force in the right direction will have their 
currents weakened ; but on withdrawal from the field 
they will regain their normal strength. 

(6) If the currents naturally flowing in conducting 
channels be feeble or nil, the act of immersion of the 
substance in a magnetic field will reverse them or 
excite opposite currents, which will last so long as 
the body remains in the field, but will be destroyed 
by its removal. 

(7) The molecular currents so magnetically in- 



150 MAGNETISM CHAP, vm 

duced are sufficient to explain the phenomena of 
diamagnetism . 

79. Let us first just recall to mind the well-known 
elementary facts of current induction. A conducting 
circuit, such as a ring or a coil of wire, suddenly 
brought near a current-conveying coil or a magnet, has 
a momentary current induced in it in the opposite 
direction to the inducing current : in other words, 
such as to cause momentary repulsion between the 
two. So long as it remains steady, nothing further 
happens ; but on withdrawing it, another momentary 
current is induced in it, in the contrary direction to 
the first excited. The shortest way of expressing 
the facts quite generally is to say, that, while from 
any cause the magnetic field through a conductor 
is increasing in strength, a momentary current is 
excited in it tending to drive it out of the field ; 
and that, whenever the magnetic field decreases again 
to its old value, a reverse flow of precisely the 
same quantity of electricity occurs. Fig. 28 shows 
a mode of illustrating these facts. A copper disk 
is supported at the end of a torsion arm and brought 
close to the face of an unexcited bar electro-magnet 
On exciting the magnet the disk is driven violently 
away: to be sucked back again, however, whenever 
the magnetism ceases. 

80. Now, why are all these effects so temporary ? 
What makes the induced current cease so soon after 
excitation ? Nothing but dissipation of energy : only 
the friction of imperfect conductivity. There is nothing 
to maintain the current, it meets with resistance in 
its flow through the metal, and SQ it soon stops. 



8o 



NATURE OF MAGNETISM 



But in a perfect conductor, like a molecule, no such 
dissipation would occur. Electricity in such a body 
will obey the first law of motion, and continue to 
flow till stopped by applied force. Destroying the 
magnetic field will stop an induced molecular current, 
but nothing else will stop it. Hence it follows that 




FIG. 28. Stout disk of copper supported on a horizontal arm near one pole of a bar 
electro-magnet. The disk is repelled while the magnet is being excited, and is 
attracted while the magnetism is being destroyed. An alternating current round 
the magnet maintains continual repulsion, because of a lag of phase in the 
alternating current induced in the disk (cf. 202). 

the repulsion experienced by a molecule is no transi- 
tory effect like that in Fig. 28, but is as permanent as 
the magnetic field which excites and exhibits it. 

Thus, then, a body whose molecules are perfectly 
conducting, but without specific current circulating in 
them, will behave diamagnetically, i.e. will move away 
from strong parts of the field towards weak ones ; or, 
if only free to point, will set its length equatorially, 
just as bismuth is known to do. 



152 MAGNETISM CHAP, vin 

Whether this be the true explanation of diamagnet- 
ism or not, it is at least a possible one. It seems to me 
extremely probable. It is known as Weber's theory. 

It does not necessarily follow that the specific 
molecular currents of a diamagnetic substance are 
really nil ; all that is needful is that they shall be 
weaker than those induced by an ordinary magnetic 
field. By using an extremely weak field, however, the 
specific currents need not be quite neutralized, and in 
such a field the body ought to behave as a very 
feebly magnetic substance. Such an effect has been 
looked for. 1 

8 1. It is however just possible for a substance to 
possess specific molecular currents and yet not to be 
sensibly magnetic : every molecule might be so 
jammed as to be unable to turn round, and such a 
substance could hardly exhibit any noticeable mag- 
netic properties. The molecules would have got 
themselves into a state of minimum potential energy, 
and if jammed therein nothing could be got out of 
them. The induced currents of diamagnetism would 
be superposed upon them just as if no initial molecular 
currents existed. By varying the temperature of such 
a substance, however, one might expect to alter its 
molecular arrangement, and so develop magnetic 
properties in it, just as electrical properties are deve- 
loped in crystals like tourmaline by heat or by cold. 

We are now able clearly to appreciate this much 
that the molecular currents needful to explain mag- 
netism are not conceivably excited by the act of 

1 See Lodge on Permanent Magnetic Polarity in many Substances, 
Nature, vol. xxxiii. 



82 NATURE OF MAGNETISM 153 

magnetization, for they are in the wrong direction. 
Induced molecular currents will be such as to cause 
repulsion : those which cause attraction must have 
existed there before, and be merely rotated into fresh 
positions by the magnetizing force. An intense mag- 
netic field will weaken them, and thus tend to render 
a magnetic substance less magnetic. 



Function of the Iron in a Magnet. Tivo Modes of 
expressing it. 

82. We can now explain the function of iron, or 
other magnetic substance, in strengthening a magnetic 
field. Bend a long coil of wire, Fig. 29, and send 
a current round it : there is a certain field a certain 
number of lines of force between its ends. Fill the 
coil with iron, so as to make it a common electro- 
magnet, and the strength of the field is greatly in- 
creased. Why? The common mode of statement 
likens the magnetic circuit to a voltaic circuit ; there 
is a certain magneto-motive force, and a certain re- 
sistance, or, as Mr. Heaviside preferably calls it, " re- 
luctance " : the quotient gives the resulting magnetic 
induction, or total number of lines of force. Iron is 
more permeable than air say, 3000 times more 
permeable and accordingly the resistance of the iron 
part of the circuit is almost negligible in comparison 
with that of the air-gap between the poles. Thus a 
good approximation to the total intensity of field is 
obtained by dividing the magneto-motive force by the 
width of the air-gap ; or more completely and generally 
by treating the varying material and section of a 



154 MAGNETISM CHAP, vin 

magnetic circuit just as the varying material and 
section of a voltaic circuit is treated, and so obtaining 
its total resistance. Iron is thus to be regarded as a 
magnetic conductor between 100 and 10,000 times 
better than air. Its specific magnetic conductivity or 
inductivity, or, as it is more usually called (after 
Kelvin), permeability, is measured by the ratio of 




FIG. 29. A nearly closed magnetic circuit. 

the magnetization produced to the magnetizing force 
applied, and is generally denoted by the symbol /x. 1 

83. This mode of regarding the case is undoubtedly 
simple and convenient, its thorough recognition has 
greatly improved the construction of electro- magnets 
and dynamo-machines, and it is difficult to over- 
estimate its practical convenience ; nevertheless it is 
well to remember that it is not the fundamental mode. 
If we look at the matter less with a view to practical 
simplicity than with the aim of seeing what is really 
going on, we shall express it thus : 

Before the iron was inserted in the coil, there were 
a certain number of circular lines of force inside it 
1 See Appendix (b]. 



83 NATURE OF MAGNETISM 155 



due to the current alone. A piece of common iron, 
although full of polarized molecules, has no external 
or serviceable lines of force : they are all shut up, as 
it were, into little closed circuits inside the iron. But 
directly the iron finds itself in a magnetic field some 
of these open out, a chain of polarized molecules is 
formed, and the lines due to its molecular currents 
add themselves to those belonging to the current of 
the magnetizing helix. 

Thus our ring electro-magnet has now not only its 
own old lines of force, but a great many of those 
belonging to the iron also, which have sympathetically 
laid themselves alongside the first. 

Parenthetically we may make the following remark. 
So long as the iron adds some 3000 lines of its own 
(more or less according to the quality of the iron) for 
every one otherwise excited in the field, so long it 
has its maximum permeability: it is infinitely far from 
saturation. But after a certain call upon it, it begins to 
show signs of poverty, and ultimately may refuse to 
add any lines of its own at all ; it is then said to be 
completely saturated : its permeability is then just 
as if it were air. The permeability of iron is an 
extremely indefinite quantity. Not only does it vary 
with the same piece as it nears saturation, but it is 
exceedingly different for different specimens. Thus 
some manganese-steel exists with a permeability one 
and a half times that of air, or only about as much as 
zinc, while Ewing finds some iron with a permeability 
as high as 20,000, under shaking. 

The end result of either mode of regarding the 
matter is of course the same the lines of force 



156 MAGNETISM CHAP, vin 



between the poles are increased in number by the 
presence of iron ; but whereas, in the first-mentioned 
mode of treatment, the fact of permeability had to be 
accepted unexplained, in the second nothing is unex- 
plained except the fundamental facts of the subject, 
such as the reason why currents tend to set them- 
selves with their axes parallel, and other matters of 
that sort. 

Permanent Magnetism. 

84. There is one curious effect of introducing iron 
or other solid magnetic medium into a magnetic field 
which must not be overlooked. This effect depends 
on the solidity of the substance, i.e. the fixedness or 
stiffness of its molecules. In a fluid the molecules 
are free to take up any fresh arrangement with ease : 
there is no set arrangement in the internal structure 
of a fluid, any more than there is a definiteness in 
its external shape. But with a solid it is different : 
its molecules once set into any position tend to remain 
more or less in that position ; the substance may be 
elastic to a certain extent, but after large disturbances 
there will always be a certain amount of " permanent 
set." Hence it is that solid bodies have a definite 
shape, which it requires force to change ; hence it is 
also that their molecules are able to crystallize into 
geometrical patterns. 

Now, since the act of magnetization consists in 
making a number of already polarized molecules face 
round more or less in one direction, it follows that 
solid magnetic sijbstances will behave differently from 



84 NATURE OF MAGNETISM 157 

fluid ones. In fluid media the magnetized arrange- 
ment can only be maintained by a continuous 
exertion of magnetizing force ; and directly this is 
withdrawn the molecules will quickly take up their 
old higgledy-piggledy arrangement of minimum 
energy, and all trace of magnetization will cease. 
They will be perfectly easy to magnetize, and they 
will automatically demagnetize themselves. But with 
solids it is otherwise. If the molecules have been set in 
their magnetized position byonly a feeble force, they will 
spring back almost completely when the magnetizing 
force is removed ; but if they have been arranged by a 
force of some violence, the spring back will be only 
partial, and a permanent set will remain. The spring 
back portion of the whole arrangement is called 
temporary magnetism ; the set portion is called 
permanent magnetism. The difference can be illus- 
trated by bending a bit of tin plate or paper, nearly 
double, and then letting it go. 

Substances differ greatly in their power of thus 
retaining magnetization : and, as is well known, steel 
has the property well ^developed ; but all substances 
exhibit it more or less. 1 Moreover, many substances 
can retain a little of the set if they are left carefully 
undisturbed, but they lose it if shaken or heated : 
sometimes even if gently touched. A long thin bar 
of soft iron is most instructive in this respect. It can 
be easily magnetized by the earth's magnetism if held 
vertically and struck with a finger. If then inverted 
slowly and cautiously, it will retain nearly the whole 
of the induced magnetism ; but if struck again, or 

1 See Lodge in Nature, vol. xxxiii. p. 484. 



158 MAGNETISM CHAP, vin 

even if the fingers are shuffled on it (so sensitive some 
bars are), the whole is immediately reversed. Soft 
iron can, in fact, retain enormously more magnetism 
than steel can, but it retains it in a very feeble and 
loose manner. Its magnetism can only be styled 
sub-permanent. 

85. A short thick bar can retain much less magnet- 
ism than a long thin one ; in fact, if a stout bar is 
made of the softest iron, it can retain hardly any. 
A piece of iron shaped like Fig. 29 would have a 
much better chance of retaining its magnetism, and if 
the gap were closed by another piece of iron, called a 
" keeper," it would retain it very well ; while if the last 
trace of air-gap, the air-films between keeper and mag- 
net, be abolished by making the whole one welded 
ring, then its magnetism is retained almost perfectly. 
There is some demagnetizing force even in this 
case, for a fluid magnetized as a ring would not 
remain magnetized, and I find that tapping or 
beating such a ring does appreciably weaken its 
magnetization, but the demagnetizing force is very 
small compared with what it is when there is an 
air-gap. 

Hence we learn that the specially demagnetizing 
portion of a magnetic circuit is the fluid portion 
the air portion ; and the greater the proportion of the 
fluid portion to the whole, the more easily is de- 
magnetization accomplished fluids having no power 
of their own to retain magnetism. If lines of force 
are forcibly maintained in air or other fluid in the 
neighbourhood of a solid magnet, all the strain of 
upholding not only its own magnetism, but all the 



86 NATURE OF MAGNETISM 159 

rest of the magnetism in the field the strain of 
keeping the molecules faced round in opposition to 
their mutual restoring forces has to be thrown upon 
the rigidity and retentivity of the solid. 

86. All the known facts of magnetism have had new 
life and interest put into them by the researches 
of Ewing. The long-known fact that solid substances 
store up in their structure any previous arrangement 
of their molecules, so that traces of the effect are re- 
cognizable long after the cause of the effect has been 
withdrawn, is not indeed by any means confined to 
magnetism ; it is a general property of solids, and 
constitutes a considerable difficulty in dealing with 
them theoretically. The properties of all fluids, 
whether liquids or gases, depend upon their state at 
the moment, and upon nothing else ; not at all upon 
how they reached that state, or upon what has happened 
to them in past times. " Hydrogen at o C. and 76 
centimetres pressure " is a perfectly definite substance- 
"Water at 50 C. and one atmosphere 'pressure " is 
again a quite complete specification. And the same 
is very nearly true of some crystalline solids. Quartz 
or ice at given temperature and pressure is generally 
considered quite a definite statement, though perhaps 
it is not so exactly definite as we imagine. But glass, 
or steel, or copper, at a specified temperature, is by no 
means a definite substance. If it has been cooled 
down to that temperature it will not be the same as if 
it had been warmed up to it. We must be told whether 
it has been hardened, or tempered, or annealed, and 
so on. The properties of a solid body depend on 
its past history as well as on its present state. 



160 MAGNETISM CHAP, vin 

All this is pre-eminently true of magnetization. 
To understand completely the behaviour of a magnet 
we must not only know its present state, but we must 
know how it got to that state. A piece of steel once 
magnetized and then demagnetized is not in the same 
condition as if it had never been in a magnetic field : 
unless, indeed, it has been re-melted and made afresh. 

This much, however, must be granted, that if every- 
thing were known about the instantaneous state of a 
body there would be no need to go back upon its past 
history : it might even be possible to deduce some of 
its past history from its present state. But it is pre- 
cisely because a knowledge of the position and relation 
of every individual molecule is impossible, and because 
we have to put up with a few salient features of 
information, that an inquiry into past history is 
necessary. A few salient features are sufficient in the 
case of fluids: they are, in general, not sufficient in 
the case of solids. 

I have laid stress upon this matter because it is an 
important general distinction between states which 
are ' self-contained, 5 so to speak, and states which are 
' led up to.' The first condition represents what is 
expressed in mathematical language by saying that 
a function " is a complete differential," or " has a 
potential." 

87. A further detail of the distinction is that, in 
solids, a direct and a return series of changes are not 
usually the same ; a precisely inverse cause does not 
precisely invert the effect. Take a body from one 
self-contained state to another. Then, whenever you 
reverse the series of operations which brought it there, 



87 NATURE OF MAGNETISM 161 

it will return through the same series of states to its 
previous condition ; everything will be as it had been 
and no work need have been done on the whole. Not so 
with the led-up-to states. Magnetize a piece of steel by 
one series of operations, and then perform the same 
operations in reverse order : it will not return by the 
same series at all, nor will it return to its original con- 
dition. Continue the process of magnetization and 
reverse magnetization several times, and you may at 
length succeed in getting the body to go through 
a cycle of changes, at least approximately. But it 
will go by one * path/ so to speak, and return by 
another. 

Now when a body of any kind is taken from a state 
A to a state B by one path, and back from B to A by 
some other path, as steam is, for instance, between 
boiler and condenser in a steam-engine, the result 
is always that some work is either done by the 
substance, or has to be done upon it, in performing 
the cycle. The return by a different path is optional 
in the case of steam, and accordingly work may 
or may not be done on it ; but in the case of a 
magnetized solid it is not optional. The ascending 
curve of increasing magnetism, and the descending 
curve of decreasing magnetism, do not coincide, and 
cannot be made to coincide. Consequently, whenever 
a piece of iron is taken round a cycle of magnetic 
changes, some work is necessarily done. 

This work, in general, results in a production of 
heat, and accordingly a piece of iron magnetized and 
demagnetized several times, in rapid succession, gets 
slightly warm. This direct heating effect is, however, 

M 



1 62 MAGNETISM CHAP, vm 

very small, though it is becoming perceptible in the 
" transformers " of modern electrical engineering. 1 

All this behaviour of iron and other substances 
with regard to magnetism is called by Ewing 
hysteresis (from vcrrepeco, to lag behind). 

8/A. A remarkable discovery has been recently made 
with regard to the mechanism of the act of magnetiza- 
tion and the nature of the forces in a solid to which 
magnetic retentivity is due. It used to be thought 
that a kind of friction, or other mechanical constraint, 
was necessary to hold the molecules in place 
permitting their elastic yield up to a certain limit 
as magnetic forces rotated them, while beyond that 
limit a sort of viscosity set in and fixed them more or 
less permanently in their new positions. But Ewing 
has experimentally shown that all the fixity required 
in a solid composed of polarized molecules is the 
fixity of their pivots or axes of rotation. If the atoms 
are capable of locomotion, as in a fluid, then no doubt 
permanent magnetization will be impossible ; but if 
locomotion be prevented, so that the atoms can only 
rotate on fixed centres, then every detail of the 
magnetic behaviour of a solid can be explained by 
the magnetic forces among the polarized but other- 
wise perfectly free atoms. No mechanical constraint 
is necessary to explain magnetic retentivity and 
hysteresis. It all follows at once from the behaviour 
of a number of magnetic particles rotating within the 
range of each other's influence ; and it can be very 

1 The large indirect heating by induced currents the so-called 
Foucault current is too familiarly known to need any other statement 
than that it is quite distinct from what we are here discussing. 



88 NATURE OF MAGNETISM 163 

sufficiently imitated by a flock of pivoted compass- 
needles. This is a considerable step towards remov- 
ing complexity and indistinctness from the constitu- 
tion of solids. The properties of liquids and gases 
are known to merge into one another, and to be 
explicable by a number of similar particles approach- 
ing more or less within each other's range of force. 
The properties of solids will probably follow too, by 
adding to the above conception that of fixity of centre, 
or absence of locomotion. 



Electrical Momentum once more. 

88. There is just one point to which I must stop 
here to call attention. The theories of magnetism and 
diamagnetism, which I have given according to 
Ampere, Weber, and Maxwell, require as their found- 
ation that in a perfect conductor electricity shall obey 
the first law of motion shall continue to flow until 
stopped by force. But the property of matter which 
enables it to do this is called inertia ; the law is called 
the law of inertia ; and anything which behaves in 
this way must be granted to possess inertia. 

It would not do to deduce so important a fact from 
a yet unverified theory ; but at least one must notice 
that momentum is essentially involved in Ampere's 
theory of magnetism. It is the only theory of mag- 
netism yet formulated, and it breaks down unless 
electricity possesses inertia. 

Nevertheless it is a fact that an electro-magnet does 
not behave in the least like a fly-wheel or spinning- 

M 2 



1 64 MAGNETISM CHAP, vm 

top : there is no momentum mechanically discover- 
able ( 39). Supposing this should turn out to be 
strictly and finally true, we must admit that a molecular 
electric current consists of two equal opposite streams 
of the two kinds of electricity : one must begin to 
regard negative electricity not as merely the negation 
or defect of positive, but as a separate entity. Its 
relation to positive may turn out to be something 
more like that of sodium to chlorine than that of cold 
to heat. 

89. I said that no effect due to electric inertia was 
mechanically discoverable ; and on the hypothesis that 
an electric current consists of a pair of equal opposite 
currents of positive and negative electricity respec- 
tively this is very natural. Think of a couple of 
india-rubber pipes tied together so as to form a double 
tube, and through each propel a current of water, one 
in an opposite direction to the other. The double 
current has no gyrostatic properties, and the only way 
the water can exhibit momentum is by its resistance 
to change of velocity like the " extra-current " effects 
in electricity ( 38). 

So long as one considered the flow of electricity in 
ordinary conductors, we could partially avoid the 
question of inertia by considering it urged forward at 
every point with a force sufficient to overcome the 
resistance there and no more ; but though this ex- 
plained the shape of the stream-lines ( 49), yet it 
did not suffice to render clear the phenomena of self- 
induction the lag of the interior electricity in a wire 
behind the outside until definitely pushed ( 43-48) ; 



89 NATURE OF MAGNETISM 165 

still less does it explain its temporary persistence in 
motion after the pushing force has ceased. 

But, now that we are dealing with perfect conduc- 
tors, with no pushing force at all, the persistence of 
molecular currents without inertia, or an equivalent 
property so like it as to be rightly called by the same 
name at present, becomes inexplicable. Inertia must 
be recognized as an essential property of electricity. 



CHAPTER IX 

STRUCTURE OF A MAGNETIC FIELD 

Representation of a Magnetic Field. 

91. The disturbance called magnetism, which we 
have shown in Chap. VII. to be something of the 
nature of a spin a rotation about an axis is con- 
spicuously not limited to the steel or iron of the 
magnet : it spreads out through all adjacent space, 
and constitutes what is called the magnetic field. A 
map of the field is afforded by the use of iron filings, 
which cling end to end and point out the direction of 
the force at every point (Fig. 33). 

These lines of force so mapped are to be regarded 
as the axes of molecular whirls (Fig. 30). They are 
continuous with similar lines in the substance of the 
steel, and every line really forms a closed curve, of 
which a portion is in the steel and a portion in the air. 
In a wire helix, such as Figs. 16 or 29, the lines are 
wholly in the air, but in one part of their course they 
thread the helix, and in another part they spread out 
more or less between its faces. 

But according to Ampere's theory of molecular cur- 
rents there is no essential difference between such a 



91 STRUCTURE OF A MAGNETIC FIELD 167 

helix and a steel magnet ; directly the currents in the 
molecules of the magnet are considered, everything 
resolves itself into chains of molecular currents, thread- 
ing themselves along a common closed curve or axis. 
Each atom, whether in the steel or in the air, is the 
seat of a whirl of electricity, more or less faced round 
so as on the average to have its plane at right angles 





FIG. 30. A, an element of a magnetic line of force with the electric whirl round it; 
B, a bit of an electric circuit with one of its magnetic lines of force shown round 
it, and the electric whirl round this : each magnetic line of force round a current 
being an electric vortex ring. Compare Fig. 39. 

to the lines of force. The simplest plan of avoiding 
having to consider those only partially faced round, is 
to imagine the whole number divided into a set which 
face accurately in the right direction, and a set which 
look any way at perfect random ; and to neglect this 
latter set. 

92. Well, now try and picture a chain of whirls 
like beads spinning on a wire threading them all, and 



1 68 MAGNETISM CHAP, ix 

think of the effect of a material fluid thus rotating. 
Obviously it would tend to whirl itself fatter, and to 
shorten its length, as in Fig. 31 or 32. An assemblage 
of parallel straight whirls would thus squeeze each 
other laterally, or cause a lateral pressure, and would 
tend to drag their free ends together, causing a longi- 
tudinal tension. 

Such whirls cannot in truth have free ends except 
at the boundary of a medium as at the free surface of 
a liquid. Magnetic whirls are in reality all closed 
curves ; but inasmuch as part of them may be in a 
mobile fluid like air, and part of them in a solid like 
iron or steel, it is convenient to distingush between 
their two portions ; and one may think of the air whirls 
alone, as reaching from one piece of iron to another, 
and by their shortening tendency or centrifugal force 
pulling the two pieces together. 

The arrangement shown in Fig. 31 illustrates the 
kind of force exerted by a spinning elastic framework, 
along and perpendicular to its axis of rotation. 

One can easily see this effect of a whirl in a tea-cup 
or inverted bell-jar full of liquid. Stir it vigorously 
and leave it. It presses against the walls harder than 
before, so that if they were flexible they would bulge 
out with the lateral pressure, and it raises the level of 
the peripheral liquid ; moreover, it sucks down the 
top or free end of its axis of rotation, producing 
quite a depression or hollow against the force of 
gravity. 

A cyclone model can be made by a horizontal disk 
with vanes spinning fast on a vertical axis. Paper 
and other light objects can be sucked up by the axis 



92 STRUCTURE OF A MAGNETIC FIELD 169 

of the whirl of air below it. Or, as another illustration, 
make the apparatus sketched in Fig. 32. 

Two circular boards joined by a short wide elastic 
tube or drum : a weight hung to the lower board, the 
top board hung from a horizontal whirling table, the 




FIG. 31. A " shape of the earth " model which, when whirled, exerts a tension along 
its axis, pulling up the weight attached to it, and a pressure at right angles, by 
reason of its bulging out. Its semicircular springs hold up a ring sliding on a 
vertical guide-rod. 

drum filled with water, and the whole spun round. 
The weight is raised by the longitudinal tension ; the 
sides bulge out with the lateral pressure. 

There is no need for the whole vessel to rotate. 
If the liquid inside rotates, the same effect is 
produced. 



170 



MAGNETISM 



CHAP. IX 



93. Imagine now a medium composed of a mul- 
titude of such cells with rotating liquid inside ; let 
the cells be either very long, or else be joined end to 
end so as to make a chain a series of chains side by 
side and you have a picture of a magnetic medium 
traversed by a field of force. End-boundaries of the 
field will be dragged together, thus representing mag- 
netic attraction ; while sideways, the lines of force 




T 




FIG. 32. An elastic-walled cylindrical vessel full of liquid hanging from a whirling 
table, and, by reason of centrifugal force, raising a weight and bulging out 
literally when spun, thereby illustrating a tension along the axis of rotation and 
a pressure in every perpendicular direction. The vessel is sketched twice first 
stationary and then spinnnig. 

(axis of whirl) squeeze each other apart, thus illus- 
trating repulsion. This is Clerk-Maxwell's view of 
an electro-magnetic medium, and of the mode in 
which magnetic stress, and magnetic attractions and 
repulsions between bodies, arise. 

Wherever lines of force reach across from one body 
to another, those bodies are dragged together as if 
pulled by so many elastics (Fig. 33) ; but wherever 
lines of force from one body present their sides to 
lines of force proceeding from another body, then 
those bodies are driven apart. 



93 STRUCTURE OF A -MAGNETIC FIELD 171 




Attraction. 




pei 

the other pair apart, 



CHAPTER X 

MECHANICAL MODELS OF A MAGNETIC FIELD 

First Representation of the Field due to a Current. 

94. RETURN now to the consideration of a simple 
circuit, or, say, a linear conductor, and start a current 
through it ; how are we to picture the rise of the lines 
of force in the medium ? how shall we represent the 
spread of magnetic induction ? First think of the 
current as exciting the field (instead of the field as 
exciting the current, which may be the truer plan 
ultimately). 

If we can think of electricity in the several mole- 
cules of the insulating medium as something con- 
nected like so many cog-wheels gearing into one 
another, and also gearing into those of the metal 
conductor, it is easy to picture a sideways-spread of 
rotation brought about by the current, just as a 
moving rack will rotate a set of pinions gearing into 
it and into each other (Fig. 34). But then half the 
wheels will be rotating one way, and half the other 
way, which is not exactly right 



94 MODELS OF MAGNETIC FIELD 173 

How is it possible for a set of parallel whirls to be 
all rotating in the same direction, as in Fig. 35 ? 




FIG. 34. 



If there is any sort of connexion between them they 
will stop each other, because they are moving in 



ooo 
ooo 



FIG. 35. 

opposite directions at their nearest points : and yet, 
if there is no connexion, how can the whirl spread 
through the field ? 

Well, return to the old models by which we endea- 
voured to explain electrostatics, and think whether 
they will help us if we proceed to superpose upon 
them a magnetic whirl in addition to the properties 
they already possess. Looking at Figs. 5, 6, and /A, 
we remember we were led to picture atoms and elec- 
tricity like beads threaded on a cord. And these 
cords had to represent, alternately, positive and 
negative electricity, which always got displaced in 
different directions (see 90). 

We are forced to a similar sort of notion in respect 
of the wheels at present under discussion ; in order 



174 MAGNETISM CHAP, x 

that they may co-operate properly, they must repre- 
sent positive and negative electricity alternately. If 
they then rotate alternately in opposite directions, all 
is well, and the electrical circulation or rotation in the 
field is all in one direction. Each wheel gears into 
and turns the next, and so the spin gets propagated 
right away through the medium, at a speed depend- 
ing on the elasticity and density concerned in such 
disturbances. 

In Fig. 36 the dotted lines may be taken to represent 
cords on which the axles of the wheels are supported, 




FIG. 36. Rows of cells alternately positive and negative, geared together and free 
to turn about fixed axles. It is an extension of Fig. JA. 

with their axis of rotation at right angles to the 
cord ; not after the fashion of simple and ordinary 
beads. The figure may then be looked at sideways 
and compared with Fig. 7 A 18, and also with Figs. 
6, 8, and 13. 

Any longitudinal shift of the cords positive and 
negative always moving in opposite directions con- 
stitutes an electric shear of the medium and must be 
accompanied by some rotation of the wheels ; very 
small if the shift is small and elastically resisted, as in 



94 MODELS OF MAGNETIC FIELD 175 

a dieletric ; large up to any amount if the movement 
of the cords is continuous, constituting a steady cur- 
rent, as in a conductor. Thus a current of every 
kind, even a mere displacement, has some magnetic 
concomitant. 

The case of an electric charge moving through the 
medium may be typified by the rack of Fig. 39. An 
uncharged body would be like a perfectly smooth bar, 
without a rack. The only grip or attachment known 
between matter and ether is afforded by an electric 




FIG. 37. Section of a magnetic field perpendicular to the lines of force. Alternate 
cells rotating oppositely. Another mode of drawing Fig. 36. 

charge ; and even then its only result is to set up a 
magnetic rotation or an electric shear, not to induce 
locomotion after the fashion of a viscous drag. 
(Cf. Lecture V. below.) 

It is not convenient at the present stage to ask the 
question whether the wheels represent atoms of matter 
or merely electricity. It may be that each atom is 
electrostatically charged and itself rotates, in which 
case it \vould carry its charge round with it, and 
thereby constitute the desired molecular current. 



176 MAGNETISM CHAP, x 

The apparent inertia of electricity would thus be 
explained simply enough, as really the inertia of the 
spinning atoms themselves ; and the absence of any 
moment of momentum in an electro-magnet, as tested 
mechanically, would be equally explained by the 
simultaneous opposite rotation of adjacent atoms. 
A question may arise as to why the opposite 
molecules should have exactly equal opposite mo- 
menta ; as they have, else a fluid magnetized medium 
would bodily rotate ; and there may be other 
difficulties connected with a bodily rotation of 
electrostatically charged molecules ; but we need not 
lay any stress on the idea that the matter itself need 
move. For our present purpose a spin of the electri- 
city inside each atom, or even independently of any 
atoms, is quite sufficient. Besides, since magnetic 
induction can spread through a vacuum quite easily, 
the wheel-work has to be largely independent of 
material atoms. 

If any difficulty is felt concerning the void spaces 
in Fig. 36, it is only necessary to draw it like Fig. 37, 
which does every bit as well, and reduces the difficulty 
to any desired minimum. The original model of 
Clerk Maxwell, referred to in 155, is the basis of all 
these modes of representing the equations of an 
electro-magnetic field. 

Representation of an Electric Current. 

95. Now notice that in a medium so constituted 
and magnetized that is, with all the wheel-work 
stationary and revolving uniformly there need be 



95 MODELS OF MAGNETIC FIELD 177 

nothing of the nature of an electric current proceeding 
in any direction whatever. For, at every point of 
contact of two wheels, the positive and negative 
electricities are going at the same rate in the same 
direction ; and this is no current at all. Only when 
positive is going one way and negative going the 
opposite way, or standing still, or at least going at a 
different rate, can there be any advance of electricity, 
or anything of the nature of a current. 

A current is nevertheless able to be represented by 
mechanism such as that of Fig. 36 or 37 : if the 
wheels are permitted to gear imperfectly and to work 
with slip. At any such slipping-place the positive is 
going faster than the negative, or vice versd, and so 
there is a current there. This slip must be supposed 
passed on from one pair of wheels to the next in the 
direction of the slip, as oil or dust might be trans- 
ferred. A line of slip among the wheels corresponds 
to a linear current. 

Now a little attention here makes it quite plain 
that such a line of slip must always have a closed 
contour, in other words that electricity must flow in 
a closed circuit ( 4). For, if only one wheel slip, 
then the circuit is limited to its circumference ; if 
a row slip, then the direct and return circuit are on 
opposite sides of the row ; while if a large area of 
any shape, with no slip inside it, is enclosed by a line 
of slip, then this gives us a circuit of any shape, but 
always closed. Understand : one is not here thinking 
of a current as analogous to a locomotion of the wheels 
their axes may be quite stationary on this mode of 
thought the slip contemplated is that of one rim on 

N 



i;8 MAGNETISM CHAP, x 

another ; and any locomotion or progression there 
may be only relates to whatever it is which the rims 
push forward and hand on from one to another : actual 
electrons, as we are now beginning to think. This 
progress forward is typified by the rack of Fig. 39 ; it 
would be propelled by the spinning wheels. 

Imagine all the wheels inside the empty contour of 
Fig. 38 to be rotating, the positive clockwise, the 




FIG. 38. Diagram of a peripheral current partitioned off from surrounding medium 
by a perfect conductor, which transmits no motion, and therefore acts as a perfect 
magnetic screen. See also 101, and Fig. 41. Shaded wheels stand for positive. 
The current is flowing clockwise round the outside of the inner nine wheels. 



negative counter clockwise, and let all those outside 
the contour be either stationary or rotating at a 
different rate or in an opposite direction ; then the 
boundary of the inside region is a line of slip, along 
which the positive rims are all travelling clockwise 
with the negative rims going the other way, and hence 
it represents a clockwise positive current round the 
inside of the empty contour. 

But it may be said that the spin inside the contour, 
if maintained, must sooner or later rotate the wheels 



MODELS OF MAGNETIC FIELD 179 

outside as fast as themselves, and that then all slip 
will cease. Yes, that is so, unless there is a complete 
breach of connexion at the contour, as in Fig. 38 
there is. If the outer region has any sort of con- 
nexion with the inner one, the slip at its boundary 
can only be temporary lasting during the era of 
acceleration. 



Distinction between a Dielectric and a Metal^ as affected 
by a spreading Magnetic Field. 

96. In a dielectric the connexion between the 
atoms is definite and perfect. If one rotates, the next 
must rotate too ; there may be instantaneous delay due 
to elasticity, but there is no slip between the geared 
surfaces. A dielectric is a case of cogged wheels like 
Fig. 36, a conduction-current in it is impossible. 

But in a metallic conductor say a metallic mass 
immersed in a growing magnetic field the gearing is 
imperfect ; it is a case of friction-gearing, with more or 
less lubrication and slip, so that turning one wheel 
only starts the next gradually it may be very quickly 
but not instantaneously, and there is a motion of a 
positive rim incompletely compensated by an equal 
similar motion of a negative rim while getting up 
speed ; in other words, there is a momentary electric 
current, lasting till the wheels have fairly started. 

In a perfect conductor all gearing is absent ; the 
lubrication is so perfect that the atoms are quite 
free of one another, and accordingly a spin ceases to 
be transmitted into such a medium at all. The only 
possible current in a perfect conductor is a skin-deep 

N 2 



i8o MAGNETISM CHAP, x 

phenomenon. 1 This is the case depicted in Fig. 38. 
(See also Chap. V. and 104.) 

A magnetized medium, of whatever sort, is thus to 
be regarded as full of spinning wheels, the positive 
rotating one way and the negative the other way. 
If the medium is not magnetized, but only magnetic 
i.e. capable of being magnetized it may be thought 
of either as having its wheels stationary, or as having 
them facing all ways at random ; the latter being 
probably the truer the former the easier repre- 
sentation, at least to begin with. 

Whether the medium be conducting or insulating 
makes no difference to the general fact of spinning 
wheels inside it wherever lines of force penetrate it. 
But the wheels of a conductor are imperfectly cogged 
together ; and accordingly, in the variable stages of a 
magnetic field, while its spin is either increasing or 
decreasing, there is a very important distinction to 
be drawn between insulating and conducting matter. 
During the accelerating era conducting matter is full 
of slip, and a certain time elapses before a steady state 
is reached. A certain time may be necessary for the 
propagation of spin in a dielectric, but it is excessively 
short, and the process is unaccompanied by slip, only 
by slight distortion and recovery. (See 103 and 

1590 
97. As for strongly magnetic substances, like iron, 

1 It has been suggested to me that although no current can be excited 
inside a perfect conductor by the lateral action of the surrounding 
dielectric, yet that thermo-electric currents excited by contact force might 
be possible in its interior. I have ventured to surmise however ( 62) 
that no thermo-electric forces can exist in a perfect conductor, because 
it cannot get a grip of the electricity. 



98 MODELS OF MAGNETIC FIELD 181 

nickel, and cobalt, one must regard them as con- 
stituted in the same sort of way, but with wheels 
greatly more massive, or very much more numerous, 
or both. The quantity which we have called perme- 
ability in 82, 83, and denoted by the symbol ^, may 
now be thought of as physically equivalent to a 
density of the magnetic medium ; so that substances 
with a large /*, like iron, have their magnetic mechanism 
or wheel-work exceedingly massive, but not spinning 
at any different rate. 

Phenomena connected with a varying Current. Nature 
of Self-induction. 

98. Proceed now to think what happens in the 
region round a conductor in which a current is rising : 
as partially described in Chapter V. Without attempt- 
ting a complete and satisfactory representation of 
what is going on, we can think of mechanical 
arrangements which have some close analogy with 
electrical processes. 

Take first a system of wheel-work connected 
together and moved at some point by a rack (Fig. 39). 
Attend especially to alternate wheels, as representing 
positive electricity. The intermediate negative wheels 
are necessary for the transmission of the motion, and 
they also serve to neutralize all systematic advance of 
positive electricity in any one direction, except where 
slip occurs, but they need not otherwise be specially 
attended to. 

Remember that every wheel is endowed with inertia 
like a fly-wheel ( 88). 



1 82 MAGNETISM CHAP, x 

Directly the rack begins to move, the wheels begin 
to rotate, and in a short time they will all be going 
full speed. Until they are so moving, the motion of 
the rack is opposed, not by friction or ordinary 
resistance, but by the inertia of the wheel-work. 

This inertia represents what is called self-induction, 
and the result of it is what has been called the " extra- 
current at make," or, more satisfactorily, the opposing 
E.M.F. of electro-magnetic inertia or self-induction. 

Having once started the rack, so long as it moves 
steadily forward, the wheel-work has no further effect 
upon it ; but, directly we try to stop it, we find it 
cannot be stopped suddenly without great violence : 
its motion is prolonged for a short time by the inertia 
of the wheel-work, and we have what is known as the 
" extra-current at break." 

99. If the rack is for a moment taken to represent 
the advancing electricity in a copper wire, then the 
diagram may be regarded as a section of the complete 
field : the complete field being obtained from it by 
rotating the diagram (Fig. 39) round the axis of the 
rack. Imagining this done : we see that the axis of 
each wheel becomes prolonged into a circular core, 
and each wheel into a circular vortex ring surround- 
ing the rack, and rolling down it as it moves 
forward, as when a stick is pushed through a tight- 
fitting umbrella-ring held stationary (see Fig. 30, B, 
p. 167).- 

As one goes further and further from the rack the 
lengths of the vortex cores increase, but there is only 
a given amount of rotation to be shared among more 
and more stuff, hence it is not difficult to imagine the 



ioo MODELS OF MAGNETIC FIELD 183 

rate of spin diminishing as the distance increases, so 
that at a reasonable distance from the conductor the 
medium is scarcely disturbed. A slightly magnetized 
medium has to be considered on this mode of 
representation as if it consisted of a few spinning 
and many quiescent wheels : an unsatisfactory 
feature of the model. 

ioo. To perceive how much rotation of the medium 
is associated with a given circuit, one must consider 




FIG. 39. A provisonal representation of a current surrounded by dielectric medium, 
either propelling or being propelled. Section through the wire. Compare 
Fig. 30, B. 

the shape of its contour the position of the return 
current. 

Take first a long narrow loop and send a current 
up one side and down the other. The rotations 
belonging to each are superposed, and though they 
agree in direction for the space enclosed by the loop, 
they oppose each other outside, and so there is barely 
any disturbance of the medium outside such a looped 
conductor ; very little dielectric is disturbed at all, and 



184 MAGNETISM CHAP, x 

accordingly the inertia or self-induction is very small 
(see Fig. 40). 

If the loop opens out as as to enclose an area, as the 
centrifugal force of the wheels will tend to make it do, 
then there is a greater amount of rotation, a greater 
moment of momentum inside it, and accordingly its 
self-induction is increased. The axis of every wheel 
is, however, continuous, and must return outside the 




FIG. 40. Diagram of a direct and return current close together, showing distribution 
of rotation and of slip in the thickness of the conductor, and in the dielectric 
between. The dielectric outside is very little disturbed. The smooth wheels 
represent a metallic conductor, the cog wheels represent a thin layer of dielectric 
between the direct and return part of the metallic circuit. 
For two currents going in the same direction, see Fig. 44. 

loop : so the outside region is somewhat affected 
by rotation, but of a kind opposite to that inside. 

101. Figs. 38 and 41 show the state of things for a 
closed circuit conveying a current. The free space in 
Fig. 38 represented a perfect conductor, or perfect 
breach of connection. Along the inner boundary of 
this space positive electricity was seen streaming in 
the direction of the arrows, and it may be streaming 



ioi 



MODELS OF MAGNETIC FIELD 



185 



quite independently on the outer boundary also, but 
nothing happens in the interior of the perfect con- 
ductor which was therefore only represented by 
empty space. 



QQQQOOOQ 




oooooooo 



FIG. 41. Diagram of simple conducting circuit like a galvanometer ring, with the 
alternate connecting wheels omitted. The same number of dielectric wheels are 
drawn outside as inside, to indicate the fact that the total spin is equal inside 
and out, though the outside is so spread out as to be much less intense. The 
diagram shows a clockwise current flowing steadily round the ring ; with the 
accompanying distribution of magnetism. Section in the plane of the ring. 
The length of the arrows must be attended to as well as their direction. 



The corresponding portion in Fig. 41 is intended for 
an ordinary conductor, full of wheels capable of slip. 
And slip in this case is a continuous necessity, for the 
rotation on one side of the conductor is in an opposite 



1 86 MAGNETISM CHAP, x 

direction to that on the other side ; so the atoms of 
the conductor have to accommodate themselves as 
best they can to the conditions, some of them rotating 
one way, some the other, and some along a certain 
neutral line of the conductor being stationary. If we 
imagine a number of square-contoured, closed circuits 
of thread, of regularly increasing size, surrounding the 
groups of wheels in Fig. 41 symmetrically something 
like the lettered contours of Fig. 43 we shall see that 
the inner one, being totally surrounded by cogged 
wheels, will not move ; since it is urged equally in 
opposite directions by the wheels inside it and the 
wheels outside it. But the next larger contour will travel 
forward clockwise, since the inner wheels which drive 
it forward are rotating faster than the outer ones 
which push it back : and, as always, we assume the 
force of propulsion to be proportional to the speed, in 
accordance with our interpretation of Ohm's law. The 
still larger thread contour will likewise be pushed 
forward in the same direction, and so will the thread 
enclosing all the nearly smooth or conductor-repre- 
senting wheels. But outside, in the outer dielectric, 
there will be no progressive movement, any more 
than there was in the interior, and for the same 
reason. 

Specifying the rim speed inside all the metal con- 
tours as u Qy that outside all as 11 ' , and the rim speed of 
successive metal contours as u v & 2 , etc., we should have 
in general u u v slightly greater than u + u 2 ; this 
slightly greater than z/ 2 + # 3 , etc. ; and slightly the 
least current, namely, u' Q u n on the outside of the 
ring, where the journey is furthest. 



ioi MODELS OF MAGNETIC FIELD 187 

But if we consider the simple case of a thin wire 
contour with uniform distribution of current through- 
out, we shall have 

U Q - U = U^ + U 2 = U 2 + U% = . . . . = - U n 

wherefore all the conductor u's are then equal, being 
equal to J of U Q ; and the neutral line of no spin is in 
the middle of the wire. 

In any case the annular space, filled by what repre- 
sents the conductor in the diagram, is characterised by 
a progressive tendency round and round the ring, 
throughout its thickness ; and this is the easiest mode 
of thinking of the electric current, represented by the 
slipping of the wheels and corresponding to the given 
distribution of magnetism. 

If a conductor is straight and infinitely long, the 
neutral line of no rotation is in the middle. If it be a 
loop, the neutral line is nearer the outside than the 
inside, because the rotation of the medium inside is 
the strongest. If the loop be shut up to nothing, the 
neutral line is its outer boundary, or nearly so (Fig. 
40). If, again, the circuit is wound round and round 
the core of a ring, as string might be lapped upon a 
common curtain-ring so as to cover it, then the axes 
of whirl are wholly enclosed by the wire, and there is 
no rotation outside at all. (See Fig. 47.) 

Fig. 42 shows a section of this last-mentioned con- 
dition, and here the wheels of the dielectric outside 
are not rotating at all. The inside is revolving, it may 
be furiously, and so between the inner and outer layers 
of the conductor we have a great amount of slip and 
dissipation of energy in other words, a strong 



i88 



MAGNETISM 



CHAP. X 



current, but the outside is not revolving at all : the 
magnetism is wholly interior and annular. 

1 02. The process of slip which we have depicted 
goes on in all conductors conveying a current, whether 
steady or variable, and in fact is the current. The 
slip is necessarily accompanied by dissipation of 
energy and production of heat : only in a perfect con- 
ductor can it occur without friction. In a steady cur- 
rent the slip is uniformly distributed throughout the 




FIG. 42. Section of a closed magnetic circuit, or electric vortex-ring, or hollow bent 
solenoid like Fig. 29. enclosing an anchor-ring air space ; the axis of the ring being 
A B. the section of its core being E and F. The lengths of the arrows indicate 
the intensity of the spin, i.e. of the magnetic field, which is a maximum at the 
middle of each section and nothing at all outside. If the core contains iron 
instead of air, its wheels have to be from 100 to 10,000 times as massive : slipping 
wheels, if of solid iron ; cogged wheels, if a bundle of fine varnished iron wires. 
Cf. Fig. 47. 

section of the conductor ; in the variable stages it is 
unequally distributed, being then more concentrated 
near the periphery of the wire ; as the diagrams 
endeavour to show, and as has been described in 

43- 

When a current is started in a wire, the outer 
layers start first, and it gradually though very quickly 



102 MODELS OF MAGNETIC FIELD 189 

penetrates to the axis. Hence the lag is greater as 
the wire is thicker, and also as it is made of better 
conducting substance. If it is of iron, the massive- 
ness of the wheels is so great that the lag is much 
increased, and the spin momentum of its outer layers 
is great enough to produce the experimental effects 
discovered by Prof. Hughes 1 ; but we are not to sup- 
pose that the velocity is greater inside a magnetic 
material, otherwise the model would suggest current 
phenomena which have no existence. 

One must never confuse the slip with the spin. Slip 
is current, spin is magnetism. There is no spin at the 
axis of a straight infinite wire conveying a current, 
and the spin occurs in opposite directions as you recede 
from the axis either way ; arranging itself in circular 
vortex cores round the axis (Fig. 30, B). But the slip 
is uniformly distributed all through the wire, as soon 
as the current has reached a steady state. The slip is 
wholly in the direction of the wire, and is of a loco- 
motive character. The axes of spin are all at right 
angles to that direction, and are stationary. 

1 Address to Society of Telegraph Engineers, January 1886. 



CHAPTER XI 

MECHANICAL MODELS OF CURRENT INDUCTION 

Rise of Induced Current in a Secondary Circuit. 

103. To study the way in which a magnetic field, 
excited in any manner, spreads itself into and through 
a conducting medium, look at Fig. 43, and suppose the 
region inside the contour A B C D to be an ordinary 
conducting region that is, full of wheels imperfectly 
geared together, and capable of slip. 

Directly the rack begins to move, all the wheels 
outside A B C D begin to rotate, and quickly get up full 
speed. The outer layer of wheels inside the contour 
likewise begins to rotate, but not at once ; there is a 
slight delay in getting them into full motion. For the 
next inner layer the delay is rather greater, and so on. 
But ultimately the motion penetrates everywhere 
equally, and everything is in a steady state. 

But while the process of starting the wheels was 
going on, a slip took place round the contour A B C D, 
and round every concentric contour inside it ; the 
periphery of the positive wheels moving in a direction 
opposite to that of the wheels in contact with the 



io 3 MODELS OF CURkENT INDUCTION 191 

rack, and so suggesting the opposite induced current 
excited at " make " in the substance of a conductor 
near a growing current, or, generally, in an increasing 
magnetic field. 

The penetration of the motion deeper and deeper, 
and the gradual dying away of all slip, illustrate also 
the mode in which this induced current arises and 
gradually dies away ; becoming nil as soon as the 




H 11 



FIG. 43. Diagram illustrating the way in which an induced current arises in a mass 
of metal immersed in an increasing magnetic field : also how it decays. A B c D 
is the boundary of the mass of metal. The dotted lines ABCD, EFGH, IJKL 
are successive lines of slip. 

magnetic field (i.e. the rotation) has penetrated to the 
interior of all conductors and become permanently 
established there as elsewhere. 

Suppose the motion of the rack now stopped : all 
the cogged wheels stop too, though it may be with a 
jerk and some violence and oscillation due to their 
momentum ; but those inside the contour ABCD will 
continue moving for a little longer. The outside layer 
of this region will slip in such direction as to illustrate 



192 MAGNETISM CHAP, x 

the direct induced current at " break," and will begin 
to stop first ; the slip and the stop will then gradually 
penetrate inwards, just as happened during the inverse 
process, until all trace of rotation ceases. This inverse 
slipping process is the direct induced current at 
" break/' 

Let us imagine threads round the successive lettered 
contours of Fig. 43, rubbed in a definite manner by 
alternate wheels, all those of one sign, as we did in 
connexion with Fig. 41, and then suppose the rack of 
Fig. 43 to start moving in a downward slanting direc- 
tion ; the threads would all experience a clockwise 
impulse forward, the outer one A B C D first, then the 
next E F G H, and so on. But as soon as the interior 
is fully reached, all motion of the threads will cease, 
notwithstanding the continued motion of the rack, 
because now all the wheels are rotating at the same 
pace. 

This condition of electrically generated magnetism, 
with cessation of induced current at make, will 
last until the rack stops or slackens in speed. When- 
ever this happens, the threads will be moved counter- 
clockwise a certain amount ; until, in the simplest 
case, they regain their initial position soon after the 
rack has stopped. This simplest case obtains when 
the conductor contains nothing disobedient to Ohm's 
law nothing like an air-gap or other valve-like 
device. 

In that last more complex case, the ultimate dis- 
placement of each thread may depend on whether the 
force of propulsion is adequate to overcome equally- 
well both-ways whatever obstacle it encounters in its 



104 MODELS OF CURRENT INDUCTION 193 

course. If the force is not equally effective both ways 
there will be a permanent shift or displacement of 
electricity in the circuit, due to the valve-like action ; 
and a Leyden jar might thereby become charged. In 
what I have called a pertinacious current, this action is 
employed to maintain currents of immense electromo- 
tive force which can be applied to keep air electrified 
for agriculture, for deposition of fume, and for other 
practical purposes. (Proc. Roy. Inst., March 1905.) 

104. Through a perfect conductor the disturbance 
could never pass, for the slip of the dielectric wheels 
on its outer skin would be perfect, and would never 
penetrate any deeper. A superficial current lasting 
for ever, or rather as long as the magnetic field (the 
rotation of the dielectric wheels) lasts, is all that would 
be excited, and it would be a perfect magnetic screen 
to any dielectric beyond and enclosed by it. Such a 
perfect conductor is represented by the empty space 
in Fig. 38. A magnetic field or spin excited outside 
that space would never reach the set of wheels 
enclosed and protected by it ; and vice versa ( 153). 

105. It will now be perceived that a fly-wheel in 
rotation is the mechanical analogue of magnetism, 
or more definitely of a section of a line (or tube) of 
magnetic force ; and that a brake, either an accelerat- 
ing or retarding brake, applied to such a fly-wheel, 
with consequent slip, dissipation of energy, and pro- 
duction of heat, is a mechanical analogue of an 
electric current. 

A magnetic field is to be regarded as full of geared 
elastic vortices or whirls, some of which are cogged 
together, so to speak, while others are merely pressed 

O 



194 MAGNETISM CHAP, xi 

together by smooth rims. It is among these latter 
that slip is possible, and it is in the regions occupied by 
these that currents exist ; the energy dissipated here 
being transmitted through the non-slippery or di- 
electric regions from the source of power, just as energy 
is transmitted from a steam-engine through mill-work 
or shafting to the various places where it is dissipated 
by friction. (Refer back to 42 44.) 



Transfer of Energy to a Distance. 

1 06. Let us now try to understand the use of a 
telegraph wire from this point of view. Given the 
means of exciting a magnetic field at one place, how 
can we transmit it to another place so as to move 
magnetic needles or make other signals there ? The 
first idea of a method might be that inasmuch as no 
perfect conductor, or absolute magnetic screen, in- 
tervenes, the field of any magnet is infinite in extent, 
and consequently already reaches the distant place. 
Have here a long iron bar capable of being magnetized 
and demagnetized at pleasure, and have there a very 
sensitive magnetometer, and the thing is done. I 
see no reason why, under certain circumstances, this 
mode of signalling without wire over short distances 
should not be attempted. But an obvious objection 
to it is that the effect produced by a given magnet 
varies inversely with the cube of its distance ; so that, 
a few miles away, the force of a magnet, even several 
yards long, is terribly weak. And even a coil of wire 
several acres in area has not very much distant power : 



io6 MODELS OF CURRENT INDUCTION 195 

though something can be done by this means, when 
the effect is collected by another similar coil. 1 

The next idea might be to carry some of the 
magnetic lines of force to the distant place by means 
of an iron rod or wire. A soft iron wire transmits 
them so much better than air, that an arrangement 
of a very elongated loop of iron, with a magnetizing 
coil upon it at one end, and a receiving coil upon 
it at the other, might serve to establish some con- 
nexion between the two places, and enable primary 
currents at this end to produce secondary induced 
currents at that. This would be a magnetic telegraph 
in which a magnetic whirl only is propagated along 
the wire, and a current is excited by it at the distant 
place. 

The current loop and the magnetic loop are, how- 
ever, reciprocal ; and the next idea might be that 
instead of a long magnetic loop with little current 
loops threading it at either end, it might be better to 
have a long current loop with little magnetic loops 
threading it at either end ; and that is exactly what 
we have in the electric telegraph. 

It is a better plan for this reason. Iron conducts 
magnetism, say, a thousand times better than air, but 
by no means infinitely better ; hence from a long loop 
of iron a great many lines of force would leak, and 
would take a shorter circuit back through air. But 
copper wire conducts electricity almost infinitely 
better than gutta-percha or porcelain. That is why 
an electric telegraph is better than a magnetic. Lead 

1 Lodge, on " Magnetic Space-Telegraphy," Proc. Inst. E. E. 
Dec. 1898. 

O 2 



196 MAGNETISM CHAP, xi 

or German silver conducts a million times better than 
dilute sulphuric acid, and yet it would be very un- 
satisfactory to have to signal through an ocean of 
dilute acid with an uncovered lead or German silver 
wire immersed in it. The percentage of loss in the 
case of a corresponding magnetic circuit of iron would 
be far greater still. 

107. But now what is it precisely that the wire of 
an electric telegraph does ? By its means a magnetic 
field at this end is made to excite a magnetic field at 
the other end, with very little loss ; it is nearly all 
concentrated upon the other end by means of the 
wire. Somehow the wire enables us to transmit the 
magnetic effect exactly in the direction we wish, and 
to reproduce it where we please. 

It is not very easy to draw a diagram of the 
arrangement, because it entails such a number of 
wheels, and because the diminution of spin with 
distance is not well represented by them. But the 
diagram may be imagined thus : 

Let the rack in Fig. 39 be regarded as an infini- 
tesimal portion of a long circuit extending to New 
York and back. At some distant point, by means of 
a battery, or a dynamo, or any other electromotive 
arrangement, excite in a few of the wheels the motion 
which in Fig. 39 would occur if the rack were pushed. 
Out and away through the dielectric the motion trans- 
mitted by the cog-wheels spreads, at a pace which for 
the present we may consider infinite, but which we 
shall learn later is the velocity of light. At length it 
reaches the rack, which is immediately propelled 
forward. In the wire, and in the wire only, exists the 



io8 MODELS OF CURRENT INDUCTION 197 

current, surrounded by a strong magnetic field. Right 
along the wire flashes the vortex-like motion ; all the 
way along it is surrounded with rings of whirl, as in 
Fig. 30, B ; and by concentrating some quantity of this 
whirl into small compass at the distant station, a visible 
motion or signal is produced. 

That is the function of the wire, it guides the effect 
transmitted through the dielectric. The wire transmits 
the electricity only, not any energy the insulating 
sheath it is which transmits all the energy : the wire 
directs it on its way, by holding asunder the mutually 
opposing gearing of the dielectric through which the 
disturbance arrives. 

Later on we shall learn that the gearing in a 
dielectric is not rigid, but elastic, and that this is 
why a certain time is required for transmission why 
a definite velocity of transmission exists. We may 
be said to have learnt it already, indeed ; and we have 
also learned that some dielectrics are less rigid than 
others gutta-percha less than air, for instance (see 
1 6 and 23), and accordingly transmit the disturb- 
ances more slowly ; but always, as we shall find, 
with the velocity of light appropriate to that parti- 
cular medium, so far as the dielectric alone is concerned 
( J 33 



Mechanical Force acting on a Conductor conveying a 
Current. 

108. In Fig. 41 the conducting portion of a circuit 
is shown with its appropriate opposite rotations on 
either side of it. Now superpose a uniform rotation 



198 MAGNETISM CHAP, xi 

all in one direction upon this, so as to increase the 
spin on one side of the conductor and diminish it on 
the other ; in other words, immerse the circuit in a 
magnetic field. Immediately the extra centrifugal 
force on one side will urge any movable part of the 
conductor from the stronger to the weaker portion of 
the field. And whether there be any movable portion 
or not, the whole circuit will tend to expand if the 
superposed magnetic whirl agree in direction with the 
whirl already inside ; while it will tend to contract if 
the superposed whirl agree with that outside. 

The field for a direct and return circuit may be 
similarly drawn by superposition of their separate 
whirls (see Fig. 40). In this case there is strong 
centrifugal force of the whirl between the wires, while 
outside there is next to no whirl at all. Hence the 
wires tend to get driven apart ; and so it becomes 
evident why a circuit tends to expand so as to enclose 
the largest possible area, even if no other magnetic 
field than its own be acting on it. The circuit shown 
in Fig. 41, for instance, tends to expand even without 
any superposed magnetic field, simply because the 
whirl inside is more concentrated, and therefore more 
intense, than the whirl outside. 

In Fig. 44, on the other hand, the conductors are 
carrying currents in the same direction, and the 
whirl is greater in the region outside them (above and 
below in the diagram) than it is between them ; 
hence they tend to be pressed together, a fact which 
is expressed briefly by saying that ' similar currents 
attract/ 

109. As for the effect of iron introduced into a 



109 MODELS OF CURRENT INDUCTION 199 

circuit, it brings into the region of space it occupies 
some hundred or thousand times as many lines of 
whirl as were there before, and these naturally con- 
tribute mightily to the effects, both those exhibiting 
mechanical force and those exhibiting inertia. 

When one says, as roughly one may do, that iron 
brings 1000 fresh lines into the field, one means that, 
for every whirl otherwise excited, 1000 more are faced 




FiG. 44. Two parallel conductors conveying equal currents in one direction and 
getting pushed together by the centrifugal force of the outside whirls, no whirl 
existing between them. The length of the arrows again suggests the distribution 
of magnetism in the conductors. Fig. 40 showed the correlative repulsion of 
opposite currents. 

round in the iron. And this process goes on while 
the field is increasing in strength until the total 
number of whirls in the iron begins to be called 
upon ; when this point is reached the rate of addition 
is not maintained, and the iron is said to show signs 
of saturation. Ultimately, if ever all its whirls were 
faced round, the iron would be quite saturated ; but 
long before this point is reached another cause is 
likely to make itself felt, viz. the falling off in the 
strength of the whirls already faced round, by the 



200 MAGNETISM CHAP, xi 

action of the strong magnetic induction, which is all 
the time acting so as to weaken the iron currents 
so far as it is able. And thus at a certain point, 
hitherto unreached by experiment, the iron may not 
only fail to increase the strength of the field any 
more, but may actually begin to diminish it. That 
is to say, its permeability may conceivably and excep- 
tionally become less than I, as if it were a diamagnetic 
substance (cf. 81). 

The easiest way to picture the effect of iron is to 
think of its wheels as some hundred or thousand 
times as massive as those of air, so that their energy 
and momentum are very great although their speed is 
the same (cf. 97). 

That which is commonly called magnetic permea- 
bility, and denoted by //, ( 82), may in fact be thought 
of as a kind of inertia, an inertia per unit volume ; 
in other words, a density an ethereal density ; though 
how it comes to pass that the ether inside iron is 
endowed with so great inertia one cannot say. 
Perhaps it is that the iron atoms themselves revolve 
with the electricity ( 94), perhaps it is something 
quite different. Whatever the peculiar behaviour of 
iron, nickel, &c., be due to, it must be something pro- 
foundly interesting and important as soon as our 
knowledge of their molecular structure enables us to 
perceive its nature. 

Representation of an Electrostatic Field again , and 
superposition of it on a perpendicular Magnetic Field 

in. An electrostatic strain is, we know, caused by 
a displacement of positive electricity one way along 



in MODELS OF CURRENT INDUCTION 201 

the lines of force, and by an equal displacement of 
negative the other way. The process was indicated 
crudely in Fig. /A ; we may now represent it rather 
more fully with the help of our elastic cells by 
Fig. 46. 

Here the positive cells have been pulled one way, 
the negative the other way ; and when the distorting 
force is removed, the medium tends to spring back to 
its normal condition, exerting an obvious tension on 




FIG. 46. A portion of an electrostatic field between two oppositely charged bodies 
with its lines of force going from right to left, and showing a tension along and a 



field was being excited and propagated from below upwards, a slight strain would 
be produced in the elastic cells, like but immensely less than that shown ; as con- 
trasted with its normal condition (Fig. 37). Conversely, while this electrostatic 
strain was being produced, the postitive whirls would be infinitesimally quickened 
and the negative ones retarded during the displacement, thus producing a minute 
magnetic effect. If the medium is not magnetized, the whirls are not necessarily 
absent, only faced all ways. 

bodies attached to it in the direction of its lines of 
force, its elongated direction, and an obvious pressure 
in all perpendicular directions, its compressed direc- 
tions. 

Now, if all the cells are full of parallel whirls, as in 
the preceding magnetic diagrams, it is not improbable 
that this electrostatic distortion or " shear " of the 
medium may affect its magnetic properties slightly, 
and that, if the direction of electrostatic strain were 
rapidly reversed, a small magnetic oscillation would 



202 MAGNETISM CHAP, xi 

also ensue ; but the exact details of these mutual 
actions are difficult to specify at present. 

There is a point here that deserves attention. If 
the cells in Fig. 46 represented a set of stretched 
elastic threads pulling the opposite boundaries (which 
may be thought of as metal plates) together, motion of 
the plates to or fro would either relax or increase the 
tension in the threads ; but in the case really 
intended that of the strained electric medium it is 
not to be supposed that motion of the plates would 
either relax or intensify the strain in each cell. 

If the plates move farther apart, more cells are 
thrown into the field ; but so long as the field ^intensity 
is uniform which will be true except near the edges 
of the plates when they are big compared with their 
distance apart the strain of each cell is unaltered. So 
also when the plates move nearer together, they wipe 
out the strain from one cell after another, but leave 
the rest unchanged. Whenever lines of force are 
spoken of as elastic threads, this peculiarity, in which 
the analogy breaks down, must be remembered ; other- 
wise the energy under different conditions will be 
wrongly estimated. 

Disruptive Discharge. 

112. Disruptive discharge may be thought of as a 
pulling of the shaded cells violently along past the 
others ; the process being accompanied by a true 
disruption a sort of electrolysis of the medium, 
and a passage of the two electricities in opposite 
directions along the line of discharge. 



ii3 MODELS OF CURRENT INDUCTION 203 

Consider the locomotion of any one horizontal row 
of shaded cells in Fig. 46 during the occurrence of such 
a disruption of the medium. The row .of cells slides 
towards the right, and, as it slides, the spin of the 
negative cells above it is retarded while that of those 
below it is accelerated ; consequently a true magnetic 
effect is produced, just like that accompanying a 
current, in rings round the line of locomotion ; and so 
a disruptive discharge has all the magnetic properties 
of a current. 



Effects of a Moving Charge. 

113. This locomotion of a set of positive cells, or 
of negative cells the other way, as just considered, is 
very near akin to the motion of a charge through a 
dielectric medium. 

When a charged body moves along with extreme 
rapidity, it can be thought of as exciting a rotation in 
the cells most closely in contact with it, greater than 
that which it excites in the opposite kind of cells ; and 
it thus produces the whirl proper to a magnetic field. 
It behaves like the rack in Fig. 39. Thus does a 
moving charge behave just like a current of a certain 
strength. 

It may be, indeed, that this is the customary way of 
exciting a voltaic current ; for the chemical forces in 
a cell cause a locomotion of charged atoms, and thus 
set up a field, which, spreading out in the way Prof. 
Poynting has sketched ( 42), reaches every part of the 
metallic circuit and excites the current there. 



204 MAGNETISM CHAP, xi 

Electrostatic Effects of a Moving or Varying Magnetic 

Field. 

114. Just as we have seen that a moving or varying 
electrostatic field may produce slight magnetic effects, 
it can be perceived that a moving or varying magnetic 
field brings about something of the nature of an 
electrostatic strain. 

For a spreading out field is continually propagating 
the rotation on from one layer of wheels to the next. If 
there is any slip, we thus get induced currents (Fig, 43), 
and the rate of propagation is comparatively slow, being 
a kind of diffusion ; but even if there is not any slip, 
yet unless the wheel-work is absolutely rigid, the rate 
of propagation will not be infinite. The actual rate 
of propagation is very great, which shows that the 
rigidity or elasticity of the wheels is very high in 
proportion to their inertia ; but it is not infinite, and 
accordingly the propagation of rotation is accompanied 
by a temporary strain. One part of the field is in full 
spin, another more distant part is as yet unreached by 
the spin ; between the two we have the region of strain, 
the wheel-work being distorted a little while taking up 
the motion. Thus does a spreading out magnetic 
field cause a slight and temporary electrostatic strain, 
at right angles both to the direction of the lines offeree 
and to the direction of their advance. 

Generation of a Magnetic Field. Induction in Closed 
Circuits. 

115. Picture to yourself an unmagnetized piece of 
iron : its whirls are all existent, but they are shut up 






us MODELS OF CURRENT INDUCTION 205 

into little closed circuits, and so produce no external 
effect ; magnetize it slightly, and some of the closed 
circuits open out and expand, with one portion of 
them in the air. Magnetize it strongly, and we have 
a whole set of them opened out into vortex cores, 
still with the whirl round them, constituting the 
common magnetic lines of force. There is no need to 
think of iron and steel in this connexion. In air or 
any substance the whirls are still present, though much 
fewer or feebler, and their axes ordinarily form little 
closed circuits it may be inside the atoms themselves. 




FIG. 47. Closed magnetic circuit like Fig. 42, with a single-ring secondary circuit, 
and another open secondary loop ; also with a short conducting-rod standing up 
in it. 

But wrap a current-conveying wire round them, and 
at once they open out into the lines of force proper to 
a circular current. 

Again, think of an iron ring, or a hank of iron wire 
as bought at an ironmonger's : wrap a copper wire 
several times round it, as each segment of a Gramme 
ring is wound (Fig. 47), and pass a current. The closed 
vortices in the iron at once expand : a portion of each 
flashes out and across the air-space inclosed by the 
ring (not by any means confining itself to a plane, of 
course), and enters the ring on the opposite side, 
having expanded and travelled laterally across the 
intervening space ; so that, directly the current is 



206 MAGNETISM CHAP, xi 

steady, the lines all lie inside the iron again, but 
now inclosing an area the area of the ring 
instead of being shut up into infinitesimal links. 
It may be said that the iron is still apparently 
unmagnetized ; for its lines of force still form closed 
contours within it, and none protrude any part of 
themselves into the air, except for irregularities. But 
in reality it is highly and permanently magnetized, 
round and round in itself: the magnetism being not 
easy to get out of it again, except by judiciously 
graduated alternating currents. 

It is now like one great electric vortex ring, instead 
of like a confused jumble of microscopic ones. Its 
section was shown in Fig. 42. (See also Appendix 
(d) and ().) 

During the variable period, while the current is in- 
creasing in strength, or while it is being reversed, the 
whole region near the ring is full of myriads of ex- 
panding lines of force, flashing across, broadside on, 
from one side of the iron to the other, and there 
stopping. It is the presence of these moving lines, 
changing rapidly from a u simply-connected " into a 
" multiply-connected " space, or vice versa, which 
causes the powerful induced currents of " secondary 
generators." 

In every case of varying magnetic field, in fact, we 
have lines moving broadside on, propagating their 
whirl, and more or less disturbing the medium through 
which they move. They never can move end-on nor 
thread a needle like a thread, for they never have free 
ends : they are always closed curves ; they can only 
move laterally. 



us MODELS OF CURRENT INDUCTION 207 

Next consider a moving or spinning magnet. Its 
lines travel with it, and, being closed curves, they also 
must move broadside through the field ; so that in this 
case we may expect just the same effect as can be 
obtained from a varying magnetic field. 

If a broadside-moving line of force cut across a 
conductor, its motion is delayed, for its wheels slip 
and only gradually get up a whirl inside the ill-geared 
substance ; thus, as we know, causing an induced 
current (see 103). 

If a conducting circuit is looped with the iron ring 
previously mentioned, as a snap hook is looped with 
an eye, so as to make what may be called a secondary, 
like the circle in Fig. 47, then every expanding vortex, 
while the ring is being magnetized, has necessarily to 
cut through the conductor or secondary circuit once 
and no more, no matter what its shape or size. 
The total or integrated E.M.F. induced in the 
secondary is therefore perfectly definite in this case, 
being proportional to the total number of magnetic 
lines transferred across the open space near the iron 
ring. At every exact reversal this would be just twice 
the number of lines of magnetic induction circulating 
inside the core of the ring. If the secondary consists 
of a wire of several turns, with joined ends, then this 
same E.M.F. is induced in each turn of wire ; so the 
previous result has now to be multiplied by the 
number of turns made by the secondary looped round 
the core of the ring. 

Instead of supposing a closed conducting secondary 
circuit, imagine an open one, likewise shown in 
Fig. 47 : there is still an E.M.F. in it, though it is 



2o8 MAGNETISM CHAP, xi 

slightly less than before, because a few of the expanding 
lines flash through the gap and produce no effect ; but 
there can be now no continuous current, only an 
accumulation of opposite charges at the gap terminals 
during the impulse ; so while the ring is being 
magnetized and demagnetized, electricity must surge 
to and fro in the conductor as water surges to and fro 
in a suddenly tilted trough, and a small condenser 
attached to the free ends will be alternately charged 
and discharged. 

In another case, the gap might become so large that 
nothing is left but a short rod (Fig. 47) : in this also 
similar electric oscillations would occur. 

But now suppose no secondary conductor at all ; 
nothing but dielectric inclosed by the ring. In the 
dielectric itself an electric displacement must be 
excited every time the magnetism of the ring is 
reversed. The displacement will be in one direction 
during rise of magnetism, and in an opposite direction 
during reversal of magnetism. A charged body 
delicately suspended within the ring may feel the 
effect of the minute electrostatic strain so magnetically 
produced, 1 and a needle with oppositely charged ends, 
if light enough, must be infinitesimally deflected. 

1 1 6. To see the mode in which an electrostatic 
displacement arises in the space embraced by the ring 
we have only to turn to Fig. 42, and look at the set 
of wheels along the line A B separating one half the 
section from the other. They cannot steadily rotate 
either way, for they are urged in opposite directions by 

1 For the experimental detection of these effects, see Lodge, Phil. Mag. 
June 1889. 



ii6 MODELS OF CURRENT INDUCTION 209 

the two halves ; in other words, there is no magnetic 
field anywhere outside such a ring, as is well known ; 
but the wheels of one sign are being urged in one 
direction by the spin inside the ring ; consequently 
there is a slight electric displacement along the axis 
of the ring, which lasts as long as the ring is 
magnetized, and recovers or reverses directly the 
magnetism ceases or is reversed. This is the only 
effect which changes in the ring's magnetism produce 
in the surrounding dielectric a slight and almost 
imperceptible electric field. 



PART IV 

RADIA TION 



P 2 



CHAPTER XII 

RELATION OF ETHER TO ELECTRICITY 

117. So far as we have been able to understand 
and explain electrical phenomena, it has been by 
assuming the existence of a medium endowed with 
certain mechanical or quasi-mechanical properties^ 
such as mobility ( 12), incompressibility or infinite 
elasticity of volume ( 5), combined with a certain 
amount either of plasticity or of finite elasticity of 
shape ( 9). We also imagined the medium to be com- 
posed of two opposite constituents, which we called 
positive and negative electricity respectively ( 90), 
and which were connected in such a way that what- 
ever one did the other tended to do the precise 
opposite. Further, we were led to endow each of 
these constituents with a certain amount of inertia 
( 38 and 88), and we recognized something of the 
nature of friction between each constituent and 
ordinary matter ( 28 and 63). 

Broadly speaking, we may say 

(i) That friction makes itself conspicuous in the 
discussion of current-electricity or the properties of 
conductors ; and that the laws of it are summarized in 



214 RADIATION CHAP, xn 

the statement known by the name of Ohm, viz. that 
the current through a given conductor is proportional 
to the force that drives it, or that the opposition force 
exerted by a conductor upon a current is simply 
proportional to the strength of that current. 

(2) That elasticity is recognized as necessary when 
studying the facts of electrostatics or the properties of 
insulators electric displacement and recoil, or charge 
and discharge : the laws having been studied by 
Faraday, and the relative pliability (or shearability, if 
there were such a word for the reciprocal of rigidity 
or elasticity of figure) of the medium in different 
substances being measured and stated in terms of 
that of empty space as their specific inductive capacity, 
K ; or, rather, for numerical specification, as the ratio 

K/K Q . 

(3) That inertia is brought into, prominence by the 
facts of magnetism, studied chiefly perhaps by 
Lord Kelvin ; who has called the relative density of the 
medium in different substances their magnetic ' per- 
meability ' or magnetic inductive capacity, /u ; the 
ratio of its value for any substance to its value for 
space being strictly denoted by JU,/JU, Q . 

(4) That the doubleness of constitution of the 
medium its being composed of two precisely opposite 
entities is suggested by the facts of electrolysis, by 
the absence of mechanical momentum in currents and 
magnets, and by the difficulty of otherwise conceiving 
a medium endowed with rigidity which yet is perfectly 
fluid to masses of matter moving through it. 

1 1 8. With the hypothesis of doubleness of con- 
stitution this last-mentioned difficulty disappears. 



ii8 ETHER AND ELECTRICITY 215 

The ether as a whole may be perfectly fluid, and allow 
bodies to pass through it without resistance, while its 
two components may be elastically attached together 
and may resist any forces tending to separate them, as 
if it possessed the characteristic rigidity of a solid. 
It is like the difference between passing one's hand 
through water, and chemically decomposing it ; it is 
like the difference between waving a piece of canvas 
about, and tearing it into its constituent threads. 

To put the matter boldly and baldly : we are 
familiar with the conceptions of matter and of ether, 
and it is known that the two things react on each 
other in some way, so that although a free portion of 
the ether appears to move freely through matter, yet 
another portion appears to move with matter as if 
bound to it. This mode of regarding the facts is as 
old as Fresnel ( 184). We now proceed a step farther, 
and analyze the ether into two constituents two equal 
opposite constituents each endowed with inertia, and 
each connected to the other by elastic ties : ties 
which the presence of gross matter in general weakens 
and in some cases dissolves. The two constituents 
are called positive and negative electricity respectively ; 
and of these two electricities we imagine the ether to be 
composed. The tie between them is dissolved in 
metals : it is relaxed or made less rigid in ordinary 
insulators. The specific inductive capacity of a 
substance means the reciprocal of the rigidity of 
its doubly constituted ether. The measure of this 
rigidity is probably ^TT/K. 

The neighbourhood of gross matter seems also to 
renders ether more dense. It is difficult to suppose 



216 RADIATION CHAP, xn 

that it can really condense an incompressible fluid ; 
but it may load it, or otherwise modify it, so as to 
produce the effect of increased density. In iron this 
internal ethereous density reaches its highest known 
value, and in all substances the density or inertia per 
unit volume of their ether may be denoted by 4 TT p 
and is proportional to their magnetic permeability. 
The hypothetical ethereous density here spoken of 
is not to be confused with the familiar density 
of ordinary material. It cannot be estimated by 
weighing. 

119. Let it be understood what we are doing. 
In Part I. we discussed effects very analogous to 
those which would be produced by an elastic incom- 
pressible medium (roughly like india-rubber or jelly) ; 
that is, we were led to postulate a medium possessing 
elasticity, or something very analogous to elasticity. 
In Parts II. and III. we discussed effects suggesting, 
and more or less necessitating, the idea of a property 
of the medium very 'analogous to inertia ; and we 
were also led to postulate a doubleness of constitution 
for the medium, so that shearing stresses may occur in 
it and yet it be perfectly fluid as a whole. We are 
now pushing these analogies and ideas into greater 
definiteness and baldness of statement. We already 
know of a continuous, incompressible fluid filling all 
space, and we call it the ether. Let us suppose that 
it is composed of, and by electromotive force ana- 
lyzable into, two constituents ; let these constituents 
cling together with a certain tenacity, so that the me- 
dium shall have an electromotive elasticity, though it is 
mechanically quite fluid ; and let each constituent pos- 



120 ETHER AND ELECTRICITY 217 

sess inertia, or something so like inertia as to produce 
similar effects. Making these hypotheses, electrical 
effects are to a certain extent explained. Not 
ultimately, indeed few things can be explained 
ultimately : not even as ultimately as could be 
wished ; for the nature of the connection between the 
two constituents of the ether, and between the ether 
and gross matter the nature of the force, that is, and 
the nature of the inertia remains untouched. This is 
a limitation to be clearly admitted ; but if that were 
the only one if all else in the hypothesis were true 
we should do well, and a distinct step would have been 
gained. It is hardly to be hoped that this is so 
hardly to be expected that the bald statement above 
is more than a kind of parody of the truth ; neverthe- 
less, supposing it only a parody, supposing what we 
call electromotive elasticity and inertia are things 
capable of clearer conception and more adequate state- 
ment ( 156), yet, inasmuch as they correspond to and 
represent a real analogy, and inasmuch as we find that 
a medium so constructed would behave in a very elec- 
trical manner, and might in conjunction with matter 
be capable of giving rise to all known electrical 
phenomena, we are bound to follow out the conception 
into other regions, and see whether any other abstruse 
phenomena, not commonly recognized as electrical, will 
not also fall into the domain of this hypothetical 
substance and be equally explained by it ( 8). This 
is what we shall now proceed to do. 

1 20. Before beginning, however, let me just say 
what I mean by " electromotive elasticity." It might 
be called chemical elasticity, or molecular elasticity. 



218 RADIATION CHAP.XII 

There is a well-known distinction between electro- 
motive force and ordinary matter-moving force. The 
one acts upon electricity, straining or moving or, in 
general, " displacing" it ; the other acts upon matter, 
displacing it. The nature of neither force can be 
considered known, but crudely we may say that as 
electricity is to matter, so is electric force to common 
mechanical force ; so also is electric elasticity to 
the common shape-elasticity or rigidity of ordinary 
matter; so perhaps, once more, may electrical inertia 
be to ordinary inertia. 

Inertia is defined as the ratio of force to ac- 
celeration ; similarly electric inertia can be defined as 
the ratio of electromotive force to the acceleration of 
electric displacement. It is quite possible that 
electric inertia and ordinary inertia are the same 
thing, just as electric energy is the same with mechan- 
ical energy. If this were known to be so, it would 
be a step upward towards a mechanical explanation ; 
but it is by no means necessarily or certainly so ; and 
whether it be so or not, the analogy undoubtedly 
holds, and may be fruitfully pursued. 

And as to " electromotive elasticity," one may say 
that pure water or gas is electromotively elastic, though 
mechanically limpid ; each resists electric forces, up to 
a certain limit of tenacity, beyond which it is broken ; 
and it recoils when they are withdrawn. Glass acts 
in the same way, but that happens to be mechanically 
elastic too. Its mechanical elasticity and tenacity 
have, however, nothing to do with its electric elasticity 
and tenacity. 

We perceive in a general way why fluids can be 



I2i ETHER AND ELECTRICITY 219 

electrically, or chemically, or atomically elastic : it is 
because their molecules are doubly or multiply com- 
posed ; and the constituent atoms cling together, while 
the several molecules are free of one another. Mechani- 
cal forces deal with the molecule as a whole, and to 
them the substance is fluid ; electrical or chemical 
forces deal with the constituents of the molecule, 
setting up between them a shearing strain and en- 
deavouring to tear them asunder. To such forces, 
therefore, the fluid is elastic and tenacious, up to a 
certain limit. Extend this view of things to the 
constitution of the ether, and one has at least a 
definite position whence to proceed further. 

121. It may be convenient here to say that a 
student might find it a help to re-read Parts I. and 
II. in the light of what has just been said : remem- 
bering that, for the sake of simplicity, only the simple 
fact of an elastic medium was at first contemplated 
and insisted on ; no attempt being made to devise a 
mechanism for its elasticity by considering it as com- 
posed of two constituents. Hence the manifest arti- 
ficiality of such figures as Fig. 6, p. 35, where fixed 
beams are introduced to serve as the support of the 
elastic connexions. But it is pretty obvious now, and 
it has been already indicated in Fig. /A, p. 41, that a 
closer analogy will be obtained by considering two 
sets of beads, arranged in alternate parallel rows, 
connected by elastic threads, and displaced simulta- 
neously in opposite directions. A still further pro- 
gressive analogy is attempted in Fig. 46, p. 201. 
We have gradually passed, therefore, from a sort of 
one-fluid theory to a modified two-fluid theory ; 



220 RADIATION CHAP, xn 

believing it to be, in some sense or other, nearer the 
truth. 

Recovery of the Medium from Strain 

122. We have now to consider the behaviour of a 
medium endowed with an elastic rigidity, 4ir//c, and a 
density, 4717),' subject to displacements or strains. One 
obvious fact is that when the distorting force is 
removed the medium will spring back to its old 
position, overshoot it on the other side, spring back 
again, and thus continue oscillating till the original 
energy is rubbed away by viscosity or internal friction. 
If the viscosity is very considerable, it will not be 
able so to oscillate ; it will then merely slide back, in 
a dead-beat manner, towards its unstrained state ; 
taking a theoretically infinite time to get completely 
back, but practically restoring itself to something 
very near its original state in what may be quite a 
short time. The recovery may in fact be either a 
brisk recoil or a leak of any degree of slowness, 
according to the amount of viscosity as compared 
with the inertia and elasticity ( 19). 

The matter is one of simple mechanics. It is a 
case of harmonic motion modified by a friction 
proportional to the speed. The electrical case is 
simpler than any mechanical one, for two reasons : 
first, because so long as ' capacity ' is constant (and 
no variation has yet been discovered) Hooke's law 
will be accurately obeyed restoring force will be 
accurately proportional to displacement ; secondly, 
because for all conductors which obey Ohm's law 



123 ELASTICITY AND VISCOSITY 221 

(and no true conductor is known to disobey it) the 
friction force is accurately proportional to the first 
power of velocity. 

123. There are two, or perhaps one may say three, 
main cases. First, where the friction is great. In 
that case the recovery is of the nature of a slow leak, 
according to a decreasing geometrical progression 
or a logarithmic curve ; the logarithmic decrement 
being independent of the inertia, and being equal 
to the quotient of the elasticity and the resistance 
coefficients. 

As the resistance is made less, the recovery be- 
comes quicker and quicker, until inertia begins pro- 
minently to assert its effect, and to lengthen out 
the time of final recovery once more, by carrying 
the recoiling matter beyond its natural position, and 
so prolonging the disturbance by oscillations. The 
quickest recovery possible is obtained just before 
these oscillations begin ; and it can be shown that 
this is when the resistance coefficient is equal to twice 
the geometric mean of the elasticity and the inertia. 
One may consider this to be the second main case. 
The recoil is then exactly dead-beat, and occurs in the 
minimum of time. 

The third principal case is when the resistance is 
quite small, and when the recovery is therefore 
distinctly oscillatory. If the viscosity were really 
zero, the motion would be simply harmonic for ever, 
unless some other mode of dissipating energy were 
provided ; but if any such mode is provided, or 
if the viscosity has a finite value, then the vibra- 
tions will be simply harmonic with a dying out 



222 RADIATION CHAP. XII 

amplitude, the extremities of all the swings lying on 
a logarithmic curve. In such a case as this, the rate 
of swing is practically independent of friction ; it 
depends only on elasticity and inertia ; and, as is well 
known for simple harmonic motion, the time of a 
complete swing is 2?r times the square root of the 
ratio of inertia and elasticity coefficients. 

124. Making the statement more electrically con- 
crete, we may consider a circuit with a certain amount 
of stored-up potential energy or electrical strain in it ; 
for instance, a charged Leyden jar provided with a 
nearly complete discharge circuit. The main elastic 
coefficient, here, is the reciprocal of the capacity of the 
jar : the more capacious the jar the more "pliable" 
it is the less force of recoil for a given displacement ; 
so that capacity is the inverse of rigidity. The main 
inertia coefficient is that which is known electrically 
as the " self-induction " of the circuit : it involves the 
inertia of all the displaced matter and ether of every- 
thing which will be moved or disturbed when the jar 
is discharged. It is not a very simple thing to calcu- 
late its value in any given case ; still it can be done, 
and the general idea is plain enough without under- 
standing the exact function and importance of every 
portion of the surrounding space. (See Appendix.) 

Corresponding, then, to the well-known simple 

harmonic T = 2?r /-r-, we have, writing L for the 

self-induction or inertia of the circuit, and S for its 
capacity or inverse elasticity constant, 

T = 27iVLS: 
This, therefore, is the time of a complete swing. 



125 ELASTICITY AND DENSITY 223 

Directly the jar is discharged, these oscillations begin, 
and they continue like the vibrations of a tuning-fork 
until they are damped out of existence by viscosity 
and other modes of dissipation of energy. 

125. But now just consider a tuning-fork. Suppose 
its substance were absolutely unviscous, would it go 
on vibrating for ever ? In a vacuum it might : in air 
it certainly would not. And why not ? Because it is 
surrounded by a medium capable of taking up vibra- 
tions and of propagating them outwards without limit. 
The existence of a vibrating body in a suitable 
medium means the carving of that medium into a 
succession of waves, and the transmission of these 
waves away into space or into absorbing obstacles. 
It means, therefore, the conveyance away of the 
energy of the vibrating body, and its subsequent 
appearance in some other form wherever the radiating 
waves are quenched ( 141). 

The laws of this kind of wave-propagation are well 
known ; the rate at which waves travel through the 
medium depends not at all on any properties of the 
original vibrating body, the source of the disturbance ; 
it depends solely on the properties of the medium, if 
it be homogeneous. They travel at a rate precisely 
equal to the square root of the ratio of its elasticity to 
its density. 

Although the speed of travel is thus fixed inde- 
pendently of the source, the length of the indivi- 
dual waves is not so independent. The length 01 
the waves depends both on the rate at which they 
travel and on the rate at which the source vibrates. 
It is well known, and immediately obvious, that the 



224 RADIATION CHAP, xil 

length of each wave is simply equal to the product 
of the speed of travel into the time of one vibration. 

126. But not every medium is able to convey every 
kind of vibration. It may be that the mode of 
vibration of a body is .^entirely other than that which 
the medium surrounding it can convey : it that case 
no dissipation of energy by wave-propagation can 
result, no radiation will be excited. The only kind 
of radiation which common fluids are mechanically 
able to transmit is well known : it is that which 
appeals to our ears as sound. The elasticity, con- 
cerned in such disturbance as this, is mere volume 
elasticity or incompressibility. But electrical ex- 
periments (the Cavendish experiment, 4 and 14 A, 
and Faraday's ice-pail experiment) prove the ether 
to be enormously perhaps absolutely incom- 
pressible ; and if so, such vibrations as these would 
travel with infinite speed and not carve proper waves 
at all. 

Conceivably (I should like to say probably) 
gravitation is transmitted by such longitudinal im- 
pulses or thrusts, and in that case it is nearly or 
quite instantaneous ; and the rate at which it travels, 
if infinite, can be determined by a still more accurate 
repetition of the Cavendish experiment (the electrical 
one, not the customary gravitation-constant deter- 
mination) than has yet been made. But true radiation 
transmitted by the ether cannot be of this longi- 
tudinal character. The elasticity possessed by the 
ether is of the nature of rigidity : it has to do 
with shears and distortions ; not mechanical stresses 
to them it is quite limpid and resistless but 



126 WAVES 225 

electromotive stresses : it has an electrical rigidity, 
and it is this which must be used in the transmission 
of wave-motion. 

But the oscillatory discharge of a Leyden jar is 
precisely competent to apply to the ether these 
electromotive vibrations ; it will shake the medium in 
the mode suitable for transmission ; and accordingly, 
from a discharging circuit, waves of electrical distor- 
tion, or transverse waves, will spread in all directions, 
at a pace depending on the properties of the medium. 

Thus, then, even with a circuit of perfect conduc- 
tivity the continuance of the discharge would be 
limited, the energy would be dissipated ; not by 
friction there would in such a circuit be no direct 
production of heat it would be dissipated by 
radiation, dissipated in the same way as a hot body 
cooling, in somewhat the same way as a vibrating 
tuning-fork mounted on its resonant box. The energy 
of the vibrating body would be transferred gradually 
to the medium, and would by this be conveyed out 
and away ; its final destination being a separate ques- 
tion, depending on the nature and position of the 
material obstacles it meets with. (See 160; and 
also a lecture on Discharge of Leyden Jar, below.) 



CHAPTER XIII 

CONSTANTS OF THE ETHER 

Velocity of Electrical Radiation. 

127. THE pace at which the radiation-waves travel, 
depends, as we have said, solely on the properties of 
the medium, solely on the relation between its elasticity 
and its density. The elasticity considered must be of 
the kind concerned in the vibrations ; but the vibra- 
tions are in this case electrical, and so electrical 
elasticity is the pertinent kind. This kind of elasticity 
is the only one the ether possesses, of finite value, and 
its value can be measured by electrostatic experiments. 
Not absolutely, unfortunately : only the relative elas- 
ticity of the ether as modified by the proximity of 
gross substances has yet been measured ; its reciprocal 
being called their specific inductive capacity, or di- 
electric constant, K. The absolute value of the quan- 
tity K is at present unknown, and so a convention has 
arisen whereby in air it is called I. This convention 
is the basis of the artificial electrostatic* system of 
units. No one supposes, or at least no one has a right 
to suppose, that its value is really I. The only rational 



127 CONSTANTS OF THE ETHER 227 

guess at ethereal rigidity is one by Lord Kelvin, long 
ago, 1 which, if we reckon it as 47r//c, is equivalent 
to saying that in air the value of re is probably not 
greater than 140. Whether known or not, and at 
present it is certainly not known, the absolute value 
of the dielectric constant is manifestly a legitimate 
problem, which may any year be solved. It now seems 
more likely to be a totally different order of mag- 
nitude from the above something exceedingly small. 

The other thing on which the speed of radiation 
waves depends, is the medium's density its electric 
density, if so it must be distinguished. Here again 
we do not know its absolute value. Its relative or 
apparent amount inside different substances is mea- 
sured by magnetic experiments, and called their 
specific magnetic capacity, or permeability, and is 
denoted by /u. 

Being unknown, another convention has arisen, quite 
incompatible with the other convention just mentioned, 
that its value in air shall be called i. This convention 
is the basis of the artificial electro-magnetic system 
of units volts, ohms, amperes, farads, and the like. 
Both of these conventions cannot be true : no one 
has the least right to suppose either true. A rational 
guess at an upper limit for ethereal free density was 
made by Lord Kelvin, on the basis of a reasonable 
maximum amplitude for waves of light ; and on this 
estimate the value of p in air would come out not less 
than 8 x io~ 24 . On the other hand, a great deal of still 

1 Trans. R. S. Edin. xxii. 60; see also article "Ether," in the 
Encyc. Brit. ; and the concluding line in Vol. III. of Mathematical 
and Physical Papers. 

Q 2 



228 RADIATION CHAP, xin 

more recent work suggests that the density of the 
ether is exceedingly great much denser than of any 
known material and accordingly that its rigidity 
must be enormous. See Chap. XVII. 

128. Very well, then ; it being clearly understood 

that these two great ethereal constants, k or -, and /it, 

K 

are neither of them at present known, but are both of 
them quite knowable, and may at any time become 
known when a suitable experiment has been devised, 
it remains to express the speed of wave transmission 
in terms of them. But it is well known that this 
speed is simply the square root of the ratio of elasticity 
to density, so 

v =V? r vofe- 3 .; 

This then is the speed with which waves leave the 
discharging Leyden jar circuit, or any other circuit 
conveying alternating or varying currents, and travel 
out into space. 

Not knowing either K or //,, we cannot calculate 
this speed directly, but we can try to observe it 
experimentally. 

129. The first and crudest way of making the 
attempt would be to arrange a secondary circuit near 
our oscillating primary circuit, and see how soon the 
disturbance reached it. For instance, we might take 
a nearly closed loop, make it face a Leyden jar circuit 
across a measured distance, and then look for any 
interval of time between the spark of the primary 
discharge and the induced spark of the secondary 
circuit ; using a revolving mirror or what we please. 



130 CONSTANTS OF THE ETHER 229 

But in this way we should hardly be able to detect 
any time at all : the propagation is too quick. 

130. Since this was first written, Dr. Hertz, of 
Karlsruhe, has succeeded in making a measure of 
velocity on this very plan. He did not indeed 
actually measure the time which elapsed between 
the closing of the primary circuit and the start of the 
induced current in the secondary, nor did he use an 
ordinary Leyden ; but he converted the advancing 
waves from an electrically oscillating arrangement, 
excited by means of an induction coil, into stationary 
waves, by means of reflection at a plane metallic wall. 
Just as waves travelling along a rope or stretched 
cord are converted into stationary waves, or nodes 
and loops, by the interference of direct and reflected 
pulses : reflection taking place from the fixed end of 
the cord ; so waves advancing from an electrostatic 
oscillator, or charged body connected with the ter- 
minals of an induction coil, were reflected at the wall 
of the room (lined with sheet zinc on purpose to 
make it a conductor, and therefore a good reflector, 
see 164). Then interference with the direct waves 
converted them into stationary nodes and loops : 
the interval between two nodes being half a wave- 
length. 

By now moving the secondary circuit about, be- 
tween the primary and the wall, places of maximum 
and minimum disturbance could be found, and thus 
the wave-length measured. By calculating the oscilla- 
tion period of the primary circuit (or part-circuit, for it 
was unclosed) an indirect measure of the velocity of 
propagation was arrived at. So far as could be told, 



230 RADIATION CHAP, xm 

it agreed with measurements made by other means, 
such as those now to be described. 

131. We might next make use of the principle of 
the electric telegraph, viz. the propagation of a dis- 
turbance round a single circuit from any one point of 
origin. Consider a large closed circuit, either convey- 
ing or not conveying a current : introduce at any one 
point a sudden change a sudden E.M.F., for instance, 
or a sudden resistance if there be a current already. 
Out from that point a disturbance will spread into the 
ether, just as happens in air when a blow is struck or 
gun-cotton fired. A regular succession of disturbances 
would carve the ether into waves : a single disturb- 
ance will merely cause a pulse or shock ; but the rate 
of transmission is the same in either case, and we may 
watch for the reception of the pulse at a distant 
station. If the station has to be very distant in order 
to give an appreciable lapse of time, a speaking-tube 
is desirable to prevent spreading out in all directions 
to concentrate the disturbance at the desired spot. 
What a speaking-tube is to sound, that is the wire of 
the circuit the telegraph-wire to ethereal pulses. 

It is a curious function, this of the telegraph wire : 
it does not convey the pulses, it directs them. They 
are conveyed wholly by the ether, at a pace deter- 
mined by the properties of the ether, modified as it 
may be by the neighbourhood of gross matter. Any 
energy which enters the wires is rapidly dissipated 
into heat, and gets no further : it is the insulating 
medium round it which transmits the pulses to the 
distant station. 

All this was mentioned in Part III., and an attempt 



132 CONSTANTS OF THE ETHER 231 

was made to explain the mechanism of the process, 
and to illustrate in an analogical way what is going on 
(Chap. XL). 

The point of the matter is that currents are not 
propelled by end-thrusts, like water in a pipe or air in 
a speaking-tube, but by lateral propulsion, as by a 
series of rotating wheels with their axes all at right 
angles to the wire, surrounding it as a central core, and 
slipping with more or less friction at its surface. This 
is characteristic of ether modes in general : the ether 
does not convey longitudinal waves or end-thrust 
pulses, like sound, but it conveys transverse vibrations 
of lateral pulses, like light ( 42). 

132. Without recapitulating further, we can perceive 
then that the transmission of the pulse round the 
circuit to its most distant parts depends mainly on 
the medium surrounding it. The process is somewhat 
as follows : Consider two long straight parallel wires, 
freely suspended, and at some great distance joined 
together. At the near end of each, start equal opposite 
electromotive impulses, as by suddenly applying to 
them the poles of a battery ; or apply a succession of 
such pulses by means of an alternating machine. Out 
spread the pulses into space, starting in opposite 
phases from the two wires, so that at a distance from 
the wires the opposite pulses interfere with each other, 
and are practically non-existent ; just as but little 
sound is audible at a distance from the two prongs of 
a freely suspended tuning-fork. But near the wires, 
and especially between them, the disturbance may be 
considerable. The energy emitted by the source as 
it reaches each wire is dissipated, and so a fresh supply 



232 RADIATION CHAP, xm 

of energy goes on continually arriving at the wires, 
always flowing in from outside, to make up the de- 
ficiency. If the wires are long enough, hardly any 
energy may remain by the time their distant ends are 
reached ; but whatever there is will still be crowding 
in upon the wires and getting dissipated, unless by 
some mechanism it be diverted and utilized to effect 
some visible or audible or chemical change, and so to 
give the desired signal ( 107). 

133. Now the pace at which this transmission of 
energy goes on in the direction of the wires is pretty 
much the same as in free space. 1 There are various 
circumstances which can retard it ; there are none 
which can accelerate it. The circumstances which can 
retard it are, first, constriction of the medium by too 
great proximity of the two conducting wires, as, for 
instance, if they consist of two flat ribbons close 
together with a mere film of dielectric between, or if one 
be a small-bore tube and the other its central axis or 
core. In such cases as this the general body of ether 
takes no part in the process, the energy has all to be 
transmitted by the constricted portion of dielectric, 
and the free propagation of ethereal pulses is interfered 
with : the propagation is then no longer a true wave- 
propagation, it approximates more or less closely to 
a mere diffusion creep : rapid it may be, and yet 
without definite velocity ; like the conduction of heat, 
or like the diffusion of a salt into water. One well- 
known effect of this is to merge successive disturb- 
ances into one another, so that their individuality, 
and consequently the distinctness of signalling, is lost. 

1 Appendix (o). 



134 CONSTANTS OF THE ETHER 233 

134. Another circumstance which can modify rate 
of transmission of the pulses is ethereal inertia in the 
substance of the conducting wires : especially extra 
great inertia ; as, for instance, if they are made of iron. 
For the dissipation of energy does not go on accu- 
rately at their outer surface, it has usually to penetrate 
to a certain depth ; and, until it is dissipated, the fresh 
influx of energy from behind does not fully occur. 
Now, so long as the value of fju for the substance of 
the wires is the same as that of air or free space, no 
important retardation is thus caused, unless the wires 
are very thick ; but directly the inertia in the sub- 
stance of the wires is some hundred or thousand times 
as big as that outside, it stands to reason that more 
time is required to get up the needful magnetic spin 
in its outer layer ; and so the propagation of pulses 
is more or less retarded. At the same time this cir- 
cumstance does not alter the character of the pro- 
pagation, it does not change it from true wave 
velocity to a diffusion, it leaves its character unaltered ; 
and so the signals, though longer in coming, may 
arrive quite clear, independent^ and distinct. It is 
much the same, indeed, as if the density of the 
surrounding medium had been slightly increased. 

I have several times mentioned the name of Prof. 
Poynting as one who has developed Maxwell's equa- 
tions, and thrown great light upon the mode in which 
electro-magnetic energy is transmitted ; in the same 
connexion, and also still more prominently in con- 
nexion with the general theory of telegraphy and 
of electro-magnetic waves, I must mention with due 
emphasis the name of Mr, Oliver Heaviside, the 



234 RADIATION CHAP, xm 

scope of whose mathematical investigation into these 
difficult fields of research is remarkable. 

135. These then are the main circumstances which 
affect the rate of transmission of a pulse from one 
part of a closed circuit to another : extra inertia, or 
so-called magnetic susceptibility in the conducting 
substance, especially in its outer layers ; and undue 
constriction or throttling of the medium through 
which the disturbance really has to go. Both these 
circumstances diminish the rate of transmission ; and 
one (the last mentioned) modifies the law, and tends 
to obliterate individual features and to destroy 
distinctness. 

Of course, besides these, the nature of the insulating 
medium will have an effect on the rate of propagation ; 
but that is obvious all along ; it is precisely the rate 
at which any given medium transmits pulses that we 
want to know, and on which we are thinking of making 
experiments. If we use gutta-percha (more accu- 
rately the ether inside gutta-percha) as our transmit- 
ting medium in an experiment, we are not to imagine 
that we have obtained a result for air. 

136. The circumstances we have considered as 
modifying the rate of transmission are both of them 
adventitious circumstances , independent of the nature 
of the medium ; and they are entirely at our own 
disposal. If we like to constrict our medium, or to use 
thick iron wires, we can do so, but there is no com- 
pulsion ; and if we wish to make the experiment in 
the simplest manner, we shall do no such thing. We 
shall use thin copper wires (the thinner the better), 
arranged parallel to one another a fair distance apart, 



137 ELECTRIC WAVES ON WIRES 235 

and we shall then observe the time which an electro- 
motive impulse communicated at one end takes to 
travel to the other. Instead of using two wires, we 
may if we like use what comes to much the same 
thing, viz. a single wire suspended at a reasonable 
height above the ground, as in a common land tele- 
graph ; with circuit completed through the earth. 

The experiment, if it -could be accurately made, 
would result in the observation of a speed of pro- 
pagation equal to 3 x io 10 centimetres (300,000 kilo- 
metres, or about 185,000 miles per second. The 
actual speed in practice may be less than this, by 
reason of the various circumstances mentioned, but 
it can never be greater. This, then, is the rate of 
transmission of transverse impulses, and therefore of 
transverse waves, through ether as free as it can be 
easily obtained. 

Electric Waves along Wires. 

137. In 1888 I succeeded in making a rough pre- 
liminary determination on this very plan ; but avoided 
the necessity for excessive lengths of wire, by using 
the principle of reflection and interference to ob- 
tain stationary waves in a pair of parallel wires of 
known length attached as lateral appendages to a 
Leyden jar circuit : such waves being excited at every 
discharge an arrangement afterwards improved by 
Prof. Lecher. Experiments on these stationary waves 
in wires, observation of their nodes and loops, and 
measurements of their wave-length, were going on 
at Liverpool, simultaneously with, but independently 



236 RADIATION CHAP, xm 

of, the experiments of Hertz in Carlsruhe on waves 
in space ; and they were both described to the British 
Association at Bath in I888. 1 Alternating pulses 
travelled along the wires, and were reflected at their 
far ends, just as pulses travel along a string attached 
to a tuning-fork in Melde's experiment. Reflection 
of the pulses at the free ends of the wires is not 
accomplished without a considerable recoil or kick, 
which can be observed by the brightness of the brush 
or the length of spark it gives. The length of the 
wires, or else the size of the discharging circuit, is 
adjusted until the recoil kick is a maximum, and the 
length of each wire is then considered to be half a 
wave-length. Knowing the rate of oscillation proper 
to the particular Leyden jar circuit employed, a deter- 
mination of the velocity of the pulses can at once 
be made. It agrees with what is said above. 

Added 1906. This is the experiment of the recoil 
kick, described more in detail to the Royal Society in 
1891 (Proc. Roy. Soc. vol. 1. p. 23), and now applied 
in various systems of syntonized wireless telegraphy, 
as well as in D'ArsonvaPs apparatus for medical 
electrification at high frequency the special appli- 
ance commonly used being often called an Oudin 
resonator. 

In all these cases there is a long wire appendage to 
an oscillating Leyden jar discharging circuit, the wire 
in some cases being coiled compactly, and reinforcing 
the effect by its own self-induction ; in other cases 

1 Lodge, Phil. Mag. August 1888 ; sm&Electricianvt 1888 and 1889, 
especially vol. xxi. pp. 607 9. See also Proc. Roy. Soc. vol. 1. ; and 
Appendix (o) below. 



138 ELECTRIC WAVES ON WIRES 237 

being extended up into the air to a considerable eleva- 
tion. A development of the above idea of transmitting 
waves along a pair of wires, reflecting them and observ- 
ing the loops and nodes, so as to measure their length 
using vacuum-tubes and other detectors is now well 
known, and is commonly referred to as Prof. Lecher's 
method. 1 

138. There are many methods known to physicists 
by which an indirect experimental determination of 
this velocity of pulses along wires can be made. These 
methods have been more largely practised than the one 
described, but they do not determine directly the speed 
with which electrical pulses or waves travel : they 
directly determine the ratio of elasticity to density, or, 
what is the same thing, the product K/J, ; and it is left 
to theory to say that this is really the velocity of 
electrical pulses in free ether. It is unnecessary to say 
more about these plans here. They are generally 
referred to as methods of determining the ratio " z/," 
or the number of electrostatic units of quantity in 
an electro-magnetic unit ; which is a roundabout and 
forced mode of expression, but it serves. 

The oscillatory discharge of a Leyden jar has been 
applied by Lodge and Glazebrook to determine the 
above ratio " v." See Stokes Memorial Volume (Camb. 
Phil. Soc., Univ. Press, 1899). 

1 The waves on wires described by the writer to the British Association 
in 1888 (See E 'lee trician, 14 Sept. 1888, vol. xxi. p. 608) were exhibited 
with luminous nodes and loops, to the Physical Society of London in 
June 1891 ; as partially reported under an erroneous heading in the 
Electrician , vol. xxvii. p. 200. See also 198, and Nature ', vol. xxxix. 
p. 548. 



CHAPTER XIV 

ELECTRICAL RADIATION, OR LIGHT 

139. HAVING now described one or two possible 
methods of measuring the velocity of electric wave 
propagation, and therefore at least the ratio of the 
two ethereal constants k and //, (or, what is the same 
thing, the product of the two constants K and //,), we 
return to the consideration of an ordinary small dis- 
charging Leyden jar or other alternating current 
circuit of a moderate size, it may be a few yards or 
a foot or an inch in diameter. 

If the alternating currents are produced artificially 
by some form of alternating machine, their "frequency" 
is, of course, arbitrary ; but if they be automatically 
caused by the recoil of a given Leyden jar in a given 
circuit, their frequency the reciprocal of their time- 
period is, as we have already said ( 124), 

TTy-QN-per second : 

27T /V /(LS) r 

where L is the electrical inertia, or self-induction, of 
the circuit ; and where S is the capacity, or reciprocal 
of the elasticity-constant, of the jar. 



I4i ELECTRICAL RADIATION 239 

140. It is not convenient here to go into the deter- 
mination of the quantity L, but roughly one may 
say that for an ordinary open single-loop circuit 
it is a quantity somewhat comparable with ten or 
twelve times its circumference multiplied by the 
constant /j,. 1 

The value of S has to do with the area, A, and thick- 
ness, z, of the dielectric of the condenser ; being, as is 

A 

well known, multiplied by the constant K. 

4.7TZ 

The product LS in the above expression contains 
therefore two factors, each of linear dimensions, ex- 
pressing the sizes of circuit and jar ; and likewise 
contains a factor /JL/C expressing the properties of the 
surrounding medium. Hence, so far as the ether is 
concerned, the above expression for frequency of 
vibration demands only a knowledge of the product 
of its two constants K and fi ; and since this is known 
by the previous velocity experiments, it is easy to 
calculate the rate of oscillation of any given con- 
denser-discharge. It is also easy to calculate the 
wave-length ; for if there are n waves produced per 
second, and each travels with the velocity v, the length 

of each wave is 
n 

Hence the wave-length is 27r A /( ). 

\ \H ' K / 

141. Now, if we go through these numerical calcu- 
lations for an ordinary Leyden jar and discharger, we 
shall find waves something like, say, 50 or 100 yards 
long. They may plainly be of any length, according 

1 See Appendix (e). 



240 RADIATION CHAP, xiv 

to the size of the jar and the size of the circuit ; 
the bigger both these are the longer will be the 
waves : 

A condenser of I microfarad capacity, discharging 
through a coil of self-induction I henry, will give rise 
to ether waves 1900 kilometres or 1200 miles long; 
and the rate of its oscillation is 160 complete swings 
per second. 

A common pint Leyden jar, discharging through a 
pair of tongs, may start a system of ether waves each 
not longer than about 1 5 or 20 metres ; and its rate 
of oscillation will be something like ten million per 
second. 

A tiny thimble-sized jar overflowing its edge may 
propagate waves only about 2 or 3 feet long. (See 
also 143, 157 and 158, and Appendix (ft)). 

141 A. Professor J. J. Thomson has shown that 
when electricity oscillates from pole to pole of a 
conducting sphere, the waves emitted are 1*4 times 
the diameter of the sphere in length. I have suc- 
ceeded in recognizing waves emitted by spheres only 
2 inches in diameter ; but it is easier to get them 
clearly from bigger spheres. The waves emitted by 
spheres 6 inches or a foot in diameter, when their 
electricity is disturbed by a sudden spark, are easily 
recognizable several yards away. 1 

When the earth receives a spark, as a flash of 
lightning, its charge may oscillate between Antipodes 
and back 17 times a second. The time of one electric 
oscillation on the sun is 6\ seconds ; and the waves 

1 See Nature, vol. xli. p. 462. 



142 ELECTRICAL RADIATION 241 

are therefore a little over a million miles long. 
These great waves are possibly part of the cause of 
our terrestrial " magnetic storms." 

142. The oscillations of current, thus recognized as 
setting up waves, have only a small duration, unless 
there is some means of maintaining them. How long 
they will last depends partly upon the conductivity of 
the circuit ; but even in a circuit of infinite conduc- 
tivity they must die out, if left to themselves, from the 
mere fact that they dissipate their energy by radiation. 
One may get 10 or 20, or perhaps even 100, percep- 
tible oscillations of gradually decreasing amplitude, 
but the rate of oscillation is so great that their whole 
duration may still be an extremely small fraction of 
a second. For instance, to produce ether waves a 
metre in length requires 300,000,000 oscillations per 
second. 

To keep up continuous radiation naturally requires 
a supply of energy, and unless it is so supplied the 
radiation rapidly ceases. Commercial alternating 
machines are artificial and cumbrous contrivances for 
maintaining electrical vibrations in circuits of finite 
resistance, and in despite of loss by radiation. 

In most commercial circuits the loss by radiation 
is so small a fraction of the whole dissipation of 
energy as to be practically negligible; but we are 
not limited to the consideration of commercial 
circuits, or to alternating machines as at present 
invented and used. It may be possible to devise 
some less direct method some chemical method, per- 
haps for supplying energy to an oscillating circuit, 

R 



242 RADIATION CHAP, xiv 

and so converting what would be a mere discharge or 
flash into a continuous source of radiation. 

143. So far we have only considered ordinarily prac- 
ticable electrical circuits, and have found their waves 
in all cases pretty long, but getting distinctly shorter 
the smaller we take the circuit. Continue the process 
of reduction in size further, and ask what sized circuit 
will give w r aves 6000 tenth-metres (three-fifths of a 
micron, or 25 millionths of an inch) long. We have 

only to put 27r /v /(-.-J = O'OOOo6; and we find 

that the necessary circuit must have a self-induction, 
in electro-magnetic units, and a capacity, in electro- 
static units, such that their geometric mean is IO" 5 
centimetre (one-tenth of a micron). This gives us at 
once something near atomic dimensions for the circuit, 
and suggests immediately that those short ethereal 
waves which are able to affect the retina, and which we 
are accustomed to call " light," may be really excited 
by electrical oscillations or surgings in circuits of 
atomic dimensions ( 157-9). 

If after the vibrations are once excited there is 
no source of energy competent to maintain them, 
the light-production will soon cease, and we shall 
have merely the temporary phenomenon of phos- 
phorescence ; but if there is an available supply 
of suitable energy the electrical vibrations may 
continue, and the radiation may become no longer 
an evanescent brightness, but a steady and permanent 
glow. 



H5 ELECTRICAL RADIATION, OR LIGHT 243 

Velocity of Electrical Radiation compared with Velocity 
of Light, in Free Space and in Material Substances. 

144. We have thus imagined the now well-known 
Maxwellian theory of light, viz. that it is produced by 
electrical vibrations, and that its waves are electrical 
waves. 

But what justification is there for such an hypothesis 
beyond the mere fact which we have here insisted 
on, viz. that waves in all respects like light-waves 
except size, i.e. transverse vibrations travelling at a 
certain pace through ether, can certainly be produced 
temporarily in practicable circuits by familiar and 
very simple means, and could be produced of exactly 
the length proper to any given kind of light if only 
it were feasible to deal with circuits ultra-micro- 
scopic in size ? The simplest point to consider 
is : Does light travel at the same speed as the 
electrical disturbances we have been considering? 
We described one method of measuring how fast 
electrical radiation travels in free space, and there 
are many other methods : the result was 300,000 
kilometres per second. Does light travel at the same 
pace? 

Methods of measuring the velocity of light have 
long been known, and the result of those measure- 
ments in free space or air is likewise 300,000 kilo- 
metres a second. The two velocities agree in free 
space. Hence surely light and electrical radiation 
are identical. 

145. But there is a further test. The speed of 
electrical radiation was not the same in all media : it 

R 2 



244 RADIATION CHAP, xiv 

depended on the electrical elasticity and the ethereal 
density of the transparent substance ; in other words, 
it was equal to the reciprocal of the geometric mean 
of its specific inductive capacity and its magnetic 
permeability 



Now, although the absolute value of neither K 
nor fjb is known, yet their values relatively to air 
are often measured, and are known for most 
substances. 

Also, it is easy to compare the pace at which light 
goes through any substance with its velocity in free 
space : the operation is called finding the refractive 
index of a substance. The refractive index means, in 
fact, simply the ratio of the velocity of light in space 
to its velocity in the given substance. The reciprocal 
of the index of refraction is therefore the relative 
velocity of light. Calling the index of refraction n, 
therefore, we ought, if the electrical theory of light 
be true, to find that # 2 = /<://,; or that the index of 
refraction of any substance is the geometric mean of 
its electrostatic and magnetic specific capacities. 

146. That this is precisely true for all substances 
cannot at present be asserted. There are some sub- 
stances for which it is very satisfactorily true : there 
are others which are apparent exceptions. It remains 
to examine whether they^are not only apparent but 
real exceptions, and, if so, to what their exceptional 
behaviour is due. 

It must be understood what the essential point is. 
It has been proved by various methods, and with 



146 ELECTRICAL RADIATION, OR LIGHT 245 

greater approach to exactness as the accuracy of the 
methods is improved, that electrical disturbances 
such as the long waves emitted by any alternating 
machine travel through air or free space with exactly 
the same velocity as light ; in other words, that there 
is no recognizable difference in speed between waves 
several hundred miles long and waves so small that a 
hundred thousand of them can lie in an inch. This is 
true in free ether, and it is a remarkable fact. If it 
proves anything concerning the structure of the ether, 
it proves that it is continuous, homogeneous, and 
simple beyond any other substance ; or at least that if 
it does possess any structural heterogeneity, the parts 
of which it is composed are so nearly infinitesimal that 
a hundred miles and a hundred-thousandth of an 
inch are quantities of practically the same order of 
magnitude so far as they are concerned : its parts are 
able to treat all this variety of wave-length in the 
same manner. 

But directly one gets to deal with ordinary gross 
matter we know that this is certainly not the case. 
Ordinary matter is composed of molecules which, 
though small, are far from being infinitesimal. Atoms 
are much smaller than light-waves, indeed, but not 
incomparably smaller. Hence it is natural to suppose 
that the ether as modified by matter will be modified 
in a similarly heterogeneous manner ; and will accord- 
ingly not be able to treat waves of all sizes in the 
same way. 

The speed of all waves is retarded by entering gross 
matter, but we should expect the smallest waves to be 
retarded most. The phenomenon is well marked, even 



246 RADIATION CHAP, xiv 

within the range of such light-waves as can affect the 
retina : the smaller waves those which produce the 
sensation of blue are more retarded, and travel a 
little slower through, say, glass or water, than the 
somewhat larger ones which produce the sensation of 
red. This phenomenon has long been known, and 
is called dispersion. One result of it is that it is 
not always easy to say at what rate waves a few inches 
or a few yards or miles long ought to travel, merely by 
knowing at what rate the ultra-microscopic light-waves 
travel. 

147. But there is even more to be said than this. 
There is not only dispersion, there is selective absorp- 
tion possessed by matter : not only does it transmit 
different-sized waves at different rates, but it absorbs 
and quenches some much faster than others. Few 
substances, perhaps none, are equally transparent to 
all sizes of waves. Glass, for instance, which transmits 
readily the assortment of waves able to affect the 
retina, is practically quite opaque to waves two or 
three times as long or as short. And whenever this 
selective absorption occurs, the laws of dispersion are 
extraordinary so extraordinary that the dispersion 
is often spoken of as " anomalous " ; which of course 
means, not that it is lawless, but that its laws are com- 
plicated. Dispersion in any case is an obscure and 
little understood subject, but dispersion modified by 
selected absorption is still worse. 1 Until the theory of 
dispersion is better understood, no one is able to say 
at what speed waves of any given length ought to 

1 It was Dr. John Hopkinson who called attention to the applicability 
of this fact to the present subject (B. A. Report, 1886, page 309). 



148 ELECTRICAL RADIATION, OR LIGHT 247 

travel. One can only examine experimentally at 
what rate they do travel. This has been done for long 
electrical waves, and it has been done for short light- 
waves : in the case of some substances the speed is the 
same, in the case of others it is different. But that 
the speed should be different is, as I have now ex- 
plained, very natural, and can by no means be twisted 
into an admission that light-waves and electrical waves 
are not essentially identical. That the speed of both 
should agree at all is noteworthy ; the agreement 
appears to be exact in air, and practically exact in 
such simple substances as sulphur, and in the class 
of hydrocarbons known as paraffins ; whereas in 
artificial substances like glass, and in organic sub- 
stances like fats and oils, the agreement is less 
perfect. 

148. So much for the vital question of the speed at 
which electrical and optical disturbances travel. In 
some cases the speeds are accurately the same, in 
no case are they entirely different ; and in those cases 
where the agreement is only rough, an efficient and 
satisfactory explanation of the difference is to hand, 
in the very different lengths of wave which have at 
present been submitted to experiment. To compare 
the speeds properly, we must learn either to shorten 
electrical waves, or to lengthen light-waves, or both, 
and then compare the two things together when of 
the same size. 

It cannot be seriously doubted that they will turn 
out identical. 



248 RADIATION CHAP, xiv 

Manufacture of Light 

149. The conclusions at which we have arrived, that 
light is an electrical disturbance, and that light-waves 
are excited by electric oscillations, must ultimately } 
and may shortly, have a practical import. 

Our present systems of making light artificially are 
wasteful and ineffective. We want a certain range of 
oscillation, between 700 and 400 billion vibrations per 
second : no other is useful to us, because no other has 
any effect on our retina ; but we do not know how to 
produce vibrations of this rate. 1 We can produce a de- 
finite vibration of one or two hundred or thousand per 
second ; in other words, we can excite a pure tone of 
definite pitch ; and we can command any desired range 
of such tones continuously, by means of bellows and a 
keyboard. We can also (though the fact is less well 
known) excite momentarily definite ethereal vibrations 
of some million per second, as I have explained at 
length ; but we do not at present know how to main- 
tain this rate at all continuously. To get much 
faster rates of vibration than this we have to fall back 
upon atoms. We know how to make atoms vibrate : it 
is done by what we call " heating " the substance ; and 
if we could deal with individual atoms unhampered by 
others, it is possible that we might get a pure and 
simple mode of vibration, if we had any means of 
violently disturbing or ' clanging ' them. It is possible, 
but unlikely ; for atoms, even when isolated, have a 

1 The number of vibrations per second needed to generate visible 
light may be represented by the number of vibrations executed by a 
tuning-fork sounding a note two octaves above the middle C, continuing 
for 12,000 years. 



150 LIGHT 249 

multitude of modes of vibration special to themselves, 
of which only a few are of practical use to us, and we 
do not know how to excite some without also the 
others. However, we do not at present even deal with 
individual atoms ; we treat them crowded together in 
a compact mass, so that their modes of vibration are 
really infinite. 

We take a lump of matter, say a carbon filament or 
a piece of quick-lime, and by raising its temperature 
we impress upon its atoms higher and higher modes 
of vibration not transmuting thelower into the higher, 
but superposing the higher upon the lower until at 
length we get such rates of vibration as our retina is 
constructed for, and are satisfied. But how waste- 
ful and indirect and empirical is the 'process. We 
want a small range of rapid vibrations, and we know 
no better than to make the whole series leading up to 
them. It is as though, in order to sound some little 
shrill octave of pipes in an organ, we were obliged to 
depress every key and every pedal, and to blow a 
young hurricane. 

150. I have purposely selected as examples the 
more perfect methods of obtaining artificial light, 
wherein the waste radiation is only useless and not 
noxious. But the old-fashioned plan was cruder even 
than this : it consisted simply in setting something 
burning ; whereby not only the fuel but the air was 
consumed ; whereby also a more powerful radiation 
was produced, in the waste waves of which we were 
content to sit stewing, for the sake of the minute, 
almost infinitesimal, fraction of it which enabled 
us to see. 



250 RADIATION CHAP, xiv 

Every one knows now, however, that combustion is 
not a pleasant or healthy mode of obtaining light ; but 
everybody does not realize that neither is incandes- 
cence a satisfactory or unwasteful method which is 
like to be practised for more than a few decades, or 
perhaps a century. 

Look at the furnaces and boilers of a great steam- 
engine driving a group of dynamos, and estimate the 
energy expended ; and then look at the incandescent 
filaments of the lamps excited by them, and estimate 
how much of their radiated energy is of real service to 
the eye. It will be as the energy of a pitch-pipe to an 
entire orchestra. 

It is not too much to say that a boy turning a handle 
could, if his energy were properly directed, produce 
quite as much real light as is produced by all this mass 
of mechanism and consumption of material. 

151. There might, perhaps, be something contrary to 
the laws of Nature in thus hoping to get and utilize 
some specific kind of radiation without the rest, but 
Lord Rayleigh has shown in a short communica- 
tion to the British Association at York 1 that it is 
not so, and that therefore we have a right to try 
to do it. 

We do not yet know how, it is true, but it is one of 
the things we have got to learn. 

Any one looking at a common glow-worm must be 
struck with the fact that not by ordinary combustion, 
nor yet on the steam-engine and dynamo principle, 
is that easy light produced. Very little waste radia- 
tion is there from phosphorescent things in general. 
1 . A. Report, 1881, p. 526. 



152 LIGHT 251 

Light of the kind able to affect the retina is directly 
emitted ; and for this, for even a large supply of this, 
a modicum of energy suffices. Cf. 200. 

Solar radiation consists of waves of all sizes, it 
is true ; but then solar radiation has innumerable 
things to do besides making things visible. The 
whole of its energy is useful. In artificial lighting 
nothing but light is desired ; when heat is wanted 
it is best obtained separately by combustion. And 
so soon as we clearly recognize that light is an 
electrical vibration, so soon shall we begin to beat 
about for some mode 'of exciting and maintaining 
an electrical vibration of any required degree of 
rapidity. When this has been accomplished, the 
problem of artificial lighting will have been solved. 



Mechanism of Electrical Radiation. 

152. In forming a mental image of an electrical 
wave, we have to note that three distinct direc- 
tions are involved. There is (i) the direction of 
propagation the line of advance of the waves ; (2) 
the direction of the electric displacements, at right 
angles to this ; and (3) the direction of the magnetic 
axis, at right angles to each of the other two. 

One may get a rough mechanical idea of the 
process of electrical radiation (at any rate in a plane) 
by means of the cog-wheel system already used in 
Part III. Imagine a series of elastic wheels, in one 
plane, all geared together, and let one of them be 



252 RADIATION CHAP, xiv 

made to twist to and fro on its axis ; from it, as 
centre, the disturbance will spread out in all direc- 
tions, each wheel being made to oscillate similarly 
and to transmit its oscillation to the next. Looking 
at what is happening at a distance from the source : 
we shall see the pulses travelling, say, from left to right, 
while the electrical displacement is up and down, and 
the oscillating axes of the wheels are to and fro, or 
at right angles to the plane containing the wheels. 
It is therefore appropriately spoken of as electro- 
magnetic radiation rather than merely as electric 
radiation. The energy of the electrostatic strain is 
just as great as that of the electro-magnetic motion ; 
in fact, the energy alternates from the potential to 
the kinetic form, or vice versa, at every quarter swing, 
just like every other case of vibration ; and on the 
average the two energies, in any considerable portion 
of space in which the waves are travelling, are 
necessarily exactly equal. If they are not equal, there 
will be reflection and complications. 

153. In ordinary cases, the oscillations are all 
very small. For just consider that the wheel-work 
extends right away to infinity in all directions ; how 
is any moderate force going to make one of these 
wheels oscillate ? If they were rigid it would be impos- 
sible, but as they are elastic it is possible, though only 
with a very small amplitude of vibration ; and it sets 
up a strain all round, which rapidly spreads out, as we 
have said, in all directions from the source. If the 
source were inclosed in a perfect conductor 01 
moderate dimensions if, for instance, we proceed 
to rotate to and fro one of a bounded series of 



154 ELECTRICAL RADIATION, OR LIGHT 253 

wheels, such as those inside the empty contour of 
Fig. 38 we can do it easily : the wheels are then 
limited in number, and can be easily got to oscillate 
considerably by a feeble source of disturbance. 

This is commonly spoken of as concentration of 
light by reflection : the conductor is said to act as a 
perfect mirror ; and, since none of the light escapes, 
any amount of illumination can be produced inside a 
closed spherical mirror of perfect conductivity. Such 
illumination would not be much use, however ; for, 
directly a bit of matter is introduced to receive the 
benefit of it, dissipation goes on at its surface, and 
the violence of the ethereal disturbance is brought 
down to something more moderate. Nevertheless, 
even when dissipation is allowed, and when the 
reflecting surface is by no means perfectly conduct- 
ing, but is bright silver, which is the best conductor 
we know, a considerable increase in illumination is 
caused : by reflection, if we choose to say so by 
limitation, in at least some directions, of the extent of 
ethereal medium to be affected by a given source, as 
we might now prefer to express it ( 164). 

154. Prof. Fitzgerald, of Dublin, devised a model 
of the ether, which by help of a little artificiality 
represents the two kinds of displacement the electric 
and magnetic very simply and clearly. 

His wheels are separated from one another by a 
certain space, and are geared together by elastic bands. 
They thus turn all in one direction, and no mention 
need be made of positive and negative electricity as 
separate entities. 

But, the wheels being massive, a rotatory disturb- 



254 



RADIATION 



CHAP. XIV 



ance given to one takes time to spread through 
the series, at a pace depending on the elasticity of 
the bands and the inertia of the wheels ; and during 
the period of acceleration one side of every elastic 
is stretched, while the other side is relaxed and there- 
fore thickened. This thickening of the elastics goes 
on in one direction, and corresponds to an electric 
displacement in that direction ; the direction being 




FIG. 48. Fitzgerald's Ether Model. A set of massive-rimmed brass wheels on 
fixed pivots with their axles connected by common elastic bands. If the bands 
are taken off any region, it becomes a perfect conductor, into which disturbances 
cannot penetrate. (Cf. Fig. 38.) Proc. Roy. Dub. Sec., January 1885. 

perpendicular both to the direction of advance of 
the disturbance and to the axes of the wheels. A 
row of wheels corresponds to a section of a wave- 
front ; the displacements of india-rubber and the 
rotating axes, i.e. the electric and the magnetic dis- 
turbances, both lie in the wave-front. 

155. Clerk Maxwell's originally suggested repre- 
sentation was not unlike this. 1 It consisted of a 
series of massive wheels, connected together not by a 
series of elastic bands but by a row of elastic particles 
or " idle wheels." These particles represented " elec- 
tricity"; their displacement during the period of ac- 
celeration corresponding to the one-sided thickening 
of the elastic bands in Fitzgerald's model. The 
object of the idle wheels was to enable all the main 

1 Phil. Mag. April 1861. 



156 ELECTRICITY AND ETHER 255 

wheels to rotate in the same sense, as desiderated in 
Fig- 35, P- 173- 

Added 1906. It is worth while to reconsider Max- 
well's model in the light of modern electron theory ? 
and to realise how near the truth his intuition led 
him. 

I have proposed to contemplate a double series of 
wheels geared directly into one another, representing 
positive and negative electricity respectively, because 
it seems to me that so many facts point to the 
existence of these two entities ; and because then no 
distinction has to be drawn between one part of the 
medium which is ether and another part which is 
electricity, but the whole is ether and may be also 
electricity; while, nevertheless, a much-needed distinc- 
tion can be drawn between a motion of the ether as a 
whole, and a relative motion of its component parts : 
a distinction between forces able to move ether i.e. to 
displace the centre of gravity of some finite portion of 
it and forces which shear it and make its components 
slide past each other in opposite senses ; these latter 
forces being truly electromotive ( 120). 

156. If it be asked how the elasticity of the ether is 
to be explained, we must turn to the vortex sponge 
theory, suggested by Prof. Fitzgerald, 1 as well as by 
Prof. Hicks 2 and recently elaborated by Lord Kelvin. 3 
But this is too complicated a matter to be suited for 
popular exposition just at present. It must suffice 

1 Proc. R. Dub. Soc. Jan. 1885. See also his Collected Memoirs. 

2 Brit. Assoc. Report ', 1885, Aberdeen, p. 930. 

3 Ibid. 1887, Manchester, p. 486. Also Phil. Mag. October 1887. 



256 RADIATION CHAP, xiv 

to indicate that the points here left unexplained are 
not necessarily at the present time unexplainable, 
but that the explanations have not yet been so com- 
pletely worked out that an easy grasp can be obtained 
of them by simple mechanical illustrations and con- 
ceptions. At the same time, the general way in 
which motion is able to simulate the effects of elas- 
ticity will be found popularly illustrated, by help 
of gyrostats, in Lord Kelvin's article " Elasticity " 
in the Encyclopedia Britannica ; x and the fact 
that elastic rigidity of a solid can be produced 
by impressing motion on a homogeneous and 
otherwise structureless fluid must be regarded as 
one of the most striking among his many funda- 
mental discoveries. 

We have found it necessary all through Part III. 
to imagine the ether as composed of cells containing 
electricity in rotation, and the act of magnetiza- 
tion as consisting in facing these whirls round. Lord 
Kelvin has taught us that a medium containing 
whirls like this will simulate the behaviour of an 
elastic solid, and in fact that whirling motion is all 
that is required to explain elasticity (Fig. 46). With 
this hint, which might be developed at greater length, 
I must leave this part of the subject. 

157. We have seen that, to generate radiation, an 
electrical oscillation is necessary and sufficient ; and we 
have attended mainly to one kind of electric oscilla- 
tion, viz. that which occurs in a condenser circuit when 
the distribution of its electricity is suddenly altered 

1 Also in a volume of the Nature Series Popular Lectures and 
Addresses, vol. i. " The Constitution of Matter." 



157 ELECTRICAL RADIATION, OR LIGHT 257 

as, for instance, by a discharge ( 124, 141). But the 
condenser circuit need not be thrown into an obviously 
Leyden jar form ; one may have a charged cylinder 
with a static charge accumulated mainly at one end, 
and then suddenly released. The recoil of the charge 
is a true current, though a weak one ; a certain amount 
of inertia is associated with it, and accordingly oscilla- 
tions will go on, the charge surging from end to end of 
the cylinder, like the water in a tilted bath suddenly 
levelled : though the electrical surgings in this case are 
few. 

In a spherical or any other conductor, the like 
electric oscillations may go on ; and the theory of these 
oscillations has been treated with great mathematical 
power both by Sir Wm. Niven and by Prof. Lamb. 1 

Essentially, however, the phenomenon is not distinct 
from a Leyden jar or condenser circuit ; for the ends of 
the cylinder have a certain capacity, and the cylinder 
has a certain self-induction ; the difficulty of the 
problem may be said to consist in finding the values 
of these things for the given case. The period of 
an oscillation may still be written 2?r x /(LS) ; only, 
since L and S are both very small, the " frequency " 
of vibration is likely to be excessive. And when we 
come to the oscillation of an atomic charge the 
frequency may perhaps surpass the rate of vibration 
which can affect the eye. The damping out of such 
vibrations, if left to themselves, will be also a very 
rapid process, because the initial energy is but small. 
It can be calculated that the oscillation of an atomic 

1 Phil. Trans. 1881 and 1883. Also by Prcf. J. J. Thomson, 
Math. Soc. Proc. April 1884. 

S 



258 RADIATION CHAP, xiv 

charge would give rise to ultra-violet rays. It is 
probably because these ultra-violet rays synchronize 
with the period of vibration of atomic charges that 
they have such extraordinarily powerful chemical 
effects ( 187). 

I5/A. The waves of visible light, if produced by 
electric oscillation in atoms, must be caused, not by 
simple oscillations as in conducting spheres, but by 
surgings in conducting channels, more after the fashion 
of a Leyden jar circuit. But it may be held more 
probable that the vibrations to which ordinary light 
is due are mechanical vibrations of the substance 
of the atoms, and that the electrical disturbance 
accompanying it and recognized by the retina is a 
secondary effect. Light is now believed to be emitted 
by perturbations in the orbital revolution of electronic 
constituents of an atom. 

I have also pointed out l that the rods and cones in 
the retina are of a diameter suitable for responding 
directly to electric oscillations of the frequency of 
visible light ; and by means of a graduated series of 
metallic cylinders facing a source of electric radiation 
end-on, have so to speak imitated a retina which is 
able to select and respond to vibrations of assorted 
pitch, after the manner of Corti's fibres in the ear. 
(See Fig. 60 below.) But at present there is no phy- 
siological basis for such an assortment of size among 
the rods and cones as would correspond to the three 
colour sensations red, green, and violet. Perhaps it 
will be looked for. 

The correspondence of the diameter of the rods 

1 Nature, vol. xli. p. 462, March 1890. 



158 ELECTRICAL RADIATION, OR LIGHT 259 

and cones to the wave-length of ordinary light may 
be accidental, but it seems hardly likely to be mere 
coincidence. 

In all probability sight is a chemical sense, some 
molecular complexes being shaken asunder by the 
impact of synchronous ether waves, and thus stimu- 
lating the associated nerve-fibres. The nourishment 
supplied by the tissues may be trusted to build up 
the substances again almost as fast as they are shaken 
asunder ; leaving, however, a little margin for lag, to 
correspond to the observed fact of retinal " fatigue." 

158. Whether the charge oscillates in a stationary 
conductor, or whether a charged body vibrates as 
a whole, it equally constitutes an alternating current 
and can equally well be treated as a source of radiation. 
Now when we were considering the subject of elec- 
trolysis, we were led to think of molecules as com- 
posed of two atoms or groups of atoms, each charged 
with equal quantities of opposite kinds of electricity. 
Under violent disturbance we may suppose the com- 
ponents of the molecules to be set in vibration the 
rate of vibration depending on and being characteristic 
of the contents of the particular molecule. The atoms 
being charged, however, an electric radiation is excited 
and propagated outwards. These vibrations would 
appear to be often of the frequency suited to our 
retina ; hence these vibrating atoms indirectly 
constitute our usual source of light. The 
" frequency " of the visible radiation can be 
examined and determined by optical means (some 
form of interference experiment, usually a diffraction 
grating), and hence many of the rates of vibration 

S 2 



260 RADIATION CHAP, xiv 

possible to the atoms of a given molecule under given 
circumstances become known : and this is the founda- 
tion of the science of spectroscopy. 

It is possible that the long duration of some kinds 
of phosphorescence may be due to the atoms receiving 
indirectly some of the ethereal disturbance, and so 
prolonging it by their inertia, instead of leaving it to 
the inertia of the ether alone. It is possible 
also that the definite emissivity of some fluorescent 
substances is due to periods of vibration proper to 
their atoms, which, being disturbed in an indirect way 
by receipt of radiation, re-emit the same radiation in 
a modified, and, as it were, laden manner. 

159. To get some further idea concerning the way in 
which an oscillating charge, or an oscillating charged 
body, can propagate radiation, refer back to Fig. 39, 
Part III., and imagine the rack oscillating to and fro. 
It will produce rotatory oscillation in the wheels 
gearing into it, these again in the next, and so on. I f 
the wheel-work were rigid, the propagation would go 
on at an infinite speed to the most distant wheels ; but 
if it be elastic, then the pace of propagation depends 
on the elasticity and the density, in a way we have 
already said enough about. The line of rack is the 
direction of electric oscillation, the axes of the wheels 
the direction of magnetic rotatory oscillation, and at 
right angles to both these is the direction of advance 
of the waves. True, the diagram is not a space 
representation, it is a mere section, and is only a crude 
suggestion of a mechanical analogy to what may be 
taking place. 

The wheels, being perfectly geared together and into 



160 ELECTRICAL RADIATION, OR LIGHT .261 

the rack, represent an insulator or dielectric ; there is 
no slip or frictional dissipation of energy in other 
words, there are no true electric currents. The electric 
oscillation is a mere displacement-oscillation due to 
elasticity and temporary give of the elastic wheels ; 
whereby, during each era of acceleration, they are 
thrown slightly into the state represented in Fig. 46 
as contrasted with Fig. 37. 

Effects of encountering a New Medium. 

1 60. Now contemplate an advancing system of 
waves, and picture their encounter with an obstacle ; 
say, a medium of greater density, or less elasticity, or 
both. If the new medium is a perfect insulator, it 
must be considered as having its wheels thoroughly 
geared up, both with each other and with those of 
the initial medium, so that there is no slip or dissipa- 
tion of energy at the surface. In this case none of 
the radiation will be lost : some will be reflected, 
and some transmitted, according to ordinary and 
well-known mechanical laws. The part transmitted 
will suddenly begin to travel at a slower pace, and 
hence if the incidence were oblique would pursue 
a somewhat different path. Also, at the edges of 
the obstacle, or at the boundary of any artificially 
limited portion of the wave, there will be certain 
effects due to spreading out and encroaching on 
parts of the medium not lying in the direct path. 
These refraction and diffraction effects are common to 
all possible kinds of wave propagation, and there is 
nothing specially necessary to be said concerning 



262 RADIATION CHAP, xiv 

electrical radiation on these heads which is not to be 
found in any work on the corresponding parts of 
optics, except that the velocity in every non-absorp- 
tive substance is characterized by the two ethereal 
constants, the electric K and the magnetic /JL ; being 
equal, in fact, to the reciprocal of their geometric 
mean. 

161. Concerning the amount and direction of the 
reflected vibrations there is something to be said, 
however, and that something very important j 1 but it 
is no easy subject to tackle, and I fear must be 
left, so far as I am concerned, as a distinct, but 
perhaps subsequently-to-be-filled-up, gap. 

If the gearing between the new medium and the 
old is imperfect if, for instance, there were a layer of 
slippery wheels between them, representing a more or 
less conducting film then some of the radiation would 
be dissipated at the surface, not all would be reflected 
and transmitted, and the film would get to a certain 
extent heated. By such a film the precise laws of re- 
flection might be profoundly modified, as they would 
be also if the transition from one medium to another 
were gradual instead of abrupt. But all these things 
must remain, for the present, part of the unfilled gap. 

Electric Radiation encountering a Conductor. 

162. We will proceed now to the case of a conducting 
obstacle that is, of waves encountering a medium 
whose electrical parts are connected, not by elasticity 
but by friction. It is plain here that not only at the 

1 Lord Rayleigh, Phil. Mag. Sept. 1888. 



1 63 ELECTRICAL RADIATION, OR LIGHT 263 

outer layer of such a medium, but at every subse- 
quent layer, a certain amount of slip will occur during 
every era of acceleration, and hence that in penetrating 
a sufficient thickness of a medium endowed with any 
metallic conductivity the whole of the incident radia- 
tion must be either reflected or destroyed : none can 
be transmitted ( 104). 

Refer back to Fig. 43, and think of the rack in that 
figure as oscillating. Through the cog-wheels the dis- 
turbance spreads without loss, but at the outer layer ' 
of the conducting region A B c D a finite slip occurs, 
and a less amount of radiation penetrates to the next 
layer, E F G H, and so on. Some thickness or other, 
therefore, of a conducting substance must necessarily 
be impervious to electric radiation : that is, it must be 
opaque. And since a good conductor dissipates very 
little energy, it will not act by absorbing the radiation, 
but will reflect it. (See 153 and 164.) 

Conductivity is not the sole cause of opacity. It 
would not do to say that all opaque bodies must be 
conductors. But conductivity is a very efficient cause 
of opacity, and it is true to say that all conductors of 
electricity are necessarily opaque to light ; under- 
standing, of course, that the particular thickness of 
any homogeneous substance which can be considered 
as perfectly opaque must depend on its conductivity. 
It is a question of degree, and a minute but 
specifiable fraction of an original disturbance may be 
said to get through any obstacle. Practically, how- 
ever, it is well known that a thin, though not the 
thinnest, film of metal is quite impervious to light. 

163. The statement that conductivity is not the sole 



264 RADIATION CHAP, xiv 

cause of opacity has reference to the kind of opacity 
caused by heterogeneity. A confused mass of per- 
fectly transparent substance may be quite opaque ; 
witness foam, powdered glass, chalk, etc. 

Hence, though a transparent body must indeed be 
an insulator, the converse is not necessarily true ; an 
insulator need not necessarily be transparent. A 
homogeneous, flawless insulator must, however, be 
transparent to some, though not necessarily to all, 
wave-lengths. A homogeneous and flawless opaque 
body, if really opaque to all wave-lengths, must be a 
conductor. 

These, then, are the simple connexions between 
two such apparently distinct things as conducting 
power for electricity and opacity to light, .which 
Maxwell's theory points out ; and it is possible 
to calculate the theoretical opacity of any given 
simply-constructed substance by knowing its specific 
electric conductivity. There are, however, reasons for 
asserting that the opacity as so calculated does not 
agree very well with that actually observed. There 
is a field for work here. 1 



Fate of the Radiation. 

164. To understand what happens to radiation 
impinging on a conducting body, it is most simple to 
proceed to the limiting case at once and consider a 
perfect conductor. In the case of a perfect conductor 
the wheels are connected not even by friction ; they 

1 For a further discussion of opacity, see Lodge, Phil. Mag. April 
1899. 



1 66 ELECTRICAL RADIATION, OR LIGHT 265 

are not connected at all. Consequently the slip at 
the boundary of such a conductor is perfect, and there 
is no dissipation of energy accompanying it. The 
blank space in Fig. 38 represented a perfectly con- 
ducting layer. Ethereal vibrations impinging on a 
perfect conductor practically arrive at an outer con- 
fine of their medium : beyond, there is nothing capable 
of transmitting them ; the outer wheels receive an 
impetus which they cannot get rid of in front, and 
which they therefore return back, the way it came, to 
those behind them : the radiation is totally reflected. 
It is like what happens when a sound-pulse reaches the 
open end of an organ-pipe ; like what happens when 
sound tries to go from water to air ; like the last of a 
row of connected balls along which a knock has been 
transmitted ; except that in the case of these longi- 
tudinal pulses the phase is reversed, whereas for 
transverse pulses under the same circumstances it is 
just not reversed. Our massive elastic wheels, espe- 
cially the wheels of Fig. 48, are able to represent the 
act of reflection quite properly. 

165. The reflected pulses will be superposed upon 
and interfere with the direct pulses, and accordingly, 
if the distances are properly adjusted, we can have 
the familiar formation of fixed nodes and stationary 
waves ( 130). 

1 66. The point of main interest, however, is to 
notice that a perfect conductor of electricity, if there 
were such a thing, would be utterly impervious to 
light : no light could penetrate its outer skin, it 
would all be reflected back : the substance would 
be a perfect reflector for ether waves of every size. 



266 RADIATION CHAP, xiv 

Thus with a perfect conductor, as with a perfect 
non-conductor, there is no dissipation. Radiation 
impinging on them is either all reflected, or some 
reflected and some transmitted. It is the cases of 
intermediate conductivity which destroy some of the 
radiation and convert its ethereal vibrations into 
atomic vibrations, i.e. which Convert it into heat. 

167. The mode in which radiation or any other 
electrical disturbance diffuses with continual loss 
through an imperfect conductor can easily be appre- 
ciated by referring to 103 again. The successive 
lines of slip, A B C D, E F G H, &c., are successive layers 
of induced currents. An electromotive impulse loses 
itself in the production of these currents, which are 
successively formed deeper and deeper in the material, 
according to laws of diffusion. 

If the waves had impinged on one face of a slab, a 
certain fraction of them would emerge from the other 
face a fraction depending on the thickness of the 
slab, according to a logarithmic or geometrical-progres- 
sion law of decrease. (See also Chap. XVI.) 

It may occur to a student to ask what happens to 
the energy of light in cases of " interference." Inter- 
ference is often popularly expressed by saying that 
two lights under certain conditions, when superposed, 
produce darkness ; two sounds produce silence, and 
so on. Hence it is natural to inquire what becomes 
of the energy. The answer is that in all such cases 
the energy is merely redistributed and localized, not 
affected in quantity. Interference phenomena always 
distribute themselves in bands, either in space or 
time. In the case of light there are visible " inter- 



167 INTERFERENCE 267 

ference-bands," the light being removed from certain 
dark strips and piled up in intermediate extra-lumin- 
ous strips, but the total illumination is unaltered ; it 
is only a question of distribution. In the case of 
sound a similar phenomenon shows itself as "beats/ 5 
and the periods of silences are compensated by extra 
loudness in the alternating sounds. The same is true 
of the criss-cross pattern when surface waves of water, 
generated by throwing in two stones, cross each 
other. It is a familiar fact that after the crossing 
the waves proceed uninjured and unaffected. Inter- 
ference is not the same as destruction ; the inter- 
ference is only local and temporary. 



CHAPTER XV 

ELECTRIC AND MAGNETIC INFLUENCES ON LIGHT 

168. WE must now mention one or two phenomena 
which depend entirely upon a modification of ether 
by the neighbourhood of matter, and which we have 
reason to believe would not occur in free ether at all. 
These are the optical phenomena of Faraday and Kerr, 
and the electric phenomenon of Hall. 

Faraday discovered, long before there was any other 
connexion known between electricity and light 
except indeed the familiar luminous phenomena which 
accompany electric discharges of various kind that 
the plane in which light-vibrations occur could be 
rotated by transmitting light through certain magnet- 
ized substances along the lines of magnetic force. To 
make this effect easily manifest, it is necessary to use 
plane-polarized light, and to transmit it through a 
fair length of magnetized substance, analyzing it after 
emergence, and showing that, though it remains plane- 
polarized, the plane has been rotated, possibly through 
a right angle or more. 

Now in a general way it is easy to imagine that, 



1 69 ELECTRO-OPTIC EFFECTS 269 

inasmuch as something of the nature of a rotation is 
going on in a magnetic field round the lines of force, 
vibrations travelling into such a field along these lines 
should be twisted round corkscrew fashion and emerge 
vibrating in a different plane. But on trying to follow 
out this process into detail, it turns out not quite so 
simple a matter. It has, however, no business to be 
a very simple and obvious consequence of the 
existence of a magnetic rotation round the rays of 
light, else would it occur in free space, and in the 
same direction in all media. But the facts are that 
in free space that is, in free ether it does not occur 
at all ; and the direction of rotation is not the same 
for all media : substances can, in fact, be divided into 
two groups, according to the way in which given 
magnetization shall rotate the plane of polarized 
light passing through them. 

169. Similar statements can be made concerning the 
electrostatic optical effect discovered by Dr. Kerr ; who 
showed that plane-polarized light transmitted across 
the lines of force in an electrostatic field could, in cer- 
tain media, come out elliptically polarized. Now, in- 
asmuch as an electric field is a region of strain, and 
strain in transparent bodies is well known to make 
them slightly doubly refracting and able to turn plane- 
polarized into elliptically polarized light, it is very 
easy to imagine such a result in an electric field to be 
natural and probable. But the explanation is not so 
simple as that, else it ought to be a large effect, occur- 
ring in all sorts of media in the same direction, and 
likewise in free space. But the facts are that it does 
not occur at all in free space, and it occurs in different 



270 RADIATION CHAP, xv 

senses in different substances ; so that again they can 
be grouped into two classes, according to the sign of 
the Kerr effect. 

Thus, then, the rotatory effect of a magnetic field 
upon light, discovered by Faraday, and the doubly 
refracting effect of an electrostatic field upon light, 
discovered by Kerr, agree in this : that they are both 
small or residual effects, depending on the existence of 
a dense or material medium, and both varying in sign 
according to the nature of the medium. 

170. The only substance in which the Faraday effect 
is large, is iron ; including with iron the other highly 
magnetic substances. The discovery of the effect in 
these bodies was likewise made by Kerr. The diffi- 
culty of dealing with them is that they are very 
opaque, and hence that the merest film of them can 
be used. The film can be used either by way of 
transmission or by way of reflection, it matters not 
which, but reflection is the way in which it was first 
done. Light reflected from the polished face of a 
magnet has indeed barely penetrated at all into the 
substance of the iron before being sent back ; still, it 
has penetrated deep enough to be distinctly rotated by 
the tremendous magnetic whirl which it finds there. 

171. All these highly magnetic substances are 
metallic conductors, and are therefore very opaque. 
Whether there is any real connexion between high 
magnetic susceptibility and conductivity is more than 
I can say. But it is quite natural, and indeed 
necessary, that the greatest portion of light should 
be reflected on entering a highly magnetic medium ; 
because in such a medium the ethereal density, 47r//,, 



172 



ELECTRO-OPTIC EFFECTS 



271 



is so great, and hence the velocity of wave trans- 
mission must undergo a sudden and immense decrease 
a circumstance always causing a great amount of 
reflection ; just as when sound tries to pass from any 
one medium to a much denser one. 

But the opacity of iron and other magnetic sub- 
stances may be explained by the mere fact of their 
conducting power, just like other metals, and no 
noteworthy effect of their large value of p need be 
detectable optically. 

If a non-conducting, highly magnetic, substance 
could be found, it would probably reflect a great deal 
of light at its surface, though it would not dissipate 
that which entered it. Such a substance would be 
most interesting to submit to experiment, but perhaps 
its existence presupposes a combination of impossible 
properties. Certainly it has not yet been discovered. 

As to the phenomenon detected by Hall, it appears 
intimately associated with that of Faraday, and it 
will be most simple to omit all reference to it for the 
present. 

172. A general idea of what is happening in the 
Faraday and Kerr phenomena can be given thus. 
D g p 





C Q 

FIG. 49. FIG. 50. 

A simple vibration, like a pendulum-swing, or any 
other oscillation in one plane, can be resolved into 



272 RADIATION CHAP, xv 

two others in an infinite variety of ways ; just as one 
force can be resolved into any number of pairs of 
equivalent forces. The two most useful modes of 
analyzing a simple vibration into a pair of constituents 
are these : (i) two equal components, likewise plane 
vibrations, each inclined at 45 to the original, as 
when P Q is resolved into A B and C D (Fig. 49) ; and 
(2) two equal circular or rotatory oscillations in oppo- 
site directions, as when P Q is resolved into P M Q and 
P N Q (Fig. 50). The first method of resolution is useful 
in explaining Kerr's effect, the second in explaining 
Faraday's. 

Of the two component vibrations, A B and C D, into 
which P Q can be supposed analyzed, let some cause, 
no matter what, make one gain upon the other, so that 
in travelling along a line perpendicular to the paper 
one goes a little the quicker : the effect at once is to 
change the character of the vibration into which they 
will recompound. After the gain, they no longer re- 
produce the original simple vibration P Q : they give 
rise to elliptic, or it may be to circular, vibrations ; this 
last, if the retardation is equal to a quarter period. 

These are matters fully treated in any elementary 
treatise on polarized light, and they are quite easily 
illustrated by means of a simple pendulum. We 
may assume them known. 

Similarly with the second system of analyzing the 
vibration into two opposing circular ones. If the 
components travel through any interposed medium 
at the same rate, they will, on emergence, reproduce 
the original vibration in its original position ; but if 
one travels quicker than the other, they recombine 



173 ELECTRO-OPTIC EFFECTS 273 

into a vibration of the same character as at first, but 
turned through a certain angle. Thus, anything which 
retards one of the rectangular components behind the 
other changes the character of the vibration from 
plane into elliptical; while anything which retards one 
of the circular components behind the other leaves 
the character of the vibration unaltered, but rotates it 
through a certain angle. 

173. So far we have said nothing but the simplest 
mechanics. The next point to consider is, What deter- 
mines the rate at which light travels through any sub- 
stance? This we have discussed at length ( 128), and 

shown to be 77 r. Anything which increases either 



the electric or the magnetic permeability of the 
medium decreases the velocity of light. Now, when 
a medium is already subject to a violent strain in any 
one direction, it is possibly less susceptible to further 
strain in that direction and responds less readily. Not 
necessarily so at all : such an effect would only be 
produced when the strain was excessive, when the 
medium was beginning to be overdone, and when its 
properties began thereby to be slightly modified. 
There are reasons for believing the specific inductive 
capacity of most media to be very constant ; of some 
media, perhaps, precisely constant ; but if there were 
any limit beyond which the strain could not pass, it is 
probable that on nearing the limit the specific induc- 
tive capacity would be altered possibly increased, 
possibly diminished one could hardly say which. 
Ouincke has investigated this matter, and has shown 
that the value of K is affected by great electric strain. 

T 



274 RADIATION CHAP, xv 

Suppose now that a dielectric is subject to a violent 
electrical stress, so that its properties along the lines 
of force become slightly different from its properties 
at right angles to those lines. The value of K will 
not be quite the same along the lines of strain as across 
them, and accordingly the rectangular component of 
a vibration resolved along the lines of force will travel 
rather quicker or rather slower than the component 
at right angles ; because the velocity of transmission 
depends upon K, as already explained: such a medium 
at once acquires the necessary doubly-refractive 
character, and will show Kerr's effect. 

174. Similarly with magnetization. It is well 
known that for many media p, is not constant. Take 
iron, for instance. For very small magnetizing forces 
the susceptibility is moderate, and increases as they 
increase; at a certain magnetization it reaches a maxi- 
mum, and then steadily decreases. But not only is 
it thus very inconstant, its ascending and descending 
values are not the same. To forces tending to mag- 
netize it more, the susceptibility has one value ; to 
forces tending to demagnetize it, it has another and 
in general smaller value. This property has been 
specially studied by Ewing, and has been called by 
him "hysteresis." Slightly susceptible substances 
cannot be magnetized to anything like the same 
extent, and hence the property in them has been less 
noticed, perhaps not noticed at all. Nevertheless it 
must exist in every substance which exhibits a trace 
of permanent magnetism, and every substance I have 
tried appears to show some such trace. 1 

1 See Nainre^ vol. xxxiii. p. 484. 



175 ELECTRO-OPTIC EFFECTS 275 

175. The direction of rotation will depend on 
whether the value of p is greater for small relaxa- 
tions, or for small intensifications, of magnetizing 
force ; and diamagnetic substances may be expected 
to be opposite in this respect to paramagnetic ones. 
Any substance for which /JL is absolutely constant, 
whatever the strength of magnetic polarization to 
which it is submitted, can hardly be expected to ex- 
hibit any hysteresis ; the ascending and descending 
curves of magnetization will coincide, being both 
straight lines, and such a substance will show no 
Faraday effect. Similarly, any substance for which K 
is absolutely constant, whatever the electric polariza- 
tion to which it is submitted, can show no Kerr's effect. 
Free space appears to be of this nature; and gases 
approach it very nearly, but not quite. 

In iron, p is greater for an increasing than for a 
decreasing force, as is shown by the loops in Ewing's 
curves ; hence the circular component agreeing in 
direction with the magnetizing current will travel 
slower than the other component, and hence the 
rotation in iron will be against the direction of the 
magnetizing current. The same appears to hold in 
most paramagnetic substances, and the opposite in 
most diamagnetic ; but the mere fact of paramag- 
netism or diamagnetism is not sufficient to tell us 
the sign of the effect in any given substance. We 
must know the mode in which its magnetic per- 
meability is affected by waxing and by waning 
magnetization respectively. 



T 2 



276 RADIATION CHAP, xv 



Possible Electrical Method of detecting the Faraday 

Effect. 

176. Thus far we have considered the rotation of 
electric displacement by a magnetic field as being 
examined optically ; the displacements being those 
concerned in light, and the rotation being detected 
by a polarizing analyzer suitable for determining the 
direction in which the vibrations occur, before and 
after the passage of light through a magnetized sub- 
stance. This is the only way in which the effect has 
at present been observed in transparent bodies. But 
we ought not to be limited to an optical method of 
detection. 

Electrical displacements are easily produced in any 
insulator, and if it be immersed in a strong magnetic 
field, so that the electric and magnetic lines of force 
are at right angles to each other, every electric dis- 
turbance ought to experience a small rotation. A 
steady strain will not be affected ; it is the variable 
state only which will experience an effect ; but every 
fresh electric displacement should experience a slight 
rotatory tendency, just like the displacements which 
occur in light. 

Now to rotate a displacement A B into the position 
A C requires the combination with it of a perpen- 
dicular displacement B C (Fig. 51). Hence the effect 
of the magnetic field upon an electric displacement, 
A B, may be said to be the generation of a small per- 
pendicular E.M.F., B C, which, compounded with the 
original one, has the resultant effect A C. It will be 



176 



ELECTRO-OPTIC EFFECTS 



277 



only a temporary effect, lasting while the displace- 
ment is being produced, and ceasing directly a steady 
state of strain is set up. 

An inverse E.M.F., A D, will be excited by the same 
magnetic field directly the displacement is reversed. 



L 

D 



FIG. 51. 



And so, if a continual electric oscillation is kept up 
between A and B in a magnetic field, an accompanying 
very minute transverse oscillation may be expected 
and may be looked for electrically. 

Some such arrangement as that here shown (Fig. 52) 
may be employed : a square of heavy glass, per- 
forated with four holes towards the centre, supplied 



D 

FIG. 52 

with electrodes : one pair of electrodes, A, B, to be 
connected with the poles of some alternating machine, 
and the other pair, C, D, connected to a telephone or 
other detector of minute oscillatory disturbance. So 
soon as a strong steady magnetic field is applied, by 



278 RADIATION CHAP, xv 

placing the glass slab between the poles of a strong 
magnet, the telephone ought to be slightly affected 
by the transverse oscillations. This effect has not yet 
been experimentally observed, though at Liverpool I 
had a bored piece of glass ready to look for it, for some 
time, but it seems to me a certain consequence of the 
Faraday rotation of the plane of polarization of 
light. 

Hall Effect. 

177. Although the existence of this transverse 
E.M.F., excited by a magnetic field in substances 
undergoing varying electric displacement, has at 
present only been detected optically in transparent 
bodies, i.e. in insulators, yet in conductors the corre- 
sponding effect with a steady current has been dis- 
tinctly observed electrically. By many persons it had 
been looked for (by Prof. Carey Foster and the writer, 
among others, though unfortunately they were not 
sufficiently prepared for its extreme smallness) ; 
by Prof. E. H. Hall, then at Baltimore, was it first 
successfully observed. 

In conductors it is natural to use a conduction- 
current instead of a displacement-current. A steady 
current can be maintained in a square or cross of 
gold-leaf, or other thin sheet of metal, between the 
electrodes A, B ; and a minute transverse E.M.F. can be 
detected, causing a very weak steady current through 
a galvanometer connected to the terminals c, D, so soon 
as a strong magnetic field is applied perpendicularly 
to the plate. Fig. 53 will sufficiently indicate the 



177 ELECTRO-OPTIC EFFECTS 279 

arrangement. The poles of the magnet are one above 
and one below the paper. 

In iron it is easy to see which way the transverse 
E.M.F. ought to be found. It has been shown ( 175) 
that a displacement will be rotated in iron against the 
magnetizing current ; hence to rotate the displace- 
ment A B to A C (Fig. 51) requires in iron a clockwise 
magnetizing current. Such a current, or, what is the 
same thing, a south pole below the paper, a north pole 




FIG. 53. The direction of the transverse E.M.F. excited by the earth's vertical 
magnetic field in this conductor, conveying a current as shown, is c D if it 
represents gold, D c if it represents iron. 

above, excites, in the cross of Fig. 53, an E.M.F. in 
the direction D c ; and this by Ampere's rule is just the 
direction in which the conductor itself is urged by the 
magnetic forces acting on the current-conveying sub- 
stance. Most diamagnetic substances should exhibit a 
transverse E.M.F. in the opposite sense. This trans- 
verse E.M.F., excited in conductors conveying a 
current in a magnetic field, is the effect known by 
the name of Hall. It is, as Prof. Rowland and 
others have pointed out, intimately connected with 
the Faraday rotation of light (Cf. Fig. 51.) 



280 RADIATION CHAP, xv 

178. Unfortunately a pure and simple Hall effect is a 
difficult thing to observe. Magnetism affects the con- 
ductivity of metals in a rather complicated manner ; 
and strains affect their thermo-electric properties. 
Now a metal conveying a current in a magnetic field is 
certainly more or less strained by mechanical forces ; 
and hence heat will be developed unequally in different 
parts, by a sort of Peltier effect ; and the result of this 
will be to modify the resistance in patches, and so to 
produce a disturbance of the flow which may easily 
result partly in a transverse E.M.F. This has been 
pointed out by Mr. Shelford Bidwell. 

The more direct effect of magnetism on conductivity 
may be negligibly small in many metals, but in bis- 
muth it is certainly large. Both of these spurious 
effects seem to be large in bismuth, and probably 
quite mask any true Hall effect there may be in that 
metal. In all cases the existence of these spurious 
effects makes it difficult to be sure of the magnitude 
and sign of the real rotational effect. 

179. But, it may be asked, what right have we to 
distinguish between a real and a spurious Hall effect ? 
If a transverse E.M.F. can be predicted by reason of 
known strains and thermo-electric properties, as well 
as by known rotation of light effects, why should the 
two things be considered different? Why should they 
not be different modes of regarding one and the same 
phenomenon ? 

In other words, may not the Faraday rotation of 
light be due to infinitesimal temporary strains and 
heatings in the medium, caused by the fact that 
minute electric displacements are occurring in a 



I79A ELECTRO-OPTIC EFFECTS 281 

violent magnetic field ? May not the Hall effect also be 
accounted for by the action of magnetism on thermo- 
electric properties ? These are questions capable of 
being answered by a quantitative determination of 
the amounts and direction of the effects to be expected, 
and a comparison with those actually observed. I do 
not know of data, at present obtained, sufficient to 
enable us to answer it. If their answers should turn 
out in the affirmative, several apparently distinct 
phenomena will be linked together. 

In the Philosophical Magazine for May 1885, Mr. 
Hall gives some more measurements, showing that in 
bismuth the effect is enormous, and in the same direc- 
tion as in copper, whereas in antimony it is also great 
and in the same direction as in iron. All these things 
seem to point to some thermo-electric connexion. 



Effect of Magnetic Field on Resistance. 

I/9A. Another interesting effect of magnetism upon 
conductors, which demands more theoretical attention 
than it has yet received, is the change in conductivity 
which metals exhibit when subjected to a magnetic 
field across the current flow. An indication of this 
phenomenon was detected by Professor S. P. Thomp- 
son in 1883 during a plotting of equipotential lines in 
a metal sheet immersed in a magnetic field. The 
resistance of tin and gold appeared to decrease about a 
third and two-thirds of one per cent, and that of iron 
seemed to increase about two parts in ten thousand. 

These results were communicated as a paper to the 



282 RADIATION CHAP, xv 

Physical Society of London on March 22, 1884, but 
it was not published. 

In the same year, quite independently, Righi dis- 
covered a much larger effect in bismuth, as well as 
in other metals. In bismuth the effect is so remark- 
able that Lenard converted it into a method of quickly 
measuring the strength of a magnetic field, by observ- 
ing its effect on the resistance of a flat spiral of bismuth 
wire, inserted between the pole pieces. 

The resistance of bismuth, whether pure or not, 
increases when it is placed in a magnetic field, especially 
when the magnetic field crosses the line of current. In 
the latter case the increase of resistance can be as much 
as one-eighth of its value. 

Lenard found the resistance of a spiral of bismuth 
wire, placed normal to the magnetic lines of force, to 
increase from i, its normal value, to 1*32 for a field of 
8oooc.g.s., and r8 for a field of 16000 c.g.s. 

When a bismuth wire is parallel to the lines of force, 
the change is distinctly less ; and he considered that 
he observed some difference between the resistance to 
alternating and that to direct currents. 

Other Outstanding Problems. 

184. Outstanding problems bristle all over the sub- 
ject, and if I pick out any for special mention it will 
only be because I happen to have made some experi- 
ments in their direction myself, or otherwise have had 
my thoughts directed to them. 

Referring back to 66 at the end of Part II., a " cur- 
rent regarded as a moving charge," it is natural to ask, 



1 86 ELECTRO-OPTIC EFFECTS 283 

Is this motion to be absolute, or relative to the ether 
only, or must it be relative to the indicating magneto- 
meter ? In other words, if a charged body and a mag- 
netic needle are flying through space together, as, for 
instance, by reason of the orbital motion of the earth, 
will the needle experience any deflecting couple ? 

It is one of many problems connected with the 
ether and its motion near gross matter problems 
which the experiment of Fizeau (showing that a 
variable part of ether was bound with matter and 
transmitted with it, while another constant portion 
was free and blew through it) began to illumin- 
ate : aberration problems such as exercised the 
genius of Sir George G. Stokes ; problems connected 
with the motion of ether near great masses of matter, 
like those which Michelson so skilfully attacked experi- 
mentally ; it is among these that we must probably 
relegate the questions whether absolute or relative 
motion of electric charges is concerned in the production 
of magnetic field, and what absolute motion through the 
ether precisely means. (Phil. Mag. Oct. 1898.) 

1 86. Then, again, there is the influence of light on 
conductivity. Annealed selenium, and perhaps a few 
other things, improve in conductivity enormously 
when illuminated. The cause of this is unknown at 
present, and whether it is a general property of 
matter, possessed by metals and other bodies to a 
slight degree, is uncertain ; for the experiments of 
Bornstein, with an affirmative result for the case of 
metals, have been seriously criticised. 

Even though metals show no effect, yet electrolytes 
might possibly do so, but the effect, if any, is small ; 



284 RADIATION CHAP, xv 

and it is particularly difficult in their case to distinguish 
any direct radiation effect from the similar effect of 
mere absorbed radiation or heat. 

The writer has found that a glass test-tube kept 
immersed in boiling water conducted distinctly better 
when the blinds of a room were raised than when 
they were lowered, though nothing but diffuse daylight 
fell upon it. But as the effect could have been pro- 
duced by a rise in temperature of about the tenth of 
a degree, and as the absorption of diffuse daylight is 
competent to produce a rise of temperature as great 
as this in the glass of a thermometer-bulb, even though 
immersed in boiling water, he feels constrained to 
regard the result, though very clear and distinct, as 
after all a negative one, and has accordingly not 
published it. 

187. The fact that ultra-violet waves have a period 
of vibration synchronous with probable electric vibra- 
tion in molecules ( 157) seems to cause a multitude 
of effects now being discovered. Hertz noticed 
that the light of one spark influenced another at a 
distance, so that a sparking interval was virtually 
shortened when illuminated. Wiedemann and Ebert 
have further investigated this, and obtained several in- 
teresting results, distinctly proving that it is ultra-violet 
light which is effective. Hallwachs has discovered 
that a clean metallic plate can become electrified 
when light falls upon it. And there are a number of 
other similar facts, some long known, some recent, 
which all illustrate the molecular effects of light. It 
appears probable that they all depend on some 
synchronized disturbance set up in the constitution 



1 89 ELECTRIC WAVES 285 

of atoms or molecules. These physical effects appear 
to be of the same order as those other familiar but 
vaguely grasped facts summed up under the category 
of the chemical or actinic power of light. For that 
light affects silver salts, ebonite, hydrogen, and chlorine 
&c., is an old story. Some progress is now likely to 
be made in ascertaining the precise mode in which 
these changes occur ( 33) ; cf. Lecture VI. below. 

1 88. Before 1888 I should have put in a prominent 
position among outstanding problems the production 
of electric radiation of moderate wave-length ; and the 
performance, with this radiation, of all the ordinary 
optical experiments reflection, refraction, inter- 
ference, diffraction, polarization, magnetic rotation, 
and the like ( i). But a great part of this has now 
been done, and so these things come to be mentioned 
under a different heading : 

Electric Waves in Space. 

" Conclusion " seemed an absurd word to write in 
1889, when the whole subject was astir with life, and 
when every month seemed to bring out some fresh 
aspect, to develop more clearly some already glimpsed 
truth. The only proper conclusion to a book dealing 
with electricity at such a time was to herald the 
advent of the very latest discoveries, and to prepare 
the minds of readers for more. 

189. Referring back to Chap. XIV., to I and 8, 
and all Part IV., we spoke confidently of a radiation 
being excited by electric oscillations, a radiation which 
travelled at the same rate as light, which was reflected 



286 RADIATION CHAP, xv 

and refracted according to the same laws, and which, 
in fact, was identical with the radiation able to affect 
our retina, except in the one matter of wave-length. 
Such a radiation, the world now knows, was definitely 
obtained and examined by Dr. Hertz, at Karlsruhe, 
afterwards Professor at Bonn ; and in Decemer, 1888, 
Prof, von Helmholtz communicated to the Physical 
Society of Berlin an account of Dr. Hertz's researches. 

For electric waves along wires, observed by me in 
1888 likewise, see Chap. XIII. above, pp. 235-7. 

The step in advance which enabled Dr. Hertz 
to do easily that which others had long wished to do, 
was the invention of a suitable receiver. Light 
when it falls on a conductor excites first electric 
currents and then heat. The secondary thermal effect 
was what we had thought of looking for ; but Dr. 
Hertz boldly took the bull by the horns, looked 
for the direct electric effect, and found it manifesting 
itself in the beautifully simple form of microscopic 
sparks across a gap between two conductors, or 
between the ends of a looped conductor. 

He took a brass cylinder, some inch or two in 
diameter, and a foot or so long, divided into two 
halves with a small sparking 'interval between, and 
connected the halves to the terminals of a small induc- 
tion-coil ; every spark of the coil caused a charge in 
the cylinder to surge to and fro about five hundred 
million times a second, and to disturb the ether in a 
manner precisely equivalent to a diverging beam of 
plane-polarized light, with waves about thrice the 
length of the cylinder. 

The radiation, so emitted, can be reflected by plane 



1 89 ELECTRIC WAVES 287 

conducting surfaces, and can be concentrated by 
metallic parabolic mirrors ; the mirror ordinarily 
used being a large parabolic cylinder of sheet zinc, 
with the electric oscillator situate along its focal line. 
By this means the effect of the wave could be felt 
at a fair distance, the receiver consisting of a synchro- 
nized pair of straight conductors with a microscopic 
spark-gap between them, across which the secondary 
induced sparks were watched for. By using a second 
mirror, like the first, to catch the parallel rays and 
reconverge them to a focus, the effect could be ap- 
preciated at a distance of 20 yards. If the receiving 
mirror were rotated through a right angle, it lost its 
converging power on this particular light. 

Apertures in a series of interposed screens proved 
that the radiation travelled in straight lines (roughly 
speaking, of course). 

A gridiron of metallic wires is transparent to the 
waves when arranged with the length perpendicular 
to the electric oscillations, but it reflects them when 
rotated through a right angle ; so that the oscillations 
are proved to take place along the conducting wires. 
The wire grid thus represented a kind of analyser, and 
established the polarization of the radiation. The 
receiver itself also acted as analyser, for if rotated 
much it failed to feel the disturbance. 

Conducting sheets, even thin ones, were very 
opaque to the electrical radiation ; but non-con- 
ducting obstacles, even such as wood, interrupt it 
very little, and Dr. Hertz remarks, " not without 
wonder," that the door separating the room contain- 
ing the source of radiation from that containing 



288 RADIATION CHAP, xv 

the detecting receiver might be shut without inter- 
cepting the communication. The secondary sparks 
were still observed. 

But the most crucial test yet applied was that of 
refraction. A great prism of pitch was made, its 
faces more than a yard square, and its refracting 
angle about 30. This being interposed in the path 
of the electric rays, they were lost to the receiver 
until it was shifted considerably. Adjusting it till 
its sparks were again at a maximum, it was found 
that the rays had been bent by the pitch prism, when 
set symmetrically, some 22 out of their original 
course ; and hence that the pitch had an index of 
refraction, for these 2-foot waves, about 1*7. 

190. These are great experiments. When the first 
edition of this book was written, the latest of them 
were but a month or two old, and they were only 
a beginning. Most of the experiments were simple, 
and had already been repeated. 1 They seem likely 
to settle many doubtful points. There has been a 
long-standing controversy in optics, nearly as old 
as the century, as to whether the direction of the 
vibrations was in, or was perpendicular to, the plane 
of polarization ; in other words, whether it was the 
elasticity or the density of the ether which varied in 
dense media ; or, in the language of Maxwell's theory 
whether it was the magnetic or the electric disturb- 
ance that coincided with that plane. This point had 
indeed, by the exertion of extraordinary genius, been 

1 See Fitzgerald and Trouton, Nature, Vol. xxxix. p. 391 ; also 
Dr. Dragoumis, on vacuum-tube detectors, Nattwe, Vol. xxxix. p. 548. 
Also Lodge and Howard, who concentrated the radiation by a suitable 
lens ; Phil. Mag. July 1889. 



iQo ELECTRIC WAVES 289 

almost settled already, through the consideration 
of common optical experiments ; but now that we 
are able electrically to produce radiation with a 
full knowledge of what we are doing, of its direc- 
tions of vibration and all about it, the complete 
solution of this and of many another recondite 
optical problem may be expected during the next 
decade to drop simply and easily into our hands. 
197 (see page 304). 

We have now a real undulatory theory of light, no 
longer based on analogy with sound, and its inception 
and early development are among the most tremendous 
of the many achievements of the latter half of the 
nineteenth century. 

In 1865, Maxwell stated his theory of light. Before 
the close of 1888 it is utterly and completely verified. 
Its full development is only a question of time, and 
labour, and skill. The whole domain of Optics is now 
annexed to Electricity, which has thus become an 
imperial science. 



CHAPTER XVI 

RECENT PROGRESS (1892) 

191. THUS far in the first edition (1889). Since 
then much has been written concerning Hertz's ex- 
periments and the consequences deducible from them. 1 
A brief account of the essential principles involved 
may be here suitable. 

The discovery of Hertz, which made all the rest 
possible, was the fact that sparks could be excited in 
properly arranged conductors exposed to the electric 
radiation from an alternate current generator of 
sufficient rapidity. The oscillatory character of a 
Leyden jar discharge was referred to in 19, and 
again in 124; and the easiest way of displaying the 
spark-producing power of electric radiation is the 
plan I described in Nature of February 20, 1890, with 
the following diagram (Fig. 54). 

1 For instance, a set of three articles by Prof. Fitzgerald in Nature , 
vol. xliii. p. 536, and vol. xliv. pp. 12 and 31. Subsequently, in 1894, 
my own book "on the work of Hertz and his successors" appeared, 
as an expansion of a Royal Institution Lecture, and of a demonstration 
before the British Association at Oxford, showing how to signal by very 
small power to a fair distance, across space, using a Kelvin mirror- 
instrument as receiver. 



ELECTRIC RESONANCE 



291 



Two similar Leyden jars are connected up to cir- 
cuits of equal size ; but so that, while the circuit of A 
is interrupted by a spark gap, that of B is complete, 
the B circuit being conveniently adjustable in size, 
till it corresponds with the other, by means of a 
slider, S. 

If, now, A is charged and discharged in the ordinary 
way, the oscillations at every discharge disturb the 
circuit of the jar B, exciting in it similar but feebler 
electric oscillations. If the two jars and their circuits 
are properly synchronized, as can be done by shifting 



O 





FIG. 



54. A couple of Leyden jars arranged so that a discharge in one causes the 
other short-circuited one to overflow along the easy tinfoil path, c. 
(See Lodge in Nature, vol. 41, p. 368. See also 194.) 



the slider S to and fro, the oscillations in B maybe 
sufficiently violent to make it overflow. It cannot be 
expected to overflow its lip, unless this happens to be 
very much shallower than usual ; but a little tinfoil 
strip, c, pasted on the glass, and reaching over from the 
inner coating till it nearly touches the outer coating, 
provides an easy overflow path, of which the disturb- 
ance readily makes use. Accordingly, when the two 
circuits are arranged with their planes parallel, and 
their distance not more than two or three times their 
diameter, a bright little spark at the air-gap of the 

U 2 



292 RADIATION CHAP, xvi 

overflow path on the jar B can be seen at every dis- 
charge of the jar A. 

A slight motion of the slider either way usually 
suffices to throw the jars out of tune, and to stop the 
effect. 

192. The phenomenon is commonly spoken of as 
illustrating electric " resonance," which it certainly 
does very well; for the discharge of such ajar has many 
oscillations per spark, and accordingly there is oppor- 
tunity for " beats " and destruction of incipient effect, 
unless the tuning is pretty exact ; bu t the name' 1 reson- 
ance" is too suggestive of some acoustic reverberation 
phenomenon to be very expressive. It is at present 
commonly used to express the sympathetic response 
of similarly tuned or timed vibrators in general, but 
even when thus used in acoustics it hardly conveys 
this meaning, except by reason of habit. The 
essential thing to be connoted is the synchronizing 
of the vibration-period of two things, and this is 
well expressed by the adjective "syntonic," which 
was suggested to me for the purpose by the late 
Dr. A. T. Myers. That which has been styled 
resonance I propose, therefore, in future, to call 
" syntony." 

193. The alternations in the case of a circuit like this 
are what it is now customary to call moderately quick 
for instance, about a million per second. It is easy 
to get them either slower or quicker. To get them 
quicker the wire part of the circuit may be diminished 
in length,and the coatings of the jar may be diminished 
in size and removed to a greater distance from each 
other. 



i 9 4 HERTZ EMITTER 293 

The figures in the diagram below show the gradual 
transition from a Leyden jar to a Hertz oscillator as 
ordinarily made. The smaller anddumpier the arrange- 
ment the quicker will be its frequency of vibration. 
It is easy to make one to emit waves only a yard long, 
in which case its charge vibrates at the rate of three 
hundred million times a second. Quicker vibrators 
still can be obtained by using spheres. One sphere 
is sufficient, but sometimes several equal ones give a 

T 





FIG. 55. Showing the relationship between Hertz oscillator and Leyden jar. 

stronger effect ; 1 and I have succeeded in working 
with a sphere two inches in diameter, which vibrates 
twelve or thirteen times as rapidly as the one just 
mentioned above. But it is not feasible at present 
to produce recognisable vibrations of any such rate 
as a billion per second ; whereas to affect the retina 
they must be at least 400 billions per second. 2 

194. Failing the retina, the principle of syntony 
enables waves to be detected : the " electric eye " 



/, vol. xli. p. 462, March 1890. See also Fig. 64 below. 
2 A seconds pendulum beats a million times in twelve days ; or a thou- 
sand million times in thirty years. A tuning fork sounding a note two 
octaves above middle C (one thousand a second) in thirty years vibrates 
a billion times. Hence before such a fork can execute the number of 
vibrations which are achieved in one second by dull red light a period of 
twelve thousand years must elapse. 



294 RADIATION CHAP, xvi 

being some conductor with the same rate of vibration. 
It may be alike in all respects, but usually there is a 
difference between sender and receiver. For instance, 
in Fig. 54, the circuit of the sender has a spark-gap 




FIG. 56. 

in it, but the circuit of the receiver must be closed ; 
else the feeble beginnings of electric surgings, which 
are to be worked up by properly timed impulses, can- 
not occur. It is useless, therefore, to use as receiver a 
Leyden jar arranged like A, Fig. 54. The Leyden jar 




FIG. 57. 

B must have its circuit carefully completed, and a side 
disturbance or overflow must be utilized as indication 
of the induced effect. (See Fig. 54.) 

Some forms of Hertz oscillators, and methods for 
emitting from them a parallel beam, are shown in 
figures 56 to 58. 

The circle in Fig. 57 represents primarily a magnetic 



194 



HERTZ DETECTOR 



295 



receiver ; for it is best held so 
that the magnetic oscillation 
occurs along its axis, i.e. per- 
pendicular to its plane, and then 
sparklets or scintillae may occur 
across its microscopic air-gap. 

If the circle be turned so that 
the line joining its knobs is 
parallel to the electric force, this 
also can excite sparks ; for it 
may be then regarded as a 
branched conductor, with one 
branch interrupted, and therefore 
ineffective. But as an explorer 
of electric force a linear receiver 
is simpler and is preferable. I 
commonly used a pair of stiff 
wires, each half a wave-length 
long, supported end to end on a 
long bar of mahogany, with their 
near ends pointed, and one of 
them movable in the direction 
of its length by a screw operating 
on it through a glass tube. With 
the wood near the points black- 
ened and otherwise shaded from 
light, excessively minute sparks 
can be seen (Fig. 58). 

A bulkier but more demon- 
strative receiver is an arrange- 
ment exactly like the oscillator, 
as in Fig. 56, but with an entire 



I" 



-S 

eg 



Jj 



V* 

!! 
*i 

u 



296 



RADIATION 



CHAP. XVI 



metal rod connecting the plates instead of one inter- 
rupted by knobs. Oscillations begin and increase in 
this, under the operation of syntony or resonance, and 
the consequent sparking or overflow can be detected 
by holding a penknife or other point to either of 
the plates. 

A receiver exactly like an oscillator, with spark-gap 
in same position, really answers very fairly ; but it 

< 70 C. > 



Emitter. Side elevation. 



Receiver. 




Plan. 

FIG. 59. Small Hertz vibrater and receiver, arranged in focus of parabolic mirror for 
emission and reception of parallel beam. 

must be responding to a subordinate oscillation, 
chiefly one in each half of the emitter ; or else is 
not really syntonic. Precise tuning is never very 
important in these short-waved vibrators, because 
their rate of damping is so enormous. With greater 
capacity, as in the Leyden jars of 191, tuning is 
essential. 

195. Sometime as oscillator I used a simple sphere, 



'95 



ELECTRIC EYE 



297 



sparked to, on either side, by knobs connected with an 
induction coil ; and a precisely similar sphere acted as 
receiver, a penknife being held to it or lightly dragged 
over it at the end of the proper diameter. The 




FIG. 60. Lodge's Electric Retina- a set of copper cylinders 40 centims. long, and of 
diameters 14, 13, 12, n, 10, 9 centims. ; facing a source of radiation, as do the 
rods and cones in the eye, and responding by diametral electric vibration to the 
appropriate frequency. Nature^ vol. 41, p. 462 ; cf. also p. -258 above. 

arrangement was unexpectedly sensitive, and affords 
very short waves (Figs. 60, 61 and 64). 

At a distance, it is easier to detect the radiation 
from a large vibrator, i.e. one possessing an extensive 

!_ L 




FIG. 61. Robinson's electric harp, responding to vibrations through a considerable 
range of wave-length. Only a few of the tinfoil strips are shown. There are 
really fifty-one of them, closely packed together. Their length varies from about 
30 centims. to i centim. The scintillae occur at the razor-cuts marked S. The 
slant line of large gaps is simply to terminate the strips in. a graduated series. 

field, because of the greater initial energy electrostatic- 
ally stored before a discharge begins, and because 
of its sort of " sounding-board " action ; but there is 



2 9 8 



RADIATION 



CHAP. XVI 



no particular advantage in a great length of spark, 
for its resistance wastes energy. Half an inch or less 
is usually about the best length. 

The smaller the oscillator, the more polished must 
be the knobs, else the requisite suddenness of start- 




FIG. 62. Small oscillator, drawn to scale , used for experiments with the pitch lens. 
(Lodge and Howard, Phil. Mag.> July 1889. Rate of vibration, 300 million 
per second. 

ing will not be attained. Excessively sudden dis- 
turbances are necessary to excite waves in such rapidly 
vibrating things. Otherwise, as when a beer-barrel is 
carefully tilted, the potential has time to equalize itself 
leisurely, and no waves are caused. Ultra-violet 
light falling on either knob, especially on the cathode, 



-o o 




FIG. 63. Gigantic oscillator, drawn to scale \, for violent and distant effects. When 
this is excited by very large coil, most gas and water pipes in the building, and 
some wire fencing also, give off sparks to conductors brought close to them or to 
each other ; and a telephone joined to any conductors in the neighbourhood is 
affected. See Lodge, on "Signalling across Space without Wires" (Electrician 
Co.) Rate of vibration, 10 million per second. Average radiation activity, while 
it lasts, 64 horse-power. 

is to be avoided, for it enables the spark to occur 
too easily, and apparently with a less sudden break- 
down. Anything of the nature of an electric brush 
or glow is especially deadly ; and that is one reason, I 
think, why a Voss or Wimshurst machine is usually 
less effective than a coil, even if it be able to give as 



196 



SPHERICAL RADIATOR 



299 



many sparks per second. The quality of the spark at 
the oscillator is very important, and is exceedingly 
capricious. Some days it declines to emit any reason- 
able amount of radiation, and on other days the 
receiver responds at surprising distances. Practice 
however, tends to reduce these caprices, and there is 




FIG. 64. Arrangement for triple-sourced radiation from six-inch spheres Suitable for 
exciting the electric retina, Fig. 60. (Nature, vol. 41, page 462.) 

seldom any difficulty in working with a good-sized 
oscillator, such as one with foot-disks a yard apart and 
connected by a rod and spark gap. 



Process of Radiation. 

196. Consider now in more detail what is happening. 
To and fro round the circuit of the jar A (Fig. 54) an 
electric current alternates, charging up the coatings 



300 RADIATION CHAP, xvi 

with alternately plus and minus charges. As a result 
of this electric current, a magnetic oscillation is going 
on along the axis of the circuit, perpendicular to the 
plane of the diagram ; and this it is which may be 
thought of as disturbing the ether at a distance. It 
is, in fact, primarily a magnetic oscillator, and it 
excites induced currents in the ordinary way. But it 
does more than that; for it throws off true waves, 
which do not return to it, but convey its energy out 
into space. 

A linear vibrator, as No. 3 in Fig. 55, may be called 
primarily an electric oscillator ; for the electric surg- 
ings, from one end to the other and back, are the pro- 
minent feature. Its ends are alternately positive and 
negative at the end of each half-swing ; and thence 
arises the electric strain which is felt by the surround- 
ing ether. In mid-swing there is no electrostatic 
strain, but there is a linear current which exerts 
magnetic influence all round itself, causing circular 
lines of magnetic force surrounding the conductor. 

Ring vibrators, such as No. I, Fig. 55, or such 
as Fig. 57, only cause electric disturbance as a 
secondary consequence of the magnetic oscillations 
along their axis ; which is their primary effect. Linear 
oscillators excite primarily an electric oscillation. But 
in neither case do the two occur simultaneously ; one 
lags a quarter-period behind the other in time, just as 
the maximum velocity and maximum excursion of a 
simple pendulum differ by a quarter-period. 

This lag is, however, strange to say, very soon made 
up or compensated by reason of a start in space. 
The magnetic disturbance being considered to start 



1 96 RADIATION THEORY 301 

at the oscillator, the electric starts a quarter of a wave- 
length in advance ; and accordingly, though it occurs a 
quarter-period later, the two travel together every- 
where except within the first quarter wave-length. 
They are not, indeed, coincident^ because they occur 
in planes at right angles to each other, but their 
maxima and zeros agree in position. If one is 
likened to a series of crests and troughs, the other is 
that same series rotated through a right angle. 

But within the first quarter wave-length all this is 
different. The electric and magnetic forces there 
sometimes agree in phase and sometimes differ, one 
being occasionally even opposite in sign to the other. 

This fact has important though not obvious conse- 
quences. In a varying electro-magnetic field, energy 
flows, as Poynting showed, in a direction determined 
by the direction of the electric and magnetic forces ; 
it flows at right angles to both of them, and is 
reversed in direction if either of them is reversed 
in sign. Now, within the first quarter wave-length 
from an oscillator, the electric and magnetic forces 
sometimes have the same sign and sometimes opposite 
sign ; hence the energy there oscillates to and fro ; 
more, indeed, travels outward than returns, but 
some returns, and assists in maintaining the next 
oscillation. 1 

Beyond the first quarter wave-length, however, the 
magnetic and electric forces wax and wane together ; 
neither is ever reversed without the other, and conse- 
quently the flux of energy there is steadily outward : 

1 Hertz's own treatment of the theory of these actions, with full 
diagrams, will be found translated by me in Nature, vol. xxxix. p. 451. 



RADIATION 



CHAP. XVI 



no more returns, but all travels out into space as 
radiation. 

Whether it be much or little that thus travels out 
is a question of wave-length. With long waves the 




\\ 




* \ 



V 



7 




_MA6NE 
_ELEC' 



'1C FORCE 
)IC FORCE 



FIG. 65. Diagram of the electric and magnetic forces concerned in radiation. 
(Drawn by Mr. Tronton, see Nature, vol. 42, p. 172.) 

point of departure, or origin of radiation, is a long way 
from the vibrator, and accordingly the disturbance 
there is but slight. Most of the energy in that case 
returns to the oscillator, whose rate of damping will 



i 9 7 RADIATION THEORY 303 

be controlled only by resistance and the production 
of local heat. Thus a commercial alternator, with 
frequency 100 per second, emits waves 3,000 kilo- 
metres (about 2,000 miles) long. Consequently, the 
radius where its energy begins to radiate away is 
something like the distance of the Shetlands from 
London ; and accordingly no appreciable loss of 
energy is due to uncompensated emission of waves. 
But a small Hertz vibrator, whose wave-length is a 
few feet, radiates powerfully far more powerfully than 
sunshine, for instance, while it lasts ; and the rapid 
dying out of its vibrations is due almost entirely to 
this cause. 

It is to be observed that the effective source of 
the emitted radiation is not at the oscillator itself, 
but at a quarter wave-length in advance of it. If it 
be assumed to start at the oscillator, the light will 
seem to travel too quickly ; and this Hertz found to be 
the case. It travels with a quarter wave-length start, 
and so its phase at any distance is a quarter-period 
more than would have been expected. 

The fact that there was this sort of quarter-wave 
acceleration of phase in ordinary light was discovered 
mathematically by Sir George Stokes, 1 and had 
become an accepted part of optics ; the reason of 
it, in Maxwell's electric theory, is now clear. 

Direction of Vibration. 

197. The electric disturbance emitted by a linear 
oscillator, like that in Fig. 62, occurs parallel to the 

1 Dynamical Theory of Diffraction, 1849 ; Stoke? s Paper s> vol. ii. 
pp. 280, 321. 



304 RADIATION CHAP, xvi 

axis of the oscillator ; the magnetic force is at right 
angles to it ; and the direction of flow of energy, or ray, 
is out from the oscillator, perpendicular to both the 
electric and magnetic disturbances. Since the direc- 
tion of vibration is thus definite, the beam is what is 
spoken of in optics as " polarized " ; though whether 
the polarization retains its precision unimpaired to 
great distances may be uncertain. 

An important experiment, on radiation thus obtained, 
was made by Mr. Trouton, in Prof. Fitzgerald's labora- 
tory, Dublin, 1 to ascertain which is the plane of polari- 
zation, through the interposition of a non-conducting, 
i.e. transparent, obstacle in the beam, at appropriate 
obliquity, tan" 1 p. It is found that if, under these 
circumstances, the plane of incidence is arranged to 
coincide with the direction of the electric disturbance, 
none of the radiation is reflected ; whereas, if the 
magnetic disturbance is in the plane of incidence, a 
portion of the radiation is reflected. Now, Fresnel- 
who was, of course, unacquainted with the fact that 
there are two distinct kinds of vibration to be 
attended to, and who treated light as if it were a 
mechanical disturbance in an elastic solid surmised 
that the direction of vibration of light, reflected at the 
polarizing angle from a transparent substance, was 
perpendicular to the plane of incidence at which it 
was reflected. Hence we see that the disturbance 
contemplated by Fresnel corresponds to the electric 
disturbance ; whereas the vibration contemplated in 

1 See lecture by Prof. Fitzgerald at the Royal Institution, March 21, 
1890, reported in Nature, vol. xlii. p. 172. See also Mr. Trouton in 
Nature, vol. xxxix. p. 391. 



i 9 7 DIRECTION OF VIBRATION 305 

the opposition theory advanced by McCullagh cor- 
responds to the magnetic disturbance. 

Thus Maxwell's theory that the magnetic force is 
in the plane of polarization is directly verified. 

In most cases, as in the action of light on chemicals, 
diffraction by small particles, and so on, the electric 
oscillation is the more effective of the two ; and it 
turns out to be this which has been properly treated 
as the direction of vibration by Stokes and others, 
from the point of view of the elastic solid theory. 

When trying the experiment on reflection at the 
polarizing angle, it was found that a thin plate, such 
as a slab of glass, would not give any reflection ; the 
reason being that its two surfaces are so close together 
that reflection from one is masked by interfering reflec- 
tion from the other ; just as happens in the black spot 
of Newton's rings. But by using as reflector a wall 
3 feet thick the surfaces were farther separated ; and, 
the transparency being imperfect, reflection from the 
back surface was also much weakened, so that un- 
interrupted reflection from the front surface could now 
be dealt with ; and this it was which varied with the 
angle of incidence and with the direction of vibration 
relatively to the plane of incidence. 

Mr. Trouton has tried a number of other inter- 
ference experiments, 1 imitating Newton's rings, &c., 
and has obtained some diffraction phenomena when 
a reflector was smaller than a wave-length in size : 
obtaining results analogous to Stokes's " experi- 
mentum crucis " with respect to direction of vibration. 
The readiest way in which I obtained interference 

1 See Nature, vol. xl. p. 398 ; also Phil. Mag. July 1891. 

X 



306 RADIATION CHAP, xvi 

effects was when using a spherical oscillator, a metal 
sphere about 6 inches in diameter supported by a 
glass rod 2 feet high on a teak table : one of those in 
Fig. 64. The receiver being a precisely similar sphere, 
it was found to respond at certain distances along the 
table and not at others ; and moving it steadily and 
gradually towards the oscillator, from a distance of 12 
feet, a succession of silences, or at least minima, were 
observed. On measuring the positions of these, they 
were found to correspond well with a difference in path 
amounting to one, three, and five half wave-lengths 
between the source and its image in the table. 

The radiation from a sphere seems to me to be a 
purer tone, more monochromatic, than the radiation 
from a dumbbell-shaped conductor with a spark-gap 
in the middle of it. I prefer, as a rule, to have the 
spark-gap beyond the oscillator, instead of in the 
middle of it, and to depend on oscillations excited 
in a regularly-shaped conducting body when thus 
suddenly supplied with an electric charge. 

Other Receivers or Detectors of Radiation. 

198 Since Hertz's discovery that little sparks could 
be excited in a conductor exposed to electric radiation, 
various other methods of detecting radiation have 
been devised. Dr. Dragoumis, working in my 
laboratory, has used vacuum tubes, and has shown 
that they glow in the oscillating field near an 
electric vibrator without being attached to wires or 
any form of conductor ; l also that they furnish 

1 Nature, vol. xxxix. p. 548. 



198 VARIOUS RECEIVERS 307 

a fairly convenient means of displaying Hertz's 
experiment to a small audience when attached to an 
ordinary syntonized receiver. They are not bright 
enough to be visible far away, nor except in a dark 
room ; but when they contain some of Sir William 
Crookes's phosphorescent powders they are very fairly 
bright. The brightest I possess is one sent me in 1889 
by Dr. Lenard, from Heidelberg, containing strontic 
sulphide rendered impure with calcic fluoride and a 
trace of copper, which last impurity seems necessary 
to the brilliant glow. 1 I estimate the light 'when its 
terminals are attached to the terminals of a small coil 
as about half a candle, and I have been accustomed 
to exhibit it to my class as 'possibly the embryo of 
the light of trie future. 

Sir W. Crookes long worked at the subject of 
phosphorescence with this idea in view; and Mr. 
Nikola Tesla constructed alternating dynamos of 
extraordinarily high frequency, which may be con- 
sidered as a means of maintaining electric oscillations, 
and thereby making vacuum tubes and other bodies 
glow with considerable brightness, when connected 
with, or even when disconnected from, the source. The 
vapour of mercury has been shown by Cooper Hewitt 
to be very bright when conveying a steady current. 

The oscillations maintained by Tesla's dynamos 
must at present be considered " slow," for they do 
not, I believe, rise to a hundred thousand per second ; 
but they are vastly quicker than any which have been 
steadily maintained before. The Leyden jar and Hertz 

1 Wied. Ann., xxxviii. p. 90 ; abstract in Phil. Mag. November 
1889, p. 428. 

X 2 



3o8 RADIATION CHAP, xvi 

vibrations, of millions or thousands of millions per 
second, do not last an appreciable instant of time. 
Three or four, or a few dozen, vibrations, and they 
are done. The intervals for refreshment bear an 
appalling proportion to the moments of activity, even 
when hundreds of sparks are excited per second ; hence 
it is that any illumination or other effect that they 
can produce is so feeble. To maintain them at their 
maximum intensity would indeed demand a good 
many horse-power ; but the effects then obtainable 
would be prodigious. 1 

Another mode of detecting electric radiation is by 
its heating effects. The late Mr. Gregory, of Coopers 
Hill, constructed a very delicate arrangement of a 
fine wire attached to a spiral shaving and mirror, 
whereby extremely minute changes of length were 
translated into perceptible rotations. This receiver, 
held in an oscillating field with its length parallel to 
the direction of the electric disturbance, was heated 
by the electric currents induced in it, and produced 
a measurable deflection of a beam of light. Mr. 
Gregory's hope was with such an instrument to make 
a series of metrical determinations in the neighbour- 
hood of a Hertz oscillator ; but the difficulty 
of obtaining consistent results lay not in the 
receiver, which seemed well enough adapted for the 
purpose, but in the capricious behaviour of the spark 
at the emitter. If every spark could be depended on 
as being like every other, many difficulties and distrac- 

1 The power of tropical sunshine (by which I mean sunshine at the 
earth's distance, but unaffected by the British climate, natural or artificial) 
is about 2 horse-power per square yard. 



198 BEGINNINGS OF COHERER 309 

tions,at that time inseparable from the experimental in- 
vestigation of electric radiation, would cease. Recently 
by employing a series of very short air-gaps this 
difficulty has been largely overcome. 

Another thermal method has been tried by 
several experimenters, viz. the interposition of a 
thermal junction in, or juxtaposition of it to, the thin 
wire receiver. This is heated by the induced currents, 
and can be connected with a galvanometer. 

For making the effects merely visible at a distance, 
a method was discovered by Prof. Fitzgerald, who 
found that, if the two halves of an ordinary Hertz 
receiver were connected with a delicate fine wire 
galvanometer, its needle was disturbed whenever the 
little scintillse occurred. A telephone does still better. 

Similarly, Mr. Blyth, of Glasgow, has used a 
quadrant electrometer in the same sort of way; while 
Prof. Boltzmann made the little spark effect moment- 
ary contact between a charged jar and an electroscope, 
and thus display itself. 

The effect observed by Fitzgerald depends on the 
induced electrostatic charging of the plates of the 
receiver connected through the galvanometer wire. 
If there is no spark, their charge and discharge 
quantities both pass through the galvanometer, and 
produce no effect ; but, if they spark, the charge alone 
passes through it, and causes a minute effect. 

I found something apparently of the same sort in 
some experiments on lightning guards. When the 
knobs of the guard were exceedingly close together, 
and were used to protect a galvanometer, that 
galvanometer was liable to capricious disturbance 



310 RADIATION CHAP, xvi 

whenever sparks passed across the microscopic air- 
gap. If the air-gap was too big, there was nothing ; 
and if it was too small, i.e. zero, there was nothing ; but 
when just short of zero there was a very pronounced 
effect, and sometimes the jaws of the gap were found 
afterwards to be feebly at a single point cohering. [See 
Journ. Ins t. E.E. April, 1890, vol. xix. pp. 352, 4], 

Coherer. 

199 This (published as above in 1892) was the 
beginning of what soon developed into the " coherer," 
which was the name I gave to the minute gap between 
a pair of metals in ostensible but incomplete contact. 
The slightest electric splash or scintilla across the 
gap caused cohesion, thereby establishing electric 
connexion, and enabling a battery current to pass and 
give any desired signal. The cohesion was easily 
broken again by a slight tap, which made it ready for 
a fresh signal. Already M. Branly, of Paris, had 
observed a similar effect with porphyrized copper and 
with metal filings ; so, before long, in my laboratory, 
sometimes a coherer and sometimes a Branly filings 
tube was employed as the detector of electric wave 
radiation, as described in my book, published in 1894, 
called The Work of Hertz and his Successors \pr 
Signalling across Space without Wires'] ; which was 
intended as a memorial to the great physicist Heinrich 
Hertz : these later developments being regarded as a 
legitimate and natural outcome of his work. 

Since that time all these modes of detection, 
whether in the form of a single specially arranged 
contact-gap, or the form of a multiplicity of loose con- 



200 COHERER 311 

tacts, have been called " coherers " indiscriminately ; 
and in 1896 Signer Marconi came to this country for 
the purpose of applying the device to official and 
commercial wireless telegraphy, which he has done 
with conspicuous success. 

Luminescence. 

200 The fact that the light of phosphorescence 
such light as that of the glow-worm, for instance is far 
more economical than any other known kind of lumin- 
escence (in other words, consists wholly of the kinds 
able to affect the retina, and of little else, as mentioned 
in 151) has now been directly ascertained by Prof, 
Langley, of the Smithsonian Institute, Washington. 1 

He has examined the spectrum of a firefly with his 
bolometer, and found that the whole intensity of radia- 
tion is concentrated in its visible portion ; while of 
infra-red or ultra-violet rays there is hardly a trace. 
It is, perhaps, rather singular that the eyes of fire- 
flies to which this light is of course intended to 
appeal should thus prove to respond to just the 
same wave-lengths as do ours. Without this fact it 
would have been rash to assert that other animals 
might not often be emitting radiation, unperceived 
by us, but visible enough to them. The observation 
of Prof. Langley, although it does not of course 
negative this possibility, yet seems to me to render it 
improbable, and to suggest that the molecular struc- 
tures adapted to respond to light in animal tissues are 
fairly definite, and not susceptible of variety. 

1 Langley and Very, " On the Cheapest Form of Light, from Studies 
at the Allegheny Observatory," Phil, Mag. September 1890, p. 278. 



312 RADIATION CHAP, xvi 

The structures susceptible to light in vegetable 
economy also respond to much the same range of 
vibration ; for plants are injured by ultra-violet light 
of too high an order, such as is given by the electric 
arc, and I believe that chlorophyll is not known to 
be affected by infra-red rays. The range of vision 
of algae and other subaqueous vegetation has not, I 
think, yet been ascertained. 



Opacity. 

201 The possibility of experimenting with radiations 
of great variety of wave-length throws a new meaning 
on the subject of opacity. It is well known that thin 
layers of anything, even gold if thin enough, will 
transmit some light, but that in sufficient thickness 
most things become opaque. 

Opacity, however, on Maxwell's theory, ought to 
depend, not only on the nature of the medium, but 
also on the frequency of vibration of the light which 
is trying to penetrate it. For the quenching takes 
place after a certain number of vibrations, and not 
instantaneously. Suppose a substance so opaque as 
to destroy or reflect the greater part of an incident 
beam in the course of three vibrations, it must de- 
pend entirely on the wave-length of those vibrations 
how great a thickness is needed to be effectively 
opaque. If the waves are each the hundred - 
thousandth of an inch in length, like ordinary light, a 
mere film is sufficient. If the waves are each a foot 
long, then a wall a yard thick would be necessary. 



201 OPACITY 313 

And if the wave-lengths are reckoned in miles, no 
practicable thickness will stop them. 

So it is experimentally found. 

Hertz waves can get through deal doors and stone 
walls not, of course, without some loss, but they 
are stopped dead by a copper plate, or even by tinfoil. 
I am unable to detect them on the other side of 
paper covered with Dutch metal, when they are 
emitted by a small sphere ; but a coating of 
plumbago lets them partially through. Very long 
waves, such as are emitted by an electro-magnet fed 
with an ordinary alternating current, can get through 
not only walls and buildings, but through sheet 
copper ; though again not without some loss. And 
if the copper be thick enough, not very much 
will get through. A perfect conductor, if such a 
thing existed, would reflect and stop everything, how- 
ever thin it might be. Still slower waves, such as are 
generated by slowly waving a magnet about, can get 
through copper of any reasonable thickness, and 
affect compass needles on the other side. But the 
more rapidly the magnet moves, the more will its 
effect be screened ; and the best way of showing the 
experiment is not to move the magnet, but the 
copper. Lord Kelvin has inclosed a fixed magnet 
in a thick copper box, and shown that when the box 
is revolved a compass needle outside is much less 
deflected then when it is stationary. 1 

M. Pictet at Berlin ascertained that screens are 
singularly transparent to the "heat" radiation emitted 
by bodies 200 Centigrade degrees below zero. 

1 Proc. Roy. Soc., vol. xlix. p. 422. 



314 RADIATION CHAP, xvi 

Thus there are all possible gradations of opacity, 
depending on the frequency of the incident vibration. 
To very long waves almost everything is transparent. 
To very short waves almost everything is opaque. 1 

Pressure of Light. 

202 Maxwell predicted that light falling upon a re- 
flecting metallic surface must repel it or press upon it 
with a minute but calculable force. Thus tropical 
sunshine falling normally on a square centimetre of 
silver must press on it with about the ten-thousandth 
part of a dyne : that being about the force needed to 
reverse the tide of energy. The pressure of sunlight 
on the whole earth amounts to about 100,000 tons 
weight. On a small enough dust particle in a comet's 
tail the solar radiation could overpower the solar 
gravitative attraction and drive the tail away from 
the sun. 2 

This pressure, though looked for by Sir William 
Crookes, had not in 1892 been detected with 
ordinary light, because it is altogether masked by the 
far greater effect of the air molecules, which cannot 
be got rid of, however good the so-called vacuum be 
made. Sir William Crookes, indeed, was fortunate 
enough to discover this molecular effect, and to 
invent the radiometer. If the direct light effect is 
ever to be seen, it must either be looked for with a 

1 For a full discussion of "opacity," see Lodge, Presidential address 
to Physical Society, 1899 (Phil. Mag. April 1899). 

2 See note by the writer at Cardiff, 1891, reported briefly in Nature, 
vol. xliv. p. 454. I find that Prof. Fitzgerald had discussed the same 
thing before (Proc. Roy. Dub. Soc. May 1882). 



202 PRESSURE OF LIGHT 315 

very bright surface ; or else with some black surface 
not beaten back by rebounding molecules. 

But there is no real need to hesitate about 
the reality of the direct light effect, because, by 
using longer waves, it becomes very conspicuous. 

It is, indeed, essentially nothing else than the fact 
discovered by Faraday as to the odd behaviour of 
metallic masses near the pole of a magnet : the 
effect described and illustrated in Fig. 28, p. 151. 

A copper disk held near a magnet fed by an 
alternating current is repelled by reason of the 
currents induced in it, and by their reaction on the 
field. That is precisely why light exerts pressure 
on a conducting surface. It induces currents in it 
and then repels them because it disagrees with them 
in phase. The repulsion of a disk near an alternat- 
ing magnet is not, indeed, a steady or one-directioned 
effect ; there may be moments of attraction, but on 
the whole there is distinct balance of repulsion, and 
repulsion alone is observed if the alternations are 
rapid and sinuously regular. 

The strength of the alternating currents now 
readily available has enabled Prof. Elihu Thomson 
to throw this kind of repulsion into various striking 
forms, making it support rings, rotate spheres, and 
exert forces expressible in pounds weight. 

Added 1906. Recently (1904) Messrs. Nicholls and 
Hull in America have succeeded in verifying the exis- 
tence of Maxwell's light-pressure, and in measuring its 
amount They avoided the radiometer-like effect 
by adjusting the vacuum so as to give as nearly as 
possible zero molecular disturbance ; since the effect 



316 RADIATION CHAP, xvi 

of residual air changes sign at a certain stage of 
exhaustion. They thus were able to show that light 
exerted a normal pressure on absorbing or reflecting 
surfaces, of amount equal to that predicted by 
Maxwell. 

Prof. Poynting of Birmingham has now gone 
further, and perceived that there must be a tangential 
force, or surface drag, when light impinges on an 
absorbing surface obliquely ; and that this is easier to 
observe than the normal pressure, since it is less 
perturbed by other causes. For a reflecting surface 
it is zero, because the direct and reflected beams act 
oppositely in a tangential direction, while their 
normal pressures are added. By suspending a couple 
of small surfaces on a quartz fibre, in such a way that 
the normal forces neutralize each other while the 
tangential forces co-operate, Foynting has measured 
this tangential or grazing force of light, and shown 
that its order of magnitude also corresponds with 
theory (Phil. Trans. 1904, vol. ccii. page 537). 

Future Experiments. 

203 Experiments on optical " dispersion," with a 
view of understanding the mechanism of the process, 
and affording information concerning the interaction of 
material atoms and the ether in which they are em- 
bedded, have not at present, so far as I know, been 
made. But Fitzgerald and Trouton have suggested 
filling pitch or paraffin with bullets, or some such plan, 
and studying the behaviour of a prism of it to 
Hertz waves of assorted wave-lengths not enormously 



203 MANUFACTURE OF LIGHT 317 

greater than the size of the metallic balls embedded 
in the transparent substance. 

They have also many instructive ideas on the sub- 
ject of maintaining an electrical vibration of a high 
degree of frequency, 1 without those disastrous intervals 
of darkness, or silence, or rest, which at present exist 
between one spark and the next : intervals which are 
to the time of one vibration as a year is to a minute. 
You cannot expect to make much impression on the 
universe by a performance, however brilliant, of five 
minutes' duration repeated once a year. 

Progress towards the direct manufacture of light, 
suggested in 149-151, is thus apparently going 
on in two different directions. One is by the at- 
tempt to construct mechanism, or discover principles, 
able to cause and maintain electric oscillations of the 
desired frequency ; the other is to depend on the 
properties of certain molecules disturbed and thrown 
into vibration by comparatively slow electric oscilla- 
tions and thus caused to phosphoresce. 

This last may seem a short cut to the desired 
result, but it has rather too strong a family likeness 
to the present blindfold method of coaxing molecules 
to radiate, viz. by jogging them with heat, to be quite 
satisfactory. The direct maintenance method may lie 
dormant longer, and be farther from the achievement 
of success, but it seems to have a deeper comprehension 
of actual conditions latent within it. 

This latter method fails at present from insufficient 
frequency ; the phosphorescence method fails at pre- 
sent from insufficient intensity. 

1 e.g. Abstract in Natzire, vol. xlv. p. 358. 



CHAPTER XVII 

MOST RECENT VIEWS CONCERNING THE 
ETHER (1907) 

204 HITHERTO, though we have constantly re- 
ferred to the two ethereal constants /JL and K, we have 
spoken as if there were no clue to their values ; but 
quite recently 1 have come to the conclusion that facts 
now known enable us to make a reasonable and plausi- 
ble estimate of the orders of magnitudes of these im- 
portant quantities. This matter I will now explain, and 
incidentally will endeavour to indicate some of the 
steps whereby it is hoped ultimately to show how the 
whole of the material universe is conceivably explic- 
able on the postulate of a continuous incompressible 
perfect fluid, thoughout space, possessing only the two 
fundamental attributes (a) inertia, and (b) intrinsic 
rotational kinetic energy, the latter involving two 
related but opposite kinds of motion. Such an. indi- 
cation is perforce unsatisfactory and incomplete at 
present, and the chief feature of the present chapter 
must be an estimate of the orders of magnitude for 
the inertia and for the intrinsic energy per unit 
volume which will be needed in the explanation. 



205 STRUCTURE OF ETHER 319 

For the ideas of some eminent workers in this field 
reference may be made to the Phil. Mag. for April 
and June, 1907. 

In conjunction with this chapter a perusal of the 
appended article, called " Lecture 5, " below, " On the 
Interstellar Ether," is recommended ; and the author 
feels it not improbable that some vital or psychical 
consequences of considerable importance such, 
perhaps, as may have been intended by the authors of 
" The Unseen Universe " may sooner or later flow 
from the immense ethereal energy and density which 
physicists are now challenged to contemplate or 
controvert. (See also Appendix s). 

Structure of the Ether. 

205 What, then, is the conclusion of the whole 
matter, so far as a conclusion is possible at present? 

The material universe seems to consist of a perfectly 
continuous incompressible and inextensible medium, 
filling all space without interstices or breach of 
continuity ; not of a molecular or discrete structure, 
and as a whole completely at rest : as frictionless 
moreover, and unresisting to all ordinary motion of 
what we call matter through it, as is the mathematical 
conception a perfect fluid. But in spite of immobility 
as a whole, it possesses that property of " rigidity," or 
elastic resilience to " shear," which is characteristic of 
what we ordinarily call a solid ; wherefore it would 
appear that it must be, throughout, in such a state of 
excessively fine-grained turbulent motion as would 
confer this property upon it. And the resilience is so 
complete and instantaneous, without any delay or 



320 VIEW OF THE ETHER (1907) CH. xvn 

permanent set, that the elasticity must be described as 
" perfect." It is the gyrostatic kind of elasticity, dis- 
covered dynamically and applied ethereally by Lord 
Kelvin, whereby a perfect fluid can kinetically acquire 
some of the properties of a perfect solid. (See 156, 
Chap. XIV. p. 256, above.) 

It is well known that every solid possesses two 
kinds of elasticity elasticity of bulk and elasticity of 
shape. The first or volume elasticity may also be 
called " the incompressibility," and is common to all 
forms of matter fluid as well as solid. In the case of 
the ether, however, the value of this quantity appears 
to be infinite : it is, at any rate, greater than we have 
as yet been able- to appreciate by specially directed 
experiments, meaning especially the Cavendish 
experiment referred to in 4 and I4A. The 
elasticity of figure, or shape-elasticity, is only possessed 
by solids, and is technically called " rigidity " ; it is 
small in the case of india-rubber, great in the case of 
steel or glass ; it is the property on which spiral 
springs and torsion-balances depend. The two kinds 
of elasticity are quite independent of each other 
quite independent also of anything akin to viscosity, 
which in the case of the ether appears to be zero. 

Now something analogous to shape-elasticity the 
ether possesses. It does not possess ordinary mecha- 
nical rigidity, because that is an affair of molecules ; but 
it possesses something which may be called an electric 
rigidity, or electromotive elasticity. It is identical 
with the electromotive elasticity of a dielectric, it is 
the property which causes recoil after charge ; and it 
has been denoted by 47r/tf, where K is the absolute 



205 PROPERTIES OF ETHER 321 

Faraday's dielectric constant, or specific inductive 
capacity, for free space. 

The property thus analogous to rigidity, or shape- 
elasticity, is accompanied by another property, akin 
to inertia. This is the property to which magnetism 
is due ; it is a magnetic inertia, to pair with electric 
rigidity, and it has been denoted throughout by 4?r/^, 
where ^ is the absolute magnetic permeability of free 
space. The self-induction or quasi-inertia associated 
with every electric current, of which 4.77 p, is the non- 
geometrical and essential factors explicable, up to a 
point, as due to the magnetic field excited by electric 
motion ; but it would seem as if ultimately it must 
necessarily be dependent on an unexplained and 
fundamental kind of inertia possessed by the ether 
itself; so that the ether may be said to have a certain 
density, or mass per unit volume, a massiveness so 
like ordinary material specific-gravity or density that 
we have to call it by the same name. 

By reason of these two properties electric elasticity 
and magnetic density transverse electromagnetic 
waves are transmitted through and by the ether, at a per- 
fectly definite and known speed. This speed of wave 
propagation is far greater than any we are accustomed 
to in connexion with matter ; and if ever the motion 
of matter can be made to approach this speed, it must 
encounter a reaction, or impedance, or opposition to 
further acceleration, which ultimately, in the limit, 
amounts to a practically infinite obstruction at the 
actual critical speed. 

This obstruction is not of the nature of friction, it 
is not resistance proportional to the velocity, or in any 

Y 



322 VIEW OF THE ETHER (1907) CH. xvn 

way dependent on the velocity : it solely opposes 
acceleration, and is of the nature of impedance or 
inertia. 

The fact of inertia enables an oscillatory wave- 
process to go on in the ether, and endows those 
oscillations with a particular kind of alternating kinetic, 
as well as with potential, energy. 

The energy of strained or distorted ether is always 
potential energy, and is all the potential energy there 
is ; but accessible or convertible kinetic energy is 
usually only possessed by those individualised and 
discriminated regions, or ethereal structures, which 
possess the power of locomotion, and which in their' 
aggregate appeal to our senses as " matter." 

During the passage of waves, the ethereal medium 
is sheared to and fro ; not with any movement as a 
whole, but with equal opposite movement of two 
aspects, or elements, or conditions, of its structure : 
such shear being equivalent to what is called an electric 
displacement, and being subject to a restoring force 
accurately proportional to that displacement. 

This elasticity is " perfect " in free space, apart from 
matter ; until a critical shear, of unknown value, is 
reached. If strained beyond that, it maybe supposed 
that a separation, or dislocation, or decomposition, of 
the ether into two components or constituents would 
occur ; constituents generated, as it were, by means 
of the shear, and probably not existing, as such, in the 
unperturbed ether. One of these components we call 
positive, and the other negative, electricity. Once 
formed they do not disappear again : they may 
combine or approach each other so closely that they 



206 MASSIVENESS OF ETHER 323 

neutralize each other's effects at a distance ; but they 
are still readily separable by electromotive force. They 
do not combine in the sense of destroying each 
other, they do not re-form the original substance out 
of which they were produced. 

The negative electricity, when separated, is freely 
mobile and easily isolated : it is what we experience 
as an electron. The positive constituent does not 
appear thus in an isolated manner, but is only known 
to exist in a mass, a mass very like a hydrogen atom 
possibly an inseparable aggregate of opposite charges 
matted together and going about as a whole. Some 
of these aggregates may unite into larger ones ; others, 
when too large, may split up into smaller ones ; and 
so finally a set of sub-permanent stable aggregates are 
formed which we recognize as the atoms of the so- 
called u elements " of matter : each with its appropriate 
degree of stability. 

These masses or aggregates may temporarily 
acquire, or may lose, one or more of the free electrons ; 
and by thus becoming amenable to electrical or 
chemical attractions and repulsions, constitute what 
we call " ions," so long as the unbalanced or electrified 
condition lasts. 



Massiveness of the Ether. 

206 Each electron, moving like a sphere through a 
fluid, has a certain mass associated with it ; dependent 
on its size, and, at very high speeds, on its velocity 
also. 

Now how shall that mass be treated ? 

Y 2 



324 VIEW OF THE ETHER (1907) CH. xvn 

Shall we deal with it .on the analogy of a sphere 
moving through a perfect irrotational liquid, without 
examining into details any further ? 

Or shall we consider it as generating circular lines 
of magnetic induction by its movement, by reason of 
the rotational properties of the ether, and attribute all 
its inertia to the magnetic whirl thus caused round 
its path : possibly treating the whirl as an actual cir- 
culation of fluid excited by the locomotion ? 

Both methods may be adopted, to see whether they 
will agree. 

Now treating it by the first method, and considering 
the electron merely as a sphere moving through a 
perfect liquid, its behaviour is exactly as if its mass 
were increased by half that of the fluid displaced and 
the surrounding fluid were annihilated. It has been 
argued in 126, from the result of the Cavendish 
surface-charge experiment, and from the phenomena 
of gravitation, that the ether is incompressible, to a 
high degree of exactness ; and accordingly the density 
of fluid inside and outside an electron must be the 
same. So that, treating it in this simplest fashion, 
the resultant inertia is half as great again as that of 
the volume of fluid corresponding to the electron : that 
is to say, is 27r/>a 3 , where p is the uniform density. If 
an electron is of some other shape than a sphere, 
then the numerical part is modified, but remains of 
the same order of magnitude. 

Now treat it by the other, or magnetic whirl, 
method. 

Let a spherical electron e of radius a be flying at 



206 MASSIVENESS OF ETHER 325 

speed ?/, so that the magnetic field at any point, r #, 
outside, is 

TT eus'mO 



and the energy per unit volume everywhere is yL6H 2 /8?r. 

It has been shown by Lord Kelvin, Mr. Heaviside, 
G. F. FitzGerald, and Prof. Larmor, that a magnetic 
field may be thought of, hypothetically, as a circulation 
of fluid along the lines of magnetic induction which 
are always closed curves at some unknown velocity^. 

Consider the energy per unit volume anywhere : it 
can be represented by the equivalent expressions 

i 2_/^ H2 _ V A 2 sin 2 (9 
^ := 8^~8^ ^~ 
wherefore 

e sm ^ 



On the cog-wheel analogy ( 113) the highest velocity 
will be that in contact with the moving charge ; and 
there is some reason to suppose that the maximum 
velocity w at the equator of the moving sphere may be 
equal to the speed u. Elsewhere it will decrease 
with the inverse square of the distance, just as H 
decreases. 

But without any hypothesis, if there be a circulation 
at all, its velocity must be a maximum at the equator 
of the sphere, where r = o and # = 90; so, calling 



this ze/ , 



iv I 

' = V. 



4-Trp 



326 VIEW OF THE ETHER (1907) CH. xvn 

and 

w _# 2 sin 

o 

and therefore the major part of the circulation is 
limited to a region not far removed from the surface 
of the electron. 

The energy of this motion is 

l/o I \w 2 . 2irr sin rdO . dr, 

whence, substituting the above value of w, the energy 
comes out equal to |- 7rpa?w Q 2 . 

Comparing this with a mass moving with speed u, 

= 8 

This agrees with the simple hydrodynamic estimate of 
effective inertia if WQ = %^/3u, that is to say, if the 
whirl in contact with the equator of the sphere is of 
the same order of magnitude as the longitudinal rack- 
motion or cog-wheel spin ( 99) at the same place. 

Now for the real relation between w and u we 
must make a hypothesis. If the two are considered 
equal, the effectively disturbed mass comes out as 
twice that of the bulk of the electron. If w is 
much smaller than u, then the mass of the effectively 
disturbed fluid is much less even than the bulk of an 
electron ; and in that case the estimate of the fluid- 
density p must be exaggerated enormously, in order 
to supply the required energy. It is difficult to suppose 
the equatorial circulation w greater than it, since it is 
generated by it ; and it is not unreasonable to treat 



206 MASSIVENESS OF ETHER 327 

them both as of the same order of magnitude. So, 
taking them as equal, 

*>r/w 



and m = twice the spherical mass. 

Hence all the estimates of the effective inertia of an 
electron are of the same order of magnitude, being all 
comparable with that of a mass of ether equal to the 
electron in bulk. 

This would also be the conclusion drawn, if, instead 
of integrating the magnetic energy from a to infinity, 
we integrated from a to a larger radius b, or say na ; 
the inertia would then come out 



3^ a 

and be still of the same order of magnitude for all 
reasonable values of n ; the feason being that all the 
effective disturbance is concentrated in the neighbour- 
hood of the charge. 

Now the linear dimension of an electron is io~ 13 
centimetre diameter, and its mass is of the order io~ 27 
gram, being about the i/7OOth part of the atom of 
hydrogen. Consequently, if its mass were due to its 
contents, the density of its material must be of the 
order 

IO -27_i_ IO - 39 = io 12 grams per cubic centimetre. 

This, truly, is enormous, but any 'reduction in the 
estimate of the circulation speed, below that of an 
electron, would only go to increase it ; and since 
electrons move sometimes at a speed not far below 



328 VIEW OF THE ETHER (1907) CH. xvn 

that of light we cannot be accused of underestimating 
the probable velocity of magnetic spin by treating it 
as of the same order of magnitude, at the bounding 
surface of the electron : a relation suggested, though 
not enforced, by the cog-wheel and gyrostat analogy. 

Incidentally, we may notice how enormous is the 
magnetic field surrounding the equator of an electron 
moving along an axis with, say, one- thirtieth the speed 
of light : it amounts to io 15 C.G.S. lines per sq. centi- 
metre. And the magnetic energy there is correspond- 
ingly enormous, being 4X io 28 ergs per c.c. At the 
velocity of light it would equal the constitutional 
energy of the ether itself. 

The centrifugal force of the whirl outside a flying 
electron, treating it as that of a set of concentric 
spherical shells, is 



where w = the equatorial speed = u. 



Wherefore F = 



that is to say the same force as would be exerted on 
its constraints by a particle of the same mass as an 
electron revolving in its own circumference at its own 
speed. 

At one-thirtieth the speed of light, this amounts to 
about ten grams' weight distributed over the surface 
of the electron. At the speed of light this also rises 
to the critical io 33 dynes per sq. centimetre, correspond- 
ing to the ether's intrinsic energy. 



207 ETHER CONSTANTS 329 

Value of the Ether eous x Constants. 

207 It has been argued throughout the book that 
the ethereons density is what we know in magnetism as 
47774 ; wherefore an approximate estimate of the 
absolute value of the magnetic constant /-& for free 
space, on this view, is lo 11 grams per c.c. 

Using the value 47774 = p y we get for the charge of 
an electron e = 47ra 2 y or comparable to its superficies. 

The speed with which waves travel through the 
medium is the square root of io 21 C.G.S. ; consequently 
the elasticity of the ether must be of the order io 33 
dynes per square centimetre ; and it is what in static 
electricity we denote by ^TT/K. Wherefore an approxi- 
mate estimate of the absolute value of Faraday's 
dielectric constant, K, for free space, is io~ 32 cubic 
centimetre per erg. 

In other words, the intrinsic energy of constitutional 
motion of the ether, to which its rigidity is due, is of 
the order io 33 ergs or io 26 Joules per cubic centimetre 
about a hundred foot Ibs. per atomic volume ; which 
is equivalent to the output of a million horse-power 

1 Note on the word Ether eous. The usual word " ethereal" suggests 
something unsubstantial, and is so used in poetry ; but for the prosaic 
treatment of Physics it is unsuitable, and etheric has occasionally been 
used instead. No just derivation can be given for such an adjective, 
however ; and I have been accustomed simply to spell etherial with an i 
when no poetic meaning was intended. This alternative spelling, I am 
told by a Scholar, is not incorrect ; but he points out that Milton uses 
the variant " ethereous," in a sense suggestive of something strong and 
substantial (Par. Lost, vi. 473). This word, therefore, can be employed to 
replace " ethereal " in physics : especially as the ether is now turning out 
to be by far the most substantial body known, in comparison with which 
the hitherto contemplated material universe is like a vapour of extreme 
tenuity, a barely perceptible filmy veil. 



330 VIEW OF THE ETHER (1907) CH. xvn 

working for forty million years, in every cubic milli- 
metre of space. It can otherwise be expressed as the 
energy of a thousand tons per cubic millimetre, moving 
with the velocity of light ; but of course the motion 
really contemplated is all internal and circulatory. 



Transmission of Waves. 

208 Wherever electrons and atoms exist, they 
modify the ether in their immediate neighbourhood, so 
that waves passing through a portion of space contain- 
ing them are affected by their presence, as if the ether 
were more or less loaded by them ; because the electric 
displacements which go on in the unseparated and 
still perfectly united constituents of free ether are also 
shared to some extent by the separated peculiarities, 
especially by those of the electrons which are not too 
embedded in or surrounded by a positive charge for 
instance, like a nucleus in a shell. These might be 
inert, and without influence on the light, except as 
small fixed mechanical obstructions ; but all those 
charges which possess externally-reaching lines of 
force must share in the motion of the waves, without 
having the requisite amount of resilience to compensate 
for their inertia ; consequently they, to that extent, 
constitute a retarding, and either an absorbing or a 
reflecting, agency. 

Furthermore, their motions of vibration and rotation 
during the epochs of acceleration, however caused, 
encounter the inertia of the medium, and thereby 
excite waves in it waves of oscillatory electric 
displacement with magnetic concomitant ; and this 



2o8 ETHER STRUCTURE 331 

electromagnetic radiation is transmitted out into 
space. But it is insignificant in amount unless the 
acceleration is violent ; it is proportional to the 
square of the acceleration. 

The positive and negative constituents, when they 
combine or cohere, do not destroy each other and revert 
into plain ether again ; on the contrary, they retain 
their individuality and persist, in either a combined or 
separate state. We do not know how to produce or 
to destroy these peculiarities ; and though atoms of 
matter are composed of them, and though all electrical 
phenomena and the excitation of radiation are due to 
their presence and behaviour, it is no more and perhaps 
not much less correct to say that the main bulk of the 
ether is composed of them, than it is to say that actual 
sodium and chlorine exist in undissociated common 
salt. These elements only make their appearance 
when the original substance is decomposed. But 
certainly matter can be dissociated with extreme ease, 
whereas the dissociation of ether is unknown and 
hypothetical, save as represented by its apparent 
results. 

Nevertheless, it must be the case that the slight, 
almost infinitesimal, shear, which goes on in the light 
wave, is of the nature of incipient and temporary 
electric separation ; and all electromotive force tends 
to drive one constituent in one direction and the other 
in the other ; thus beginning that individualisation or 
separate manifestation of the two ingredients, without 
a knowledge of which the original fluid would have 
appeared to be of a perfectly uniform and homogeneous 
character. 



332 VIEW OF THE ETHER (1907) CH. xvn 

It is quite possible that the actually double aspect 
of ether is not only manifested, but really generated, 
by an electromotive force applied to it, just as the 
elastic recoil is so generated. It appears possible 
that a sufficiently violent E.M.F., applied to the ether 
by some method unknown to us at present, must be 
the kind of influence necessary to shear it beyond the 
critical value and leave its components permanently 
distinct ; such constituents being opposite electric 
charges, which, when once thoroughly separated, only 
combine to form matter, and do not recoil into 
ordinary ether again. 

Hypothetical Longitudinal Stress. 

209 Every attempt at separation of this kind, even 
if no stronger than exists in ordinary light, seems to be 
accompanied by a slight longitudinal force at right 
angles both to the displacement and to the orbital 
axis of the excursion a force which is known as the 
normal pressure of light, or Maxwell's pressure, 
perpendicular to an advancing wave-front : the inertia 
of the constantly encroached upon region of free ether 
having the effect of momentum. Cf. 202. 

If the disturbance could be made so extreme as to 
result in permanent dislocation, this pressure might 
leave behind it, as permanent residue, a longitudinal 
pressure, extending throughout space inversely as the 
distance ; whereby all the dislocated material would 
thereafter be urged together with a force which we 
know as gravitation, proportional to the rate of varia- 
tion of this pressure, and proportional in any piece of 
matter to the number of dislocated centres which go 



210 AMPLITUDE OF WAVES 333 

to compose it, and therefore proportional to its mass, 
irrespective of secondary accidents of a physical or 
chemical constitution. See Appendix (r). 



Amplitude of Light Wave. 

2 10 If a is the amplitude of shear during the passage 
of a wave of light, and if u is the maximum velocity 
of recovery, then 

u a 



where v is the velocity of light and \ the wave-length. 

The total energy per unit volume is |/o?/ 2 , where p 
is the density of the medium ; for this represents twice 
the average kinetic energy, and of this quantity one- 
half is really kinetic, the other half potential. 

Now direct thermal measurements such as those 
conducted by Pouillet give, as the energy of sunlight 
near the earth, 4 x io~ 5 erg per c.c. ; and consequently 
in the region of intense light near the solar surface 
the energy of radiation must be about 2 ergs per 
cubic centimetre. There may be more intense light 
than this, but this is the most intense we know of; so 
it is instructive to consider the amplitude of the shear 
corresponding to such violent illumination. Let us 
therefore put |/)& 2 = 2, whence 



It follows, therefore, that 

/# = X io~ 16 , 
and accordingly the amplitude, of the most intense 



334 VIEW OF THE ETHER (1907) CH. xvn 

visible light we are acquainted with is only io~ 17 of a 
wave-length. The maximum strain is 2?r times this 
fraction ; and so the tangential stress thus called out 
in the medium, its rigidity being io 33 , may be estimated 
as comparable with io 17 dynes per square centimetre, 
or io 11 atmospheres. 

The ordinary electrostatic unit of charge, on this 
estimate of ethereous elasticity, becomes io~ 16 square 
centimetre, or the superficial dimension of an atom- 
This also corresponds with the estimate above, that 
the electronic charge is equal to the superficies of an 
electron ; since one should be io 10 times the other. 

The pressure of light has been represented by Prof. 
Poynting as a travelling momentum, like that of a jet 
of water, resulting in a pressure pcv ; where c is the 
velocity of longitudinal motion or circulation in the 
light beam, and v is the velocity of light. Taking the 
pressure of intensest light as 2 dynes per square centi- 
metre, this gives ^=io~ 22 centimetre per second: 
excessively small, therefore, even in that extreme 
case. 



Hypothetical Flow along Lines of Magnetic Induction. 

211 It has long been a working hypothesis with 
some mathematical physicists (see, for instance, the 
April 1907 number of the Philosophical Magazine) that 
there was probably something of the nature of a flow 
an ethereous flow along lines of magnetic induction ; 
and the fact that these are always closed curves, in all 
known circumstances, is in favour of such an idea. 
The energy of the field would then be attributed to 



2ii HYPOTHETICAL FLOW 335 

the energy of this flow ; and though it is possible that 
the flow might be of the nature of components moving 
in opposite directions, the movement is hardly likely 
to be of this nature, since that would correspond with 
merely an electric current. 

Fourteen years ago, in 1893, having rather perfect 
appliances for examining the effect of drift on the 
velocity of light, I carefully looked for some longi- 
tudinal flow along lines of magnetic force ; repeating 
the experiments still more anxiously when I learnt 
that something of the kind was seriously suspected by 
Dr. Larmor. 

Applying a field of 1400 C.G.S. units over a length 
of light path of about 14 metres in the aggregate, 
things were so arranged that a drift of I foot a second, 
or about io~ 9 th of the velocity of light, would have 
been observed by a fractional shift of micrometically 
viewed interference bands, if it had occurred. But no 
effect whatever on the interference bands could be 
detected ; nor was anything observed when with less 
perfect vision, in that case, owing to increassd 
difficulties the air along the field and path of light 
was replaced by bisulphide of carbon ; except that, of 
course, if plane-polarised light was used, the plane was 
then rotated by a very large amount. Sufficient 
details of this series of negative experiments are 
given in the April 1907 issue of the Philosophical 
Magazine. 

The result was to show that if the magnetic energy 
were to be accounted for in the assumed kinetic 
fashion, the density of the ether must be very con- 
siderable in fact about 180 times that of water, in 



336 VIEW OF THE ETHER (1907) CH. xvn 

order to give the actual energy with a velocity below 
what could be observed in this way. 

I have now, however, as described above, made a 
theoretical estimate of the density of the ether- 
arriving at the tentative conclusion that it is of the 
order io~ 12 and we can therefore proceed to calculate 
what velocity of hypothetical ethereous drift is to be 
expected in any given magnetic field. It will come 
out, of course, exceedingly slow ; for, on this view, the 
electromagnetic unit of field is /*-*, which equals 
3 x io~ 6 centimetre per second, and the velocity to be 
expected is the 2?rth of that. 

So, for instance, the field inside a solenoid, 
surrounded by a current of 100 amperes circulating 
100 times round every centimetre of it, being 477-^0, 
will equal 12,000 C.G.S. ; which corresponds with a 
velocity of 0^003 centimetre per second, or about 4 
inches an hour. In fact the ampere-turns per inch, in 
any solenoid, measures the speed of magnetic circula- 
tion along its axis, no matter what the material of the 
core may be, in millimicrons per second. 

When iron is substituted for air, the speed is the 
same ; but the ethereous density is virtually increased, 
by the loading due to the molecular whirls in the 
iron. 

It may seem difficult to reconcile this very slow 
velocity, in any ordinary field, with the great velocity 
of the very same character, already postulated in the im- 
mediate neighbourhood of an electron ; where it is sup- 
posed that the magnetic circulation is equal to, or at any 
rate of the same order of magnitude as, the locomotion 
speed which it is well known may easily be i/3Oth 



2ii HYPOTHETICAL FLOW 337 

of the velocity of light, without departing appreciably 
from the simply calculated inertia. But that great 
speed, in the immediate neighbourhood of an electron, 
can be fully admitted ; and there is nothing really 
inconsistent in that with the slow speed observed at 
any ordinary distance. For instance, if, close to the 
equator of a flying electron, the ethereal magnetic 
speed is i/3Oth of the velocity of light, or io 9 centi- 
metres per second ; then, at a distance of i millimetre 
away, the speed is reduced to io~ 24 th of that value, 
and is, therefore, even at that small distance, only 
io~ 15 centimetre per second, or one-third of a micron 
per thousand years. 

The speed at the axis of a solenoid is, of course, far 
greater than that, because of the immense number of 
electrons in any ordinary current surrounding it ; but 
in order to get up a drift-velocity of i centimetre per 
second in a solenoid, a thousand amperes would have 
to circulate three thousand times round every centi- 
metre of it ; which seems hardly practicable. 

The optical arrangements, in my experiment above 
spoken of, could doubtless be improved sufficiently to 
show an ether drift of i centimetre per second ; but I 
do not see how to produce a field of the required 
intensity to give even this leisurely flow. Such a 
field would have to be about four million C.G.S. 
units, and must exist throughout a great length of 
air. 

The experimental verification of the above theoreti- 
cal estimate of ethereous density seems therefore to be 
beyond the reach of this form of experiment. Never- 
theless, I feel reasonably convinced that there is a 

z 



338 VIEW OF THE ETHER (1907) CH. xvn 

justification for assuming the ether to have properties 
such as can only for the present be represented, in 
analogy with the properties of matter, by saying that 
its behaviour consistently indicates something typified 
by its possession of an immense elasticity or rigidity, 
io 33 dynes per square centimetre, caused by its intrinsic 
constitutional energy ; combined with a property 
analogous to, and resulting in, material inertia, and 
typified by attributing to it a density of the order io 12 
grams per cubic centimetre. The ethereous property 
here called elasticity is certainly the source and 
origin of every kind of material elasticity and potential 
energy ; for the only real static effect producible in 
the particles of matter is a change in their arrangement 
or configuration. All stress must exist really in the 
ether. 

Although the experimental methods so far 
suggested have proved themselves unable to test the 
magnitudes involved in these high values, some other 
method of inquiry may be suggested, and the theory 
may yet be brought to the test of experiment. 



Other Anticipations of Great Ether Density. 

212 I pointed out in an appendix to my Philosophical 
Magazine paper for April 1907, that many great 
authorities had anticipated a high ethereous density, 
partly on the basis of a flow along magnetic lines 
of force. But Dr. O. W. Richardson of Princeton 
has since shown, in Nature vol. Ixxvi. p. 78, that, 
without assuming a flow of that precise kind, a 
similar calculation may be made on another basis ; 



212 ANTICIPATIONS 339 

namely, the basis laid by J. J. Thomson, of a flow 
at right angles to both sets of lines of force and 
along the Poynting vector. This mode of regarding 
the matter results numerically in a density of the 
same order of magnitude as that given by the other 
calculation, and is therefore confirmatory of the 
general propositions sustained in the present chapter ; 
though the question of how far the hypothetical 
magnetic flow here spoken of is a reality must be 
regarded as sub judice, and in fact as a question for 
the mathematical physicists who originated the idea. 
Prof. Larmor sustains the reality of such a flow, 
as a physical interpretation of mathematical results 
rigorously arrived at by the principle of " Least 
Action" (Nature, July 18, 1907, page 270); but it 
is desirable, before leaving the subject, briefly to 
indicate the other or alternative mode of regarding 
the matter, and of arriving at a not dissimilar con- 
clusion, as popularly explained by J. J. Thomson 
in his American Lectures in the following way : 



Extracts from /. /. Thomson's book " Electricity and 
M. atter" (1904). 

" The momentum in unit volume of the medium 
is 47Tyu,N 2 v sin 0, and is in the direction of the 
component of the velocity of the Faraday tubes [lines 
of electric force] at right angles to their length. 
Now this is exactly the momentum which would 
be produced if the tubes were to carry with them, 
when they move at right angles to their length, a 
mass of the surrounding medium equal to 47r/iN 2 per 

z 2 



340 VIEW OF THE ETHER (1907) CH. xvn 

unit volume, the tubes possessing no mass themselves 
and not carrying any of the medium with them 
when they glide through it parallel to their own 
length." .... 

" We shall call the mass 47Tyu.N 2 , carried by the 
tubes in unit volume, the mass of the bound ether. 
It is a very suggestive fact that the electrostatic 
energy, E, in unit volume is proportional to the mass 
of the bound ether in that volume .... thus E is 
equal to the kinetic energy possessed by the bound 
mass when moving with the velocity of light." .... 

" On this view of the constitution of matter, part of 
the mass of any body would be the mass of the 
ether dragged along by the Faraday tubes stretching 
across the atom between the positively and negatively 
electrified constituents. The view I wish to put 
before you is that it is not merely a part of the mass 
of a body which arises in this way, but that the whole 
mass of any body is just the mass of ether surrounding 
the body which is carried along by the Faraday tubes 
associated with the atoms of the body. In fact, that 
all mass is mass of the ether, all momentum, 
momentum of the ether, and all kinetic energy, 
kinetic energy of the ether. This view, it should be 
said, requires the density of the ether to be im- 
mensely greater than that of any known substance." 

Summary. 

213 Thus, our hypothesis is as follows : Through- 
out the greater part of space we find simple unmodified 
ether, elastic and massive, squirming and quivering 



213 ETHICS AND MATTER 341 

with energy, but stationary as a whole. Here and 
there, however, we find specks of 'electrified ether , isolated 
yet connected together by fields of force, and in a 
state of violent locomotion. 

These ' specks ' are what, in the form of prodigious 
aggregates, we know as " matter " ; and the greater 
number of sensible phenomena, such as viscosity, heat, 
sound, electric conduction, absorption and emission of 
light, belong to these differentiated or individualized 
and dissociated or electrified specks, which are either 
flying alone or are revolving with orbital motion in 
groups. The " matter " so constituted, built up of 
these well separated particles, with interstices enormous 
in proportion to the size of the specks, must be an 
excessively porous or gossamer-like structure, like a 
cobweb, a milky way, or a comet's tail ; and the inertia 
of matter that is, the combined inertia of a group 
of electrified ether particles must be a mere residual 
fraction of the mass of the main bulk of undiffer- 
entiated continuous fluid occupying the same space ; 
of which fluid the particles are hypothetically 
composed, and in which they freely move. 



APPENDED LECTURES 

The following lectures bearing on the subject of this book are 
here conveniently appended.} 



LECTURE I 

THE RELATION BETWEEN ELECTRICITY AND 
LIGHT l 

WHEN a person is setting off to discuss the relation 
between electricity and light it is very natural and 
very proper to pull him up short with the two 
questions : What do you mean by electricity ? and, 
What do you mean by light ? These two questions I 
intend to try briefly to answer. And here let me 
observe that in answering these fundamental questions 
I do not necessarily assume a fundamental ignorance 
on your part of these two agents, but rather the 
contrary ; and must beg you to remember that if 
I quote well-known and simple experiments it is for 
the purpose of directing attention to their real mean- 
ing and significance, not to their obvious and super- 
ficial characteristics : in the same way that I might 
repeat the exceedingly familiar experimentof dropping 
a stone to the earth if we were going to define what 
we meant by gravitation. 

Now then we will ask first, What is Electricity ? 
and the strict answer must be, We don't exactly know. 

1 Portions of a lecture delivered at the London Institution on 
December 16, 1880. 



346 APPENDED LECTURES LECT. i 

Well, but this need not necessarily be depressing. If 
the same question were asked about Matter, or about 
Energy, we should have likewise to reply, No one 
knows. 

But then the term Matter is a very general one, 
and so is the term Energy. They are heads, in fact, 
under which we classify more special phenomena. 

Thus if we were asked, What is sulphur ? or What 
is selenium ? we should at least be able to reply, A 
form of matter ; and then proceed to describe its pro- 
perties, i.e. how it affected our bodies and other 
bodies. 

Again, to the question, What is heat ? we can reply, 
A form of energy ; and proceed to describe the 
peculiarities which distinguish it from other forms 
of energy. 

But to the question, What is electricity ? we have 
no answer pat like this. We cannot assert that it is 
a form of matter, neither can we deny it ; on the 
other hand, we certainly cannot assert that it is a form 
of energy, and I should be disposed to deny it. It 
may be that electricity is an entity per se, just as 
matter is an entity per se. 

Nevertheless I can tell you what I mean by 
electricity by appealing to its known behaviour. 

Here is a battery that is, an electricity pump : it 
will drive electricity along. Some people say that it 
generates electricity ; but that is exactly what neither 
it nor anything else can do. It is as impossible to 
generate electricity in the sense I am trying to give 
the word, as it is to generate matter, though it is 
perfectly possible to generate heat, e.g. by simple 



LECT. I ELECTRICITY AND LIGHT 347 

friction : that is by converting some other form of 
energy into heat. 

I want you then to regard this battery and all 
electrical machines and batteries as kinds of electricity 
pumps, which drive the electricity along through the 
wire very much as a water-pump can drive water along 
pipes ; and that no electric machine can manufacture 
electricity, any more than a pump can manufacture 
water. 

Conversely an electric motor driven by a current 
consumes none of the current ; just as much flows 
away from it by the low-potential wire as was supplied 
by the high-potential wire. In this respect it is like a 
water-wheel or turbine, where the tail-water is equal 
in quantity to the head-water : the only loss being a 
loss of level and consequent loss of available energy. 

Based on this conservation of current several 
attempts at perpetual motion were made at one time, 
by people with half knowledge, who conceived that the 
unexpended current from one motor could be used to 
drive another and so on ad infinitum ; thus gaining 
any amount of energy from a single supply. This 
can be done if the electromotive force of the generator 
is high enough, but it must always exceed the sum of 
the back E.M.F.'s of all the motors ; and accordingly 
the power that has to be expended, in driving the 
generator or dynamo, is always in excess of the 
power obtained from the series of motors ; and in fact 
the series is no more efficient than a single good one 
which develops an adequate back E.M.F. The case is 
quite analogous to a pump used to elevate water, the 
high pressure water being then used to drive a turbine 



348 APPENDED LECTURES LECT. i 

or series of turbines. It is only a case of transmission 
of power, with, of course, some loss in the process ; 
though surprisingly little loss in the electrical case, if 
well designed. 

While the flow of electricity is going on, the wire 
manifests a whole series of properties, which are 
called the properties of the current. 

[Here were shown an ignited platinum wire, the 
electric arc between two carbons, an electric machine 
spark, an induction-coil spark, and a vacuum tube 
glow. Also a large nail was magnetized by being 
wrapped, in the current, and two helices were sus- 
pended and seen to direct and attract each other.] 

To make a magnet, then, we only need a current of 
electricity flowing round and round in a whirl. A 
vortex or whirlpool of electricity is, in fact, a magnet ; 
and vice versd. And these whirls have the power of 
directing and attracting other previously existing 
whirls according to certain laws, called the laws of 
magnetism. And, moreover, they have the power of 
exciting fresh whirls in neighbouring conductors, and 
of repelling them according to the laws of dia- 
magnetism. The theory of the actions is known ; 
though the nature of the whirls, as of the simple 
stream of electricity, is at present unknown. 

[Here was shown a large electro-magnet and an 
induction-coil vacuum discharge spinning round and 
round when placed in its field (Fig. 24).] 

So much for what happens when electricity is made 
to travel along conductors, i.e. when it travels along 
like a stream of water in a pipe, or spins round and 
round like a whirlpool 



LECT. I ELECTRICITY AND LIGHT 349 

But there is another set of phenomena, usually re- 
garded as distinct and of another order, but which 
are not so distinct as they appear, which manifest 
themselves when you join the pump to a sheet of glass 
or any non-conductor and try to force the electricity 
through that. You succeed in driving some through, 
but the flow is no longer like that of water in an open 
pipe ; it is as if the pipe were completely obstructed 
by a number of elastic partitions, or diaphragms. 
The water cannot move without straining and bending 
these diaphragms, and if you allow it, these strained 
partitions will recover themselves and drive the water 
back again. [Here was explained the process of 
charging a Leyden jar, and the model (Fig. n, 24) 
was shown.] The essential thing to remember is 
that we may have electrical energy in two forms, 
the static and the kinetic ; and it is therefore also 
possible to have the rapid alternation from one of 
those forms to the other, called vibration. 

Now we will pass to the second question : What 
do you mean by Light ? And the first and obvious 
answer is, Everybody knows. And everybody that is 
not blind does know to a certain extent. We have a 
special sense-organ for appreciating light, whereas we 
have none for electricity. Nevertheless, we must 
admit that we really know very little about the inti- 
mate nature of light very little more than about 
electricity. But we do know this, that light is a form 
of energy ; and, moreover, that it is energy rapidly 
alternating between the static and the kinetic forms 
that it is, in fact, a special kind of energy of vibra- 
tion. We are absolutely certain that light is a periodic 



350 APPENDED LECTURES LECT. i 

disturbance in some medium, periodic both in space 
and time ; that is to say, the same appearances regu- 
larly recur at certain equal intervals of distance at 
the same time, and also present themselves at equal 
intervals of time at the same place ; that in fact it 
belongs to the class of motions called by mathemati- 
cians undulatory or wave motions. 

A row of a large number of equidistant similar 
pendulums serves well to illustrate a disturbance 
periodic both in space and time. The pendulums 
themselves, at rest, may be said to be periodic in 
space ; like a row of palings, or a grating, or many 
other things. The motion of each pendulum separ- 
ately is periodic in time, repeating itself at regular 
intervals called the time-period. If now such a set of 
pendulums be successively disturbed, one after the 
other, in quick succession, the resulting motion is 
periodic both in time and space ; and if the experi- 
ment is neatly done by steadily withdrawing a deflector 
the series of swinging bobs will appeal to the eye very 
distinctly as a wave motion, the waves advancing 
from one end of the row to the other, constantly coming 
in at one end and passing out at the other, although 
manifestly there can be no material transfer. 

The velocity of transmission in this case is some- 
what arbitrary : it is equal to the wave-length divided 
by the time of swing, as usual, but the wave-length 
depends only upon the way the pendulums were set 
in motion ; so that it is the rate of removal of the 
initial deflexion which determines the velocity. 
In that respect it does not correspond to natural 
waves in a medium ; but it is otherwise a good illus- 



LECT. I ELECTRICITY AND LIGHT 351 

tration. It is instructive to remark that in this model 
the velocity is directly proportional to the wave- 
length ; in large water waves the velocity is propor- 
tional to the square root of the wave-length ; in sound 
and light waves it is independent of the wave-length 
being determined solely by the properties of the 
medium, provided the medium is homogeneous. In 
so far as there is any dependence of speed on wave- 
length in the case of light, the result is dispersion, and 
the cause of it is a lack of homogeneity in the medium : 
for instance, a jelly with plums embedded in it is 
not a homogeneous medium. There are other cases 
known, where the velocity of a wave is actually 
greater as the wave-length is shorter ; such cases can 
be found in minute ripples of liquid, and in the 
impulses transmitted by rigid bars. 

The wave motion in a well-known model, called 
Powell's wave apparatus, results from the simple up- 
and-down motion popularly associated with the term 
wave. But when a mathematician calls a thing a 
wave he means that the disturbance is represented by 
a certain general type of formula ; not that it is an up- 
and-down motion, or that it looks at all like those 
things on the top of the sea. The motion of the 
surface of the sea falls within that formula, and hence is 
a special variety of wave motion. But the term " wave v 
has acquired in popular use this signification and 
no other ; so that when one speaks ordinarily of a 
wave, or undulatory motion, people immediately think 
of something heaving up and down, or even perhaps 
of something breaking on the shore. But when we 
assert that the form of energy called light is undula- 



352 APPENDED LECTURES LECT. i 

tory y we by no means intend to assert that anything 
whatever is moving up and down, or that the motion, 
if we could see it, would be anything at all like what 
we are accustomed to on the ocean. The kind of 
motion is unknown ; we are not even sure that there 
is anything like " motion " in the ordinary sense of the 
word at all. 

Now how much connexion between electricity and 
light have we perceived in this glance into their 
natures ? Not much, truly. It amounts to about this : 
That on the one hand electrical energy may exist in 
either of two forms the static form, when insulators 
are electrically strained by having had electricity 
driven partially through them (as in the Leyden jar), 
which strain is a form of energy because of the 
tendency to discharge and do work ; and the kinetic 
form, where electricity is moving bodily along through 
conductors or whirling round and round inside them 
which motion of electricity is a form of energy, 
because the conductors and whirls can attract or repel 
each other and thereby do work. 

And, on the other hand, that light is the rapid alter- 
nation of energy from one form to another from the 
static form where the medium is strained, to the 
kinetic form where it moves. It is just conceivable 
then that the static form of the energy of light is 
electro-static that is, that the medium is electrically 
strained and that the kinetic form of the energy of 
light is electro-Y\\\eti.Q. that is, that the motion is not 
ordinary motion, but electrical motion ; in fact that 
light is an electrical vibration, not a material one. 

On November 5, 1879, there died at Cambridge a 



LECT. i ELECTRICITY AND LIGHT 353 

man in the full vigour of his faculties such faculties 
as do not appear many times in a century whose 
chief work had been the establishment of this very 
fact, the discovery of the link connecting light and 
electricity ; and the proof for it certainly amounts to 
a proof that they are different manifestations of one 
and the same class of phenomena : that light is, in 
fact, an electro-magnetic disturbance. The premature 
death of James Clerk Maxwell is a loss to science 
which appears at present utterly irreparable, for he 
was engaged in researches that no other man can hope 
as yet adequately to grasp and follow out : but for- 
tunately it did not occur till he had published his 
book on Electricity and Magnetism, one of those 
immortal productions which exalt one's conception of 
the mind of man, and which has been mentioned by 
competent critics in the same breath as the Principia 
itself. 

The main proof of the electro-magnetic theory of 
light is this. The rate at which light travels has been 
measured many times, and is pretty well known. 
The rate at which an electro-magnetic wave-disturb- 
ance would travel, if such could be generated, can be 
also determined by calculation from electrical measure- 
ments. The two velocities agree exactly. This is 
the great physical constant known as the ratio " z;," 
which so many physicists have been measuring, and 
are likely to be measuring for some time to come. 

Many and brilliant as were Maxwell's discoveries, 
not only in electricity, but also in the theory of the 
nature of gases, and in molecular science generally, I 

A A 



354 APPENDED LECTURES LECT. i 

cannot help thinking that if one of them is more 
striking and more full of future significance than the 
rest, it is the one I have just mentioned the theory 
that light is an electrical phenomenon. 

The first glimpse of this splendid generalization 
was caught in 1845, five and thirty years ago, by that 
prince of pure experimentalists, Michael Faraday, 
His reasons for suspecting some connexion between 
electricity and light are not clear to us in fact they 
could not have been clear to him ; but he seems to 
have felt a conviction that if he only tried long enough, 
and sent all kinds of rays of light in all possible 
directions across electric and magnetic fields, in all 
sorts of media, he must ultimately hit upon some- 
thing. Well, that is very nearly what he did. With 
a sublime patience and perseverance, which remind 
one of the way Kepler hunted down guess after 
guess in a different field of research, Faraday com- 
bined electricity, or magnetism, and light, in all 
manner of ways ; and at last was rewarded with a 
result. 

A singularly out-of-the-way result it seemed. First 
you have to get a most powerful magnet and very 
strongly excite it ; then you have to pierce its two 
poles with tunnels, in order that a beam of light may 
travel from one to the other along the lines of force ; 
then, as ordinary light is no good, you must get a 
beam of plane-polarized light and send it between the 
poles. But still no result is obtained ; until, finally, you 
interpose a piece of a rare and out-of-the-way material 
which Faraday had himself discovered and made a 
kind of glass which contains borate of lead, and which 



LECT. I ELECTRICITY AND LIGHT 355 

is very heavy, or dense, and which must be perfectly 
annealed. 

And now, when all these arrangements are 
completed, what is seen is simply this, that if an 
analyzer is arranged to stop the light and make the 
field quite dark before the magnet is excited, then, 
directly the battery is connected and the magnet 
called into action, a faint and barely perceptible 
brightening of the field occurs ; which will disappear 
if the analyzer be slightly rotated. [The experiment 
was then shown.] Now no wonder that no one 
understood this result. Faraday himself did not 
understand it ; he seems to have thought that the 
magnetic lines of force were rendered luminous, or 
that the light was magnetized ; he had no clear idea 
of its real significance. Nor had anyone. Continental 
philosophers experienced some difficulty and several 
failures before they were able to repeat the experiment. 
It was in fact discovered too soon, before the scientific 
world was ready to receive it ; and it was reserved for 
Lord Kelvin briefly but clearly to point out, and 
for Clerk Maxwell more fully to develop, its most im- 
portant consequences. 

This is the fundamental experiment which prob- 
ably suggested Clerk Maxwell's theory of light ; 
but of late years many fresh facts and relations 
between electricity and light have been discovered, 
and at the present time they are tumbling in in great 
numbers. 

It was found by Faraday that many other trans- 
parent media besides heavy glass would show the 
phenomenon if placed between the poles : only in a 

A A 2 



356 APPENDED LECTURES LECT. i 

less degree ; and the very important observation that 
air itself exhibits the same phenomenon, though to an 
exceedingly small extent, has been made by Kundt 
and Rontgen in Germany. 

Dr. Kerr, of Glasgow, has extended the result to 
opaque bodies, and has shown that if light be passed 
through magnetized iron its plane is rotated. The film 
of iron must be exceedingly thin, because of its opacity ; 
and hence, though the intrinsic rotating power of iron 
is undoubtedly very great, the observed rotation is 
exceedingly small and difficult to observe ; and it was 
only by very remarkable patience and care and in- 
genuity that Dr. Kerr obtained his result. Prof. 
Fitzgerald has examined the question mathematically, 
and has shown that Maxwell's theory would have 
enabled Dr. Kerr's result to be predicted. 

Another requirement of the theory is that bodies 
which are transparent to light must be insulators or 
non-conductors of electricity, and that conductors of 
electricity are necessarily opaque to light. Simple 
observation amply confirms this ; metals are the best 
conductors, and are the most opaque bodies known. 
Insulators such as glass and crystals are transparent, 
whenever they are sufficiently homogeneous ; and the 
very remarkable researches of Prof. Graham Bell in 
the last few months (1880) have shown that even ebonite -, 
one of the most opaque insulators to ordinary vision, is 
certainly transparent to some kinds of radiation, and 
transparent to no small degree. 

[The reason why transparent bodies must insulate, 
and why conductors must be opaque, was here 
illustrated by mechanical models. 



LECT. I ELECTRICITY AND LIGHT 357 

The model which represented a dielectric has already 
been depicted in Fig. 8 ; and when the cord thread- 
ing all the elastically supported balls is vibrated, 
waves travel readily through it. 

The model which represented a metallic conductor 
is shown here in Fig. 66. It has its wooden balls 
sliding on smooth brass rods so that they have no 
tendency to recoil to a settled position, but remain 
where placed. On shaking the cord connecting these 




FIG. 66. Rude model to pair with Fig. 8 (p. 43) and to call attention to some of 
the differences between a metal and an insulator. 

balls, the waves penetrate a certain small depth into 
the medium, but fail to get through it. 

The two models were connected in series, and waves 
which had been transmitted along the cord by one, 
were partly quenched, partly reflected, by the o