A History of The Theories of Aether & Electricity 1

A History of the Theories of
AETHER
and
ELECTRICITY
SIR EDMUND WHITTAKER
VOLUME ONE
The ClassicalTheories
HARPER TORCHBOOKS / The ScienceLibrary
HARPER f5 BROTHERS, NEW YORK
C0?7i..t· ,L-~ I , ~I'n~

.I
Sci.''Aueobi.ncio jn~ecnati.onaL
Centee of Educaii.on.
PONDICHF.RRY·.2 (INDIA)
Preface
IN 1910 I publisheda work under the title A History of the Theories
of Aether and Electriciry, from the age of Descartes to the closeof
the nineteenthcentu~'Y.When the original edition wasexhausted,I
felt that any new issueshould describethe origins of relativity
and quantum-theory, and their development since 1900. My
opportunitieswerehowevernot sufficientto enablemeto prepare
an accurateand fully-documented accountof this very creative
period, and I was compelled to lay the plan aside. Retire-
ment from my professorialchair has made it possiblefor me
to take up this project again; it will occupy three volumes,
of which this, the first, deals with the classicaltheories. The
volume of 1910 has been to a considerableextent rewritten,
with the incorporation of much additional material; and in
the secondvolume, the story will be continued to 1926.
A word might be said about the title Aether and Electriciry.
As everyoneknows,the aetherplayed a great part in the physics
of the nineteenth century; but in the first decade of the
twentieth, chiefly asa result of the failure of attemptsto observe
the earth's motion relative to the aether, and the acceptanceof
the principle that such attempts must always fail, the word
, aether' fell out of favour, and it becamecustomary to refer to
the interplanetary spacesas 'vacuous'; the vacuum being
conceivedas mere emptiness,having no properties except that
of propagating electromagneticwaves. But with the develop-
ment of quantum electrodynamics, the vacuum has come
to be regardedasthe seatof the' zero-point' oscillationsof the
electromagneticfield, of the' zero-point' fluctuations of electric
charge and current, and of a 'polarisation' corresponding
to a dielectric constant different from unity. It seemsabsurd
to retain the name ' vacuum' for an entity so rich in physical
properties,and the historical word' aether' may fitly beretained.
My grateful thanks are due to ProfessorsE. T. Copsonof the
University ofSt Andrews andJ. M. Whittaker of the University
of Liverpool for help in reading the proofs.
E. T. WHI'ITAKER
48 George Square
Edinburgh, April 1951
CONTENTS
Davy's Chemical Theory of the Pile 75

Voltaic Polarisation 76
Grothuss' Chain 76
De La Rive's Hypothesis . 77
Berzelius' Scheme of Electra-chemistry 78
Early Attempts to discover a connection between Electricity and Magnetism 80
Oersted's Experiment: His Explanation of it 81
The Law of Biot and Savart 82
The Researchesof Ampere on Electrodynamics 83
Seebeck'sPhenomenon 89
Davy's Researcheson Conducting Power go
Ohm's Theory: Electroscopic Force 91
IV THE LUMINIFEROUS MEDIUM FROM BRADLEY TO FRESNEL

Bradley discoversAberration 94
Johann Bernoulli's Model of the Aether 95
Maupertuis and the Principle of Least Action 97
Euler's Aether . 98
Views of Courtivron and Melvill 99
Radiant Heat and Ultra-violet Light 99
Young defends the Wave Theory and explains the Colours of Thin Plates 100
Laplace supplies a Corpuscular Theory of Double Refraction 104
Young proposesa Dynamical Theory of Light in Crystals 104
Researchesof Malus on Polarisation 105
Brewster's recognition of Biaxal Crystals 107
Fresnel successfullyexplains Diffraction 107
His Theory of the Relative Motion of Aether and Matter loB
Young suggeststhe Transversality of the Vibrations of Light 114
Fresnel discussesthe Dynamics of Transverse Vibrations 116
Fresnel's Theory of the Propagation of Light in Crystals 117
Hamilton predicts Conical Refraction 122
Fresnel's Theory of Reflection 123
Experiments of Airy, Foucault and Fizeau 126
V THE AETHER AS AN ELASTIC SOLID

Astronomical objection to the Elastic-solid Theory: Stokes'sHypothesis 128
Navier and Cauchy discover the Equation of Vibration of an Elastic Solid 129
Poissondistinguishes Condensational and Distortional Waves 131
Cauchy's First and Second Theories of Light in Crystals 133
Cauchy's First Theory of Reflection 134
His Second Theory of Reflection 136
The Theory of Reflection of MacCullagh and Neumann 137
Green discovers the correct Conditions at the Boundaries 139
Green's Theory of Reflection: objections to it 140
MacCullagh introduces a New Type of Elastic Solid . 142
W. Thomson's Model of a Rotationally-elastic Body . 145
Cauchy's Third Theory of Reflection: the Contractile Aether . 145
Later Work ofW. Thomson and others on the Contractile Aether 147
Green's First and Second Theories of Light in Crystals 148
Influence of Green: Neglect of his work during his Lifetime . 153
Researchesof Stokeson the relation of the Direction of Vibration of Light
to its Plane of Polarisation 154
The Hypothesis of Aeolotropic Inertia 156
Rotation of the Plane of Polarisation of Light by Active Bodies 158
MacCullagh's Theory of Natural Rotatory Power 159
CONTENTS
MacCulIagh and Cauchy's Theory of Metallic Reflection
Extension of the Elastic-solid Theory to Metals
Lord Rayleigh's objection
Cauchy's Theory of Dispersion
Boussinesq'sElastic-solid Theory
VI FARADAY
Discovery of Induced Currents: Lines of Magnetic Force 171
Self-induction 174
Identity of Frictional and Voltaic Electricity: Faraday's views on the
Nature of Electricity 175
Electro-chemistry 177
Controversy between the adherents of the Chemical and Contact Hypotheses 180
The Properties of Dielectrics 184
Theory of Dielectric Polarisation: Faraday, W. Thomson and Mossotti. 186
The Connection between Magnetism and Liltht 190
Airy's Theory of Magnetic Rotatory Polarisation 191
Faraday's ThoughtsonRay-Vibrations 193
Researchesof Faraday and Pliicker on Diamagnetism 194
VII THE MATHEMATICAL ELECTRICIANS OF THE MIDDLE OF THE

NINETEENTH CENTURY
F. Neumann's Theory oflnduced Currents: the Electrodynamic Potential 198
W. Weber's Theorv of Electrons 201
Electrostatic and Electromagnetic Units 202
Criticisms ofWeber's Theory 203
Deduction of the Formulae for the Induction of Currents 204
Riemann's Law . 206
Proposals to modify the Law of Gravitation . 207
Weber's Theory of Paramagnetism and Diamagnetism: late,' Theories 208
Joule's Law: Energetics of the Voltaic Cell. 212
The Mechanical Equivalent of Heat 213
Helmholtz and the Conservation of Energy . 214
Carnot's Principle: Cyclical Processes 215
Caloric, Entropy and Absolute Temperature 215
W. Thomson's Equation of Available Energy 217
Researchesof Helmholtz on Electrostatic and Electrodynamic Energy 217
W. Thomson distinguishes the Circuital and Irrotational Magnetic Vectors 219
His Theory of Magnecrystallic Action 220
His Formula for the Energy of a Magnetic Field 222
Extension of the Formula to the caseof Fields produced by Currents 223
Kirchhoff identifies Ohm's Electroscopic Force with Electrostatic Potential 225
The Discharge of a Leyden Jar: W. Thomson's Theory 226
The Velocity of Electricity and the Propagation of Telegraphic Signals. 227
Clausius' Law of Force between Electric Charges :" Crucial Experiments 234
Nature of the Current 235
The Thermo-electric Researchesof Peltier and W. Thomson 236
VIII MAXWELL
Gauss and Riemann on the Propagation of Electric Actions
Analogies suggestedby W. Thomson
Maxwell's Hydrodynamical Analogy
The Vector Potential
Linear and Rotatory Interpretations of Magnetism
Maxwell's Mechanical Model of the Electromagnetic Field
CONTENTS
ElectricDisplacement 249
Similarity of ElectricVibrationsto thoseof Light 252
Connectionof RefractiveIndex and SpecificInductiveCapacity 253
Maxwell'sMemoir of 1864 255
The inquiry regardingGravitation 258
Maxwell'sTreatmentof Bodiesin Motion 258
The Propagationof ElectricDisturbances in Crystalsandin Metals 259
AnomalousDispersion 261
The Maxwell-SellmeierTheoryof Dispersion 262
Imperfectionsof the Electromagnetic Theory of Light 266
The Theory of L. Lorenz 267
Maxwell'sTheoryof Stressin the ElectricField 271
The Pressure of Radiation 273
Recoilproducedby the Emissionof Light 275
Maxwell'sTheoryof the MagneticRotationof Light. 276
IX MODELS OF THE AETHER
Analogiesin which a RotatoryCharacteris attributed to Magnetism 279
Modelsin which MagneticForceis represented asa Linear Velocity 280
Researchesof W. ThoInson, Bjerknes,and Leahy on Pulsatingand
OscillatingBodies 284
MacCullagh'sQuasi-elasticSolidasa modelof the ElectricMedium 286
The Hall Effect . 289
Modelsof Riemannand FitzGerald 291
Helmholtz'discoveryregardingVortex-motion 293
Connectionwith the Atomic Theory of Matter: W. ThoInson'sVortex
AtoIns 293
The Vortex-spongeTheory of the Aether: Researches of W. Thomson,
FitzGeraldand Hicks 295
Larmor on the Aether 303
X THE FOLLOWERS OF MAXWELL
Helmholtzand Lorentzsupplyan Electromagnetic Theoryof Reflection 304
Crucial Experimentsof Helmholtzand Schiller 305
The Moving ChargedSphere: Researches of J. J. ThoInson,FitzGerald
and Heaviside 306
MechanicalForceon an ElectricChargemovingin a MagneticField 310
Conductionof Rapidly-alternatingCurrents. 311
FitzGeralddevisesthe MagneticRadiator 312
Poynting'sTheorem 313
PoyntingandJ. J. ThoInsondevelopthe theoryof Moving Linesof Force 315
MechanicalMomentumin the ElectromagneticField 317
New Derivationof Maxwell'sEquationsby Hertz 319
Hertz'sAssumptions and Weber'sTheory 321
Experimentsof Hughesand Hertz on ElectricWaves. 322
The memoirsof Hertz and Heavisideon Fieldsin which MaterialBodies
are in Motion 328
The Current of DielectricConvection 330
Kerr's Magneto-opticPhenomenon 331
Rowland'sTheoryof Magneto-optics 332
The Rotationof thePlaneof Polarisationin Naturally-activeBodies 334
XI CONDUCTION IN SOLUTIONS AND GASESFROM FARADAY TO
THE DISCOVERY OF THE ELECTRON
The Hypothesisof WilliaInsonand Clausius. 335
Migration of the Ions 336
CONTENTS
The Researchesof Hittorf and Kohlrausch 336

Polaris••tion of Electrodes. 337
Electrocapillarity 338
Single Differences of Potential 340
Helmholtz' Theory of Concentration-cells 342
Arrhenius' Hypothesis 344
The Researchesof Nernst 345
Earlier investigations of the Discharge in Rarefied Gases 349
Faraday observesthe Dark Space . 350
Researchesof Pliicker, Hittorf, Goldstein and Varley on the Cathode Rays 351
Crookes and the Fourth State of Matter 352
Objections and Alternatives to the Charged-particle Theory of Cathode
Rays 353
Giese's and Schuster's Ionic Theory of Conduction in Gases 355
Photo-electricity. 356
J.J. Thomson measuresthe Velocity of Cathode Rays 357
Discovery of X-rays: Hypotheses regarding them 357
Furthcr Researchesof J.J. Thomson on Cathode Rays 360
Prout's Hypothesis 361
The Ratio mle 362
Canal Rays 363
Vitreous and Resinous Electricity 363
Determination of the Ionic Charge by J. J. Thomson and Townsend 364
Electrons. Nature of Ionisation 365
Larmor's Argument on the Discrete Nature of Electric Charge. 365
XII CLASSICAL RADIATION-THEORY

Early discoveriesregarding Spectra 367
Doppler's Principle 368
Connection of Emission and Absorption of Spectral Lines 369
Prevost's Law of Exchanges 371
Halfour Stewart and Kirchhoff on Emission and Absorption: Black Bodies 371
Intrinsic Brightness of a Surface 373
Stefan's Law and Boltzmann's proof of it 374
Introduction of Wave-numbers by Stoney and Hartley 375
Recognition of Seriesin Spectra 375
Balmer's Formula and Rydberg's Constant 376
The Seriesof the Alkali Metals 377
Rydberg and Ritz's Combination-principle 378
Wien's TheoreIllS on Black-body Radiation 379
The Equipartition of Energy 382
Rayleigh's Law of Black-body Radiation 383
XIII CLASSICAL THEORY IN THE AGE OF LORENTZ

Stokes'sTheory of Aethereal Motion near Moving Bodies 386
Astronomical phenomena in which the Velocity of Light is involved 387
Crucial Experiments relating to the Optics of Movin& Bodies 389
Lorentz' Theory of Electrons 392
The Scalar and Vector Potentials in Lorentz' Theory. 394
Loss of Energy by Radiation from a Moving Charge . 396
Lorentz' Theory of Dielectrics 397
Rontgen's Experiment on a Dielectric moving in an Electrostatic Field 399
The Electronic Theory of Dispersion 401
Deduction of Fresnel's Formula from the Theory of Electrons • 403
CONTENTS
Experimental verification of Lorentz' Hypothesis 403
FitzGerald's explanation of Michelson's Experiment 404
Lorentz' Treatise of 1895 . 405
Modification of Fresnel's Formula 406
Lorentz' Local Time 406
Expression of the Potentials in Terms of Discrete Electronic Charges 408
The Two Scalar Potentials 409
The Zeeman Effect 410
Lorentz' Theory of it. 411
Experiments inconsistent with the Theory 413
Connection of the Zeeman Effect with the Magnetic Rotation of Light 414
The Optical Properties of Metals 416
The Electronic Theory of Metals 418
Thermionics 425
429
Chapter I
THE THEORY OF THE AETHER TO THE DEATH
OF NEWTON
THE earliest science,that of the ancient Greeks,had its origin in

the speculationsof philosophersregarding the nature of Being and
Becoming. Modern sciencelikewise was at first closely related to
philosophy,and a history of its beginningsmust be prefacedby some
accountof the philosophicalsituation at the closeof the Middle Ages.
As was pointed out by Aristotle,! there are fundamentally two
types of philosophy, according as the approach to philosophical
problems is made dialectically ('\OytKWS) or 'naturally' (epVUtKWS).
To the first type belong all thosephilosopherswho have speculated
in the grand manner, from Plato to Hegel; to the secondbelong
Democritus,Aristotle himself, and mostof the modern men of science
who have written on philosophicalsubjects.
The philosophersof the first type have produced schemes,more
or lessindependentof experience,which haveclaimed to comprehend
all Beingin their framework; thoseof the secondtype proceedmore
cautiously-their fundamental conceptsare obtained by systematis-
ing their experienceof the external world. In other words, they
proceedto their general philosophy by extending their conceptual
analysisof natural science(that this is true of Aristotle is obvious
when his Metaphysics is comparedwith his Physics). Of coursesuch a
methodimplies the principle that philosophicaltruth is attained only
progressively,and that none of the all-embracing systemsof the first
type is likely to be true.
Aristotle's plan for constructing a philosophy was, then, one
which would be fully approved by a modern man of science. But
unfortunately the cosmologicaland physical notions from which he
started were entirely false. He believed that there was a radical
differencein kind betweenthe heavenlybodieson the one hand and
terrestrial objectson the other: the former were changeless,uncom-
poundedand eternal, moving in circular paths, while the latter were
corruptible, synthetic, transient and moved naturally in straight
lines. The earth was at rest, while the stars were carried round
on celestial spheres,whose motions were directed by immaterial
1 De generationeet corruptione, i, 2, 316a
I
THE AETHER
intelligences. His word for ' matter' (vAT) did not connote matter
in our sense,but meant something out of which something else
could be made by imposing on it a further condition or structural
principle, called a form. Ordinary bodies he supposedto be con-
stituted of four elements,fire, air, water, earth; but he believed
that these could be transmuted into each other, and were there-
fore merely different 'forms' associatedwith a 'prime matter'
which was itself inaccessibleto experience. His explanation of the
motion of bodieswas,that the' substantialform' or' nature' of each
body impels it to someparticular place: thus, fire movesupwards,
while other bodies tend to move towards the centre of the earth.
All things, in fact, seekan ' end,' or, in Aristotelian terminology, a
final cause; and the ' final cause' of a movement in spaceis some
definite location. It may be added that Aristotle studied nothing
quantitatively; it is only fair to him to say that he insistedon the
necessityfor observation and experience, but the injunction was
lost sight of by many of his followers in later ages.
A typical example of the Aristotelian natural philosophy is its
interpretation of light in terms of the metaphysicalnotionsof potency
(Dvva}J-tS') and act (EV€pyEta or EVTf: A€XEta). A transparent body has
a ' potency' for transmitting light, but it doesnot becomeactually
transparent until light is passedthrough it and thereby brings the
transparencyinto action. Therefore light was defined by Aristotle
as ' the act of a transparent body, inasmuch asit is transparent.'
Under the influencechiefly ofSt ThomasAquinas,Aristotelianism
was adopted generally throughout westernEurope in the thirteenth
century. Like Aristotle, St Thomas declared his philosophy to be
of what we have called the secondtype. 'The origin of our know-
ledge,' he said,l ' is in the senses,evenof thosethings that transcend
sense;' and' metaphysicshas receivedits name, i.e. beyondphysics,
becauseto us, who naturally arrive at the knowledgeof things im-
material by meansof things sensible,it offers itself by rights as an
object of study after physics.' 2 Metaphysicsis ' the last of the parts
of philosophyto be studied,, because'it is necessaryto haveprevious
knowledgeof many things.' 3
All this is very sound, but it was impossible to extract a true
metaphysics from the false Aristotelian physics. The evil was
aggravatedby the medievalschoolmen,who regardedAristotelianism
as an encyclopaedicbody of knowledge,in no needof being supple-
1 Summacontragentiles,i, ch. 12
• Opusc.lx (ed. 1571),in Boet. de Trinit., quo 5, art. 1

• Summacontragentiles,i, ch. 4
(82S) 2
TO THE DEATH OF NEWTON
mentedby further experiments. They abandoned8t Thomast view

of the relation betweenphysicsand metaphysics. 'It is not,' they
said, ' the province of physicsto theoriseon its own facts and laws,
or to undertakea reconstructionof cosmologyor metaphysics,'1 and
, if a physical theory is inconsistent with received metaphysical
teaching,it cannot be admitted, becausemetaphysicsis the supreme
natural science,not physics.'2 They interpreted the external world
only by applying formal logic in order to make deductions from
obscureand sterileprinciples, which really representeda petrifaction
of Aristotle's erroneousphysics: a procedurewhich yielded nothing
but prolix sophistry.3 As Erasmus said, 'There are innumerable
nicetiesconcerningnotions, relations, instants, formalities, quiddities and
haecceities,which no-one can pry into, unlesshe has eyesthat can
penetratethe thickest darkness,and there can seethings that have
no existencewhatever.' t Until this tradition could be overthrown,
there wasno possibility of any right understandingof nature.
For the birth of modern science,then, a necessarycondition was
an emancipationfrom Thomist philosophy. This wasbrought about
chiefly as a result of the work of William of Ockham, an English
Franciscanfriar, who lived in the first half of the fourteenth century.
In metaphysics, Ockham introduced principles which need not
concern us here, but which undermined Aristotelianism and led
ultimately to that generalrejectionof medievalphilosophywhich was
associated with the Renaissance;and it washewho madea beginning
of rational dynamics, which after his death was developed by his
followers,and on which all subsequentprogresswas founded. The
most decisive advanceswere the new conception of the relation
betweenthe earth and the sun, proposed by Copernicus in 1543,
and the successof Kepler, in the early yearsof the next century, in
disprovingthe Aristotelian principle that celestialbodiesare different
in kind from terrestrial. But the age of decadentscholasticismwas
long drawn out, and nearly three centuries after Ockham the
Aristoteliansstill had enoughinfluence to procure the condemnation
of Galileo.o
It wasa younger contempor-aryof Galileo, Rene Descartes,who
1 T. Harper, The Metaphysicof the &hool (London, 1879-84), i, p. 41

I ibid. ii, p. 249
• This is not true of Robert Greathead and of Roger Bacon, but they were senior to
St Thomas in age and may be said to belong to the pre-Aristotelian era; and it is not
true of the followers of William of Ockham, but they were not 1homists.
• Erasmus, Moriae Encomium(1509)
• The decree of condemnation contains the following: ehe la terra non ria centro del
mondo,M immobile,rna ehesi moueetiandio di nwto diuTllo, e propositioneassurde,efalsa infilosofia.
C82a) 3
THE AETHER
first attempted a general reconstruction of ideas regarding the

physical universe,and it is with an accountof his work that the history
properly begins.
Descartes was born in 1596, the son of Joachim Descartes,
Counsellor to the Parliament of Brittany. As a young man he
followed the professionof arms, and served in the campaigns of
Maurice of Nassauand the Emperor; but his twenty-fourth year
brought a profound mental crisis, apparently not unlike thosewhich
have been recorded of many religious leaders, and he resolvedto
devotehimself thenceforwardto the study of philosophy.
The agewhich precededthe birth of Descartes,and that in which
he lived, was marked by eventswhich greatly altered the prevalent
conceptionsof the world. The discovery of America, the circum-
navigation of the globeby the companionsof Magellan, the invention
of the telescope,the overthrow of the Ptolemaicsystemof astronomy
and the generaldissatisfactionwith scholasticism,all helpedto loosen
the old foundations and to make plain the needfor a new structure.
It was this that Descartessethimself to erect. His aim was nothing
lessthan to createfrom the beginning a theory of the universe,worked
out asfar as possiblein every detail.
Of sucha systemthe basismust necessarilybe metaphysical; and
it is by this part of Descartes'work that he is most widely known.
Here we shall refer only to those of his philosophical principles
which determined his representationof the external world.!
The first step was to discard the futile methods of the Middle
Ages,the attempts to interpret Nature in terms of act and potency,
matter and form, substanceand accident,the ten categories,and the
like. This had already been proposedby many philosophers,and
new ways had beenindicated, particularly by FrancisBacon(1561-
1626) and Galileo Galilei 2 (1564-1642). They rejectedaltogether
the Aristotelian practice of referring everything in physicsto Final
Causes, S and asked,How? rather than Why? They dwelt on the
necessityfor observingthe external world, and moreoverinsistedthat
besidesthe study of suchoccurrrencesasare naturally presentedto us,
1 Of the works which bear on our subject, the Dioptrique was published (anonymously)
in 1637, the MIt/ores in 1638, and the Principia Philosophiae at Amsterdam in 1644, six
years before the death of its author. The Discoursdela methode,published in 1637, was an
essayprefatory to the Dioptrique and the Mltlores.
• The predominant part played by Galileo in destroying the Aristotelian-Thomistic
conception of the world is very well brought out by Dr A. C. Crombie in a monograph,
, Galileo's " Dialogues concerning the two principal systems of the world," , Dominican
Studies, 1950, 34 pp.
• Causarumjinalium inquisitia sterilis est, Bacon, De Augment (1623), iii, 5. Bacon's
principal philosophical works were written about eighteen years before thoseof Descartes.
4
there should be an interrogation of Nature by meansof carefully

devisedexperiments. (This recommendationwas not superfluous,
for experimentation had hitherto been associated chiefly with
alchemy,and had no goodreputation.) Bacon'steachingwasin one
important respectdefective,for he failed to appreciatethe importance
of the quantitative in contradistinction to the qualitative aspectsof
phenomena; the samechargecannot be made against Galileo, but
he was an experimentalrather than a mathematical physicist, and
it wasa third contemporary,JohannesKepler (1571-1630) who laid
stresson the supremeimportance of mathematics in the study of
Nature.
Kepler's teaching provided the chief inspiration of Descartes,
whoseresearcheswere dominated by a conviction that the theorems
of mathematicshad a precision, an indubitability and a universal
acceptance,which werenot to befound in other fieldsof study. Soto
. thesefeatureshe attachedthe highestimportance, laying it down as
an axiom that clarity and certainty were marks of all genuineknow-
ledge. Now one of the problems of natural philosophy was to
accountfor actionstransmitted betweenbodies not in contact with
each other, such as those indicated by the behaviour of magnets,
or by the connectionbetweenthe moon'sposition on the one hand
and the rise and fall of the tides on the other. To accept theseas
, occult' influenceswould havebeencontrary to his principles; and
he concludedthat they must be effectedby the agencyof the only
types of action between bodies which were perfectly intelligible,
namely, pressureand impact. This implied that bodiescan act on
each other only when they are contiguous; in other words, he
deniedaction at a distance; 1 and this had the further consequence,
that the spacebetween the moon and the earth, and indeed the
whole of space,could not be void. It is occupiedpartly by ordinary
materialthings-air and tangiblebodies; but the intersticesbetween
the particlesof these,and the whole of the restof space,must be filled
with particlesof a much more subtle kind, which everywherepress
upon, or collide with, each other: they are the contrivance intro-
ducedin order to accountfor all physicalhappenings. Spaceis thus,
in Descartes'view, a plenum, being occupied by a medium which,
though imperceptible to the senses;is capableof transmitting force,
and exerting effectson material bodiesimmersedin it-the aether, as
it is called. The word 2 had meant originally the blue sky or upper
1 The doctrine that force cannot be communicated except by pr~sure or impact had
been maintained by the ancient Greek atomists and also by Aristotle and St Thomas, but
had been denied by Duns Scotus and his followers. I alO'JP
5
THE AETHER
air (asdistinguishedfrom the lower air at the level of the earth), and
had beenborrowed from the Greek by Latin writers,l from whom it
had passedinto French and English in the Middle Ages. In ancient
cosmologyit wassometimesusedin the senseof that which occupied
celestialregions; and when the notion of a medium filling the inter-
planetary void was introduced, aether was the obvious word to use
for it. Before Descartes,it had connoted merely the occupancyof
somepart of space: he wasthe first to bring the aetherinto science,
by postulating that it had mechanicalproperties. In his view, it was
to be regarded as the solitary tenant of the universe,savefor that
infinitesimal fraction of spacewhich is occupiedby ordinary matter.
Descartesassumedthat the aether particles are continually in
motion. As howeverthere wasno empty spacefor moving particles
to moveinto, he inferred that they moveby taking the placesvacated
by other aether particles which are themselvesin motion. Thus the
movementof a single particle of the aether involved the motion of
an entire closedchain of particles; and the motionsof theseclosed
chains constituted vortices, which performed important functions in
his picture of the cosmos.
Holding, then, that the effectsproducedby meansof contactsand
collisions were the simplest and most intelligible phenomenain
the external world, Descartesadmitted no other agencies. He did
not claim for his schemean accurate quantitative agreementwith
experiment,in the spirit of a modernscientific textbook. He trusted
more to the clarity and distinctnessof a speculation than to its
consistencewith observedfact 2: rather is his work to be regardedas
a stupendouseffort of the imaginative intellect-a myth, in the same
senseas the Timaeus of Plato is a myth-intended to show that all
the contentsand happeningsof the universe might be exhibited as
parts of a logically co-ordinatedmechanicalscheme,dependingsolely
on very elementary types of physical action, and having (oncethe
premisseswereaccepted)the completecertitude of deductionthat had
hitherto been associatedonly with mathematics. Descarteswasthe
originator of the Mechanical Philosophy, Le. of the doctrine that the
externalinanimate world may for scientificpurposesberegardedasan
automatic mechanism,and that it is possibleand desirableto imagine
a mechanicalmodel of every physical phenomenon.
Such a conceptioncould not well haveoriginated earlierthan the
Renaissance,becausethe ancientsand the men of the Middle Ages
1 e.g.innubilis aether (cloudlesssky), Lucr. 3, 21
• That explainswhy he preferredhis own accountof the circulationof the blood to
Harvey's.
6
TO THE DEATH OF NEWfON
had little or no experienceof mechanismsthat were self-contained

and could work independently of human direction; they were
acquainted only with tools, which in order to achieve definite
purposes required intelligent guidance. Any manifestation of
regularity in performanceseemedto our forefathersto indicate, as
its directing cause,an activity of mind; indeed it was preciselythe
order and harmony observedin the movementsof the heavenly
bodiesthat led the Greekthinkersto regard them aspossessing souls.!
Movementssuch as the fall of a material body to the ground were
accountedfor, as we have seen,by supposingthat heavy matter
tended to seekits natural place, the centre of the universe. This
explanationbecameunsatisfyingwhen the Copernicantheory of the
solar systemwas accepted,since the earth was now in motion in
infinite space,and no point could be identified as the centre of the
universe. It was at this juncture that Descartesput forward his
revolutionary suggestion,that the cosmosmight be thought of as an
immense machine. This entailed the general principle that all
happeningsin the material world can be predicted by mathematical
calculation, and this principle has proved the element of greatest
value in the Cartesianphilosophyof nature.
Descarteshimselfseemsto havegonefurther, and to haveheld the
doctrine of epistemologicalrationalism, that is, the assertionthat physics
can, like Euclidean geometry, be derived entirely from a priori
principles,without any dependenceon observationand experiment.
At any rate, his generalpracticewas to representphenomenaas the
effectsof preconceiveddispositionsand causes. In this respecthe
departedfrom the sounddoctrinesthat had beenpreacheda genera-
tion earlier by Francis Baconand Galileo, and he drew down upon
himself the criticism of Huygens2: 'Descartes,' wrote Huygens,
, who seemedto me to be jealous of the fame of Galileo, had the
ambition to be regarded as the author of a new philosophy, to be
taught in academiesin place of Aristotelianism. He put forward
his conjecturesas verities, almost as if they could be proved by his
affirming them on oath. He ought to pave presentedhis systemof
physicsasan attempt to showwhat might be anticipated asprobable
1 I. may beremarkedthat modernmachines can simulateintelligencein a wonderful
way. For instance,longdivision,astaught to schoolchildren,requiresan act of intelligent
choice,in order to obtain eachnew digit in the quotient; but machinesnow existwhich
perform long division without any human guidance during the process: when the
machineis set, that is to say,furnishedwith the problemto be solved,then it is necessary
only to pressa button which switcheson an electriccurrent, and presentlythe required
quotientappearson a dial.
I It wasgivenin his annotationson Baillet'sLife of Descartes(a bookwhich appeared
in 1691),whichare printed in V. Cousin'sFragmentsphilosophiques,tomeii, p. 155.

7
THE AE1HER
in this science,when no principles but those of mechanicswere

admitted: this would indeedhavebeenpraiseworthy; but he went
further, and claimed to have revealed the precise truth, thereby
greatly impeding the discoveryof genuineknowledge.'
In putting forward an all-embracingtheory of the universebefore
he had studiedany of its processes in detail, Descarteswascontinuing
the tradition of the ancient Greeks,rather than treading in the new
paths struck out by Tycho, Kepler and Galileo: he never really
grasped the principle that true knowledge can only be acquired
piecemeal,by the patient interrogation of nature. A further weakness
in his systemwas involved in the assumptionthat force cannot be
communicatedexceptby actual pressureor impact, a principle which
compelledhim to provide an explicit mechanismin order to account
for eachof the known forcesof nature. This taskis evidentlymuch
more difficult than that which lies before thosewho are willing to
admit action at a distanceasan ultimate property of matter.
The many defectsof Descartes'method led to the rejection of
almost all his theoriesin lessthan a century. It must be said,how-
ever,that the grandeurof his plan, and the boldnessof its execution,
stimulated scientificthought to a degreebeforeunparalleled. 'Give
me matter and motion,' he cried, , and I will constructthe universe.'
Matter, in the Cartesian philosophy, is characterisednot by
impenetrability, or by any quality recognisableby the senses,
but simply by extension; extensionconstitutesmatter, and matter
constitutesspace. The basisof all things is a primitive, elementary,
unique type of matter, boundlessin extent and infinitely divisible.
In the processof evolution of the universethree distinct formsof this
matter have'originated, correspondingrespectivelyto the luminous
matter of the sun, the transparent matter of interplanetary space,
and the denseopaquematter of the earth. 'The first is constituted
by what has been scrapedoff the other particles of matter when
they were rounded; it moveswith so much velocity that when it
meetsother bodiesthe force of its agitation causesit to be broken
and divided by them into a heap of small particles that are of such
a figure as to fill exactly all the holes and small intersticeswhich
they find around thesebodies. The next type includesmost of the
rest of matter: its particles are spherical,and are very small com-
pared with the bodieswe seeon the earth; but neverthelessthey
have a finite magnitude, so that they can be divided into othersyet
smaller. There existsin addition a third type exemplifiedby some
kinds of matter-namely, thosewhich, on account of their sizeand
figure, cannot be soeasilymovedasthe preceding. I will endeavour
8
to show that all the bodies of the visible world are composed of
thesethree forms of matte., as of three distinct elements; in fact,
that the sun and the fixed stars are formed of the first :of these
elements,the interplanetary spacesof the second,and the earth, with
the planets and comets,of the third. For, seeingthat the sun and
the fixed stars emit light, the heavenstransmit it, and the earth,
the planets and the comets reflect it, it appears to me that there
is ground for using thesethree qualities of luminosity, transparence
and opacity, in order to distinguish the three elementsof the visible
world '.1
According to Descartes' theory, the sun is the centre of an
immense vortex formed of the first or subtlest kind of matter. I
The vehicle of light in interplanetary spaceis matter of the second
kind or element,composedof a closelypackedassemblageof globules
whosesize is intermediate between that of the vortex matter and
that of ponderable matter. The glol>ulesof the second element,
and all the matter of the first element, are constantly straining away
from the centresaround which they turn, owing to the centrifugal
force of the vortices 3 ; so that the globules are pressedin contact
with each other and tend to move outwards, although they do not
actually so move.' It is the transmission of this pressure which
constituteslight; the action of light therefore extends on all sides
round the sun and fixed stars, and travels instantaneously to any
distance.5 In the Dioptrique,8 vision is compared to the perception
of the presenceof objects which a blind man obtains by the use
of his stick; the transmissionof pressurealong the stick from the
object to the hand being analogousto the transmissionof pressure
from a luminous object to the eye by the secondkind of matter.
Descartessupposedthe 'diversities of colour and light' to be
due to the different ways in which the matter moves.7 In the
MItlores,8 the various colours are connected with different rotatory
velocities of the globules, the particles which rotate most rapidly
giving the sensationof red, the slower onesof yellow and the slowest
of green and blue-the order of colours being taken from the rain-
bow. The assertionof the dependenceof colour on periodic time
is a curious foreshadowingof a great discovery which was not fully
establisheduntil much later.
1 Principia, pt. iii, § 52
• It is curiousto speculateon the impressionwhich would have beenproducedhad
the spirality of nebulaebeendiscoveredbfo.fore the overthrow of the Cartesiantheory of
vortices.
• ibid. §§55-9 • ibid. § 63 • ibid. § 64 • Discours premier
, Principia, pt. iv, § 195 • Discours lwitiJnu
9
THE AETHER
The generalexplanationoflight on theseprincipleswasamplified

by a more particular discussionof reflection and refraction. The
law of reflection-that the anglesof incidence and reflection are
equal-had been known to the Greeks. Kepler in his Dioptrics
(1611) discussedrefraction through lenses; he had found experi-
mentally that for glass,when the incidenceis almostperpendicular,
the anglesof incidence and refraction are nearly in the ratio of
3 to 2; and assumingthis value, he had found the correct result
that the focus of a double convex lens, whosetwo faceshave the
samecurvature, is the centre of curvature of the side nearestthe
object. He did not, however, succeedin discoveringthe general
law of refraction-that the sinesof the angles of incidence and
refraction are to each other in a ratio dependingon the media-
which was now published for the first time.! Descartesgave it as
his own: but he seemsto have beenunder considerableobligations
to Willebrord Snell (1591-1676), Professorof Mathematics at
Leyden, who had discoveredit experimentally (though not in the
form in which Descartesgaveit) about 1621. Snell did not publish
his result, but communicatedit in manuscript to severalpersons,
and Huygens affirms 2 that this manuscript had been seen by
Descartes.
Descartespresentsthe law as a deduction from theory. This,
however,he was able to do only by the aid of analogy; when rays
meet ponderablebodies,, they are liable to be deflectedor stopped
in the sameway as the motion of a ball or a stoneimpinging on a
body': for' it is easyto believe that the action or inclination to
move, which I have said must be taken for light, ought to follow
in this the samelaws as motion.' 8 Thus he replaceslight, whose
velocity of propagation he believesto be alwaysinfinite, by a pro-
jectile whose velocity varies from one medium to another. The
law of refraction is then proved substantiallyasfollows' :
Supposethat a ray of light is refracted acrossa plane interface
from onemediuminto another. Let a light corpuscle,whosevelocity
in the first medium is Vi, be incident on the interface, making an
angle i with the normal to the interface, and let it be refracted at
an angle r into the second medium, in which its velocity is v,.
Descartesassumedthat the ratio of the velocitiesVi and v, depends
only on the nature of the media: (V,/Vi) = fL, say. He assumed
1 Diop/rique, Discou,s second

• cr. Cousin's Fragmentsphilosophiques,tome ii, p. 162
• Diop/rique, Discours premier
• ibid. Discours second
10
also that the component of velocity parallel to the interface 15

unaffectedby the refraction, sowe have
vjsini=vrsinr
Combining theseequations,we have
sin i = I-' sin ,.
which is the law of refraction.
Theseequationsimply that if i > ,.(e.g. if the refraction is from
air into glass)the velocity is greaterin the secondor densermedium.
As we shall see,this consequenceof the corpuscular theory in its
primitive form is in contradiction with experimentalfact.l
Crude though the Cartesiansystemwas in many features,there
is no doubt that by presenting definite mechanical conceptionsof
physical activity, and applying them to so wide a range of pheno-
mena,it stimulatedthe spirit of inquiry and in somedegreeprepared
the way for the more accuratetheoriesthat comeafter. In its own
day it met with great acceptance;a the confusionthat had resulted
from the destruction of the old order was now, as it seemed,
endedby a reconstructionof knowledgein a systemat oncecredible
and complete. Nor, as we shall seelater, did its influence quickly
wane.
So far as the theory of light was concerned, Descartes'con-
ceptions rapidly displaced those which had been current in the
Middle Ages. The validity of his explanation of refraction was,
however, called in question by his fellow-countryman, Pierre de
Fermat (1601-65),3 and a controversy followed, which was kept
1 Descartes' idea can however be modified so as to avoid erroneous consequences in
the following way. Instead of considering the velocities of the light-corpuscle in the two
media, consider its momenta, say Pi and pr (for a light-corpuscle, as will be seen later, the
momentum is not measured by the product of the mass and the velocity). Let us assume
that the ratio of pr to p, depends only on the nature of the media, and that the component
of momentum parallel to the interface is unaffected by the transition from one medium
to the other, then as in Descartes' proof we have
(Mpi) =~
P~ sin i = Pr sin r
and therefore
sini=~sinr
These equations are correct; the momentum is actually greater in the optically denser
medium.
• Though it appears from the Optica jlTomota of James Gregory, which was published
in 1663, and which contains an account of the reflecting telescope known by his name,
that Gregory had not heard of the discovery of the law of refraction, which had been given
in Descartes' Dioptrique twenty-five years earlier.
• Renati DescartesEpistolae, Pars tertia (Amsterdam, 1683). The Fermat correspondence
is comprised in letters xxix to xlvi.
II
THE AETHER
Up by the Cartesianslong after the death of their master. Fermat

eventually introduced a new fundamental law, from which he
proposedto deducethe pathsof rays oflight. This wasthe celebrated
Principle of Least Time, enunciated1 in the form, ' Nature alwaysacts
by the shortestcourse.' From it the law of reflection can readily
be derived, since the path describedby light betweena point on
the incident ray and a point on the reflected ray is the shortest
possible consistent with the condition of meeting the reflecting
surfaces.' In order to obtain the law of refraction, Fermat assumed
that 'the resistanceof the media is different,' and applied his
, method of maxima and minima' to find the path which would be
describedin the least time from a point of one medium to a point
of the other. In 1661 he arrived at the solution.s 'The result of
my work,' he writes, 'has been the most extraordinary, the most
unforeseen,and the happiest, that ever was; for, after having
performed all the equations, multiplications, antitheses,and other
operationsof my method, and having finally finished the problem,
I have found that my principle gives exactly and precisely the
sameproportion for the refractions which Monsieur Descarteshas
established.' His surprisewas all the greater, as he had supposed
light to move more slowly in densethan in rare media, whereas
Descarteshad (as we have seen)beenobliged to makethe contrary
supposition.
Although Fermat's result was correct, and, indeed, of high
permanent interest, the principles from which it was derived were
metaphysical rather than physical in character, and consequently
were of little usefor the purposeof framing a mechanicalexplana-
tion of light. The influence of Descartes'theory was therefore
scarcelyat all diminished as a result of Fermat's work. But a new
attack soon opened from another quarter: nothing less, indeed,
than the rise of a new schoolof philosophy,hostile to Aristotelian-
ism and to Cartesianism alike. Its author was Pierre Gassendi
(1592-1655), Professorat the Collegede France in Paris,a follower
1 Epist. xlii, written at Toulouse in August 1657to Monsieur de la Chambre;

reprinted in (EuVTesde Fmnat (ed. 1891),ii, p. 354.
• That reflectedlight follows the shortestpath was no new result, for it had been
affirmed in a Commentaryon Aristotle's Meteor, iii, by Olympiodorw, who lived at
Alexandriain the middle of the sixth century A.D. : and alsoin the K€~aw. TWV mCKwv
of Damianus,a pupil of Heliodorusof Larissa,a work of which severaleditionswere
publishedin the seventeenthcentury. Damianusgivesin his first bookthe opticalwork
of Hero of Alexandria,to whom he attributesthe theoremregardingthe path of reflected
light.
• Epist. xliii, written at Toulouseon I Jan. 1662; reprinted in (EuVTesde Fmnat, ii,
p. 457; i, pp. 170, 173
12
of Copernicusand Galileo, who re-introduced the doctrine of the

ancient atomists,namely that the universe is formed of material
atoms,eternal and unchangeable,moving about in a spacewhich
exceptfor them is empty. The most formidable objection to the
new teaching was that in the Graeco-Romanworld it had been
associatedwith the moral and theological views of Epicurus and
Lucretius. Gassendi,who was himself a priest, worked hard to
showthat it had no necessaryconnection with religious error, and
that it could be acceptedas a basisfor physicsby Christian men.
After a sharp controversywith Descartesin 1641-6, he had such
successthat, as we shall see, his doctrine was accepted not long
afterwardsby Newton, and in fact becamethe departure point for
all subsequentnatural philosophy.
One great advantageof it was, that sincespacewas not now a
plenum,therewasno needfor moving particlesto form closedchains,
and the Cartesianvorticescould be done away with. As a matter
of fact, most of the later developmentsof the theory postulated an
aether (or sometimesmore than one), and so in a sensereverted to
the idea of spaceas a plenum; but theseaetherswere conceived
as somethingwhich offered no resistanceto the motion of ordinary
matter through them, so that material particles could be treated
asif they were situatedin a vacuum.
The next event of importance in the theory of light was the
publication in 16671 of the Micrographia of Robert Hooke
(1635-1703).2 Hooke, who had begun his scientific career as the
assistantof Robert Boyle, and who was ultimately secretaryof the
Royal Society,madetwo experimentaldiscoverieswhich concernour
presentsubject; but in both of these,as it appeared,he had been
anticipated. The first 8 wasthe observationof the iridescentcolours
which are seenwhen light falls on a thin layer of air betweentwo
glassplatesor lenses,or on a thin film of any transparentsubstance.
Thesearegenerallyknownasthe' coloursof thin plates,'or' Newton's
rings'; they had been previously observedby Boyle." Hooke's
secondexperimental discovery,5 made after the date of the Micro-
graphia, was that light in air is not propagated exactly in straight
lines, but that there is some illumination within the geometrical
shadowof an opaque body. This observationhad been published
in 1665 in a posthumouswork 1 of Fr. FrancescoMaria Grimaldi
1 The imprimatur of Viscount Brouncker,P.R.S., is dated 23 Nov. 1664.

I On Hooke, cf. the excellent Wilkins Lecture by E. N. da C. Andrade, printed in
Proc. R.S., cci (1950),p. 439 I Micrographia, p. 47
• Boyle'sWorks (ed. 1772),i, p. 742 • Hooke'sPosthumousWorks, p. 186

13
THE AETHER
(1613-63), a Jesuit, who had given to the phenomenonthe name

diffraction.
Hooke's theoretical investigations on light were of great
importance,representingasthey do the transition from the Cartesian
systemto the fully developedtheory of waves. He beginsby attack-
ing Descartes'proposition, that light is a tendencyto motion rather
than an actual motion. 'There is,' he observes,2'no luminous
Body but has the parts of it in motion more or less'; and this
motion is ' exceedingquick.' Moreover, sincesomebodies (e.g.the
diamond when rubbed or heatedin the dark) shinefor a considerable
time without being wasted away, it follows that whatever is in
motion is not permanently lost to the body, and therefore that the
motion must be of a to-and-fro or vibratory character. The ampli-
tude of the vibrations mustbe exceedinglysmall,sincesomeluminous
bodies (e.g. the diamond again) are very hard, and so cannot yield
or bend to any sensibleextent.
Concluding, then, that the condition associatedwith the emission
of light by a luminous body is a rapid vibratory motion of very
small amplitude, Hooke next inquires how light travels through
space. 'The next thing we are to consider,' he says' is the way
or manner of the trajection of this motion through the interpos'd
pellucid body to the eye: And here it will be easilygranted:
'First, that it must be a body susceptible and impartible of this
motion that will deservethe name of a Transparent; and next,
that the parts of sucha body must be homogeneous,or of the samekind.
'Thirdly, that the constitution and motion of the parts must
be suchthat the appulseof the luminousbody may be communicated
or propagated through it to the greatestimaginable distancein the
least imaginable time, though I seeno reasonto affirm that it must
be in an instant.
, Fourthly, that the motion is propagatedevery way through an
Homogeneousmedium by direct or straight lines extendedevery way like
Rays from the centre of a Sphere.
'Fifthly, in an Homogeneous medium this motion is propagated
every way with equal velocity, whence necessarilyevery pulse or
vibration of the luminous body will generate a Sphere, which will
continually increase,and grow bigger, just after the samemanner
(though indefinitely swifter) as the waves or rings on the surface
of the water do swell into bigger and bigger circles about a point
of it, where by the sinking of a Stonethe motion wasbegun,whence
1 Plrysico-Mathesis de lumine, coloribus, et iride (Bologna, 1665), i, prop. i
• Micrographia, p. 55
14
it necessarilyfollows, that all the parts of theseSpheresundulated

through an Homogeneousmedium cut the Rays at right angles.'
Here we have a fairly definite mechanical conception. It
resemblesthat of Descartesin postulating a medium as the vehicle
of light; but accordingto the Cartesianhypothesisthe disturbance
is a statical pressurein this medium, while in Hooke's theory it is a
rapid vibratory motion of small amplitude. In the above extract
Hooke introduces, moreover,the idea of the wave-front, or locus at
any instant of a disturbance generatedoriginally at a point, and
affirms that it is a sphere, whose centre is the point in question,
and whoseradii are the rays of light issuingfrom the point.
Hooke's next effort was to produce a mechanical theory of
refraction, to replace that given by Descartes. 'Because,' he says,
, all transparentmediums are not Homogeneous to one another, there-
fore we will next examinehow this pulse or motion will be propa-
gatedthrough differingly transparentmediums. And here, according
to the mostacute and excellent Philosopher Des Cartes, I suppose
the sine of the angle of inclination in the first medium to be to the
sine of refraction in the second,as the density of the first to the
densityof the second. By density, I mean not the density in respect
of gravity (with which the refractions or transparencyof mediums
hold no proportion), but in respectonly to the trajection of the Rays
of light, in which respect they only differ in this, that the one
propagates the pulse more easily and weakly, the other more
slowly, but more strongly. But as for the pulses themselves,they
will by the refraction acquire anotherproperty, which we shall now
endeavourto explicate.
, We will suppose,therefore, in the first Figure, ACFD to be a
physical Ray, or ABC and DEF to be two mathematical Rays,
o
A
L L
0 N
M
M
K
15
THE AETHER
trajected from a very remote point of a luminous body through an

Homogeneous transparentmedium LL, and DA, EB, FC, to be small
portions of the orbicular impulses which must therefore cut the
Rays at right angles,these Rays meeting with the plain surface
NO of a medium that yields an easiertransitus to the propagationof
light, and falling obliquely on it, they will in the medium MM be
refracted towards the perpendicular of the surface. And because
this medium is more easily trajected than the former by a third,
thereforethe point C of the orbicular pulseFC will be movedto H
four spacesin the sametime that F, the other end of it, is movedto
three spaces,therefore the whole refracted pulse to H shall be
oblique to the refractedRays CHK and GI.'
Although this is not in all respectssuccessful,it representsa
decidedadvanceon tIie treatment of the sameproblem by Descartes,
which rested on a mere analogy. Hooke tries to determine what
happensto the wave-front when it meetsthe interface betweentwo
media; and for this end he introducesthe correct principle that the
sideof the wave-frontwhich first meetsthe interfacewill go forward
in the secondmedium with the velocity proper to that medium,
while the other side of the wave-front, which is still in the first
medium, is still moving with the old velocity; so that the wave-
front will be deflectedin the transition from one medium to the
other.
This deflection of the wave-front was supposedby Hooke to
be the origin of the prismatic colours. He regarded natural or
white light as the simplest type of disturbance, being constituted
by a simple and uniform pulse at right anglesto the direction of
propagation, and inferred that colour is generatedby the distortion
to which this disturbanceis subjectedin the processof refraction.
, The Ray,' 1 he says,, is dispersed,split, and openedby its Refrac-
tion at the Superficiesof a secondmedium, and from a line is opened
into a diverging Superficies,and soobliquated,wherebythe appear-
ancesof Coloursare produced.' 'Colour,' he saysin anotherplace,J
, is nothing but the disturbance of light by the communication of
the pulse to other transparent mediums, that is by the refraction
thereof.' His precise hypothesis regarding the different colours
was S ' that Blue is an impressionon the Retinia of an oblique and
confus'd pulse of light, whose weakestpart precedes,and whose
strongestfollows. And, that Red is an impressionon the Retina of
1 Hooke's PosthumousWorks, p. 82
• To the Royal Society, 15 February 1671-72 • Micrographia, p. 64

J6
an oblique and confus'dpulseoflight, whosestrongestpart precedes,

and whoseweakestfollows.'
Hooke's theory of colour was completely overthrown, within
a few yearsof its publication, by one of the earliest discoveriesof
Isaac Newton (1642-1727). Newton, at that time a scholar of
Trinity College,Cambridge, had in the beginning of 1666 obtained
a triangular prism, 'to try therewith the celebrated Pha~nomena
of Colours.' For this purpose,, having darkenedmy chamber, and
madea smallholein my window-shuts,to let in a convenientquantity
of the Sun'slight, I placed my Prismeat his entrance,that it might
be thereby refracted to the opposite wall. It was at first a very
pleasingdivertisement,to view the vivid and intensecoloursproduced
thereby; but after a while applying myself to consider them more
circumspectly,I becamesurprised to seethem in an oblong form,
which, according to the received laws of Refraction, I expected
shouldhavebeencircular.' The length of the colouredspectrumwas
in fact about five times asgreat asits breadth.
This puzzling fact he set himself to study; and after more
experiments the true explanation was discovered-namely, that
ordinary white light is really a mixture 1 of rays of every variety of
colour,and that the elongationof the spectrumis dueto thedifferences
in the refractive power of the glassfor thesedifferent rays.
, Amidst thesethoughts,' he tells US,2 , I was forced from Cam-
bridge by the intervening Plague'; this was in 1666, and his
memoir on the subjectwas not presentedto the Royal Societyuntil
five years later. In it he propounds a theory of colour directly
opposedto that of Hooke. 'Colours,' he says,, are not Qualifications
of light derived from Refractions, or Reflections of natural Bodies
(as 'tis generally believed), but Original and connateproperties, which
in divers Rays are divers. SomeRays are disposedto exhibit a red
colour and no other: somea yellow and no other, somea greenand
no other, and so of the rest. Nor are there only Rays proper and
particular to the more eminent colours, but even to all their inter-
mediategradations.
'To the same degree of Refrangibility ever belongs the same
1 The word mixture must not be taken to imply that the rays of different colours, when
compounded together, preserve their separate existence and identity unaltered within the
compound, like the constituents of a mechanical mixture. On the contrary, as was shown
by Gouy in 1886 (Journal de Physique, (.) v, p. 354), natural white light is to be pictured,
in the undulatory representation, as a successionof short pulses, out of which any spectro-
scopic apparatus such as a prism manufactures the different monochromatic rays, by a
process which is physically equivalent to the mathematical resolution of an arbitrary
function into periodic terms by Fourier's integral theorem.
• Phil. Trans. vi (19 February 1671-2), p. 3075
17
THE AETHER
colour, and to the same colour ever belongs the same degree of
Refrangibility.
, The speciesof colour, and degreeof Refrangibility proper to
any particular sort of Rays, is not mutable by Refraction, nor by
Reflection from natural bodies,nor by any other cause,that I could
yet observe. When anyone sort of Rays hath beenwell parted from
thoseof other kinds, it hath afterwardsobstinatelyretained its colour,
notwithstanding my utmost endeavoursto changeit.'
The publication of the new theory gave rise to an acute con-
troversy.I As might be expected,Hooke was foremost among the
opponents,and led the attack with somedegreeof asperity. When
it is rememberedthat at this time Newton was at the outset of his
career, while Hooke was an older man, with an establishedreputa-
tion, suchharshnessappearsparticularly ungenerous; and it is likely
that the unpleasantconsequences which followed the announcement
of his first great discoveryhad much to do with the reluctancewhich
Newton ever afterwardsshowedto publish his resultsto the world.
In the courseof the discussionNewton found occasionto explain
more fully the views which he entertained regarding the nature of
light. Hooke charged him with holding the doctrine that light is a
material substance. Now Newton had, as a matter of fact, a great
dislike of the more imaginative kind of hypotheses; he altogether
renouncedthe attempt to construct the universefrom its foundations
after the fashion of Descartes,and aspired to nothing more than a
formulation of the lawswhich directly governthe actual phenomena.
His desire in regard to optics was to present a theory free from
speculation as to the hidden mechanismof light. Accordingly, in
reply to Hooke's criticism, he protestedI that his views on colour
were in no way bound up with any particular conception of the
ultimate nature of optical processes.
Newton was,however,unable to carry out his plan of connecting
together the phenomenaof light into a coherentand reasonedwhole
without having recourseto hypotheses. The hypothesisof Hooke,
that light consistsin vibrations of an aether, he rejectedfor reasons
which at that time were perfectly cogent,and which indeedwerenot
successfullyrefuted for over a century. One of thesewas the in-
competenceof the wave-theoryto account for the rectilinear propa-
gation of light, and another wasits inability to embracethe facts-
1 The astronomer John Flamsteed explained his objections to it in a letter to Collin.
of date 17 April 1672, printed in Rigaud's Correspondence

of scientific menof the XVIlth century
ii, p. 134.
• Phil. Trans. vii (1672), p. 5086
18
discovered,asweshallpresentlysee,by Huygens,and first interpreted

correctly by Newton himself-of polarisation. On the whole, he
seemsto havefavoureda schemeof which the following may be taken
asa summary 1 :
All spaceis permeatedby an elastic medium or aether, which is
capableof propagatingvibrations in the sameway asthe air propa-
gatesthe vibrations of sound,but with far greater velocity.
This aether pervadesthe poresof all material bodies,and is the
causeof their cohesion; its densityvariesfrom one body to another,
beinggreatestin the free interplanetary spaces. It is not necessarily
a singleuniform substance: but just asair containsaqueousvapour,
so the aether may contain various ' aethereal spirits,' adapted to
producethe phenomenaof electricity, magnetismand gravitation.
The vibrations of the aether cannot, for the reasonsalready
mentioned,be supposedin themselvesto constitute light. Light is
thereforetakento be ' somethingof a different kind, propagatedfrom
lucid bodies. They, that will, may supposeit an aggregateof various
peripatetic qualities. Others may supposeit multitudes of un-
imaginablesmall and swift corpusclesof varioussizes,springing from
shining bodiesat great distancesone after another; but yet without
any sensibleinterval of time, and continually urged forward by a
principle of motion, which in the beginning acceleratesthem, till
the resistanceof the aethereal medium equals the force of that
principle, much after the manner that bodies let fall in water are
acceleratedtill the resistanceof the water equalsthe force of gravity.
But they, that like not this, may supposelight any other corporeal
emanation, or any impulse or motion of any other medium or
aetherealspirit diffused through the main body of aether, or what
elsethey can imagine proper for this purpose. To avoid dispute,
and makethis hypothesisgeneral,let everyman here take his fancy ;
only whateverlight be, I supposeit consistsof rays differing from one
another in contingent circumstances,as bigness,form or vigour.' 2
In any case,light and aether are capableof mutual interaction;
aether is in fact the intermediary between light and ponderable
matter. When a ray of light meetsa stratum of aether denseror
1 cr. Newton's memoir in Phil. Trans. vii (1672); his memoir presented to the Royal
Society in December 1675, which is printed in T. Birch's History of the Royal Society of
London (4 vols., 1756-7), iii, p. 247; his Opticks, especially queries 18, 19,20,21,23,29 ;
the Scholium at the end of the Principia; and a letter to Boyle, written in February 1678-9,
which is printed in Horsley's Newtoni Opera,p. 385, and in Rigaud's Correspondence, ii, p. 407.
In the Principia, bk. i, § xiv, the analogy between rays of light and streams of
corpuscles is indicated; but Newton does not commit himself to any theory of light
based on this.
• Royal Society, 9 Dec. 1675 IE-,rch iii p. 255)
(8"23) 19
THE AETHER
rarer than that through which it has lately been passing,it is, in
general, deflected from its rectilinear course; and differencesof
density of the aether between one material medium and another
accounton theseprinciples for the reflection and refraction oflight.1
The condensationor rarefaction of the aetherdue to a material body
extendsto somelittle distancefrom the surfaceof the body, so that
the inflection due to it is really continuous, and not abrupt; and
this further explains diffraction, which Newton took to be ' only a
new kind of refraction, caused,perhaps, by the external aether's
beginning to grow rarer a little before it came at the opakebody,
than it was in free spaces.'
Although the regular vibrations of Newton's aether were not
supposedto constitute light, its irregular turbulence seemsto have
representedfairly closelyhis conceptionof heat. He supposedthat
when light is absorbedby a material body, vibrations are set up in
the aether,and are recognisableasthe heat which is alwaysgenerated
in suchcases. The conduction of heat from hot bodiesto contiguous
cold ones he conceivedto be effected by vibrations of the aether
propagated between them; and he supposedthat it is the violent
agitation of aetherealmotions which excitesincandescentsubstances
to emit light.
Assuming with Newton that light is not actually constitutedby
the vibrations of an aether, even though such vibrations may exist
in closeconnection with it, the most definite and easily conceived
suppositionis that rays of light are streamsof corpusclesemitted by
luminous bodies. Although this wasnot the hypothesisof Descartes
himself, it was so thoroughly akin to his general schemethat the
scientific men of Newton's generation, who were for the most part
deeply imbued with the Cartesianphilosophy, instinctively selected
it from the wide choice of hypotheseswhich Newton had offered
them; and by later writers it wasgenerallyassociatedwith Newton's
name. A curiousargument in its favour wasdrawn from a phenom-
enon which had then been known for nearly half a century:
Vincenzo Cascariolo,a shoemakerof Bologna,had discovered,about
1630, that a substance,which afterwardsreceivedthe nameof Bologna
stone or Bologna phosphorus, has the property of shining in the dark
after it hasbeenexposedfor sometime to sunlight; and the storage
of light which seemedto be hereinvolved wasm0r.:eeasilyexplicable
on the corpuscular theory than on any other. The evidencein this
quarter, however, pointed the other way when it was found that
1 Newton's proof of the law of refraction, in Opticks, i, prop. 6, doesnot differ greatly
in prindple from Descartes'proof.

20
phosphorescentsubstancesdo not necessarilyemit the samekind of

light as that which wasusedto stimulate them.
In accordancewith his earliest discovery, Newton considered
colour to be an inherent characteristic of light, and inferred that it
must be associatedwith somedefinite quality of the corpusclesor
aethervibrations. The corpusclescorrespondingto different colours
would, he remarked, like sonorousbodies of different pitch, excite
vibrations of different types in the aether; and' if by any means
those[aether vibrations] of unequal bignessesbe separatedfrom one
another, the largest beget a Sensationof a Red colour, the least or
shortestof a deep Violet, and the intermediate ones,of intermediate
colours; much after the manner that bodies, according to their
severalsizes,shapesand motions, excite vibrations in the Air of
various bignesses,which, according to thosebignesses,make several
Tonesin Sound.' 1
This sentenceis the first enunciation of the great principle that
homogeneouslight is essentiallyperiodic in its nature, and that
differencesof period correspondto differencesof colour. The obvious
analogy with sound was expresslymentioned by Huygens in his
Trait! dela lumiere, of which morehereafter2 : and it may beremarked
in passingthat Newton's theory of periodic vibrations in an elastic
medium, which he developed3 in connection with the explanation
of sound,would alone entitle him to a place among thosewho have
exercisedthe greatestinfluence on the theory of light, evenif he had
madeno direct contribution to the latter subject.
Newton devoted considerable attention to the colours of thin
plates, and determined the empirical laws of the phenomena with
great accuracy. In order to explain them, he supposedthat' every
ray of light, in its passagethrough any refracting surface,is put into
a certain transient constitution or state, which, in the progressof
the ray, returns at equal intervals, and disposesthe ray, at every
return, to be easily transmitted through the next refracting surface,
and, betweenthe returns, to be easilyreflectedby it." The interval
between two consecutive dispositions to easy transmission or
, length of fit' varies, as he found, with the colour, being greatest
for red light and leastfor violet. If then a ray of homogeneouslight
falls on a thin plate, its fortunesasregardstransmissionand reflection
at the two surfaceswill depend on the relation which the length of
1 Phil. Trans. vii (1672), p. 5088
I The hypotheses that the colour of monochroI:latic light depends on the period of
vIbration, and that the brightness increaseswith the amplitude of the vibration, seem to
have been first clearly enunciated in 1699 by N. de Malebranche (1638-1715).
• Newton's PrilUipia, ii, props. xliii-1 • Opticks, ii, prop. 12
21
THE AETHER
fit bears to the thicknessof the plate; and on this basishe built
up a theory of the coloursof thin plates. It is evidentthat Newton's
, length of fit' correspondsin somemeasureto the quantity which
in the undulatory theory is called the wave-length of the light;
but the theory of fits of easy transmission and ea.ryreflection, though it
served to explain successfullyphenomenawhich were not known
until after Newton's time, was abandonedafter the triumph of thf"
wave-theoryin the nineteenthcentury.!
Newton'stheory oflight led to a correspondence with Fr. Ignace-
Gaston Pardies, S.J. (163~73),2 Professorof Mathematics in the
College de Clermont (later known as the College Louis-Ie-Grand)
at Paris. As a result, Pardies,who was a Cartesian,seemsto have
become convinced of the truth of Newton's doctrine. He was,
however,strongly in favour of the wave-hypothesis(which wasnot
irreconcilablewith it), and beforehis early deathwrote a dissertation
on wave-motion which seemsto have influenced Huygens, who
saw it in Paris. Pardies adumbrated the notion of aberration,
which was not discovereduntil more than half a century later, and
he rejectedDescartes'suppositionof an infinite velocity for light.
At the time of the publication of Hooke's Micrographia, and
Newton's theory of colours, it was not known whether light is
propagatedinstantaneouslyor not. An attempt to settle the ques-
tion experimentally had been made many years previously by
Galileo,3 who had stationedtwo men with lanternsat a considerable
distance from each other; one of them was directed to observe
when the other uncoveredhis light, and exhibit his own the moment
he perceivedit. But the interval of time required by the light for
its journey was too small to be perceived in this way; and the
discovery was ultimately made by an astronomer. In 1675 Olaf
Roemer' (1644-1710), a Dane who was at the time resident in
Parisasprofessorof mathematicsto the Dauphin of France,observed
that the eclipsesof the first satellite of Jupiter were apparently
affectedby an unknowndisturbing cause; the time of the occurrence
of the phenomenonwasretarded when the earth andJupiter, in the
courseof their orbital motions, happenedto be most remote from
1 It 'Was however a remarkable anticipation of the twentieth-century quantum-
theory explanation : the" fits of easy transmission and easy reflection" correspond to
the transition probabilities of the quantum theory.
• Two letters from Pardies are printed in Phil. Trans. vii (1672), pp. 4087, 5012,
followed by Newton's comments.
• Discorsi e dimostrazioni matematiche,p. 43 of the Elzevir edition of 1638
• Mem. de ['Acad. x (1666-99), p. 575. Roemer, who was afterwards appointed
Director of the Copenhagen Observatory, was the inventor of many astronomical instru-
ments, notably the equatorial and the transit instrument.
22
each other, and was acceleratedin the contrary case. Roemer

explainedthis by supposingthat light requires a finite time for its
propagationfrom the satelliteto the earth; and by observationsof
eclipses,he calculated the interval required for its passagefrom
the sunto the earth (the light equation asit is called) to be 11minutes.!
Shortly after Roemer's discoverythe wave-theoryof light was
greatly improved and extendedby Christiaan Huygens(1629-95).
Huygens,who at the time was living in Paris, communicated his
results in 1678 to Cassini, Roemer, De la Hire and the other
physicistsof the French Academy, and prepared a manuscript of
considerablelength on the subject. This he proposedto translate
into Latin, and to publish in that languagetogetherwith a treatise
on the optics of telescopes; but the work of translation making
little progress,after a delay of twelve years he decided to print
the work on wave-theoryin its original form. In 1690 it appeared
at Leyden,Z under the title Traiti de la lumiere OU sont expliqudes les
causesde ce qui luy arrive dans la rijlexion et dans la rifraction. Et parti-
culierementdans l'dtrange rifraction du cristal d'Islande. Par C.H.D.Z.3
The truth of Hooke'shypothesis,that light is essentiallya form
of motion, seemedto Huygensto be proved by the effectsobserved
with burning-glasses; for in the combustioninduced at the focus
of the glass,the moleculesof bodiesare dissociated; which, as he
remarked, must be taken as a certain sign of motion, if, in con-
formity to the Cartesianphilosophy,we seekthe causeof all natural
phenomenain purely mechanicalactions.4.
The questionthen arisesas to whether the motion is that of a
medium, as is supposedin Hooke's theory, or whether it may be
comparedto that of a flight of arrows, asin the corpusculartheory.
Huygens decided that the former alternative is the only tenable
1 It was soon recognised that Roemer's value was too large; and the astronomers of
the succ('~dinghalf-century reduced it to 7 minutes. Delambre, by an investigation whose
details appear to have been completely destroyed, published in 1817 the value 493'2s,
from a discussionof eclipsesof Jupiter's satellites during the previous 150years. Glasenapp,
in an inaugural dissertation published in 1875, discussed the eclipses of the first satellite
between 1848 and 1870, and derived, by different assumptions, values between 496s and
SOl", the most probable value being 500·8s. Sampson, in 1909, derived 498·640 from his
own readings of the Harvard Observations, and 498·795from the Harvard readings with
probable errors of about ± 0.029• The inequalities of Jupiter's surface give rise to some
difficulty in exact determinations.
• Huygens had by this time returned to Holland.
• i.e. Christiaan Huygens de Zuylichem. The custom of indicating names by initials
was not unusual in that age. English translation by S. P. Thompson, London, 1912.
• It would seem from Berkeley's First Dialogue between Hylas and Philonom that
this opinion was favoured in England in 1713. 'Hylas: I tell you, Philonous, external
light is nothing but a thin fluid substance, whose minute particles, being agitated with a
brisk motion, communicate different motions to the optic nerves.'
23
THE AKTHER
one, since beamsof light proceedingin directions inclined to each

other do not interfere with eachother in any way.
Moreover, it had previously been shownby Torricelli that light
is transmitted as readily through a vacuum as through air; and
from this Huygens inferred that the medium or aether in which
the propagation takes place must penetrate all matter, and be
present even in all so-called vacua. Light, therefore, consistsof
disturbances,propagated with great velocity, in a highly elastic
medium composedof very subtle matter.
The processof wave-propagationhe discussedby aid ofa principle
which was now 1 introduced for the first time, and has sincebeen
generally known by his name. It may be stated thus: considera
wave-front,2 or locus of disturbance,as it existsat a definite instant
to; then each surface-elementof the wave-front may be regarded
asthe sourceof a secondarywave, which in a homogeneousisotropic
medium will be propagated outwards from the surface-elementin
the form of a spherewhose radius at any subsequentinstant t is
proportional to (I-to); and the wave-front which representsthe
whole disturbance at the instant t is simply the envelopeof the
secondarywaves which arise from the various surface-elementsof
the original wave-front. The introduction of this principle enabled
Huygensto succeedwhere Hooke and Pardieshad failed, in achiev-
ing the explanation of refraction and reflection. His method was
to combine his own principle with Hooke's device of following
separately the fortunes of the right-hand and left-hand sidesof a
wave-front when it reachesthe interface betweentwo media. The
actual explanation for the caseof reflection is asfollows :
Let AB represent the interface at which reflection takes place,
AHC the incident wave-front at an instant to, GMB the position
which the wave-front would occupy at a later instant t if the
propagationwere not interrupted by reflection. Then by Huygens'
principle the secondarywave from A is at the instant t a sphere
RNS of radius equal to AG; the disturbancefrom H, after meeting
the interface at K, will generate a secondarywave TV of radius
equal to KM, and similarly the secondarywave correspondingto
any other element of the original wave-front can be found. It is
obvious that the envelope of these secondarywaves, which con-
stitutes the· final wave-front, will be a plane BN, which will be
1 Traitl de la lurn., p. 17
I It may be remarked that Huygens' • waves' are really what modern writers, follow-
ing Hooke, call' pulses' ; Huygens never considered true wave-trains having the property
of periodicity.
inclined to AB at the same angle as AC. This gives the law of

reflection.
The law of refraction is establishedby similar reasoning,or the

supposition that the velocity of light dependson the medi~~ in
which it is propagated. Since a ray which passesfrom air to glass
is bent inwards towards the normal, it may be inferred that light
travels more slowly in glassthan in air.
Huygens offered a physical explanation of the variation in
velocity of light from one medium to another, by supposingthat
transparent bodiesconsistof hard particles which interact with the
aetherealmatter, modifying its elasticity. The opacity of metalshe
explained by an extension of the same idea, supposingthat some
of the particles of metals are hard (these account for reflection)
and the rest soft; the latter destroy the luminous motion by damp-
ing it.
The secondhalf of the Traiti de La Lumiere is concernedwith a
phenomenonwhich had been discovereda few years previously by
a Danish philosopher, ErasmusBartholin (1625-98). A sailor had
brought from Iceland to Copenhagena number of beautiful crystals
which he had collected in the Bay of Roerford. Bartholin, into
whosehands they passed,noticed 1 that any small object viewed
through one of these crystals appeared double, and found the
immediate causeof this in the fact that a ray of light entering the
crystal gaverise in generalto two refracted rays. One of theserays
wassubjectto the ordinary law of refraction, while the other, which
was called the extraordinary ray, obeyed a different law, which
Bartholin did not succeedin determining.
1 Experimenla crista/Ii lslandici disdiadastici, 1669
25
THE AETHER
The matter had arrived at this stagewhen it was taken up by

Huygens. Since in his conception each ray of light correspondstl)
the propagation of a wave-front, the two rays in Iceland spar must
correspondto two different wave-frontspropagatedsimultaneously.
In this idea he found no difficulty; ashe says: 'It is certain that
a spaceoccupied by more than one kind of matter may permit the
propagation of severalkinds of waves,different in velocity; for this
actually happensin air mixed with aetherealmatter, where sound-
wavesand light-wavesare propagatedtogether.'
Accordingly he supposedthat a light disturbancegeneratedat
any spot within a crystal of Iceland spar spreadsout in the form of
a wave-surface,composedof a sphere and a spheroid having the
origin of disturbanceascentre. The sphericalwave-frontcorresponds
to the ordinary ray, and the spheroid to the extraordinary ray;
and the direction in which the extraordinary ray is refractedmay be
determinedby a geometrical construc-
tion, in which the spheroid takesthe
place which in the ordinary construc-
tion is taken by the sphere.
A T Thus, let the plane of the figure be
at right angles to the intersection of
the wave-front with the surfaceof the
crystal; let AB representthe trace of
the incident wave-front; and suppose
that in unit time the disturbancefrom
B reachesthe interface at T. In this unit interval of time the
disturbancefrom A will have spread out within the crystal into a
sphere and spheroid; so the wave-front corresponding to the
ordinary ray will be the tangent plane to the sphere through the
line whose trace is T, while the wave-front correspondingto the
extraordinary ray will be the tangent plane to the spheroidthrough
the sameline. The points of contact Nand M will determinethe
directionsAN and AM of the two refracted rays1 within the crystal.
Huygens did not in the Theorie de la lumiere attempt a detailed
physical explanation of the spheroidal wave, but communicated
one later in a letter to Papin,2written in December 1690. 'As to
the kinds of matter containedin Iceland crystal,' he says,, I suppose
one composedof small spheroids,and another which occupiesthe
1 The word ray in the wave-theory is always applied to the line which goes from the
centre of a wave (i.e. the origin of the disturbance) to a point on its surface, whatever may
be the inclination of this line to the surface-element on which it abuts; for this line has
the optical properties of the ' rays' of the emission theory.
• Huygens' CEuvres (ed. 1905), x, p. 177
26
interspacesaround thesespheroids,and which servesto bind them

together. Besidesthese,there is the matter of aether permeating
all the crystal, both between and within the parcels of the two
kinds of matter, just mentioned; for I supposeboth the little
spheroids,and the matter which occupiesthe intervals around them,
to be composedof small fixed particles, amongstwhich are diffused
in perpetual motion the still finer particles of the aether. There
is now no reasonwhy the ordinary ray in the crystal should not be
due to wavespropagatedin this aetherealmatter. To account for
the extraordinary refraction, I conceive another kind of waves,
which havefor vehicle both the aetherealmatter and the two other
kinds of matter constituting the crystal. Of theselatter, I suppose
that the matter of the small spheroidstransmits the wavesa little
more quickly than the aethereal matter, while that around the
spheroidstransmits thesewavesa little more slowly than the same
aetherealmatter.... Thesesamewaves,when they travel in the
direction of the breadth of the spheroids,meet with more of the
matter of the spheroids,or at least passwith lessobstruction, and
soarepropagateda little morequickly in this sensethan in the other;
thus the light disturbanceis propagatedasa spheroidalsheet.'
Huygensmade another discovery1 of capital importance when
experimentingwith the Iceland crystal. He observedthat the two
rays which are obtained by the double refraction of a single ray
afterwardsbehavein a way different from ordinary light which has
not experienceddoublerefraction; and in particular, if one of these
rays is incident on a secondcrystal of Iceland spar, it givesrise in
somecircumstancesto two, and in othersto only one, refractedray.
The behaviour of the ray at this secondrefraction can be altered
by simply rotating the secondcrystal about the direction of the ray
asaxis; the ray undergoingthe ordinary or extraordinary refraction
accordingasthe principal sectionof the crystalisin a certain direction
or in the direction at right anglesto this.
The first stagein the explanation of Huygens' observationwas
reachedby Newton, who in 1717showed2 that a ray obtained by
doublerefractiondiffersfrom a ray of ordinary light in the sameway
that a long rod whosecross-section is a rectanglediffers from a long
rod whosecross-sectionis a circle; in other words, the properties
of a ray of ordinary light are the samewith respectto all directions
at right anglesto its direction of propagation,whereasa ray obtained
by double refraction must be supposedto have sides, or properties
1 Theoneae la lumiere, p. 89
• The second edition of Newton's Opticlcs, query 26
27
THE AETHER
related to specialdirectionsat right anglesto its own direction. The
refraction of such a ray at the surfaceof a crystal dependson the
relation of its sidesto the principal plane of the crystal.
That a ray of light should possesssuch properties seemedto
Newton 1 an insuperableobjection to the hypothesiswhich regarded
wavesof light asanalogousto wavesof sound. On this point he was
in the right; his objections are perfectly valid against the wave-
theory as it was understood by his contemporaries, I although not
againstthe theory 3 which wasput forward a century later by Young

and Fresnel.
Although Huygens' chief researchesin optics came later than
Newton's discoveriesof 1666-72, Huygens was actually senior to
Newton by thirteen years, and he was, as regardshis physical con-
ceptions,in the main a Cartesian. His aether was, like Descartes',
constituted of subtle particles; and he attempted to interpret
gravity in termsof it, supposingit to be in rapid rotation in the region
of spacesurroundingthe earth-to form, in fact, a Cartesianvortex-
and, by the effectof its centrifugal force, to displaceterrestrialobjects
towards the centre of the earth. As we have seen,however (p. 7),
he was well aware of the errors of principle involved in Descartes'
method of approach to Nature.
The correct law of gravitation, which is valid for the whole
universe,and which made it possibleto calculate the motionsof all
the heavenly bodies, was published in 1687 in Newton's Principia:
namely, that any two particlesare attracted to eachother with a force
whosemagnitude is proportional to the product of their massesand
the inversesquareof their distanceapart. Newton claimed nothing
more for his discoverythan that it provided the necessaryinstrument
for mathematicalprediction, and he pointed out that it did not touch
on the question of the mechanism of gravity. As to this, he
conjectured that the densityof the aether might vary from place to
place, and that bodiesmight tend to move from the denserparts of
the medium towards the rarer; but whether this were the true
explanation or not, at any rate, he said, to suppose' that one body
may act upon another at a distancethrough a vacuum, without the
mediation of anything else, . . . is to me sogreat an absurdity,that
I believe no man, who has in philosophical matters a competent
faculty for thinking, can ever fall into.'
1 Opticks, query 28
I In which the oscillationsareperformedin the directionin which thewaveadvances.
I In which the oscillationsare performedin a direction at right anglesto that in
which the waveadvances.

The work of Newton was of such unparalleled brilliance that

one is surprised to learn how slowly it displaced Descartes'theory
of vortices. Long after the publication of the Principia, the text-
book of natural philosophy used at Cambridge continued to be
a translation (into Latin from French) of the Physics of Rohault, a
work entirely Cartesian. Whiston hasrecordedthat, having returned
to Cambridge after his ordination in 1693, he resumedhis studies
there, , particularly the Mathematicksand the CartesianPhilosophy,
which wasalonein Vogue with us at that Time. But it wasnot long
before I, with immense Pains, but no Assistance,set myself with
the utmost Zeal to the study of Sir Isaac Newton's wonderful
Discoveries.'1 The change in the character of the official teaching
wasbrought about in a very curious manner. Dr Samuel Clarke, a
zealousNewtonian, published about the year 1718a new translation
of Rohault, with a running commentary of notc:s which, while
avoiding the language and appearance of controversy, actually
constituteda completerefutation of the text. This edition superseded
the older one in current use,and the younger generation peacefully
adoptedthe new knowledge.
The first British university to incorporate Newtonianism into its
formal instruction seemsto have been Edinburgh, where David
Gregory, a friend of Newton, was professorfrom 1683-90. To
quote Whiston again, , He causedseveralof his Scholarsto keepActs,
aswe call them, upon severalBranchesof the Newtonian Philosophy;
while we at Cambridge,poor Wretches,were ignominiously studying
the fictitious Hypothesesof the Cartesian.' 2
On the continent, the change took place still more slowly. 'A
Frenchman who arrives in London,' wrote Voltaire in 1730 ' will
find Philosophy,like everything else,very much changedthere. He
had left the world a jJ!enum, and now he finds it a vacuum.' 3 For this
he gave a most surprising explanation. 'It is,' he said,' 'the
languageused,and not the thing in itself, that irritates the human
mind. If Newton had not usedthe word attraction in his admirable
philosophy,everyonein our Academy would have openedhis eyesto
the light; but unfortunately he usedin London a word to which an
idea of ridicule was attached in Paris; and on that alone he was
judged adversely,with a rashnesswhich will someday be regarded
asdoing very little honour to his opponents.' Ii
I W. Whiston's Memoirs (1749), i, p. 36 • ibid. p. 36
I Voltaire, ut/res philosophiques,quatorzieme letlre
• In a letter to M. de la Condamine, of date 22 June 1734, printed in (Euvrescompletes
ie M. de Voltaire (Deux-Ponts, 1791-2), tome 79, p. 219
• Presumably the Cartesian word impulsion would have satisfied them.
29
THE AETHER
In Germany, Leibnitz described the Newtonian formula as a

return to the discreditedscholasticconceptof occult qualities; 1 and as
late as the middle of the eighteenth century Euler and two of the
Bernoullis basedthe explanation of magnetismon the hypothesisof
vortices.2
The rejection of the inverse-squarelaw of gravitation by the
French Cartesiansantagonisedthe younger disciplesof Newton to
such an extent that the latter hardenedinto an oppositionnot only
to the vorticesbut to the whole body of Cartesiannotions,including
the aether. In the secondedition (1713) of the Principia, there is a
prefacewritten by Roger Cotes(1682-1716), in which the Newtonian
law of action at a distanceis championedasbeing the only formula-
tion of the facts of experiencewhich doesnot introduce unverifiable
and uselesssuppositions. The principle which Cotesnow affirmed
was that the aim of theoretical physicsis simply the prediction of
future events,and that everything which is not strictly neededfor this
purpose, and which is not directly deducible from observedfacts,
should be pruned away. In this assertionthere is clearly a question
involved which is fundamental in the philosophy of nature, and
which it will be worth while to look at more closely.
The investigationsof the theoreticalphysicistareconcernedpartly
with events which can actually be observed-these will be called
phenomena-and partly also with eventswhich cannot be detected
in themselves,but which are assumedto exist in order to set up
continuity between separated events that are in fact observed;
continuity is felt to be necessaryif the picture of the world is to be
logically coherent. Thesehypothetical eventsare called by modern
writers 3 inter-phenomena. An illustration may be drawn from optics:
light cannot be perceivedexceptwhen it impingeson matter: these
impacts are phenomena; but since the velocity of light is finite,
there areintervalsof time betweensuccessive encounterswith matter;
and if we adopt the belief that the luminous disturbancecontinuesto
exist in someform betweenits leaving one point and appearing at
another, then during this intermediate state it is classifiedas an
inter-phenomenon. Another illustration is provided by gravity:
assuming(what is in fact now established)that its speedofpropaga-
tion is finite, the question as to what is happening while it is in
course of propagation belongs to the realm of inter-phenomena.
1 S. Clarke, Papm which passedbetweenMr Leibnit~ and Dr Clarke (1717),p. 265

• Their memoirsshareda prize of the FrenchAcademyin 1743,and were printed in
1752in the &cueil despiecesqui ant remportllts prix de [' Acad., tome v.
• e.g. H. Reichenbach
30
Cotes's principle recommends that inter-phenomena should be

ignored, on the ground that they are unobservable and that
formulae of prediction can be establishedwithout making use of
them; that they belong to metaphysics rather than to physics.
On the other hand, the SChOOl. of which Descartesand Huygens
were the first illustrious members, and to which almost all the
physicistsof the next two centuries belonged, regarded the under-
standing of inter-phenomenaas one of the primary aims of natural
philosophy; as the carrying of ideas beyond phenomena into
regionswhich were more profound, and which still had ontological
validity.
In Cotes'sday the propagationof gravitation wasassumedto be
instantaneous,and consequentlythere was no span of time to be
bridged by inter-phenomena. When the eighteenth-centurynatural
philosophersfound by experience that the Newtonian law was
marvellouslypowerful, yielding formulae by which practically every
observablemotion in the solar systemcould be predicted, while on
the other hand the searchfor an explanation of inter-phenomena
led to no practical result, opinion set in favour of Cotes's
attitude, which came to prevail widely; and in the middle of
the century R. G. Boscovich (1711-87), a Croatian Jesuit, who
was the first exponent of Newtonian ideas in Italy, attempted
to account for all known physical effectsin terms of action at a
distance between point particles.1 The luminiferous aether was
abandoned,in spite of Huygens' brilliant developmentof the wave-
hypothesis,and a corpuscular theory of light was adopted almost
universally.
Even during this age,however,the hope of understandinginter-
phenomenawas never entirely abandoned. A curious attempt to
CartesianiseNewtonianism was made by a French-Swiss,George
Louis Le Sage,2who proposedto accountfor gravitation by meansof
an aether of Descartes'original type, that is to say, a cloud of
excessivelyminute particles,ultra-mundane corpusclesashe called them.
These,which resembledthe neutrinosof the modern atomic physicist,
he supposedto existin great numbersin all parts of space,and to be
moving with great speedin all directions; their diameters were
further assumedto be so small in comparison with their distances
apart that collisionsbetweenthem wereexceedinglyrare. Then two
particlesof ordinary matter would to someextent screeneach other
from bombardmentby the corpuscles,each particle receiving fewer
1 Boscovich, Theona PhilosophitU Naturalis, Venice, 1763
• Mim. deBerlin for 1782 (Berlin, 1784), p. 404
31
THE AETHER
impacts on the side facing the other than on the reverseside; and
Le Sageshowed that this effect would be equivalent to a force of
attraction between them, varying inversely as the square of their
mutual distance; which force he identified with that discoveredby
Newton.
Chapter II
ELECTRIC AND MAGNETIC SCIENCE PRIOR TO
THE INTRODUCTION OF THE POTENTIALS
THE ancientswere acquainted with the curious properties possessed

by two minerals, amber (ifA€K7pOV) and magnetic iron ore ('Ij >.t8o~
Mayvi}'Tts). The former, when rubbed, attracts light bodies; the
latter has the power of attracting iron.
The useof the magnet for the purposeof indicating direction at
seawas not known in classicalantiquity; and the questionsas to
where, when and by whom the compasswas invented, cannot be
answeredwith complete certainty. Until recent years the general
opinion was that it originated in China, and was brought by the
Arabs to the Mediterranean and so came to the knowledge of the
Crusaders. This, however,wasnot so; the Chinesewere acquainted
with the directive property of a magnet by the end of the eleventh
century, but did not apply it to purposesof navigation until at least
the end of the thirteenth. There is no evidencethat the Arabs had
any sharein the invention or first transmissionof knowledgeregarding
the compass. The earliest referenceto it is in a work written, most
probably in 1186, by Alexander Neckam (1157-1217), a monk of
St Albans, and he doesnot refer to it as something new. There
seemslittle doubt that it was known in north-western Europe,
probably in England, earlier than elsewhere.!
Magnetism was one of the few sciencesthat made any progress
during the Middle Ages; for in the thirteenth century Pierre de
Maricourt, a native of Picardy, made a discovery of fundamental
importance.2
Taking a natural magnet or lodestone,which had been rounded
into a globular form, he laid it on a needleand marked the line along
which the needlesetitself. Then laying the needleon other parts of
the stone,he obtained more lines in the sameway. When the entire
surfaceof the stone had been coveredwith such lines their general
1 The early history of magnetismis treated very fully and with great learning by
A. Crichton Mitchell, Terrestrial Magnetism and AtmosphericElectricity, xxxvii (1932),105;

xlii (1937),241; xliv (1939),77; Ii (1946),323.
I His Epistola Petri Peregrini d! Maricourt de magnetewaswritten in 1269. At that time
both 5t Albert the Great and 5t Thomas were living, but they and their immediate
~uccessors do not referto it.
33
ELECTRICITY AND MAGNETISM
disposition became evident; they formed circles which girdled the

stone in exactly the sameway as meridians of longitude girdle the
earth; and thereweretwo points at oppositeendsof the stonethrough
which all the circlespassed,just asall the meridianspassthrougb the
north and south poles of the earth.! Struck by the analogy,
Peregrinusproposedto call thesetwo points the poles of the magnet;
and he observedthat the way in which magnetsset themselvesand
attract each other dependssolely on the position of their poles,as if
thesewere the seatof the magnetic power. Such was the origin of
thosetheoriesof polesand polarisationwhich in later ageshaveplayed
so great a part in natural philosophy.
The observationsof Peregrinuswere greatly extended not long
before the time of Descartesby William Gilberd or Gilbert 2 (1540-
1603). Gilbert was born at Colchester; after studying at Cam-
bridge he took up medical practice in London, and had the honour
of being appointed physician to Queen Elizabeth. In 1600 he
published a work 3 on magnetism and electricity, with which the
modern history of both subjectsbegins.
Of Gilbert's electrical researcheswe shall speak later; in
magnetismhe made the capital discoveryof the reasonwhy magnets
set in definite orientations with respect to the earth; which is,
that the earth is itself a great magnet, having one of its polesin high
northern and the other in high southern latitudes. Thus the prop-
erty of the compasswasseento be included in the general principle,
that the north-seeking pole of every magnet attracts the south-
seekingpole of every other magnet,and repelsits north-seekingpole.
Gilbert went further, and conjectured that magnetic forces were
capable of accounting for the earth's gravity and the motions of the
planets.
Descartesattempted' to account for magnetic phenomena by
his theory of vortices. Adopting a suggestionof Gilbert's, he pos-
tulated a vortex of fluid matter round eachmagnet,the matter of the
vortex entering by one pole and leaving by the other: this matter
wassupposedto act on iron and steelby virtue of a specialresistance
to its motion afforded by the moleculesof thosesubstances.
In the seventeenthcentury severaljesuits-N. Cabeo,A. Kircher,
V. Leotaud-published works on magnetismwhich, while acknow-
1 'Procul dubio omnes lineae hujusmodi in duo puncta concurrent sicut omnes orbes
meridiani in duo concurrunt polos mundi oppositos.'
• The form in the Colchester records is Gilberd.
• Gulielmi Gilberti de Magnete, Magneti&isque corporibus, et ue magno magnete tdlure,
London, 1600. An English translation by P. F. Mottelay was published in 1893.
• Principia, pt. iv, § 133 sqq.
34
TO THE INTRODUCTION OF THE POTENTIALS
ledging the discoveriesof Gilbert, attempted to bring the subject

within the framework of the Aristotelian-scholasticphilosophy.!
The magnetic discoveriesof Pierre de Maricourt and Gilbert,
and the vortex hypothesisby which Descarteshad attempted to
explain them, had raised magnetism to the rank of a separate
scienceby the middle of the seventeenthcentury. The kindred
scienceof electricity was at that time in a less developed state;
but it had beenconsiderablyadvancedby Gilbert, whoseresearches
in this direction will now be noticed.
For two thousandyearsthe attractive power of amber had been
regarded as a virtue peculiar to that substance,or possessed by at
mostone or two others.2 Gilbert proved 3 this view to be mistaken,
showingthat the sameeffectsare induced by friction in quite a large
classof bodies,among which he mentioned glass,sulphur, sealing-
wax and variouspreciousstones.
A force which was manifestedby so many different kinds of
matter seemedto needa name of its own; and accordingly Gilbert
gaveto it the name electric, which it has ever sinceretained.'
Between the magnetic and electric forces Gilbert remarked
many distinctions. The lodestonerequires no stimulus of friction
suchas is neededto stir glassand sulphur into activity. The lode-
stone attracts only magnetisable substances,whereas electrified
bodies attract everything. The magnetic attraction between two
bodies is not affected by interposing a sheet of paper or a linen
cloth, or by immersing the bodies in water; whereasthe electric
attraction is readily destroyed by screens. Lastly, the magnetic
force tends to arrange bodies in definite orientations; while the
electricforcemerelytendsto heapthem togetherin shapeless clusters.
Thesefactsappearedto Gilbert to indicate that electric phenom-
ena are due to something of a material nature, which under the
influence of friction is liberated from the glassor amber in which
under ordinary circumstancesit is imprisoned. In support of this
view he adducedevidencefrom other quarters. Being a physician,
he was well acquainted with the doctrine 5 that the human body
1 It should be said for Cabeo, however, that in his commentary on Aristotle's

Meteorology he frequently insists on the desirability of following experience rather than
Aristotle.
• Bede in his Ecclesiastical History mentions that jet, when warmed by rubbing, , holds
fast whatever is applied to it, like amber.'
• De Magnete, lib. ii, cap. 2
• The substantive electricity was introduced by Sir Thomas Browne in his Pseudodoxia
epidemica(1646), p. 79.
• The doctrine of humours had been borrowed by the physicians of western Europe
from the Arabs, who in turn had derived it from the Greeks.
(823) 35
contains various humours or kinds of moisture-phlegm, blood,

choler and melancholy-which, as they predominated, were sup-
posed to determine the temper of mind,l and when he observed
that electrifiable bodies were almost all hard and transparent, and
therefore (according to the ideas of that time) formed by the con-
solidation of watery liquids, he concluded that the common men-
struum of theseliquids must be a particular kind of humour, to tht.
possessionof which the electrical properties of bodies were to be
referred. Friction might be supposedto warm or otherwise excite
or liberate the humour, which would then issuefrom the body as
an effluvium and form an atmosphere around it. The effluvium
must, he remarked, be very attenuated, for its emissioncannot be
detectedby the senses.
The existenceof an atmosphereof effluvia round everyelectrified
body might indeed have been inferred, according to Gilbert's ideas,
from the single fact of electric attraction. For he believed that
matter cannot act where it is not; and henceif a body acts on all
surrounding objects without appearing to touch them, something
must have proceededout of it unseen.
The whole phenomenon appeared to him to be analogousto
the attraction which is exercisedby the earth on falling bodies;
for in the latter case he conceived of the atmosphericair as the
effluvium by which the earth draws all things downwardsto itself.
Gilbert's theory of electrical emanations commended itself
generally to such of the natural philosophers of the seventeenth
century as were interested in the subject; among whom were
numbered Niccolo Cabeo (1585-1650), who was perhaps the first
to observe that electrified bodies repel as well as attract ; 2 the
English royalist exile, Sir Kenelm Digby (1603-65); and the
celebrated Robert Boyle (1627-91). There were, however, some
differencesof opinion as to the manner in which the effluvia acted
on the small bodies and set them in motion towards the excited
electric; Gilbert himself had supposedthe emanationsto have an
inherent tendency to reunion with the parent body; Digby likened
their return to the condensationof a vapour by cooling; and other
writers pictured the effluvia as forming vortices round the attracted
bodiesin the Cartesianfashion.
There is a well-known allusion to Gilbert's hypothesisin Newton's
Opticks.3
1 Phlegmatic, sanguine, choleric and melQluholic were terms describing excessof the different
humours.
• The first publication of this phenomenon seems to have been by O. von Guericke
in 1663. • query 22
36
, Let him also tell me, how an electrick body can by friction
emit an exhalation so rare and subtle,1 and yet so potent, as by
its emissionto causeno sensiblediminution of the weight of the
electrickbody, and to be expandedthrough a sphere,whosediameter
is above two feet, and yet to be able to agitate and carry up leaf
copper, or leaf gold, at a distanceof above a foot from the electrick
body? '
It is, perhaps,somewhatsurprising that the Newtonian doctrine
of gravitation shouldnot have proved a severeblow to the emanation
theory of electricity; but Gilbert's doctrine was now so firmly
establishedas to be unshakenby the overthrow of the analogy by
which it had been originally justified. It was, however, modified
in one particular about the beginning of the eighteenth century.
In order to account for the fact that electrics are not perceptibly
wastedaway by excitement,the earlier writers had supposedall the
, emanations to return ultimately to the body which had emitted
them; but the corpusculartheory of light accustomedphilosophers
to the idea of emissionsso subtle as to causeno perceptible loss;
and after the time of Newton the doctrine of the return of the
electric effluvia gradually lost credit.
Newton died in 1727. Of the expositions of his philosophy
which were published ~n his lifetime by his followers, one at least
deservesto be noticed for the sake of the insight which it affords
into the state of opinion regarding light, heat and electricity in the
first half of the eighteenth century. This was the Physices elementa
mathematica experimentis confirmata of Wilhelm Jacob 'sGravesande
(1688-1742), published at Leyden in 1720. The Latin edition was
afterwards reprinted several times, and was, moreover, translated
into French and English; it seemsto have exerciseda considerable
and, on the whole, well-deserved influence on contemporary
thought.
'sGravesande supposed light to consist in the projection of
corpusclesfrom luminous bodies to the eye; the motion being
very swift, as is shown by astronomical observations. Since many
bodies, e.g. the metals, becomeluminous when they are heated, he
inferred that every substancepossesses a natural store of corpuscles,
which are expelledwhen it is heated to incandescence; conversely,
corpusclesmay becomeunited to a material body; as happens,for
instance,when the body is exposedto the rays of a fire. Moreover,
since the heat thus acquired is readily conducted throughout the
1 • Subtlety,' saysJohnson,• which in its original import meansexility of particles,is
takenin its metaphoricalmeaningfor nicety of distinction.'

37
substanceof the body, he concluded that corpusclescan penetrate

all substances,however hard and densethey be.
Let us here recall the ideas then current regarding the nature
of material bodies. After the publication in 1661 of The Sceptical
Chemist, by Robert Boyle (1626-91) it had beenrecognisedgenerally
that substancesperceptible to the sensesmay be either elementsor
compounds or mixtures; the compounds being chemical individuals,
distinct from mere mixtures of elements. But the substancesat that
time accepted as elements were very different from those which
are now known by the name. Air and the calces1 of the metals
figured in the list, while almost all the chemical elements now
recognisedwere omitted from it; some of them, such as oxygen
and hydrogen, becausethey were as yet undiscovered,and others,
such as the metals, becausethey were believed to be compounds.
Among the chemical elements,it became customary after the
time of Newton to include light corpuscles.s That something
which is confessedlyimponderable should ever have been admitted
into this class may at first sight seemsurprising. But it must be
remembered that questionsof ponderability counted for very little
with the philosophersof the period. Three-quartersof the eighteenth
century had passedbefore Lavoisier enunciated the fundamental
doctrine that the total weight of the substancesconcerned in a
chemical reaction is the same after the reaction as before it. As
soon as this principle came to be universally applied, light parted
company from the true elementsin the schemeof chemistry.
We must now consider the views which were held at this time
regarding the nature of heat. These are of interest for our present
purpose, on account of the analogieswhich were set up between
heat and electricity.
The various conceptions which have been entertained con-
cerning heat fall into one or other of two classes,according as heat
is represented as a mere condition producible in bodies, or as a
distinct speciesof matter. The former view, which is that universally
held at the present day, was advocated by the great philosophers
of the seventeenth century. Bacon maintained it in the Novum
Organum: 'Calor,' he wrote, 'est motus expansivus,cohibitus, et
nitens per partes minores.' 8 Boyle' affirmed that the' Nature of
1 i.e. oxides
• Newton himself (Opticks, p. 349) suspectedthat light corpuscles and ponderable
matter might be transmuted into each other; much later, Boscovich (Tluoria, pp. 215,
217) regarded the matter of light as a principle or element in the constitution of natural
bodies.
• NOJl. Org., Lib. u, Aphor. :xx • Mechanical Productionof Heat and Cold
38
TO THE. INTRODUCTION OF THE POTENTIALS
Heat' consistsin ' a various, vehement, and intestine commotion

of the Parts among themselves.' Hooke 1 declared that ' Heat is a
property of a body arising from the motion or agitation of its parts.'
And Newton 2 asked: 'Do not all fixed Bodies,when heatedbeyond
a certain Degree, emit light and shine; and is not this Emission
performedby the vibrating Motion of their Parts? ' and, moreover,
suggestedthe converseof this namely, that when light is absorbed
by a material body, vibrations are setup which are perceivedby the
sensesasheat.
The doctrine that heat is a material substancewas maintained
in Newton's lifetime by a certain school of chemists. The most
conspicuousmember of the school was Wilhelm Homberg (1652-
1715) of Paris, who a identified heat and light with the sulphureolls
principle, which he supposedto be one of the primary ingredients of
all bodies, and to be present even in the interplanetary spaces.
Betweenthis view and that of Newton it might at first seemas if
nothing but sharp opposition was to be expected.' But a few years
later the professedexponentsof the Principia and the Opticks began
to develop their systemunder the evident influence of Homberg's
writings. This evolution may easily be traced in 'sGravesande,
whosestarting-point is the admittedly Newtonian idea that heat
bears to light a relation similar to that which a state of turmoil
bearsto regular rectilinear motion; whence,conceiving light as a
projection of corpuscles,he infers that in a hot body the material
particles and the light corpusclesIi are in a state of agitation, which
becomesmore violent as the body is more intenselyheated.
'sGravesandethus holds a position between the two opposite
camps. On the one hand he interprets heat as a mode of motion;
but on the other he associatesit with the presenceof a particular
kind of matter, which he further identifies with the matter of light.
After this the materialistic hypothesismade rapid progress. It was
frankly advocatedby anothermemberof the Dutch school,Hermann
Boerhaaves (1668-1738), Professorin the University of Leyden,
whosetreatiseon chemistrywas translated into English in 1727.
1 Mia-ographia, p. 37 • Op/icks, Query VIII
• Mem. de l'Acad. (1705), p. 88
• Though it reminds us of a curious conjecture of Newton's : • Is not the strength and
vigour of the action between light and sulphureous bodies one reason why sulphureous
bodies take fire more readily and burn more vehemently than other bodies do ?' (Op/icks,
Query VIII)
• I have thought it best to translate 'sGravesande's ignis by • light corpuscles.' This
is, I think, fully justified by such of his statements as Quandoignis per lineasr"1as oculosIIOs/ros
in/rat, ex motu quemfibris in fundo oculi communica/ideam luminis excilaJ.
• Boerhaave followed Homberg in supposing the matter of heat to be present in all
IO-Calledvacuous spaces.
39
Somewhatlater it was found that the heating effectsof the rays

from incandescentbodies may be separated from their luminous
effectsby passingthe rays through a plate of glass,which transmits
the light, but absorbsthe heat. After this discoveryit wasno longer
possibleto identify the matter of heat with the corpusclesof light;
and the former wasconsequentlyacceptedasa distinct element,under
the name of caloric.l In the latter part of the eighteenthand early
part of the nineteenth centuries 2 caloric wasgenerally conceivedas
occupying the intersticesbetweenthe particles of ponderablematter
-an idea which fitted in well with the observation that bodies
commonly expand when they are absorbingheat, but which wasless
competent to explain the fact 3 that water expandswhen freezing.
The latter difficulty was overcomeby supposingthe union between
a body and the caloric absorbedin the processof melting' to be of a
chemical nature; so that the consequentchangesin volume would
be beyond the possibility of prediction.
As we have already remarked, the imponderability of heat did
not appear to the philosophersof the eighteenth century to be a
sufficient reasonfor excluding it from the list of chemical elements;
and in any casethere was considerabledoubt as to whether caloric
was ponderable or not. Some experimentersbelieved that bodies
were heavier when cold than when hot; others that they were
heavier when hot than when cold. The century was far advanced
before Lavoisier and Rumford finally proved that the temperature
of a body is without sensibleinfluence on its weight.
Perhaps nothing in the history of natural philosophy is more
amazing than the vicissitudes of the theory of heat. The true
hypothesis,after having met with general acceptancethroughout
a century, and having been approved by a successionof illustrious
men, was deliberately abandonedby their successors in favour of a
conception utterly false,and, in someof its developments,grotesque
and absurd.
1 Scheele in 1777 supposed c~loric to be a compound of oxygen and phlogiston, and
light to be oxygen combined with a greater proportion of phlogiston. Euler alone among
the philosophers of the mid-eighteenth century refused to regard heat as a material
substance.
I In spite of the experiments of Benjamin Thompson, Count Rumford (1753-1814),
in the closing years of the eighteenth century. These should have sufficed to re-establish
the older conception of heat.
I This had been known since the time of Boyle.
• Latent heat was discovered by Joseph Black (1728-99) in 1762. Black was the first
to make a dear distinction between quantity of heat and temperature, although G. D. Fahren-
heit (1686-1736) had devised a thermometer which was practically useful as a measuring
instrument in 1708-17.
TO THE INTRODUCTION OF THE POTENT.ALS
We must now return to 'sGravesande'sbook. The phenomena

of combustionhe explained by assumingthat when a body is suffi-
ciently heatedthelight corpusclesinteract with the material particles,
someconstituentsbeing in consequenceseparatedand carried away
with the corpusclesasflame and smoke. This view harmoniseswith
the theory of calcination which had been developedby Becherand
his pupil Stahl at the end of the seventeenthcentury, according to
which the metalswere supposedto be composedof their calcesand
an elementphlogiston. The processof combustion,by which a metal
is changedinto its calx, wasinterpreted asa decomposition,in which
the phlogiston separated from the metal and escaped into the
atmosphere; while the conversionof the calx into the metal was
regardedasa union with phlogiston.l
'sGravesandeattributed electric effectsto vibrations induced in
effluvia, which he supposedto be permanently attached to such
bodiesasamber. 'Glass,' he asserted,, containsin it, and hasabout
its surface,a certain atmosphere,which is excited by Friction and
put into a vibratory motion; for it attracts and repelslight Bodies.
The smallestparts of the glassare agitated by the Attrition, and by
reasonof their elasticity, their motion is vibratory, which is com-
municatedto the Atmosphereabove-mentioned; and thereforethat
Atmosphereexerts its action the further, the greater agitation the
Parts of the Glassreceive when a greater attrition is given to the
glass.'
The English translator of 'sGravesande'swork was himself
destinedto playa considerablepart in the history of electricalscience.
Jean Theophile Desaguliers(1683-1744) was an Englishman only
by adoption. His father had beena Huguenot pastor,who, escaping
from France after the revocation of the Edict of Nantes, brought
away the boy from La Rochelle,concealed,it is said, in a tub. The
young Desagulierswas afterwards ordained, and becamechaplain
to that Duke of Chandoswho wassoungratefully ridiculed by Pope.
In this situation he formed friendships with some of the natural
philosophersof the capital, and amongstothers with StephenGray,
I The correct idea of combustion had been advanced by Hooke. • The dissolution of
inflammable bodies,' he assertsin the lt1icrographia, • is performed by a substance inherent
in and mixed with the air, that is like, if not the very same with, that which is fixed in
saltpetre.' But this statement met with little favour at the time, and the doctrine of the
compound nature of metals survived in full vigour until the discovery of oxygen by
Priestley and Scheelein 1771-5. In 1775 Lavoisier reaffirmed Hooke's principle that a
metallic calx is not the metal minus phlogiston, but the metal plus oxygen; and this idea,
which carried with it the recognition of the elementary nature of metals, was generally
accepted by the end of the eighteenth century.
41
an experimenterof whom little is known 1beyondthe fact that hewas

a pensionerof the Charterhouse.
In 1729 Gray communicated, as he says,2'to Dr Desagulien;
and someother Gentlemen' a discoveryhe had lately made,, show-
ing that the Electrick Vertue of a GlassTube may be conveyedto
any other Bodiesso as to give them the sameProperty of attracting
and repelling light Bodiesasthe Tube does,when excitedby rubbing;
and that this attractive Vertue might be carried to Bodiesthat were
many feet distant from the Tube.'
This wasa result of the greatestimportance, for previousworkers
had known of no other way of producing the attractive emanations
than by rubbing the body concerned.3 It was found that only a
limited classof substances,amongwhich the metalswereconspicuous,
had the capacity of acting aschannelsfor the transport of the electric
power; to theseDesaguliers,who continued the experimentsafter
Gray's death in 1736,gave' the name non-electrics or conductors.
Mter Gray's discovery it was no longer possibleto believe that
the electric efHuviaare inseparably connectedwith the bodiesfrom
which they are evoked by rubbing; and it became necessaryto
admit that these emanations have an independent existence,and
can be transferred from one body to another. Accordingly we find
them recognised,under the name of the electric fluid,5 as one of the
substancesof which the world is constituted. The imponderability
of this fluid did not, for the reasonsalready mentioned, prevent its
admissionby the side of light and caloric into the list of chemical
elements.
The question was actively debated as to whether the electric
fluid was an element sui generis, or, as somesuspected,was another
1 Those who are interested in the literary history of the eighteenth century will recall
the controversy as to whether the verses on the death of Stephen Gray were written by
Anna Williams, whose name they bore, or by her patron Johnson.
Now, hoary Sage, pursue thy happy flight

With swifter motion haste to purer light
Where Bacon waits with Newton and with Boyle
To hail thy genius, and applaud thy toil;
Where intuition breaks through time and space
And mocks experiment's successiverace;
Seestardy science toil at Nature's laws
And wonders how th' effect obscures the cause.
• Phil. Trans. xxxvii (1731), pp. 18,227,285,397
3 Otto von Guericke (1602-86) had, as a matter of fact, observed the conduction of
electricity along a linen thread; but this experiment does not seemto have been followed
up. cf. Experimenta nova magdeburgica,1672.
• Phil. Trans. xli (1739), pp. 186, 193, 200, 209; Dissertation conr.erningElectricivr. 1742
• The Cartesians defined a fluid to be a body whose minute parts are in a continual
agitation. The word fluid was first used as a noun in English by Boyle.
42
manifestationof that principle whoseoperationisseenin the phenom-

ena of heat. Thosewho held the latter view urged that the electric
fluid and heat can both be induced by friction, can both induce
combustionand can both be transferredfrom one body to another
by mere contact; and moreover,that the best conductorsof heat
are alsoin generalthe best conductorsof electricity. On the other
hand it was contendedthat the electrification of a body doesnot
causeany appreciablerisein its temperature; and an experimentof
StephenGray'sbrought to light a yet morestriking difference. Gray,1
in 1729, madetwo oakencubes,one solid and the other hollow, and
showedthat when electrifiedin the sameway they producedexactly
similar effects; whencehe concluded that it was only the surfaces
which had taken part in the phenomena. Thus while heat is dis-
seminatedthroughout the substanceof a body, the electric fluid
residesat or nearits surface. In the middle of the eighteenthcentury
it was generally compared to an enveloping atmosphere. 'The
electricity which a non-electric of great length (for example, a
hempenstring 800 or 900 feet long) receives,runs from one end to
the other in a sphereof electrical EjJluvia,' saysDesaguliersin 17402 ;
and a report of the French Academy in 1733 says3: 'Around an
electrified body there is formed a vortex of exceedinglyfine matter
in a state of agitation, which urges towards the body such light
substancesaslie within its sphereof activity. The existenceof this
vortex is morethan a mereconjecture; for when an electrified body
is brought close to the face it causesa sensationlike that of en-
counteringa cobweb.' &
The report from which this is quoted waspreparedin connection
with the discoveriesof Charles-Franc;ois du Fay (1698-1739), super-
intendentof gardensto the King of France. Du Fay 6 accountedfor
the behaviour of gold leaf when brought near to an electrified glass
tube by supposingthat at first the vortex of the tube envelopesthe
gold leaf, and so attracts it towards the tube. But when contact
occurs, the gold leaf acquires the electric virtue, and so becomes
surroundedby a vortex of its own. The two vortices, striving to
extend in contrary senses,repel each other, and the vortex of the
tube, being the stronger,drives away that of the gold leaf. 'It is
then certain,' saysdu Fay,6 ' that bodieswhich havebecomeelectric
1 Phil. Trans.xxxvii, p. 35 • Phil. Trans.xli, p. 636
• Hist. de l' Acad. (1733),p. 6
• This observationhad beenmadefirst by Hauksbeeat the beginning of the century.
• Mim. de l'Acad. (1733),pp. 23, 73, 233, 457; (1734),pp. 341, 503; 1737,p. 86 ;
Phil. Trans.xxxviii (1734),p. 258
• Mim. de [,Acad. (1733),p. 464
43
by contact are repelledby thosewhich haverenderedthem electric;

but are they repelled likewise by other electrified bodies of all
kinds? And do electrified bodies differ from each other in no
respect save their intensity of electrification? An examination of
this matter has led me to a discovery which I should never have
foreseen,and of which I believe no-one hitherto has had the least
idea.'
He found, in fact, that when gold leaf which had beenelectrified
by contact with excited glasswas brought near to an excited piece
of copal,l an attraction was manifested between them. 'I had
expected,'he writes, 'quite the oppositeeffect, since, accordingto
my reasoning,the copal and gold leaf, which were both electrified
should have repelled each other.' Proceedingwith his experiments
he found that the gold leaf, when electrified and repelled by glass,
was attracted by all electrified resinoussubstances,and that when
repelled by the latter it was attracted by the glass. 'We see,then,'
he continues,'that there are two electricities of a totally different
nature-namely, that of transparentsolids,suchasglass,crystal, &c.
and that of bituminous or resinousbodies, such as amber, copal,
sealing-wax,&c. Each of them repelsbodieswhich havecontracted
an electricity of the samenature as its own, and attracts thosewhose
electricity is of the contrary nature. We seeeventhat bodieswhich
are not themselveselectricscan acquire either of theseelectricities,
and that then their effectsare similar to thoseof the bodieswhich
have communicatedit to them.'
To the two kinds of electricity whoseexistencewas thus demon-
strated, du Fay gave the namesvitreous and resinous, by which they
have ever sincebeenknown.
An interest in electrical experimentsseemsto have spreadfrom
du Fay to other membersof the Court circle of Louis XV; and from
1745 onwards the Memoirs of the Academy contain a seriesof papers
on the subject by the Abbe Jean-Antoine Nollet (1700-70), after-
wardspreceptorin Natural Philosophyto the Royal Family. Nollet
attributed electricphenomenato the movementin oppositedirections
of two currents of a fluid, ' very subtle and inflammable,' which he
supposedto be presentin all bodiesunder all circumstances.2 When
an electric is excited by friction, part of this fluid escapesfrom
its pores,forming an e.ffluentstream; and this lossis repaired by an
aifluent stream of the same fluid entering the body from outside.
I A hard transparentresin,usedin the preparationof varnish.
I cf. Nollet's R~ch~rch~s
(1749),p. 245. Nollet'slecturesin Ex/J~rimentalPhilosophywere
translatedby John Colsonin 1748.
44
Light bodiesin the vicinity, being caught in one or other of these

streams,are attracted or repelled from the excited electric.
Nollet's theory was in great vogue for some time; but six or
sevenyearsafter its first publication, its author came acrossa work
purporting to be a French translation of a book printed originally in
England, describingexperimentssaid to have been made at Phila-
delphia, in America, by one Benjamin Franklin. 'He could not at
first believe,' as Franklin tells us in his Autobiography, 'that such a
work came from America, and said it must have been fabricated
by his enemiesat Paristo decry his system. Mterwards, having been
assured that there really existed such a person as Franklin at
Philadelphia, which he had doubted, he wrote and published a
volume of letters,chiefly addressedto me, defending his theory, and
denying the verity of my experiments,and of the positions deduced
from them.
We must now trace the events which led up to the discovery
which soperturbed Nollet.
In 1745 Pieter van Musschenbroek(1692-1761), Professorat
Leyden, attempted to find a method of preservingelectric charges
from the decay which was observedwhen the charged bodies were
~\Urroundedby air. With this purpose he tried the effect of sur-
rounding a charged massof water by an envelope of some non-
conductor, e.g. glass. In one of his experiments,a phial of water
was suspendedfrom a gun barrel by a wire let down a few inches
into the water through the cork; and the gun barrel, suspended
on silk lines, was applied so near an excited glassglobe that some
metallic fringes inserted into the gun barrel touched the globe in
motion. Under thesecircumstancesa friend named Cunaeus,who
happened to grasp the phial with one hand, and touch the gun
barrel with the other, received a violent shock; and it became
evident that a method of accumulating or intensifying the electric
power had beendiscovered.1
Shortly after the discoveryof the Leyden phial, as it was named
by Nollet, had becomeknown in England, a London apothecary
named William Watson (1715-87) 2 noticed that when the experi-
ment is performedin this fashionthe observerfeelsthe shock' in no
other partsof his body but his armsand breast' ; whencehe inferred
that in the act of discharge there is a transferenceof something
1 The discovery was made independently (and probably somewhat earlier) by Ewald
Georg von Kleist, Dean of Kammin in Pomerania. Kleist did not publish it, and it was
first described in a work Gmhichttd. Erdt in d. alltralttstm Ztit (1746) by J. G. KrUger,
Professorof Medicine at Halle.
I Watson afterwards rose to eminence in the medical profession, and was knighted.
45
which takes the shortestor best conducting path between the gun
barrel and the phial. This idea of transferenceseemedto him to
bear somesimilarity to Nollet's doctrine of afflux and efflux; and
there can indeed be little doubt that the Abbe's hypothesis,though
totally falsein itself, furnished someof the ideasfrom which Watson
with the guidance of experiment, constructed a correct theory.
In a memoir 1 read to the Royal Society in October 1746he pro-
pounded the doctrine that electrical actions are due to the presence
of an ' electrical aether,' which in the charging or dischargingof a
Leyden jar is transferred, but is not created or destroyed. The
excitation of an electric, according to this view, consistsnot in the
evoking of anything from within the electric itself without com-
pensation, but in the accumulation of a surplus of electrical aether
by the electric at the expenseof someother body, whosestock is accord-
ingly depleted. All bodies were supposed to possessa certain
natural store, which could be drawn upon for this purpose.
, I have shewn,' wrote Watson, , that electricity is the effect of
a very subtil and elastic fluid, occupying all bodiesin contact with
the terraqueous globe; and that every-where,in its natural state,
it is of the samedegreeof density; and that glassand other bodies,
which we denominate electrics per se, have the power, by certain
known operations,of taking this fluid from one body, and conveying
it to another, in a quantity sufficient to be obviousto all our senses;
and that, under certain circumstances,it was possibleto render the
electricity in some bodies more rare than it naturally is, and, by
communicating this to other bodies, to give them an additional
quantity, and make their electricity more dense.'
In the same year in which Watson's theory was proposed, a
certain Dr Spence,who had lately arrived in America from Scotland,
was showing in Boston some electrical experiments. Among his
audiencewasa man who already at forty yearsof agewasrecognised
as one of the leading citizens of the English colonies in America,
Benjamin Franklin of Philadelphia (1706-90). Spence'sexperiments
'were,' writes Franklin,2 'imperfectly performed, as he was not
very expert; but, being on a subject quite new to me, they equally
surprisedand pleasedme.' Soonafter this, the ' Library Company'
of Philadelphia (an institution founded by Franklin himself) received
from Mr Peter Collinson of London a presentof a glasst~be, with
someaccount of its use. In a letter written to Collinson on 11July
1 Phil. Trans. xliv, p. 718. It may herebe noted that it wasWatsonwho improved
the phial by coatingit nearlyto the top, both insideand outside,with tinfoil.
I Franklin'sAutobiography
1747,1Franklin described experiments made with this tube, and

certain deductionswhich he had drawn from them.
If one personA, standing on wax so that electricity cannot pass
from him to the ground, rubs the tube, and if another person B,
likewise standing on wax, passeshis knuckle along near the glass
so as to receive its electricity, then both A and B will be capable
of giving a spark to a third personC standing on the floor; that is,
they will be electrified. If, however, A and B touch each other,
either during or after the rubbing, they will not be electrified.
This observationsuggestedto Franklin the samehypothesisthat
(unknown to him) had been propounded a few months previously
by Watson: namely, that electricity is an element present in a
certain proportion in all matter in its normal condition; so that,
before the rubbing, each of the personsA, Band C has an equal
share. The effectof the rubbing is to transfersomeof A's electricity
to the glass,whenceit is transferredto B. Thus A has a deficiency
and B a superfluity of electricity; and if either of them approaches
C, who hasthe normal amount, the distribution will be equalisedby
a spark. If, however, A and B are in contact, electricity flows
betweenthem soasto re-establishthe original equality, and neither
ie; then electrified with referenceto C.
Thus electricity is not created by rubbing the glass,but only
transferred to the glassfrom the rubber, so that the rubber loses
exactly as much as the glassgains; the total quantiry of electriciry in
any insulated system is invariable. This assertionis usually known as
the principle of conservationof electric charge.
The condition of A and B in the experiment can evidently be
expressedby plus and minus signs: A having a deficiency - e and
B a superfluity +e of electricity. Franklin, at the commencement
of hisown experiments,wasnot acquaintedwith du Fay'sdiscoveries;
but it is evidentthat the electric fluid of Franklin is identical with the
vitreous electricity of du Fay, and that du Fay's resinouselectricity
is, in Franklin's theory, merely the deficiency of a stock of vitreous
electricitysupposedto bepossessed naturally by all ponderablebodies.
In Franklin's theory we are sparedthe necessityfor admitting that
two quasi-materialbodiescan by their union annihilate each other,
asvitreous and resinouselectricity were supposedto do.
Some curiosity will naturally be felt as to the considerations
which induced Franklin to attribute the positive character to
vitreous rather than to resinous electricity. They seem to have
been founded on a comparison of the brush discharges fronl
1 Franklin's New Experimmts and Observationson E/ectri&ity, letter ii
47
conductorscharged with the two electricities; when the electricity

was resinous,the dischargewas observedto spreadover the surface
of the opposite conductor 'as if it flowed from it.' Again, if a
Leydenjar whoseinner coating is electrified vitreously is discharged
silently by a conductor, of whosepointed endsone is near the knob
and the other near the outer coating, the point which is near the
knob is seenin the dark to be illuminated with a star or globule,
while the point which is near the outer coating is illuminated with
a pencil of rays; which suggestedto Franklin that the electric fluid,
going from the inside to the outside of the jar, entersat the former
point and issuesfrom the latter. And yet again, in somecasesthe
flame of a wax taper is blown away from a brassball which is dis-
charging vitreous electricity, and towards one which is discharging
resinouselectricity. But Franklin remarks that the interpretation
of these observations is somewhat conjectural, and that whether
vitreous or resinous electricity is the actual electric fluid is not
certainly known.
Regarding the physical nature of electricity, Franklin held much
the sameideas as his contemporaries; he pictured it as an elastic1
fluid, consisting of 'particles extremely subtile, since it can per-
meate common matter, even the densestmetals,with such easeand
freedom as not to receiveany perceptibleresistance.' He departed,
however, to some extent from the conceptionsof his predecessors,
who were accustomed to ascribe all electrical repulsions to the
diffusion of effluvia from the excited electric to the body acted on ;
so that the tickling sensationwhich is experiencedwhen a charged
body is brought near to the human face was attributed to a direct
action of the effluvia on the skin. This doctrine, which, aswe shall
see,practically ended with Franklin, bearsa suggestiveresemblance
to that which nearly a century later was introduced by Faraday;
both explained electrical phenomena without introducing action
at a distance,by supposingthat somethingwhich forms an essential
part of the electrified systemis presentat the spot where any electric
action takes place; but in the older theory this something was
identified with the electric fluid itself, while in the modern view it is
identified with a stateof stressin the aether. In the interval between
the fall of one school and the rise of the other, the theory of action
at a distancewas dominant.
The germsof the last-mentionedtheory may be found in Frank-
lin's own writings. It originated in connectionwith the explanation
of the Leyden jar, a matter which is discussedin his third letter to
1 i.e. repulsiveof its own particles
48
Collinson,of date 1 September1747. In charging the jar, he says,

a quantity of electricity is taken away from one side of the glass,
by meansof the coating in contact with it, and an equal quantity
is communicatedto the other side, by meansof the other coating.
The glassitself he supposesto be impermeableto the electric fluid,
so that the deficiencyon the one side can permanently coexistwith
the redundancyon the other, so long as the two sidesare not con-
nected with each other; but, when a connection is set up, the
distribution of fluid is equalised through the body of the experi-
menter,who receivesa shock.
Compelled by this theory of the jar to regard glass as im-
penetrableto electric effluvia, Franklin wasneverthelesswell aware 1
that the interposition of a glassplate between an electrified body
and the objectsof its attraction doesnot shield the latter from the
attractiveinfluence. He wasthusdriven to suppose2 that the surface
of the glasswhich is nearestthe excited body is directly affected,
and is able to exert an influence through the glasson the opposite
surface; the latter surface,which thus receivesa kind of secondary
or derivedexcitement,is responsiblefor the electric effectsbeyondit.
This idea harmonisedadmirably with the phenomenaof thejar;
for it was now possibleto hold that the excessof electricity on the
inner face exercisesa repellent action through the substanceof the
glass,and so causesa deficiencyon the outer facesby driving away
the electricity from it. 3
Franklin had thus arrived at what was really a theory of action
at a distancebetweenthe particlesof the electric fluid; and this he
was able to support by other experiments. 'Thus,' he writes,4
'the stream of a fountain, naturally dense and continual, when
electrified, will separateand spread in the form of a brush, every
drop endeavouringto recedefrom every other drop.' In order to
account for the attraction between oppositely charged bodies, in
one of which there is an excessof electricity as compared with
ordinary matter, and in the other an excessof ordinary matter as
compared with electricity, he assumedthat 'though the particles
of electrical matter do repel each other, they are strongly attracted
by all other matter' ; sothat' commonmatter is asa kind of spunge
to the electrical fluid.' It absorbs electricity until saturation is
leached, after which any further excessof electricity must lie on
or near the surfaceof the body.
These repellent and attractive powers he assignedonly to the
I New Experiments(1750), § 28 • ibid. (1750), § 34
• ibid. (1750), § 32 • Letter v
49
actual (vitreous) electric fluid; and when later on the mutual

repulsion of resinously electrified bodies became known to him,l
it causedhim considerableperplexity.2 As we shall see,the difficulty
was eventually removedby Aepinus.
In spiteof his belief in the power of electricity to act at a distance,
Franklin did not abandon the doctrine of effluvia. 'The form of
the electrical atmosphere,'he says,3' is that of the body it surrounds.
This shape may be rendered visible in a still air, by raising
a smoke from dry rosin dropt into a hot teaspoon under the
electrified body, which will be attracted, and spread itself equally
on all sides,covering and concealing the body. And this form it
takes,becauseit is attracted by all parts of the surfaceof the body,
though it cannot enter the substancealready replete. Without this
attraction, it would not remain round the body, but dissipatein
the air.' He observed,however,that electrical effluvia do not seem
to affect, or be affected by, the air, since it is possibleto breathe
freely in the neighbourhood of electrified bodies; and moreover a
current of dry air doesnot destroyelectricattractionsand repulsions.'
Regarding the suspectedidentity of electricity with the matter
of heat,asto which Nollet had takenthe affirmative position,Franklin
expressedno opinion. 'Common fire,' he writes,5 ' is in all bodies,
more or less,aswell aselectrical fire. Perhapsthey may be different
modificationsof the sameelement; or they may be different elements
The latter is by somesuspected. If they are different things, yet they
may and do subsisttogether in the samebody.'
Franklin's wo~k did not at first receive from European philos-
ophers the attention which it deserved; although Watson gener-
ously endeavoured to make the colonial writer's merits known,8
and inserted some of Franklin's letters in one of his own papers
communicated to the Royal Society. But an account of Franklin's
discoveries,which had been printed in England, happened to fall
into the hands of the naturalist Buffon, who was so much impressed
that he securedthe issueof a French translation of the work; and
it was this publication which, as we have seen,gave suchoffenceto
Nollet. The successof a plan proposed by Franklin for drawing
lightning from the cloudssoonengagedpublic attention everywhere;
and in a short time the triumph of the one-fluid theory of electricity, as
the hypothesis of Watson and Franklin is generally called, was
1 He refers to it in his Paper read to the Royal Society, 18 Dec. 1755.
• cf. letters xxxvii and xxxviii, dated 1761 and 1762
• New Experiments (1750), § 15 • Letter, vii, 1751 • Letter v
• Phil. Trans. xlvii, p. 202. Watson agreed with Nollet in rejecting Franklin's theory
of the impermeability of glass.
5°
complete. Nollet, who was obdurate, , lived to seehimself the last

of his sect,exceptMonsieur B-- of Paris, his eleveand immediate
disciple.' 1
The theory of effluvia was finally overthrown, and replaced by
that of action at a distance, by the laboyrs of one of Franklin's
continental followers,Franz Ulrich Theodor Aepinus 2 (1724-1802).
Aepinus,after taking his doctor's degreein 1747 at the university of
his native city of Rostock,was appointed in 1755 a member of the
Berlin Academyof Sciences. In 1751 a young SwedenamedJohan
Carl Wi1cke3 (1732-96) came to Rostock and attended Aepinus'
lecturesthere, following him to Berlin in 1755. Here they set up
housetogether, and came very much under the influence of Euler,
who was in Berlin from 1741 to 1766. In 1757 Aepinus was trans-
lated to St Petersburg,where the rest of his active life was spent,
while Wi1ckein the sameyear returned to Sweden.
The doctrine that glassis impermeable to electricity, which had
formed the basis of Franklin's theory of the Leyden phial, was
generalisedby Aepinus" and Wi1cke into the law that all non-
conductorsare impermeableto the electric fluid. That this applies
evento air they proved by constructing a machine analogousto the
Leyden jar, in which, however, air took the place of glassas the
medium between two oppositely charged surfaces. The successof
this experiment led Aepinus to deny altogether the existence of
electric effluvia surrounding charged bodies5 : a position which he
regardedasstrengthenedby Franklin's observation,that the electric
field in the neighbourhoodof an excitedbody is not destroyedwhen
the adjacentair is blown away. The electric fluid must thereforebe
supposednot to extend beyond the excited bodiesthemselves. The
experimentof Gray, to which we have already referred, showedthat
it doesnot penetratefar into their substance; and thus it became
necessaryto supposethat the electric fluid, in its stateof rest, is con-
fined to thin layerson the surfacesof the excitedbodies. This being
granted, the attractions and repulsionsobservedbetweenthe bodies
compel us to believe that electricity acts at a distance acrossthe
intervening air.
Since two vitreously charged bodies repel each other, the force
1 Franklin'sAutQbiography
I This philosopher'ssurnamehad beenhellenisedfrom its original form Hoeckor Hoch
to al7TEtVO, by oneof his ancestors,a distinguishedtheologian.
• cr. C. W. Oseen,Johan Carl Wileke, Experimental-Fysiker,Uppsala, 1939
• F. V. T. Aepinus,Ten/amenTheonaeEleetrici/atis et Magne/ismi, St Petersburg,1759
I This was also maintained about the same time by Giacomo Battista Beccaria
(1716-81)of Turin.
(823) .51
between two particles of the electric fluid must (on Franklin's one-
fluid theory, which Aepinus adopted) be repulsive; and sincethere
is an attraction betweenoppositelychargedbodies,the forcebetween
electricity and ordinary matter must be attractive. Theseassump-
tions had beenmade, aswe have seen,by Franklin; but in order to
account for the repulsion between two resinously charged bodies,
Aepinus introduced a new supposition-namely, that the particles
of ordinary matter repel each other.l This, at first, startled his
contemporaries; but, as he pointed out, the' unelectrified' matter
with which we are acquainted is really matter saturated with its
natural quantity of the electricfluid, and the forcesdue to the matter
and fluid balance each other; or perhaps,as he suggested,a slight
want of equality betweentheseforcesmight give, as a residual, the
force of gravitation.
Assuming that the attractive and repellent forces increase as
the distancebetweenthe acting chargesdecreases,Aepinus applied
his theory to explain a phenomenonwhich had been more or less
in~efinitely observedby many previouswriters, and speciallystudied
a short time previouslyby John Canton 2 (1718-72) and by Wilcke 3
-namely, that if a conductor is brought into the neighbourhoodof
an excited body without actually touching it, the remoter portion of
the conductor acquiresan electric charge of the samekind as that
of the excitedbody, while the nearerportion acquiresa chargeof the
oppositekind. This effect,which is known asthe induction of electric
charges,had been explained by Canton himself and by Franklin 4
in terms of the theory of electric effluvia. Aepinus showedthat it
followed naturally from the theory of action at a distance,by taking
into account the mobility of the electric fluid in conductors; and
by discussingdifferent cases,so far as was possiblewith the means
at his command, he laid the foundationsof the mathematicaltheory
of electrostatics. Wilcke followed up his work of 1757 by presenting
to the SwedishAcademy in 1762 a memoir Ii in which he described
many experifnentsrelating to electric induction (including a method
of generatingany desiredamount of staticalelectricity by this means),
and also discussedthe physicsof non-conductors,adumbrating the
theory proposed long afterwards by Faraday, that a dielectric
1 It will be seenthat the relationsof' electricity' and' ordinary-matter' in Aepinus'
theory correspondcloselyto thoseof' electrons'and' atomicnuclei' in modernatomic

physics.
• Phil. Trans. xlviii (1753),p. 350
• Disputatio phyma experimentalisde electricitatibusconJrariis,Rostock,1757
• In his paperread to the Royal Societyon 18Dec. 1755
• K. Vetensk.Acad. Handlingar, vol. xxiii, 1762
52
TO THE INTRODUCTION OF THE POTkNTIALS
exposedto an electric field is thereby thrown into a state of electric

polarisation.
Aepinus did not succeedin determining the law according to
which the forcebetweentwo electric chargesvarieswith the distance
betweenthem; and the honour of having first accomplishedthis
belongsto Joseph Priestley (1733-1804),the discovererof oxygen.
Priestley,who wasa friend of Franklin's, had beeninformed by the
latter that he had found cork balls to be wholly unaffectedby the
electricityof a metal cup within which they wereheld; and Franklin
desiredPriestleyto repeat and ascertainthe fact. Accordingly, on
21 December1766,Priestley instituted experiments,which showed
that, whena hollow metallic vesselis electrified,thereis no chargeon
the inner surface(exceptnear the opening), and no electric force in
the air inside. From this he at once drew the correct conclusion,
which waspublishedin 1767.1 'May we not infer,' he says,, from
this experiment that the attraction of electricity is subject to the
samelaws with that of gravitation, and is thereforeaccordingto the
squaresof the distances; sinceit is easily demonstratedthat were
the earth in the form of a shell, a body in the inside of it would not
be attracted to onesidemore than another? '
This brilliant inferenceseemsto have beeninsufficiently studied
by the scientific men of the day; and, indeed, its author appears
to have hesitatedto claim for it the authority of a complete and
rigorousproof. Accordingly we find that the questionof the law of
forcewasnot regardedasfinally settledfor eighteenyearsafterwards.
In 1769Dr John Robison (1739-1805)of Edinburgh determinedthe
law of force by direct experiment, and found it to be that of the
inverse 2'06th power of the distancefor the repulsion of two like
charges,while for the attraction of two unlike chargeshe found the
inverseof a power lessthan the second: he conjectured that the
correct power wasthe inversesquare.2
By Franklin's law of the conservation of electric charge, and
Priestley'slaw of attraction betweenchargedbodies,electricity was
raisedto the position of an exact science. It is impossibleto men-
tion the namesof thesetwo friends in such a connection without
reflecting on the curious parallelism of their lives. In both men
there was the samecombination of intellectual boldnessand power
with moral earnestness and public spirit. Both of them carried on a
1 J. Priestley, The History and Present State of Electricity, with Original Experiments
(London, 1767), p. 732. That electrical attraction follows the law of the inverse square
had been suspectedby Daniel Bernoulli in 1760; cf. Socin's Experiments, Acta Helvetica,
iv, p. 214.
I Robison's lvlechanical Philosophy, vol. iv, pp. 73--4
53
long and tenacious struggle with the reactionary influenceswhich

dominated the English Government in the reign of George III ;
and both at last, when overpowered in the conflict, reluctantly
exchangedtheir native flag for that of the United Statesof America.
The namesof both have been held in honour by later generations,
not more for their scientific discoveriesthan for their servicesto the
causeof religious, intellectual, and political freedom.
The most celebrated electrician of Priestley's contemporaries
in London was the Hon. Henry Cavendish (1731-1810), whose
interest in the subject was indeed hereditary, for his father, Lord
Charles Cavendish,had assistedin Watson'sexperimentsof 1747.1
In 1771 CavendishI presentedto the Royal Societyan ' Attempt to
explain someof the principal phenomenaof Electricity, by meansof
an elastic fluid.' The hypothesisadopted is that of the one-fluid
theory, in much the sameform asthat of Aepinus. It was,ashe tells
us, discoveredindependently, although he becameacquaintedwith
Aepinus' work before the publication of his own paper.
In this memoir Cavendish makesno assumptionregarding the
law of force betweenelectric charges,except that it is ' inverselyas
somelesspower of the distancethan the cube'; but he evidently
inclinesto believein the law of the inversesquare. Indeed, he shows
it to be ' likely, that if the electric attraction or repulsionis inversely
as the squareof the distance1almost all the redundant fluid in the
body will be lodged close to the surface, and there pressedclose
together,and the restof the body will be saturated' ; which approxi-
matescloselyto the discoverymadefour yearspreviouslyby Priestley.
Cavendishdid, asa matter of fact, rediscoverthe inverse-squarelaw
shortly afterwards; but, indifferent to fame, he neglectedto com-
municate to others this and much other work of importance. The
value of his researcheswas not realiseduntil the middle of the nine-
teenth century, when William Thomson (Lord Kelvin) found in
Cavendish'smanuscriptsthe correct value for the ratio of the electric
chargescarried by a circular disk and a sphereof the sameradius
which had beenplacedin metallic connection. Thomsonurged that
the papers should be published; which came to pass3 in 1879,a
hundred years from the date of the great discoverieswhich they
enshrined. It was then seen that Cavendish had anticipated his
successors in severalof the ideaswhich will presentlybe discussed-
amongstothers,thoseof electrostaticcapacity and specificinductive
capacity.
1 Phil. Trans. xlv (1750), p. 67 • Phil. Trans. Ixi (1771), p. 584
• The Electrical Restarchesof the Han. Henry Cavendish, edited by J. Cleek Maxwell, 1879
54
In the published memoir of 1771 Cavendish worked out the

consequences of his fundamental hypothesismore completely than
Aepinus; and, in fact, virtually introduced the notion of electric
potential, though, in the absenceof any definite assumptionas to the
law of force,it wasimpossibleto developthis idea to any great extent.
One of the investigationswith which Cavendishoccupiedhimself
was a comparison between the conducting powers of different
materialsfor electrostaticdischarges. The question had been first
raised by Beccaria,who had shown 1 in 1753that when the circuit
through which a dischargeis passedcontains tubes of water, the
shockis morepowerful whenthe cross-section of the tubesis increased.
Cavendishwent into the matter much more thoroughly, and was
able, in a memoir presentedto the Royal Societyin 1775,2to say:
, It appearsfrom someexperiments,of which I proposeshortly to lay
an account before this Society, that iron wire conducts about 400
million timesbetter than rain or distilled water-that is, the electricity
meetswith no more resistancein passingthrough a pieceof iron wire
400,000,000incheslong than through a column of water of the same
diameter only one inch long. Sea-water,or a solution of one part
of salt in 30 of water, conducts 100times, or a saturatedsolution of
sea-saltabout 720 times, better than rain-water.'
The promised account of the experimentswas published in the
volume edited in 1879. It appears from it that the method of
testing by which Cavendishobtained theseresults was simply that
of physiologicalsensation; but the figures given in the comparison
of iron and seawater are remarkably exact.
While the theory of electricity was being establishedon a sure
foundation by the great investigatorsof the eighteenth century, a
no lessremarkable development was taking place in the kindred
scienceof magnetism,to which our attention must now be directed.
The law of attraction between magnetswas investigated at an
earlier date than the corresponding law for electrically charged
bodies. Newton, in the Principia,3 says: 'The power of gravity
is of a different nature from the power of magnetism. For the
magnetic attraction is not as the matter attracted. Some bodies
are attracted more by the magnet, others less; most bodies not at
all. The power of magnetism,in one and the samebody, may be
increasedand diminished; and is sometimesfar stronger, for the
quantity of matter, than the power of gravity; and in receding
from the magnet, decreasesnot in the duplicate, but almost in the
1 G. B. Beccaria,Dell' elettricismoartijiciale e naturale (Turin, 1753),p. 113
• Phil. Trans. !xvi (1776),p. 196 • Bk. iii, prop. vi, cor. 5
55
triplicate proportion of the distance, as nearly as I could judge

from somerude observations.'
The edition of the Principia which was published in 1742 by
Thomas Le Seur and Francis Jacquier contains a note on this
corollary, in which the correct result is obtained that the directive
couple exercisedon one magnet by another is proportional to the
inversecube of the distance.
The first discoverer1 of the law of force betweenmagneticpoles
wasJohn Michell (1724-93), at that time a youngFellow of Queens'
College, Cambridge,2who in 1750published A Treatise of Artificial
Magnets; in which is shown an easy and expeditious method of making
them superior to the best natural ones. In this he statesthe principles of
magnetictheory asfollows 3 :
'Wherever any Magnetism is found, whether in the Magnet
itself, or any piece of Iron, etc., excited by the Magnet, there are
always found two Poles, which are generally called North and
South; and the North Pole of one Magnet always attracts the
South Pole, and repelsthe North Pole of another: and vice versa.'
This is of courseadoptedfrom Gilbert.
, Each Pole attracts or repelsexactly equally, at equal distances,
in everydirection.' This, it may be observed,overthrowsthe theory
of vortices, with which it is irreconcilable. 'The Magnetical
Attraction and Repulsion are exactly equal to each other.' This,
obvious though it may seemto us, was really a most important
advance, for, as he remarks, 'Most people, who have mention'd
any thing relating to this property of the Magnet, have agreed,not
only that the Attraction and Repulsion of Magnets are not equal
to each other, but that also, they do not observethe samerule of
increaseand decrease.
'The Attraction and Repulsion of Magnets decreases,as the
Squaresof the distancesfrom the respectivepoles increase.' This
great discovery,which is the basisof the mathematical theory of
1 A suggestion of the inverse-square law of attraction between magnetic poles had
however been made by the Cardinal Nicholas of Cusa (1401-<i4) in 1450; cf. Nicolai de
Cusa Opera Omnia, ed. L. Bauer (Lipsiae, 1937), v, p. 127. Of course it was the
statical notion of force (measured by balancing weights) that was involved, as the
dynamical notion was not yet discovered.
• Michell had taken his degree only two years previously. Later in life he was on
terms of friendship with Priestley, Cavendish and William Herschel. The plan of
determining the density of the earth, which was carried out by Cavendish in 1798and is
generally known as the ' Cavendish Experiment,' was due to Michell. Michell was the
first inventor of the torsion balance; he also made many valuable contributions to
astronomy. In 1767 he became Rector of Thornhill, Yorks, and lived tllcre until his
death. cf. Memoirs of John Mich411by Sir A. Geikie, Cambridge, 1918.
• loco cit. p. 17
56
Magnetism, was deduced partly from his own observations,and

partly from those of previous investigators (e.g. Dr Brook Taylor
and P. Musschenbroek),who, as he observes,had made accurate
experiments,but had failed to take into account all the considera-
tions necessaryfor a sound theoretical discussionof them.
After Michell the law of the inverse square was maintained
by Tobias Mayer 1 (1723-62) of Gottingen, better known as the
author of Lunar Tables which were long in use; and by the cele-
brated mathematician,Johann Heinrich Lambert 2 (1728-77).
The promulgation of the one-fluid theory of electricity, in the
middle of the eighteenth century, naturally led to attempts to
construct a similar theory of magnetism; this was effectedin 1759
by Aepinus,3who supposedthe 'poles' to be places at which a
magnetic fluid was present in amount exceedingor falling short of
the normal quantity. The permanenceof magnetswas accounted
for by supposingthe fluid to be entangled in their pores, so as to
be with difficulty displaced. The particles of the fluid were assumed
to repel each other, and to attract the particles of iron and steel;
but, as Aepinus saw, in order to satisfactorily explain magnetic
phenomena it was necessaryto assumealso a mutual repulsion
amongthe material particles of the magnet.
Subsequentlytwo imponderable magnetic fluids, to which the
names boreal and austral were assigned,were postulated by the
Hollander, Anton Brugmans (1732-89) and by Wilcke. These
fluids were supposedto have properties of mutual attraction and
repulsionsimilar to thosepossessedbyvitreousand resinouselectricity.
The writer who next claims our attention for his servicesboth
to magnetism and to electricity is the French physicist, Charles
Augustin Coulomb' (1736-1806). By aid of the torsion balance,
which was independently invented by Michell and himself, he
verified in 1785 Priestley'sfundamental law that the repulsive force
betweentwo small globescharged with the samekind of elec:tricity
is in the inverseratio of the squareof the distanceof their centres.
In the secondmemoir he extended this law to the attraction of
oppositeelectricities.
Coulomb did not accept the one-fluid theory of Franklin,
Aepinus and Cavendish,but preferred a rival hypothesiswhich had
1 Noticedin Gallinger Gelehrter Anzeiger (1760); cf. Aepinus,Nov. Comm. Acad. Petrop.
(1768)and Mayer's Opera lnedila, herausg.von G. C. Lichtenberg

• Histoire de [,Acad. de Berlin (1766),pp. 22,49
• In the Ten/amen, to which referencehasalreadybeenmade
• Coulomb's First, Secondand Third Memoirs appear in Mem. de l'Acad., 1785;
the Fourth in 1786,the Fifth in 1787,the Sixth in 1788and the Seventhin 1789.
57
beenproposedin 1759by Robert Symmer1 asa developmentof the

ideas published by du Fay in 1733. 'My notion,' said Symmer,
, is that the operationsof electricity do not depend upon one single
positive power, according to the opinion generally received; but
upon two distinct, positive, and active powers, which, by contrast-
ing, and, as it were, counteracting each other produce the various
phenomena of electricity; and that, when a body is said to be
positively electrified, it is not simply that it is possessed
of a larger
shareof electric matter than in a natural state; nor, when it is said
to be negatively electrified, of a less; but that, in the former case,
it is possessed of a larger portion of one of thoseactive powers,and
in the latter, of a larger portion of the other; while a body in its
natural state remains unelectrified, from an equal ballance of those
two powerswithin it.'
Coulomb developed this idea: 'Whatever be the cause of
electricity,' he says,2' we can explain all the phenomenaby suppos-
ing that there are two electric fluids, the parts of the samefluid
repelling eachother according to the inversesquareof the distance,
and attracting the parts of the other fluid according to the same
inversesquarelaw.' ' The suppositionof two fluids,' he adds, 'is
moreover in accord with all those discoveriesof modern chemists
and physicists,which have made known to us various pairs of gases
whose elasticity is destroyed by their admixture in certain pro-
portions-an effect which could not take place without something
equivalent to a repulsion between the parts of the same gas,
which is the causeof its elasticity, and an attraction between the
parts of different gases,which accountsfor the loss of elasticity on
combination.'
According, then, to the two-fluid theory, the 'natural fluid'
contained in all matter can be decomposed,under the influence of
an electric field, into equal quantities of vitreous and resinous
electricity, which, if the matter be conducting, can then fly to the
surfaceof the body. The abeyanceof the characteristicproperties
of the opposite electricities when in combination was sometimes
further compared to the neutrality manifestedby the compoundof
an acid and an alkali.
The publication of Coulomb's views led to some controversy
between the partisansof the one-fluid and two-fluid theories; the
latter wassoongenerallyadoptedin France,but wasstoutly opposed
in Holland by Van Marum and in Italy by Volta. The chief
difference between the rival hypothesesis that, in the two-fluid
1 Phil. Trans.Ii (1759), p. 371 • Sixth Memoir, p. 561
58
theory, both the electric fluids are movable within the substanceof
a solid conductor; while in the one-fluid theory the actual electric
fluid is mobile, but the particles of the conductor are fixed. The
dispute could therefore be settled only by a determination of the
actual motion of electricity in discharges; and this was beyond
the reachof experiment.
In his Fourth Memoir Coulomb showed that electricity in
equilibrium is confined to the surface of conductors, and does
not penetrate to their interior substance; and in the Sixth
Memoir 1 he virtually establishesthe result that the electric
force near a conductor is proportional to the surface-densityof
electrification.
Sincethe overthrowof the doctrine of electriceffluvia by Aepinus,
the aim of electricianshad been to establishtheir scienceupon the
foundation of a law of action at a distance,resembling that which
had led to such triumphs in Celestial Mechanics. When the law
first stated by Priestley was at length decisively established by
Coulomb, its simplicity and beauty gaverise to a generalfeeling of
completetrust in it asthe best attainable conceptionof electrostatic
phenomena. The result was that attention was almost exclusively
focusedon action-at-a-distancetheories,until the time, long after-
wards, when Faraday led natural philosophersback to the right
path.
Coulomb rendered great servicesto magnetic theory. It was
he who in 1777, by simple mechanical reasoning, completed the
overthrow of the hypothesisof vortices.2 He also, in the second
of the Memoirsalready quoted,3 confirmed Michell's law, according
to which the particles of the magnetic fluids attract or repel each
other with forcesproportional to the inversesquareof the distance.
Coulomb, however,went beyond this, and endeavouredto account
for the fact that the two magnetic fluids, unlike the two electric
fluids, cannot be obtained separately; for when a magnetis broken
into two pieces,one containing its north and the other its south
pole, it is found that eachpieceis an independentmagnetpossessing
two poles of its own, so that it is impossibleto obtain a north or
south pole in a state of isolation. Coulomb explained this by sup-
posing4 that the magneticfluids are permanently imprisonedwithin
the moleculesof magnetic bodies,so as to be incapable of crossing
from one moleculeto the next; each moleculetherefore under all
1 p. 677 I Mem. presentes
par diversSavans,ix (1780), p. 165
I Mem. de ['Acad. (1785), p. 593. Gaussfinally establishedthe law by a much more
refined method.
• In his SeventhMemoir, Mem. de l'Acad. (1789),p. 488
59
circumstancescontainsas much of the boreal asof the austral fluid,

and magnetisationconsistssimply in a separationof the two fluids
to opposite ends of each molecule. Such a hypothesisevidently
accountsfor the impossibility of separatingthe two fluids to opposite
endsof a body of finite size.
In spite of the advanceswhich have beenrecounted,the mathe-
matical developmentof electric and magnetic theory was scarcely
begun at the closeof the eighteenthcentury; and many erroneous
notions were still widely entertained. In a Report 1 which was
presentedto the French Academy in 1800, it was assumedthat the
mutual repulsion of the particles of electricity on the surfaceof a
body is balanced by the resistanceof the surrounding air; and for
long afterwards the electric force outside a charged conductor was
confusedwith a supposedadditional pressurein the atmosphere.
Electrostatical theory was, however,suddenlyadvancedto quite
a mature stateof developmentby SimeonDenisPoisson(1781-1840),
in a memoir which was read to the French Academy in 1812.2 As
the opening sentencesshow, he accepted the conceptions of the
two-fluid theory.
, The theory of electricity which is most generally accepted,'he
says, 'is that which attributes the phenomena to two different
fluids, which are contained in all material bodies. It is supposed
that moleculesof the same fluid repel each other and attract the
moleculesof the other fluid; theseforcesof attraction and repulsion
obey the law of the inversesquareof the distance; and at the same
distancethe attractive poweris equalto the repellentpower; whence
it follows that, when all the parts of a body contain equal quantities
of the two fluids, the latter do not exert any influence on the fluids
contained in neighbouring bodies, and consequentlyno electrical
effects are discernible. This equal and uniform distribution of the
two fluids is called the natural state; when this state is disturbed
in any body, the body is said to be electrified, and the variousphenom-
ena of electricity begin to take place.
, Material bodiesdo not all behavein the sameway with respect
to the electric fluid. Some,suchasthe metals,do not appearto exert
any influence on it, but permit it to move about freely in their
substance; for this reasonthey are called conductors; others,on the
contrary-very dry air, for example-oppose the passageof the
electric fluid in their interior, so that they can prevent the fluid
accumulatedin conductorsfrom being dissipatedthroughout space.'
1 On Volta's discoveries
• Mlm. del'Institut (181l), pt. i, p. 1; pt. ii, p. 163
60
When an excessof one of the electric fluids is communicatedto a

metallic body, this charge distributes itself over the surfaceof the
body, forming a layer whosethicknessat any point dependson the
shapeof the surface. The resultant force due to the repulsionof all
the particles of this surface-layermust vanish at any point in the
interior of the conductor, sinceotherwisethe natural state existing
there would be disturbed; and Poissonshowedthat by aid of this
principle it is possiblein certain casesto determine the distribution
of electricity in the surfacelayer. For example,a well-known propo-
sition of the theory of attractions assertsthat a hollow shell whose
bounding surfacesare two similar and similarly situated ellipsoids
exercisesno attractive force at any point within the interior hollow;
and it may thencebe inferred that if an electrifiedmetallic conductor
has the form of an ellipsoid, the charge will be distributed on it
proportionally to the normal distancefrom the surfaceto an adjacent
similar and similarly situated ellipsoid.
Poissonwent on to show that this result was by no meansall
that might with advantagebe borrowed from the theory of attrac-
tions. Lagrange,in a memoir on the motion of gravitating bodies,
had shown1 that the componentsof the attractive force at any point
can be simply expressedas the derivates of the function which is
obtained by adding together the massesof all the particles of an
attracting system,eachdivided by its distancefrom the point; and
Laplacehad shown 2 that this function V satisfiesthe equation
o=V + ?=V + ozV = 0

ax· Oy. oz=
in spacefree from attracting matter. Poissonhimself showedlater,
in 1813,3that when the point (x,y, z) is within the substanceof the
attracting body, this equation of Laplace must be replacedby
o;~+ ~~~+ ~~~ = _ %p,
where p denotesthe density of the attracting matter at the point.

In the presentmemoir Poissoncalled attention to the utility of this
function V in electrical investigations,remarking that its value over
the surfaceof any conductor must be constant.
1 Mtm. de Berlin (1777). The theorem'tVasafterwardspublished,and ascribedto

Laplace,in a memoirby Legendreon the Attractionsof Spheroids,which will be found
in the Mtm. par divers S(Wans,publishedin 1785.
• Mem. de l'Acad. (1782,publishedin 1785).p. 113
• Bull. de la Soc. Philomathique iii (1813),p. 388
61
The known formulae for the attractions of spheroidsshow that

when a charged conductor is spheroidal, the repellent force acting
on a small chargedbody immediately outsideit will be directed at
right anglesto the surfaceof the spheroid,and will be proportional
to the thicknessof the surface-layerof electricity at thisplace. Poisson
suspectedthat this theoremmight be true for conductorsnot having
the spheroidal form-a result which, as we have seen,had been
already virtually given by Coulomb; and Laplace suggestedto
Poissonthe following proof, applicable to the general case. The
force at a point immediately outside the conductor can be divided
into a part s due to the part of the charged surfaceimmediately
adjacent to the point, and a part S due to the rest of the surface.
At a point closeto this, but just inside the conductor, the force S
will still act; but the force s will evidently be reversedin direction.
Since the resultant force at the latter point vanishes,we must have
S = s; so the resultant force at the exterior point is 2s. But s is
proportional to the chargeper unit area of the surface,asis seenby
considering the case of an infinite plate; which establishesthe
theorem.
When several conductors are in presenceof each other, the
distribution of electricity on their surfacesmay be determinedby the
principle, which Poissontook as the basisof his work, that at any
point in the interior of anyone of the conductors,the resultantforce
dueto all the surface-layersmustbezero. He discussed, in particular,
one of the classicalproblemsof electrostatics,namely that of deter-
mining the surface-densityon two chargedconductingspheresplaced
at any distance from each other. The solution depends on
double gamma functions in the general case; when the two
spheresare in contact, it dependson ordinary gamma functions.
Poissongave a solution in terms of definite integrals, which is
equivalent to that in terms of gamma functions; and after re-
ducing his results to numbers, compared them with Coulomb's
experiments.
The rapidity with which in a singlememoir Poissonpassedfrom
the barest elementsof the subject to such recondite problems as
thosejust mentioned may well excite admiration. His successis,
no doubt, partly explainedby the high stateof developmentto which
analysis had been advanced by the great mathematiciansof the
eighteenth century; but even after allowance has been made for
what is due to his predecessors,Poisson'sinvestigation must be
accounteda splendid memorial of his genius.
Some years later Poissonturned his attention to magnetism;
62
and, in a masterlypaper 1presentedto the French Academyin 1824,

gavea remarkably completetheory of the subject.
His starting-point is Coulomb's doctrine of two imponderable
magnetic fluids, arising from the decompositionof a neutral fluid,
and confined in their movementsto the individual elementsof the
magneticbody, soasto be incapableof passingfrom one elementto
the next.
Supposethat an amount m of the positive magnetic fluid is
located at a point (x,y, z) ; the componentsof the magnetic intensiry,
or force exerted on unit magnetic pole, at a point (t, "1, ') will
evidently be
- m ~(~)
ot r ' -m ~0"1(~)
r '
- m ~(~),
0' r
where r denotes {(t - x)t + ("1 - y)t + (' - Z)t}l. Hence if we
considernext a magnetic element in which equal quantities of the
two magnetic fluids are displaced from each other parallel to the
x-axis,the componentsof the magnetic intensity at (t, "1, ') will be
the negative derivates,with respect to t, "1, , respectively,of the
function
A ~(~),
ox r
where the quantity A, which does not involve (~, "1, {), may be
called the magnetic moment of the element; it may be measuredby
the couple required to maintain the element in equilibrium at a
definite angular distancefrom the magnetic meridian.
If the displacementof the two fluids from each other in the
element is not parallel to the axis of x, it is easily seen that the
expressioncorrespondingto the last is
A ~(~)
ox r + B~(~)
(!J r
+ C oz
~(~)r '
wherethe vector (A, B, C) now denotesthe magneticmoment of the
element.
Thus the magnetic intensity at an external point (~, "1, ') due to
any magneticbody has the components
(- ~~,- ~¥,- ~~),

HI + + G)
where
V = (A~ B~ C~) fix dy tk
integrated throughout the substanceof the magnetic body, and
I Mem. de Z'Aead. v, p. 247
63
where the vector (A, B, C) or 11 representsthe magneticmoment

per unit-volume, or, as it is generally called, the magnetisation. The
function V wasafterwardsnamedby Green the magnetic potential.
Poisson,by integrating by parts the precedingexpressionfor the
magneticpotential, obtained it in the form
V = IH (I. dS) - HH div 1 dx dy dZ,2
the first integral being taken over the surface S of the magnetic
body, and the secondintegral being taken throughout its volume.
This formula showsthat the magnetic intensity produced by the
body in external spaceis the same as would be produced by a
fictitious distribution of magnetic fluid, consistingof a layer over
its surface,of surface-charge(I . dS) per elementdS, togetherwith
a volume distribution of density - div 1 throughout its substance.
These fictitious magnetisationsare generally known as Poisson's
equivalent surface- and volume-distributions q[ magnetism.
Poisson,moreover, perceived that at a point in a very small
cavity excavatedwithin the magneticbody, the magneticpotential
hasa limiting value which is independentof the shapeof the cavity
asthe dimensionsof the cavity tend to zero; but that this is not true
of the magnetic intensity, which in such a small cavity depends
on the shapeof the cavity. Taking the cavity to be spherical, he
showedthat the magneticintensity within it is
grad V +i 7T1,3
where1 denotesthe magnetisationat the place.

This memoir also contains a discussionof the magnetismtem-
porarily induced in soft iron and other magnetisablemetals by the
approach of a permanent magnet. Poisson accounted for the
properties of temporary magnets by assumingthat they contain
embeddedin their substancea great number of small spheres,which
are perfect conductorsfor the magneticfluids; so that the resultant
magnetic intensity in the interior of one of thesesmall spheresmust
be zero. He showedthat such a sphere,when placed in a field of
1 In the present work, vectors will generally be distinguished by heavy type.
I If the components of a vector a are denoted by (ax. ay, at), the quantity axbx + avby
+ azh. is called the scalarproduct of two vectors a and b, and is denoted by (a . b).
The quantity 0:: + 0:; + ~. is called the divergenceof the vector a, and is denoted
by diva.
OV OV OV.
I The vector whose components are - ox ' - 8.1-' - 0.1;; IS denoted by grad V.
64
magneticintensity F, must acquire a magnetic moment of amount

~ F X the volume of the sphere,in order to counteract within the
spherethe forceF. Thus if kp denotethe total volumeof thesespheres
contained within a unit volume of the temporary magnet, the
magnetisationwill be I, where
!111 = kpF,
and F denotes the magnetic intensity within a spherical cavity
excavatedin the body. This is Poisson's law of induced magnetism.
It is known that some substancesacquire a greater degree of
temporary magnetisation than others when placed in the same
circumstances. Poissonaccounted for this by supposingthat the
quantity kp varies from one substanceto another. But the experi-
mental data show that for soft iron kp must have a value very near
unity, which would obviouslybe impossibleif kp is to mean the ratio
of the volume of spherescontained within a region to the total
volume of the region.l The physical interpretation assignedby
Poissonto his formulae must therefore be rejected, although the
formulae themselvesretain their value.
Poisson'selectricaland magneticalinvestigationsweregeneralised
and extended in 1828 by George Green 2 (1793-1841). Green's
treatment is basedon the propertiesof the function already usedby
Lagrange,Laplace and Poisson,which representsthe sum of all the
electric or magneticchargesin the field, divided by their respective
distancesfrom somegiven point. To this function Green gave the
namepotential, by which it has alwayssincebeenknown.3
Near the beginning of the memoir is establishedthe celebrated
formula connecting surface and volume integrals, which is now
generallycalled Green's Theorem, and of which Poisson'sresult on the
equivalent surface-and volume-distributions of magnetisation is a
particular application. By using this theorem to investigate the
propertiesof the potential, Green arrived at many resultsof remark-
able beauty and interest. We need only mention, as an example
of the power of his method, the following: supposethat there is a
hollow conducting shell, bounded by two closedsurfaces,and that a
number of electrified bodies are placed, some within and some
1 This objection was advanced by Maxwell in sect. 430 of his Treatise. An attempt to
overcome it was made by E. Betti; cr.
his Teoria d.fone Newtoniane (Pisa, 1879).
• An essayon the application of mathematical ana!ysis to the theoriesof electricity and magnetism
(Nottingham, 1828) ; reprinted in The Mathematical Papers of the late GeorgeGreen, p. I.
• Euler in 1744 (De methodis inveniendi ... ) had spoken of the vis potentialis-what
would now be called the potential energy-possessed by an elastic body when bent.
65
without it; and let the inner surfaceand interior bodiesbe called
the interior system,and the outer surface and exterior bodies be
called the exterior system. Then all the electrical phenomenaof the
interior system,relative to attractions, repulsionsand densities,will
be the sameasif there were no exterior system,and the inner surface
were a perfect conductor, put in communication with the earth;
and all thoseof the exterior systemwill be the sameasif the interior
systemdid not exist,and the outer surfacewerea perfect conductor,
containing a quantity of electricity equal to the whole of that
originally contained in the shell itself and in all the interior bodies.
It will be evident that electrostaticshad by this time attained
a state of development in which further progresscould be hoped
for only in the mathematical superstructure, unless experiment
should unexpectedly bring to light phenomenaof an entirely new
character. This will therefore be a convenient place to pauseand
considerthe rise of another branch of electrical philosophy.
66
Chapter III
GALVANISM, FROM GALVAN I TO OHM.
UNTIL the last decadeof the eighteenth century, electricians were

occupied solely with statical electricity. Their attention was then
turned in a different direction.
In a work entitled Recherchessur l'origine des sentiments agreables et
disagreables, which was published 1 in 1752,Johann Georg Sulzer
(1720-79) had mentioned that if two piecesof metal, the one of
lead and the other of silver, be joined together in such a manner
that their edgestouch, and if they be placed on the tongue, a taste
is perceived 'similar to that of vitriol of iron,' although neither
of thesemetals applied separatelygives any trace of such a taste.
, It is not probable,' he says,'that this contact of the two metals
causesa solution of either of them, liberating particles which might
affect the tongue; and we must thereforeconcludethat the contact
setsup a vibration in their particles, which, by affecting the nerves
of the tongue,producesthe tastein question.'
This observationwas not suspectedto have any connection with
electrical phenomena,and it played no part in the inception of the
next discovery,which indeed was suggestedby a mere accident.
Luigi Galvani, born at Bologna in 1737, occupied from 1775
onwards a chair of Anatomy in his native city. For many years
before the event which made him famous he had been studying
the susceptibility of the nerves to irritation; and, having been
formerly a pupil of Beccaria, he was also interested in electrical
experiments. One day in the latter part of the year 1780he had,
as he tells US,2 'dissected and prepared a frog, and laid it on a
table, on which, at some distance from the frog, was an electric
machine. It happenedby chancethat one of my assistantstouched
the inner crural nerve of the frog, with the point of a scalpel;
whereuponat oncethe musclesof the limbs wereviolently convulsed.
, Another of thosewho usedto help me in electrical experiments
thought he had noticed that at this instant a spark was drawn from
the conductor of the machine. I myself was at the time occupied
with a totally different matter; but when he drew my attention to
J Mem. de ['A cad. de Berlin (1752), p. 356
• Aloysii Galvani, De Viribus Electricitatis in Motu Musculari: Commentarii Bononiensi,
vii (1791), p. 363
(823) 67
GALVANISM
this, I greatly desired to try it for myself, and discoverits hidden

principle. So I, too, touched one or other of the crural nerveswith
the point of the scalpel,at the sametime that one of thosepresent
drew a spark; and the samephenomenonwas repeatedexactly as
before.' 1
After this, Galvani had the idea of trying whether the electricity
of thunderstormswould induce muscular contractions equally well
with the electricity of the machine. Having successfullyexperi-
mented with lightning, he 'wished,' ashe writes,2' to try the effect
of atmosphericelectricity in calm weather. My reasonfor this was
an observation I had made, that frogs which had been suitably
prepared for theseexperimentsand fastened,by brasshooksin the
spinal marrow, to the iron lattice round a certain hanging-garden
at my house,exhibited convulsionsnot only during thunderstorms,
but sometimeseven when the sky was quite serene. I suspected
theseeffectsto be due to the changeswhich take place during the
day in the electric state of the atmosphere; and so, with some
degree of confidence, I performed experiments to test the point;
and at different hours for many days I watched frogs which I had
disposedfor the purpose; but could not detect any motion in their
muscles. At length, weary of waiting in vain, I pressedthe brass
hooks, which were driven into the spinal marrow, against the iron
lattice, in order to see whether contractions could be excited by
varying the incidental circumstancesof the experiment. I observed
contractions tolerably often, but they did not seem to bear any
relation to the changesin the electrical state of the atmosphere.
, However, at this time, when asyet I had not tried the experiment
except in the open air, I came very near to adopting a theory that
the contractions are due to atmospheric electricity, which, having
slowly entered the animal and accumulated in it, is suddenly dis-
charged when the hook comes in contact with the iron lattice.
For it is easyin experimenting to deceiveourselves,and to imagine
we seethe things we wish to see.
'But I took the animal into a closed room, and placed it on
an iron plate; and when I pressedthe hook which was fixed in
the spinal marrow against the plate, behold! the samespasmodic
contractions as before. I tried other metals at different hours on
various days, in several places, and always with the same result,
1 According to a story which has often been repeated, but which rests on no sufficient
evidence, the frog was one of a number which had been procured for the Signora Galvani
who, being in poor health, had been recommended to take a soup made of these animals
as a restorative.
I loco cit. p. 377
68
FROM GALVANI TO OHM
except that the contractions were more violent with some metals
than with others. After this I tried various bodies which are not
conductorsof electricity, such as glass,gums, resins,stonesand dry
wood; but nothing happened. This was somewhatsurprising, and
led me to suspectthat electricity is inherent in the animal itself.
This suspicionwas strengthenedby the observation that a kind of
circuit of subtle nervousfluid (resemblingthe electric circuit which
is manifestedin the Leyden jar experiment) is completed from the
nervesto the muscleswhen the contractions are produced.
'For, while I with one hand held the prepared frog by the
hook fixed in its spinal marrow, so that it stood with its feet on
a silver box, and with the other hand touched the lid of the box,
or its sides,with any metallic body, I was surprised to seethe frog
becomestrongly convulsedevery time that I applied this artifice.' 1
Galvani thus ascertainedthat the limbs of the frog are convulsed
whenever a connection is made between the nerves and muscles
by a metallic arc, generally formed of more than one kind of metal ;
and he advancedthe hypothesisthat the convulsionsare causedby
the transport of a peculiar fluid from the nervesto the muscles,the
arc acting as a conductor. To this fluid the names Galvanism and
Animal Electricity were soon generally applied. Galvani himself
consideredit to be the same as the ordinary electric fluid, and,
indeed, regarded the entire phenomenonas similar to the discharge
of a Leydenjar.
The publication of Galvani's views soon engagedthe attention
of the learned world, and gave rise to an animated controversy
between those who supported Galvani's own view, those who
believed galvanism to be a fluid distinct from ordinary electricity,
and a third school who altogether refused to attribute the effects
to a supposedfluid contained in the nervous system. The leader
of the last-namedparty wasAlessandroVolta (1745-1827), Professor
of Natural Philosophy in the University of Pavia, who in 1792 put
forward the view 2 that the stimulus in Galvani's experiment is
derived essentiallyfrom the connection of two different metals by a
moist body. 'The metals used in the experiments, being applied
to the moist bodiesof animals,can by themselves,and of their proper
virtue, excite and dislodge the electric fluid from its state of rest;
so that the organs of the animal act only passively.' At first he
inclined to combine this theory of metallic stimulus with a certain
degree of belief in such a fluid as Galvani had supposed; but
1 This observationwas made in 1786.
• Phil. Trans. !xxxiii (1793), pp. 10,27
69
GALVANISM
after the end of 1793 he denied the existenceof animal electricity

altogether.
From this standpoint Volta continued his experiments and
worked out his theory. The following quotation from a letter 1
which he wrote later to Gren, the editor of the Neues Journal d.
Physik, setsforth his view in a more developedform:
, The contact of different conductors,particularly the metallic,
including pyrites and other minerals, as well as charcoal, which
I call dry conductors, or of the first class, with moist conductors,
or conductors of the second class, agitates or disturbs the electric
fluid, or gives it a certain impulse. Do not ask in what manner:
it is enough that it is a principle, and a general principle. This
impulse, whether produced by attraction or any other force, is
different or unlike, both in regard to the different metals and to
the different moist conductors; so that the direction, or at least
the power, with which the electric fluid is impelled or excited, is
different when the conductor A is applied to the conductor B, or
to another C. In a perfect circle of conductors,where either one
of the secondclassis placed betweentwo different from eachother
of the first class,or, contrariwise, one of the first classis placed
betweentwo of the secondclassdifferent from eachother, an electric
streamis occasionedby the predominating force either to the right
or to the left-a circulation of this fluid, which ceasesonly when
the circle is broken, and which is renewedwhen the circle is again
renderedcomplete.'
Another philosopher who, like Volta, denied the existenceof
a fluid peculiar to animals, but who took a somewhatdifferent
view of the origin of the phenomenon, was Giovanni Fabroni
(1752-1822) of Florence,who in 1796, having placed two platesof
different metalsin water, observedI that one of them waspartially
oxidised when they were put in contact; from which he rightly
concludedthat somechemical action is inseparablyconnectedwith
galvanic effects.
The feeble intensity of the phenomena of galvanism, which
compared poorly with the striking displays obtained in electro-
statics,wasresponsiblefor somefalling off of interestin them towards
the end of the eighteenth century; and the last years of their
illustrious discovererwere clouded by misfortune. Being attached
to the old order which was overthrown by the armiesof the French
Revolution, he refusedin 1798 to take the oath of allegianceto the
1 Phil. Mag. iv (1799), pp. 59, 163, 306
• Phil. Journal, 4to, iii, p. 308; iv, p. 120; Journal de Physique, vi, p. 348
7°
FROM GALVANI TO OHM
newly constituted Cisalpine Republic, and was deposedfrom his

professorialchair. A profound melancholy,which had beeninduced
by domesticbereavement,wasaggravatedby poverty and disgrace;
and, unable to survive the lossof all he held dear, he died broken-
heartedbeforethe end of the year.l
Scarcelymore than a year after the death of Galvani, the new
sciencesuddenlyregainedthe eagerattention of philosophers. This
renewal of interest was due to the discoveryby Volta, in the early
spring of 1800,of a meansof greatly increasingthe intensity of the
effects. Hitherto all attempts to magnify the action by enlarging
or multiplying the apparatushad endedin failure. If a long chain
of different metalswas usedinsteadof only two, the convulsionsof
the frog were no more violent. But Volta now showed2 that if any
number of couples,eachconsistingof a zinc disk and a copper disk
in contact, were taken, and if each couple was separatedfrom the
next by a disk of moistened pasteboard (so that the order was
copper,zinc, pasteboard,copper, zinc, pasteboard,etc.), the effect
of the pile thus formed was much greater than that of any galvanic
apparatus previously introduced. When the highest and lowest
disks were simultaneouslytouched by the fingers, a distinct shock
was felt; and this could be repeated again and again, the pile
apparently possessing within itself an indefinite power of recupera-
tion. It thus resembleda Leyden jar endowed with a power of
automaticallyre-establishingits stateof tensionafter eachexplosion;
with, in fact, 'an inexhaustible charge, a perpetual action or
impulsion on the electric fluid.'
Volta unhesitatingly pronounced the phenomena of the pile
to be in their nature electrical. The circumstancesof Galvalli's
original discoveryhad prepared the minds of philosophersfor this
belief, which was powerfully supported by the similarity of the
physiologicaleffectsof the pile to thoseof the Leyden jar, and by
the observationthat the galvanic influence was conducted only by
those bodies-e.g. the metals-which were already known to be
goodconductorsof static electricity. But Volta now supplied a still
more convincing proof. Taking a disk of copper and one of zinc,
he held each by an insulating handle and applied them to each
other for an instant. Mter the diskshad beenseparated,they were
brought into contact with a delicate electroscope,which indicated
by the divergenceof its strawsthat the diskswere now electrified-
1 A decree of reinstatement had been granted, but had not come into operation at
the time of Galvani's death.

S Phil. Trans. (1800), p. 403
GALVANISM
the zinc had, in fact, acquired a positive and the copper a negative
electric charge.1 Thus the mere contact of two different metals,
such as those employed in the pile, was shown to be sufficient for
the production of effectsundoubtedly electrical in character.
On the basisof this result a definite theory of the action of the
pile was put forward. It was not originally published by Yolta
himself, but it is contained in a Report which Biot wrote for a
committee appointed to examine Yolta's work.2 Supposefirst that
a disk of zinc is laid on a disk of copper, which in turn restson an
insulating support. The experimentjust describedshowsthat the
electric fluid will be driven from the copper to the zinc. We may
then, according to Yolta, represent the state or 'tension' of the
copper by the number - t, and that of the zinc by the number
+ t, the difference being arbitrarily taken as unity, and the sum
being (on accountof the insulation) zero. It will be seenthat Volta's
idea of ' tension' was a mingling of two ideas, which in modern
electric theory are clearly distinguished from each other-namely,
electric chargeand electric potential.
Now let a disk of moistenedpasteboardbe laid on the zinc, and
a disk of copper on this again. Since the uppermost copper is not
in contact with the zinc, the contact action does not take place
between them; but sincethe moist pasteboardis a conductor, the
copper will receivea chargefrom the zinc. Thus the stateswill now
be representedby -1for the lower copper, + ! for the zinc, and
+ 1for the upper copper,giving a zero sum asbefore.
If, now, another zine disk is placed on the top, the stateswill be
representedby - 1 for the lower copper, 0 for the lower zine and
upper copper, and +
1 for the upper zinc.
In this way it is evident that the differencebetweenthe numbers
indicating the tensionsof the uppermost and lowest disksin the pile
will alwaysbe equal to the number of pairs of metallic diskscontained
in it. If the pile is insulated, the sum of the numbersindicating the
statesof all the disksmust be zero; but if the lowestdisk is connected
to earth, the tension of this disk will be zero, and the numbers
indicating the statesof all the other disks will be increasedby the
sameamount, their mutual differencesremaining unchanged.
The pile as a whole is thus similar to a Leydenjar; when the
experimenter touches the uppermost and lowest disks, he receives
1 Abraham Bennet (1750-99) had previously shown (New Experiments in Electricity
[t 789], pp. 86-102) that many bodies, when separated after contact, are oppositely
electrified; he conceived that different bodies have different attractions or capacities
for ('1ectricity.
I cf. Opere di Alessander Volta, cd. Naz. (Milan, 1918- ), ii, p. 110
72
FROM GALVANI TO OHM
the shock of its discharge, the intensity being proportional to the

number of disks.
The moist layers played no part in Volta's theory beyond that
of conductors.l It wassoonfound that when the moisture is acidified
the pile is more efficient, but this was attributed solely to the
superiorconducting power of acids.
Volta fully understood and explained the impossibility of con-
structing a pile from disks of metal alone, without making use of
moist substances. As he showedin 1801,if disks of various metals
are placed in contact in any order, the extreme metalswill be in the
samestate as if they touched each other directly without the inter-
vention of the others; so that the whole is equivalent merely to a
singlepair. When the metalsare arrangedin the order silver, copper,
iron, tin, lead, zinc, each of them becomespositive with respectto
that which precedesit, and negativewith respectto that which follows
it; but the moving force from the silver to the zinc is equal to the
sumof the moving forcesof the metals comprehendedbetweenthem
in the series.
When a connection was maintained for some time between the
extreme disks of a pile by the human body, sensationswere ex-
perienced which seemedto indicate a continuous activity in the
entire system. Volta inferred that the electric current persistsduring
the whole time that communication by conductors existsall round
the circuit, and that the current is suspendedonly when this com-
munication is interrupted. 'This endlesscirculation or perpetual
motion of the electric fluid,' he says,' may seemparadoxical, and
may prove inexplicable; but it is none the lessreal, and we can, so
to speak,touch and handle it.'
Volta announced his discovery in a letter to Sir Joseph Banks,
dated from Como, 20 March 1800. Sir Joseph, who was then
Presidentof the Royal Society, communicated the news to William
Nicholson (1753-1815), founder of the Journal which is generally
known by his name, and his friend Anthony Carlisle (1768-1840),
afterwards a distinguished surgeon. On the 30th of the following
month, Nicholson and Carlisle setup the first pile made in England.
In repeating Volta's experiments, having made the contact more
secureat the upper plate of the pile by placing a drop of water
there, they noticed 2 a disengagementof gas round the conducting
wire at this point; whereupon they followed up the matter by
1 Volta had inclined, in his earlier experiments on galvan~m, to locate the seat of
power at the interfaces of the metals with the moist conductors. Cf. his letter to Gren,
Phil. Mng. iv (1799), p. 62.
I Nidwlson's Journal (4to), iv, p. 179 (1800) ; Phil. Mag. vii, p. 337 (1800)
73
GALVANISM
introducing a tube of water, into which the wires from the terminal
of the pile were plunged. Bubbles of an inflammable gas wer
liberated at one wire, while the other wire becameoxidised; whe
platinum wires were used, oxygen and hydrogen were evolvedin
free state,one at eachwire. This effect, which wasnothing lessthan:
the electric decomposition of water into its constituent gases,was
obtained on 2 May 1800.1
Although it had long been known that frictional electricity is
capableof inducing chemical action,2the discoveryof Nicholsonand
Carlisle was of the first magnitude. It was at once extended by
William Cruickshank (1745-1800) of Woolwich, who showed3 that
solutionsof metallic salts are also decomposedby the current; and
William Hyde Wollaston (1766-1828) seizedon it as a test" of the
identity of the electric currents of Yolta with thoseobtained by the
discharge of frictional electricity. He found that water could be
decomposedby currentsof either type, and inferred that all differences
betweenthem could be explainedby supposingthat voltaic electricity
as commonly obtained is ' lessintense,but produced in much larger
quanti ty.' Later in the sameyear (1801), Martin van Marum (1750-
1837) and Christian Heinrich Pfaff (1773-1852) arrived at the same
conclusion by carrying out on a large scale5 Yolta's plan of using
the pile to charge batteries of Leyden jars; while Johann Wilhelm
Ritter (1776-1810) showed6 that the electricities furnished at the
poles of the pile attract each other, and the electricities at the
correspondingpolesof two similar piles repel.
The discoveryof Nicholson and Carlisle made a great impression
on the mind of Humphry Davy (1778-1829), a young Cornishman
who about this time was appointed Professorof Chemistry at the
Royal Institution in London. Davy at once began to experiment
with voltaic piles, and in November 1800,7 showed that it gives no
current when the water betweenthe pairs of platesis pure, and that
its power of action is 'in great measureproportional to the power
of the conducting fluid substancebetweenthe double platesto oxy-
date the zinc.' This result, as he immediately perceived, did not
1 It was obtained independently four months later by J. W. Ritter.
• Beccaria (uttere dell' eiettricismo [Bologna, 1758], p. 282) had reduced mercury and
other metals from their oxides by discharges of frictional electricity; and Priestley had
obtained an inflammable gas from certain organic liquids in the same way. Cavendish in
1781had established the constitution of water by electrically exploding hydrogen and oxygen.
• Nicholson's Journal (4to), iv (1800), pp. 187,245; Phil. Mag. vii (1800), p. 337
• Phil. Trans. xci (1801), p. 427 • Phil. Mag. xii (1802), p. 161
• Gilbert's Annalen, viii (1801), p. 385
7 Nicholson's Journal (4to), iv (1800), pp. 275, 326, 337, 380, 394, 527; Davy's
Works, ii, p. 155
74
FROM GALVANI TO OHM
harmonisewell with Volta's viewson the sourceof electricity in the

pile,but was,on the other hand, in agreementwith Fabroni'sidea that
galvanic effectsare always accompaniedby chemical action. After
a seriesof experimentshe definitely concluded that 'the galvanic
pile of Volta acts only when the conducting substancebetweenthe
platesis capableof oxydating the zinc; and that, in proportion as
a greater quantity of oxygen entersinto combination with the zinc
in a given time, soin proportion is the power of the pile to decompose
water and to give the shockgreater. It seemstherefore reasonable
to conclude,though with our presentquantity of factswe are unable
to explain the exact mode of operation, that the oxydation of the
zinc in the pile, and the chemical changesconnectedwith it, are
somehow the causeof the electrical effectsit produces.' This principle
of oxidation guided Davy in designingmany new typesof pile, with
elementschosenfrom the whole range of the known metals.
Davy's chemicaltheory of the pile wassupportedby Wollaston 1
and by Nicholson,2the latter of whom urged that the existenceof
piles in which only one metal is used (with more than one kind of
fluid) is fatal to any theory which placesthe seatof the activity in the
contact of dissimilar metals.
Davy afterwards proposed3 a theory of the voltaic pile which
combinesideas drawn from both the 'contact' and 'chemical'
explanations. He supposedthat before the circuit is closed, the
copper and zinc disks in each contiguous pair assumeopposite
electrostaticstates,in consequenceof inherent' electrical energies'
possessed by the metals; and when a communication is made
between the extreme disks by a wire, the opposite electricities
annihilate each other, as in the dischargeof a Leyden jar. If the
liquid (which Davy compared to the glassof a Leyden jar) were
incapable of decomposition, the current would cease after this
discharge; but the liquid in the pile is composedof two elements
which have inherent attractions for electrified metallic surfaces,
henceariseschemicalaction, which removesfrom the disksthe outer-
most layers of molecules,whose energy is exhausted,and exposes
new metallic surfaces. The electrical energiesof the copperand zinc
are consequentlyagain exerted, and the processof electromotion
continues. Thus the contact of metalsis the causewhich dislurbs the
equilibrium, while the chemical changes continually restore the
conditionsunder which the contact energycan be exerted.
In this and other memoirs Davy assertedthat chemical affinity
1 Phil. Trans. xci (1801), p. 427 • Nicholson's Journal, i (1802), p. 142
I Phil. Trans. xcvii (1807), p. 1
75
GALVANISM
is essentially of an electrical nature. 'Chemical and electrical

attractions,' he declared,l ' are produced by the samecause,acting
in one caseon particles, in the other on masses,of matter; and the
sameproperty, under different modifications,is the causeof all the
phenomenaexhibited by different voltaic combinations.'
The further elucidation of this matter came chiefly from
researcheson electro-chemicaldecomposition,which we must now
consider.
Ritter in 1803 2 constructed a column formed of disks of silver
alternating with disksof damp cloth, and sent through this column
the current from a voltaic pile. When the connectionwith the pile
was broken, the silver-and-cloth column acted as a voltaic pile on
its own account. Ritter himself misunderstoodthe phenomenon,
comparing the action of the original current to that of a current
which is used to charge a condenser. The correct explanation was
given in 1805 by Volta,3 who showedthat so long as the original
current is passingthrough Ritter's column, the water in the cloth
disksis being decomposed,oxygen and hydrogen being collectedat
the oppositefacesof any disk. When the charging current is cut off,
theseproducts of decompositionsetup a reverseelectromotiveforce,
the Ritter column now acting as a pile formed with two different
fluids and one metal. This was the discoveryof voltaic polarisation;
the Ritter column may be regarded as a precursor of the modern
, accumulator' cell.
A phenomenon which had greatly surprised Nicholson and
Carlislein their early experimentswasthe appearanceof the products
of galvanic decomposition at placesremote from each other. The
first attempt to account for this was made in 1806 by Theodor von
Grothuss' (1785-1822) and by Davy,6 who advanceda theory that
the terminals at which water is decomposedhave attractive and
repellent powers; that the pole whenceresinouselectricity issueshas
the property of attracting hydrogen and the metals,and of repelling
oxygen and acid substances,while the positive terminal has the
power of attracting oxygen and repelling hydrogen; and that these
forces are sufficiently energetic to destroy or suspend the usual
operation of chemical affinity in the water moleculesnearest the
terminals. The force due to eachterminal wassupposedto diminish
with the distance from the terminal. When the molecule nearest
1 Phil. Trans. cxvi, pt. 3 (1826),p. 383 • Voigt's Magadnf. Nalurk. vi (1803),p. 181
• Gilbert's Annalen, xix (1805),p. 490 • Annates de ehimie, lviii (1806),p. 54
• Bakerianlecturefor 1806,Phil. Trans. xcvii (1807),p. I. A theorysimilar to that
of Grothussand Davy wascommunicatedby Peter Mark Roget (1779-1869)in 1807to
the PhilosophicalSocietyof Manchester; cf. Roget'sGalvanism, § lOG.
76
FROM GALVANI TO OHM
one of the terminals has been decomposedby the attractive and

repellent forcesof the terminal, one of its constituentsis liberated
there, while the other constituent, by virtue of electrical forces (the
oxygenand hydrogenbeing in oppositeelectrical states),attacksthe
next molecule,which is then decomposed. The surplus constituent
from this attacks the next molecule, and so on. Thus a chain of
decompositionsand recompositionswassupposedto be setup among
the moleculesintervening betweenthe terminals.
The hypothesisof Grothussand Davy was attacked in 1825 by
Auguste De La Rive 1 (1801-73) of Geneva,on the ground of its
failure to explain what happenswhen different liquids are placed in
seriesin the circuit. If, for example, a solution of zinc sulphate is
placedin onecompartment, and water in another, and if the positive
pole is placed in the solution of zinc sulphate,and the negativepole
in the water, De La Rive found that oxide of zinc is developedround
the latter; although decomposition and recomposition of zinc
sulphate could not take place in the water, which contained none
of it. Accordingly, he supposedthe constituentsof the decomposed
liquid to be bodily transportedacrossthe liquids, in closeunion with
the moving electricity. In the electrolysisof water, one current of
electrified hydrogen was supposedto leave the positive pole, and
'becomedecomposedinto hydrogen and electricity at the negative
pole, the hydrogen being there liberated as a gas. Another current
in the sameway carried electrified oxygenfrom the negative to the
positivepole. In this schemethe chain of successive decompositions
imagined by Grothuss does not take place, the only molecules
decomposedbeing thoseadjacent to the poles.
The appearanceof the products of decompositionat the separate
poles could be explained either in Grothuss' fashion by assuming
dissociationsthroughout the massof liquid, or in De La ttive's by
supposingparticular dissociatedatoms to travel considerable dis-
tances. Perhapsa preconceivedidea of economyin Nature deterred
the workersof that time from acceptingthe two assumptionstogether,
when either of them separatelywould meet the case. Yet it is to this
apparent redundancy that later researcheshave pointed as the
truth. Nature is what sheis, and not what we would make her.
De La Rive was one of the most thoroughgoing opponents of
Volta's contact theory of the pile; evenin the casewhen two metals
are in contact in air only, without the intervention of any liquid,
he attributed the electric effectwholly te.the chemical affinity of the
air for the metals.
1 Anna/esde Chimie, xxviii (1825), p. 190
77
GALVANISM
During the long interval between the publication of the rival

hypothesesof Grothuss and De La Rive, little real progresswas
made with the specialproblems of the cell; but meanwhileelectric
theory was developing in other directions. One of these,to which
our attention will first be turned, was the electro-chemicaltheory
of the celebratedSwedishchemist,JonsJacob Berzelius(1779-1848).
Berzelius founded his theory,! which had been in one or two
of its features anticipated by Davy,2 on inferences drawn from
Volta's contact effects. 'Two bodies,' he remarked, 'which have
affinity for each other, and which have been brought into mutual
contact, are found upon separation to be in opposite electrical
states. That which has the greatest affinity for oxygen usually
becomespositively electrified, and the other negatively.'
This seemedto him to indicate that chemical affinity arisesfrom
the play of electric forces,which in turn spring from electric charges
within the atoms of matter. To be precise,he supposedeachatom
to possesstwo poles, which are the seat of oppositeelectrifications,
and whoseelectrostaticfield is the causeof chemical affinity.
By aid of this conception Berzelius drew a simple and vivid
picture of chemical combination. Two atoms, which are about
to unite, disposethemselvesso that the positive pole of one touches
the negative pole of the other; the electricities of thesetwo poles
then dischargeeach other, giving rise to the heat and light which
are observed to accompany the act of combination.3 The dis-
appearanceof tht:se leaves the compound molecule with the two
remaining poles; and it cannot be dissociatedinto its constituent
atoms again until somemeansis found of restoring to the vanished
poles their charges. Such a meansis afforded by the action of the
galvanic pile in electrolysis: the oppositeelectricitiesof the current
invade the moleculesof the electrolyte, and restore the atoms to
their original stateof polarisation.
If, as Berzelius taught, all chemical compoundsare formed by
the mutual neutralisation of pairs of atoms, it is evident that they
must have a binary character. Thus he conceived a salt to be
compoundedof an acid and an oxide, and each of theseto be com-
pounded of two other constituents. Moreover, in any compound
the electropositivemember would be replaceableonly by another
electropositivemember, and the electro-negativemember only by
another member also electronegative; so that the substitution of,
1 Memoirs q[ the Acad. of Stockholm, 1812; Nicholson's Journal, xxxiv (1813), pp. 142,
153,240,319; xxxv, pp. 38, 118, 159

I Phil. Trans. xcvii (1807), p. 1 I This idea wasDavy's.
78
FROM GAL V ANI TO OHM
for example, chlorine for hydrogen in a compound would be

impossible-an inference which was overthrown by subsequent
discoveriesin chemistry.
Berzeliussucceededin bringing the most curiously diversefacts
within the scopeof his theory. Thus' the combination of polarised
atomsrequiresa motion to turn the opposite poles to each other;
and to this circumstance is owing the facility with which com-
bination takes place when one of the two bodies is in the liquid
.state, or when both are in that state; and the extreme difficulty,
or nearly i1l1possibility,of effecting an union between bodies, both
of which are solid. And again, since each polarised-particle must
have an electric atmosphere,and as this atmosphere is the pre-
disposingcauseof combination, as we have seen,it follows, that the
particles cannot act but at certain distances,proportioned to the
intensity of their polarity; and henceit is that bodies,which have
affinity for each other, alwayscombine nearly on the instant when
mixed in the liquid state,but lesseasilyin the gaseousstate,and the
union ceasesto be possibleunder a certain degreeof dilatation of
the gases; as we know by the experimentsof Grothuss, that a
mixture of oxygenand hydrogen in due proportions, when rarefied
to a certain degree, cannot be set on fire at any temperature
whatever.' And again: 'Many bodies require an elevation of
temperature to enable them to act upon each other. It appears,
therefore, that heat possessesthe property of augmenting the
polarity of thesebodies.'
Berzeliusaccountedfor Volta's electromotiveseriesby assuming
the electrification at one pole of an atom to be somewhatmore or
somewhatlessthan what would be required to neutralisethe charge
at the other pole. Thus each atom would possessa certain net or
residualcharge,which might be of either sign; and the order of the
elementsin Volta's seriescould be interpreted simply as the order
in which they would standwhen ranged accordingto the magnitude
of this residual charge. As we shall see,this conception was after-
wardsoverthrown by Faraday.
Berzeliuspermitted himself to publish somespeculationson the
nature of heat and electricity, which bring vividly before us the
outlook of an able thinker in the first quarter of the nineteenth
century. The great question, he says,is whether the electricities
and caloric are matter or merely phenomena. If the title of matter
is to be granted to such things as are ponderable, then theseprob-
lematic entities are certainly not matter; but thus to narrow the
application of the term is, he believes,a mistake; and he inclines
79
GALVANISM
to the opmlOn that caloric is truly matter, possessingchemical

affinities without obeying the law of gravitation, and that light and
all radiations consist in modesof propagating such matter. This
conclusion makes it easier to decide regarding electricity. 'From
the relation which exists between caloric and the electricities,' he
remarks, 'it is clear that what may be true with regard to the
materiality of one of them must also be true with regard to that
of the other. There are,however,a quantity of phenomenaproduced
by electricity which do not admit of explanation without admitting
at the sametime that electricity is matter. Electricity, for instance,
very often detacheseverything which covers the surface of those
bodies which conduct it. It, indeed, passesthrough conductors
without leaving any trace of its passage; but it penetratesnon-
conductors which oppose its course, and makes a perforation
preciselyof the samedescriptionaswould have beenmade by some-
thing which had need of place for its passage. We often observe
this when electric jars are broken by an overcharge,or when the
electric shock is passedthrough a number of cards, etc. We may
therefore, at least with some probability, imagine caloric and the
electricities to be matter, destitute of gravitation, but possessing
affinity to gravitating bodies. When they are not confined by these
affinities,they tend to placethemselvesin equilibrium in the universe.
The sunsdestroy at every moment this equilibrium, and they send
the re-united electricities in the form of luminous rays towards the
planetary bodies,upon the surfaceof which the rays, being arrested,
manifest themselvesas caloric; and this last in its turn, during the
time required to replace it in equilibrium in the universe,supports
the chemical activity of organic and inorganic nature.'
It was scarcely to be expected that anything so speculativeas
Berzelius' electric conception of chemical combination would be
confirmed in all particulars by subsequentdiscovery; and, as a
matter of fact, it did not as a coherent theory survive the lifetime
of its author. But someof its ideashave persisted,and amongthem
the conviction which lies at its foundation, that chemical affinities
are, in the last resort, of electrical origin.
While the attention of chemists was for long directed to the
theory of Berzelius,the interest of electricianswas diverted from it
by a discoveryof the first magnitude in a different region.
That a relation of some kind subsistsbetween electricity and
magnetismhad beensuspectedby the philosophersof the eighteenth
century. The suspicionwas basedin part on somecurious effects
produced by lightning, of a kind which may be illustrated by a
80
FROM GALVANI TO OHM
paper publishedin the Philosophical Transactions in 1735.1 A trades-

man of Wakefield, we are told, ' having put up a great number of
knives and forks in a large box, and having placed the box in the
cornerof a large room, there happen'd in July, 1731,a suddenstorm
of thunder, lightning, etc., by which the corner of the room was
damaged,the Box split, and a good many knives and forks melted,
the sheathsbeing untouched. The owner emptying the box upon a
Counter where someNails lay, the Personswho took up the knives,
that lay upon the Nails, observedthat the knives took up the Nails.'
Lightning thus cameto be credited with the power of magnetis-
ing steel; and it was doubtlessthis which led Franklin 2 in 1751
to attempt to magnetisea sewing-needleby meansof the discharge
of Leyden jars. The attempt was indeed successful; but, as Van
Marum afterwardsshowed,it was doubtful whether the magnetism
wasdue directly to the current.
More experiments followed.s In 1805 Jean-Nicolas-Pierre
Hachette (1769-1834) and Charles-BernardDesormes(1777-1862)
attempted to determine whether an insulated voltaic pile, freely
suspended,is oriented by terrestrial magnetism; but without
positive result. In 1807 Hans Christian Oersted (1777-1851),
Professorof Natural Philosophy in Copenhagen, announced his
intention of examining the action of electricity on the magnetic
needle; but it was not for someyearsthat his hopeswere realised.
If one of his pupils is to be believed,4 he was 'a man of genius,
but a very unhappy experimenter; he could not manipulate instru-
ments. He must always have an assistant,or one of his auditors
who had easyhands,to arrange the experiment.'
During a courseof lectures which he delivered in the winter of
1819-20 on 'Electricity, Galvanism and Magnetism,' the idea
occurredto him that the changesobservedwith the compassneedle
during a thunderstorm might give the clue to the effect of which
he was in search; and this led him to think that the experiment
should be tried with the galvanic circuit closedinstead of open, and
to inquire whether any effect is produced on a magnetic needle
when an electric current is passedthrough a neighbouring wire.
At first he placed the wire at right anglesto the needle,but observed
no result. After the end of a lecture in which this negative experi-
ment had been shown, the idea occurred to him to place the wire
parallel to the needle; on trying it, a pronounced deflection was
1 Phil. Trans. xxxix (1735), p. 74 I Letter vi from Franklin to Collinson
I In 1774 the Electoral Academy of Bavaria proposed the question, , Is there a real
and physical analogy between electric and magnetic forces? ' as the subject of a prize.
• cf. a letter from Hanstcen inserted in BenceJones's Life of Faraday, ii, p. 395
81
GALVANISM
observed, and the relation between magnetism and the electric

current was discovered. After confirmatory experimentswith more
powerful apparatus, the public announcementwas made in July
1820.1
Oersted did not determine the quantitative laws of the action,
but contentedhimself with a statementof the qualitative effect and
someremarks on its cause,which recall the magnetic speculations
of Descartes; indeed, Oersted's conceptionsmay be regarded as
linking thoseof the Cartesianschoolto thosewhich wereintroduced
subsequentlyby Faraday. 'To the effect which takesplace in the
conductor and in the surrounding space,'he wrote, , we shall give
the name of the conflict of electriciry.' 'The electricconflict actsonly
on the magneticparticlesof matter. All non-magneticbodiesappear
penetrableby the electric conflict, while magneticbodies,or rather
their magnetic particles, resist the passageof this conflict. Hence
they can be moved by the impetus of the contendingpowers.
'It is sufficiently evident from the preceding facts that the
electric conflict is not confined to the conductor, but dispersed
pretty widely in the circumjacent space.
'From the preceding facts we may likewise collect, that this
conflict performs circles; for without this condition, it seems
impossiblethat the one part of the uniting wire, when placedbelow
the magneticpole, shoulddrive it toward the east,and when placed
above it toward the west; for it is the nature of a circle that the
motions in oppositeparts should have an oppositedirection.'
Oersted'sdiscoverywas describedat the meeting of the French
Academy on 11 September1820 by an academician(Arago) who
had just returned from abroad. Several investigatorsin France
repeatedand extendedhis experiments; and the first preciseanalysis
of the effectwaspublishedby two of these,Jean-BaptisteBiot (1774-
1862) and Felix Savart (1791-1841), who, at a meeting of the
Academyof Scienceson 30 October 1820 announced2 that the action
experiencedby a pole of austral or boreal magnetism,when placed
at any distancefrom a straight wire carrying a voltaic current, may
be thus expressed: 'Draw from the pole a perpendicular to the'
wire; the force on the pole is at right anglesto this line and to
the wire, and its intensity is proportional to the reciprocal of the
distance.' This result wassoonfurther analysed,the attractive force
being divided into constituents,each of which was supposedto be
1 Experimenta circa dJectum conjlictus electrici in acum magneticam (Copenhagen,1820);
German trans. in Schweigger'sJournal Jilr Chemie und Physik, xxix (1820), p. 275;
Englishtrans.in Thomson'sAnnals of Philosophy, xvi (1820),p. 273
• Annales de Chimie, xv (1820),p. 222; Journal de Phys. xci (1820),p. 151
82
FROM GAL V ANI TO OHM
due to someparticular elementof the current; in its new form the

law maybe stated thus: the magneticforce due to an elementdsof a circuit,
in which a current i is flowing, at a point whose vector distancefrom ds is r,
is (in suitable units)
?
[ds. r] 1 or cur .'\r lids I
-.
r
It was now recognisedthat a magnetic field may be produced

as readily by an electric current as by a magnet; and, as Arago
soonshowed,3this,like anyother magneticfield, iscapableof inducing
magnetisationin iron. The questionnaturally suggesteditself as to
whether the similarity of properties between currents and magnets
extended still further, e.g. whether conductors carrying currents
would, like magnets,experienceponderomotive forceswhen placed
in a magneticfield, and whether suchconductorswould consequently
like magnets,exert ponderomotiveforceson each other.
The first step towards answering these inquiries was taken by
Oersted'" himself. 'As,' he said, 'a body cannot put another in
motion without beingmovedin its turn, when it possesses the requisite
mobility, it is easyto foreseethat the galvanic arc must be moved
by the magnet'; and this he verified experimentally.s
The next step came from Andre-Marie Ampere (1775-1836),
who at the meetingof the Academy on 18 September,exactly a week
after the newsof Oersted'sfirst discoveryhad arrived, showedthat
two parallel wirescarrying currents attract eachother if the currents
are in the samedirection, and repel eachother if the currents are in
oppositedirections. During the next three yearsAmpere continued
to prosecutethe researchesthus inaugurated, and in 1825published
his collectedresultsin one of the most celebrated memoirs e in the
history of natural philosophy.
Ampere introduces his work by proclaiming himself a follower
1 If a and b denote two vectors, the vector whose components are (aybz - a,by,
a.b:z- axb" axhy - aybx) is called the vectorprodw:t of a and b, and is denoted by ra,
b].
Its direction is at right angles to those of a and b, and its magnitude is represented by
twice the area of the triangle formed by them .
• If a denotes any vector, the vector whose components are oa. oa, oax oa,
0)1 - 0,,' 0" - ox'
~I! _ oax is denoted by curl a.
ox By
• Annales de Chimie, xv (1820), p. 93
« Schweigger's }ournal filr Chern. u. Phys. xxix (1820), p. 364; Thomson's Annals oj
Philosophy, xvi (1820), p. 375
• Davy found in 1821 (Phil. Trans. cxi [1821], p.425) that the electric arc is del!ected
by a magnet.
o MIm. de [,Acad. vi (1825), p. 175
(823) 83
GALVANISM
of that schoolwhich explained all physical phenomenain terms of

equal and oppositelydirectedforcesbetweenpairs of particles; and
he renouncesthe attempt to seekmore speculative,though possibly
more fundamental, explanationsin terms of the motionsof ultimate
fluids and aethers. Nevertheless,he indicatestwo conceptionsof this
latter character,on which suchexplanationsmight be founded.
In the first 1 he suggeststhat the ponderomotiveforcesbetween
circuits carrying electric currentsmay be due to ' the reactionof the
elastic fluid which extendsthroughout all space,whosevibrations
produce the phenomenaof light,' and which is ' put in motion by
electric currents.~ This fluid or aether can, he says,'be no other
than that which resultsfrom the combinationof the two electricities.'
In the secondconception,2Ampere suggeststhat the interspaces
between the metallic moleculesof a wire which carries a current
may be occupiedby a fluid composedof the two electricities,not in
the proportions which form the neutral fluid, but with an excessof
that one of them which is oppositeto the electricity peculiar to the
moleculesof the metal, and which consequentlymasksthis latter
electricity. In this inter-molecularfluid the oppositeelectricitiesare
continually being dissociatedand recombined; a dissociationof the
fluid within one inter-molecular interval having taken place, the
positive electricity thus produceduniteswith the negativeelectricity
of the interval next to it in the direction of the current, while the
negative electricity of the first interval unites with the positive
electricity of the next interval in the other direction. Such inter-
changes,accordingto this hypothesis,constitutethe electric current.
Ampere's memoir is, however,but little occupiedwith the more
speculativeside of the subject. His first aim was to investigate
thoroughly by experiment the ponderomotive forces on electric
currents.
, When,' he remarks,, M. Oersteddiscoveredthe action which a
current exerciseson a magnet, one might certainly have suspected
the existenceof a mutual action between two circuits carrying
currents; but this was not a necessaryconsequence; for a bar of
soft iron also acts on a magnetisedneedle, although there is no
mutual action betweentwo bars of soft iron.'
Ampere, therefore, submitted the matter to the test of the
laboratory, and discoveredthat circuits carrying elecfric currents
exert ponderomotiveforceson each other, and that ponderomotive
forces are exerted on such currents by magnets. To the science
1 Recueil d'obseTvatwnsliectro-dY'Ulmiques,p. 215; and the memoir just cited, pp. 285, 370
• Recueil d'obseTvatwnsllectro-dyrwmiques, pr. 297, 300, 371
84
FROM GALVANI TO OHM
which deals with the mutual action of currents he gave the name
electrodynamics 1; and he showedthat the action obeysthe following
laws:
(I) The effect of a current is reversedwhen the direction of the
current is reversed.
(2) The effect of a current flowing in a circuit twisted into small
sinuositiesis the sameas if the circuit were smoothedout.
(3) The forceexertedby a closedcircuit on an elementof another
circuit is at right anglesto the latter.
(4) The force between two elements of circuits is unaffected
when all linear dimensions are increased proportionately, the
current-strengthsremaining unaltered.
From these data, together with his assumption that the force
betweentwo elementsof circuits acts along the line joining them,
Ampere obtained an expressionof this force: the deduction may
be made in the following way:
Let ds, ds' be the elements,r the line joining them and i, i' the
current strengths. From (2) we seethat the effect of ds onds' is the
vectorsum of the effectsof dx, dy, dz on ds', where theseare the three
componentsof ds; sothe required force must be of the form:
r X a scalar quantity which is linear and homogeneousin ds ;
and it must similarly be linear and homogeneousin ds'; so using
(I), we seethat the force must be of the form
F = ii'r {(ds . ds') 4> (r) + (ds . r) (ds' . r) ifJ (r)},
where 4> and ifJ denoteundetermined functions of r.
From (4) it follows that when ds, ds', r are all multiplied by the
samenumber, F is unaffected: this showsthat
4>(r) =~ and ifJ(r) = ~'
whereA and B denoteconstants. Thus we have
F = ..,
nr
{A(ds. ds')
r + B(ds. r)y6.(ds' • r)}
Now, by (3), the resolved part of F along ds' must vanish when
integrated round the circuit s, i.e. it must be a completedifferential
whendr is taken to be equal to - ds. That is to say,
A(ds • ds') (r. ds') + B(ds • r) (ds' • r)'
r r6
1 Loc. cit. p. 298
85
GALVANISM
must be a complete differential; or
- i;.ad(r. ds')t + ~(ds • r) (r • dS')1
must be a complete differential; and therefore

A B
d 2,,=-?(ds.r),
3A B
or ---dr=-dr
2r' r"
or B = - iA.
Thus finally we have
F = Constant X ii'r{~(ds • ds') - ~(ds • r) (ds' • r)}.
This is Ampere's formula: the multiplicative constant dependsof

courseon the units chosenand may be taken to be - 1.
The weaknessof Ampere's work evidently lies in the assumption
that the force is directed along the line joining the two elements;
for in the analogouscaseof the action betweentwo magnetic mole-
cules,we know that the force is not directed along the line joining the
molecules. It is thereforeof interest to find the form ofF when this
restriction is removed.
For this purpose we observethat we can add to the expression
already found for F any term of the form
4>(r) • (ds • r) . ds',
where 4>(r) denotesany arbitrary function of r; for since
(ds• r) = - r . ds . ~,
this term vanisheswhen integrated round the circuit s; and it

containsds and ds' linearly and homogeneously,as it should. We
can alsoadd any terms of the form
d{r . (ds' • r) . X(r)},
where x(r) denotes any arbitrary function of r, and d denotes
differentiation along the arc s, keeping ds' fixed (so that
dr = - ds) ; this differential may be written
- ds . (ds' • r) . X(r) - rX(r) (ds' • ds) - !x'(r) ,. r (ds. r) (ds' • r).

86
FROM GALVAN! TO OHM
In order that the law of Action and Reaction may not be violated,
we must combine this with the former additional term so asto obtain
an expressionsymmetrical in ds and ds'; and hence we seefinally
that the general value of F is given by the equation
F = - ii'r{ ~(ds • ds') -~ (ds • r) (ds'. r)}

+ X(r) (ds' • r) ds + X(r) (ds • r) ds' + X(r) (ds • ds') r
+ ~'(r) (ds • r) (ds' • r) r.
The simplestform of this expressionis obtained by taking

ii'
X(r) = ,s'
when we obtain
F = ~'{(ds. r) ds' + (ds' • r)ds - (ds. ds')r}.
The comparatively simple expressionin bracketsis the vector part

of the quaternion product of the three vectorsds, r, dS'.l
From any of these values of F we can find the ponderomotive
force exerted by the whole circuit s on the element ds'; it is, in
fact, from the last expression,
ii'f$ ~ {Cds'• r) ds - (ds • ds') r},
..,( [d , [ds.r]]
or ll)s s. ,s ,
or i' Cds' • B],
where B = i.~~[ds.r].
Now this value of B is precisely the value found by Biot and
Savart 2 for the magnetic intensity at ds' due to the current t m
the circuit s. Thus we seethat the ponderomotive force on a current
elementds' in a magnetic field B is i' [ds'. B].
1 The simpler form of F given in the text is, if the term in ds' be omitted, the form
given by Grassmann, Ann. d. Phys. !xiv (1845), p. I. F.>r further work on his subject cr.
Tait, Proc. R. S. Edin. viii (1873), p. 220; Helmholtz, Berl. Akad. Monalsber. (1873), p. 91;
and Korteweg, Journal/lir Malh. xc (1881), p. 49. Helmholtz assumesthat the interaction
between two current elements is derivable from a potential, and this entails the existence
of a couple in addition to a force along the line joining the elements. cr. the discussion
in chap. vii of the present work.
• Seeante, p. 82
GALVANISM
Ampere developed to a considerable extent the theory of the

equivalenceof magnetswith circuits carrying currents; and showed
that an electric current is equivalent, in its magnetic effects,to a
distribution of magnetismon any surfaceterminated by the circuit,
the axesof the magnetic moleculesbeing everywherenormal to this
surface: 1 such a magnetisedsurfaceis called a magnetic shell. He
preferred, however, to regard the current rather than the magnetic
fluid as the fundamental entity, and consideredmagnetism to be
really an electrical phenomenon; each magnetic moleculeowesits
properties, according to this view, to the presencewithin it of a
small closed circuit in which an electric current is perpetually
flowing.
The impression produced by Ampere's memoir was great and
lasting. Writing half a century afterwards, Maxwell speaksof it
as ' one of the most brilliant achievementsin science.' 'The whole,'
he says, 'theory and experiment seemsas if it had leaped, full-
grown and full-armed, from the brain of the "Newton of electricity."
It is perfect in form and unassailablein accuracy; and it is summed
up in a formula from which all the phenomenamay be deduced,
and which must always remain the cardinal formula of electro-
dynamics.'
Heaviside, however, in 1888 expresseda different opinion 2 :
'It has been stated, on no less authority than that of the great
Maxwell, that Ampere's law of force between a pair of current
elementsis the cardinal formula of electrodynamics. If so, should
we not be always using it? Do we ever use it? Did Maxwell in
his Treatise? Surely there is somemistake. I do not in the least
mean to rob Ampere of the credit of being the father of electro-
dynamics; I would only transfer the name of cardinal formula to
another due to him, expressingthe mechanicalforce on an element
of a conductor supporting current in any magnetic field-the
vector product of current and induction. There is something real
about it; it is not like his force betweena pair of unclosedelements;
it is fundamental; and, as everybody knows, it is in continual use,
either actually or virtually (through electromotive force), both by
theoristsand practicians.'
Not long after the discovery by Oersted of the connection
between galvanism and magnetism, a connection was discovered
between galvanism and heat. In 1822 Thomas Johann Seebeck
1 Loc. cit. p. 367

• Electrician (28 Dec. 1888), p. 229; Heavisidc's Electrical Papers, ii, p. 500
88
FROM GALVANI TO OHM
(1770-1831), of Berlin discovered1 that an electric current can be

set up in a circuit of metals, without the interposition of any liquid,
merely by disturbing the equilibrium of temperature. Let a ring
be formed of copper and bismuth soldered together at the two
extremities; to establish a current it is only necessaryto heat the
ring at one of thesejunctions. To this new class of circuits the
name thermo-electric was given.
It was found that the metals can be arranged as a thermo-electric
series,in the order of their power of generating currents when thus
paired, and that this order is quite different from Yolta's order of
electromotive potency. Indeed antimony and bismuth, which are
near each other in the latter series,are at opposite extremities of
the former.
The currents generated by thermo-electric means are generally
feeble; and the mention of this fact brings us to the question,
which was about this time engaging attention, of the efficacy of
different voltaic arrangements.
Comparisonsof a rough kind had been instituted soon after the
discovery of the pile. The French chemists Antoine-Franc;ois de
Fourcroy (1755-1809), Louis-Nicolas Yauquelin (1763-1829) and
Louis-JacquesThenard (1777-1857) found 2 in 1801, on varying
the sizeof the metallic disksconstituting the pile, that the sensations
producedon the human frame wereunaffectedsolong asthe number
of disks remained the same; but that the power of burning finely
drawn wire wasaltered; and that the latter power wasproportional
to the total surface of the disks employed, whether this were dis-
tributed among a small number of large disks, or a large number
of small ones. This was explained by supposing that small plates
give a small quantity of the electric fluid with a high velocity, while
large platesgive a larger quantity with no greater velocity. Shocks,
which were supposedto depend on the velocity of the fluid alone,
would therefore not be intensified by increasing the size of the
plates.
The effect of varying the conductorswhich connectthe terminals
of the pile was also studied. Nicolas Gautherot (1753-1803)
observed3 that water containedin tubeswhich have a narrow opening
1 Abhandt. d. Berlin Akad. x (11322-3),p. 265. Ann. d. Ph),s. \xxiii (1823), pp. 115,430;
vi (1826), pp. I, 133, 253
Volta had previously noticed that a silver plate whose ends were at different tempera-
tures appeared to act like a voltaic cell.
Further experiments were performed by James Cumming (1777-1861), Professor of
Chemistry at Cambridge, Trans. C~mb. Phil. Soc. ii (1823), p. 47 ; and by Antoine Cesar
Becquere1(1788-1878), Annates de Chimie, xxxi (1826), p. 371.
• Ann. de CMmie, xxxix (1801), p. 103 • Annates de Chim. xxxix (1801), p. 203
89
GALVANISM
doesnot conductvoltaic currentssowell aswhen the openingis more

considerable. This experiment is evidently very similar to that
which Beccaria had performed half a century previously1 with
electrostaticdischarges.
As we have already seen, Cavendish investigated very com-
pletely the power of metals to conduct electrostatic discharges;
their power of conducting voltaic currents was now examinedby
Davy.2 His method was to connect the terminals of a voltaic
battery by a path containing water (which it decomposed),and also
by an alternativepath consistingof the metallic wire under examina-
tion. When the length of the wire waslessthan a certain quantity,
the water ceasedto be decomposed; Davy measuredthe lengths
and weights of wires of different materials and cross-sectionsunder
these limiting circumstances; and, by comparing them, showed
that the conducting power of a wire formed of anyone metal is
inversely proportional to its length and directly proportional to its
sectional area, but independent of the shapeof the cross-section. 3
The latter fact, as he remarked, showedthat voltaic currents pass

through the substanceof the conductor and not along its surface.
Davy, in the same memoir, compared the conductivities of
various metals, and studied the effect of temperature; he found
that the conductivity varied with the temperature, being' lower
in someinverseratio asthe temperaturewashigher.'
He also observedthat the samemagneticpower is exhibited by
every part of the samecircuit, even though it be formed of wires
of different conducting powers pieced into a chain, so that' the
magnetismseemsdirectly as the quantity of electricity which they
transmit.'
The current which flows in a given voltaic circuit evidently
dependsnot only on the conductors which form the circuit, but
also on the driving-power of the battery. In order to form a com-
plete theory of voltaic circuits, it was thereforenecessaryto extend
Davy's lawsby taking the driving-powerinto account. This advance
waseffectedin 1826 by GeorgSimon Ohm" (1787-1854).
Ohm had already carried out a considerableamount of experi-
mental work on the subject, and had, e.g., discoveredthat if a
1 Seep. 55
2 Phil. Trans. cxi (1821),p. 425. His resultswereconfirmedafterwardsby Becquerel,
Annales de ehimie, xxxii (1825),p. 423.
• Theseresultshad beenknown to Cavendish.
• Ann. d. Phys. vi (1826),p. 459; vii, pp. 45, 117; Die GalvanischeKettt mathematisch
bearbeitet, Berlin, 1827; translatedin Taylor's Scientific Memoirs, ii (1841),p. 401. cr. also
subsequentpapersby Ohm in Kastner's Archiv flir d. ges. Naturlehre, and Schweigger's
Jahrbuch.
go
FROM GALVANI TO OHM
number of voltaic cells are placed in seriesin a circuit, the current

is proportional to their number if the external resistanceis very
large, but is independentof their number if the external resistance
is small. He now essayedthe task of combining all the known
resultsinto a consistenttheory.
For this purposehe adopted the idea of comparing the flow of
electricity in a current to the flow of heat along a wire, the theory
of which had been familiar to all physicistssince the publication
of Fourier's Thiorie analytique de la chaleur in 1822. 'I have pro-
ceeded,'he says,'from the suppositionthat the communication of
the electricity from one particle takes place directly only to the
one next to it, so that no immediate transition from that particle
to any other situate at a greater distanceoccurs. The magnitude
of the flow betweentwo adjacent particles, under otherwiseexactly
similar circumstances,I have assumedto be proportional to the
differenceof the electric forces existing in the two particles; just
as, in the theory of heat, the flow of caloric betweentwo particles
is regardedasproportional to the differenceof their temperatures.'
The comparisonbetweenthe flow of electricity and the flow of
heat suggestedthe propriety of introducing a quantity whose
behaviourin electrical problemsshouldresemblethat of temperature
in the theory of heat. The differencein the valuesof sucha quantity
at two points of a circuit would provide what was so much needed,
namely, a measureof the' driving-power' acting on the electricity
between these points. To carry out this idea, Ohm recurred to
Volta's theory of the electrostatic condition of the open pile. It
was customary to measurethe 'tension' of a pile by connecting
one terminal to earth and testing the other terminal by an electro-
scope. Accordingly Ohm says: 'In order to investigatethe changes
which occur in the electric condition of a body A in a perfectly
definite manner, the body is each time brought, under similar
circumstances,into relation with a secondmoveablebody of invari-
able electrical condition, called the electroscope; and the force with
which the electroscopeis repelled or attracted by the body is deter-
mined. This force is termed the electroscopicforce of the body A.
'The same body A may also serve to determine the electro-
scopicforce in various parts of the same body. For this purpose
take the body A of very small dimensions,so that when we bring
it into contact with the part to be testedof any third body, it may
from its smallnessbe regarded as a substitute for this part; then
its electroscopicforce, measuredin the way described, will, when
it happens to be different at the various places, make known
91
GALVANISM
the relative differenceswith regard to electricity between these

places.'
Ohm assumed,as was customaryat that period, that when two
metalsare placedin contact, , they constantlymaintain at the point
of contact the same difference between their electroscopicforces.'
He accordingly supposedthat each voltaic cell possesses a definite
tension,or discontinuity of electroscopicforce,which is to beregarded
as its contribution to the driving-force of any circuit in which it
may be placed. This assumptionconfersa definite meaningon his
useof the term' electroscopicforce' ; the forcein questionisidentical
with the electrostaticpotential. But Ohm and his contemporaries
did not correctly understand the relation of galvanic conceptions
to the electrostatic functions of Poisson. The electroscopicforce
in the open pile was generally identified with the thicknessof the
electrical stratum at the place tested; while Ohm, recognisingthat
electric currents are not confined to the surfaceof the conductors,
but penetrate their substance,seemsto have thought of the electro-
scopic force at a place in a circuit as being proportional to the
volume-densityof electricity there-an idea in which he was con-
firmed by the relation which, in an analogouscase,existsbetween
the temperatureof a body and the volume-densityof heat supposed
to be containedin it.
Denoting, then, by S the current which flows in a wire of con-
ductivity y, when the difference of the electroscopicforces at the
terminals is E, Ohm writes
S =yE.
From this formula it is easy to deduce the laws already given by

Davy. Thus, if the area of the cross-sectionof a wire is A, we can
by placing n such wires side by side construct a wire of cross-section
nA. If the quantity E is the samefor each,equal currentswill flow
in the wires; and thereforethe current in the compoundwire will
be n times that in the single wire; so when the quantity E is un-
changed, the current is proportional to the cross-section; that is,
the conductivity of a wire is directly proportional to its cross-section,
which is one of Davy's laws.
In spite of the confusion which was attached to the idea of
electroscopicforce, and which was not dispelled for some years,
the publication of Ohm's memoir marked a great advance in
electrical philosophy. It was now clearly understood that the
current flowing in any conductor dependsonly on the conductivity
inherent in the conductor and on another variable which bearsto
92
FROM GALVANI TO OHM
electricity the samerelation that temperature bears to heat; and,

moreover, it was realised that this latter variable is the link con-
necting the theory of currents with the older theory of electrostatics.
These principles were a sufficient foundation for future progress;
and much of the work which was published in the secondquarter
of the century was no more than the natural development of the
principles laid down by Ohm.l
It is painful to relate that the discovererhad long to wait before
the merits of his great achievementwere officially recognised.2 He
was, however, awarded the Copley Medal of the Royal Society in
1841: and twenty-two years after the publication of the memoir
on the galvanic circuit, he was promoted to a university professor-
ship; this he held for the five yearswhich remained until his death
in 1854.
1 Ohm's theory was confirmed experimentally by several investigators, among whom
may be mentioned Gustav Theodor Fechner (1801-87), Maassbestimmungen iiber die

GalvanischeKelte (Leipzig, 1831), and Charles Wheatstone (1802-75), Phil. Trans. clxxxiv
(1843), p. 303.
3 An interesting account of the reception of Ohm's electrical researchesby his con-
temporaries is given by H.J.J. Winter in Phil. Mag. xxxv (1944), p. 371.
93
Chapter IV
THE LUMINIFEROUS MEDIUM, FROM BRADLEY
TO FRESNEL
ALTHOUGH Newton, as we have seen, refrained from committmg

himself to any doctrine regarding the ultimate nature of light, the
writers of the next generationinterpreted his criticism of the wave-
theory asequivalent to an acceptanceof the corpuscularhypothesis.
As it happened,the chief optical discoveryof this period tended to
support the latter theory, by which it was first and most readily
explained. In 1728JamesBradley (1692-1762), at that time Savilian
Professorof Astronomy at Oxford, sent to the Astronomer Royal
(Halley) an ' Account of a newdiscoveredmotion of the Fix'd Stars.' 1
In observingthe star')' in the head of the Dragon, with the object
of discovering its parallax, he had found that during the winter
of 1725-6 the transit across the meridian was continually more
southerly, while during the following summer its original position
was restoredby a motion northwards. Such an effect could not be
explained as a result of parallax; and eventually Bradley guessed
it to be due to the gradual propagation of light.2
c Thus, let CA denote a ray of light, falling on the
line BA; and supposethat the eye of the observeris
travelling along BA, with a velocity which is to the
velocity of light as BA is to CA. Then the corpuscleof
light, by which the object is discernible to the eye at A,
would have been at C when the eye was at B. The
tube of a telescopemust therefore be pointed in the
direction BC, in order to receive the rays from an
object whoselight is really propagated in the direction
CA. The angle BCA measuresthe difference between
B the real and apparent positions of the object; and
it is evident from the figure that the sine of this angle is to the
sine of the visible inclination of the object to the line in which the
eye is moving, as the velocity of the eye is to the velocity of light.
Observationssuch as Bradley's will therefore enable us to deduce
1 Phil. Trans. xxxv (1728), p. 637
• Roemer, in a letter to Huygens of date 30 Dec. 1677, mentions a suspected displace-
ment of the apparent position of a star, due to the motion of the earth at right angles to
the line of sight. cr. Correspondanced6 Huygens, viii, p. 53.
94
FROM BRADLEY TO FRESNEL
the ratio of the mean orbital velocity of the earth to the velocity
of light, or, as it is called, the constant of aberration; from its value
Bradley calculated that light is propagated from the sun to the
earth in 8 minutes 12seconds,which, ashe remarked, , is asit were
a Mean betwixt what had at different times been determined
from the eclipsesof Jupiter's satellites.'1
With the exceptionof Bradley's discovery,which was primarily
astronomical rather than optical, the eighteenth century was
decidedly barren, as regardsboth the experimental and the theo-
retical investigationof light; in curiouscontrast to the brilliance of
its recordin respectof electricalresearches.But someattention must
be given to a suggestivestudy 2 of the aether,for which the younger
John Bernoulli (1710-90) was in 1736 awarded the prize of the
FrenchAcademy. His ideasseemto have beenoriginally suggested
by an attempt 3 which his father, the elder John Bernoulli (1667-
1748), had made in 1701 to connectthe law of refraction with the
mechanicalprinciple of the compositionof forces. If two opposed
forceswhoseratio is p, maintain in equilibrium a particle which is
free to move only in a given plane, it follows from the triangle of
forcesthat the directionsof the forcesmust obey the relation
sin i = p.sin r,
where i and r denotethe anglesmade by thesedirections with the

normals to the plane. This is the same equation as that which
expresses the law of refraction, and the elder Bernoulli conjectured
that a theory of light might be basedon it; but he gave no satis-
factory physical reasonfor the existenceof forcesalong the incident
and refractedrays. This defecthis sonnow proceededto remove.
All space,according to the young Bernoulli, is permeatedby a
fluid aether, containing an immense number of excessivelysmall
whirlpools. The elasticity which the aether appearsto possess, and
in virtue of which it is able to transmit vibrations, is really due to
the presenceof thesewhirlpools; for, owing to centrifugal force,
1 Struve in 1845 found for the constant of aberration the value 20"445, which he
afterwards corrected to 20'·463. This was superseded in 1883 by the value 20'·492,
determined by M. Nyren. The observations of both Struve and Nyren were made with
the transit in the prime vertical. The method now generally used depends on the
measurement of differences of meridian zenith distances. In the present century values
have been found which range from 20"·523 down to 20'·445. The value. calculated from
the solar pflrallax and the velocity oflight is 20"511, if for the parallax we take the value
8'·790 found by H. Spencer Jones in 1941, and for the velocity of light take the value
2·99776 X 10" em/sec.
• Printed in 1752, in the Recue;l despiecesqui ont remportls les~rk l' Acad., tome iii
• Acta eruditorum (1701), p. 19
95
THE LUMINI.FEROUS MEDIUM
eachwhirlpool is continually striving to dilate, and sopressesagainst

the neighbouring whirlpools. It will be seenthat Bernoulli is a
thorough Cartesian in spirit; not only doeshe reject action at a
distance,but he insiststhat eventhe elasticity of his aethershall be
explicablein terms of matter and motion.
This aggregate of small vortices, or 'fine-grained turbulent
motion,' asit cameto be called a century and a half later,! is inter-
spersedwith solid corpuscles,whosedimensionsare small compared
with their distancesapart. These are pushedabout by the whirl-
pools whenever the aether is disturbed, but never travel far from
their original positions.
A sourceof light communicatesto its surroundingsa disturbance
which condenses the nearestwhirlpools; theseby their condensation
displacethe contiguouscorpusclesfrom their equilibrium position;
and these in turn produce condensationsin the whirlpools next
beyond them, so that vibrations are propagatedin every direction
from the luminous point. It is curiousthat Bernoulli speaksof these
vibrations as longitudinal, and actually contraststhem with thoseof
a stretched cord, which, 'when it is slightly displaced from its
rectilinear form, and then let go, performs transversevibrations in a
direction at right anglesto the direction of the cord.' When it is
rememberedthat the objection to longitudinal vibrations, on the
score of polarisation, had already been clearly stated by Newton,
and that Bernoulli's aether closely resemblesthat which Maxwell
invented in 1861-2 for the expresspurposeof securingtransversality
of vibration, one feelsthat perhapsno man ever so narrowly missed
a great discovery.
Bernoulli explained refraction by combining these ideas with
those of his father. Within the pores of ponderable bodies the
whirlpools are compressed,so the centrifugal force must vary in
intensity from one medium to another. Thus a corpusclesituated
in the interface betweentwo media is acted on by a greater elastic
force from one medium than from the other; and by applying the
triangle of forces to find the conditions of its equilibrium, the law
of Snell and Descartesmay be obtained.
Not long after this, the echoesof the old controversybetween
Descartesand Fermat about the law of refraction were awakened2
by Pierre-Louis-Moreaude Maupertuis (1698-1759).
It will be rememberedthat according to Descartes,the velocity
of light is greatestin densemedia, while according to Fermat the
1 cf. Lord Kelvin's vortex-sponge aether, described later in this work
• Mem. de l'Acad. (1744), p. 417

96
propagation is swiftest in free aether. The arguments of the cor-

puscular theory convinced Maupertuis that on this particular point
Descarteswas in the right; but neverthelesshe wished to retain
for sciencethe beautiful method by which Fermat had derived his
result. This he now proposedto do by modifying Fermat's principle
so as to make it agree with the corpuscular theory; instead of
assuming that light follows the quickest path, he supposed that
, the path describedis that by which the quantity of action is the
least'; and this action he defined to be proportional to the sum
of the spacesdescribed,each multiplied by the velocity with which
it is traversed. Thus instead of Fermat's expression
or
f~
(where t denotestime, v velocity and ds an element of the path)
Maupertuis introduced Jv ds
as the quantity which is to assumeits mInImUm value when the

path of integration is the actual path of the light. Since Maupertuis'
v, which denotesthe velocity according to the corpuscular theory, is
proportional to the reciprocal of Fermat's v, which denotes the
velocity according to the wave-theory, the two expressionsare
really equivalent,and lead to the samelaw of refraction. Maupertuis'
memoir is, however, of great interest from the point of view of
dynamics; for his suggestion was subsequently developed by
himself and by Euler and Lagrange into a general principle which
coversthe whole range of Nature, so far as Nature is a dynamical
system.
The natural philosophersof the eighteenth century for the most
part, like Maupertuis, accepted the corpuscular hypothesis; but
the wave-theory was not without defenders. Franklin declared 1
for it; and the celebratedmathematician Leonhard Euler (1707-83)
ranged himself on the same side, being impressed by the notion
that the emissionof particles would causea diminution in the mass
of the radiating body, which was not observed,while the emission
of wavesinvolved no such consequence. In a work entitled Nova
Theoria Lucis et Colorum, published 2 while he was living under the
patronage of Frederic the Great at Berlin, he insisted strongly on
the resemblancebetween light and sound; the whole of the space
through which the heavenly bodies move is filled with a subtle
1 Letter xxiii, written in 1752
• L. Euleri, Opuscula varii argumenti (Berlin.• 1746), p. 169

97
THE LUMINIFEROUS MEDIUM
matter, the aether, and light consistsin vibrations of this aether;

, light is in the aether the samething as sound in air.' Accepting
Newton's doctrine that colour dependson wave-length, he in this
memoir supposedthe frequency greatestfor red light, and least for
violet; but a few yearslater 1 he adopted the oppositeopinion.
The chief novelty of Euler's writings on light is his explanation
of the manner in which material bodies appear coloured when
viewed by white light; and, in particular, of the way in which
the coloursof thin platesare produced. He deniedthat suchcolours
are due to a more copious reflection of light of certain particular
periods, and supposedthat they represent vibrations generated
within the body itself under the stimulus of the incident light. A
colouredsurface,accordingto this hypothesis,containslargenumbers
of elastic molecules,which, when agitated, emit light of period
dependingonly on their own structure. The coloursof thin plates
Euler explained in the same way; the elastic responseand free
period of the plate at any place would, he conceived, dependon
its thicknessat that place; and in this way the dependenceof the
colour on the thicknesswas accounted for, the phenomenaas a
whole being analogousto well-known effectsobservedin experiments
on sound.
Euler, who remained at Berlin from 1741 to 1766, numbered
among his pupils a nieceof Frederic, the Princessof Anhalt-Dessau,
to whom in 1760-1 he wrote a seriesof letters setting forth his
views on natural philosophy. He anticipated Maxwell in asserting
that the sourceof all electrical phenomenais the sameaether that
propagateslight: electricity is nothing but a derangementof the
equilibrium of the aether. 'A body must becomeelectrical,' he
said, 'whenever the aether contained in its pores becomesmore
or lesselastic than that which is lodged in adjacent bodies. This
takesplace when a greater quantity of aether is introduced into the
pores of a body, or when part of the aether which it contained is
forced out. In the former case, the aether becomesmore com-
pressed,and consequentlymore elastic; in the other, it becomes
rarer, and losesits elasticity. In both cases,it is no longer in
equilibrium with that which is external; and the efforts which it
makes to recover its equilibrium, produce all the phenomenaof
electricity.2
Not only electrical, but also gravitational phenomena, were
explained in terrns of the aether.8 The explanation dependedon
1 Mem. tk l' Acad. tk Berlin (1752), p. 262 • Letter of 4 July 1761
• uonhardi Euleri Opera Omnia, seriestertia i, pp. 4, 149
98
supposingthat the pressureof the aetherincreaseswith the distance

from the centreof the earth, sayas Constant - l/r, sothat the force
pressinga body towards the earth is stronger than that directed
away from it, the balance of theseforces being the weight of the
body. On this hypothesis,the force on each atom would be pro-
portional to the volume of the atom, and therefore the weight of
an atom must be proportional to its volume, that is to say, the
densitiesof all atomsare equal; the fact that the densitiesof bodies
differ from eachother is accountedfor by assumingthat the atoms
are not in contact.
An attempt to improve the corpuscular theory in another
direction was made in 1752 by the Marquis de Courtivron,l and
independentlyin the following year by T. MelvilI.2 Thesewriters
suggested,as an explanation of the different refrangibility of dif-
ferent colours,that' the differently colour'd rays are projected with
different velocities from the luminous body: the red with the
greatest,violet with the least, and the intermediate colours with
intermediatedegreesof velocity.' On this supposition,asits authors
pointed out, the amount of aberration would be different for every
different colour; and the satellitesof Jupiter would changecolour,
from white through green to violet, through an interval of more
than half a minute beforetheir immersioninto the planet's shadow;
while at emersion the contrary successionof colours should be
observed,beginning with red and ending in white. The testimony
of practical astronomerswas soon given that such appearancesare
not observed; and the hypothesiswas accordingly abandoned.
In spite of Euler's assertionthat a single aether should serve
for all purposes,then~wasin the latter part of the eighteenthcentury
a tendencyin somequarters to add to the number of aethersby
postulating one whosefunction was to transmit radiant heat, e.g.
from a fire through the air to surroundingbodies. It was,however,
shownin 18003 by Sir William Herschel (1738-1822) that rays of
ordinary light transmit a certain amount of heat, that this effect
is more marked for light at the red end of the spectrum than for
light at the violet end, and that beyond the red end there are rays
which transmit heat but do not affect the human senseof vision:
theserays can be reflected and refracted like ordinary light. An
obviousinference(which Herschelat first accepted,but later rejected
for reasonswhich need not be discussedhere) is that radiant heat
is essentiallyof the same nature as light, and therefore does not
1 Courtivron's Trait! d'optique (1752) I Phil. Trans. xlviii (1753), p. 262
• Phil. Trans. xc (1800), pp. 255, 284, 293, 437
(biS) 99
require a separateaether. This identification wasfor long regarded

as doubtful, but in the courseof the next half-century it was shown
experimentally 1 that radiant heat hasall the characteristicproperties
of light (polarisation, double refraction, interference), and the view
originally suggestedby Herschel's work came to be accepteduni-
versally. It was shown in the following year by J. W. Ritter 2 that
invisible rays exist also beyond the violet end of the spectrum,
capable of chemical actions such as blackening chloride of silver:
and Young showed3 by projecting Newton's rings on paper covered
with silver chloride) that theseinvisible rays have the property of
interference (for which seebelow).
The fortunes of the wave-theory began to brighten at the end
of the century, when a new champion arose. Thomas Young, born
at Milverton in Somersetshirein 1773, and trained to the practice
of medicine, began to write on optical theory in 1799. In his first
paper" he remarked that, according to the corpuscular theory, the
velocity of emissionof a corpuscle must be the same in all cases,
whether the projecting force be that of the feeble spark produced
by the friction of two pebbles,or the intenseheat of the sun itself-
a thing almost incredible. This difficulty does not exist in the
undulatory theory, since all disturbances are known to be trans-
mitted through an elastic fluid with the samevelocity. The reluc-
tance which somephilosophersfelt to filling all spacewith an elastic
fluid he met with an argument which strangely foreshadowsthe
electric theory of light: 'That a medium resembling in many
properties that which has been denominated ether doesreally exist,
is undeniably proved by the phenomenaof electricity. The rapid
transmissionof the electrical shock showsthat the electric medium
is possessed of an elasticity as great as is necessaryto be supposed
for the propagation of light. Whether the electric ether is to be
consideredthe samewith the luminous ether, if such a fluid exists,
may perhaps at some future time be discovered by experiment:
hitherto I have not been able to observethat the refractive power
of a fluid undergoesany changeby electricity.'
Young then proceedsto show the superior power of the wave-
1 J. E. Berard,Ann. ,him. !xxxv (1813),p. 309
A. M. Ampere,Bibl. univ. xlviii (1832),p. 225; Ann. d. phys. xxvi (1832),p. 161;
Ann. ,him. Iviii (1835),p. 432 •
J. D. Forbes,Phil. Mag. v, p. 209; vii (1835),p. 349
M. Melloni, Ann. ,him. !xv (1837),p. 5, and many other papersin the samejournal
H. L. Fizeauand L. Foucault,ComplesRendusxxv (1847),p. 447
F. de la Provostayeand P. Desains,Ann. ,him. xxvii (1849),p. 109
• Ann. d. Phy.f. vii (1801),p. 527 • Phil. Trans.xciv (1804),p. 1
" Phil. Trans.xc (1800),p. 106
100
theory to explain reflection and refraction. In the corpuscular

theory it is difficult to seewhy part of the light should be reflected
and anotherpart of the samebeamrefracted; but in the undulatory
theory there is no trouble, as is shown by analogy with the partial
reflectionof soundfrom a cloud or denserstratum of air: 'Nothing
more is necessarythan to supposeall refracting media to retain,
by their attraction, a greater or lessquantity of the luminous ether,
so as to make its density greater than that which it possesses in a
vacuum, without increasing its elasticity.' This is precisely the
hypothesisadoptedlater by Fresneland Green.
In 1801Young made a discoveryof the first magnitude 1 when
attempting to explain Newton's rings on the principles of the wave-
theory. Rejecting Euler's hypothesis of induced vibrations, he
assumedthat the colours observedall exist in the incident light,
and showedthat they could be derived from it by a processwhich
wasnow for the first time recognisedin optical science.
The idea of this processwas not altogethernew, for it had been
used by Newton in his theory of the tides. 'It may happen,' he
wrote,2' that the tide may be propagated from the ocean through
different channelstowards the sameport, and may passin lesstime
through some channels than through others, in which case the
samegeneratingtide, being thus divided into two or moresucceeding
one another, may produce by composition new types of tide.'
Newton applied this principle to explain the anomalous tides
at Batsha in Tonkin, which had previously been described by
Halley.3
Young's own illustration of the principle is evidently suggested
by Newton's. 'Suppose,' he says,' 'a number of equal waves of
water to move upon the surfaceof a stagnant lake, with a certain
constantvelocity, and to enter a narrow channel leading out of the
lake; supposethen another similar causeto have excited another
equal seriesof waves,which arrive at the samechannel, with the
samevelocity, and at the sametime with the first. Neither series
of waveswill destroythe other, but their effectswill be combined;
if they enter the channelin sucha manner that the elevationsof one
seriescoincide with thoseof the other, they must together produce
a seriesof greater joint elevations; but if the elevations of one
seriesare so situated as to correspond to the depressionsof the
other, they must exactly fill up those depressions,and the surface
of the water must remain smooth. Now I maintain that similar
1 Phil. Trans. xcii (1802), pp. 12, 387 • Principia, bk. iii, prop. 24
• Phil. Trans. xiv (1684), p. 681 • Young's Works, i, p. 202
101
effectstake place whenever two portions of light are thus mixed;

and this I call the generallaw of the interferenceof light.'
Thus, 'whenever two portions of the samelight arrive to the
eye by different routes, either exactly or very nearly in the same
direction, the light becomesmost intense when the difference of
the routes is any multiple of a certain length, and least intensein
the intermediate state of the interfering portions; and this length
is different for light of different colours.'
Young's explanation of the colours of thin plates as seenby
reflection was, then, that the incident light givesrise to two beams
which reach the eye: one of thesebeamshas beenreflectedat the
first surfaceof the plate, and the other at the secondsurface; and
thesetwo beamsproduce the coloursby their interference.
One difficulty encountered in reconciling this theory with
observation arosefrom the fact that the central spot in Newton's
rings (where the thicknessof the thin film of air is zero) is black
and not white, as it would be if the interfering beamswere similar
to each other in all respects. To account for this Young showed,
by analogy with the impact of elastic bodies, that when light is
reflected at the surfaceof a densermedium, its phaseis retarded
by half an undulation: so that the interfering beamsat the centre
of Newton'sringsdestroyeachother. The correctness of this assump-
tion he verified by substituting essenceof sassafras (whoserefractive
index is intermediate betweenthoseof crown and flint glass)for air
in the spacebetweenthe lenses; ashe anticipated, the centreof the
ring systemwasnow white.
Newton had long before observedthat the rings are smaller
when the medium producing them is optically more dense. Inter-
preted by Young's theory, this definitely proved that the wave-
length of light is shorter in dense media, and therefore that its
velocity is less.
The publication of Young's papers occasioneda fierce attack
on him in the Edinburgh Review, from the pen of Henry Brougham,
afterwards Lord Chancellor of England. Young replied in a
pamphlet, of which it is said 1 that only a single copy was sold;
and therecanbe no doubt that Broughamfor the time beingachieved
his object of discrediting the wave-theory.2 Its failure to make
much impression in the years immediately following its appear-
ance was strikingly shown in 1807-10, when Sir William Herschel
1 Peacock'sLife of Young
• 'Strange fellow,' wrote Macaulay,when half a centuryafterwardshe foundhimself
sitting besideBroughamin the Houseof Lords, , his powen gone: his spiteimmortal.'
102
publishedthreepapersIon Newton'srings. HerschelignoredYoung's

theory altogether, and, in place of Newton's fits of easyreflection
and easy transmission, proposed an explanation based on 'the
principle of a critical separationof the colours which takesplace at
certain anglesof incidence, admitting the reflection of somerays at
the sameangleof incidenceat which othersare transmitted.'
Young now turned his attention to the fringes of shadows. In
the corpuscular explanation of these, it was supposed that the
attractive forceswhich operate in refraction extend their influence
to somedistancefrom the surfacesof bodies, and inflect such rays
as passclose by. If this were the case, the amount of inflection
should obviously depend on the strength of the attractive forces,
and consequentlyon the refractive indices of the bodies-a propo-
sition which had been refuted by the experimentsof'sGravesande.
The causeof diffraction effects was thus wholly unknow!1, until
Young, in the Bakerian lecture for 1803,2showedthat the principle
of interference is concernedin their formation; for when a hair
is placed in the cone of rays diverging from a luminous point, the
internal fringes (i.e. thosewithin the geometricalshadow)disappear
when the light passingon one side of the hair is intercepted. His
conjectureasto the origin of the interfering rayswasnot sofortunate;
for he attributed the fringes outside the geometrical shadow to
interferencebetweenthe direct rays and raysreflectedat the diffract-
ing edge; and supposedthe internal fringes of the shadow of a
narrow object to be due to the interferenceof rays inflected by the
two edgesof the object.
The successof so many developmentsof the wave-theory led
Young to inquire more closely into its capacity for solving the
chief outstanding problem of optics-that of the behaviour of light
in crystals. The beautiful construction for the extraordinary ray
given by Huygenshad lain neglectedfor a century; and the degree
of accuracywith which it representedthe observationswasunknown.
At Young's suggestionWollaston 3 investigated the matter experi-
mentally, and showedthat the agreemep..t betweenhis own measure-
ments and Huygens' rule was remarkably close. 'I think,' he
wrote, , the result must be admitted to be highly favourable to the
Huygenian theory; and, although the existenceof two refractions
at the sametime, in the samesubstance,be not well accountedfor,
and still lesstheir interchangewith each other, when a ray of light
1 Phil. Trans.xcvii (1807), p. 180; xcix (1809), p. 259; c (1810), p. 149
• Phil. Trans.xciv (1804), p. I ; Young's Works,i, p. 179
• Phil. Trans.xcii (1802), p. 381
103
is made to passthrough a secondpiece of spar situated transversely

to the first, yet the oblique refraction, when consideredalone,seems
nearly aswell explainedasany other optical phenomenon.'
Meanwhile the advocatesof the corpuscular theory were not
idle; and in the next few years a successionof discoverieson
their part, both theoretical and experimental, seemed likely to
imperil the good position to which Young had advancedthe rival
hypothesis.
The first of thesewas a dynamical explanation of the refraction
of the extraordinary ray in crystals, which was published in 1808
by Laplace.l His methodis an extensionof that by which Maupertuis
had accounted for the refraction of the ordinary ray, and which
sinceMaupertuis' day had beensodevelopedthat it wasnow possible
to apply it to problemsof all degreesof complexity. Laplaceassumes
that the crystalline medium acts on the light corpusclesof the
extraordinary ray so as to modify their velocity, in a ratio which
dependson the inclination of the extraordinary ray to the axis of
the crystal; so that, in fact, the difference of the squaresof the
velocitiesof the ordinary and extraordinary rays is proportional to
the squareof the sine of the angle which the latter ray makeswith
the axis. The principle of least action then leadsto a law of refrac-
tion identical with that found by Huygens' construction with the
spheroid; just asMaupertuis' investigationled to a law of refraction
for the ordinary ray identical with that found by Huygens' con-
struction with the sphere.
The law of refraction for the extraordinary ray may also be
deduced from Fermat's principle of least time, provided that the
velocity is taken inversely proportional to that assumedin the
principle of least action; and the velocity appropriate to Fermat's
principle agreeswith that found by Huygens, being, in fact, pro-
portional to the radius of the spheroid. Theseresultsare obvious
extensionsof thosealready obtainedfor ordinary refraction.
Laplace'stheOlYwaspromptly attackedby Young,2who pointed
out the improbability of sucha systemofforces aswould be required
to impressthe requisite changeof velocity on the light corpuscles.
If the aim of controversial matter is to convincethe contemporary
world, Young's paper must be counted unsuccessful; but it per-
manently enriched scienceby proposing a dynamical foundation
for double refraction on the principles of the wave-theory. 'A
1 Mem. de l'Inst. (1809), p. 300; Journal de PhysiqUi!, Ixviii (Jan. 1809),p. 107;
Mem. de la Soc. d'Arcueil, ii
I Quarterly Review, Nov. 1809; Young's Works, i, p. 220
10 4
solution,' he says,, might be deducedupon the Huygenian principles,

from the simplestpossiblesupposition,that of a medium more easily
compressiblein one direction than in any direction perpendicular
to it, as if it consistedof an infinite number of parallel plates con-
nected by a substancesomewhatlesselastic. Such a structure of
the elementaryatomsof the crystal may be understoodby comparing
them to a block of wood or of mica. Mr Chladni found that the
mereobliquity of the fibres of a rod of Scotchfir reducedthe velocity
with which it transmitted sound in the proportion of 4 to 5. It is
therefore obvious that a block of such wood must transmit every
impulse in spheroidal-that is, oval-undulations; and it may also
be demonstrated,as we shall show at the conclusionof this article,
that the spheroid will be truly elliptical when the body consists
either of plane and parallel strata, or of equidistant fibres, supposing
both to be extremely thin, and to be connected by a less highly
elasticsubstance; the spheroidbeing in the former caseoblate and
in the latter oblong.' Young then proceedsto a formal proof that
, an impulse is propagated through every perpendicular section of
a lamellar elastic substancein the form of an elliptic undulation.'
This must be regarded as the beginning of the dynamical theory
of light in crystals. It was confirmed in a striking way not long
afterwards by Brewster,l who found that compression in one
direction causesan isotropic transparent solid to become doubly
refracting.
Meanwhile, in January 1808, the French Academy had proposed
as the subject for the physical prize in 1810, ' To furnish a mathe-
matical theory of double refraction, and to confirm it by experiment.'
Among those who resolved to compete was Etienne-Louis Malus
(1775-1812), a colonel of engineerswho had seen service with
Napoleon's expedition to Egypt. While conducting experiments
towards the end of 1808 in a housein the Rue desEnfers in Paris,
Malus happenedto analysewith a rhomb of Iceland spar the light
of the setting sun reflected from the window of the Luxembourg,
and was surprised to notice that the two images were of very
different intensities. Following up this observation,he found that
light which had beenreflectedfrom glassacquirestherebya modifica-
tion similar to that which Huygens had noticed in rays which have
experienceddouble refraction, and which Newton had explained
by supposingrays of light to have' sides.' This discoveryappeared
so important that without waiting for the prize competition he
communicatedit to the Academy in December 1808, and published
1 Phil. Trans.cv (1815), p. 60
105
it in the following month. 'I have found,' he said, 'that this

l
singular disposition,which has hitherto beenregardedasone of the

peculiar effectsof double refraction, can be completely impressed
on the luminous moleculesby all transparent solids and liquids.
For example, light reflected by the surface of water at an angle
of 52 45' has all the characteristicsof one of the beamsproduced
0
by the double refraction of Iceland spar, whoseprincipal section

is parallel to the planewhich passesthrough the incident ray and the
reflectedray. If wereceivethis reflectedrayon any doubly refracting
crystal, whoseprincipal sectionis parallel to the plane of reflection,
it will not be divided into two beamsasa ray of ordinary light would
be, but will be refractedaccordingto the ordinary law.'
Mter this Malus found that light which has been refracted at
the surfaceof any transparent substancelikewise possesses in some
degreethis property, to which he gave the namepolarisation. The
memoir 2 which he finally submitted to the Academy, and which
contains a rich storeof experimentaland analytical work on double
refraction, obtainedthe prize in 1810; its immediateeffectasregards
the rival theoriesof the ultimate nature of light was to encourage
the adherents of the corpuscular doctrine; for it brought into
greater prominence the phenomenaof polarisation, of which the
wave-theorists,still misled by the analogy of light with sound,were
unable to give any account.
The successful discovererwaselectedto the Academyof Sciences,
and becamea member of the celebratedclub of Arcueil.3 But his
health, which had been undermined by the Egyptian campaign,
now broke down completely; and he died, at the ageof thirty-six,
in the following year.
The polarisation of a reflected ray is in general incomplete-
i.e. the ray displaysonly imperfectly the properties of light which
has been polarised by double refraction; but for one particular
angle of incidence, which depends on the reflecting body, the
polarisation of the reflectedray is complete. Malus measuredwith
considerable accuracy the polarising angles for glass and water,
and attempted to connect them with the other optical constantsof
thesesubstances,the refractive indices and dispersivepowers,but
without success. The matter was afterwards taken up by David
1 Nouveau Bulletin des Sciences, par La Soc. PhiLomatique, i (1809), p. 266; Memoires de la
Soc. d'Arcueil, ii, 1809
• Mem. prisentls Ii l'Inst. par divers Savans, ii (1811), p. 303
• So called from the village near Paris where Laplace and Berthollet had their country
houses, and where the meetings took place. The club consisted of a dozen of the most
celebrated scientific men in France.
106
Brewster(1781-1868), who in 1815 1 showedthat there is complete

polarisation by reflection when the reflected and refracted rays
satisfythe condition of being at right anglesto eachother.
Almost at the same time Brewster made another discovery
which profoundly affected the theory of double refraction. It had
till then been believed that double refraction is always of the type
occurring in Iceland spar,to which Huygens'construction is applic-
able. Brewsternow found this belief to be erroneous,and showed
that in a large classof crystals there are two axes,instead of one,
along which there is no double refraction. Such crystalsare called
biaxal, the simpler type to which Iceland spar belongsbeing called
uniaxal.
The wave-theory at this time was still encumbered with diffi-
culties. Diffraction was not satisfactorily explained; for polarisa-
tion no explanation of any kind was forthcoming; the Huygenian
construction appearedto require two different luminiferous media
within doubly refracting bodies; and the universality of that con-
struction had been impugned by Brewster's discovery of biaxal
crystals.
The upholdersof the emissiontheory, emboldenedby the success
of Laplace's theory of double refraction, thought the time ripe for
their final triumph; and as a step to this, in March 1817 they
proposedDiffraction asthe subjectof the Academy'sprize for 1818.
Their expectation was disappointed; and the successfulmemoir
afforded the first of a seriesof reversesby which, in the short space
of sevenyears,the corpusculartheory was completely overthrown.
The author was Augustin Fresnel (1788-1827), the son of an
architect, and himself a civil engineer in the Government service
in Normandy. During the brief dominance of Napoleon after his
escapefrom Elba in 1815, Fresnelfell into trouble for having enlisted
in the small army which attempted to bar the exile's return; and
it was during a period of enforced idlenessfollowing on his arrest
that he commencedto study diffraction. In his earliest memoir 2
he propoundeda theory similar to that of Young, which wasspoiled
like Young's theory by the assumptionthat the fringes depend on
light reflected by the diffracting edge. Observing, however, that
the blunt and sharp edgesof a knife produce exactly the same
fringes, he became dissatisfiedwith this attempt, and on 15 July
1816, presentedto the Academy a supplement to his paper,3 in
which, for the first time, diffraction-effects are referred to their
1 Phil. Trans. cv (1815), p. 125
I Annates de chimie (2), i (1816), p. 239; CEuvres, i, p. 89 I CEuvres, i, p. 129
107
true cause-namely, the mutual interferenceof the secondarywaves

emitted by those portions of the original wave-front which have
not been obstructed by the diffracting screen. Fresnel'smethod of
calculation utilised the principles of both Huygens and Young;
he summedthe effectsdue to different portions of the sameprimary
wave-front.
The sketch presentedto the Academy in 1816 was during the
next two years developed into an exhaustive memoir,l which was
submitted for the Academy'sprize.
It so happened that the earliest memoir, which had been pre-
sented to the Academy in the autumn of 1815, had been referred
to a Commissionof which the reporter was Franc;oisArago (1786-
1853) ; Arago was so much impressedthat he soughtthe friendship
of the author, of whom he waslater a strenuouschampion.
A champion was indeed needed when the larger memoir was
submitted; for Laplace, Poisson and Biot, who constituted a
majority of the Commission to which it was referred, were all
zealoussupportersof the corpuscular theory. During the examina-
tion, however, Fresnel was vindicated in a somewhatcurious way.
He had calculated in the memoir the diffraction-patterns of a
straight edge, of a narrow opaque body bounded by parallel sides,
and of a narrow opening bounded by parallel edges,and had shown
that the results agreed excellently with his experimental measures.
Poisson,when reading the manuscript, happenedto notice that the
analysiscould be extended to other cases,and in particular that it
would indicate the existenceof a bright spot at the centre of the
shadow of a circular screen. He suggestedto Fresnelthat this and
some further consequencesshould be tested experimentally; this
was done,2and the results were found to confirm the new theory.
The concordanceof observation and calculation was so admirable
in all caseswhere a comparison was possible that the prize was
awarded to Fresnelwithout further hesitation.
In the sameyear in which the memoir on diffraction was sub-
mitted, Fresnel published an investigation 3 of the influence of the
earth's motion on light. We have already seenthat aberration was
explained by its discoverer in terms of the corpuscular theory;
and it was Young who first showed' how it may be explained on
the wave-hypothesis. 'Upon considering the phenomena of the
1 Mim. de l'Acad. v (1826),p. 339; (Euvres, i, p. 247
• The bright spot in the centreof the shadowhad beennoticed in the early part of
the eighteenthcentury by J. N. Delisle.
• Annales de Chimie, ix (1818),p. 57; (Euvres, ii, p. 627
• Phil. Trans. xciv (1804),p. 1; Young's Works, i, p. 188
108
aberrationof the stars,'he wrote, , I am disposedto believethat the

luminiferousaetherpervadesthe substanceof all material bodieswith
little or no resistance,asfreely perhapsasthe wind passes through a
grove of trees.' In fact, if we supposethe aether surrounding the
earth to be at rest and unaffectedby the earth's motion, the light
waveswill not partake of the motion of the telescope,which we
may supposedirected to the true place of the star, and the image
of the star will therefore be displacedfrom the central spider-line
at the focusby a distanceequal to that which the earth describes
while the light is travelling through the telescope. This agreeswith
what is actually observed.
But a hostof further questionsnow suggestthemselves. Suppose,
for instance,that a slab of glasswith a plane face is carried along
by the motion of the earth, and it is desiredto adjust it so that a
ray of light coming from a certain star shall not be bent when it
entersthe glass; must the surfacebe placed at right anglesto the
true direction of the star asfreedfrom aberration, or to its apparent
direction as affected by aberration? The question whether rays
comingfrom the starsare refracteddifferently from rays originating
in terrestrial sourceshad been raisedoriginally by MichellI; and
Robison and Wilson 2 had assertedthat the focal length of an
achromatic telescopeshould be increa.sedwhen it is directed to a
star towardswhich the earth is moving, owing to the changein the
relative velocity of light. Arago 3 submitted the matter to the test
of experiment,and concludedthat the light coming from any star
behavesin all casesof reflection and refraction preciselyasit would
if the star were situatedin the place which it appearsto occupy in
consequence of aberration, and the earth were at rest; so that the
apparent refraction in a moving prism is equal to the absolute
refractionin a fixed prism.
Fresnelnow set out to provide a theory capable of explaining
Arago's result. To this end he adopted Young's suggestion,that
the refractive powers of transparent bodies depend on the con-
centration of aether within them; and made it more preciseby
assumingthat the aetherealdensity in any body is proportional to
the squareof the refractive index. Thus, if c denote the velocity
of light in vacuo, and if C1 denoteits velocity in a given material body
at rest, so that I-' = C/Cl is the refractive index, then the densities
1 Phil. Trans. Ixxiv (1784), p. 35 • Trans. R. S. (Edin.), i, Hist., p. 30
• Arago communicated his results to the Institute in 1810, but did not publish them
until much later, in ComptesRendusviii (1839), 326 and xxxvi (1853), 38. cf. Biot, Astron.
phys. (3rd edn.), v, p. 364. The accuracy of Arago's experiment can scarcely have been
such as to demonstrate absolutely his result.
109
P and PI of the aethcr in interplanetary spac~and In the body

respectivelywill be connectedby the relation
PI = I-'Ip•
Fresnel further assumedthat, when a body is in motion, part
of the aether within it is carried along-namely, that part which
constitutesthe excessof its density over the density of aether in
vacuo; while the rest of the aetherwithin the spaceoccupiedby the
body is stationary. Thus the density of aether carried along is
(PI - p) or (1-'2 - l)p, while a quantity of aether of density P
remains at rest. The velocity with which the centre of gravity of
the aetherwithin the body movesforward in the direction of propa-
gation is therefore 1-'1 _ 1
--- w,
1-'1
where w denotesthe componentof the velocity of the body in this
direction. This is to be added to the velocity of propagation of the
light waveswithin the body; sothat in the moving body the absolute
velocity of light is 1
+ I-'
1
Cl --1-
- w.
I-'
Many years afterwards Stokes1 put the samesuppositionin a
slightly different form. Supposethe whole of the aether within
the body to move together, the aether entering the body in front,
and being immediately condensed,and issuing from it behind,
where it is immediately rarefied. On this assumptiona masspW
of aethermust passin unit time acrossa plane of area unity, drawn
anywhere within the body in a direction at right angles to the
body's motion; and therefore the aether within the body has a
drift-velocity - Wp/Pl relative to the body; so the velocity of
light relative to the body will be C1 - wp/Pl> and the absolutevelocity
of light in the moving body will be
C1 +w _ wp,
PI
or C1 + 1-'1I-'-
--1-
1 w, asbefore.
This formula was experimentally confirmed in 1851 by H.

Fizeau,2 who measured the displacement of interference-fringes
1 Phil. Mag. xxviii (1846), p. 76
• AnnaZesde ehimie, lvii (1859), p. 385. Also by A. A. Michelson and E. W. Morley,
Am. Journ. Science,xxxi (1886), p. 377.
110
formed by light which had passedthrough a column of moving

water.
The same result may easily be deduced from an experiment
performedby Hoek.l In this a beam of light was divided into two
portions, one of which was made to passthrough a tube of water
AB and was then reflectedat a mirror C, the light
being afterwards allowed to return to A without A
passing through the water; while the other
portion of the bifurcated beam was made to de-
scribe the same path in the reverse order, i.e.
passing throug-h the water on its return journey B
from C instead of on the outward journey. On
causingthe two portions of the beam to interfere,
Hoekfound that no differenceof phasewasproduced
----c
between them when the apparatus was oriented in the direction of
the terrestrial motion.
Let w denote the velocity of the earth, supposedto be directed
from the tube towards the mirror. Let clf' denote the velocity
of light in the water at rest, and clf' + 4> the velocity of light
in the water when moving. Let l denote the length of the tube.
The magnitude of the distance BC doesnot affect the experiment,
so we may supposeit zero.
The time taken by the first portion of the beam to perform its
journey is evidently
l +_l_,
clf' + 4> - w c+ w
while the time for the secondportion of the beam is
l l
c- w + -c/-f'---4>-+-w·
The equality of theseexpressionsgives at once, when terms of
higher ordersthan the first in wlc are neglected.
4> = (f'1 - 1) wlf",
which is Fresnel'sexpression.2
On the basis of this formula, Fresnel proceeded to solve the
problem of refraction in moving bodies. Suppose that a prism
1 ArchivesNlerl. iii (1868),p. 180
• Fresnel'slaw may alsobe deducedfrom the principle that the amount of light trans-
mitted by a slabof transparentmatter mustbe the samewhetherthe slabis at restor in
motion; otherwise the equilibrium of exchangesof radiation would be vitiated.
ef. Larmor, Phil. Trans.clxxxv (1893),p. 775.
II I
AoCoBois carried along by the earth's motion in vacuo, its face AoCo
being at right angles to the direction of motion; and that light
from a star is incident normally on this face. The rays experience
no refraction at incidence; and we have only to consider the effect
produced by the secondsurfaceAoBo.Supposethat during an interval
T of time the prism travels from the position AoCoBoto the position
A1C1B1,while the luminous disturbance at Co travels to B1,and the
AI
luminous disturbance at Ao gives rise (according to Huygens'
Principle) to a sphere,so BID, the tangent from B1to this sphere,is
the emergent wave-front. 1
+
1 _
Then since the absolute velocity in the glass IS c1 1J-_1_ W,

IJ-
we have
A
If we write C1A1B1= i, and denote the total deviation of the
wave-front by 81,we have
AID = AoD - A}Ao cos 8} = TC - TW cos81>
C}B} = CoB}- CoC}= T(C} + IJ-IIJ-I1 W - W) = T(C} - W;}I) sincelJ-=~
112
and therefore (neglectingsecond-orderterms in wle)
+ S,)
sin (i
--.-.-
sm z
e
=-
e,
+ -we -w cos0,
- e, ~
Denoting by S the value of S, when w is zero, we have therefore

sin (i
--.--.-
+
S) _ c
--.
SInz e,
Subtracting this equation from the preceding, we have
(S - S,) cos(i + S) = _ ~ + ~ cosS = _ ~ + ~ sin (i ~ ?) cosS
SInz e c, c e SInz
whence
Now the telescopeby which the emergent wave-front BID is

receivedis itself being carried forward by the earth's motion; and
we must therefore apply the usual correction for aberration in order
to find the apparent direction of the emergentray. But this correc-
tion is w sin Sic, and precisely counteracts the effect which has
been calculated as due to the motion of the prism. So finally we
seethat the motion of the earth has no first-order influence on the
refraction of light from the stars.
Fresnelinferred from his formula that if observationswere made
with a telescopefilled with water, the (;lberrationwould be unaffected
by the presenceof the water-a result which was verified by Airy I
in 1871. He showed, moreover, that the apparent positions of
terrestrial objects,carried along with the observer,are not displaced
by the earth's motion; that experiments in refraction and inter-
ference are not influenced by any motion which is common to the
source, apparatus and observer; and that light travels between
given points of a moving material systemby the path of least time.
Thesepredictionshavealsobeenconfirmed by observation: Respighi 2
in 1861,and Hoek 3 in 1868, experimenting with a telescopefilled
with water and a terrestrial source of light, found that no effect
was produced on the phenomena of reflection and refraction by
altering the orientation of the apparatus relative to the direction
1 Proc. R. S. xx (1871), p. 35 I Mem. Accad. Sci. Bologna, ii (1861), p. 279
a Asl. Nach. Ixxiii (1868), p. 193
113
of the earth's motion. E. Mascart 1 in 1872studied experimentally

the question of the effect of motion of the source or recipient of
light in all its bearings, and showedthat the light of the sun and
that derived from artificial sourcesare alike incapable of revealing
by diffraction-phenomenathe translatory motion of the earth.
The greatestproblem now confronting the investigatorsof light
was to reconcile the facts of polarisation with the principles of the
wave-theory. Young had long been pondering over this, but had
hitherto beenbaffled by it. In 1816he receiveda visit from Arago,
who told him of a new experimental result which he and Fresnel
had lately obtained 2-namely, that two pencils of light, polarised
in planes at right angles, do not interfere with each other under
circumstancesin which ordinary light showsinterference-phenomena,
but alwaysgive by their reunion the sameintensity oflight, whatever
be their differenceof path.
Arago had not long left him when Young, reflecting on the
new experiment, discoveredthe long-sought key to the mystery:
it consistedin the very alternative which Bernoulli had rejected
eighty years before, of supposing that the vibrations of light are
executedat right anglesto the direction of propagation.
Young's ideas were first embodied in a letter to Arago,3 dated
12 January 1817. 'I have been reflecting,' he wrote, 'on the
possibility of giving an imperfect explanation of the affection of
light which constitutes polarisation, without departing from the
genuine doctrine of undulations. It is a principle in this theory,
that all undulations are simply propagated through homogeneous
mediums in concentric spherical surfaceslike the undulations of
sound, consistingsimply in the direct and retrograde motionsof the
particles in the direction of the radius, with their concomitant con-
densationand rarefactions. And yet it is possibleto explain in this
theory a transversevibration, propagatedalsoin the direction of the
radius, and with equal velocity, the motions of the particles being
in a certain constantdirection with respectto that radius; and this
is a polarisation.'
In an article on ' Chromatics,' which was written in September
1 Ann. de l'&ole !formale (2), i (1872), p. 157
• It was not published until 1819, in Annales de Chimie, x; Fresnel's (Euvres, i, p. 509.
By means of this result, Fresnel was able to give a complete explanation of a class of
phenomena which Arago had discovered in 1811, viz. that when polarised light is trans-
mitted through thin plates of sulphate of lime or mica, and afterwards analysed by a
prism oflceland spar, beautiful complementary colours are displayed. Young had shown
that these effects are due essentially to interference, but had not made clear the part
played by polarisation.
• Young's Works, i, p. 380
of the sameyear 1 for the supplementto the Encyclopaedia Britannica,

he saysZ: 'If we assumeas a mathematical postulate, on the
undulating theory, without attempting to demonstrateits physical
foundation,that a transversemotion may be propagatedin a direct
line, we may derive from this assumptiona tolerable illustration of
the subdivisionof polarisedlight by reflection in an oblique plane,'
by , supposingthe polar motion to be resolved' into two constituents,
which fare differently at reflection.
In a further letter to Arago, dated 29 April 1818Young recurred
to the subject of transversevibrations, comparing light to the
undulationsof a cord agitated by one of its extremities.3 This letter
wasshownby Arago to Fresnel,who at once saw that it presented
the true explanation of the non-interferenceof beamspolarised in
perpendicularplanes,and that the latter effect could evenbe made
the basisof a proof of the correctnessof Young's hypothesis; for if
the vibration of each beam be supposedresolvedinto three com-
ponents,one along the ray and the other two at right anglesto it,
it is obviousfrom the Arago-Fresnelexperimentthat the components
in the direction of the ray must vanish; in other words, that the
vibrationswhich constitutelight are executedin the wave-front.
It must be rememberedthat the theory of the propagation of
wavesin an elastic solid was as yet unknown, and light was still
alwaysinterpreted by the analogy with the vibrations of sound in
air, for which the direction of vibration is the sameasthat of propa-
gation. It was thereforenecessaryto give somejustification for the
new departure. With wonderful insight Fresnel indicated & the
precisedirection in which the theory of vibrations in ponderable
bodiesneededto be extendedin order to allow of wavessimilar to
those of light: 'the geometers,'he wrote, 'who have discussed
the vibrations of elastic fluids hitherto have taken account of no
accelerating forces except those arising from the difference of
condensationor dilatation betweenconsecutivelayers.' He pointed
out that if we also supposethe medium to possessa rigidity, or
power of resisting distortion, such as is manifestedby all actual
solid bodies,it will be capableof transversevibration. The absence
1 Peacock'sLife of roung, p. 391 I Young's Works, i, p. 279
• This analogy had been given by Hooke in a communication to the Royal Society
on 15 Feb. 1671-2. But there seemsno reason to suppose that Hooke appreciated the
point now advanced by Young.
« Annales de Chimie, xvii (1821), p. 180; CEuvres,i, p. 629. Young had already drawn
attention to this point. 'It is difficult,' he saysin his LecturesonNa/ural Philosophy (ed. 1807),
vol. i, p. 138, , to compare the lateral adhesion, or the force which resists the detrusion
of the parts of a solid, with any form of direct cohesion. This force constitutes the rigidity
or hardness of a solid body, and is wholly absent from liquids.'
(8~3) 115
of longitudinal wavesin the aether he accountedfor by supposing

that the forces which opposecondensationare far more powerful
than those which oppose distortion, and that the velocity with
which condensationsare propagatedis so great compared with the
speed of the oscillations of light, that a practical equilibrium of
pressureis maintained perpetually.
The nature of ordinary non-polarisedlight was next discussed.
, If then,' Fresnelwrote,! , the polarisation of a ray of light consists
in this, that all its vibrations are executedin the samedirection, it
resultsfrom any hypothesison the generation of light waves,that
a ray emanating from a single centre of disturbance will always
be polarised in a definite plane at any instant. But an instant
afterwards, the direction of the motion changes,and with it the
plane of polarisation; and these variations follow each other as
quickly as the perturbations of the vibrations of the luIninous
particle; so that evenif we could isolatethe light of this particular
particle from that of other luminous particles, we should doubtless
not recognisein it any appearanceof polarisation. If we consider
now the effectproducedby the union of all the waveswhich emanate
from the different points of a luminous body, we seethat at each
instant, at a definite point of the aether, the generalresultant of all
the motionswhich commingletherewill havea determinatedirection,
but this direction will vary from one instant to the next. So direct
light can be consideredas the union, or more exactly as the rapid
succession, of systemsof wavespolarisedin all directions. According
to this way of looking at the matter, the act of polarisation consists
not in creating thesetransversemotions, but in decomposingthem
in two invariable directions, and separatingthe componentsfrom
each other; for then, in each of them, the oscillatory motions take
place alwaysin the sameplane.'
He then proceededto consider the relation of the direction of
vibration to the plane of polarisation. 'Apply theseideasto double
refraction, and regard a uniaxal crystal as an elastic medium in
which the acceleratingforce which results from the displacement
of a row of moleculesperpendicularto the axis,relative to contiguous
rows,is the sameall round the axis; while the displacementsparallel
to the axis produce accelerating forces of a different intensity,
strongerif the crystal is ' repulsive,' and weakerir-it is ' attractive.'
The distinctive character of the rays which are ordinarily refracted
being that of propagating themselveswith the samevelocity in all
directions,we must admit that their oscillatory motionsare executed
1 lococit. p. 185
116
at right anglesto the plane drawn through theserays and the axis
of the crystal; for then the displacementswhich they occasion,
alwaystaking place along directionsperpendicular to the axis, will,
by hypothesis, always give rise to the same accelerating forces.
But, with the conventionalmeaningwhich is attached to the expres-
sion plane of polarisation, the plane of polarisation of the ordinary
rays is the plane through the axis; thus, in a pencil of polarised light,
the oscillatory motion is executedat right angles to theplane of polarisation.'
This result afforded Fresnel a foothold in dealing with the
problem which occupied the rest of his life; henceforth his aim
was to basethe theory of light on the dynamical properties of the
luminiferousmedium.
The first topic which he attacked from this point of view was
the propagation of light in crystalline bodies. Since Brewster's
discoverythat many crystalsdo not conform to the type to which
Huygens' construction is applicable, the wave-theory had to some
extent lost credit in this region. Fresnel,now, by what wasperhaps
the most brilliant of all his efforts,1 not only reconqueredthe lost
territory, but added a new domain to science.
He had, as he tells us himself, never believed the doctrine that
in crystalsthere are two different luminiferousmedia,one to transmit
the ordinary and the other the extraordinary waves. The alter-
native to which he inclined wasthat the two velocitiesof propagation
werereally the two roots of a quadratic equation, derivable in some
way from the theory of a single aether. Could this equation be
obtained, he was confident of finding the explanation, not only of
double refraction, but also of the polarisation by which it is always
accompanied.
The first step was to take the caseof uniaxal crystals, which
had been discussedby Huygens, and to see whether Huygens'
sphereand spheroid could be replaced by, or made to depend on,
a single surface.2
Now a wave propagated in any direction through a uniaxal
crystal can be resolved into two plane-polarised components;
oneof these,the' ordinary ray,' is polarisedin the principal section,
1 His first memoir on double refraction was presented to the Academy on 19 Nov.
1821, but has not been published except in his collected works, (Euvres, ii, p. 261. It was
followed by other papers in 1822; and the results were finally collected in a memoir
which was printed in 1827, Mem. de I'Acad. vii, p. 45; (Euvres, ii, p. 479.
• In attempting to reconstruct Fresnel's course of thought at this period, the present
writer has derived much help from the Life prefixed to the (EuvresdeFresnel. Both Fresnel
and Young were singularly fortunate in their biographers: Peacock's Life of Young,and
this notice of Fresnel, which was the last work of Verdet, are excellent reading.
117
and has a velocity VI, which may be representedby the radius of

Huygens' sphere-say, VI = b ;
while the other, the < extraordinary ray,' is polarised in a plane

at right anglesto the principal section,and has a wave-velocity V2,
which may be representedby the perpendicular drawn from the
centre of Huygens' spheroid on the tangent-planeparallel to the
plane of the wave. If the spheroidbe representedby the equation
y2 + Z2 +-x2
--- = I,
a 2 b i
and if (l, m, n) denote the direction cosinesof the normal to the

plane of the wave, we have therefore
Vi2 = a2(m2 + n2) + b2l2•
But the quantities Ilvl and Ilv2' as given by theseequations,are
easilyseento be the lengths of the semi-axesof the ellipsein which
the spheroid b2(y2 + Z2) + a2x2 = I
is intersectedby the plane
lx + my + nz = 0 ;
and thus the constructionin terms of Huygens' sphereand spheroid
can be replaced by one which dependsonly on a single surface,
namely the spheroid b2(y2 Z2) +
a2x2 = 1. +
Having achieved this reduction, Fresnel guessedthat the case
of biaxal crystals could be covered by substituting for the latter
spheroidan ellipsoid with three unequal axes-say,
x2 y2 Z2
--+- +-
£1
= 1.
£2 £3
If livi and IIv2 denote the lengths of the semi-axesof the ellipse
in which this ellipsoid is intersectedby the plane
lx + my + nz = 0,
it is well known that VI and V2 are the roots of the equation in v
r m n 2
--+ --+ --
2
I I I 2 2 2
=0;
--V --V --V
£1 £2 £8
and accordingly Fresnelconjecturedthat the roots of this equation

representthe velocities,in a biaxal crystal, of the two plane-polarised
waveswhosenormals are in the direction (l, m, n).
118
Having thus arrived at his result by reasoningof a purely geo-

metrical character,he now deviseda dynamical schemeto suit it.
The vibrating medium within a crystal he supposed to be
ultimately constituted of particles subjected to mutual forces;
and on this assumptionhe showedthat the elasticforce of restitution
when the systemis disturbed must dependlinearly on the displace-
ment. In this first proposition a difference is apparent between
Fresnel'sand a true elastic-solidtheory; for in actual elasticsolids
the forcesof restitution depend not on the absolute displacement,
but on the strains,i.e. the relative displacements.
In any crystal there will exist three directions at right angles
to each other, for which the force of restitution acts in the same
line asthe displacement: the directionswhich possess this property
are namedaxes of elasticiry. Let thesebe taken as axes,and suppose
that the elasticforcesof restitution for unit displacementsin these
three directions are I jE1 I JEI 1JEI respectively. That the elasticity
shouldvary with the direction of the moleculardisplacementseemed
to Fresnelto suggestthat the moleculesof the material body either
take part in the luminousvibration, or at any rate influencein some
way the elasticity of the aether.
A unit displacementin any arbitrary direction (a., fJ, y) can be
resolvedinto componentdisplacements(cosa., cosfJ, cosy) parallel
to the axes,and eachof theseproducesits own effectindependently;
sothe componentsof the force of restitution are
cos a cos f3 ~o~'(.
El
, -~-,
~I Ea
This resultant force has not in generalthe samedirection as the

displacementwhich producedit; but it may alwaysbe decomposed
into two other forces,one parallel and the other perpendicular to
the directionof the displacement; and the former of theseis evidently
cos· a
-- + --
cost f3 + cos· y
--.
El EI EI
The surface
will thereforehave the property that the squareof its radius vector
in any direction is proportional to the componentin that direction
of the elasticforce due to a unit displacementin that direction: it
is called the surface of elasticiry.
Considernow a displacementalong one of the axesof the section
on which the surfaceof elasticity is intersectedby the plane of the
119
wave. It is easilyseenthat in this casethe componentof the elastic

force at right angles to the displacement acts along the normal
to the wave-front; and Fresnel assumesthat it will be without
influence on the propagation of the vibrations, on the ground of his
fundamental hypothesisthat the vibrations of light are performed
solely in the wave-front. This step is evidently open to criticism;
for in a dynamical theory everything should be deducedfrom the
laws of motion without special assumptions. But granting his
contention,it followsthat sucha displacementwill retain its direction,
and will be propagated as a plane-polarisedwave with a definite
velocity.
Now, in order that a stretchedcord may vibrate with unchanged
period, when its tension is varied, its length must be increased
proportionally to the square root of its tension; and similarly the
wave-length of a luminous vibration of given period is proportional
to the squareroot of the elasticforce (per unit displacement),which
urges the molecules of the medium parallel to the wave-front.
Hence the velocity of propagation of a wave, measuredat right
angles to its front, is proportional to the square root of the com-
ponent, along the direction of displacement,of the elasticforce per
unit displacement; and the velocity of propagation of sucha plane-
polarised wave as we have consideredis proportional to the radius
vector of the surfaceof elasticity in the direction of displacement.
Moreover, any displacement in the given wave-front can be
resolved into two, which are respectivelyparallel to the two axes
of the diametral sectionof the surfaceof elasticityby a plane parallel
to this wave-front; and it follows from what has been said that
each of these component displacementswill be propagated as an
independent plane-polarised wave, the velocities of propagation
being proportional to the axesof the section,1 and thereforeinversely
proportional to the axesof the sectionof the inversesurfaceof this
with respectto the origin, which is the ellipsoid
But this is preciselythe result to which, aswe have seen,Fresnel

had been led by purely geometrical considerations; and thus his
geometrical conjecture could now be regarded as substantiatedby
a study of the dynamicsof the medium.
1 It is evidentfrom this that the optic axes,or linesof singlewave-velocity,alongwhich
there is no double refraction, will be perpendicularto the two circular sectionsof the
surfaceof elasticity.
120
It is easyto determine the wave-surfaceor locus at any instant

-say, t = l-of a disturbance originated at someprevious instant
-say, t = O-at someparticular point-say, the origin. For this
wave-surfacewill evidently be the envelopeof plane wavesemitted
from the origin at the instant t = G-that is, it will be the envelope
of planes Ix + my + nz - v = 0,
wherethe constantsI, m, n, v are connectedby the identical equation
II +m +n l l
= I,
and by the relation previouslyfound-namely,
II m n
--+ --+ --
l l
=0.
1 1
--v I
--v III
--v
El EI E.
By the usual procedure for determining envelopes,it may be

shownthat the locusin questionis the surfaceof the fourth degree
y. z.
--+
Xl
Elr· 1 -
----+
1
E.r· -
--- 1
=0
' Ear· -
which is called Fresnel's wave-surface.1 It is a two-sheetedsurface,

as must evidently be the case from physical considerations. In
uniaxal crystals,for which EI and Ea are equal, it degeneratesinto
the sphere I -
r -
1/E.,
and the spheroid

E.XI + El (yl + Zl) = 1.
It is to thesetwo surfacesthat tangent planes are drawn in the
construction given by Huygens for the ordinary and extraordinary
refracted rays in Iceland spar. As Fresnel observed, exactly the
sameconstruction applies to biaxal crystals, when the two sheets
of the wave-surfaceare substitutedfor Huygens'sphereand spheroid.
, The theory which I have adopted,' saysFresnel at the end of
this memorablepaper, ' and the simple constructionswhich I have
deducedfrom it, havethis remarkablecharacter,that all the unknown
quantities are determined together by the solution of the problem.
We find at the sametime the velocities of the ordinary ray and of
the extraordinary ray, and their planes of polarisation. Physicists
1 Another construction for the wave-surfaceis the following, which is due to
MacCullagh,Coll. Works, p. 1. Let the ellipsoid

t'lX' + t'Jy'+ £a(;' =1
be intersectedby a planethrough its centre,and on the perpendicularto that planetake
lengthsequalto the semi-axesof the section. The locusof theseextremitiesis the wave
surface.
121
who have studied attentively the laws of Nature will feel that such
simplicity and such close relations between the different elements
of the phenomenon are conclusivein favour of the hypothesison
which they are based.'
The question as to the correctnessof Fresnel'sconstruction was
discussedfor many years afterwards. A striking consequenceof it
was pointed out in 1832 by William Rowan Hamilton (1805-65),
Royal Astronomer of Ireland, who remarked 1 that the surface
defined by Fresnel's equation has four conical points, at each of
which there is an infinite number of tangent planes; consequently,
a single ray, proceeding from a point within the crystal in the
direction of one of thesepoints, must be divided on emergenceinto
an infinite number of rays, constituting a conical surface. Hamilton
also showed that there are four planes, each of which touchesthe
wave-surfacein an infinite number of points, constituting a circle of
contact: so that a corresponding ray incident externally should
be divided within the crystal into an infinite number of refracted
rays, again constituting a conical surface.
These singular and unexpected consequenceswere shortly
afterwardsverified experimentally by Humphrey Lloyd,2 and helped
greatly to confirm belief in Fresnel'stheory. It should, however,
be observedthat conical refraction only showshis form of the wave-
surface to be correct in its general features, and is no test of its
accuracy in all details. But it was shown experimentally by Stokes
in 1872,3 Glazebrook in 18794 and Hastings in 1887,5 that the
construction of Huygens and Fresnel is certainly correct to a very
high degreeof approximation; and Fresnel'sfinal formulae have
since been regarded as unassailable. The dynamical substructure
on which he based them is, as we have seen,open to objection;
but, as Stokesobserved6: 'If we reflect on the state of the subject
as Fresnel found it, and as he left it, the wonder is, not that he
failed to give a rigorous dynamical theory, but that a single mind
was capableof effecting so much.'
In a secondsupplementto his first memoir on double refraction,
presentedto the Academy on 26 November 1821,7Fresnelindicated
the lines on which his theory might be extendedsoasto take account
1 Trans. Roy. Irish Acad. xvii (1833), p. I ; Hamilton's Math. Papers, i, p. 164
I Trans. Roy. Irish Acad. xvii (1833), p. 145. Strictly speaking, the bright cone which
is usually observed arises from rays adjacent to the singular ray; the latter can, however,
be observed, its enfeeblement by dispersion into the conical form causing it to appear dark.
I Proc. R. S. xx (1872), p. 443 • Phil. Trans. c1xxi (1879), p. 421
• Am. Jour. Sci. (3), xxxv (1887), p. 60 • Brit. Assoc. Rep. (1862), p. 254
, CEuvres,ii, p. 438
122
of dispersion. • The molecular groups, or the particles of bodies,'

he wrote, • may be separatedby intervals which, though small, are
certainly not altogetherinsensiblerelatively to the length of a wave.'
Such a coarse-grainedness of the medium would, as he foresaw,
introduce into the equations terms by which dispersion might be
explained; indeed, the theory of dispersionwhich was afterwards
given by Cauchy was actually based on this principle. It seems
likely that, towards the closeof his life, Fresnelwas contemplating a
great memoir on dispersion,1 which was never completed.
Fresnelhad reasonat first to be pleasedwith the reception of
his work on the optics of crystals, for in August 1822 Laplace
spokehighly of it in public; and when at the end of the year a
seat in the Academy becamevacant, he was encouragedto hope
that the choice would fall on him. In this he was disappointed.2
Meanwhile his researcheswere steadily continued; and in January
1823the very month of his rejection, he presentedto the Academy
a theory in which reflection and refraction 3 are referred to the
dynamical propertiesof the luminiferous media.
As in his previous investigations,he assumesthat the vibrations
which constitute light are executedat right anglesto the plane of
polarisation. He adopts Young's principle, that reflection and
refraction are due to differencesin the inertia of the aether in
different material bodies, and supposes(as in his memoir on
aberration) that the inertia is proportional to the inverse square
of the velocity of propagation of light in the medium. The con-
ditions which he proposesto satisfy at the interface between two
media are that the displacementsof the adjacent molecules,resolved
parallel to this interface, shall be equal in the two media; and that
the energy of the reflected and refracted waves together shall be
equal to that of the incident wave.
On theseassumptionsthe intensity of the reflected and refracted
light may be obtained in the following way:
Consider first the casein which the incident light is polarised
in the plane of incidence,so that the displacementis at right angles
to the plane of incidence; let the amplitude of the displacement
1 cr. the biography in CEuvresde Fresnel, i, p. xcvi
• Writing to Young in the spring of 1823, he says: • Tous cesmemoires, que derniere-
ment j'ai presentescoup sur coup a l'Academie des Sciences,ne m'en ont pas cependant
ouvert la porte. C'est M. Dulong qui a ete nomme pour remplir la place vacante dans la
section de physique. . . . Vous voyez, Monsieur, que la theorie des ondulations ne m'a
point porte bonheur: mais cela ne m'en degoiite pas: etje me console de ce malheur en
m'occupant d'optique avec une nouvelle ardeur.'
• The MSS was for some time believed to be lost, but was ultimately found among
the papers of Fourier, and printed in Mem. tU l'Acad. xi (1832), p. 393; CEuvres,i, p. 767.
123
at a given point of the interface be f for the incident ray, g for the
reflected ray and h for the refracted ray.
The quantities of energy propagated per second across unit
cross-section of the incident, reflected and refracted beams are
proportional respectively to
where 'I) C2) denote the velocities of light, and Ph P2 the densities
of aether, in the two media; and the cross-sectionsof the beams
which meet the interface in unit area are
cos t, cos i, cosr

respectively. The principle of conservation of energy therefore
gives
The equation of continuity of displacement at the interface is
f+g =h.
Eliminating h between thesetwo equations, and using the formulae
sinI r _ cl' PI
c.- -P ,
_
sin'i - 1 I
we obtain the equation
f __ sin (i - r)
-i - sin (i + r)"
Thus when the light is polarised in the plane of reflection the ampli-
tude of the reflected wave is
s~n((~- r)) X the amplitude of the incident vibration.

SIn t r+
Fresnel showsin a similar way that when the light is polarised
at right anglesto the plane of reflection, the ratio of the amplitudes
of the reflected and incident wavesis
tan (i - r)
tan (i +r)"
Theseformulae are generally known as Fresnel'ssine law and Fresnel's

tangent law respectively. They had, however, been discovered
124
experimentally by Brewster some years previously. When the

incidenceis perpendicular, so that i and r are very small, the ratio
of the amplitudesbecomes
· .
LlOUt Z - r
Z +
-. -,
r
or iLl - iLl
iLz + iL/
where iLl and iLl denote the refractive indices of the media. This
formula had been given previously by Young 1 and Poisson,2on
the suppositionthat the elasticity of the aether is of the samekind
asthat of air in sound.
When i + r = 90°, tan (i +
r) becomesinfinite; and thus a
theoretical explanation is obtained for Brewster's law, that if the
incidence is such as to make the reflected and refracted rays per-
pendicular to eachother, the reflectedlight will be wholly polarised
in the plane of reflection.
Fresnel'sinvestigationcan scarcelybe called a dynamical theory
in the strict sense,as the qualities of the medium are not defined.
His method was to work backwardsfrom the known properties of
light, in the hope of arriving at a mechanismto which they could
be attnbuted; he succeededin accounting for the phenomenain
terms of a few simple principles, but was not able to specify an
aether which would in turn account for these principles. The
, displacement' of Fresnelcould not be a displacementin an elastic
solid of the usualtype, sinceits normal componentis not continuous
acrossthe interfacebetweentwo media.3
The theory of ordinary reflection was completedby a discussion
of the casein which light is reflected totally., :This had formed the
subject of some of Fresnel'sexperimental researchesseveral years
before; and in two papers' presentedto the Academyin November
1817and January 1818,he had shown that light polarised in any
plane inclined to the plane of reflection is partly , depolarised' by
total reflection, and that this is due to differencesof phasewhich
are introduced between the components polarised in and per-
pendicular to the plane of reflection. 'When the reflection is total,'
hesaid,, rayspolarisedin the plane of reflection are reflected nearer
Article' Chromatics,'Eru;ycl. Brill. Suppl.

I I Mim. de l'/nsl. ii, 1817,p. 305
Fresnel'stheory of reflectioncan, however,be reconciledwith the electromagnetic
I
theoryof light, by identifying his ' displacement' with the electricforce.

• (Euvres de Fresnel, i, pp. 441, 487
125
the surfaceof the glassthan those polarised at right anglesto the

sameplane, sothat there is a differencein the paths described.'
This changeof phasehe now deducedfrom the formulae already
obtained for ordinary reflection. Considering light polarised in
the plane of reflection, the ratio of the amplitudes of the reflected
and incident light is, as we have seen,
sin (i - r) •
sin (i +r) ,
when the sineof the angle of incidence is greater than fJ-Z/fJ-l' sothat
total reflection takesplace, this ratio may be written in the form
8v-=-1
e
where 8 denotesa real quantity defined by the equation
( fJ-l z sOn2 Zo - fJ-z2)1 .
t an"2"18 _
-
1
fJ-l cos t
0.
Fresnelinterpreted this expressionto mean 1 that the amplitude
of the reflected light is equal to that of the incident, but that the
two wavesdiffer in phaseby an amount O. The caseof light polarised
at right anglesto the plane of reflection may be treated in the same
way, and the resulting formulae are completely confirmed by
experiment.
A few monthsafter the memoir on reflection had beenpresented,
Fresnelwas electedto a seatin the Academy; and during the rest
of his short life honourscameto him both from France and abroad.
In 1827the Royal Society awarded him the Rumford medal; but
Arago, to whom Young had confided the mission of conveying the
medal, found him dying; and eight days afterwards he breathed
his last.
By the genius of Young and Fresnel the wave-theory of light
was establishedin a position so strong that henceforth the cor-
puscular hypothesiswas unable to recruit any adherentsamong the
younger men.2 Two striking experiments which were devised
later seemedto be decisivein the controversy. The first was due
to G. B. Airy (1801-92), who remarked 3 that according to the
corpuscular theory the colours of thin plates are produced solely
by the light reflectedfrom the secondsurfaceof the plate, whereasin
the wave-theorythey dependon interferencebetweenlight reflected
1 On this interpretation cr. De Morgan's Trigonometry and Double Algebra, pp. 119,139
• Brewster(b. 1781)still inclined to the corpusculartheory in the 1830's.
• Trans. Camb. Phil. Soc. iv (1833),p. 279 (read 14Nov. 1831)
126
at the first and secondsurfaces. Hence if, in the production of

Newton's rings, we can by any means prevent reflection at the
first surface,the rings ought not to appear according to the wave-
theory, but ought still to be seenif the other hypothesisis true.
Airy arranged an experiment in which a lens was laid upon a
polished metallic surface. By using light polarised at right angles
to the plane of reflection, and incident at the polarising angle, he
securedthat the light coming to the eye should be that reflected
from the metal, unmixed with other light: the rings were not seen,
sothe verdict wasin favour of the wave-theory.
The secondexperimentwasperformedin 1850,when Foucault 1
and Fizeau,2carrying out a plan long before imagined by Arago,
directly measuredthe velocity of light in air and in water, and
found that on the questionsolong debatedbetweenthe rival schools
the adherentsof the undulatory theory had beenin the right.
1 ComptesRmdus, xxx (1850), p. 551 2 ibid. p. 562
127
Chapter V
THE AETHER AS AN ELASTIC SOLID
WHEN Young and Fresnelput forward the view that the vibrations
of light are performedat right anglesto its direction of propagation,

they at the same time pointed out that this peculiarity might be
explained by making a new hypothesisregarding the nature of
the luminiferous medium; namely, that it possesses the power of
resisting attempts to distort its shape. It is by the possessionof
such a power that solid bodiesare distinguishedfrom fluids, which
offer no resistanceto distortion; the idea of Young and Fresnelmay
thereforebe expressedby the simple statementthat the aether behaves
as an elastic solid. Mter the death of Fresnel this conception was
developedin a brilliant seriesof memoirs to which our attention
must now be directed.
The elastic-solidtheory meetswith one obviousdifficulty at the
outset. If the aether has the qualities of a solid, how is it that the
planets in their orbital motions are able to journey through it at
immensespeedswithout encountering any perceptible resistance?
This objection wasfirst satisfactorilyansweredby Sir GeorgeGabriel
Stokes1 (1819-1903), who remarked that such substancesas pitch
and shoemaker'swax, though so rigid as to be capable of elastic
vibration, are yet sufficiently plastic to permit other bodies to
passslowly through them. The aether, he suggested,may have
this combination of qualities in an extreme degree, behaving
like an elastic solid for vibrations so rapid as those of light, but
yielding like a fluid to the much slower progressivemotionsof the
planets.
Stokes'sexplanation harmonisesin a curious way with Fresnel's
hypothesisthat the velocity of longitudinal wavesin the aether is
indefinitely great compared with that of the transversewaves;
for it is found by experiment with actual substancesthat the ratio
of the velocity of propagation of longitudinal waves to that of
transversewaves increasesrapidly as the medium becomessofter
and more plastic.
1 Trans. Camb. Phil. Soc. viii (1845), p. 287

128
In attempting to set forth a parallel between light and the

vibrations of an elastic substance, the investigator is compelled
more than once to make a choice between alternatives. He may,
for instance,supposethat the vibrations of the aether are executed
either parallel to the plane of polarisation of the light or at right
anglesto it; and he may supposethat the different refractive powers
of different media are due either to differencesin the inertia of the
aether within the media, or to differencesin its power of resisting
distortion, or to both thesecausescombined. There are, moreover,
severaldistinct methodsfor avoiding the difficulties causedby the
presenceof longitudinal vibrations; and as, alas! we shall see,a
further sourceof diversity is to be found in that liability to error
from which no man is free. It is therefore not surprising that the
list of elastic-solidtheoriesis a long one.
At the time when the transversality of light was discovered,no
generalmethodhad beendevelopedfor investigating mathematically
the propertiesof elastic bodies; but under the stimulus of Fresnel's
discoveries,someof the best intellects of the age were attracted to
the subject. The volume of Memoirs of the Academy which contains
Fresnel'stheory of crystal-opticscontains also a memoir by Claud
Louis-Marie-Henri Navier 1 (1785-1836), at that time Professorof
Mechanics in Paris, in which the correct equations of vibratory
motion for a particular type of elastic solid were for the first
time given. Navier supposed the medium to be ultimately
constituted of an immense number of particles, which act on
each other with forces directed along the lines joining them, and
depending on their distancesapart; and showedthat if e denote
the (vector) displacementof the particle whoseundisturbed position
is (x, y, Z), and if p denote the density of the medium, the
equation of motion is
ere
p otl
= - 3n grad div e - n curl curl e,
where n denotesa constant which measuresthe rigidity, or power

of resisting distortion, of the medium. All such elastic properties
of the body asthe velocity of propagationof wavesin it mustevidently
dependon the ratio nip.
Among the refereesof one of Navier's papers was Augustine-
Louis Cauchy (1789-1857), one of the greatest analysts of the
1 Mem. de l'Acad. vii, p. 375. The memoir was presentedin 1821,and publishedin
1827.
129
AETHER AND ELECTRICITY
nineteenth century,l who, becoming interested in the question,

published in 18282 a discussionof it from an entirely different point
of view. Instead of assuming,asNavier had done, that the medium
is an aggregateof point-centresof force, and thus involving himself
in doubtful molecular hypotheses,he devised a method of directly
studying the elastic properties of matter in bulk, and by its means
showedthat the vibrations of an isotropic solid are determined by
the equation
ote
Pot2 = - (k + in) grad dive - n curl curl e ;
here n denotes,as before,the constantof rigidity; and the constant

k, which is called the modulus of compression,3denotesthe ratio of a
pressure to the cubical compressionproduced by it. Cauchy's
equation evidently differs from Navier's in that two constants,k
and n, appear instead of one. The reasonfor this is that a body
constituted from point-centres of force in Navier's fashion has its
moduli of rigidity and compressionconnectedby the relation ~
Actual bodies do not necessarilyobey this condition; e.g. for

1 Hamilton's opinion, written in 1833, is worth repeating: • The principal theories
of algebraical analysis (under which I include Calculi) require to be entirely remodelled;
and Cauchy has done much already for this great object. Poissonalso has done much;
but he does not seem to me to have nearly so logical a mind as Cauchy, great as his
talents and clearness are; and both are in my judgment very far inferior to Fourier,
whom I place at the head of the French School of Mathematical Philosophy, even above
Lagrange and Laplace, though I rank their talents abov.ethose of Cauchy and Poisson.'
(Life of Sir W. R. Hamilton, ii, p. 58.)
William Thomson (Lord Kelvin) and Heaviside were warm admirers of Fourier. In
May 1840Thomson, being then aged fifteen, borrowed Fourier's Theone de la Chaleur from
the Glasgow University library, and read it through in a fortnight. It was the strongest
of all the influences that shaped his mind .
• No-one admires Fourier more than I do,' said Heaviside (Electromagnetic Theory,
ii, p. 32).
2 Cauchy, Exercicesde Mathematiques, iii (1828), p. 160
• This notation was introduced at a later period, but is used here in order to avoid
subsequent changes.
• In order to construct a body whoseelastic properties are not limited by this equation,
William John Macquorn Rankine (1820-72) considered a continuous fluid in which a
number of point-centres of force are situated; the fluid is supposed to be partially con-
densed round these centres, the elastic atmosphere of each nucleus being retained round
it by attraction. An additional volume elasticity due to the fluid is thus acquired; and
no relation between k and 11 is now necessary. ef. Rankine's Miscellaneous Scientific Papers,
pp. 81 sqq.
Sir William Thomson (Lord Kelvin), in 1889, formed a solid not obeying Navier's
condition by using pairs of dissimilar atoms. ef. Thomson's Papers, iii, p. 395; ef. also
Baltimore Lectures, pp. 123 sqq.
india-rubber k is much larger than 5n/3,1and there seemsto be no

reasonwhy we should imposeit on the aether.
In the sameyear Poisson2 succeededin solving the differential
equationwhich had thus beenshownto determinethe wave-motions
possiblein an elastic solid. The solution, which is both simple and
elegant, may be derived as follows: let the displacementvector e
be resolved into two components,of which one c is circuita!, or
satisfiesthe condition d· 0
IV C = ,
while the other b is irrotational, or satisfiesthe condition

curl b = O.
The equation takesthe form
0 + c) -
1
Pats(b nv'c -
,(
k + 34)n v'b = o.
The terms which involve b and thosewhich involve c must be
separately zero, since they represent respectively the irrotational
and the circuital parts of the equation. Thus, c satisfiesthe pair
of equations
div c = OJ
while b is to be determinedfrom
curl b = O.
A particular solution of the equationsfor c is easilyseento be
c" = A sin A (~ - tJ ~), c" = B sin A (~ - tJ~), c, = 0,
which representsa transverseplane wave propagated with velocity

Y(n/p). It can be shownthat the generalsolution of the differential
equationsfor c is formed of such waves as this, travelling in all
directions, superposedon eachother.
A particular solution of the equationsfor b is
b" = 0, by = 0, b, = C sin A (~ - t J ~n),

k ~
1 It may, however, be objected that india-rubber and other bodies which fail to fulfil
Navier's relation are not true solids. On this historic controversy, cf. Todhunter and
Pearson's Hiswry of Elasticity, i, p. 496.
1 Mlm. de l'Acad. viii (1828), p. 623. Poissontakes the equation in the restricted form
given by Navier; but this does not affect the question of wave-propagation.
(823) 131
which represents a longitudinal wave propagated with velocity

v(k + {n)/p ;
the general solution of the differential equation for b is formed by
the superpositionof such wavesas this, travelling in all directions.
Poissonthus discovered that the waves in an elastic solid are
of two kinds: those in c are transverse,and are propagated with
velocity (nip)! ; while thosein b arelongitudinal, and arepropagated
with velocity { (k + {n)lp}!. The latter are 1 wavesof dilatation and
condensation,like soundwaves; in the c waves,on the other hand,
the medium is not dilated or condensed,but only distorted in a
manner consistentwith the preservationof a constantdensity.2
The researcheswhich have been mentioned hitherto have all
been concernedwith isotropic bodies. Cauchy in 1828 3 extended
the equations to the caseof crystalline substances. This, however,
he accomplishedonly by reverting to Navier's plan of conceiving
an elastic body as a cluster of particles which attract each other
with forces depending on their distancesapart; the aelotropy he
accountedfor by supposingthe particles to be packed more closely
in somedirections than in others.
The general equations thus obtained for the vibrations of an
elastic solid contain twenty-one constants; six of thesedepend on
the initial stress,so that if the body is initially without stress,only
fifteen constantsarc involved. If, the initial stressbeing retained, the
medium is supposedto be symmetricalwith respectto three mutually
orthogonal planes,the twenty-one constantsreduce to nine, and the
equationswhich determinethe vibrations may be written in the form (
?ot2e2 x = (a + G) o2ex
ox2
+ (h + H) 02ex
oy2
+ (g + I) 02e
oZ2
x
~ (aoex + h oey + oez),

+ 2 ox ox Oy g oZ
1 cr. Stokes,• On the Dynamical Problemof Diffraction,' Comb. Phil. Trans. ix, 1849,
p. 1 : Math. and Phys. Papers ii, p. 243

I It may easilybeshownthat any disturbance,in eitherisotropicor crystallinemedia,
for which the direction of vibration of the moleculeslies in the wave-frontor surfaceof
constantphase,must satisfythe equation
dive = 0,
wheree denotesthe displacement; if, on the other hand, the direction of vibration of the
moleculesis perpendicularto the wave-front,the disturbancemust satisfythe equation
curl e = O.
Theseresultswereproved by M. O'Brien, Trans. Comb. Phil. Soc. vii (1842),p. 397
• Exercices tk Math. iii (1828),p. 188
• Theseare substantiallyequations(68) on p. 208of the third volumeof the Exercices.
132
and two similar equations. The three constantsG, H, I represent

the stressesacrossplanes parallel to the co-ordinate planes in the
undisturbedstateof the aether.1
On the basisof theseequations, Cauchy worked out a theory
of light, of which an instalment relating to crystal-opticswas pre-
sented to the Academy in 1830.2 Its characteristic features will
now be sketched.
By substitution in the equationslast given, it is found that when
the wave-front of the vibration is parallel to the plane of yz, the
velocity of propagation must be (h + G)~ if the vibration takes
placeparallel to the axis ofy, and (g + G)~ if it takesplace parallel
to the axis of z. Similarly when the wave-front is parallel to the
plane of zx, the velocity must be (h +
H)~ if the vibration is parallel
to the axis of x, and (1+ H)~ if it is parallel to the axis of z; and
when the wave-frontis parallel to the plane of xy, the velocity must
be (g +I)t if the vibration is parallel to the axis of x, and (1 I)t +
if it is parallel to the axis of y.
Now it is known from experiment that the velocity of a ray
polarised parallel to one of the planes in question is the same,
whetherits direction of propagationis along one or the other of the
axes in that plane; so, if we assumethat the vibrations which
constitute light are executedparallel to the plane of polarisation,
we must have
1+ H = 1+ I, g +I =g + G, h +H =h +G;
or, G = H = I.
This is the assumptionmade in the memoir of 1830; the theory
basedon it is generally known as Cauchy's First Theory 3 ; the equili-
brium pressuresG, H, I being all equal, are taken to be zero.
If, on the other hand, we make the alternative assumptionthat
the vibrations of the aetherare executedat right anglesto the plane
of polarisation,we must have
h +H = g + I, 1+ I = h + G, g + G = 1+ H ;
G, H, I are tensions when they are positive and pressureswhen they are negative.
1
Mem. de I'Acad. x, p. 293.

S
In the previous year (Mem. de l'Acad. ix, p. 114) Cauchy had stated that the equations
of e1asticity lead in the caseof uniaxal crystals to a wave-surface of which two sheetsare a
sphere and spheroid as in Huygens' theory.
• The equations and results of Cauchy's First Theory of crystal-optics were independ-
ently obtained shortly afterwards by Franz Ernst Neumann (1798-1895); cr. Ann. d.
Phys. xxv (1832), p. 418, reprinted as no. 76 of Ostwald's Klassiker der exakten Wwenschaften,
with notes by A. Wangerin.
133
the theory based on this supposition is known as Cauchy's Second

Theory; it waspublishedin 1836.1
In both theoriesCauchy imposesthe condition that the section
of two of the sheetsof the wave-surfacemade by anyone of the
co-ordinate planesis to be formed of a circle and an ellipse, as in
Fresnel'stheory; this yields the three conditions
3be = f(b + e +f); 3ea = g(e + a + g); 3ab = h(a + b + h).
Thus in the first theory we havethesetogetherwith the equations
G = 0, H = 0, I = 0,
which expressthe condition that the undisturbed stateof the aether
is unstressed; and the aethereal vibrations are executedparallel
to the plane of polarisation. In the secondtheory we have the
three first equations,togetherwith
f-G=g-H=h-I;
and the plane of polarisation is interpreted to be the plane at right
anglesto the direction of vibration of the aether.
Either of Cauchy's theories accounts tolerably well for the
phenomenaof crystal-optics; but the wave-surface(or rather the
two sheetsof it which correspondto nearly transversewaves)is not
exactly Fresnel's. In both theories the existenceof a third wave,
formed of nearly longitudinal vibrations, is a formidable difficulty.
Cauchy himself anticipated that the existenceof thesevibrations
would ultimately be demonstrated by experiment, and in one
place 2 conjectured that they might be of a calorific nature. A
further objection to Cauchy's theoriesis that the relations between
the constantsdo not appear to admit of any simple physical inter-
pretation, being evidently assumedfor the sole purposeof forcing
the formulae into some degree of conformity with the results of
experiment. And further difficulties will appear when we proceed
subsequentlyto compare the properties which are assignedto the
aether in crystal-optics with those which must be postulated in
order to accountfor reflection and refraction.
To the latter problem Cauchy soon addressedhimself, his
investigations being in fact published 3 in the same year (1830)
as the first of his theoriesof crystal-optics.
At the outsetof any work on refraction, it is necessaryto assign
a causefor the existenceof refractive indices, i.e. for the variation
1 ComptesRtndus, ii (1836), p. 341 ; Mem. de l'Acad. xviii (1839), p. 153
• Mim. de Z'Acad. xviii (1839, published 184-2),p. 153
• Bull. des SciencesMath. xiv (1830), p. 6
134
in the velocity of light from one body to another. Huygens, as we

have seen,suggestedthat transparent bodiesconsistof hard particles
which interact with the aethereal matter, modifying its elasticity.
Cauchy in his earlier papers1 followed this lead more or lessclosely,
assumingthat the density p of the aether is the samein all media,
but that its rigidity n varies from one medium to another.
Let the axis of x be taken at right anglesto the surfaceof separa-
tion of the media, and the axis of z parallel to the intersectionof this
interface with the incident wave-front; and suppose,first, that the
incident vibration is executedat right anglesto the plane of incidence,
so that it may be representedby
e, = f (- x cosi - y sin i + J~t),

where i denotesthe angle of incidence; the reflected wave may
be representedby
e, = F (x cosi - y sin i + J] t),

and the refracted wave by
e, = ft (-x cosr - y sin r + J~t),

where r denotesthe angle of refraction, and n' the rigidity of the
secondmedium.
To obtain the conditions satisfied at the reflecting surface,
Cauchy assumed (without assigning reasons) that the x and y
componentsof the stressacrossthe ~ plane are equal in the media
on either side of the interface. This implies in the presentcasethat
the quantities 0 0
n -.!..=. and n-!!.
ox Oy
are to be continuous acrossthe interface: so we have
n cosi. (f' - F') = n' cosr .ft'; n sin i. (f' + F') = n' sin r .ft'.
Eliminating ft', we have
F' _ sin (r - i)
f' - sin (r +
if
Now this is Fresnel's sine law for the ratio of the intensity of
the reflected ray to that of the incident ray; and it is known that
the light to which it appliesis that which is polarised parallel to the
t As will appear,his viewson this subjectsubsequentlychanged.
135
plane of incidence. Thus Cauchywasdriven to the conclusionthat,

in order to satisfythe known facts of reflection and refraction, the
vibrations of the aether must be supposedexecutedat right angles
to the plane of polarisation of the light.
The caseof a vibration performed in the plane of incidencehe
discussedin the same way. It was found that Fresnel'stangent
law could be obtained by assumingthat ex and the normal pressure
acrossthe interface have equal valuesin the two contiguousmedia.
The theory thusadvancedwasencumberedwith manydifficulties.
In the first place, the identification of the plane of polarisationwith
the plane at right anglesto the direction of vibration was contrary
to the only theory of crystal-opticswhich Cauchy had as yet pub-
lished. In the secondplace, no reasonswere given for the choice
of the conditions at the interface. Cauchy's motive in selecting
these particular conditions was evidently to securethe fulfilment
of Fresnel'ssine law and tangent law; but the resultsare incon-
sistent with the true boundary conditions,which were given later
by Green.
It is probable that the results of the theory of reflection had
much to do with the decision,which Cauchy now made,1to reject
the first theory of crystal-optics in favour of the second. After
1836 he consistentlyadheredto the view that the vibrations of the
aether are performed at right anglesto the plane of polarisation.
In that year he made another attempt to frame a satisfactorytheory
of reflection,2basedon the assumptionjust mentioned, and on the
following boundary conditions: at the interface between two
media curl e is to be continuous,and (taking the axis of x normal
to the interface) oex/ox is alsoto be continuous.
Again we find no very satisfactory reasonsassignedfor the
choice of the boundary conditions; and as the continuity of e
itself acrossthe interface is not included amongst the conditions
chosen, they are obviously open to criticism; but they lead to
Fresnel'ssine and tangent equations, which correctly expressthe
actual behaviour of light.3 Cauchy remarksthat in order to justify
them it is necessaryto abandonthe assumptionof his earlier theory,
that the density of the aetheris the samein all material bodies.
1 Comptes&ndus, ii (1836), p. 341
I Comptes Rendus, ii (1836), p. 341: • Memoire sur la dispersion de la lumihe'
(Nouveaux eXeTcices de Math. [1836], p. 203)
I These boundary conditions of Cauchy's are, as a matter of fact, satisfied by the
electric force in the electromagnetic theory of light. The continuity of curl e is equivalent
to the continuity of the magnetic vector acrossthe interface, and the continuity of 8e,;/8x
leads to the same equation as the continuity of the component of electric force in the
direction of the intersection of the interface with the plane of incidence.
136
It may be remarked that neither in this nor in Cauchy's earlier

theory of reflection is any trouble caused by the appearance of
longitudinal waves when a transverse wave is reflected, for the
simple reasonthat he assumesthe boundary conditions to be only
four in number; and thesecan all be satisfiedwithout the necessit.y
for introducing any but transversevibrations.
These features bring out the weaknessof Cauchy's method of
attacking the problem. His object was to derive the properties
of light from a theory of the vibrations of elastic solids. At the
outset he had already in his possession the differential equationsof
motion of the solid, which were to be his starting-point, and the
equationsof Fresnel,which were to be his goal. It only remained
to supplythe boundary conditionsat an interface,which are required
in the discussionof reflection, and the relations betweenthe elastic
constantsof the solid, which are required in the optics of crystals.
Cauchy seemsto have considered the question from the purely
analytical point of view. Given certain differential equations,what
supplementaryconditions must be adjoined to them in order to
producea given analytical result? The problem when statedin this
form admits of more than one solution; and hence it is not sur-
prising that within the spaceof ten years the great French mathe-
matician produced two distinct theoriesof crystal-optics and three
distinct theoriesof reflection,1 almost all yielding correct or nearly
correctfinal formulae, and yet mostly irreconcilable with eachother,
and involving incorrect boundary conditions and improbable
relations betweenelasticconstants.
Cauchy's theories,then, resembleFresnel'sin postulating types
of elasticsolid which do not exist, and for whoseassumedproperties
no dynamical justification is offered. The sameobjection applies,
though in a lessdegree,to the original form of a theory of reflection
and refraction which was discovered about this time 2 almost
simultaneouslyby James MacCullagh (1809-47) of Trinity College,
Dublin, and Franz Neumann (1798-1895) of Konigsberg. To
theseauthorsis due the merit of having extendedthe laws of reflec-
tion to crystalline media; but the principles of the theory were
lOne yet remains to be mentioned.
I The outlines of the theory were published by MacCullagh in Brit. Assoc.Rep., 1835 ;
and his results were given in Phil. Mag. x (Jan. 1837) and in Proc. Roy. Irish Acad. xviii, Jan.
1837; ef. Hamilton's memoir of MacCullagh in Proc. R. S., v (1847), p. 712. Neumann's
memoir was presented to the Berlin Academy towards the end of 1835, and published in
1837 in Abh. Berl. Ak. aus demJahre 1835, Math. Klasse, p. I. So far as publication is con·
cerned, the priority would seem to belong to MacCullagh; but there are reasons fOI
believing that the priority of discovery really rests with Neumann, who had arrived at his
equations a year before they were communicated to the Berlin Academy.
137
originally derived in connection with the simpler caseof isotropic

media, to which our attention will for the presentbe confined.
MacCullagh and Neumann felt that the great objection to
Fresnel'stheory of reflection was its failure to provide for the con-
tinuity of the normal component of displacementat the interface
betweentwo media; it is obvious that a discontinuity in this com-
ponent could not exist in any true elastic-solidtheory, sinceit would
imply that the two media do not remain in contact. Accordingly,
they made it a fundamental condition that all three componentsof
the displacement must be continuous at the interface, and found
that the sine law and tangent law can be reconciled with this
condition only by supposingthat the aether vibrations are parallel
to the plane of polarisation: which supposition they accordingly
adopted. In place of the remaining three true boundary conditions,
however,they usedonly a singleequation, derived by assumingthat
transverseincident wavesgive rise only to transversereflected and
refracted waves,and that the conservationof energyholds for these
-i.e. that the massesof aether put in motion, multiplied by the
squaresof the amplitudes of vibration, are the same before and
after incidence. This is, of course, the same device as had been
usedpreviously by Fresnel; it must, however,be remarkedthat the
principle is unsound as applied to an ordinary elastic solid; for
in such a body the refracted and reflectedenergy would in part be
carried away by longitudinal waves.
In order to obtain the sine and tangent laws, MacCullagh and
Neumannfound it necessaryto assumethat the inertia of the lumini-
ferous medium is everywherethe same,and that the differencesin
behaviour of this medium in different substancesare due to differ-
ences in its elasticity. The two laws may then be deduced in
much the sameway as in the previousinvestigationsof Fresneland
Cauchy.
Although to insist on continuity of displacementat the interface
was a decided advance, the theory of MacCullagh and Neumann
scarcelyshowed asyet much superiority over the quasi-mechanical
theoriesoftheir predecessors.Indeed, MacCullagh himselfexpressly
disavowed any claim to regard his theory, in the form to which
it had then been brought, as a final explanation of the properties
oflight. 'If we are asked,'he wrote, , what reasonscan be assigned
for the hypotheseson which the preceding theory is founded, we
are far from being able to give a satisfactory answer. We are
obliged to confessthat, with the exceptionof the law of vis viva, the
hypothesesare nothing more than fortunate conjectures. These
138
conjecturesare very probably right, since they have led to elegant

laws which are fully borne out by experiments; but this is all we
can assertrespecting them. We cannot attempt to deduce them
from first principles; because,in the theory of light, suchprinciples
are still to be soughtfor. It is certain, indeed,that light is produced
by undulations, propagated, with transversalvibrations, through a
highly elastic aether; but the constitution of this aether, and the
laws of its connection (if it has any connection) with the particles
of bodies,are utterly unknown.'
The needful reformation of the elastic-solidtheory of reflection
was effectedby Green, in a paper 1 read to the Cambridge Philo-
sophical Society in December 1837. Green, though inferior to
Cauchyas an analyst,was his superior in physical insight. Instead
of designingboundary equationsfor the expresspurposeof yielding
Fresnel'ssine and tangent formulae, he set to work to determine
the conditions which are actually satisfiedat the interfacesof real
elasticsolids.
These he obtained by means of general dynamical principles.
In an isotropic medium which is strained, the potential energy per
unit volume due to the stateof stressis
rP = ~
2
(k + 4.3 n)( oe%+ oey + oez)B+ !n {(oez + oey)2 + (oe%+ oez)1
ox Oy oz 2 Oy oz oz ox
+ (oey + oe%)1_ 4 oer oez_ 4 oezoe%_ 4 oc%~}
ox oy oy oz oz ox ax Oy ,
where e denotes the displacement, and k and n denote the two
elastic constants already introduced; by substituting this value
of rP in the generalvariational equation
JH {a;;: oex + ~;:y oey + ~2t~Zoez} dxdydz = - HJ 0rP dxdydz

p
(where p denotes the density), the equation of motion may be

deduced.
But this method does more than merely furnish the equation
of motion
p ~:~ = - (k + In) grad div e - n curl curl e ;

or,
p ~t~ = - (k + ~n) grad dive + nY' e, 2
which had already been obtained by Cauchy; for it also yields

1 Trans. Camb. Phil. Soc. vii (1838), pp. I, 113; Greeo's Math. Papers, p. 245
139
the boundary conditions which must be satisfied at the interface

between two elastic media in contact; these are, as might be
guessedby physical intuition, that the three componentsof the
displacement1 and the three componentsof stressacrossthe inter-
face are to be equal in the two media. If the axis of x be taken
normal to the interface, the latter three quantities are
(k - ~3 n) d· e + 2n oex
IV ox' ox + oe
n (oe,
x)
0;:;'
and
n (~x
Oy
+ oe y
ox· )
The correct boundary conditions being thus obtained, it was a
simple matter to discussthe reflection and refraction of an incident
wave by the procedureof Fresneland Cauchy. The result found by
Greenwasthat if the vibration of the aetherealmoleculesis executed
at right anglesto the plane of incidence,the intensity of the reflected
light obeys Fresnel's sine law, provided the rigidity n is assumed
to be the same for all media, but the inertia p to vary from one
medium to another. Since the sine law is known to be true for
light polarised in the plane of incidence, Green's conclusion con-
firmed the hypothesesof Fresnel, that the vibrations are executed
at right angles to the plane of polarisation, and that the optical
differencesbetweenmedia are due to the different densitiesof aether
within them.
I t now remained for Green to discussthe case in which the
incident light is polarised at right anglesto the plane of incidence,
so that the motion of the aetherealparticles is parallel to the inter-
sectionof the plane of incidencewith the front of the wave. In this
caseit is impossibleto satisfyall the six boundary conditionswithout
assumingthat longitudinal vibrations are generatedby the act of
reflection. Taking the plane of incidence to be the plane of yz,
and the interface to be the plane of xy, the incident wave may be
representedby the equations
ey = A~f(t + I;:;+ ~v); e, = - A.~lf(t + lz + my);

where, if i denotethe angle of incidence,we have
l = J~cosi, m= - J~sin i.
There will be a transversereflectedwave,
ey = B~f(t -I;:; + my); e, = - B{f(t -I;:; + my);

I Thesefirst three conditionsare of coursenot dynamicalbut geometrical.
140
and a transverserefracted wave,
e. = - C ayf(t
a + lIz + my),
where,sincethe velocity of transversewavesin the secondmedium
is vn/p2' we can determine11from the equation
there will alsobe a longitudinal reflectedwave,

a
ey = D-ayf(t - Az + my); a
e. = D &f(t - Az + my),
where>'is determinedby the equation
A' +m2 =
kl +PI ~n'•
and a longitudinal refracted wave,
a a
ey = E iyf(t + A1Z + my); e. = E bi.f(t + AlZ + my),
where>'1is determinedby
AI' + m' = k I ~ an 4 •
Substituting thesevalues for the displacementin the boundary

conditions which have been already formulated, we obtain the
equations which determine the intensities of the reflected and
refracted waves; in particular, it appears that the amplitude of
the reflectedtransversewave is given by the equation
A - B _ 11Pl +m 2
(PI - PI)'
A + B - lp. T p. (API + AlPl)"
Now if the elasticconstantsof the media are suchthat the velocities
of propagation of the longitudinal waves are of the same order
of magnitude as thoseof the transversewaves,the direction cosines
of the longitudinal reflected and refracted rays will in general have
real values, and these rays will carry away some of the energy
which is brought to the interface by the incident wave. Green
avoided this difficulty by adopting Fresnel's suggestion that the
resistanceof the aether to compressionmay be very large in com-
parison with the resistanceto distortion, asis actually the casewith
suchsubstances asjelly and caoutchouc; in this casethe longitudinal
141
wavesare degradedin much the sameway asthe transverserefracted

ray is degradedwhen there is total reflection, and so do not carry
away energy. Making this supposition,so that k1 and k2 are very
large, the quantities A and Al have the value m and we have v~
A - B =!.!-f2 _ '!!: (PI - P2)2 v-I
A B + l PI l P2(PI + PI) •
Thus we have
B + 1)1(PI _.!2)1 + (~_ 1)· m:

II
2 (~
PI PI l PI l
A = (~ + 1)1(PI +-.b)1 + (~ _ 1)4 m 1
2
•
PI PI l PI l
This expressionrepresentsthe ratio of the intensity of the trans-
versereflectedwave to that of the incident wave. It doesnot agree
with Fresnel'stangent formula; and both on this account and also
because(aswe shall see)this theory of reflection doesnot harmonise
well with the elastic-solidtheory of crystal-optics, it must be con-
cluded that the vibrations of a Greeniansolid do not furnish an exact
parallel to the vibrations which constitute light.
The successof Green's investigation from the standpoint of
dynamics, set off by its failure in the details last mentioned,
stimulated MacCullagh to fresh exertions. At length he succeeded
in placing his own theory, which had all along been free from
reproach so far as agreement with optical experiments was con-
cerned, on a sound dynamical basis; thereby effecting that recon-
ciliation of the theories of light and dynamics which had been the
dream of every physicistsincethe daysof Descartes.
The central feature of MacCullagh's investigation,1 which was
presentedto the Royal Irish Academy in 1839,is the introduction
of a new type of elastic solid. He had, in fact, concluded from
Green'sresults that it was impossibleto explain optical phenomena
satisfactorily by comparing the aether to an elastic solid of the
ordinary type, which resistscompressionand distortion; and he
saw that the only hope of the situation was to devise a medium
which should be asstrictly conformableto dynamical lawsasGreen's
elastic solid, and yet should have its properties specially designed
to fulfil the requirementsof the theory of light. Such a medium
he now described.
If as before we denote by e the vector displacementof a point
of the medium from its equilibrium position, it is well known that
1 Trans. Roy. Irish Acad. xxi (1848) p. 17; MacCullagh's Coli. Works, p. 145
142
the vector curl e denotestwice tIle rotation of the part of the solid
in the neighbourhoodof the point (x, y, z) from its equilibrium
orientation. In an ordinary elastic solid, the potential energy of
strain dependsonly on the changeof size and shapeof the volume
elements-on their compression and distortion, in fact. For
MacCullagh'snew medium, on the other hand, the potential energy
dependsonly on the rotation of the volume elements.
Since the medium is not supposedto be in a state of stressin
its undisturbed condition, the potential energy per unit volume
must be a quadratic function of the derivatesof e; so that in an
isotropic medium this quantity 4> must be formed from the only
invariant which dependssolely on the rotation and is quadratic in
the derivates,that is from (curl e)2; thus we may write
..I.
'I'
= 1
'II-'
{(aeOys _ aev)1+ (aeazz _~z)1
az ax + (ae
y
ax _ aex)l}
Oy ,
whereI-' is a constant.
The equation of motion is now to be determined,as in the case
of Green'saether,from the variational equation
HI g;:x Sez + a;;fSey + ~;~Se,} dx t!J dz

p = - HIS
4> dxt!J dz;
the result is ale
p atl
= - ,.,.
curl curl e.
It is evident from this equation that if div e is initially zero

it will always be zero; we shall supposethis to be the case,so
that no longitudinal wavesexist at any time in the medium. One
of the greatestdifficulties which beset elastic-solidtheories is thus
completelyremoved.
The equation of motion may now be written
()Ie _ I
P 8jI- 1-'''1 e,
which showsthat transversewaves are propagated with velocity

V(I-'/p),
From the variational equation we may also determine the
boundary conditions which must be satisfied at the interface
between two media; these are, that the three components of e
are to be continous acrossthe interface, and that the two com-
ponentsof I-' curl e parallel to the interface are alsoto be continuous
acrossit. One of thesefive conditions, namely, the continuity of
143
the normal component of e, is really dependenton the other four;

for if we take the axis of x normal to the interface, the equation of
motion gives
a
a2ex
p at2 = - Oy (I-' curl e)z + aza (I-' curl e)",
and as the quantities p, (I-' curl e)z, and (I-' curl e)" are continuous
acrossthe interface, the continuity of 02ex/at2 follows. Thus the
only independent boundary conditions in MacCullagh's theory are
the continuity of the tangential componentsof e and of I-' curl e.l
It is easilyseenthat theseare equivalent to the boundary conditions
used in MacCullagh's earlier paper, namely, the equation of vis
viva and the continuity of the three componentsof e; and thusthe
, rotationally elastic' aether of this memoir furnishes a dynamical
foundation for the memoir of 1837.
The extension to crystalline media is made by assuming the
potential energy per unit volume to have, when referred to the
principal axes,the form
A (oez _ ?!1)2 + B (oex _ + C (oey _ Oe )2

i~z)1 x
Oy oz az ax ox Oy'
where A, B, C denote three constantswhich determine the optical
behaviour of the medium; it is readily seen'that the wave-surface
is Fresnel's,and that the plane of polarisation containsthe displace-
ment, and is at right anglesto the rotation.
MacCullagh's work was regarded with doubt by his own and
the succeeding generation of mathematical physicists, and can
scarcelybe said to have been properly appreciateduntil FitzGerald
drew attention to it forty years afterwards. But there can be no
doubt that MacCullagh really solved the problem of devising a
medium whosevibrations, calculated in accordancewith the correct
laws of dynamics,should have the sameproperties asthe vibrations
of light.
1 MacCullagh's equations may readily be interpreted in the electromagnetic theory
of light (see chap. viiii and ix of this book); e corresponds to the magnetic force,
,...curl e to the electric force, and curl e to the electric displacement. This interpretation
is due to Hea,oiside,Electrician, xxvi (1891), p. 360: Lord Kelvin had regarded rotation as
0:,
the analogue of magnetic force. Heaviside's discussion is as follows: let E be the torque.
The force is -curl E, so we have the equation of motion - curl E = ,... if H is the
velocity and,... the density. Also the torque is proportional to the rotation, so curl H
= co:' where c is the reciprocal of the quasi-rigidity. These are the electromagnetic
equations.
144
The hesitationwhich wasfelt in acceptingthe rotationally elastic

aetherarosemainly from the want of any readily conceivedexample
of a body endowed with such a property. This difficulty was
removed in 1889 by Sir William Thomson (Lord Kelvin), who
designed mechanical models possessedof rotational elasticity.1
Suppose,for example,2that a structure is formed of spheres,each
sphere being in the centre of the tetrahedron formed by its four
nearestneighbours. Let each spherebe joined to thesefour neigh-
bours by rigid bars, which have spherical capsat their endsso as to
slide freely on the spheres. Such a structure would, for small
deformations, behave like an incompressible perfect fluid. Now
attach to each bar a pair of gyroscopically mounted flywheels,
rotating with equal and opposite angular velocities, and having
their axesin the line of the bar; a bar thus equipped will require
a couple to hold it at rest in any position inclined to its original
position, and the structure as a whole will possessthat kind of
quasi-elasticitywhich wasfirst imagined by MacCullagh.
This particular representation is not perfect, since a system of
forceswould be required to hold the model in equilibrium if it were
irrotationally distorted. Lord Kelvin subsequentlyinvented another
structure free from this defect.3
The work of Green proved a stimulus not only to MacCullagh
but to Cauchy, who now (1839) published yet a third theory of
reflection.4 This appears to have owed its origin to a remark of
Green's,5that the longitudinal wave might be avoided in either
of two ways-namely, by supposingits velocity to be indefinitely
great or indefinitely small. Green curtly dismissedthe latter alter-
native and adopted the former, on the ground that the equilibrium
of the medium would be unstableif its compressibilitywere negative
(as it must be if the velocity of longitudinal waves is to vanish).
Cauchy, without attempting to meet Green'sobjection, took up the
study of a medium whose elastic constants are connected by the
equation
k +
~n = 0,
so that the longitudinal vibrations have zero velocity; and showed

that if the aethereal vibrations are supposed to be executed at
right anglesto the plane of polarisation, and if the rigidity of the
1 W. Thomsonwas,saidHeaviside,, mostintenselymechanical,and could not accept
any etherunlesshecould makea modelof it' (Heaviside'sElectromagnetic Theory, iii, p. 479).

I ComptesRendus,cix (16Sept.1889),p. 453: Kelvin's Math. and Phys. Papers, iii, p. 466
I Proc. R. S. Edin. xvii (17 Mar. 1890),p. 127: Kelvin's Math. and Phys. Papers, iii,
p.468
• ComptesRendus, be (25 Nov. 1839),p. 676; and (2 Dec. 1839),p. 726
• Green'sMath. Papers, p. 246
aetheris assumedto be the samein all media,a ray which is reflected

will obey the sine law and tangent law of Fresnel. The boundary
conditions which he adopted in order to obtain this result were
the continuity of the displacemente and of its derivate oe/ OX, where
the axisof x is taken at right anglesto the interface.1 Theseare not
the true boundary conditions for general elasticsolids; but in the
particular casenow under discussion,where the rigidity is the same
in the two media, they yield the sameequations as the conditions
correctly given by Green.
The aether of Cauchy'sthird theory of reflection is well worthy
of somefurther study. It is generally known as the contractile or
labile 2 aether, the names being due to William Thomson (Lord
Kelvin, 1824-1907), who discussedit long afterwards.3 It may be
defined as an elastic medium of (negative)compressibilitysuchas
to makethe velocity of the longitudinal wavezero; this impliesthat
no work is required to be donein order to give the mediumany small
irrotational disturbance. An exampleis furnished by homogeneous
foam free from air and held from collapseby adhesionto a contain-
ing vessel.
Cauchy, as we have seen, did not attempt to refute Green's
objection that sucha medium would be unstable; but, asThomson
remarked, every possible infinitesimal motion of the medium is,
in the elementarydynamics of the subject,proved to be resolvable
into coexistentwave-motions. If, then, the velocity of propagation
for each of the two kinds of wave-motion is real, the equilibrium
must be stable,provided the medium either extendsthrough bound-
lessspaceor has a fixed containing vesselasits boundary.
When the rigidity of the luminiferous medium is supposedto
have the samevalue in all bodies,the conditions to be satisfiedat
an interface reduce to the continuity of the displacemente, of the
tangential componentsof curl e, and of the scalar (k+!n) dive
acrossthe interface.
Now we have seenthat when a transversewave is incident on
an interface, it givesrise in generalto reflectedand refractedwaves
of both the transverseand the longitudinal species. In the caseof
the contractile aether,for which the velocity of propagation of the
longitudinal waves is very small, the ordinary construction for
refracted waves shows that the directions of propagation of the
1 ComptesRendus, x (2 Mar. 1840),p. 347; xxvii (1848),p. 621 ; xxviii (1849),p. 25.
Mem. de l'Acad. xxii (1848),pp. 17,29

• Labile or neutral is a term usedof such equilibrium as that of a rigid body on a
perfectlysmoothhorizontal plane.
• Phil. Mag. xxvi (1888),p. 414
reflected and refracted longitudinal waves will be almost normal

to the interface. The longitudinal waveswill therefore contribute
only to the componentof displacementnormal to the interface, not
to the tangential components; in other words, the only tangential
componentsof displacementat the interface are those due to the
three transverse waves-the incident, reflected and refracted.
Moreover,the longitudinal wavesdo not contribute at all to curl e ;
and, therefore, in the contractile aether, the conditions that the
tangential componentsof e and of n curl e shall be continuous
acrossan interface are satisfiedby the distortional part of the dis-
turbance taken alone. The condition that the component of e
normal to the interface is to be continuous is not satisfiedby the
distortional part of the disturbancetaken alone,but is satisfiedwhen
the distortional and compressionalparts are taken together.
The energycarried away by the longitudinal wavesis infinitesi-
mal, as might be expected,since no work is required in order to
generate an irrotational displacement. Hence, with this aether,
the behaviourof the transversewavesat an interfacemay be specified
without consideringthe irrotational part of the disturbance at all,
by the conditionsthat the conservationof energyis to hold and that
the tangential componentsof e and of n curl e are to be continuous.
But if we identify thesetransversewaveswith light, assumingthat
the displacemente is at right anglesto the plane of polarisation of
the light, and assumingmoreoverthat the rigidity n is the samein
all media 1 (the differencesbetweenmedia dependingon differences
in the inertia p), we have exactly the assumptionsof Fresnel'stheory
of light; whence it follows that transversewaves in the labile
aether must obey in reflection the sine law and tangent law of
Fresnel.
The great advantageof the labile aether is that it overcomes
the difficulty about securing continuity of the normal component
of displacementat an interface between two media; the light-
wavestaken alone do not satisfythis condition of continuity; but
the total disturbance consisting of light waves and irrotational
disturbancetakentogetherdoessatisfyit; and this is ensuredwithout
allowing the irrotational disturbanceto carry off any of the energy.2
William Thomson (Lord Kelvin), who devoted much atten-
tion to the labile aether, was at one time led to doubt the validity
1 This condition is in any case necessary for stability, as was shown by R. T. Glaze-
brook; cf. Thomson, Phil. Mag. xxvi, (1888), p. 500.

s The labile-aether theory oflight may be compared with the electromagnetic theory,
by interpreting the displacement e as the electric force, and pe as the electric displacement.
(S~3) 147
of this explanation of light 1; for when investigating the radiation

of energy from a vibrating rigid globe embeddedin an infinite
elastic-solid aether, he found that in some casesthe irrotational
waveswould carry away a considerablepart of the energy if the
aetherwereof the labile type. This difficulty, however,wasremoved
by the observation2 that it is sufficientfor the fulfilment of Fresnel's
laws if the velocity of the irrotational wavesin one of the two media
is very small, without regard to the other medium. Following up
this idea, Thomsonassumedthat in spacevoid of ponderablematter
the aether is practically incompressibleby the forcesconcer!ledin
light waves, but that in the spaceoccupied by liquids and solids
it has a negative compressibility, so as to give zero velocity for
longitudinal aether waves in these bodies. This assumptionwas
basedon the conception that material atoms move through space
without displacing the aether; a conception which, as Thomson
remarked, contradicts the old scholasticaxiom that two different
portions of matter cannot simultaneouslyoccupy the samespace.3
He supposedthe aether to be attracted and repelled by the atoms,
and thereby to be condensedor rarefied.4
The year 1839, which saw the publication of MacCullagh's
dynamical theory of light and Cauchy'stheory of the labile aether,
was memorable also for the appearanceof a memoir by Green on
crystal-optics.5 This really contains two distinct theories, which
respectivelyresembleCauchy'sFirst and SecondTheories: in one
of them, the stresses in the undisturbed state of the aether are sup-
posedto vanish,and the vibrations of the aetheraresupposedto be
executedparallel to the plane of polarisation of the light; in the
other theory, the initial stressesare not supposedto vanish, and
the aether vibrations are at right anglesto the plane of polarisation.
The two investigations are generally known as Green's First and
SecondTheoriesof crystal-optics.
The foundationsof both theoriesare, however,the same. Green
first of all determined the potential energyof a strained crystalline
solid; this in the most general caseinvolves 27 constants,or 21 if
1 Baltirrwre Lectures(edn. 1904), p. 214 • ibid. p. 411
• Michell and Boscovich in the eighteenth century had taught the doctrine of the
mutual penetration of matter, i.e. that two substancesmay be in the same place at the
same time without excluding each other; cr. Priestley's History, i, p. 392.
• cf. Baltimore Lectures, pp. 413-14, 463 and Appendices A and E. 'I am afraid
nothing except the complete overthrow of his whole notion of how the functions of
the ether are produced will cure Sir W. T.' (FitzGerald to Heaviside, 4 Feb. 1889).
, I had a long correspondencewith Lord Kelvin, and his last letter saysthat he gives up
everything he ever wrote about the ether. I hope he is not still quite so down in the
mouth about it.' (FitzGerald to Heaviside, 8 June 1896).
• Trans. CambridgePhil. Soc. VII (1839), p. 121; Green's Math. Papers, p. 293
148
there is no initial stress.1 If, however, as is here assumed,the

medium possesses three planesof symmetry at right anglesto each
other, the number of constantsreducesto 12, or to 9 if there is no
initial stress; if e denote the displacement, the potential energy
per unit volume may be written
q, = G oex + H oey + I oe"

ox oy 0<.
+tG {e:xr + (~;r + (~;zr}+ lH {e;z)'+ (~r + (~fr}

+ II{ e;x)' + (~;r+ (~;"r}
+ ja (oex)1 ~ (~!1')1
ox + Jjb 0)'
+ {c (~!3)1
oz
+1' oeyoe" + ' oez oez + h' oexoey
0)' OZ g ox oz ox oy
+ 1.1(oeOZy + 0)'
2
oe,,)' + (oe,,+oex)1
1.
2g ox OZ + 1
2"
h (oex +
0)'
oe'!!.)'.
ox
The usual variational equation
HI p g;:x 8ex+ 0;;: 8ey + ~;:sOez}dx dy dz = - IH 8q, dx dy tk,

then yields the differential equationsof motion, namely:
02 ex o2ex+ (g + I) o'ex
p ot' = (a + G) o2ex+
ox2 (h + H) 0)'2 OZ'
+~(a
ox ox
oex+ h ~ey+ oe,,) + ~(a oex+ h' oey+ ' oe,,)
oy g oz ox ox 0)' g oz '
and two similar equations.
These differ from Cauchy's fundamental equations in having
greater generality; for Cauchy's medium was supposed to be
built up of point-centres of force attracting each other according
to somefunction of the distance; and, as we have seen,there are
limitations in this method of construction, which render it incom-
petent to representthe most generaltype of elastic solid. Cauchy's
equationsfor crystalline media are, in fact, exactly analogousto the
equations originally found by Navier for isotropic media, which
contain only one elasticconstantinstead of two.
1 For there are 21 terms in a homogeneous function of the second degree in six
variahles.
149
The number of constantsin the above equations still exceeds

the three which are required to specify the properties of a biaxal
crystal; and Green now proceedsto consider how the number
may be reduced. The condition which he imposesfor this purpose
is that for two of the three waveswhosefront is parallel to a given
plane, the vibration of the aetherealmoleculesshall be accurately
in the plane of the wave: in other words,that two of the threewaves
shall be purely distortional, the remaining one being consequently
a normal vibration. This condition gives five relations,1 which
may be written:
a=b=c=!I-';
f' = I-' - 2f; g' = I-' - 2g; h' = I-' - 2h;
where I-' denotesa new constant. 2
Thus the potential energyper unit volume may be written
cP = G oex
ox
+ H_oeOy" + I oez
0<;
+ iG {(~:xr+ (~:vr+ (~~r}+iH {(iiY +(~;r+(~zr}

x
+ !I{(~;r+ (~ir+ (~;zr}
+.121-' (oeox + oe"
OY
+ _oez)1
0<;
+ !f{(oe" 0<; + oy
oez)1 _ 4 Oy
oe{'oe z} +! {(oez + ~!-,:)I_ 4 o;z oex'.'
0<; g ox 0<; oz ox J
+ Ih {(oeOyx + exoe,,)1 _ 4 ~~ oe,,}_
ox Oy
At this point Green'stwo theoriesof crystal-opticsdiverge from
1 As Green showed, the hypothesis of transversality really involves the existence of
planes of symmetry, so that it alone is capable of giving 14 relations between the 21

constants; and 3 of the remaining 7 constants may be removed by change of axes,
leaving only four.
" It was afterwards shown by Barre de Saint-Venant (1797-1886), Journal de Math. vii
(1863), p. 399, that if the initial stressesbe supposedto vanish, the conditions which must
be satisfied among the remaining nine constants a, b,c,f, g, h,j' g', h', in order that the
wave-surface may be Fresnel's, are the following:
(3b -f) (3c -f) = (I +f')"
(3c - g) (3a - g) = (g + g')"
{ (3a - h) (3b - h) = (h + h')"
(3a - g) (3b - h) (3c - f) + (3a - h) (3b -f) (3c - g) = 2(1 +f') (g + g') (h + hI'
These reduce to Green's relations when the additional equation b = c is assumed.
Saint-Venant disputed the validity of Green's relations, asserting that they are com-
patible only with isotropy. On this controversy cf. R. T. Glazebrook, Brit. Assoc.Report
(1885), p. 171, and Karl Pearson in Todhunter & Pearson's History if Elasticity, ii, § 147.
150
eachother. According to the first theory, the initial stressesG, H, I

are zero,sothat
-I.
'f'
=].
ox + oey
(oe"
2fL Oy
+ oe:)1
oz
+ 1f{(oeoz + ~e:)1_ 4- oey oe:} +]. {(oe: + oe,,)I_ 4- ~e: oe,,}
y
"2" oy Oy oz g ox oZ 2 oZ ox
+ i h {(Be" + oey)! _ 4- Of-" oey}.
oy ox ox Oy
This expressioncontains the correct number of constants,namely,
four: three of them represent the optical constants of a biaxal
crystal, and one (namely, fL) representsthe square of the velocity
of propagation of longitudinal waves. It is found that the two
sheetsof the wave-surfacewhich correspondto the two distortional
wavesform a Fresnel'swave-surface; the third sheet,which corre-
spondsto the longitudinal wave,being an ellipsoid. The directions
of polarisation and the wave-velocitiesof the distortional wavesare
identical with thoseassignedby Fresnel,provided it is assumedthat
the direction of vibration of the aether particles is parallel to the
plane of polarisation; but this last assumptionis of courseincon-
sistentwith Green'stheory of reflection and refraction.
In his SecondTheory, Green, like Cauchy, used the condition
that for the waveswhosefronts are parallel to the co-ordinateplanes,
the wave-velocitydependsonly on the plane of polarisation, and not
on the direction of propagation. He thus obtained the equations
alreadyfound by Cauchy
G- f = H- g = I-h.
The wave-surfacein this case also is Fresnel's,provided it is
assumedthat the vibrations of the aether are executed at right
anglesto the plane of polarisation.
The principle which underlies the Second Theories of Green
and Cauchyis that the aether in a crystal resemblesan elasticsolid
which is unequally pressedor pulled in different directions by
the unmoved ponderable matter. This idea appealed strongly to
W. Thomson (Kelvin), who long afterwards developed it further,l
arriving at the following interesting result: let an incompressible
solid, isotropic when unstrained, be such that its potential energy
per unit volume is
1
!q-+-+--3, 1 1 )
(a f3 y
1 Prot. R. s. Edin. xv (1887), p. 21; Phil. Mag. xxv (1881l),p. 116; Baltirrwre Lectures,
pp. 228-59
where q denotesits modulus of rigidity when unstraincd, and

pi, yt denote the proportions in which lines parallel to the
of strain are altered; then if the solid be initially strained in a
defined by given values of (I, p, y by forcesapplied to its surfa
and if waves of distortion be superposedon this initial strain, t!
transmissionof thesewaveswill follow exactly the laws of Fresn
theory of crystal-optics,the wave-surfacebeing
--+ +--=0.
Xl yl Zl
fJ
~ rl
q
- 1 - rl
q
- 1 r.
q
rl - 1
There is some difficulty in picturing the manner in which

moleculesof ponderablematter act upon the aethersoasto prod
the initial strain required by this theory. Lord Kelvin utilised 1
suggestionto which we have already referred, namely, that
aether may pervade the atoms of matter so as to occupy sp
jointly with them, and that its interaction with them may con
in attractions and repulsions exercised throughout the regi
interior to the atoms. Theseforcesmay be supposedto be sola
in comparison with those called into play in free aether that t
resistanceto compressionmay be overcome,and the aether m
be (say) condensedin the central region of an isolated atom, a
rarefied in its outer parts. A crystal may be supposedto consist
a group of spherical atoms in which neighbouring spheresoverl
each other; in the central regionsof the spheresthe aether will
condensed,and within the lens-shapedregions of overlapping
will be still more rarefied than in the outer parts of a solitary ato
while in the interstices between the atoms its density will be
affected. In consequenceof theserarefactions and condensatio
the reaction of the aether on the atoms tends to draw inwards t
outermost atoms of the group, which, however,will be maintain
in position by repulsionsbetweenthe atoms themselves; and th
we can account for the pull which, accordingto the presenthypot
esis,is exertedon the aether by the ponderable moleculesof crysta
Analysis similar to that of Cauchy's and Green'sSecondTheol
of crystal-optics may be applied to explain the doubly refracti
property which is possessed by strained glass; but in this casetli
formulae derived are found to conflict with the resultsof experimd
The discordanceled Kelvin to doubt the truth of the whole theor
'Mter earnest and hopeful consideration of the stresstheory I
double refraction during fourteen years,' he said,2'I am unab
I Baltimore Uctures, p. 253 • Baltimore Lectures,p. 258
to seehow it can give the true explanation either of the double

refraction of natural crystals, or of double refraction induced in
isotropic solidsby the application of unequal pressuresin different
directions.'
It is impossibleto avoid noticing throughout all Kelvin's work
evidencesof the deep impression which was made upon him by
the writings of Green. The samemay be said of Kelvin's friend and
contemporaryStokes; and, indeed, it is no exaggerationto describe
Green as the real founder of that ' Cambridge school' of natural
philosophers,of which Kelvin, Stokes, Rayleigh, Clerk Maxwell,
Lamb, J. J. Thomson, Larmor and Love were the most illustrious
membersin the latter half of the nineteenth century. In order to
understandthe peculiar position occupied by Green, it is necessary
to recall somethingof the history of mathematical studies at Cam-
bridge.
The century which elapsedbetween the death of Newton and
the scientific activity of Green was the darkest in the history of the
University. It is true that Cavendishand Young were educatedat
Cambridge; but they, after taking undergraduatecourses,removed
to London. In the entire period the only natural philosopher of
distinction who lived and taught at Cambridge was Michell; and
for somereasonwhich at this distanceof time it is difficult to under-
stand fully, Michell's researchesseemto have attracted little or no
attention among his collegiate contemporariesand successors,who
silently acquiescedwhen his discoverieswere attributed to others,
and allowed his name to perish entirely from Cambridge tradition.
A few years before Green published his first paper a notable
revival of mathematical learning swept over the University; the
fluxional symbolism, which since the time of Newton had isolated
Cambridge from the continental schools,was abandoned in favour
of the differential notation, and the works of the great French
analysts were introduced and eagerly read. Green undoubtedly
received his own early inspiration from this source, chiefly from
Poisson; but in clearnessof physical insight and concisenessof
expositionhe far excelledhis masters; and the slight volume of his
collected papers has to this day a charm which is wanting in their
voluminous writings.
The neglect of his work during his lifetime (he died in 1841)
has often been remarked. It is, however, to be remembered that
in 1840the Lucasian professorat Cambridge was a man who never
wrote anything, that the Plurnian professor was Challis, whose
attention was engagedin his hydrodynarnical researches,that the
G I5jJUt :tU-Z- {&?Z.~

Lowndeanprofessorwas Dean of Ely and lived there, that Airy had

removed to Greenwich and was almost altogether occupied with
astronomy, that Stokes was an undergraduate and that William
Thomsonhad not yet enteredthe University. Green'sgreat memoirs
of 1837and 1839on the elastic-solidtheory were referred to in the
Trans. Camb.Phil. Soc.for 1842by M. O'Brien, who was evidently
well acquainted with them; but when Stokes was writing on
hydrodynamicsin 1843he did not know that Greenhad solvedthe
problem of the motion of an ellipsoid in a fluid; while Thomson
knew nothing of Green (except from a referencein a paper by
Murphy to the electrical memoir publishedat Nottingham in 1828),
until January 1845,when Hopkins gavehim a copy of the Notting-
ham essay. Before a year was over Thomson's enthusiasmcaused
everyoneto know about Green.
In spite of the advanceswhich were made in the great memoirs
of the year 1839,the fundamental question as to whether the aether
particlesvibrate parallel or at right anglesto the plane of polarisation
was still unanswered. More light was thrown on this problem ten
yearslater by Stokes'sinvestigation of diffraction.! Stokesshowed
that on almost any conceivablehypothesisregarding the aether, a
disturbance in which the vibrations are executed at right angles
to the plane of diffraction must be transmitted round the edgeof an
opaque body with lessdiminution of intensity than a disturbance
whosevibrations are executedparallel to that plane. It follows that
when light, of which the vibrations are oblique to the plane of
diffraction, is so transmitted, the plane of vibration will be more
nearly at right angles to the plane of diffraction in the diffracted
than in the incident light. Stokeshimself performed experiments
to test the matter, using a grating in order to obtain strong light
diffracted at a large angle,and found that when the planeof polarisa-
tion of the incident light was oblique to the plane of diffraction, the
plane of polarisation of the diffracted light was more nearly parallel
to the plane of diffraction. This result, which was afterwardscon-
firmed by L. Lorenz,2 appearedto confirm decisivelythe hypothesis
of Fresnel,that the vibrations of the aetherealparticlesare executed
at right anglesto the plane of polarisation.
Three yearsafterwards Stokesindicated 3 a secondline of proof
leading to the sameconclusion. It had long been known that the
blue light of the sky, which is due to the scatteringof the sun'sdirect
1 Trans. Camb. Phil. Soc. ix (1849), p. 1 ; Stokes's Math. and Phys. Papers, ii, p. 243
• Ann. d. Phys. cxi (1860), p. 315; Phil. Mag. xxi (1861), p. 321
• Phil. Trans. (1852), p. 463; Stokes's Math. and Phys. Papers, iii, p. 267. cf. the
footnote added on p. 361 of the Math. and Phys. Papers
154
THE AETHER AS AN ELASTIC SOLIn
rays by small particles or molecules in the atmosphere, is partly

polarised. The polarisation is most marked when the'light comes
from a part of the sky distant 90° from the sun, in which caseit
must have beenscatteredin a direction perpendicular to that of the
direct sunlight incident on the small particles; and the polarisation
is in the plane through the sun.
If, then, the axis of y be taken parallel to the light incident on a
small particle at the origin, and the scatteredlight be observedalong
the axis of x, this scatteredlight is found to be polarisedin the plane
xy. Consideringthe matter from the dynamical point of view, we
may supposethe material particle to possess so much inertia (com-
pared to the aether) that it is practically at rest. Its motion relative
to the aether, which is the causeof the disturbanceit createsin the
aether, will therefore be in the sameline as the incident aethereal
vibration, but in the opposite direction. The disturbance must be
transversal,and must therefore be zero in a polar direction and a
maximum in an equatorial direction, its amplitude being, in fact,
proportional to the sine of the polar distance. The polar line must,
by considerationsof symmetry, be the line of the incident vibration.
Thus we seethat none of the light scatteredin the x-direction can
comefrom that constituentof the incident light whichvibratesparallel
to the x axis; so the light observedin this direction must consistof
vibrations parallel to the z axis. But we have seenthat the plane of
polarisation of the scatteredlight is the plane of xy; and therefore
the vibration is at right anglesto the plane of polarisation.l
The phenomenaof diffraction and of polarisation by scattering
thus agreed in confirming the result arrived at in Fresnel's and
Green's theory of reflection. The chief difficulty in accepting it
arose in connection with the optics of crystals. As we have seen,
Greenand Cauchywereunable to reconcilethe hypothesisof aethereal
vibrations at right anglesto the plane of polarisation with the correct
formulae of crystal-optics, at any rate so long as the aether within
crystalswas supposedto be free from initial stress. The underlying
reasonfor this can be readily seen. In a crystal, where the elasticity
is different in different directions,the resistanceto distortion depends
solely on the orientation of the plane of distortion, which in the
caseof light is the plane through the directions of propagation and
vibration. Now it is known that for light propagated parallel to
one of the axesof elasticity of a crystal, the velocity of propagation
dependsonly on the plane of polarisation of the light, being the
1 The theory of polarisation by small particles was afterwards investigated by Lord
Ra)'leigh, Phil. Mag. xli (1871), p. 447.

155
samewhichever of the two axeslying in that plane is the direction

of propagation. Comparing these results, we see that the plane
of polarisation must be the plane of distortion, and therefore the
vibrations of the aether particles must be executedparallel to the
plane of polarisation.1
A way of escapefrom this conclusionsuggesteditself to Stokes,2
and later to Rankine 3 and Lord Rayleigh.' What if the aether in a
crystal, insteadof having its elasticity different in different directions,
were to have its rigidity invariable and its inertia different in different
directions? This would bring the theory of crystal-optics into
complete agreementwith Fresnel'sand Green'stheory of reflection,
in which the optical differences between media are attributed to
differences of inertia of the aether contained within them. The
only difficulty lies in conceiving how aelotropy of inertia can exist;
and all three writers overcamethis obstacleby pointing out that a
solid which is immersed in a fluid may have its effective inertia
different in different directions. For instance, a coin immersedin
water movesmuch more readily in its own plane than in the direction
at right anglesto this.
Supposethen that twice the kinetic energy per unit volume of
the aether within a crystal is representedby the expression
PI (oeof + PI (oer)
x) I (oez)1
8t + ps at '
I
and that the potential energy per unit volume has the samevalue
as in spacevoid of ordinary matter. The aether is assumedto be
incompressible,so that dive is zero: the potential energy per unit
volume is therefore
where n denotesas usual the rigidity.
1 In Fresnel'stheory of crystal-optics,in which the aether vibrations are at right
anglesto the plane of polarisation,the velocity of propagationdependsonly on the

directionof vibration, not on the planethrough this and the direction of transmission.
I Stokes, in a letter to Lord Rayleigh,insertedin his Memoir and Scientific Correspondence,
ii, p. 99, explainsthat the idea presenteditself to him while he waswriting the paperon
fluid motionwhich appearedin Trans. Camb. Phil. Soc. viii (1843),p. 105. He suggested
the wave-surface to which this theoryleadsin Brit. Assoc. Rep. (1862),p. 269.
• Phil. Mag. (4), i (1851),p. 441 • Phil. Mag. (4), xli (1871),p. 519
156
The variational equation of motion is
HI {PI ~;~z Sez + pz ~:e;Sey + P3 8;~,se,} dx dy dz

= - HI { S4> - pS (~:z+ ~;+ ~;,)}dx dy dz,
where p denotesan undetermined function of (x, y, z): the term
in p being introduced on account of the kinematical constraint
expressedby the equation
dive = O.
The equationsof motion which result from this variational equation
are 81e op
PI iii: = - ox + n"ijZez,
and two similar equations. It is evident that p resemblesa hydro-
static pressure.
Substituting in these equations the analytical expressionfor a
plane wave, we readily find that the velocity V of the wave is con-
nected with the direction cosines('\, 1-', v) of its normal by the
equation AI I I
---V-I + I-' VI + V V = O.
n - PI n - pz n - P3 I
When this is compared with Fresnel's relation between the

velocity and direction of a wave, it is seenthat the new formula
differs from his only in having the reciprocal of the velocity in
place of the velocity. About 1867 Stokescarried out a seriesof
experimentsin order to determine which of the two theories was
mostnearly conformableto the facts. He found the constructionof
Huygensand Fresnelto be decidedlythe more correct, the difference
betweenthe resultsof it and the rival construction being about a
hundred timesthe probable error of observation.1
The hypothesisthat in crystals the inertia dependson direction
seemedthereforeto be discreditedwhen the theory basedon it was
compared with the results of observation. But when, in 1888,
W. Thomson (Kelvin) revived Cauchy'stheory of the labile aether,
the question naturally arose as to whether that theory could be
extendedso as to account for the optical properties of crystals:
and it was shown by R. T. Glazebrook2 that the correct formulae
1 Proc. R. S. xx (1872),p. 443. After theseexperimentsStokesgaveit ashis opinion
(Phil. Mag. xli [1871], p. 521) that the true theory of crystal-opticswasyet to be found.
On the accuracyof Fresnel'sconstructioncr. Glazebrook,Phil. Trans.c1xxi(1879),p. 421,
and Hastings,Am. Journ. Sci. (3), xxxv (1887),p. 60.
• Phil. Mag. xxvi (1888),p. 521; xxviii (1889),p. 110
157
of crystal-opticsare obtained when the Cauchy-Thomsonhypothesis

of zero velocity for the longitudinal wave is combined with the
Stokes-Rankine-Rayleighhypothesisof aelotropic inertia.
For, on referenceto the formulaewhich havebeenalreadygiven,
it is obvious that the equation of motion of an aether having these
propertiesmust be
(P1ez, p,ey, Pac,) = - n curl curl e,

where e denotesthe displacement,n the rigidity, and (Ph Pt, />3)
the inertia; and this equationleadsby the usualanalysisto Fresnel's
wave-surface. The displacemente of the aetherealparticles is not,
however, accurately in the wave-front, as in Fresnel'stheory, but
is at right anglesto the direction of the ray, in the plane passing
through the ray and the wave-normal.l
Having now traced the progressof the elastic-solidtheory so
far as it is concerned with the propagation of light in ordinary
isotropic media and in crystals, we must consider the attempts
which weremadeabout this time to accountfor the optical properties
of a more peculiar classof substances.
It was found by Arago in 1811 2 that the state of polarisation
of a beamof light is altered when the beamis passedthrough a plate
of quartz along the optic axis. The phenomenonwasstudiedshortly
afterwardsby Biot, who showed3 that the alteration consistsin a
rotation of the plane of polarisation about the direction of propaga-
tion: the angle of rotation is proportional to the thickness of
the plate and inversely proportional to the square of the wave-
length.
In somespecimensof quartz the rotation is from left to right,
in others from right to left. This distinction was shown by Sir
John Herschel' (1792-1871) in 1820 to be associatedwith differ-
encesin the crystalline form of the specimens,the two typesbearing
the samerelation to each other as a right-handed and left-handed
helix respectively. Fresnel5 and W. Thomson (Kelvin) 6 proposed
the term helical to denote the property of rotating the plane of
1 This theory of crystal-optics may be assimilated to the electromagnetic theory by
interpreting the elastic displacement e as electric force, and the vector (Ple:e,p,el! Pae.) as
electric displacement.
• Mem. de l'lnstitut (1811), pt. i, p. liS sqq.
• Mem. de l'Institut (1812), pt. i, p. 218 ~qq.; Annates de Chim. ix (1818), p. 372 ;
x (11l19), p. 63
• Trans. Camb. Phil. Soc. i (1820), p. 43
• Mem. de l'Institut, vii (11127),p. 45, at p. 73. The memoir was presented to the
Academy in three parts, on 26 Nov. 1821,22 Jan. 1822 and 22 Apr. 1822 respectively.
• Baltimore Lectures, p. 31
polarisation,exhibited by suchbodiesasquartz; the lessappropriate

term natural rotatory polarisation is, however,generally used.l
Biot showed that many liquid organic bodies, e.g. turpentine
and sugar solutions, possessthe natural rotatory property. We
might be led to infer the presenceof a helical structure in the mol-
eculesof such substances; and this inference is supported by the
study of their chemical constitution, for they are invariably of the
, mirror-image' or ' enantiomorphous' type, in which one of the
atoms (generallycarbon) is asymmetricallylinked to other atoms.
The next advancein the subjectwasdue to Fresnel,2who showed
that in naturally active bodies the velocity of propagation of
circularly polarised light is different according as the polarisation
is right-handed or left-handed. From this property the rotation
of the plane of polarisation of a plane-polarised ray may be
immediately deduced; for the plane-polarisedray may be resolved
into two rays circularly polarised in opposite senses,and these
advancein phaseby different amounts in passingthrough a given
thicknessof the substance: at any stagethey may be recompounded
into a plane-polarisedray, the azimuth of whoseplane of polarisation
varieswith the length of path traversed.
It is readily seenfrom this that a ray of light incident on a
crystal of quartz will in general bifurcate into two refracted rays,
each of which will be elliptically polarised, i.e. will be capable of
resolutioninto two plane-polarisedcomponentswhich differ in phase
by a definite amount. The directions of theserefracted rays may
be determined by Huygens'construction, provided the wave-surface
is supposedto consist of a sphereand spheroid which do not touch.
The first attempt to frame a theory of naturally active bodies
was made by MacCullagh in 1836.3 Supposea plane wave of
light to be propagated within a crystal of quartz. Let (x, y, z)
denote the co-ordinatesof a vibrating molecule,when the axis of
x is taken at right anglesto the plane of the wave, and the axis of z
at right anglesto the axis of the crystal. Using Y and Z to denotethe
displacementsparallel to the axesof y and z respectivelyat any time
t, MacCullagh assumedthat the differential equationswhich deter-
mine Y and Z are o'Y o'Y 03Z
ott = c ' ox' flo ox'
l +
a'Z a'Z f;SY
7fi2 = CIS ax' - flo -ax3
1 The term rotatory may be applied with propriety to the property discoveredby
Faraday,which will be discussed later. • Annales fk Chim. xxviii (1825),p. 147

• Trans. Roy. Irish Acad. xvii (1837),p. 461; MacCullagh'sCall. Works, p. 63
159
where I-" denotesa constant on which the natural rotatory property

of the crystal depends. In order to avoid complicationsarisingfrom
the ordinary crystalline properties of quartz, we shall supposethat
the light is propagatedparallel to the optic axis,so that we can take
&1 equal to &2'
Assuming first that the beam lS circularly polarised, let it be
representedby
Y =A sin {:1T (lx - t) }, z= ± A cos {~ (lx - t) },

the ambiguous sign being determined according as the circular
polarisation is right-handed or left-handed.
Substituting in the abovedifferential equations,we have
21T
1 = &12l2 =f 1-".-.l3,
T
or
1 1T1-"
l=-±-.
&1 T&14
Since l/l denotesthe velocity of propagation, it is evident that the

reciprocals of the velocities of propagation of a right-handed and
left-handed beam differ by the quantity
21T1-"
-T&14 •
from which it is easily shown that the angle through which the
plane of polarisation of a plane-polarised beam rotates in unit
length of path is 21T11-"
T2C14'
If we neglect the variation of Cl with the period of the light,

this expressionsatisfiesBiot's law that the angle of rotation in unit
length of path is proportional to the inverse square of the wave-
length.
MacCullagh's investigation can be scarcelycalled a theory, for
it amounts only to a reduction of the phenomena to empirical,
though mathematical, laws; but it was on this foundation that
later workersbuilt the theory which is now accepted.1
1 The later developments of this theory wiII be discussedin a subsequent chapter;
but mention may here be made of an attempt which was made in 1856by Carl Neumann,
then a very young man, to provide a rational basisfor MacCuIlagh's equations. Neumann
showed that the equations may be derived from the hypothesis that the relative displace-
ment of one aethereal particle with respect to another acts on the latter according to the
same law as an element of an electric current acts on a magnetic pole. cr. the preface to
C. Neumann's Die magnetischeDrehung der PolarisationsebenedesLichtes (Halle, 1863).
160
The great investigatorswho developedthe theory of light after

the death of Fresnel devoted considerableattention to the optical
properties of metals. Their researchesin this direction must now
be reviewed.
The moststriking propertiesof metalsare the power of brilliantly
reflecting light at all angles of incidence, which is so well shown
by the mirrors of reflecting telescopes,and the opacity, which causes
a train of wavesto be extinguishedbefore it has proceededmany
wave-lengthsinto a metallic medium. That thesetwo attributes are
connectedappearsprobable from the fact that certain non-metallic
bodies-e.g. aniline dyes-which strongly absorbthe rays in certain
partsof thespectrum,reflectthoserays with almostmetallic brilliance.
A third quality in which metals differ from transparent bodies,and
which, as we shall see, is again closely related to the other two,
is in regard to the polarisation of the light reflected from them.
This wasfirst noticed by Malus; and in 1830 Sir David Brewster 1
showed that plane-polarised light incident on a metallic surface
remainspolarisedin the sameplane after reflection if its polarisation
is either parallel or perpendicular to the plane of reflection, but that
in other casesthe reflectedlight is polarisedelliptically.
It wasthis discoveryof Brewster'swhich suggestedto the mathe-
maticians a theory of metallic reflection. For, as we have seen,
elliptic polarisation is obtained when plane-polarisedlight is totally
reflected at the surface of a transparent body; and this analogy
betweenthe effectsof total reflection and metallic reflection led to
the surmise that the latter phenomenon might be treated in the
sameway asFresnelhad treated the former, namely, by introducing
imaginary quantities into the formulae of ordinary reflection. On
theseprinciplesmathematicalformulae weredevisedby MacCullagh2
and Cauchy.3
To explain their method, we shall supposethe incident light to
be polarisedin the plane of incidence. According to Fresnel'ssine
law, the amplitude of the light (polarisedin this way) reflectedfrom
a transparent body is to the amplitude of the incident light in
the ratio . (t. - r )
sm
J = sin (i + T)'
where i denotes the angle of incidence and T IS determined from
the equation
sin i = p.sin r.
1 Phil. Trans, 1830, p. 287
• Proc. Roy. Irish Acad. i (1836), p. 2; ii (1843), p. 376: Trans. Roy. Irish Acad. xviii
(1837), p. 71 : MaeCullagh's Coll. Works, pp. 58, 132,230
• ComptesRendus,vii (1838), p. 953; viii (1839), pp. 551,658,961 ; xxvi (1848), p. 86
161
MacCullagh and Cauchy assumedthat theseequations hold good

also for reflection at a metallic surface, provided the refractive
index I-' is replaced by a complex quantity
I-' = v (1 - I( -V-=-l) say,
where v and I( are to be regarded as two constantscharacteristic
of the metal. We have therefore
J- tan i-tan r _ (1-'2 - sin2 i)t - cosi
- tan i +
tan r - (1-'2 - +
sin2 i) ~ cosi
If then we write
v2 (1 - I( y-=-l)a - sin2 i = V2e2u.J - 1,
so that equations defining U and v are obtained by equating

separately the real and the imaginary parts of this equation, we
have --
Ueuv' -1 _ cosi
J = Veuv'-1 + cos/
and this may be written in the form
IJ',e Bv' - 1
where
IJI~_ V~+ co~i _~_2V cosv cosJ
, - V2 + cos2i + 2 U cosv cosi
{ " _ 2V cosi sin v
,
tan 0 - -va - COS 2't •
The quantities IJI and 0 are interpreted in the same way as in

Fresnel's theory of total reflection: that is, we take 1J12 to mean
the ratio of the intensities of the reflected and incident light, while
o measuresthe changeof phaseexperiencedby the light in reflection.
The caseoflight polarisedat right anglesto the plane of incidence
may be treated in the sameway.
When the incidence is perpendicular, V evidently reducesto
v (I + 1(2)t, and v reducesto - arctan-II(. For silver at perpendicular
incidence almost all the light is reflected, so 1J12 is nearly unity;
this requires cosv to be small, and I( to be very large. The extreme
casein which I( is indefinitely great but v indefinitely small, so that
the quasi-indexof refraction is a pure imaginary, is generallyknown
as the caseof ideal silver.
The physical significanceof the two constantsv and I( was more
or lessdistinctly indicated by Cauchy; in fact, as the difference
between metals and transparent bodies dependson the constant K.
162
It IS evident that K must in some way measurethe opacity of the

substance. This will be more clearly seenif we inquire how the
elastic-solidtheory oflight can be extendedsoasto provide a physical
basis for the formulae of MacCullagh and Cauchy.1 The sine
formula of Fresnel,which was the starting-point of our investigation
of metallic reflection, is a consequenceof Green's elastic-solid
theory; and the differencesbetweenGreen'sresults and thosewhich
we have derived arise solelyfrom the complex value which we have
assumedfor fL' We have therefore to modify Green'stheory in such
a way as to obtain a complex value for the index of refraction.
Take the plane of incidence as plane of xy, and the metallic
surface as plane of yz. If the light is polarised in the plane of
incidence, so that the light vector is parallel to the axis of z, the
incident light may be taken to be a function of the argument
ax + by + ct,
c = - (p),
n cos . (p)' . .
where
a
1, Cb = - n sm 1 ;
here i denotesthe angle of incidence, p the inertia of the aether,

and n its rigidity.
Let the reflectedlight be a function of the argument
a1x + by + ct,
where, in order to securecontinuity at the boundary, band c must
have the samevalues as before. Since Green's formulae are to be
still applicable, we must have
a1
b = cot T,
wheresin i = fL sin T, but fL hasnow a complex value. This equation

may be written in the form
all +b l = '---.
n
U2 pCl
Let the complex value of fL 2 be written
fL2 = f!} - A V - 1,
P
the real part being written pdp in order to exhibit the analogy
with Green'stheory of transparent media; then we have
all + b2 = P..!n cl - pel

n
A V- 1.
1 This wasdone by Lord Rayleigh, Phil. Mag. xliii (1872),p. 321
(828) 163
But an equation of this kind must (as in Green'stheory) represent

the condition to be satisfiedin order that the quantity
(a.x + 6y + cl)y-=1
e
may satisfythe differential equation of motion of the aether; from
which we seethat the equationof motion of the aetherin the metallic
medium is probably of the form
02eJ
Pl otB
+ pcA at
oe: (02e: + 02e:)
= n ox2 8y2·
This equation of motion differs from that of a Greenian elastic

solid by reasonof the occurrenceof the term in Oe:/ot. But this is
evidently a ' viscous' term, representingsomethinglike a frictional
dissipation of the energy of luminous vibrations; a dissipation
which, in fact, occasionsthe opacity of the metal. Thus the term
which expresses opacity in the equation of motion of the luminiferous
mediumappearsasthe origin of the peculiaritiesof metallicreflection.1
It is curious to notice how closely this accords with the idea of
Huygens, that metals are characterised by the presenceof soft
particleswhich damp the vibrations of light.
There is, however,one great difficulty attending this explanation
of metallic reflection, which wasfirst pointed out by Lord Rayleigh.2
We have seen that for ideal silver fL 2 is real and negative; and
thereforeA must be zero and Pl negative; that is to say,the inertia
of the luminiferous medium in the metal must be negative. This
seemsto destroy entirely the physical intelligibility of the theory
as applied to the caseof ideal silver.
The difficulty is a deep-seatedone, and was not overcomefor
many years. The direction in which the true solution lies will
suggestitself when we consider the resemblancewhich has already
been noticed between metals and those substanceswhich show
, surface colour' -e.g. the aniline dyes. In the caseof the latter
substances,the light which is so copiouslyreflected from them lies
within a restricted part of the spectrum; and it therefore seems
probable that the phenomenonis not to be attributed to the existence
of dissipative terms, but that it belongs rather to the same class
of effects as dispersion, and is to be referred to the same causes.
In fact, dispersionmeansthat the value of the refractive index of a
substancewith respectto any kind of light dependson the period
1 It is easilyseenthat the amplitudeis reducedby thefactor e - 2.,.1( whenlight travels
one wavelengthin the metal; is generallycalled the cofjficient of absorption.

I(
I loco cit.
of the light; and we have only to supposethat the physical causes

which operate in dispersion causethe refractive index to become
imaginary for certain kinds of light, in order to explain satisfactorily
both the surfacecoloursof the aniline dyesand the strong reflecting
powersof the metals.
Dispersion was the subject of several memoirs by the founders
of the elastic-solidtheory. So early as 1830Cauchy's attention was
directed 1 to the possibility of constructing a mathematical theory
of this phenomenonon the basisof Fresnel's' Hypothesis of Finite
Impacts' 2-i.e. the assumption that the radius of action of one
particle of the luminiferous medium on its neighbours is so large
asto be comparablewith the wave-lengthoflight. Cauchysupposed
the medium to be formed, as in Navier's theory of elastic solids, of
a systemof point-centresof force: the force between two of these
point-centres, m at (x, y, z), and p. at (x + t1x, Y + t1y, z + t1z),
may be denoted by mp.f(r), where r denotes the distance between
m and p.. When this medium is disturbed by light waves propa-
gated parallel to the z axis, the displacement being parallel to
the x axis, the equation of motion of m is evidently
osg t1x + t1g
011 = ~p.f(r + p) r +p ,
where g denotesthe displacementof m, (g + t1g) the displacement
of p., and (r + p) the new value of r. Substituting for p its value,
and retaining only terms of the first degree in t1g, this equation
becomes olg f(r) d {f(r) } (t1x)1
aes = 'l:.p. -r- t1g + 'l:.p. dr r -, t1g.
Now, by Taylor's theorem, sinceg dependsonly on z, we have
og 1 olg 1 osg s
+ +
I
t1g = oz t1z 2T OZI (t1z) 3f ozs (t1z) •••
Substituting, and remembering that summations which involve
odd powersof t1z must vanish when taken over all the point-centres
within the sphereof influence of m, we obtain an equation of the form
o~ o~ ~g ~g
aes = a OZI + fJ oz, +y ozG + . . .,
where a, f3, 'Y • . • denoteconstants.
Each successiveterm on the right-hand side of this equation
involves an additional factor (t1z) 2/>.. 2 as compared with the pre-
1 BuU. desSc. Math. xiv (1830), p. 9; • Sur la dispersionde la 1umiere,' Nouv. Exerci&,s
deMath., 1836 I cr. p. 123
ceding term, where ,\ denotesthe wave-length of the light; so if

the radii of influence of the point-centres were indefinitely small
in comparison with the wave-length of the light, the equation
would reduceto
which is the ordinary equation of wave-propagationin one dimen-

sion in non-dispersivemedia. But if the medium is socoarse-grained
that A is not large compared with the radii of influence, we must
retain the higher derivatesof g. Substituting
27Ti(z - '1t)1>'
g =e
in the differential equation with these higher derivates retained,

we have
. .,
which showsthat el, the velocity of the light in the medium, depends
on the wave-length'\; asit should do in order to explain dispersion.
Dispersionis, then, accordingto the view of Fresneland Cauchy,
a consequenceof the coarse-grainedness of the medium. Sincethe
luminiferous medium wasfound to be dispersiveonly within material
bodies,it seemednatural to supposethat in thesebodiesthe aether
is loaded by the moleculesof matter, and that dispersiondepends
essentiallyon the ratio of the wave-length to the distancebetween
adjacent material molecules. This theory, in one modification or
another, held its ground until forty years later it was overthrown
by the facts of anomalousdispersion.
The distinction between aether and ponderable matter was
more definitely drawn in memoirs which were published inde-
pendently in 1841-2 by F. E. Neumann 1 and Matthew O'Brien.2
Theseauthors supposedthe ponderable particles to remain sensibly
at rest while the aether surgesround them, and is acted on by
them with forceswhich are proportional to its displacement. Thus 3
the equation of motion of the aether becomes
ole
p ot2 = - (k + in) grad dive - n curl curl e - Ce,
where C denotesa constant on which the phenomenaof dispersion

1 Berlin Abhandlungen aus demJahre 1841, Zweiter Teil (Berlin, 1843), p. I
• Trans. Camb. Phil. Soc. vii (1842), p. 397 • O'Brien, loc. cit. §§ 15, 28
166
depend. For polarised plane waves propagated parallel to the

axis of x, this equation becomes
ae 2 ae 2
p at = n axi -
2 Ce ;
and if e is proportional to
2'1T '\1'-=-1 (I _ ~)
e TV,
where T denotesthe period and V the velocity of the light, we have

n _ C I
VI-P-4-rrzT,
an equation which expressesthe dependenceof the velocity on

the period.
The attempt to representthe properties of the aether by those
of an elasticsolid lost someof its interest after the rise of the electro-
magnetic theory of light. But in 1867, before the electromagnetic
hypothesishad attracted much attention, an elastic-solidtheory in
many respectspreferable to its predecessorswas presentedto the
French Academy 1 by Joseph Boussinesq(1842-1929). Until this
time, aswehaveseen,investigatorshad beendivided into two parties,
accordingasthey attributed the optical propertiesof different bodies
to variations in the inertia of the luminiferous medium, or to varia-
tionsin its elasticproperties. Boussinesq,taking up a position apart
from both theseschools,assumedthat the aether is exactly the same
in all material bodies as in interplanetary space,in regard both to
inertia and to rigidity, and that the optical propertiesof matter are
due to interaction between the aether and the material particles,
as had been imagined more or less by Neumann and O'Brien.
These material particles he supposedto be disseminatedin the
aether,in much the sameway asdust-particlesfloating in the air.
If e denotethe displacementat the point (x, y, z) in the aether,
and e' the displacementof the ponderable particles at the same
place,the equation of motion of the aether is
ae iJ2e'
+ in) + nY'le
2
P "at 2 - (k grad dive - PI otZ' (1)
where P and PI denote the densities of the aether and matter

1 Journal de Math. (2), xiii (1868), pp. 313, 425; cr. also Comptes&ndus, cxvii (1893),
pp. 80, 139, 193. Equations kindred to some of those of Boussinesq were afterwards
deduced by Karl Pearson, Proc. Lond. Math. Soc. xx (1889), p. 297, from the hypothesia
that the strain-energy involves the velocities.
167
respectively,and k and n denote as usual the elastic constantsof

the aether. This differs from the ordinary Cauchy-Greenequation
only in the presenceof the term Pl02e' /ot2, which representsthe
effect of the inertia of the matter. To this equation we must adjoin
another expressingthe connectionbetweenthe displacementsof the
matter and of the aether; if we assumethat theseare simply pro-
portional to eachother-say,
e' = Ae, (2)

where the constant A dependson the nature of the ponderable
body--our equation becomes
02e
tp + pIA) lJi2 = - (k + in) grad dive + nVle,
which is essentiallythe sameequation as is obtained in thoseolder
theories which supposethe inertia of the luminiferous medium to
vary from one medium to another. So far there would seemto be
nothing very new in Boussinesq'swork. But when we proceedto
consider crystal-optics, dispersion and rotatory polarisation, the
advantage of his method becomesevident: he retains equation
(I) as a formula universally true-at any rate for bodies at rest;
while equation (2) is varied to suit the circumstancesof the case.
Thus dispersion can be explained if, instead of equation (2), we
take the relation
e' = Ae - DV1e,
where D is a constant which measuresthe dispersivepower of the
substance; the rotation of the plane of polarisation of sugarsolutions
can be explained if we supposethat in thesebodiesequation (2) is
replaced by
e' = Ae +
B curl e,
where B is a constantwhich measuresthe rotatory power; and the
optical properties of crystals can be explained if we supposethat
for them equation (2) is to be replaced by the equations
When thesevalues for the componentsof e' are substitutedin

equation (1), we evidently obtain the sameformulae aswere derived
from the Stokes-Rankine-Rayleighhypothesis of inertia different
in different directions in a crystal; to which Boussinesq'stheory of
crystal-opticsis practically equivalent.
168
The optical properties of bodies in motion may be accounted

for by modifying equation (I), so that it takesthe form
ole
Pat?: = -(k + in) grad dive + n"\'2e-PI (0at+wrax+Wy~
0 & +w%oz
&)8 e,'
where w denotesthe velocity of the ponderable body. If the body

is an ordinary isotropic one, and if we consider light propagated
parallel to the axis of z, in a medium moving in that direction, the
light-vector being parallel to the axis of x, the equation reducesto
ole
P (}t:
(}2e
= n ()/ - PIA
( ()
8t + Waz0 )' e~;
substituting
e~ = f(z - Vt),
where V denotesthe velocity of propagation of light in the medium
estimated with referenceto the fixed aether, we obtain for V the
(P +n)\
value
PIA
PIA + P + PIA w.
The absolutevelocity of light is therefore increasedby the amount
PIAw/(p+PIA) owing to the motion of the medium; and this
may be written (I.t2 - I) w/1.t2, where I.t denotes the refractive
index; so that Boussinesq'stheory leads to the same formula as
had beengiven half a century previously by Fresnel.l
It is Boussinesq'smerit to have clearly assertedthat all space,
both within and without ponderable bodies, is occupied by one
identical aether, the sameeverywhereboth in inertia and elasticity;
and that all aetherealprocessesare to be representedby two kinds
of equations, of which one kind expressesthe invariable equations
of motion of the aether, while the other kind expressesthe inter-
action between aether and matter. Many years afterwards these
ideas were revived in connection with the electromagnetic theory,
in the modern forms of which they are indeed of fundamental
importance.
a d. p. 110
169
Chapter VI
FARADAY
TOWARDS the end of the year 1812,Davy receiveda letter in which

the writer, a bookbinder's journeyman named Michael Faraday,
expresseda desire to escapefrom trade, and obtain employment in
a scientific laboratory. With the letter was enclosed a neatly
written copy 1 of notes which the young man-he was twenty-one
years of age-had made of Davy's own public lectures. The great
chemist replied courteously, and arranged an interview; at which
he learnt that his correspondenthad educated himself by reading
the volumes which came into his hands for binding. 'There
were two,' Faraday wrote later, 'that especially helped me, the
Encyclopaedia Britannica, from which I gained my first notions of
electricity,and MrsJane Marcet's Conversationson Chemistry,which gave
me my foundation in that science.' Already, before his application
to Davy, he had performed a number of chemical experiments,and
had made for himself a voltaic pile, with which he had decomposed
severalcompound bodies.
At Davy's recommendation Faraday was in the following spring
appointed to a post in the laboratory of the Royal Institution, which
had been establishedat the close of the eighteenth century under
the auspicesof Count Rumford; and herehe remainedfor the whole
of his active life, first as assistant,then (on the death of Davy in
1829) as director of the laboratories, and from 1833onwardsas the
occupant of a chair of chemistry which wasfounded for his benefit.
For many years Faraday was directly under Davy's influence,
and was occupied chiefly in chemical investigations. But in 1821,
when the new field of inquiry opened by Oersted's discovery was
attracting attention, he wrote an Historical Sketchof Electro-Magnetism, 2
as a preparation for which he carefully repeated the experiments
describedby the writers he was reviewing; and this seemsto have
beenthe beginning of the researchesto which his fame is chiefly due.
The memoir which stands first in the published volumes of
Faraday's electrical work 3 was communicated to the Royal Society
on 24 November 1831. The investigation was inspired, as he tells
J Still preservedat the Royal Institution in London.
I Publishedin Annals of Philosophy, ii (1821),pp. 195,274; iii (1822),p. 107
• Experimental Researchesin Electricity, by Michael Faraday; 3 vols.
170
FARADAY
US, by the hope of discovering analogiesbetween the behaviour of

electricity as observedin motion in currents, and the behaviour
of electricity at rest on conductors. Static electricity was known
to possessthe power of ' induction' -i.e. of causing an opposite
electrical state on bodiesin its neighbourhood; was it not possible
that electric currents might show a similar property? The idea at
first was that if in any circuit a current were made to flow, any
adjacent circuit would be traversed by an induced current, which
would persist exactly as long as the inducing current. Faraday
found that this was not the case; a current was indeed induced,
but it lasted only for an instant, being in fact perceivedonly when
the primary current was started or stopped. It depended, as he
soon convinced himself, not on the mere existenceof the inducing
current, but on its variation.l
Faraday now set himself to determine the laws of induction of
currents, and for this purpose devised a new way of representing
the state of a magnetic field. Philosophershad been long accus-
tomed 2 to illustrate magnetic power by strewing iron filings on a
sheet of paper, and observing the curves in which they dispose
themselveswhen a magnet is brought underneath. These curves
suggestedto Faraday 3 the idea of lines of magneticforce 4; or curves
whosedirection at every point coincides with the direction of the
magnetic intensity at that point; the curves in which the iron
filings arrange themselveson the paper resemble these curves so
far as is possiblesubject to the condition of not leaving the plane
of the paper.
With theselines of magnetic force Faraday conceived all space
to be filled. Every line offorce is a closedcurve, which in somepart
of its coursepassesthrough the magnet to which it belongs.6 Hence
if any small closedcurve be taken in space,the lines of force which
intersectthis curve must form a tubular surfacereturning into itself;
1 This great discovery was only narrowly missed by Ampere; ef. S. P. Thompson,
Phil. Mag. xxxix (1895), p. 534.
2 The practice goes back at least as far as Niccolo Cabeo; indeed the curves traced
by Petrus Peregrinus on his globular lodestone (cf. p. 33) were projections of lines offorce.
Among eighteenth-century writers La Hire mentions the use of iron filings, Mem. de
['Acad, 1717. Faraday had referred to them in his electromagnetic paper of 1821,
Exp. Res. ii, p. 127.
• They were first defined in Exp. Res. § 114: 'By magnetic curves, I mean the lines
of magnetic forces,however modified by the juxtaposition of poles, which could be depicted
by iron filings; or those to which a very small magnetic needle would form a tangent.'
• The term liMS of force (liMoe virtutis) had been used by Aristotelian-scholastic
philosophers in connection with magnetism, e.g. in the Phi/Qsophia magMti&a (1629) of
Niceolo Cabeo.
• Exp.&s. iii, p. 405
such a surfaceis called a tube of force. From a tube of force we may

derive information not only regarding the direction of the magnetic
intensity, but alsoregarding its magnitude; for the product of this
magnitude 1 and the cross-sectionof any tube is constant along the
entire length of the tube.2 On the basis of this result, Faraday
conceivedthe idea of partitioning all spaceinto compartmentsby
tubes,each tube being such that this product has the samedefinite
value. For simplicity, eachof thesetubesmay be called a ' unit line
of force'; the strength of the field is then indicated by the separa-
tion or concentrationof the unit lines of force,3 so that the number
of them which intersect a unit area placed at right anglesto their
direction at any point measuresthe intensity of the magneticfield
at that point.
Faraday constantlythought in terms of lines of force. 'I cannot
refrain,' he wrote, in 1851,' , from again expressingmy conviction
of the truthfulness of the representation,which the idea of lines
of force affords in regard to magnetic action. All the points which
are experimentally establishedin regard to that action-i.e. all that
is not hypothetical-appear to be well and truly representedby it.' 6
Faraday found that a current is induced in a circuit either when
the strength of an adjacent current is altered, or when a magnet is
brought near to the circuit, or when the circuit itself is movedabout
in presenceof another current or a magnet. He sawfrom the first 6
that in all casesthe induction dependson the relative motion of the
circuit and the lines of magnetic force in its vicinity. The precise
nature of this dependencewas the subjectof long-continuedfurther
experiments. In 1832 he found 7 that the currents produced by
induction under the same circumstancesin different wires are
proportional to the conducting powersof the wires-a result which
1 Within the substance of magnetised bodies we must in this connection understand

the magnetic intensity to be that experienced in a crevice whose sides are perpendicular
to the lines of magnet isation ; in other words, we must take it to be what since Maxwell's
time has been called the magnetic induction.
I Exp. Res.§ 3073. This theorem was first proved by the French geometer Michel
Chasles in his memoir on the attraction of an ellipsoidal sheet, Journal del'&ou Polyt. xv
(1837), p. 266.
I ibid. § 3122.• The relative amount of force, or oflines of force, in a given spaceis
indicated by their concentration or separation-i.e. by their number in that space.'
• Exp. Res.§ 3174
I Some of Faraday's most distinguished contemporaries were far from sharing this
conviction. • I declare,' wrote Sir George Airy in 1855, • that I can hardly imagine
anyone who practically and numerically knows this agreement' between observation and
the results of calculation based on action at a distance, • to hesitate an instant in the choiclt
between this simple and precise action, on the one hand, and anything so vague and
varying as lines of force, on the other hand.' d. BenceJones's Life of Faraday, ii, p. 353.
• Exp. Res.,§ 116 • ibid. § 213
172
FARADAY
showedthat the induction consistsin the production of a definite

electromotiveforce, independent of the nature of the wire, and
dependentonly on the intersectionsof the wire and the magnetic
curves. This electromotiveforceis producedwhetherthe wire forms
a closedcircuit (sothat a current flows) or is open (sothat electric
tensionresults).
All that now remainedwas to inquire in what way the electro-
motive force dependson the relative motion of the wire and the
linesofforce. The answerto this inquiry is,in Faraday'sown words,!
that ' whether the wire movesdirectly or obliquely acrossthe lines
of force, in one direction or another, it sumsup the amount of the
forcesrepresentedby the lines it has crossed,'so that' the quantity
of electricity thrown into a current is directly as the number of
curvesintersected.'2 The induced electromotiveforce is, in fact,
simply proportional to the number of the unit lines of magnetic
forceintersectedby the wire per second.
This is the fundamental principle of the induction of currents.
Faradayis undoubtedlyentitled to the full honour of its discovery;
but for a right understandingof the progressof electrical theory
at this period, it is necessaryto rememberthat many yearselapsed
before all the conceptionsinvolved in Faraday's principle became
clear and familiar to his contemporaries; and that in the mean-
time the problem of formulating the laws of induced currentswas
approachedwith successfrom other points of view. There were
indeedmany obstaclesto the direct appropriation of Faraday'swork
by the mathematicalphysicistsof his own generation; not being
himself a mathematician,he was unable to addressthem in their
own language; and his favourite modeof representationby moving
lines of force repelled analystswho had been trained in the school
of Laplaceand Poisson. Moreover, the idea of electromotiveforce
itself, which had been applied to currents a few years previously
in Ohm's memoir, was, as we have seen,still involved in obscurity
and misapprehension.
A curious question which arose out of Faraday's theory was
whether a bar magnetwhich is rotated on its own axis carries its
linesof magneticforcein rotation with it. Faradayhimselfbelieved
that the lines of force do not rotate.3 On this view a revolving
magnetlike the earth is to be regardedas moving through its own
lines of force, so that it must becomecharged at the equator and
poleswith electricity of oppositesigns; and if a wire not partaking
1 &po Res., § 3082 • ibid. § 3115 • ibid. § 3090

173
in the earth's rotation were to have sliding contact with the earth
at a pole and at the equator, a current would steadilyflow through
it. Experimentsconfirmatory of theseviewswere madeby Faraday
himself 1; but they do not strictly prove his hypothesisthat the
lines of force remain at rest; for it is easily seen2 that, if they
were to rotate, that part of the electromotiveforce which would be
produced by their rotation would be derivable from a potential,
and so would produce no effect in closedcircuits such as Faraday
used.
Three years after the commencementof Faraday's researches
on induced currents he was led to an important extensionof them
by an observation which was communicated to him by another
worker. William Jenkin had noticed that an electric shockmay be
obtained with no more powerful sourceof electricity than a single
cell, provided the wire through which the current passesis long
and coiled; the shock being felt when contact is broken.3 As
Jenkin did not chooseto investigate the matter further, Faraday
took it up, and showed' that the powerful momentary current,
which was observedwhen the circuit was interrupted, was really
an induced current governedby the samelaws as all other induced
currents, but with this peculiarity, that the induced and inducing
currents now flowed in the samecircuit. In fact, the current in its
steadystate establishesin the surrounding region a magnetic field,
whoselinesof force are linked with the circuit; and the removalof
theselines of force when the circuit is broken originatesan induced
current, which greatly reinforcesthe primary current just beforeits
final extinction. To this phenomenonthe name of self-induction has
beengiven.
The circumstancesattending the discovery of self-induction
occasioneda comment from Faraday on the number of suggestions
which were continually being laid beforehim. He remarked that
although at different times a large number of authorshad presented
him with their ideas,this caseof Jenkin was the only one in which
any result had followed. 'The volunteers are seriousembarrass-
mentsgenerallyto the experiencedphilosopher.' 5
The discoveriesof Oersted,Ampere and Faradayhad shownthe
t Exp. Res., §§ 218,3109, etc.
2 cr. W. Weber, Ann. d. Phys.lii (1841), p. 353; S. Tolver Preston, Phil. Mag. xix (1885),
p. 131. In 1891S. T. Preston (Phil. Mag. xxxi, p. 100) designed a crucial experiment to
test the question, but it was not tried for want of a sufficiently delicate electrometer.
• A similar observation had been made by Joseph Henry (1797-1878), and published
in the Amer. Jour. Sci. xxii (1832), p. 408. The spark at the rupture of a spirally-wound
circuit had been often observed, e.g. by Pouillet and Nobili.
• Exp. Res., § 1048 • Bence Jones's Life of Faraday, ii, p. 45
174
FARADAY
closeconnectionof magnetic with electric science. But the connec-

tion of the different branchesof electricsciencewith eachother was
still not altogetherclear. Although Wollaston'sexperimentsof 1801
had in effect proved the identity in kind of the currentsderivedfrom
frictional and voltaic sources,the question was still regarded as
open thirty years afterwards,l no satisfactory explanation being
forthcoming of the fact that frictional electricity appeared to be a
surface-phenomenon,whereas voltaic electricity was conducted
within the interior substanceof bodies. To this question Faraday
now applied himself; and in 1833he succeeded2 in showing that
everyknown effectof electricity-physiological, magnetic,luminous,
calorific, chemical and mechanical-may be obtained indifferently
either with the electricity which is obtained by friction or with that
obtainedfrom a voltaic battery. Henceforth the identity of the two
wasbeyonddispute.
Somemisapprehension,however,hasexistedamonglater writers
as to the conclusionswhich may be drawn from this identification.
What Faraday proved is that the processwhich goeson in a wire
connectingthe terminalsof a voltaic cell is of the samenature asthe
processwhich for a short time goeson in a wire by which a condenser
is discharged. He did not prove, and did not professto have proved,
that this processconsistsin the actual movementof a quasi-substance,
electricity, from one plate of the condenserto the other, or of two
quasi-substances, the resinousand vitreous electricities, in opposite
directions. The processhad been pictured in this way by many of
his predecessors, notably by Volta; and it hassincebeensopictured
by mostof his successors;but from such assumptionsFaraday him-
self carefully abstained.
What is common to all theories,and is universally conceded,is
that the rate of increasein the total quantity of electrostaticcharge
within any volume-elementis equal to the excessof the influx over
the efflux of current from it.3 This statement may be represented
by the equation
atop + d'IV S = 0, (1)
where p denotes the volume-density of electrostatic charge, and

1 cf. John Davy, Phil. Trans. (1832), p. 259; W. Ritchie, ibid., p. 279. Davy
suggestedthat the electrical power, , according to the analogy of the solar ray,' might be
, not a simple power, but a combination of powers, which may occur variously associated,
and produce all the varieties of electricity with which we are acquainted.'
• Exp. Res.,series iii
• The first satisfactory experimental proof of the conservation of electric charge was
given by Faraday in Phil. Mag. xxii (1843), p. 200; Exp. Res.,ii, p. 279.
175
s the current, at the place (x,y, z) at the time t. Volta's assumption

is really one way of interpreting this equation physically: it presents
itself when we compare equation (1) with the equation
~ + div (pv) = 0,
which is the equation of continuity for a fluid of density p and

velocity v. We may identify the two equations by supposings to
be of the samephysical nature as the product pv; and this is pre-
cisely what is done by thosewho accept Volta's assumption.
But other assumptionsmight be made which would equally
well furnish physical interpretations to equation (1). For instance,
if we supposep to be the convergenceof any vector of which s is the
time-flux, 1 equation (1) is satisfiedautomatically; we can picture
this vector as being of the nature;of a displacement. By such an
assumptionwe should avoid altogether the necessityfor regarding
the conduction current as an actual flow of electric charges,or for
speculating whether the drifting chargesare positive or negative;
and there would be no longer anything surprising in the production
of a null effect by the coalescenceof electric chargesof opposite
SIgnS.
Faraday himself wished to leave the matter open, and to avoid
any definite assumption.2 Perhapsthe best indication of his views
is afforded by a laboratory note 3 of date 1837:
, After much consideration of the manner in which the electric
forces are arranged in the various phenomena generally, I have
come to certain conclusionswhich I will endeavour to note down
without committing myselfto anyopinion asto the causeof electricity,
Le., as to the nature of the power. If electricity existindependently
of matter, then I think that the hypothesisof one fluid will not stand
against that of two fluids. There are, I think, evidently what I may
call two elementsof power, of equal force and acting toward each
other. But thesepowersmay be distinguishedonly by direction, and
may be no more separatethan the north and south forces in the
=.
1 In symbols,
divq = -p
8q
ot
where q denotes the vector in question.
• 'His principal aim,' said Helmholtz in the Faraday Lecture of 1881, • was to
express in his new conceptions only facts, with the least possible use of hypothetical
substancesand forces. This was really a progress in general scientific method, destined
to purify science from the last remains of metaphysics.'
• BenceJones's Life of Faraday, ii, p. 77
176
FARADAY
elementsof a magneticneedle. They may be the polar points of the

forcesoriginally placed in the particles of matter.'
It may be remarked that since the rise of the mathematical
theory of electrostatics,the controversy between the supporters of
the one-fluid and the two-fluid theories had become manifestly
barren. The analytical equations,in which interest wasnow largely
centred, could be interpreted equally well on either hypothesis;
and there seemedto be little prospect of discriminating between
them by any new experimental discovery. But a problem doesnot
lose its fascination becauseit appears insoluble. 'I said once to
Faraday,' wrote Stokesto his father-in-law in 1879,'as I sat beside
him at a British Association dinner, that I thought a great step
would be made when we should be able to say of electricity that
which we say of light, in saying that it consistsof undulations. He
said to me he thought we were a long way off that yet.' 1
For his next seriesof researches,2Faraday reverted to subjects
which had been among the first to attract him as an apprentice
attending Davy's lectures: the voltaic pile, and the relations of
electricity to chemistry.
It was at this time generally supposedthat the decomposition
of a solution, through which an electric current is passed,is due
primarily to attractive and repellent forcesexercisedon its molecules
by the metallic terminals at which the current entersand leavesthe
solution. Such forceshad been assumedboth in the hypothesisof
Grothussand Davy, and in the rival hypothesisof De La Rive 3 ;
the chief differencebetweenthesebeing that whereasGrothussand
Davy supposeda chain of decompositionsand recompositionsin the
liquid, De La Rive supposedthe moleculesadjacent to the terminals
to be the only onesdecomposed,and attributed to their fragments
the power of travelling through the liquid from one terminal to the
other.
To test this doctrine of the influence of terminals, Faraday
moisteneda piece of paper in a saline solution, and supported it
in the air on wax, so as to occupy part of the interval betweentwo
needle points which were connected with an electric machine.
When the machine was worked the current was conveyedbetween
the needlepoints by way of the moistenedpaper and the two air-
intervals on either side of it; and under thesecircumstancesit was
found that the salt underwent decomposition. Sincein this caseno
metallic terminals of any kind were in contact with the solution,
1 Stokes'sScient~ Co"esponluru:e,
i, p. 353 • Exp. Res.§ 450 (1833)
• cr. pp. 76, 77
177
it was evident that all hypotheseswhich attributed decomposition

to the action of the terminals were untenable.
The ground being thus cleared by the demolition of previous
theories, Faraday was at liberty to construct a theory of his own.
He retained one of the ideas of Grothuss' and Davy's doctrine,
namely that a chain of decompositionsand recombinations takes
place in the liquid; but thesemolecular processes he attributed not
to any action of the terminals, but to a power possessedby the
electric current itself, at all placesin its coursethrough the solution.
If as an example we consider neighbouring moleculesA, B, C, D,
... of the compound-say water, which was at that time believed
to be directly decomposedby the current-Faraday supposedthat
before the passageof the current the hydrogen of A would be in
close union with the oxygen of A, and also in a lesscloserelation
with the oxygenatomsof B, C, D, . . . : theselatter relationsbeing
conjectured to be the causeof the attraction of aggregationin solids
and fluids.! When an electric current is sent through the liquid,
the affinity of the hydrogen of A for the oxygenof B is strengthened
if A and B lie along the direction of the current; while the hydrogen
of A withdraws someof its bonds from the oxygen of A, with which
it is at the moment combined. So long asthe hydrogen and oxygen
of A remain in association,the state thus induced is merely one of
polarisation; but the compound molecule is unable to stand the
strain thus imposed on it, and the hydrogen and oxygen of A part
company from each other. Thus decompositionstake place, fol-
lowed by recombinations: with the result that after each exchange
an oxygen atom associatesitself with a partner nearer to the positive
terminal, while a hydrogen atom associateswith a partner nearer
to the negative terminal.
This theory explains why, in all ordinary cases,the evolved
substancesappear only at the terminals; for the terminals are the
limiting surfacesof the decomposing substance; and, except at
them, every particle finds other particles having a contrary tendency
with which it can combine. It also explains why, in numerous
cases,the atoms of the evolved substancesare not retained by the
terminals (an obvious difficulty in the way of all theories which
supposethe terminals to attract the atoms) : for the evolved sub-
stancesare expelledfrom the liquid, not drawn out by an attraction.
Many of the perplexities which had harassedthe older theories
were at once removed when the phenomena were regarded from
Faraday'spoint of view. Th us,mere mixtures(asopposedto chemical
1 &cp. Res.,§ 523
178
FARADAY
compounds)are not separatedinto their constituentsby the electric

current; although there would seem to be no reason why the
Grothuss-Davypolar attraction should not operate as well on ele-
mentscontainedin mixtures ason elementscontainedin compounds.
In the latter part of the sameyear (1833) Faraday took up the
subject again.1 It was at this time that he introduced the terms
which haveeversincebeengenerally usedto describethe phenomena
of electrochemicaldecomposition. To the terminals by which the
electric current passesinto or out of the decomposingbody he gave
the name electrodes. The electrode of high potential, at which
oxygen, chlorine, acids, etc., are evolved, he called the anode, and
the electrode of low potential, at which metals, alkalis and bases
areevolved,the cathode. Thosebodieswhich are decomposeddirectly
by the current he named electrolytes; the parts into which they are
decomposed, ions; the acid ions,which travel to the anode,he named
anions; and the metallic ions, which passto the cathode,cations.
Faraday now proceededto test the truth of a suppositionwhich
he had published rather more than a year previously,2and which
indeedhad apparently beensuspectedby Gay-Lussacand Thenard 3
so early as 1811; namely, that the rate at which an electrolyte is
decomposeddependssolely on the intensity of the electric current
passingthrough it, and not at all on the sizeof the electrodesor the
strength of the solution. Having establishedthe accuracy of this
law,( he found by a comparison of different electrolytes that the
massof any ion liberated by a given quantity of electricity is pro-
portional to its chemical equivalent, i.e. to the amount of it required
to combine with somestandard massof somestandard element. If
an element is n-valent, so that one of its atoms can hold in com-
bination n atoms of hydrogen, the chemical equivalent of this
elementmay be taken to be 1In of its atomic weight; and therefore
Faraday'sresult may be expressedby saying that an electric current
will liberate exactly one atom of the element in question in the
time which it would take to liberate n atoms of hydrogen.5
The quantitative law seemedto Faraday 6 to indicate that' the
atomsof matter are in someway endowedor associatedwith electrical
powers,to which they owe their most striking qualities, and amongst
1 Exp. Res., § 661 • ibid. § 377, Dec. 1832

I Recherchesphysico-chimiquesfaites sur la pile (Paris, 1811), p. 12
• Exp. Res., §§ 713-821
• In the modern units, 96,580 coulombs of electricity must pass round the circuit in
order to liberate of each ion a number of grams equal to the quotient of the atomic weight
by the valency.
• Exp. Res., § 852
(823) 179
them their mutual chemicalaffinity.' Looking at the factsof electro-

lytic decompositionfrom this point of view, he showedhow natural
it is to supposethat the electricity which passesthrough the electro-
lyte is the exact equivalent of that which is possessedby the atoms
separated at the electrodes; which implies that there is a certain
absolute quantity of the electric power associatedwith each atom of matter.
The claims of this splendid speculationhe advocatedwith con-
viction. 'The harmony,' he wrote,l 'which it introducesinto the
associated theories of definite proportions and electrochemical
affinity is very great. According to it, the equivalent weights of
bodies are simply those quantities of them which contain equal
quantities of electricity, or have naturally equal electric powers;
it being the ELECTRICITY which determines the equivalent number,
becauseit determines the combining force. Or, if we adopt the
atomic theory or phraseology,then the atoms of bodies which are
equivalent to each other in their ordinary chemical action, have
equal quantities of electricity naturally associatedwith them. But,'
he added, , I must confessI amjealous of the term atom: for though
it is very easyto talk of atoms, it is very difficult to form a clear
idea of their nature, especiallywhen compound bodies are under
consideration.'
These discoveries and ideas tended to confirm Faraday in
preferring, among the rival theoriesof the voltaic cell, that one to
which all his antecedentsand connectionspredisposedhim. The
controversybetweenthe supportersof Volta's contact hypothesison
the one hand, and the chemical hypothesisof Davy and 'Vollaston
on the other, had now beencarried on for a generationwithout any
very decisiveresult. In Germany and Italy the contact explanation
was generally accepted,under the influence of Christian Heinrich
Pfaff (1773-1852) of Kiel, and of Ohm, and, among the younger
men, of Gustav Theodor Fechner (1801-87) of Leipzig,2 and
StefanoMarianini (1790-1866) of Modena. Among French writers
De La Rive, of Geneva,was, as we have seen,active in support of
1 ibid. § 869
• Johann Christian Poggendorff (1796-1877) of Berlin, for long the editor of the
Annalen der Physik, leaned originally to the chemical side, but in 1838 became convinced
of the truth of the contact theory, which he afterwards actively defended. Moritz
Hermann Jacobi (1801-74) of Dorpat, is also to be mentioned among its advocates.
Faraday's first series of investigations on this subject were made in 1834 (Exp. Res.,
seriesviii). In 1836 De La Rive followed on the same side with his Recherchessur la Cause
de [,Electr. Voltaique. The views of Faraday and De La Rive were criticised by Pfaff,
Revision der Lehre vom Galuanismus (Kie1, 1837), and by Fechner, Ann. d. Phys. xlii (1837),
p. 481, and xliii (1838), p. 433; translated Phil. Mag. xiii (1838), pp. 205, 367. Faraday
returned to the question in 1840,Exp. Ru., seriesxvi and xvii.
180
FARADAY
the chemical hypothesis; and this side in the dispute had always
beenfavoured by the English philosophers.
There is no doubt that when two different metals are put in
contact, a differenceof potential is set up betweenthem without any
apparent chemical action; but while the contact party regarded
this as a direct manifestation of a ' contact force' distinct in kind
from all other known forcesof nature, the chemical party explained
it as a consequenceof chemical affinity or incipient chemical action
betweenthe metals and the surrounding air or moisture. There is
also no doubt that the continued activity of a voltaic cell is always
accompaniedby chemical unions or decompositions; but while the
chemical party assertedthat theseconstitute the efficient source of
the current, the contact party regarded them as secondaryactions,
and attributed the continual circulation of electricity to the per-
petual tendency of the electromotive force of contact to transfer
chargefrom one substanceto another.
One of the most active supportersof the chemical theory among
the English physicists immediately preceding Faraday was Peter
Mark Roget (1779-1869), to whom are due two of the strongest
argumentsin its favour. In the first place, carefully distinguishing
betweenthe quantity of electricity put into circulation by a cell and
the tension at which this electricity is furnished, he showedthat the
latter quantity dependson the' energy of the chemical action' 1_
a fact which, when taken together with Faraday's discoverythat the
quantity of electricity put into circulation dependson the amount
of chemicalsconsumed,placesthe origin of voltaic activity beyond
all question. Roget's principle was afterwards verified by Faraday 2
and by De La Rive 3; 'the electricity of the voltaic pile is pro-
portionate in its intensity to the intensity of the affinities concerned
in its production,' said the former in 1834; while De La Rive wrote
in 1836, 'The intensity of the currents developed in combinations
and in decompositionsis exactly proportional to the degreeof affinity
which subsistsbetween the atoms whosecombination or separation
hasgiven rise to thesecurrents.'
Not resting here, however, Roget brought up another argument
of far-reaching significance. 'If,' he wrote,' , there could exist a
1 C The absolute quantity of electricity which is thus developed, and made to circulate,
will depend upon a variety of circumstances, such as the extent of the surfaces in chemical
action, the facilities afforded to its transmission, etc. But its degree of intensity, or tension,
as it is often termed, will be regulated by other causes,and more especially by the energy
of the chemical action.' (Roget's Galvanism [1832], § 70.)
• Exp. Res., §§ 908,909,916,988, 1958 • A,males de c!limit!xi (1836), p. 38
• Roget's Galvanism (1832), § 113
181
r
power having the property ascribedto it by the contact] hypothesis,
namely, that of giving continual impulse to a fluid in one constant
direction, without being exhaustedby its own action, it would differ
essentiallyfrom all the other known powers in nature. All the
powers and sourcesof motion, with the operation of which we are
acquainted, when producing their peculiar effects,are expendedin
the sameproportion asthoseeffectsare produced; and hencearises
the impossibility of obtaining by their agency a perpetual effect;
or, in other words, a perpetual motion. But the electromotive
force ascribed by Volta to the metals when in contact is a force
which, aslong as a free courseis allowed to the electricity it setsin
motion, is never expended,and continuesto be exerted with undi-
minishedpower, in the production of a never-ceasingeffect. Against
the truth of sucha suppositionthe probabilities are all but infinite.'
This principle, which is little lessthan the doctrine of conserva-
tion of energy applied to a voltaic cell, was reassertedby Faraday.
The processimagined by the contact school 'would,' he wrote,
, indeed be a creation of power, like no other force in nature.' In all
known casesenergyis not generated,but only transformed. There
is no such thing in the world as ' a pure creation of force; a pro-
duction of power without a correspondingexhaustionof something
to supply it.' 1
As time went on, each of the rival theoriesof the cell became
modified in the direction of the other. The contact party admitted
the importance of the surfacesat which the metals are in contact
with the liquid, where of course the chief chemical action takes
place; and the chemical party confessedtheir inability to explain
the stateof tensionwhich subsistsbeforethe circuit is closed,without
introducing hypothesesjust asuncertain asthat of contact force.
Faraday's own view on this point 2 was that a plate of amalga-
mated zinc, when placed in dilute sulphuric acid, , haspower sofar
to act, by its attraction for the oxygen of the particles in contact
with it, as to place the similar forces already active betweenthese
and the other particles of oxygen and the particles of hydrogen in
the water, in a peculiar state of tension or polarity, and probably
also at the sametime to throw thoseof its own particles which are
in contact with the water into a similar but opposedstate. Whilst
this state is retained, no further change occurs: but when it is
relieved by completion of the circuit, in which case the forces
determined in oppositedirections, with respectto the zinc and the
electrolyte, are found exactly competent to neutralise each other,
1 Exp.&s., § 2071 (1840) • Exp.&s., § 949
182
FARADAY
then a series of decompositionsand recompositions takes place

amongstthe particles of oxygen and hydrogen which constitute the
water, betweenthe place of contact with the platina and the place
wherethe zinc is active: theseintervening particles being evidently
in close dependenceupon and relation to each other. The zinc
forms a direct compoundwith thoseparticles of oxygenwhich were,
previously, in divided relation to both it and the hydrogen: the
oxide is removedby the acid, and a freshsurfaceof zinc is presented
to the water, to renew and repeat the action.'
Theseideaswere developedfurther by the later adherentsof the
chemical theory, especiallyby Faraday's friend Christian Friedrich
Schonbein1 of Basle (1799-1868), the discovererof ozone. Schon-
bein made the hypothesismore definite by assumingthat when the
circuit is open, the moleculesof water adjacent to the zinc plate are
electrically polarised,the oxygensideof eachmoleculebeing turned
towards the zinc and being negatively charged,while the hydrogen
sideis turned away from the zinc and is positively charged. In the
third quarter of the nineteenth century, the general opinion was in
favour of somesuch conception as this. Helmholtz 2 attempted to
graspthe molecular processes more intimately by assumingthat the
different chemicalelementshave different attractive powers(exerted
only at small distances)for the vitreous and resinouselectricities;
thus potassiumand zinc have strong attractions for positive charges,
while oxygen, chlorine and bromine have strong attractions for
negativeelectricity. This differs from Volta's original hypothesisin
little elsebut in assumingtwo electric fluids where Volta assumed
only one. It is evident that the contact difference of potential
between two metals may be at once explained by Helmholtz's
hypothesis,asit wasby Volta's; and the activity of the voltaic cell
may be referred to the sameprinciples; for the two ions of which
the liquid moleculesare composedwill alsopossess different attractive
powers for the electricities, and may be supposedto be united
respectivelywith vitreous and resinouscharges. Thus when two
metals are immersed in the liquid, the circuit being open, the
positive ions are attracted to the negative metal and the negative
ions to the positive metal, thereby causinga polarised arrangement
of the liquid moleculesnear the metals. When the circuit is closed,
the positively chargedsurfaceof the positive metal is dissolvedinto
the fluid; and as the atoms carry their charge with them, the
1 Ann. d. Phys. lxxviii (1849), p. 289, translated Archivesdessc.phys. xiii (1850), p. 192.
Faraday and Schonbein for many years carried on a correspondence, which has been
edited by G. W. A. Kahlbaum and F. V. Darbishire (London, Williams and Norgate).
I In his celebrated memoir of 1847 on the Conservation of Energy
183
positive charge on the immersed surface of this metal must be

perpetually renewedby a current flowing in the outer circuit.
It will be seenthat Helmholtz did not adhereto Davy's doctrine
of the electrical nature of chemical affinity quite assimply or closely
asFaraday, who preferred it in its most direct and uncompromising
form. 'All the factsshowus,' he wrote,!' that that power commonly
called chemical affinity can be communicatedto a distancethrough
the metals and certain forms of carbon; that the electric current is
only another form of the forcesof chemical affinity; that its power
is in proportion to the chemical affinities producing it; that when
it is deficient in force it may be helped by calling in chemical aid,
the want in the former being made up by an equivalent of the
latter; that, in other words, the forces termed chemical affinity and
electricity are one and the same.'
In the interval between Faraday's earlier and later papers on
the cell, someimportant resultson the samesubjectwere published
by Frederic Daniell (1790-1845), Professorof Chemistry in King's
College, London.2 Daniell showed that when a current is passed
through a solution of a salt in water, the ionswhich carry the current
are those derived from the salt, and not the oxygen and hydrogen
ions derived from the water; this follows since a current divides
itself betweendifferent mixed electrolytesaccordingto the difficulty
of decomposingeach, and it is known that pure water can be
electrolysedonly with great difficulty. Daniell further showedthat
the ions arising from, say, sodium sulphateare not Na20 and SOa,
but Na and S04; and that in such a caseas this, sulphuric acid is
formed at the anode and soda at the cathode by secondaryaction,
giving rise to the observed evolution of oxygen and hydrogen
respectivelyat theseterminals.
The researchesof Faraday on the decomposition of chemical
compounds placed between electrodes maintained at different
potentials led him in 1837 to reflect on the behaviour of such
substancesas oil of turpentine or sulphur, when placed in the same
situation. These bodies do not conduct electricity, and are not
decomposed; but if the metallic facesof a condenserare maintained
at a definite potential difference, and if the spacebetweenthem is
occupied by one of theseinsulating substances,it is found that the
charge on either face depends on the nature of the insulating
substance. If fo:" any particular insulator the chargehas a value E
times the value which it would have if the intervening body were
air, the number E may be regarded as a measureof the influence
1 Exp. &s., § 918 I Phil. Trans. (1839), p. 97
184
FARADAY
which the insulator exertson the propagation of electrostaticaction

through it; it was called by Faraday the specific inductive capacity of
the insulator.l
The discoveryof this property of insulating substancesor dielectrics
raised the question as to whether it could be harmonisedwith the
old ideas of electrostatic action. Consider,for example, the force
of attraction or repulsion between two small electrically charged
bodies. So long as they are in air, the force is proportional to the
inversesquareof the distance; but if the medium in which they are
immersedbe partly changed-e.g. if a globe of sulphur be inserted
in the intervening space--this law is no longer valid; the change
in the dielectric affectsthe distribution of electricintensity throughout
the entire field.
The problem could be satisfactorily solved only by forming a
physical conception of the action of dielectrics; and such a con-
ception Faraday now put forward.
The original idea had been promulgated long before by his
master Davy. Davy, it will be remembered,2in his explanation of
the voltaic pile, had supposedthat at first, beforechemicaldecomposi-
tions take place, the liquid plays a part analogousto that of the glass
in a Leydenjar, and that in this is involved an electric polarisation
of the liquid molecules.3 This hypothesiswas now developed by
Faraday. Referring first to his own work on electrolysis,he asserted4
that the behaviour of a dielectric is exactly the sameas that of an
electrolyte, up to the point at which the electrolyte breaks down
under the electric stress; a dielectric being, in fact, a body which is
capableof sustainingthe stresswithout suffering decomposition.
'For,' he argued,5 'let the electrolyte be water, a plate of ice
being coatedwith platina foil on its two surfaces,and thesecoatings
connectedwith any continued sourceof the two electrical powers,
the ice will charge like a Leyden arrangement,presentinga caseof
commoninduction, but no current will pass. If the ice be liquefied,
the induction will now fall to a certain degree,becausea current
can now pass; but its passingis dependentupon a peculiar molecular
arrangement of the particlesconsistentwith the transferof the elements
of the electrolyte in opposite directions. . . . As, therefore, in the
1 Exp. Res., § 1252(1837). Cavendishhad discovered specificinductive capacitylong

before,but his paperswerestill unpublished.
• cf. p. 75
I This is expressly statedin Davy's Elements of Chemical Philosophy (1812),div. i, ~ 7,
where he lays it down that an essential' property of non-conductors' is ' to receive
electricalpolarities.'
• Exp. Res., §§ 1164,1338,1343,1621 • Exp. Res., § 1164
18S
electrolytic action, induction appearedto be thefirst step, and decom-

position the second (the power of separating these stepsfrom each
other by giving the solid or fluid condition to the electrolyte being
in our hands); as the induction was the samein its nature as that
through air, glass, wax, etc., produced by any of the ordinary
means; and as the whole effect in the electrolyte appearedto be
an action of the particles thrown into a peculiar or polarised state,
I was glad to suspectthat common induction itself was in all cases
an action of contiguous particles, and that electrical action at a distance
(i.e. ordinary inductive action) never occurred except through the
influence of the intervening matter.'
Thus at the root of Faraday'sconceptionof electrostaticinduction
lay this idea that the whole of the insulating medium through which
the action takes place is in a state of polarisation similar to that
which precedesdecomposition in an electrolyte. 'Insulators,' he
wrote,1 ' may be said to be bodies whose particles can retain the
polarised state, whilst conductors are those whoseparticles cannot
be permanently polarised.'
The conception which he at this time entertained of the polari-
sation may be reconstructed from what he had already written
concerning electrolytes. He supposed2 that in the ordinary or un-
polarised condition of a body, the moleculesconsistof atomswhich
are bound to each other by the forces of chemical affinity, these
forces being really electrical in their nature; and that the same
forces are exerted, though to a less degree,between atoms which
belong to different molecules, thus producing the phenomenaof
cohesion. When an electric field is set up, a changetakesplace in
the distribution of theseforces; some are strengthenedand some
are weakened,the effectbeing symmetricalabout the direction of the
applied electric force.
Such a polarisedcondition acquired by a dielectric when placed
in an electric field presentsan evident analogy to the condition of
magnetic polarisation which is acquired by a mass of soft iron
when placed in a magnetic field; and it was thereforenatural that
in discussingthe matter Faraday should introduce lines of electric
force, similar to the lines of magnetic force which he had employed
so successfullyin his previous researches. A line of electric force
he defined to be a curve whosetangent at every point has the same
direction as the electric intensity.
1 Exp. Res.,§ 1338

• This must not be taken to be more than an idea which Faraday mentions as present
to his mind. He declined as yet to formulate a definite hypothesis.
186
FARADAY
The changes which take place in an electric field when the

dielectric is varied may be very simply described in terms of lines
of force. Thus if a mass of sulphur, or other substanceof high
specific inductive capacity, is introduced into the field, the effect
is as if the lines of force tend to crowd into it; as W. Thomson
(Kelvin) showed later, they are altered in the same way as the
lines of flow of heat, in a caseof steady conduction of heat, would
be altered by introducing a body of greater conducting power for
heat. By studying the figures of the lines of force in a great number
of individual cases,Faraday was led to notice that they always
disposethemselvesas if they were subject to a mutual repulsion,
or asif the tubesof force had an inherent tendency to dilate.l
It is interesting to interpret by aid of theseconceptionsthe law
of Priestley and Coulomb regarding the attraction between two
oppositelycharged spheres. In Faraday's view, the medium inter-
vening betweenthe spheresis the seat of a systemof stresses,which
may be representedby an attraction or tension along the lines of
electric force at every point, together with a mutual repulsion of
theselines, or pressurelaterally. Where a line of force ends on one
of the spheres,its tension is exercisedon the sphere; in this way,
every surface-elementof each sphere is pulled outwards. If the
spheres were entirely removed from each other's influence, the
state of stresswould be uniform round each sphere, and the pulls
on its surface-elementswould balance, giving no resultant force
on the sphere. But when the two spheresare brought into each
other's presence, the unit lines of force become somewhat more
crowded together on the sidesof the sphereswhich face than on the
remote sides,and thus the resultant pull on either sphere tends to
draw it toward the other. When the spheresare at distancesgreat
compared with their radii, the attraction is nearly proportional to
the inversesquareof the distance,which is Priestley'slaw.
In the following year (1838) Faraday amplified 2 his theory of
electrostaticinduction, by making further use of the analogy with
the induction of magnetism. Fourteen yearspreviously Poissonhad
imagined 3 an admirable model of the molecular processeswhich
accompany magnetisation; and this was now applied with very
little change by Faraday to the case of induction in dielectrics.
'The particles of an insulating dielectric,' he suggested,4'whilst
under induction may be compared to a seriesof small magnetic
needles, or, more correctly still, to a series of small insulated
1 Exp. Res.,§§ 1224, 1297 (1837) • Exp. Res.,series xiv
• cr. p. 65 • Exp. Res.,§ 1679
conductors. If the spaceround a charged globe were filled with a

mixture of an insulating dialectric, as oil of turpentine or air, and
small globular conductors,asshot,the latter being at a little distance
from each other so as to be insulated, then thesewould in their
condition and action exactly resemblewhat I consider to be the
condition and action of the particles of the insulating dielectric
itself. If the globe were charged, theselittle conductorswould all
be polar; if the globe were discharged, they would all return to
their normal state, to be polarised again upon the recharging of
the globe.'
That this explanation accountsfor the phenomenaof specific
inductive capacity may be seen by what follows, which is sub-
stantially a translation into electrostatical language of Poisson's
theory of induced magnetism.1
Let p denote volume-density of electric charge. For each of
Faraday's' small shot' the integral
III pdx 4Ydz,
integrated throughout the shot, will vanish, since the total charge
of the shot is zero; but if r denotethe vector (x,y, z), the integral
III p r dx 4Ydz
will not be zero, since it representsthe electric polarisation of the
shot; if there are N shot per unit volume, the quantity
P =N IIJprdx4Ydz
will representthe total polarisation per unit volume. If d denote
the electric force, and E the averagevalue of d, then P will be
proportional to E, say P = (€ _ 1) E.
By integration by parts, assumingall the quantltles concernedto

vary continuouslyand to vanish at infinity, we have
HI(p%~x + PY;y + Pz ~) 4>(x,y, z) dx 4Ydz = - HI4> divP dx4Y dz,

where 4> denotesan arbitrary function, and the volume-integrals
are taken throughout infinite space. This equation showsthat the
polar distribution of electric charge on the shot is equivalent to a
volume distribution throughout space,of density
p = - div P.
1 W. Thomson(Kelvin), Camb. and Dub. Math. Journal, i (Nov. 1845),p. 75; reprinted
in his Papers on Electrostatics and Magnetism, § 43 sqq.; F. O. Mossotti,Arch. dessc.phys. vi

(Geneva,1847),p. 193; Mem. della Soc. Ital. ModeM (2), xiv (1850),p. 49
188
FARADAY
Now the fundamental equation of electrostaticsmay in suitable

units be written,
div d = P;
and this giveson averaging
div E = PI + P,
where PI denotes the volume-density of free electric charge, i.e.
excluding that in the doublets; or
div (E + P) = PI,
or div (E E) = Pl.
This is thefundamentalequationof electrostatics,asmodified in order

to take into account the effect of the specificinductive capacity E.
The conception of action propagated step by step through a
medium by the influence of contiguousparticles had a firm hold
on Faraday'smind, and was applied by him in almost every part of
physics. 'It appearsto me possible,'he wrote in 1838,1 , and even
probable,that magneticaction may be communicated to a distance
by the action of the intervening particles, in a manner having a
relation to the way in which the inductive forcesof static electricity
are transferred to a distance; the intervening particles assuming
for the time more or lessof a peculiar condition, which (though with
a very imperfect idea) I have severaltimes expressedby the term
electrotonic state.' 2
The sameset of ideassufficedto explain electric currents. Con-
duction, Faraday suggested,3might be 'an action of contiguous
particles,dependenton the forcesdevelopedin electricalexcitement;
theseforcesbring the particles into a state of tension or polarity; 4
and being in this state the contiguous particles have a power or
capability of communicating these forces, one to the other, by
which they are lowered and dischargeoccurs.'
Mter working strenuouslyfor the ten years which followed the
discoveryof induced currents,Faraday found in 1841 that his health
was affected5; and for four years he rested. A secondperiod of
brilliant discoveriesbeganin 1845.
1 Exp. &s., § 1729
• This name had been devised in 1831 to express the state of matter subject to
magneto-electric induction; cf. Exp. Res.,§ 60.
• Exp. Res.iii, p. 513 • As in electrostatic induction in dielectrics.
• The description of his symptoms has given rise to the conjecture that he was suffering
inter alia from mercury poisoning: in electric connections he made great use of cups of
mercury, some of which would no doubt be spilt occasionally on the floor of the laboratory
Bnd ultimately vaporised.
18g
Many experimentshad been made at different times by various

investigators1 with the purposeof discoveringa connectionbetween
magnetism and light. These had generally taken the form
of attempts to magnetise bodies by exposure in particular
ways to particular kinds of radiation; and a successfulissue
had been more than once reported, only to be negatived on re-
examination.
The true path was first indicated by Sir John Herschel. Mter
his discoveryof the connectionbetweenthe outward form of quartz
crystals and their property of rotating the plane of polarisation of
light, Herschelremarkedthat a rectilinear electric current, deflecting
a needleto right and left all round it, possesses a helicoidal dissym-
metry similar to that displayed by the crystals. The analogy,
therefore,, led me,' he wrote,2 ' to concludethat a similar connection
exists, and must turn up somehowor other, between the electric
current and polarisedlight, and that the plane of polarisation would
be deflectedby magneto-electricity.'
Herschelneversubmittedhis prediction to the testof experiment;
this wasdone by Faraday, who sofar back as 18343 had transmitted
polarised light through an electrolytic solution during the passage
of the current, in the hope of observing a change of polarisation.
This early attempt failed; but in September, 1845, he varied
the experiment by placing a piece of heavy glass between the
poles of an excited electro-magnet; and found that the plane of
polarisation of a beam of light wasrotated when the beam travelled
through the glass parallel to the lines of force of the magnetic
field.'
In the following year a Swissphysicist,E. F. Wartmann, described
experimentsindicating that a phenomenonexactly similar to Fara-
day's magnetic rotation of the plane of polarisation of light can be
observedwhen rays of radiant heat are made to passthrough rock-
salt in a strong magnetic field; a complete confirmation of this
result was obtained a little later by F. de la Provostayeand P. Q.
Desains.5
In the year following Faraday's discovery, Airy 6 suggesteda
way of representing the effect analytically; as might have been
expected, this was by modifying the equations which had been
1 e.g. by Morichini, of Rome, in 1813,Quart. ]oum. Sci. xix, p. 338; by Samuel
Hunter Christie, of Cambridge, in 1825,Phil. Trans. (1826), p. 219; and by Mary
Somervillein the sameyear, Phil. Trans. (1826),p. 132
• Sir J. Herschelin BenceJones'sLife of Faraday,p. 205
• Exp. Res.,§ 951 • lb., § 2152
• Ann. chim. (3), xxvii (1849),p. 232 • Phil. Mag. xxviii (1846),p. 469
190
FARADAY
already introduced by MacCullagh for the caseof naturally active

bodies. In MacCullagh's equations
02Y 02Y osz
fjia = C1
1
oxl + fL ox S
{ 02Z OIZ osY

W = l
Cl oxl - fL oxS'
the terms oSZloxs and oSYloxs change sign with :c, so that the
rotation of the plane of polarisationis alwaysright-handed or always
left-handed with respectto the direction of the beam. This is the
casein naturally active bodies; but the rotation due to a magnetic
field is in the same absolute direction whichever way the light is
travelling, so that the derivations with respectto x must be of even
order. Airy proposedthe equations
02Y 02Y oZ
--ai2 = Cl
l
ox + at
2 fL
{ OIZ OIZ oY
--ai2 = C1
2
ox2 - fL at'
where fL denotes a constant, proportional to the strength of the
magnetic field which is used to produce the effect. He remarked,
however,that instead of taking fLoZlot and Ji-oYlot as the additional
terms, it would be possible to take fLoSZlotS and fLosY lotS, or
fLosZlox20t and fLosY lox20t, or any other derivates in which the
number of differentiations is odd with respect to t and even with
respectto x. It may, in fact, be shown by the method previously
applied to MacCullagh's formulae that, if the equationsare
02Y 02Y ()r+sZ
at2 = Cl axl
l
fLoxr ots +
{ 02Z 02Z ()r+sY
W = C12 ox2 - fL oxrf,Ts'
where (r + s) is an odd number, the angle through which the plane
of polarisation rotatesin unit length of path is a numerical multiple
of
where T denotesthe period of the light. Now it was shown by

Verdet 1 that the magneticrotation is approximately proportional to
the inversesquareof the wave-length; and hencewe must have
r+s=3;
1 Campus Rendus, Ivi (1863), p. 630
191
SO that the only equationscapableof correctly representingF'araday's

effect are either {02Y 02Y 03Z
ot2 = C1
2
OX2 + l-'ox2ot
02Z 2
02Z o3Y
ot2 = C1 ox2 - l-'ax2ot
or
The former pair arise, as will appear later, in Maxwell's theory

of rotatory polarisation; the latter pair, which were suggestedin
1868 by Boussinesq,lfollow from that physical theory of the
phenomenonwhich was more generally accepted by later workers
in the electromagnetictheory.2 There is practically no difference
between them.
Airy's work on the magnetic rotation of light was limited in the
sameway as MacCullagh's work on the rotatory power of quartz;
it furnished only an analytical representationof the effect, without
attempting to justify the equations. The earliest endeavour to
provide a physical theory seemsto have beenmade 3 in 1858,in the
inaugural dissertation of Carl Neumann, of Halle.4 Neumann
assumedthat every elementof an electric current exertsforce on the
particles of the aether; and in particular that this is true of the
molecular currents which constitute magnetisation,although in this
casethe force vanishesexceptwhen the aetherealparticle is already
in motion. If e denotethe displacementof the aetherealparticle m,
the force in question may be representedby the term
km [~;. KJ
where K denotes the imposed magnetic field, and k denotes a
magneto-opticconstant characteristicof the body. When this term
is introduced into the equationsof motion of the aether, they take
the form which had been suggestedby Airy; whenceNeumann's
1 Journal de Malh. xiii (1868), p. 430

I Y and Z being interpreted as components of electric force
I But see two letters from Stokes to W. Thomson of dates 12 Dec. and 13 Dec. 1848,
printed by Larmor in Proc. Camb. Phil. Soc. xxii (1924), p. 76

• Explicare lenlalur, quomodofial, ul luris planum po/ansalionis per vires el. vel mag. declinelur.
Halis Saxonum (1858). The results were republished in a tract Die magnelischeDrehung de,
PolarisalioTlsebenedes Lichtes (Halle, 1863).
FARADAY
hypothesisis seento lead to the incorrect conclusionthat the rotation

is independentof the wave-length.
The rotation of plane-polarised light depends, as Fresnel had
shown,1on a differencebetweenthe velocitiesof propagation of the
right-handed and left-handed circularly polarised wavesinto which
plane-polarised light may be resolved. In the case of magnetic
rotation, this difference was shown by Verdet to be proportional to
the componentof the magneticforce in the direction of propagation
of the light; and Cornu 2 showed further that the mean of the
velocitiesof the right-handed and left-handed wavesis equal to the
velocity of light in the medium when there is no magnetic field.
From thesedata, by Fresnel'sgeometrical method, the wave-surface I
in the medium may be obtained; it is found to consistof two spheres
(one relating to the right-handed and one to the left-handed light),
each identical with the spherical wave-surfaceof the unmagnetised
medium, displacedfrom eachother along the lines of magneticforce.3
The discovery of the connection between magnetism and light
gave interest to a short paper of a speculative character which
Faraday published 4 in 1846, under the title Thoughts of Ray-
Vibrations. In this it is possibleto trace the progressof Faraday's
thought towards somethinglike an electromagnetictheory of light.
Considering first the nature of ponderable matter, he suggests
that an ultimate atom may be nothing else than a field of force-
electric, magnetic and gravitational-surrounding a point-centre;
on this view, which is substantially that of Michell 5 and Boscovich,
an atom would have no definite size, but ought rather to be con-
ceived of as completely penetrable, and extending throughout all
space; and the molecule of a chemical compound would consist
not of atoms side by side, but of' spheresof power mutually pene-
trated, and the centreseven coinciding.' 6
All space being thus permeated by lines of force, Faraday
suggestedthat light and radiant heat might be transversevibrations
propagated along theselines of force. In this way he proposed to
, dismissthe aether,' or rather to replaceit by lines of force between
centres,the centrestogether with their lines of force constituting the
particles of material substances.
1 cf. p. 159 • ComptesRendus,xcii (1881),p. 1368

a Cornu, ComptesRendus,xcix (1884),p. 1045
• Phil. Mag. (3), xxviii (1846),p. 345; Exp. Res. iii, p. 447
• Michell wasled to this opinion by reflectingon Baxter'sdoctrineof the immateriality
of the soul; cf. one of the paperswritten by Sir William Herschel at Bath and not
publisheduntil 1911,in Dreyer'sedition of Herschel'sworks.
a cr. BenceJones'sLife of Faraday, ii, p. 178
193
If the existenceof a luminiferous aether were to be admitted,

Faraday suggestedthat it might be the vehicle of magnetic force;
, for,' he wrote in 1851,1, it is not at all unlikely that if there be an
aether, it should have other usesthan simply the conveyanceof
radiations.' This sentencemay be regarded as the origin of the
electromagnetictheory of light.
At the time when the Thoughts on Ray- Vibrations were published,
Faraday was evidently trying to comprehend everything in terms of
lines of force; his confidence in which had been recently justified
by another discovery. A few weeksafter the first observationof the
magnetic rotation of light, he noticed 2 that a bar of the heavy glass
which had been usedin this investigation, when suspendedbetween
the poles of an electromagnet, set itself across the line joining
the poles: thus behaving in the contrary way to a bar of an
ordinary magnetic substance,which would tend to set itself along
this line. A simpler manifestation of the effect was obtained
when a cube or sphere of the substancewas used; in such forms
it showed a disposition to move from the stronger to the weaker
places of the magnetic field. The pointing of the bar was then
seen to be merely the resultant of the tendencies of each of its
particles to move outwards into the positions of weakestmagnetic
action.
Many other bodiesbesidesheavy glasswere found to display the
sameproperty; in particular, bismuth.3 The name diamagnetic was
given to them.
, Theoretically,' remarked Faraday, , an explanation of the move-
ments of the diamagnetic bodiesmight be offered in the supposition
that magnetic induction caused in them a contrary state to that
which it produced in magnetic matter; i.e. that if a particle of each
kind of matter were placed in the magnetic field, both would become
magnetic, and each would have its axis parallel to the resultant of
magnetic force passing through it; but the particle of magnetic
matter would have its north and south poles opposite, or facing
toward the contrary polesof the inducing magnet, whereaswith the
diamagnetic particles the reverse would be the case; and hence
would result approximation in the one substance,recessionin the
other. Upon Ampere's theory, this view would be equivalent to the
supposition that, as currents are induced in iron and magnetics
parallel to thoseexisting in the inducing magnet or battery wire, so
1 Exp. Res., § 3075 I Phil. Trans. (1846), p. 21 ; Exp. Res., § 2253
I The repulsion of bismuth in the magnetic field had been previously observed by
A. Brugmans in 1778; Antonii Brugmans Ma.l!netismus,Leyden, 1778.
194
FARADAY
in bismuth, heavyglassand diamagneticbodiesthe currentsinduced

are in the contrary direction.' 1
This explanation becamegenerally known as the ' hypothesisof
diamagnetic polarity'; it representsdiamagnetism as similar to
ordinary induced magnetismin all respects,exceptthat the direction
of the induced polarity is reversed. It was accepted by other
investigators,notably by W. Weber, Plucker, Reich and Tyndall ;
but was afterwards displaced from the favour of its inventor by
another conception, more agreeableto his peculiar views on the
nature of the magnetic field. In this secondhypothesis,Faraday
supposedan ordinary magnetic or paramagnetic 2 body to be one
which offers a specially easypassageto lines of magnetic force, so
that they tend to crowd into it in preferenceto other bodies; while
he supposeda diamagneticbody to have a low degreeof conducting
power for the lines of force, sothat they tend to avoid it. 'If, then,'
he reasoned,3, a medium having a certain conducting power occupy
the magnetic field, and then a portion of another medium or
substancebe placed in the field having a greater conducting power,
the latter will tend to draw up towards the place of greatestforce,
displacing the former.' There is an electrostatic effect to which
this is quite analogous; a charged body attracts a body whose
specificinductive capacity is greater than that of the surrounding-
medium, and repelsa body whosespecificinductive capacity is less;
in either casethe tendencyis to afford the path of bestconductance
to the lines of force.4
For sometime the advocatesof the ' polarity' and' conduction'
theories of diamagnetism carried on a controversy which, indeed,
like the controversy between the adherents of the one-fluid and
two-fluid theories of electricity, persisted after it had been shown
that the rival hypotheseswere mathematically equivalent, and that
no experimentcould be suggestedwhich would distinguish between
them.
Meanwhile new properties of magnetisablebodies were being
discovered. In 1847Julius Plucker (1801-68), Professorof Natural
Philosophyin the University of Bonn, while repeating and extending
Faraday's magnetic experiments, observed5 that certain uniaxal
crystals, when placed between the two poles of a magnet,
1 Exp. Res., § 2429-30
• This term was introduced by Faraday, Exp. Res., § 2790.
• Exp. Res., § 2798
• The mathematical theory of the motion of a magnetisable body in a non-uniform
field of force was discussedby W. Thomson (Kelvin) in 1847.
• A,m. d. Phys. !xxii (1847), p. 315; Taylor's Scientific Memoirs. v. p. 353
(8:'3) 195
tend to set themselvesso that the optic axis has the equatorial
position. At this time Faraday wascontinuing his researches; and,
while investigating the diamagnetic properties of bismuth, was
frequently embarrassedby the occurrenceof anomalousresults. In
1848 he ascertainedthat thesewere in someway connectedwith the
crystalline form of the substance,and showed1 that when a crystal
of bismuth is placed in a field of uniform magnetic force (so that no
tendency to motion arisesfrom its diamagnetism) it setsitself so as
to have one of its crystalline axesdirected along the lines of force.
At first he supposedthis effect to be distinct from that which
had been discovered shortly before by PlUcker. 'The results,' he
wrote,2 'are altogether very different from those produced by
diamagnetic action. They are equally distinct from thosediscovered
and describedby PlUcker,in his beautiful researchesinto the relation
of the optic axis to magnetic action; for there the force is equatorial,
whereashere it is axial. So they appear to presentto us a new force,
or a new form of force, in the moleculesof matter, which, for con-
venience sake, I will conventionally designate by a new word, as
the magnecrystallic force.' Later in the same year, however, he
recognised3 that' the phaenomenadiscoveredby PlUckerand those
of which I have given an account have one common origin and
cause.'
The idea of the 'conduction' of lines of magnetic force by
different substances,by which Faraday had sosuccessfullyexplained
the phenomenaof diamagnetism,he now applied to the study of the
magnetic behaviour of crystals. 'If,' he wrote,4 ' the idea of con-
duction be applied to these magnecrystallic bodies, it would seem
to satisfy all that requires explanation in their special results. A
magnecrystallicsubstancewould then be one which in the crystallised
state could conduct onwards, or permit the exertion of the magnetic
force with more facility in one direction than another; and that
direction would be the magnecrystallic axis. Hence, when in the
magnetic field, the magnecrystallic axis would be urged into a
position coincident with the magnetic axis, by a force correspondent
to that difference, just as if two different bodies were taken, when
the one with the greater conducting power displacesthat which is
weaker.'
This hypothesisled Faraday to predict the existenceof another
type of magnecrystallic effect, as yet unobserved. 'If such a
view were correct,' he wrote,5 'it would appear to follow that a
1 Phil. Trans. (1849), p. 1 ; Exp. &s., § 2454 • Exp. &s., § 2469
• ibid. § 2605 • ibid. § 2837 • Exp. &s., § 2839
196
FARADAY
diamagnetic body like bismuth ought to be lessdiamagnetic when

its magnecrystallicaxis is parallel to the magnetic axis than when it
is perpendicular to it. In the two positions it should be equivalent
to two substanceshaving different conducting powersfor magnetism,
and therefore if submitted to the differential balance ought to
present differential phaenomena.' This expectation was realised
when the matter was subjectedto the test of experiment.l
The seriesof Faraday's Experimental Researchesin Electriciry end
in the year 1855. The closing period of his life was quietly spent
at Hampton Court, in a houseplaced at his disposalby the kindness
of the Queen; and here on 25 August 1867he passedaway.
Among experimental philosophers Faraday holds by universal
consentthe foremost place. The memoirs in which his discoveries
are enshrined will never ceaseto be read with admiration and
delight; and future generationswill preservewith an affection not
lessenduring the personalrecords and familiar letters, which recall
the memory of his humble and unselfishspirit.
I ibid. § 2841
Chapter VII
THE MATHEMATICAL ELECTRICIANS OF THE
MIDDLE OF THE NINETEENTH CENTURY
WHILE Faraday was engagedin discovering the laws of induced

currents in his own way, by useof the conception of lines of force,
his contemporary Franz Neumann was attacking the sameproblem
from a different point of view. Neumann preferredto take Ampere
as his model; and in 1845 published a memoir,l in which the
laws of induction of currents were deducedby the help of Ampere's
analysis.
Among the assumptionson which Neumann based his work
was a rule which had been formulated, not long after Faraday's
original discovery, by Emil Lenz,2 and which may be enunciated
as follows: when a conducting circuit is moved in a magnetic
field, the induced current flowsin sucha direction that the pondero-
motive forceson it tend to opposethe motion.
Let ds denote an element of the circuit which is in motion,
and let Cds denotethe component,taken in the direction of motion,
of the ponderomotive force exerted by the inducing current on ds,
when the latter is carrying unit current; so that the value of C is
known from Ampere's theory. Then Lenz's rule requires that the
product of C into the strength of the induced current should be
negative. Neumann assumedthat this is becauseit consistsof a
negativecoefficientmultiplying the squareof C; that is, he assumed
the induced electromotiveforce to be proportional to C. He further
assumedit to be proportional to the velocity v of the motion; and
thus obtainedfor the electromotiveforce inducedin ds the expression
- €vCds,
where € denotesa constant coefficient. By aid of this formula,
in the earlier part 3 of the memoir, he calculatedthe inducedcurrents
in various particular cases.
But having arrived at the formulae in this way, Neumann
1 Berlin Abhandlungen (1845), p. I; (1848) p. I; reprinted as no. 10 and no. 36 of
Ostwald's Klassiker; translated Journal de Math. xiii (1848), p. 113
• Ann. d. Phys. xxxi (1834), p. 483
• §§ 1--8. It may be remarked that Neumann, in making use of Ohm's law, was (like
everyone else at this time) Wlaware of the identity of electroscopic force with electrostatic
potential.
THE MIDDLE OF THE NINETEENTH CENTURY
noticed 1 a peculiarity in them which suggesteda totally different

method of treating the subject. In fact, on examiningthe expression
for the current induced in a circuit which is in motion in the field
due to a magnet,it appearedthat this induced current dependsonly
on the alteration causedby the motion in the value of a certain
function; and, moreover, that this function is no other than the
potential of the ponderomotiveforceswhich, according to Ampere's
theory, act betweenthe circuit, supposedtraversedby unit current,
and the magnet.
Accordingly, Neumann now proposedto reconstruct his theory
by taking this potential function as the foundation.
The nature of Neumann's potential, and its connection with
Faraday's theory, will be understoodfrom the following considera-
tions :
The potential energy of a magnetic molecule M in a field of
magnetic intensity B is (B.M); and therefore the potential energy
of a current i flowing in a circuit s in this field is
i H s
(B. dS),
where S denotesa diaphragm bounded by the circuit s; as is seen

at once on replacing the circuit by its equivalent magnetic shell S.
If the field B be produced by a current i' flowing in a circuit s', we
have, by the formula of Biot and Savart,
B = i' f Cds' • rl r3
r
= .,fr Ids'
I cur-.
r
Hence,the mutual potential energyof the two currents is
""'III (
u
r 5
r.
curl ds' dS) ,
which by Stokes'stransformation may be written in the form

.., II
u (ds • ds')•
r
• s'
This expressionrepresentsthe amount of mechanical work which

must be performed against the electrodynamic ponderomotive
1 §9
199
THE MATHEMATICAL ELECTRICIANS OF
forces, in order to separatethe two circuits to an infinite distance

apart, when the current strengthsare maintained unaltered.
The above potential function has been obtained by considering
the ponderomotive forces; but it can now be connected with
Faraday's theory of induction of currents. For by interpreting the
expression
II
(B. dS)
5
in termsoflines of force, we seethat the potential function represents

the product of i into the number of unit lines of magnetic force due
to s', which pass through the gap formed by the circuit s; and
since by Faraday's law the currents induced in s depend entirely
on the variation in the number of theselines, it is evident that the
potential function supplies all that is needed for the analytical
treatment of the induced currents. This was Neumann's discovery.
The electromotive force induced in a circuit s by the motion
of other circuits s', carrying currents i', is thus proportional to the
time-rate of variation of the potential
i' H (ds ~ds')
s s'
so that if we denoteby a the vector
Z
.,J r'
ds'
s'
which, of course, is a function of the pOSItIOnof the element ds

from which r is measured,then the electromotive force induced in
any circuit element ds by any alteration in the currents which give
rise to a is
The induction of currents is therefore governed by the vector a;

this, which is generally known as the vector-potential, has from
Neumann's time onwards played a great part in electrical theory.
It may be readily interpreted in terms of Faraday's conceptions;
for (a.ds) representsthe total number of unit lines of magnetic
force which have passedacross the line-element ds prior to the
instant t. The vector-potential may in fact be regarded as the
analytical measureof Faraday'selectrotonic state.l
While Neumann was endeavouring to comprehend the laws
of induced currentsin an extendedform of Ampere's theory, another
1 cr. p. 189
200
investigatorwas attempting a still more ambitious project: no less

than that of uniting electrodynamicsinto a coherent whole with
electrostatics.
Wilhelm Weber (1804-90) wasin the earlier part of his scientific
career a friend and colleague of Gauss at Gottingen. In 1837,
however, he becameinvolved in political trouble. The union of
Hanover with the British Empire, which had subsistedsince the
accessionof the Hanoverian dynasty to the British throne, was
in that year dissolved by the operation of the Salic law; the
PrincessVictoria succeededto the crown of England, and her
uncle Ernest-Augustus to that of Hanover. The new king
revoked the free constitution which the Hanoverians had for
sometime enjoyed; and Weber, who took a prominent part in
opposingthis action, was deprived of his professorship. From 1843
to 1849,when his principal theoretical researchesin electricity were
made,he occupieda chair in the University of Leipzig.
The theory of Weber was in its origin closely connectedwith
the work of another Leipzig professor,Gustav Theodor Fechner,
who in 18451 introduced certain assumptionsregarding the nature
of electric currents. Fechnersupposedevery current to consistin a
streamingof electric charges,the vitreous chargestravelling in one
direction, and the resinouscharges,equal to them in magnitude and
number, travelling in the opposite direction with equal velocity.
He further supposedthat like chargesattract each other when they
are moving parallel to the same direction, while unlike charges
attract when they are moving in opposite directions. On these
assumptionshe succeededin bringing Faraday's induction effects
into connectionwith Ampere'slaws of electrodynamics.
In 1846 Weber,2 adopting the same assumptionsas Fechner,
analysedthe phenomenain the following way.
The formula of Ampere for the ponderomotive force between
two elementsd5, ds' of currents i, i', may be written
F -- "I
tt
d d
5 5
I (~d'J.r
r ds ds'
_!rl ds
dr dr )
dsI •
Supposenow that A units of vitreous electricity are contained in

unit length of the wire 5, and are moving with velocity u; and that
an equalquantity of resinouselectricityismoving with velocityu in the
oppositedirection; so that i = 2AU.
Let A', u', denotethe correspondingquantities for the other current;
\
1 Ann. d. Phys. !xiv (1845),p. 337
1 Leipzig Abhandl., 1846,p. 209; Ann. d. Phys. !xxiii (1848),p. 193; English transla-
tion in Taylor's ScientijU;Memoirs, v (1852),p. 489
201
and let the suffix 1 be taken to refer to the action between the
positive chargesin the two wires, the suffix 2 to the action between
the positive charge in s and the negative charge in s', the suffix 3 to
the action between the negative charge in s and the positive charge
in s', and the suffix 4 to the action between the negative chargesin
the two wires. Then we have
dr) dr , dr
( dt 1 = U lIs + U dS"
and
d2r) d2r d2r d1r
( (jt2 1 = ul {lSi + 2uu' ds ds' + U'2 ds'l·
By aid of theseand the similar equationswith the suffixes2' 3, 4, the
equation for the ponderomotive force may be transformed into the
equation { 2 2
= "A"A' ds d.c' (r dt21'2)1 _. (r d1r)
dt2 1 (d r)
r dt2 3 + (dr dt2r) 4}
rl
F
- t
(dr) 1 (dr) 1 (,!!)I (t!!)1
dt + l dt +! dt
1 dt
.I 3
J.
2 4
But this is the equation which we should have obtained had we

set out from the following assumptions: that the ponderomotive
force between two current-elements is the resultant of the force
between the positive charge in ds and the positive charge in ds', of
the force between the positive charge in ds and the negative charge
in ds', etc.; and that any two electrified particles of chargese and e',
whosedistanceapart is r, repel eachother with a force of magnitude
~' { d2r _
rl r dt2 ! (~)2}
dt •
Two such charges would, of course, also exert on each other an
electrostatic repulsion, whose magnitude in these units would be
ee'c2/r2, where c denotesa constant 1 of the dimensionsof a velocity,
1 Two different systemsof units are used in electromagnetic theory. In the first, the
electrostatic system of units, a unit electric charge is defined to be such that two unit charges
at unit distance apart repel each other with unit ponderomotive force. In the other, the
electromagnetic system of units, a unit magnetic pole is defined to be such that two unit
poles at unit distance apart repel each other with unit ponderomotive force. In this
system a unit of electric current is defined to satisfy the condition that the mutual electro-
dynamic potential of two closed circuits carrying unit currents is equal to the mutual
potential of two magnetic shells of strength unity whose boundaries are the two circuits
respectively; and the unit of electric charge is then defined to be the quantity of electricity
conveyed in unit time by a unit current. This electromagnetic unit of electric charge is
c times the unit electrostatic charge.
In the investigation in the test above, electromagnetic units are used; in the electro-
static systemof units, the ponderomotive force of repulsion between two moving electrified
charges is ee'- {' 1
r'
!:.
+ c' dt
JIr _
2
1 (t!!.) I}
2&2\dl
202
whosevalue is approximately 3 X 1010 cm.jsec. So that on these

assumptionsthe total repellent force would be
ee;2 {I + is~:: -
c2
2~2 (~r}·
This expressionfor the force between two electric chargeswas
taken by Weber asthe basisof his theory. Weber'sis the first of the
electron theories-a name given to any theory which attributes the
phenomena of electrodynamics to the agency of moving electric
charges,the forceson which depend not only on the position of the
charges(asin electrostatics),but alsoon their velocities.
The latter feature of Weber's theory led its earliestcritics to deny
that his law of force could be reconciled with the principle of
conservation of energy. They were, however, mistaken on this
point, asmay be seenfrom the following considerations. The above
expressionfor the force betweentwo chargesmay be written in the
form
aU
- -Fr" dt
d ( 8U )
+ 8(~;),
where U denotesthe expression
ee' c' h + _!_(~)2}.
r l 2c2 8t
Consider now two material particles at distance r apart, whose
mechanical kinetic energy is T, and whose mechanical potential
energy is V, and which carry chargese and e'. The equations of
motion of theseparticles will be exactly the sameasthe equationsof
motion of a dynamical systemfor which the kinetic energy is
T _ e{
2r
(?!)2
8t '
and the potential energy is
,,
V + eec •
r
To such a system the principle of conservation of energy may be
applied: the equation of energyis, in fact,
ee' (8r)2
T +V - 12 8t + pe'
~c' = constant.
The first objection made to Weber's theory is thus disposedof i
but anotherand moreseriousonenow presentsitself. The occurrence
2°3
of the negative sign with the term _ ee' (C!!:) 2 implies that a charge
2r at
behavessomewhat as if its mass were negative, so that in certain
circumstances its velocity might increase indefinitely under the
action of a force opposedto the motion. This is one of the vulnerable
points of Weber's theory, and has been the object of much criticism.
In fact, 1 supposethat one charged particle of massI-' is free to move,
and that the other chargesare spreaduniformly over the surfaceof a
hollow spherical insulator in which the particle is enclosed. The
equation of conservation of energy is
1(I-' - ep)v2 +V = constant,
where e denotesthe chargeof the particle, v its velocity, V its potential
energy with respect to the mechanical forces which act on it, and p
denotesthe quantity a If
1\
- coss(v . r) dS,
r
where the integration is taken over the sphere,and where a denotes
the surface-density; p is independent of the position of the particle
I-' within the sphere. If now the electric charge on the sphereis so
great that ep is greater than 1-', then v2 and V must increase and
diminish together, which is evidently absurd.
Leaving this objection unanswered, we proceed to show how
\Veber's law of force betweenelectronsleads to the formulae for the
induction of currents.
The mutual energy of two moving chargesis
ee'e2 1 (8r)2)
-r -
{
1 - 2e2· ai J'
or ee'e2
r
[1 _ {(r. v') - (r. v) } • ]
2c2r2 ,
where v and Vi denote the velocities of the charges; so that the

mutual energy of two current-elements containing charges e, e'
respectively of each kind of electricity, is
ee'
2,-3 [ -{ (r • Vi) - (r • v) p + { (r • v') + (r • v) }I
+ {- (r. v') - (r. v) p - { (- r • v') + (r • v) pJ,
4ee' (r • Vi) (r • v)
or
r3
1 This examplewas given by Helmholtz, Journalfiir Math. lxxv (1873),p. 35; Phil.
Mag. xliv (1872),p. 530.
204
If ds, ds' denotethe lengthsof the elements,and i,l i' the currents in
them, we have ids = 2ev, i'ds' = 2e'v' ;
sothe mutual energyof two current-elementsis
i~ (r . ds') (r • ds).
r
The mutual energyof ids with all the other currents is therefore
i (ds • a),
wherea denotesa vector-potential
f·, t
(r • ds') r
3 •
" r
By reasoning similar to Neumann's, it may be shown that the
electromotiveforce induced in ds by any alteration in the rest of
the field is
- (ds.~~).
ot '
and thus a completetheory of induced currentsmay be constructed.
The necessityfor induced currents may be inferred by general
reasoning from the first principles of Weber's theory. When a
circuit s moves in the field due to currents, the velocity of the
vitreous chargesin s is, owing to the motion of s, not equal and
oppositeto that of the resinouscharges; this givesriseto a difference
in the forcesacting on the vitreous and resinouschargesin s; and
hence the chargesof opposite sign separatefrom each other and
movein oppositedirections.
The assumptionthat positive and negative chargesmove with
equal and oppositevelocitiesrelative to the matter of the conductor
is oneto which, for variousreasonswhich will appearlater, objection
may be taken; but it is an integral part of Weber's theory, and
cannotbe excisedfrom it. In fact, if this condition werenot satisfied,
and if the law of force were Weber's, electric currents would exert
forces on electrostatic charges at rest 1; as may be seen by the
following example. Let a current flow in a closedcircuit formed by
arcsof two concentriccirclesand the portions of the radii connecting
their extremities; then, if Weber's law were true, and if only one
kind of electricity were in motion, the current would evidently exert
an electrostaticforce on a chargeplaced at the centre of the circles.
1 This remarkwasfirst madeby Clausius,Joumalfiir Math. !xxxii (1877),p. 86; the
aimpleproof given aboveis due to Grassmann,Joumalfiir Math. !xxxiii (1877),p. 57.
205
It has been shown,l indeed, that the assumptionof opposite elec-

tricities moving with equal and opposite velocities in a circuit is
almostinevitable in any theory of the type of Weber's,solong asthe
mutual action of two chargesis ,assumedto depend only on their
relative (asopposedto their absolute)motion.
The law of Weber is not the only one of its kind; an alternative
to it was suggestedby Bernhard Riemann (1826-66), in a courseof
lectures which were delivered 2 at Gottingen in 1861, and which
were published after his death by K. HattendorfI'. Riemann pro-
posed as the electrokinetic energy of two electrons e(x, y, z) and
e'(x',y', z') the expression
_ t~' [(ax _ aX')1+

r l at at
(By _ ~l)2 + (azat _
at at
az')I} •
at '
or in the electrostaticsystemof units
this differs from the correspondingexpressiongiven by Weber only

in that the relative velocity of the two electronsis substituted in
place of the component of this velocity along the radius vector.
Eventually, as will be seenlater, the laws of Riemann and Weber
were both abandonedin favour of a third alternative.
At the time, however, Weber's discoverywas felt to be a great
advance; and indeedit had, perhaps,the greatestsharein awaken-
ing mathematical physiciststo a senseof the possibilitieslatent in
the theory of electricity. Beyondthis, its influencewasfelt in general
dynamics; for Weber's electrokinetic energy, which resembled
kinetic energyin somerespectsand potential energyin others,could
not be preciselyclassifiedunder either head; and its introduction,
by helping to break down the distinction which had hitherto sub-
1 H. Lorberg, JournalJur Math. !xxxiv (1878), p. 305

" Schwere,'Elektrit;itiit und Magnetismus, nach den Vorlesungenvon B. Riemann
(Hannover, 1875), p. 326. Another alternative to Weber's law had been discovered by
Gaussso far back as 1835, but was not published until after his death; cf. Gauss' Werke,
v, p. 616. It may be remarked that Riemann's form of the Lagrangian function might be
written, when the electrostatic term is included,
L = __ ~e_e_'e" _
r v{l - (vie)"}
where v denotes the relative velocity; or in the electrostatic system of units
L= e_e'__
, v{l - (vie}"}
206
sistedbetween the two parts of the kinetic potential, prepared the

way for the modern transformation theory of dynamics.1
Another subject whosedevelopment was stimulated by the work
of Weber was the theory of gravitation. That gravitation is propa-
gated by the action of a medium, and consequently is a process
requiring time for its accomplishment, had been an article of faith
with many generationsof physicists. Indeed, the dependenceof the
forceon the distancebetweenthe attracting bodiesseemedto suggest
this idea; for a propagation which is truly instantaneous would,
perhaps, be more naturally conceived to be effected by some kind
of rigid connection between the bodies, which would be morC'likely
to give a force independent of the mutual distance.
It is obvious that, if the simple law of Newton is abandoned,
there is a wide field of rival hypothesesfrom which to chooseits
successor. The first notable attempt to discussthe question was
made by Laplace.2 Laplace supposedgravity to be produced by
the impulsion on the attracted body of a 'gravific fluid,' which
flows with a definite velocity toward the centre of attraction-say,
the sun. If the attracted body or planet is in motion, the velocity
of the fluid relative to it will be compoundedof the absolutevelocity
of the fluid and the reversedvelocity of the planet, and the force of
gravity will act in the direction thus determined, its magnitude being
unaltered by the planet's motion. This amounts to supposingthat
gravity is subject to an aberrational effect similar to that observed
in the case of light. It is easily seen that the modification thus
introduced into Newton's law may be representedby an additional
perturbing force, directed along the tangent to the orbit in the
opposite senseto the motion, and proportional to the planet's
velocity and to the inverse square of the distance from the sun.
By considering the influence of this force on the secular equation
of the moon's motion, Laplace found that the velocity of the gravific
fluid must be at least a hundred million times greater than that
of light.
The assumptionsmade by Laplace are evidently in the highest
degree questionable; but the generation immediately succeeding,
overawed by his fame, seemsto have found no way of improving
on them. Under the influence of Weber's ideas, however, astrono-
mers began to think of modifying Newton's law by adding a term
involving the velocitiesof the bodies. Tisserand 3 in 1872discussed
, cr. Whittaker, Analytical Dynamics, chaps.ii, iii, xi
• Mlcanique Clleste, livre x, chap. vii, § 22
• ComptesJUndus,Ixxv (1872),p. 760. ef. alsoComptesJUndus,ex (1890),p. 313, and
HolzmUller,;::'eitschriftfiir Math. u. Phys. (1870),p. 69.
2°7
the motion of the planets round the sun on the suppositionthat the
law of gravitation is the same as Weber's law of electrodynamic
action, so that the force is
F= fi~ J 1 _ J.. (~)I+ .3.- r d r}, 2
r2 l hi dt h2 dt2
wheref denotesthe constantof gravitation, m the massof the planet,
fLthe massof the sun, r the distance of the planet from the sun,
and h the velocity of propagation of gravitation. The equationsof
motion may berigorouslyintegratedby the aid of elliptic functions 1 ;
but the simplestprocedureis to write
F= fmfL
r2
+F I,
and, regarding F1 as a perturbing function, to find the VarIatIOn

of the constants of elliptic motion. Tisserand showed that the
perturbations of all the elementsare zero or periodic, and quite
insensible, except that of the longitude of perihelion, which has
a secularpart. If h be assumedequal to the velocity of light, the
effect would be to rotate the major axis of the orbit of Mercury
in the direct sense14" in a century.
Now, asit happened,a discordancebetweentheory and observa-
tion was known to exist in regard to the motion of Mercury's peri-
helion; for Le Verrier had found that the attraction of the planets
might be expectedto turn the perihelion 527" in the direct sense
in a century, whereas the motion actually observedwas greater
than this by 38". It is evident, however, that only i of the excess
is explainedby Tisserand'sadoption of Weber'slaw; and it seemed
therefore that this suggestionwould prove as unprofitable as Le
Verrier's own hypothesisof an intra-mercurial planet. But it was
found later 2 that! of the excesscould be explained by substituting
Riemann's electrodynamic law for Weber's, and that a combina-
tion of the laws of Riemann and Weber would give exactly the
amount desired.3
After the publication of his memoir on the law of force between
electrons,Weber turned his attention to the questionof diamagnet-
ism, and developed Faraday's idea regarding the explanation of
diamagnetic phenomenaby the effectsof electric currents induced
1Thi~ had been done in an inaugural dissertation by Seegersat Gottingen in 1864.
• By Maurice Levy, ComptesRendus,ex (1890), p. 545
3 The consequencesof adopting the electrodynamic law of Clausius (for which see
later) were discussed by Oppenheim, '<:ur Frage nach der Fortpjlan(.ungsgeschwindigkeitder

Gravitation, Wien, 1895.
208
in the diamagnetic bodies.1 Weber remarked that if, with Ampere,

we assumethe existenceof molecular circuits in which there is no
ohmic resistance,so that currents can flow without dissipation of
energy,it is quite natural to supposethat currents would be induced
in thesemolecular circuits if they were situated in a varying magnetic
field; and he pointed out that such induced molecular currents
would confer upon the substance the properties characteristic of
diamagnetism.
The difficulty with this hypothesis is to avoid explaining too
much; for, if it be accepted, the inference seemsto be that all
bodies,without exception, should be diamagnetic. Weber escaped
from this conclusion by supposingthat in iron and other magnetic
substancesthere exist permanent molecular currents, which do not
owe their origin to induction, and which, under the influence of the
impressed magnetic force, set themselvesin definite orientations.
Sincea magneticfield tendsto give such a direction to a pre-existing
current that its course becomes opposed to that of the current
which would be induced by the increase of the magnetic force,
it follows that a substancestored with such pre-existing currents
would display the phenomena of paramagnetism. The bodies
ordinarily called paramagnetic are, according to this hypothesis,
thosebodiesin which the paramagnetismis strong enough to mask
the diamagnetism.
The radical distinction which Weber postulated between the
naturesof paramagnetismand diamagnetismseemedto be confirmed
by many facts which were discovered subsequently. Thus in 1895
P. Curie showed2 that the magnetic susceptibility per gramme-
molecule is connected with the temperature by laws which are
different for paramagnetic and diamagnetic bodies. For the former
it varies in inverse proportion to the absolute temperature, whereas
for diamagnetic bodiesit is independent of the temperature.
The conclusionswhich followed from the work of Faraday and
Weber were adverse to the hypothesis of magnetic fluids; for
according to that hypothesisthe induced polarity would be in the
same direction whether due to a change of orientation of pre-
existing molecular magnets, or to a fresh separation of magnetic
fluids in the molecules. 'Through the discovery of diamagnetism,'
wrote Weber 3 in 1852,, the hypothesisof electric molecular currents
1 Leipzig Berichte, i (1847), p. 346; Ann. d. Phys. !xxiii (1848), p. 241; translated
Taylor's Scientific Memoirs, v, p. 477; Abhandl. der K. Siichs. Ges. i (1852), p. 483; Ann. d.
Phys. Ixxxvii (1852), p. 145; trans. Tyndall and Francis' Scientific Memoirs, p. 163
• Amlales de Chimie (7), v (1845), p. 289
• Ann. d. Phys. !xxxvii (1852), p. 145; Tyndall and Francis's Sci. Mem., p. 163
2°9
in the interior of bodies is corroborated, and the hypothesis of

magnetic fluids in the interior of bodies is refuted.' The latter
hypothesis was, moreover, unable to account for the phenomena
shown by bodies which are strongly magnetic, like iron; for it is
found that when the magnetising force is gradually increasedto a
very large value, the magnetisation induced in such bodies does
not increasein proportion, but tends to a saturation value. This
effect cannot be explained on the assumptionsof Poisson,but is
easily deducible from those of Weber; for, according to Weber's
theory, the magnetising force merely orients existing magnets;
and when it has attained such a value that all of them are oriented
in the samedirection, there is nothing further to be done.
H the elementary magnets were supposedto be free to orient
themselveswithout encountering any resistance,it is evident that
a very small magnetisingforce would sufficeto turn them all parallel
to each other, and thus would produce immediately the greatest
possibleintensity of induced magnetism. To overcomethis difficulty
Weber assumedthat every displacement of a molecular circuit is
resistedby a couple. The form of the hypothesiswhich ultimately
gained acceptancewas that the orientation is resisted by couples
which arise from the mutual action of the molecular magnets
themselves. In the unmagnetisedcondition the molecules' arrange
themselvesso as to satisfy their mutual attraction by the shortest
path, and thus form a complete closed circuit of attraction,' as
D. E. Hughes wrote 1 in 1883; when an external magnetisingforce
is applied, these small circuits are broken up; and at any stage
of the processa molecular magnet is in equilibrium under the joint
influence of the external force and the forces due to the other
molecules.2
This hypothesis, due originally to Weber, was advocated by
Maxwell, 3 and was later developed by J. A. Ewing; 4 its
consequences may be illustrated by the following simple example5 :
Consider two magnetic molecules,each of magnetic moment m,
whosecentres are fixed at a distance c apart. When undisturbed,
they dispose themselvesin the position of stable equilibrium, in
which they point in the same direction along the line c. Now let
an increasingmagneticforce H be made to act on them in a direction
1 Proc. R. s. xxxv (1883), p. 178
• On the history of this, see W. Peddie, Nature, cxx (1927), p. 80
, Treatise on Elect. & Mag., § 443
• Phil. Mag. xxx (1890), p. 205; Magnetic Induction in Iron and other Metals (London,
1892) ; cf. also E. Warburg, Phil. Mag. xv (1883), p. 246
• E. G. Gallop, Messengerof Math. xxvii ,1897), p. 6
210
at right anglesto the line c. The magnetsturn towardsthe direction

ofH; andwhenH attainsthevalue 3m/c3, they becomeperpendicular
to the line c, after which they remain in this position, when H is
increasedfurther. Thus they display the phenomenaof induction
initially proportional to the magnetising force, and of saturation.
If the magnetisingforce H be supposedto act parallel to the line c,
in the direction in which the axesoriginally pointed, the magnets
will remain at rest. But if H acts in the opposite direction, the
equilibrium will be stableonly solong asH is lessthan m/c3; when
H increasesbeyond this limit, the equilibrium becomesunstable,
and the magnetsturn over so as to point in the direction of H ;
when H is gradually decreasedto zero, they remain in their new
positions,thus illustrating the phenomenonof residual magnetism.
By taking a large number of such pairs of magnetic molecules,
originally oriented in all directions, and at such distancesthat the
pairs do not sensibly influence each other, we may construct a
model whosebehaviour under the influence of an external magnetic
fieldwill closelyresemblethe actual behaviourofferromagneticbodies.
In order that the magnetsin the model may come to rest in
their new positions after reversal, it will be necessaryto suppose
that they experiencesomekind of dissipativeforce which dampsthe
oscillations; to this would correspondin actual magneticsubstances
the electric currents which would be set up in the neighbouring
masswhen the molecular magnetsare suddenlyreversed; in either
case, the sudden reversals are attended by a transformation of
magneticenergyinto heat.
In 1871Weber restatedhis theory of magnetism1 in terms of a
theory of electric particle-charges: an Amperean molecular current
wasnow supposedto be constituted by an electric chargemoving in
an orbit arounda fixed electricchargeof oppositesign,a picturewhich
reappearedforty yearslater in the electron orbits of the Rutherford-
Bohr atom. In the twentieth century, however, many theoretical
physicistswereat first inclined to doubt whether the magnetictheory
of Ampere and Weber could be successfullyexpressedin terms of the
new electron theory; thesedoubts were set at restby P. Langevin,2
who gave a completeformulation. In this the diamagnetic pheno-
menon is regarded in all casesas the sole initial effect of applying
an external magnetic field; paramagnetismis attributed to subse-
quent mutual actionsbetweenmolecules.
The transformation of energy from one form to another is a
1 Leip;;.ig Ahhandl. Math. Phys. x (1873),p. I; English translationin Phil. Mag. xliii
(1872),pp. I, 119 I Annales chim. phys. v (1905),p. 70
\823 211
'THE MATHEMATICAL ELECTRICIANS OF
subject which was first treated in a general fashion shortly before

the middle of the nineteenth century. It had long been known
that the energyof motion and the energyof position of a dynamical
systemare convertible into eachother, and that the amount of their
sum remains invariable when the system is self-contained. This
principal of conservation of dynamical energy had been extended
to optic~ by Fresnel,who had assumed1 that the energybrought to
to an int.erfaceby incident light is equal to the energycarried away
from the interface by the reflected and refracted beams. A similar
conception was involved in Roget's and Faraday's defence2 of
the chemical theory of the voltaic cell; they argued that the work
done by the current in the outer circuit must be provided at the
expenseof the chemical energy stored in the cell, and showedthat
the quantity of electricity sent round the circuit is proportional to
the quantity of chemicalsconsumed,while its tensionis proportional
to the strength of the chemical affinities concernedin the reaction.
This theory was extendedand completed by James PrescottJoule,
of Manchester, in 1841. Joule acceptedthe principle which had
been maintained, and virtually establishedby Rumford and Davy
in 1798-9, that heat is producible from mechanical work and
convertible into it ; and he saw that a closer investigation of this
principle was necessaryin order to achieve a complete theory of
the voltaic cell. He therefore measured3 the amount of heat
evolved in unit time in a metallic wire, through which a current
of known strength was passed; he found the amount to be pro-
portional to the resistanceof the wire multiplied by the squareof
the current strength; or (asfollows from Ohm's law) to the current
strength multiplied by the difference of electric tensions at the
extremitiesof the wire.
The quantity of energy yielded up as heat in the outer circuit
being thus known, it becamepossibleto considerthe transferenceof
energyin the circuit asa whole. 'When,' wroteJoule, , any voltaic
arrangement, whether simple or compound, passesa current of
electricity through any substance,whether an electrolyteor not, the
total voltaic heat which is generatedin any time is proportional to
the number of atomswhich are electrolysedin eachcell of the circuit,
multiplied by the virtual intensity of the battery; if a decomposing
cell be in the circuit, the virtual intensity of the battery is reduced
in proportion to its resistanceto electrolysation.' In the sameyear
he ' enhancedthe significanceof this by showingthat the quantities
1 cf. p. 182 • cf. p. 124 • Phil. Mag. xix (1841), p. 260; Joule's ScientificPapm, i, p. 60
• Phil. Mag. xx (1841), p. 98; cf. also Phil. Mag. xxii (1843), p. 204
212
of heat which are evolvedby the combustion of the equivalentsof

bodiesareproportional to the intensitiesof their affinitiesfor oxygen,
as measuredby the electromotive force of a battery required to
decompose the oxide electrolytically.
The theory of Roget and Faraday, thus perfected by Joule,
enablesus to trace quantitatively the transformationsof energy in
the voltaic cell and circuit. The primary sourceof energy is the
chemical reaction: in a Daniell cell, ZnIZnSO,ICuSO,ICu, for
instance,it is the substitutionof zinc for copperasthe partner of the
sulphion. The strength of the chemical affinities concernedis in
this casemeasuredby the differenceof the heatsof formation of zinc
sulphateand coppersulphate; and it is this which determinesthe
electromotiveforce of the cel1.l The amount of energy which is
changedfrom the chemicalto the electricalform in a given interval
of time is measuredby the product of the strength of the chemical
affinity into the quantity of chemicalsdecomposedin that time, or
(what is the samething) by the product of the electromotiveforce of
the cell into the quantity of electricity which is circulated. This
energymay be either dissipatedasheat in conformity to Joule's law,
or otherwiseutilised in the outer circuit.
Joule was the first to determine the mechanical equivalent of
heat by methodswhich fulfil the essentialcondition that no ultimate
changeof stateis producedin the matter operatedon. The notions
of Rumford and Davy regarding a mechanicalequivalent of heat
were howeverin the minds of many of his contemporaries; there
was a generaltendencyto return to the doctrine of Boyle, Hooke,
Newton and Euler, that heat is not a particular quasi-material
entity, but is an oscillatory motion of the smallestparts of bodies;
and the 'material' theory decayed until by about 1860 it had
completely disappeared. Among the writers who contributed to
this result may be mentioned C. F. Mohr of Coblenz (1806-79) 2
and Julius Robert Mayer (1814-78),3 who was a medical man in
Heilbronn. Mayer doesnot refer to Mohr's paper, although it was
published in the samejournal and contains almost all that was of
any value in his own contribution; and neither seemsto have had
any influence on contemporary thought, since Helmholtz doesnot
1 The heat of formation of a gramme-molecule of ZnSO, is greater than the heat of

formationof a gramme-moleculeof CuSO, by about 50,000calories; and with divalent
metals,46,000caloriesper gramme-moleculecorrespondsto an e.m.f.of onevolt; sothe
e.m.f.of a Daniell cell shouldbe 50{46volts, which is nearly the case.
S Annal. der Chimie xxiv (1837),p. 141; translatedPhil. Mag. (5) ii (1876),p. 110
• Annal. der Chimie, xlii (1842),p. 233; translatedPhil. Mag. xxiv (1862),p. 371. Maye"s
paperwasfirst declinedby the editorsof the Ann. d. Phys.
213
refer to them in the work now to be mentioned. The movement

culminated in a memoir published in 1847 by Hermann von
Helmholtz 1 (1821-94), under the title On the Conservation of Force.
It was read originally to the Physical Society of Berlin 2; but
though the younger physicists of the Society received it with
enthusiasm,the prejudices of the older generation 8 prevented its
acceptancefor the Annalen der Physik; and it waseventuallypublished
asa separatetreatise.'
In the memoir it was asserted5 that the conservationof energy
is a universal principle of nature; that the kinetic and potential 6
energy of dynamical systemsmay be convertedinto heat according
to definite quantitative laws, as taught by Rumford, Davy and
Joule; and that any of theseforms of energymay be convertedinto
the chemical, electrostatic,voltaic and magneticforms.
A paper" On the Mechanical Equivalent of Heat" wasread by Joule
to a meeting of the British Association7 at Oxford in 1847. Among
the audience was William Thomson, who had taken his degreeas
SecondWrangler at Cambridge two yearspreviously, and who now
sawthat in order to bring the entire subjectof heat and heat-engines
into a form which should be adapted to mathematical treatment
and should reveal scientific laws, it was necessaryfirst of all to
discover a method of measuring temperature absolutely, which
should be altogetherindependentof the propertiesof any particular
substance. He further saw that this could be founded on a paper
which had been published more than forty years previously by
NicolasLeonard Sadi Carnot (1796-1832), eldersonof the' organiser
of victory' of the French Revolution. Carnot's work, which
1 Helmholtz, like Mayer, was a medical man. The origin and function of animal heat
had long been a matter of discussion in their profession. I On 23 July 1847
3 Based largely on the dread of a revival of something like Hegel's natural philosophy
• Berlin, G. A. Reimer. English translation in Tyndall and Francis's ScientijieMtmOirs,

p. 114. The publisher, to Helmholtz' 'great surprise,' gave him an honorarium.
cr. Hermann von HeLmhoLtz,by Leo Koenigsberger; English translation by F. A. Welby.
• Helmholtz had been partly anticipated by W. R. Grove, in his lectures on the
Correlation of Physical Forces,which were delivered in 1843 and published in 1846. Grove,
after asserting that heat is ' purely dynamical' in its nature, and that the various' physical
forces' may be transformed into each other, remarked: 'The great problem which
remains to be solved, in regard to the correlation of physical forces, is the establishment
of their equivalent of power, or their measurable relation to a given standard.'
• These terms appeared later. Kinetic energywas first used in an article by W. Thomson
and P. G. Tait in Good Words (then edited by Charles Dickens), October 1862. The term
potentiaL energy was first used by W. J. M. Rankine in 1853 [Proc. GLasgowPhiL. Soc. iii
(1853), p. 276 = Rankine's Misc. Sci. Papers (1881), p. 203]. There was however little
or nothing new in this, for the thing signified was what Lazare Carnot had called force vive
virtueLle in 1803, and it had been a familiar concept in dynamics for generations past,
while the word potentiaL had been in common use since Green had introduced it in 1828.
• Brit. Ass. &p. (1847), pt. 2, p. 55
appearedat Paris in 1824under the title Rijlexions sur la Puissance

Motrice du Feu, wasa theoreticalstudy of the machineswhich produce
mechanical work from heat. In such a machine there must be
constituent parts which servethe samepurposesas the steam, the
boiler and the condenserof a steam-engine: that is to say,a working
substance (analogousto the steam),a source of heat (analogousto the
boiler) and a sink of heat (analogousto the condenser). Carnot
considereda processin which the working substanceand all parts of
the engineare in the samestate at the end of the processas at the
beginning-what may be called a cyclical process, or cycle. He showed
that in order to obtain the maximum efficiencyin the production of
mechanicalwork, it is necessarythat there should not be any direct
flow of heat betweenbodies whosetemperaturesdiffer by a finite
amount; when this condition is satisfied,and when moreover the
engineis supposedto work without friction, the cyclical processis
reversible,and the enginemay be said to be ' perfect.'
Carnot usedfor heat the term calorique, which had been intro-
duced by Lavoisier1; and he assumedthat the amount of caloric
taken from the sourcein a reversiblecycle is equal to the amount
given up to the sink,just asthe amount of water going over the top
of a waterfall is equal to the amount that falls on the bottom. He
showed that with this assumptionthe ..mechanicalwork obtained
dependsonly on the amount of caloric and the temperaturesof the
sourceand sink; a result which is quite correct when his ' caloric'
is identified with the quantity known later as' entropy.' 2
Thomson's new idea was thus expressedin his first paper 3 :
, The characteristicproperty of the scalewhich I now proposeis,
that all degreeshave the samevalue; that is, that a unit of heat
descendingfrom a body A at the temperatureTOof this scale,to a
body B at the temperature (T - 1t, would give out the same
mechanicaleffect, whatever be the number T. This may justly be
termedan absolutescale,sinceits characteristicis quite independent
of the physicalpropertiesof any specificsubstance.'
The' heat' referredto by Thomsonin this sentencewasCarnot's
, caloric,' and he wasnot asyet clear regarding its relation to Joule's
'heat.' The key to this problem was supplied by Clausius in a
communication to the Berlin Academy of February 1850'; and
Thomson's principle could now take the form 6: 'In a Carnot
1 Traiti I/lmentaire de chimie, 1789
• The term entropy was introduced by Clausius, Ann. d. Phys. cxxv (1865), p. 390
• Proc. Camb. Phil. Soc. i (5 June 1848), p. 66; Phil. Mag. xxxiii (Oct. 1848), p. 313;
Thomson's Math. and Phys. Papers, i, p. 100 • Ann. d. Phys. !xxix (1850), pp. 368, 500
• Trans. R. S. Edin. xx (Mar. 1851), p. 261 ; Thomson's Math. and Phys. Papers, i, p. 174
215
cycle of a reversible engine, the quantity of heat-energytaken from

the source is to the quantity of heat-energygiven up to the sink, in
the ratio of the absolute temperatures of the sourceand sink. The
difference of 6ese amounts of heat-energyis given up to the external
world in the form of mechanical work.'
The zeroof absolutetemperature could thereforebe definedby the
property that if the sink were at zero absolutetemperature, the whole
of the heat-energysupplied by the sourcecould be convertedinto me-
chanical work. For convenience,the marks 0° and 100°on the scale
weretaken to be definedasin theCentigradesystem: and it wasfound
that the zero of absolute temperatnre would then be at - 273°'1.
Thomson's principle as stated above may be generalisedinto the
statement that in any reversible processwhich leaves the working
substancein the samestate as it was originally,
where T is the absolutetemperature at which the elementof heat dQ

is absorbed by the working substance. Thus the integral I dQJT
has the samevalue for every processwhich leadsfrom a stateA to a
stateB of the working substance,and which is reversible. The value
of this integral is called the increase of entropy of the working substance
in passingfrom the state A to the state B.
The importance of the principle of energy in connection with the
voltaic cell wasempha~isedby HelmholtZ in his memoir of 1847,and
by W. Thomson in 18511 : the equationshave subsequentlyreceived
only one important modification, which is due to Helmholtz. 2
Helmholtz pointed out that the electrical energy furnished by a
voltaic cell need not be derived exclusively from the energy of the
chemical reactions: for the cell may also operate by abstracting
heat-energy from neighbouring bodies, and converting this into
electrical energy. The extent to which this takesplace is determined
by a law which was discovered in 1855 by Thomson.3 Thomson
showedthat if E denotesthe' available energy,' i.e. possibleoutput
of mechanical work, of a system maintained at the absolute tem-
perature T, then a fraction T dE
EdT
of this work is obtained, not at the expenseof the thermal or chemical
1 Phil. Mag. ii (1851), pp. 429, 551; Kelvin's Math. and Phys. Papers, i, pp. 472,490
• Berlin Sitzungsber. (1882),pp. 22, 825; (1883),p. 647
• Quart. Jour. Math. i (April 1855),p. 57; Kelvin's Math. and Phys. Papers, i, p. 297,
eqn. (7)
216
energy of the systemitself, but at the expenseof the thermal energy

of neighbouring bodies. Now in the case of the voltaic cell, the
principle of Roget, Faraday and Joule is expressedby the equation
E = A,
where E denotesthe available or electrical energy, which is measured
by the electromotive force of the cell, and where A denotesthe heat
of the chemical reaction which supplies this energy. In accordance
with Thomson's principle, we must replace this equation by
dE
E= A + T dT'
which is the correct relation between the electromotive force of a
cell an<;lthe energy of the chemical reactions which occur in it. In
general the term ,\ is much larger than the term TdEfdT; but in
certain classesof cells-e,g. concentration cells-,\ is zero; in which
casethe whole of the electrical energy is procured at the expenseof
the thermal energy of the cells' surroundings.1
To establishThomson's equation, we usethe principle that if dQ
is the quantity of heat communicated to a working substanceat the
absolutetemperature T in a seriesof reversible changes,then dQfT,
being the differential of the entropy of the working substance,is a
perfectdifferential. In the caseof the cell, if a quantity e of electricity
passesin the circuit, and if U is the internal energy of the cell, so
that ,\ = -
au
ei' then
dQ = Ede + dU = (E - A) de + ~~dT.
. r
TI lerelore E - Ade + TI aT
--r au dT ' IS a peuect
.J' d'ffi . I
1 erentla,
so a
aTr-
(E - A) =
a
ai
(If au) I 02U
aT T oToe = = -
1 OA
TaT
1 oE E - A
or 'ti3T-TT=O
or E = A + T~~
aT
which is Thomson's equation.
Now returning to Helmholtz' memoir of 1847, let us consider
his discussionof the energy of an electrostatic field. I t will be
1 It is at thejunctionsthat this interchangeof thermal energywith the outsideworld
generallytakesplace: it is relatedto the Peltier effect,describedlater.
21 7
convenient to supposethat the systemhasbeenformed by continually

bringing from a very great distance infinitesimal quantities of elec-
tricity, proportional to the quantities already present at the various
points of the system; so that the charge is always distributed pro-
portionally to the final distribution. Let e typify the final chargeat
any point of space,and V the final potential at this point. Then at
any stageof the processthe charge and potential at this point will
have the values ,\e and '\V, where,\ denotesa proper fraction. At
this stagelet chargesed,\ be brought from a great distanceand added
to the charges,\e. The work required for this is
L e dA . A V,
so the total work required in order to bring the systemfrom infinite
r
dispersionto its final state is
Le V. AdA, or lLe V.
o
By reasoning similar to that used in the case of electrostatic
distributions, it may be shown that the energy of a magnetic field,
which is due to permanent magnetsand which also contains bodies
susceptibleto magnetic induction, is
t J JJ Po c/> dx dy dz,
wherePo denotesthe densityof Poisson'sequivalent magnetisation,for

the permanent magnetsonly, and c/> denotesthe magneticpotential. 1
Helmholtz, moreover, applied the principle of energy to systems
containing electric currents. For instance, when a magnet is
movedin the vicinity of a current, the energytaken from the battery
may be equated to the sum of that expendedas Joulean heat and
that communicated to the magnet by the electromagnetic force :
and this equation showsthat the current is not proportional to the
electromotive force of the battery, i.e. it reveals the existenceof
Faraday's magneto-electric induction. As, however, Helmholtz
was at the time unacquainted with the conception of the electro-
kinetic energy stored in connection with a current, his equations
were for the most part defective. But in the caseof the mutual
action of a current and a permanent magnet, he obtained the
correct result that the time-integral of the induced electromotive
force in the circuit is equal to the increasewhich takesplace in the
potential of the magnet towards a current of a certain strength in
the circuit.
1 We suppose all transitions to be continuous, so as to avoid the necessity for writing
surface-integrals separately.
218
The correct theory of the energy of magnetic and electro-

magnetic fields is due mainly to W. Thomson (Kelvin).
Thomson's researcheson this subject commencedwith one or two
short investigationsregarding the ponderomotive forces which act
on temporary magnets. In 1847he discussed1 the caseof a small
iron sphereplaced in a magnetic field, showing that it is acted on
by a ponderomotiveforce representedby - grad cR', where c denotes
a constant, and R denotesthe magnetic force of the field; such a
spheremust evidently tend to move towards the places where R'
is greatest. The same analysis may be applied to explain why
diamagnetic bodies tend to move, as in Faraday's experiments,
from the strongerto the weaker parts of the field.
Two years later Thomson presented to the Royal Society a
memoir 2 in which the results of Poisson'stheory of magnetism
were derived from experimental data, without making use of the
hypothesisof magnetic fluids; and this was followed in 1850 by
a secondmemoir,3 in which Thomson drew attention to the fact
previouslynoticed by Poisson,4that the magneticintensity at a point
within a magnetisedbody dependson the shapeof the small cavity
in which the exploring magnet is placed. Thomson distinguished
two vectors6 ; one of these, by later writers generally denoted by
B, representsthe magnetic intensity at a point situated in a small
crevice in the magnetisedbody, when the facesof the crevice are
at right angles to the direction of magnetisation; the vector B is
always circuital, that is, div B = O. The other vector, generally
denotedby H, representsthe magnetic intensity in a narrow tubular
cavity tangential to the direction of magnetisation; it is an irrota-
tional vector, that is, curl H = O. The magnetic potential tends at
any point to a limit which is independent of the shapeof the cavity
in which the point is situated; and the space-gradientof this limit
is identical with H . Thomsoncalled B the' magneticforce according
to the electromagnetic definition,' and H the 'magnetic force
according to the polar definition'; but the namesmagnetic induction
and mugneticforce, proposedby Maxwell, have been generally used
by later writers.
It may be remarked that the vector to which Faraday applied
the term ' magnetic force,' and which he representedby lines of
force, is not H, but B; for the number of unit lines of force passing
1 Camb. and Dub. Math. Journal, ii (1847),p. 230; W. Thomson'sPapers on Electro-
statics and Magnetism, p. 499; cf. alsoPhil. Mag. xxxvii (1850),p. 241
• Phil. Trans. (1851),p. 243; Thomson'sPapers on Elect. and Mag., p. 345
• Phil. Trans. (1851),p. 269; Papers on Elect. and Mag., p. 382 • cr.p. 64
• lococit. § 78 of the original paper,and § 517of the reprint
219
through any gap must depend only on the gap, and not on the
particular diaphragm filling up the gap, acrosswhich the flux is
estimated; and this can be the case only if the vector which is
representedby the lines of force is a circuital vector.
Thomson introduced a number of new terms into magnetic
science-as indeed he did into every sciencein which he was inter-
ested. The ratio of the measureof the induced magnetisation Ii'
in a temporary magnet, to the magnetisingforce H, he named the
susceptibility; it is positive for paramagnetic and negative for dia-
magnetic bodies, and is connected with Poisson'sconstant k/ by
the relation _ 3 kp
K - 417 I - kp'
where K denotesthe susceptibility. By an easyextensionof Poisson's
analysis it is seenthat the magnetic induction and magnetic force
are connectedby the equation
B =H 4171, +
where I denotesthe total intensity of magnetisation: so if 10 denote
the permanent magnetisation,we have
B = H + 417li + 4171 0,
= p.H+ 4171 0,
wherep. denotes(1 + 417K) ; p. wascalled by Thomson the permeability.

In 1851 Thomson extendedhis magnetic theory so as to include
magnecrystallicphenomena. The mathematical foundations of the
theory of magnecrystallic action had been laid by anticipation, long
before the experimental discoveryof the phenomenon,in a memoir
read by Poissonto the Paris Academy in February 1824. Poisson,
as will be remembered, had supposedtemporary magnetism to be
due to 'magnetic fluids,' movable within the infinitely small
, magnetic elements' of which he assumedmagnetisablematter to
be constituted. He had not overlooked the possibility that in
crystals these magnetic elements might be non-spherical (e.g.
ellipsoidal), and symmetrically arranged; and had remarked that
a portion of such a crystal, when placed in a magnetic field, would
act in a manner depending on its orientation. The relations con-
necting the induced magnetisation I with the magnetising force
II he had given in a form equivalent to
I" = aH" + b'Hy + c"Hz,
{ Iy = a"H" + bHy + c'Hz,
Iz = a'H" + b"Hy + cHz•
1 cr. p. 65
220
Thomson now showed1 that the nine coefficientsa, b', c" ... ,
introduced by Poisson,are not independent of each other. For
a spherecomposedof the magnecrystallinesubstance,if placed in
a uniform field of force, would be acted on by a couple: and the
work doneby this couple when the sphere,supposedof unit volume,
performs a complete revolution round the axis of x may be easily
shown to be 7TH (1 - Hx2/H2) (- b" + c'). But this work must
be zero, since the system is restored to its primitive condition;
and henceb" and c' must be equal. Similarly c" = a', and a" = b'.
By changeof axesthree more coefficientsmay be removed, so that
the equationsmay be brought to the form
Ix = K1HX' Iy = K2Hy, Iz = KaHz,
where Kl, K2, KS may be called the principal magnetic susceptibilities.
In the same year (1851) Thomson investigated the energy
which, as was evident from Faraday's work on self-induction, must
be stored in connection with every electric current. He showed
that, in his own words,2, the value of a current in a closedconductor,
left without electromotiveforce, is the quantity of work that would
be got by letting all the infinitely small currents into which it may
be divided along the lines of motion of the electricity cometogether
from an infinite distance,and makeit up. Each of these" infinitely
small currents" is of coursein a circuit which is generally of finite
length; it is the sectionof eachpartial conductor and the strength
of the current in it that must be infinitely small.'
Discussingnext the mutual energy due to the approach of a
permanent magnet and a circuit carrying a current, he arrived at
the remarkable conclusion that in this case there is no electro-
kinetic energy which dependson the mutual action; the energy is
simply the sum of that due to the permanent magnets and that
due to the currents. If a permanent magnet is causedto approach
a circuit carrying a current, the electromotive force acting in the
circuit is thereby temporarily increased; the amount of energy
dissipatedasJoulean heat, and the speedof the chemical reactions
in the cells,are temporarily increasedalso. But the increasein the
Joulean heat is exactly equal to the increasein the energy derived
from consumptionof chemicals,together with the mechanical work
done on the magnet by the operator who moves it; so that the
balanceof energyis perfect, and none needsto be added to or taken
from the electrokinetic form. It will now be evident why it was
1 Phil. Mag. (4), i (1851), p. 177; Paperson Electrostaticsand Magnetism, p. 471
• Paperson Electrostaticsand Magnetism, p. 446
221
that Helmholtz escapedin this casethe errors into which he was

led in other casesby his neglect of electrokinetic energy; for in
this casethere was no electrokinetic energyto neglect.
Two years later, in 1853, Thomson 1 gave a new form to the
expressionfor the energy of a systemof permanent and temporary
magnets.
We have seenthat the energyof such a systemis representedby
1 HI Po4> dx dy dz,
where Po denotesthe density of Poisson'sequivalent magnetisation

for the permanent magnets,and 4> denotesthe magnetic potential,
and where the integration may be extendedover the whole of space.
Substituting for Po its value - div 10,2 the expression may be
written in the form
or, integrating by parts,
-1 IH (10 • grad 4»dx dy dz, or -I IH (H. 10) dx dy dz.
Since B = fLH + 47rIo' this expression may be written in the

form
- 8~ III (H • B) dx dy dz + 8~ III fLHI dx dy dz ;
but the former of theseintegrals is equivalent to
- 8~ HI (B. grad 4» dx dv dz, or .- 8~ HI 4> div B dx dy dz,
which vanishes,since B is a circuital vector. The energy of the

field, therefore,reducesto
8~ HI fLH' dx dy ik,
integrated over all space; which is equivalent to Thomson'sform. 8
In the same memoir Thomson returned to the question of the
energywhich is possessed by a circuit in virtue of an electric current
1 Pmc. Glasgow Phil. Soc. iii (1853), p. 281 ; Kelvin's Math. and Phys. Papers, i, p. 521
I cr.
p. 64
• The form actually given by Thomson was
J; III (HI + 4: II) dx dy d<:,

which reduces to the above when we neglect that part of I" which is due to the permanent
magnetism, over which we have no control.
222
circulating in it. As he remarked, the energy may be determined

by calculating the amount of work which must be done in and
on the circuit in order to double the circuit on itself while the current
is sustainedin it with constantstrength; for Faraday's experiments
showthat a circuit doubled on itself hasno storedenergy. Thomson
found that the amount of work required may be expressedin the
form tLi2, where i denotesthe current strength, and L, which is
called the coe.fJicientof self-induction, dependsonly on the form of the
circuit.
It may be noticed that in the doubling processthe inherent
electrodynamic energy is being given up, and yet the operator is
doing positive work. The explanation of this apparent paradox is
that the energy derived from both these sourcesis being used to
savethe energy which would otherwisebe furnished by the battery,
and which is expendedin Joulean heat.
Thomson next proceeded1 to show that the energy which is
stored in connection with a circuit in which a current is flowing
may be expressedas a volume-integral extended over the whole
of space,similar to the integral by which he had already repre-
sentedthe energyof a systemof permanent and temporary magnets.
The theorem, as originally stated by its author, applied only to the
case of a single circuit; but it may be establishedfor a system
formed by any number of circuits in the following way:
If Ns denote the number of unit tubes of magnetic induc-
tion which are linked with the sth circuit, in which a curent is is
flowing, the electrokinetic energy of the system is t~Nsis; which
may be written t~I" where I, denotes the total current flowing
through the gap formed by the ,Jh unit tube of magnetic induc-
tion. But if H denote the (vector) magnetic force, and Hits
numerical magnitude, it is known that (lj47T)I Hds, integrated
along a closed line of magnetic induction, measures the total
current flowing through the gap formed by the line. The
energy is therefore (I j87T) ~ I Hdr, the summation being extended
over all the unit tubes of magnetic induction, and the integra-
tion being taken along them. But if dS denote the cross-section
of one of these tubes, we have BdS = 1, where B denotes the
numerical magnitude of the magnetic induction B: so the energy
is (1/87T)~BdS I Hdr; and as the tubes fill all space, we may
replace ~dS I dr by III dx dy d;:;. Thus the energy takes the form
1 Nichols' CyclopaLdja, 2nd edn., 1860,article' Magnetism,dynamicalrelationsof' ;

reprintedin Thomson'sPapers on Eled. and Mag., p. 447,and his Math. and Phys. Papers,
p.532
223
(I j87T) Iff BH dx dy dz, where the integration is extendedover the

whole of space; and sincein the presentcaseB = p.H, the energy
may alsobe representedby (l j87T) III p.H 2 dx dy dz.
But this is identical with the form which was obtained for a
field due to permanent and temporary magnets. It thus appears
that in all casesthe stored energy of a systemof electric currents
and permanent and temporary magnetsis
where the integration is extendedover all space.

It must, however, be remembered that this representsonly
what in thermodynamicsis called the ' available energy'; and it
must further be rememberedthat part even of this available energy
may not be convertible into mechanicalwork within the limitations
of the system: e.g. the electrokinetic energy of a current flowing
in a single closedperfectly conducting circuit cannot be converted
into any other form so long as the circuit is absolutely rigid. All
that we can say is that the changesin this stored electrokinetic
energy correspond to the work furnished by the system in any
change.
The above form suggeststhat the energy may not be localised
in the substanceof the circuits and magnets,but may be distributed
over the whole of space, an amount (p.H2j87T) of energy being
contained in each unit volume. This conception was afterwards
adopted by Maxwell, in whose theory it is of fundamental im-
portance.
While Thomson was investigating the energy stored in con-
nection with electric currents, the equationsof flow of the currents
were being generalisedby Gustav Kirchhoff (1824-87). In 1848
Kirchhoff 1 extended Ohm's theory of linear conduction to the
caseof conduction in three dimensions; this could be done without
much difficulty by making useof the analogy with the flow of heat,
which had proved so useful to Ohm. In Kirchhoff's memoir a
systemis supposedto be formed of three-dimensionalconductors,
through which steady currents are flowing. At any point let V
denote the 'tension' or 'electroscopic force '-a quantity the
significanceof which in electrostaticswas not yet correctly known.
Then, within the substanceof any homogeneousconductor, the
function V must satisfyLaplace's equation V2V = 0; while at the
air-surface of each conductor, the derivate of V taken along
I Ann. d. Phys. !xxv (1848), p. 189; Kirchhoff's Gu. Abhandl., p. 33
224
THE MIDDLE OF THE NINETEENTH CENTUR.Y
the normal must vanish. At the interface between two conductors

formed of different materials, the function V has a discontinuity,
which is measured by the value of Volta's contact force for the
two conductors; and, moreover, the condition that the current shall
be continuous acrosssuch an interface requires that koV/oN shall be
continuous, where k denotesthe ohmic specific conductivity of the
conductor, and a/oN denotesdifferentiation along the normal to the
interface. The equations which have now been mentioned suffice
to determine the flow of electricity in the system.
Kirchhoff also showed that the currents distribute themselves
in the conductors in such a way as to generate the least possible
amount of Joulean heat; as is easily seen, since the quantity of
Joulean heat generatedin unit time is
where k, as before, denotes the specific conductivity; and this

integral has a stationary value when V satisfiesthe equation
~ (k OV)
ox ox + ~
Oy
(k OV) ~ (k ?z
oy + oz V) = o.
Kirchhoff next applied himself to establish harmony between

e1ectrostaticalconceptions and the theory of Ohm. That theory
had now been before the world for twenty years, and had been
verified by numerous experimental researches; in particular, a
careful investigation was made at this time (1848) by Rudolph
Kohlrausch (1809-58), who showed1 that the difference of the
electric 'tensions' at the extremities of a voltaic cell, measured
electrostatically with the circuit open, was for different cells pro-
portional to the electromotive force measuredby the electrodynamic
effectsof the cell with the circuit closed; and, further,2 that when
the circuit was closed, the difference of the tensions, measured
electrostatically, at any two points of the outer circuit was pro-
portional to the ohmic resistanceexisting between them. But in
spite of all that had been done, it was still uncertain how' tension,'
or ' electroscopicforce,' or 'electromotive force' should be inter-
preted in the language of theoretical electrostatics; it will be
remembered that Ohm himself, perpetuating a confusion which
had originated with Volta, had identified electroscopicforce with
density of electric charge, and had assumedthat the electricity in a
1 AM. d. Phys.!xxv (1848), p. 220 • AM. d. Phys.!xxviii (1849), p. 1
225
conductor is at rest when it is distributed uniformly throughout the

substanceof the conductor.
The uncertainty was finally removed in 1849 by Kirchhoff,l
who identified Ohm's electroscopic force with the electrostatic
potential. That this identification is correct may be seenby com-
paring the different expressionswhich have been obtained for
electric energy; Helmholtz' expression2 shows that the energy
of a unit charge at any place is proportional to the value of the
electrostatic potential at that place; while Joule's result 3 shows
that the energy liberated by a unit charge in passing from one
place in a circuit to another is proportional to the difference of the
electric tensions at the two places. It follows that tension and
potential are the samething.
The work of Kirchhoff was followed by several other investiga-
tions which belong to the borderland between electrostaticsand
electrodynamics. One of the first of these was the study of the
Leyden jar discharge.
Early in the century Wollaston, in the courseof his experiments
on the decompositionof water, had observedthat when the decom-
position is effected by a dischargeof static electricity, the hydrogen
and oxygen do not appear at separateelectrodes; but that at each
electrode there is evolved a mixture of the gases,as if the current
had passed through the water in both directions. After this
F. Savary· had noticed that the dischargeofa Leydenjar ma:snetises
needlesin alternating layers, and had conjectured that' the electric
motion during the discharge consistsof a seriesof oscillations.' A
similar remark was made in connection with a similar observation
by Joseph Henry of Washington, in 1842.5 'The phenomena,'
he wrote, 'require us to admit the existence of a principal
discharge in one direction, and then severalreflex actionsbackward
and forward, each more feeblethan the preceding, until equilibrium
is restored.' Helmholtz had repeated the same suggestionin his
essayon the conservation of energy; and in 1853 W. Thomson 6
verified it, by investigating the mathematical theory of the dis-
charge, as follows :
Let C denote the capacity of the jar, Le. the measureof the
charge when there is unit difference of potential between the coat-
ings; let R denote the ohmic resistanceof the discharging circuit,
1 ibid. !xxviii (1849), p. 506; Kirchhoff's Gu. Abhandl., p. 49; Phil. Mag. (3) xxxvii
(1850), p. 463
• cf. p. 218 • ef. p. 212 • Annales de Chimie, xxxiv (1827), p. 5
6 Proc. Am. Phil. Soc. ii (1842), p. 193
• Phil. Mag. (4) v (1853), p. 400; Kelvin's Math. and Phys. Papers, i, p. 540
226
and L its coefficientof self-induction. Then if at any instant t the

charge of the condenserbe Q, and the current in the wire be i,
we have i = dQ/dt; while Ohm's law, modified by taking self-
induction into account, givesthe equation
Ri + L~~= -~.
Eliminating i, we have
d2Q
L(jj2
dQ
+ Rllt + c1 Q= 0,
an equation which showsthat when R 2C< 4L, the subsidenceof

Q to zerois effectedby oscillationsof period
27T
1
(L C - ~It!.
IfR2C> 4L, the dischargeis not oscillatory.
This simple result may be regarded as the beginning of the
theory of electric oscillations.1
Thomson was at this time much engagedin the problems of
submarinetelegraphy; and thus he was led to examine the vexed
questionof the' velocity of electricity' over long insulated wires and
cables. Various workers had made experiments on this subject
at different times, but with hopelesslydiscordant results. Their
attempts had generally taken the form of measuring the interval
of time betweenthe appearanceof sparksat two spark gaps in the
samecircuit, betweenwhich a great length of wire intervened, but
which were brought near each other in order that the discharges
might be seentogether. In one seriesof experiments,performed
by Watsonat Shooter'sHill in 1747-8,2the circuit was four miles in
length, two miles through wire and two miles through the ground ;
but the dischargesappearedto be perfectly simultaneous; whence
Watsonconcludedthat the velocity of propagation of electric effects
is too great to be measurable.
In 1834 Charles Wheatstone,3 Professor of Experimental
Philosophyin King's College,London, by examining in a revolving
mirror sparks formed at the extremities of a circuit, found the
velocity of electricity in a copper wire to be about one and a half
times the velocity of light. In 1850H. Fizeau and E. Gounelle,'
1 Thomson's theory was verified by W. Feddersen of Leipzig in a series of papers
published in 1857--66; these have been reprinted as Ostwald's Klassiker, no. 166.
Feddersen was the first to investigate the oscillating discharge fully and correctly.
I Phil. Trans. xlv (1748), pp. 49, 491 I Phil. Trans. (1834), p. 583
, CompusRendus,xxx (1850), p. 437
(828) 227
experimenting with the telegraph lines from Paris to Rouen and to

Amiens, obtained a velocity about one-third that of light for the
propagation of electricity in an iron wire, and nearly two-thirds that
of light for the propagation in a copper wire.
The reasonsfor thesediscrepancieswere not understood at the
time. The large value of the velocity found by Wheatstonewas
mainly due to the fact that his wire was not straight, but was coiled
in twenty straight windings: and the effect travels in a zig-zag or
spiral wire more quickly than in a straight one.1
An important advancewas made when Faraday,2early in 1854,
showed experimentally that a submarine cable, formed of copper
wire covered with gutta-percha, ' may be assimilatedexactly to an
immenseLeyden battery; the glassof the jars representsthe gutta-
percha; the internal coating is the surface of the copper wire,'
while the outer coating correspondsto the sea water. It follows
that in all calculations relating to the propagation of electric dis-
turbances along submarine cables,the electrostatic capacity of the
cable must be taken into account.
The theory of signalling by cable originated in a correspondence
between Stokesand Thomson in 1854.a In the caseof long sub-
marine lines, the speed of signalling is so much limited by the
electrostatic factor that electromagnetic induction has no sensible
effect; and it was accordingly neglectedin the investigation. In
view of other applications of the analysis,however,we shall suppose
that the cable has a self-induction L per unit length, and that R
denotesthe ohmic resistance,and C the capacity per unit length,
V the electric potential at a distancex from one terminal, and i the
current at this place. Ohm's law, as modified for inductance, is
expressedby the equation
_ oV _ L ~
ox - ot
+ R·l ,
.
moreover, since the rate of accumulation of charge in unit length

at x is - oil ox, and since this increasesthe potential at the rate
- (l/C) oil ox, we have oV oi
Caf=-bx"
Eliminating i betweenthesetwo equations,we have an equation
1 cr. J. Stefan, WienAnzeiger,xxviii (1891), p. 106
I Proc. Roy. Inst., 20 Jan. 1854; Phil. Mag. vii (June 1854), p. 396; Exp. Res.iii,
pp. 508, 521
I The Dover-Calais cable was laid in 1851; the Holyhead-Howth cable in 1853.
The first Atlantic cable, which lasted three weeks, was laid in 1858; the second and third
in 1866.
first obtained by Oliver Heaviside 1 (1850-1925), namely

1 O'V o'V av
C ax' = L ali + R at'
which is known as the equation of telegraphy.2
Thomson,in one of his letters 3 to Stokes,in 1854, obtained this
equation in the form which applies to Atlantic cables,i.e. with the
term in L neglected. In this form it is the sameasFourier's equation
for the linear propagation of heat: so that the known solutions of
Fourier's theory may be used in a new interpretation. If we sub-
stitute
we obtain
v
= e2ntv~ + Ax ,
A = ± (1 +
V - 1) (nCR)! ;
and thereforea typical elementarysolution of the equation is
J
V = e- (nCR)'1x sin {2nt - (nCR)h}.

The form of this solution showsthat if a regular harmonic variation
of potential is applied at one end of a cable, the phaseis propagated
with a velocity which is proportional to the square root of the
frequencyof the oscillations: sincethereforethe different harmonics
are propagatedwith different velocities,it is evident that no definite
'velocity of transmission' is to be expected for ordinary signals.
If a potential is suddenly applied at one end of the cable, a certain
time elapsesbefore the current at the other end attains a definite
percentageof its maximum value; but it may easily be shown 4
that this retardation is proportional to the resistanceand to the
capacity, and each of these is proportional to the length of the
cable: sothe retardation is proportional to the squareof the length:
thus the apparent velocity of propagation would be less,the greater
the length of cable used.
The caseof a telegraph wire insulated in the air on poles is
different from that of a cable; for here the capacity is small, and
it is necessaryto take into account the inductance. If in the general
equation of telegraphy we write
V -- eMv-=-l + ILt,
fL = - 2L
R ± (R' n')! ;
40 - CL
1 Phil. Mag. ii (Aug. 1876), p. 135
• We have neglected leakage, which is beside our present purpose.
• Proc. R. S. vii (May 1855), p. 382 ; Kelvin's Math. and Phys. Papers, ii, p. 61
• This result, indeed, follows at once from the theory of dimensions.
229
as the capacity is small, we may replace the quantity under the

radical by its secondterm: and thuswe seethat a typical elementary
solution of the equation is
Rt
V = e -2Lsin n {x - (CL)-1 t} ;
this showsthat any harmonic disturbance, and therefore any dis-

turbance whatever, is propagated along the wire with velocity
(CL)4. The difference betweenpropagation in a telegraph wire
and propagation in an oceanic cable is, as Thomson remarked,
similar to the difference between the propagation of an impulsive
pressurethrough a long column of fluid in a tube when the tube
is rigid (caseof the telegraph wire) and when it is elastic,so as to
be capable of local distension (case of the cable, the distension
correspondingto the effect of capacity); in the former case,as is
well known, the impulse is propagated with a definite velocity,
namely, the velocity of soundin the fluid.
The work of Thomson on signalling along cableswas followed
in 1857by a celebratedinvestigation1 of Kirchhoff's, on the propa-
gation of electric disturbance along a telegraph wire of circular
cross-section.
Kirchhoff assumedthat the electric charge is practically all
residenton the surfaceof the wire, and that the current is uniformly
distributed over its cross-section; his idea of the current was the
sameasthat of Fechnerand Weber,namely, that it consistsof equal
streams of vitreous and resinous electricity flowing in opposite
directions. Denoting the electric potential by V, the charge per
unit length of wire bye, the length of the wire by l, and the radius
of its cross-section
by a,heshowedthat V is determinedapproximately
by the equation 2 V = 2e log (lfa).
The next factor to be consideredis the mutual induction of the
current-elementsin different parts of the wire. Assuming with
1 Ann. d. Phys.c (1857),pp. 193,251; Kirchhoff's Ges.Abhandl.,p. 131; Phil. Mag.
xiii (1857),p. 393

• His method of obtaining this equationwas to calculateseparatelythe effectsof
(1) the portion of the wire within a distance£ on either sideof the point considered,
where£ denotesa lengthsmallcomparedwith I, but largecomparedwith a, and (2) the
restof the wire. He thusobtainedthe equation
2£
V = 2e log a
- + f e'ds'
-,,
wherethe integrationis to be takenoverall the lengthof the wire exceptthe portion2£:
the equationgiven in the text wasthen derivedby an approximation,which, howe'-er,
is opento someobjection.
230
Weber that the electromotive force induced in an element ds due

to another element ds' carrying a current if is derivable from a
vector-potential ., (r.d s') r
l r3 ,
Kirchhoff found for the vector-potential due to the entire wire the
approximatevalue W = 2i log (lfa),
where i denotesthe strength of the current 1; the vector-potential

being directed parallel to the wire. Ohm's law then gives the
equation . (oV loW)
l = - Trka
l
ox + C' at '
where k denotesthe specific conductivity of the material of which
the wire is composed; and finally the principle of conservationof elec-
tricity givesthe equation Oi oe
ax=-Ff
Denoting log Ufa) by y, and eliminating e, l, w from these four
equations,wehave OIV I OIV I oV
ox =
l l otl c
2YKTra -ai'
i +
which is, as might have been expected,the equation of telegraphy.
When the term in oV fat is ignored, as we have seenis in certain
casespermissible,the equation becomes
OIV I OIV
ox l = Ci ail'
which shows that the electric disturbance is propagated along the wire
with the velociry c.2 Kirchhoff's procedure has, in fact, involved
the calculation of the capacity and self-induction of the wire, and
is thus able to supply the definite values of the quantities which
were left undeterminedin the general equation of telegraphy.
1 This expression was derived in a similar way to that for V, by an intermediate
formula
w = 2i log..!
2
a
+f"ds'
'_T cos 8 cos 8',
where 8 and 8' denote respectively the angles made with r by ds and ds'.
I In referring to the original memoirs of Weber and Kirchhoff, it must be remembered
that the quantity which in the present work is denoted by c, and which represents the
velocity of light in free aether, was by these writers denoted by c/v'2. Weber, in fact,
denoted by c the relative velocity with which two charges must approach each other in
order that the force between them, as calculated by his formula, should vanish.
It must also be remembered that those writers who accepted the hypothesis that
currents consist of equal and opposite streams of vitreous and resinous electricity, were
accustomed to write 2i to denote the current strength.
231
The velocity c, whose importance was thus demonstrated,has

already been noticed in connection with Weber's law of force; it
is a factor of proportionality, which must be introduced when
electrodynamic phenomena are described in terms of units which
have been defined electrostatically,1 or conversely when units
which have becn defined electrodynamically are usedin the descrip-
tion of electrostaticphenomena. That the factor which is introduced
on such occasionsmust be of the dimensions (length/time), may
be easily seen; for the electrostatic repulsion between electric
chargesis a quantity of the samekind asthe electrodynamicrepulsion
betweentwo definite lengths of wire, carrying currents which may be
specified by the amount of charge which travels past any point in
unit time.
Shortly before the publication of Kirchhoff's memoir, the
value of c had been determined by Weber and Kohlrausch 2 ;
their determination rested on a comparison of the measuresof the
charge of a Leyden jar, as obtained by a method depending on
electrostatic attraction, and by a method depending on the effects
of the current produced by discharging thejar. The resulting value
was nearly
c = 3.1 X 1010 cm.jsec. ;
which was the same,within the limits of the errors of measurement,
as the speedwith which light travels in interplanetary space. The
coincidence was noticed by Kirchhoff, who was thus the first to
discover the important fact that the velocity with which an electric
disturbance is propagated along a perfectly conducting aerial wire
is equal to the velocity of light.
In a second memoir published in the same year, Kirchhoff 3
extendedthe equationsof propagation of electric disturbance to the
caseof three-dimensionalconductors.
As in his earlier investigation, he divided the electromotive
force at any point into two parts, of which one is the gradient of
the electrostatic potential 4>, and the other is the derivate with
respect to the time (with sign reversed) of a vector-potential a ;
so that if s denotethe current and k the specificconductivity, Ohm's
law is expressedby the equation
s=k (c1
grad 1> - ~~).
I cf. p. 202
• Ann. d. Phys. xcix (1856), p. 10. Weber was really following up the work of absolute
measurements begun by himself and Gauss in connection with terrestrial magnetism.
I Ann. d. Phys. cii (1857), p. 529; Ges. A'bhandl., p. 154
232
Kirchhoff calculated the value of a by aid of Weber's formula for

the inductive action of one current-elementon another; the result is
a- - fff ,3
dx' dy' dz' ( r. s') r,
where r denotesthe vector from the point (x, y, z), at which a is

measured,to any other point (x',f, z') of the conductor, at which
the current is s; and the integration is extended over the whole
volume of the conductor. The remaining general equations are
the ordinary equation of the electrostaticpotential
\l2c/> + %p = 0
(where p denotesthe density of electric charge), and the equation
of conservationof electricity
ap + d·IV S = 0•
at
It will be seenthat Kirchhoff's electrical researcheswere greatly
influenced by those of Weber. The latter investigations,however,
did not enjoy unquestionedauthority; for there wasstill a question
asto whether the expressionsgiven by Weber for the mutual energy
of two current-elements,and for the mutual energyof two electrons,
wereto be preferredto the rival formulae of Neumann and Riemann.
The matter was examined in 1870 by Helmholtz, in a series of
memoirs 1 to which referencehas already been made.2 Helmholtz
remarked that, for two elementsds,ds', carrying currents i, i', the
electrodynamicenergyis ii' (ds. ds')
r '
according to Neumann, and
..,
, (r. ds)(r. ds'),
t~
according to Weber; and that these expressionsdiffer from each

other only by the quantity
iT'dsds'
--{-
r
cos(ds. ds') + cos(r. ds) cos(r. ds') },
2
or .., d ds' d r •
tt s ds dS,- ,
smce this vanisheswhen integrated round either circuit, the two

1 Journalfiir Math. !xxii (1870), p. 57; !xxv (1873), p. 35; !xxviii (1874), p. 273
• cf. ~. 204
233
formulae give the sameresult when applied to entire currents. A

generalformula including both that of Neumann and that of Weber
is evidently .., (d d') d2
tt s. s
r
k'" _r_ d ds'
U dsds' s ,
+
where k denotesan arbitrary constant.l
Helmholtz's result suggestedto Clausius2 a new form for the
law of force betweenelectrons; namely, that which is obtained by
supposing that two electrons of chargese, e', and velocities v, v',
possess electrokinetic energy of amount
ee' (v • v')
r
+ kee' dsds'
d'r ,
vv.
Subtracting from this the mutual electrostatic potential energy,
which is ee'c2/r, we may write the mutual kinetic potential of the
two electronsin the form
ee' (OX
r at
ax: +J!)' ;y + oz oz' _
at at at at at
c1) + kee' dsds'
_dtr vv'
,
where (x,y, z) denotethe co-ordinatesof e, and (x',y', z') thoseof e'.
The unknown constant k has clearly no influence so long as
closed circuits only are considered; if k be replaced by zero, the
expressionfor the kinetic potential becomes
ee' (ox ox' +?1_ry' + ~_oz' _ c1)
r at at at at at at '
or, in the electrostaticsystemof units,
ee'(l ox ox' +
r ct at at
1 ryat ry'at +.!.-c' ozat oz'at _ 1)'
c'
which, as will appear later, closely resemblesthe corresponding
expressionin the Lorentz theory of electrons.3
Clausius' formula has the great advantage over Weber's, that
it does not compel us to assumeequal and opposite velocities for
1 cf. H. Lamb, Proc. Lond. Math. Soc. xiv (1883), p. 301
• }ournaljiir Math. lxxxii (1877), p. 85; Phil. Mag. x (1880), p. 255
• In fact, when any number of electrons are present, the part of the kinetic potential
which concerns anyone of them-say e-may be written
I.e = e (~a%!!:
c2 at +J.
c2
all at°l +-~
c·
Q.z~~
at - "')
where a and'" denote potential-functions defined by the equations
a = HY;'dx' dy'tk', '" = HJ~dx' dy'd~'.
To convert these into Lorentz's formulae, it is only necessaryto retard the potentials.
234
THE ~UDDLE OF THE NINETEENTH CENTURY
the vitreous and resinouschargesin an electric current; on the

other hand, Clausius' expressioninvolves the absolutevelocities of
the electrons,while Weber'sdependsonly on their relative motion;
and therefore Clausius' theory requires the assumptionof a fixed
aetherin space,to which the velocitiesv and Vi may be referred.
When the behaviour of finite electrical systemsis predicted
from the formulae of Weber, Riemann and Clausius, the three
laws do not always lead to concordant results. For instance, if a
circular current be rotated with constant angular velocity round
its axis, according to Weber's law there would be a development
of free electricity on a stationary conductor in the neighbourhood;
whereas,accordingto Clausius'formula there would be no induction
on a stationary body, but electrification would appear on a body
turning with the circuit as if rigidly connectedwith it. Again,!
let a magnet be suspendedwithin a hollow metallic body, and let
the hollow body besuddenlychargedor discharged; then, according
to Clausius' theory, the magnet is unaffected; but according to
Weber'sand Riemann'stheoriesit experiencesan impulsive couple.
And again, if an electrified disk be rotated in its own plane, under
certain circumstancesa steadycurrent will be induced in a neigh-
bouring circuit accordingto Weber'slaw, but not according to the
other formulae.
An interesting objection to Clausius' theory was brought for-
ward in 1879by Frohlich 2-namely, that when a charge of free
electricity and a constant electric current are at rest relatively to
each other, but partake together of the translatory motion of the
earth in space,a forceshouldact betweenthem if Clausius'law were
true. It was, however, shown by Budde 3 that the circuit itself
acquiresan electrostaticcharge,partly asa result of the sameaction
which causesthe force on the external conductor, and partly as a
result of electrostaticinduction by the charge on the external con-
ductor; and that the total force betweenthe circuit and external
conductoris thus reducedto zero.'
We have seenthat the discrimination betweenthe different laws
1 The two following crucial experiments, with others, were suggestedby E. Budde,
Ann. d. Phys. xxx (1887), p. 100.

I Ann. d. Phys. ix (1880), p. 261 I Ann. d. Phys. x (1880), p. 553
• This caseof a charge and current moving side by side was afterwards examined by
FitzGerald (Trans. R~y. Dub. Soc. i [1882], p. 319: Scient. Writings of G. F. FitzGerald,
p. III) without reference to Clausius' formula, from the standpoint of Maxwell's theory.
The result obtained was the same-namely, that the electricity induced on the conductor
carrying the current neutralises the ponderomotive force between the current and the
external charge. The question was also discussed in the light of relativity-theory by
Lorentz, Phys. <:'S. xi (1910), p. 1234.
235
of electrodynamic force is closely connected with the question

whether in an electric current there are two kinds of electricity
moving in opposite directions, or only one kind moving in one
direction. On the unitary hypothesis,that the current consistsin a
transport of one kind of electricity with a definite velocity relative
to the wire, it might be expectedthat a coil rotated rapidly about
its own axis would generate a magnetic field different from that
produced by the samecoil at rest. Experiments to determine the
matter were performed by A. Foppl 1 and by E. L. Nichols and
W. S. Franklin,2 but with negativeresults. The latter investigators
found that the velocity of electricity must be such that the quantity
conveyedpast a specifiedpoint in a unit of time, when the direction
of the current was that in which the coil was travelling, did not
differ from that transferredwhen the current and coil were moving
in opposite directions by as much as one part in ten million, even
when the velocity of the wire was 9,096 cm./sec. They considered
that they would have beenable to detect a changeof deflectiondue
to the motion of the coil, eventhough the velocity of the current had
been considerably greater than a thousand million metres per
second.
During the decadesin the middle of the century considerable
progress was made in the science of thermo-electricity, whose
beginnings we have already described.3 In Faraday's laboratory
notebook, under the date 28 July 1836, we read: 4 'Surely the
converseof thermo-electricity ought to be obtained experimentally.
Passcurrent through a circuit of antimony and bismuth.'
Unknown to Faraday, the experiment here indicated had
already been made, although its author had arrived at it by a
different train of ideas. In 1834Jean CharlesPeltier 6 (1785-1845)
attempted the task, which was afterwards performed with success
byJoule,6 of measuringthe heat evolvedby the passageof an electric
current through a conductor. He found that a current produces
in a homogeneousconductor an elevation of temperature, which is
the samein all parts of the conductor where the cross-sectionis the
same; but he did not succeedin connectingthe thermal phenomena
quantitatively with the strength of the current-a failure which was
due chiefly to the circumstancethat his attention was fixed on the
rise of temperature rather than on the amount of the heat evolved.
But incidentally the investigation led to an important discovery
1 Ann. d. Phys. xxvii (1886), p. 410 • Amer. Jour. Sci. xxxvii (1889), p. 103
• cr. pp. 88,89 • Bence Jones's Life of Faraday, ii, p. 79
• Annales de Chimie, 1vi (1834), p. 371 • cf. p. 212
-namely, that when a current was passedin successionthrough

two conductors made of dissimilar metals, there was an evolution
of heat at thejunction; and that this dependedon the direction of
the current; for if the junction washeatedwhen the current flowed
in one sense,it was cooled when the current flowed in the opposite
sense. This Peltier effect, as it is called, is quite distinct from the
ordinary Joulean liberation of heat, in which the amount of energy
setfreein the thermal form is unaffectedby a reversalof the current;
the Joulean effect is, in fact, proportional to the squareof the current
strength, while the Peltier effect is proportional to the current
strength directly. The Peltier heat which is absorbedfrom external
sourceswhen a current i flows for unit time through ajunction from
one metal B to another metal A may thereforebe denotedby
I1~
(T)i,
where T denotesthe absolute temperature of the junction. The
function I1~ (T) is found to be expressibleas the difference of two
parts, of which one dependson the metal A only, and the other on
the metal B only; thus we can write
I1~(T) = I1A(T) - I1B(T).
In 1851 a general theory of thermo-electric phenomena was
constructed on the foundation of Seebeck's1 and Peltier's dis-
coveriesby W. Thomson.2 Considera circuit formed of two metals,
A and B, and let one junction be maintained at a slightly higher
temperature (T + 8T) than the temperatureT of the otherjunction.
As Seebeckhad shown,a thermo-electriccurrent will be setup in the
circuit. Thomson saw that such a systemmight be regarded as a
heat engine, which absorbsa certain quantity of heat at the hot
junction, and convertspart of this into electrical energy,liberating
the rest in the form of heat at the cold junction. If the Joulean
evolution of heat be neglected,the processis reversible, and must
obey the secondlaw of thermodynamics; that is, the sum of the
quantities of heat absorbed, each divided by the absolute tem-
perature at which it is absorbed,must vanish. Thus we have
I1~(T +
8T) _ I1~(T) _ 0 .
T +
8T T - ,
so the Peltier effect I1~ (T) must be directly proportional to the
1 cf. pp. 88, 89
• Proc. R. S. Edin. iii (1851), p. 91 ; Phil. Mag. iii (1852), p. 529; Kelvin's Math. and
P/>ys.P(!pers,i, p. 316. cr. also Trans. R. S. Edin. xxi (1854), p. 123, reprinted in Papers, i,
p. 232; and Phil. Trans. cxlvi (1856), p, 649, reprinted in Papers, ii, p. 189.
237
absolute temperature T. This result, however, as Thomson well

knew, was contradicted by the observations of Cumming, who
had shownthat when the temperatureof the hot junction is gradually
increased,the electromotiveforce risesto a maximum value and then
decreases. The contradiction led Thomson to predict the existence
of a hitherto unrecognisedthermo-electric phenomenon-namely, a
reversible absorption of heat at placesin the circuit other than the
junctions. Supposethat a current flows along a wire which is of
the same metal throughout, but varies in temperature from point
to point. Thomson showed that heat must be liberated at some
points and absorbed at others, so as either to accentuate or to
diminish the differences of temperature at the different points of
the wire. Supposethat the heat absorbed from external sources
when unit electric charge passesfrom the absolute temperature T
to the temperature (T + 3T) in a metal A is denoted by SA(T)3T.
The thermodynamical equation now takesthe correctedform
IT~(T
T ++
3T) _ IT~(T)
3T T
+ {S BAT
(T) _ S (1')} 3T - 0
- •
Since the metalsA and B are quite independent, this gives
ITA(T +
3T) _ ITA(T) _ S (T) 3T - 0
T +
3T TAT - ,
T
or SA(1') = T d~ {ITAi )}.
This equation connects Thomson's 'specific heat of electricity'

SA(T) with the Peltier effect.
In 1870P. G. Tait 1 found experimentally that the specificheat
of electricity in pure metals is proportional to the absolute tem-
perature. We may therefore write SA(T) = uAT, where UA denotes
a constant characteristic of the metal A. The thermodynamical
equation then becomes
d {ITA (T)}
dT T = erA,
or ITA(T) = 7TAT + uAT·,

where 7T Adenotesanother constant characteristic of the metal. The
chief part of the Peltier effect arisesfrom the term 7T AT.
By the investigations which have been describedin the present
chapter, the theory of electric currents was considerably advanced
1 Proc.R. S. Edin. vii (1870), p. 308. cf. alsoBatelli, Alii della R. Ace.di Torino,xxii
(1886), p. 48, translatedPhil. Mag. xxiv (1887),p. 295.

238
in severaldirections. In all these researches,however, attention

was fixed on the conductor carrying the current as the seat of the
phenomenon. In the following period, interest was centred not so
muchon the conductorswhich carry chargesand currents, ason the
processeswhich take place in the dielectric media around them.
239
Chapter VIII
MAXWELL
SINCE the time of Descartes,natural philosophershave never ceased

to speculateon the manner in which electric and magneticinfluences
are transmitted through space. About the middle of the nineteenth
century, speculationassumeda definite form, and issuedin a rational
theory.
Among thosewho thought much on the matter wasKarl Friedrich
Gauss(1777-1855). In a letter 1 to Weber of date 19 March 1845,
Gaussremarked that he had long ago proposedto himself to supple-
ment the known forceswhich act between electric chargesby other
forces,suchaswould causeelectric actionsto be propagatedbetween
the chargeswith a finite velocity. But he expressedhimself asdeter-
mined not to publish his researchesuntil he should have deviseda
mechanism by which the transmission could be conceived to be
effected; and this he had not succeededin doing.
More than one attempt to realise Gauss' aspiration was made
by his pupil Riemann. In a fragmentary note,2 which appearsto
have been written in 1853, but which was not published until after
his death, Riemann proposed an aether whose elementsshould be
endowedwith the power of resistingcompression,and also (like the
elementsof MacCullagh's aether) of resistingchangesof oricntation.
The former property he conceived to be the causeof gravitational
and electrostatic effects, and the latter to be the causeof optical
and magneticphenomena. The theory thus outlined wasapparently
not developedfurther by its author; but in a short investigation 3
which was published posthumously in 1867,' he returned to the
question of the process by which electric action is propagated
through space. In this memoir he proposed to replace Poisson's
equation for the electrostaticpotential, namely,
V' V + %p = 0,
by the equation
1 o'V
V' V - - -
c' ot'
+ 41TP = 0,
1 Gauss' Werke, v, p. 629 • Riemann's Werke, 2' Aufl., p. 526
• Ann. d. Phys. cxxxi (1867), p. 237 j Riemann's Werke, 2' Aufl., p. 288 j Phil. Mag.
xxxiv (1867), p. 368
• It had been presented to the Gottingen Academy in 1858,but afterwards withdrawn.
240
MAXWELL
according to which the changes of potential due to changing

electrification would be propagated outwards from the charges
with a velocity c. This, so far as it goes,is in agreementwith the
view which is now acceptedas correct; but Riemann's hypothesis
was too slight to serveas the basisof a complete theory. Success
came only when the properties of the intervening medium were
taken into account.
In that power to which Gaussattached so much importance,
of devising dynamical models and analogiesfor obscure physical
phenomena, perhaps no-one has ever excelled W. Thomson 1 ;
and to him, jointly with Faraday, is due the credit of having initiated
the theory of the electric medium. In one of his earliest papers,
written at the age of seventeenwhen he was a very young freshman
at Cambridge,2Thomson compared the distribution of electrostatic
force, in a region containing electrified conductors,with the distri-
bution of the flow of heat in an infinite solid; the equipotential
surfacesin the one casecorrespondto the isothermal surfacesin the
other, and an electric chargecorrespondsto a sourceof heat.3
It may, perhaps,seemas if the value of such an analogy as this
consistedmerely in the prospect which it offered of comparing,
and thereby extending, the mathematical theories of heat and
electricity. But to the physicist its chief interest lay rather in the
idea that formulae which relate to the electric field, and which
had been deducedfrom laws of action at a distance, were shown
to be identical with formulae relating to the theory of heat, which
had been deduced from hypothesesof action between contiguous
particles. 'This paper,' as Maxwell said long afterwards, 'first
introduced into mathematical sciencethat idea of electrical action
carried on by meansof a continuousmedium, which, though it had
been announcedby Faraday, and used by him as the guiding idea
of his researches,had never been appreciated by other men of
1 As will appear from the present chapter, Maxwell had the same power in a very
marked degree. It has always been cultivated by the • Cambridge school' of natural
philosophers, founded by Green, Stokcs and W. Thomson, which was dominated by the
belief that all physical action is founded on dynamics. The value of a dynamical model
is, that it will have properties other than those which suggested its construction; the
question then arises as to whether these properties are found in nature.
I Camb. Math. Jour. iii (1842), p. 71; reprinted in Thomson's Papers on Electro-
statics and Magnetism, p. 1. Also Camb. and Dub. Math. Jour. i (Nov. 1845), p. 75; reprinted
in Papers, p. 15.
I As regards this'comparison, Thomson had been anticipated by Chasles, Journal de
l' Ec. Polyt. xv (1837), p. 266, who had shown that attraction according to Newton's law
gives rise to the same fields as the steady conduction of heat, both depending on Laplace's
equation V2V = O.
It will be remembered that Ohm had used an analogy between thermal conduction
and galvanic phenomena.
science,and wassupposedby mathematiciansto be inconsistentwith

the law of electrical action, as establishedby Coulomb, and built
on by Poisson.'
In l846-the year after he had taken his degree as second
wrangler at Cambridge-Thomson investigated1 the analogies
of electric phenomena with those of elasticity. For this purpose
he examined the equations of equilibrium of an incompressible
elastic solid which is in a state of strain; and showed that the
distribution of the vector which representsthe elastic displacement
might be assimilated to the distribution of the electric force in an
electrostatic system. This, however, as he went on to show, is not
the only analogy which may be perceived with the equations of
elasticity; for the elasticdisplacementmay equally well be identified
with a vector a, defined in terms of the magneticinduction B by the
relation
curl a = B.
The vector a is equivalent to the vector-potential which had
been used in the memoirs of Neumann, Weber and Kirchhoff, on
the induction of currents; but Thomson arrived at it independently
by a different process,and without being at the time aware of the
identification.
The results of Thomson's memoir seemedto suggesta picture
of the propagation of electric or magnetic force: might it not take
place in somewhatthe sameway as changesin the elastic displace-
ment are transmitted through an elastic solid? These suggestions
were not at the time pursued further by their author; but they
helped to inspire another young Cambridge man to take up the
matter a few years later. James Clerk Maxwell, by whom the
problem was eventually solved, was born in 1831, the son of a
landed proprietor in Kirkcudbrightshire. He was educated at
Edinburgh, and at Trinity College, Cambridge, of which society
he became in 1855 a Fellow; and in 1855-6, not long after his
election to Fellowship, he communicated to the Cambridge Philo-
sophical Society the first of his endeavours2 to form a mechanical
conception of the electromagneticfield.
Maxwell had been reading Faraday's Experimental Researches;
and, gifted ashe waswith a physical imagination akin to Faraday's,
he had been profoundly impressedby the theory of lines of force.
At the same time, he was a trained mathematician; and the dis-
tinguishing feature of almost all his researcheswas the union of the
1 Camb.and Dub. Math. Jour. ii (1847), p. 61; Thomson's Math. and Phys.Papers,
i, p. 76 I Trans. Camb.Phil. Soc.x (1864),p. 27; Maxwell's Scientift Papers,i, p. 155
242
MAXWELL
imaginative and the analytical facuItiesto produce resultspartaking

of both natures. This first memoir may be regarded as an attempt
to connect the ideas of Faraday with the mathematical analogies
which had beendevisedby Thomson.l
Maxwell consideredfirst the illustration of Faraday's lines of
force which is afforded by the lines of flow of a liquid. The lines
of force represent the direction of a vector; and the magnitude
of this vector is everywhere inversely proportional to the cross-
sectionof a narrow tube formed by suchlines. This relation between
magnitude and direction is possessed by any circuital vector; and
in particular by the vector which representsthe velocity at any
point in a fluid, if the fluid be incompressible. It is thereforepossible
to representthe magneticinduction B, which is the vector represented
by Faraday's lines of magnetic force, as the velocity of an incom-
pressiblefluid. Such an analogy had been indicated some years
previously by Faraday himself,2who had suggestedthat along the
linesof magneticforcetheremay be a ' dynamic condition,' analogous
to that of the electric current, and that, in fact, , the physical lines
of magneticforce are currents.'
The comparisonwith the lines of flow of a liquid is applicable
to electric as well as to magnetic lines of force. In this casethe
vector which correspondsto the velocity of the fluid is, in free
aether, the electric force E. But when different dielectrics are
presentin the field, the electric force is not a circuital vector, and
thereforecannot be representedby Faraday's type of lines of force;
in fact, the equation
div E = 0
is now replacedby the equation .
div (€ E) = 0,
where € denotesthe specificinductive capacity or dielectric constant
at the place (x,y, z). It is, however,evident from this equation that
the vector €E is circuital; this vector, which will be denoted by
D, bearsto E a relation similar to that which the magneticinduction
B bearsto the magneticforce H. It is the vector D which is repre-
sentedby Faraday's lines of electric force, and which in the hydro-
dynamical analogy correspondsto the velocity of the incompressible
fluid.
In comparing fluid motion with electric fields it is necessaryto
introduce sourcesand sinks into the fluid to correspond to the
1 Light is thrown on the origins of Clerk Maxwell's electric ideasby a seriesof letters
written by him to Thomsonand publishedin Proc. Camb. Phil. Soc.xxxii (1936),p. 695.
• Exp. Res.,§ 3,269(1852)
(82:l) 243
electric charges; for D is not circuital at placeswhere there is free

charge. The magneticanalogy is thereforesomewhatthe simpler.
In the latter half of his memoir Maxwell discussedhow Faraday's
, electrotonic state' might be representedin mathematicalsymbols.
This problem he solvedby borrowing from Thomson'sinvestigation
of 1847the vector a, which is defined in terms of the magnetic in-
duction by the equation curl a = B ;
if, with Maxwell, we call a the electrotonic intensity, the equation

is equivalent to the statementthat' the entire electrotonicintensity
round the boundary of any surface measuresthe number of lines
of magnetic force which passthrough that surface.' The electro-
motive forceof induction at the place (x,y, z) is - oa/ot: asMaxwell
said, 'the electromotive force on any element of a conductor is
measuredby the instantaneousrate of change of the electrotonic
intensity on that element.' From this it is evident that a is no other
than the vector-potential which had been employed by Neumann,
Weber and Kirchhoff, in the calculation of induced currents;
and we may take 1 for the electrotonic intensity due to a current i'
flowing in a circuit s' the value which resultsfrom Neumann'stheory
namely, a = i' J d:'.
It may, however,be remarked that the equation
curl a = B,
taken alone, is insufficient to determine a uniquely; for we can
choosea so asto satisfythis, and alsoto satisfythe equation
diva = ifi,
where ifi denotesany arbitrary scalar. There are, therefore, an
infinite number of possiblefunctions a. With the particular value
of a which hasbeenadopted,we have
d·IV a - -0
OX
·'i
t
"r
dx'
- + -Oy0 ·'1
t
sf
dy'
--
r
0 t·'1 dz'
+ --oZ,' -
r
= - i' 1., (dx' • 0:' + dy' . ~,+ dz' . o~,)(~)

= -i' {, dO)
= 0;
so the vector-potential a which we have chosenis circuital.
1 cr. p. 200
244
MAXWELL
In this memoir the physical importance of the operators curl

and div first becameevident 1 ; for, in addition to thoseapplications
which have been mentioned, Maxwell showedthat the connection
betweenthe strength s of a current and the magnetic field H, to
which it givesrise, may be representedby the equation
%s = curl H ;
this equationis equivalent to the statementthat' the entire magnetic
intensity round the boundary of any surfacemeasuresthe quantity
of electric current which passesthrough that surface.'
In the sameyear (1856) in which Maxwell's investigation was
published, W. Thomson2 put forward an alternative interpretation
of magnetism. He had now come to the conclusion,from a study
of the magnetic rotation of the plane of polarisation of light, that
magnetismpossesses a rotatory character; and suggestedthat the
resultant angular momentum of the thermal motions of a body 3
might be taken as the measureof the magnetic moment. 'The
explanation,' he wrote, 'of all phenomena of electromagnetic
attraction or repulsion, or of electromagneticinduction, is to be
looked for simply in the inertia or pressureof the matter of which
the motions constitute heat. Whether this matter is or is not
electricity, whether it is a continuous fluid interpermeating the
spacesbetweenmolecular nuclei, or is itself molecularly grouped ;
or whether all matter is continuous,and molecular heterogeneous-
nessconsistsin finite vortical or other relative motionsof contiguous
parts of a body; it is impossibleto decide,and, perhaps,in vain to
speculate,in the presentstateof science.'
The two interpretations of magnetism,in which the linear and
rotatory charactersrespectivelyare attributed to it, occur frequently
in the subsequenthistory of the subject. The former was amplified
in 1858, when Helmholtz published his researches 4 on vortex
motion; for Helmholtz showedthat if a magneticfield producedby

electric currents is comparedto the flow of an incompressiblefluid,
sothat the magneticvector is representedby the fluid velocity, then
the electric currents correspondto the vortex filaments in the fluid.
This analogy correlates many theorems in hydrodynamics and
1 These operators had, however, occurred frequently in the writings of Stokes,
especially in his memoir of 1849 on the Dynamical Tluory of Diffraction.
• Proc. R. S. viii (1856), p. 150; xi (1861), p. 327, footnote; Phil Mag. xiii (1857),
p. 198; Baltimore Lectures, Appendix F
• This was written shortly before the kinetic theory of gaseswas developed by Clausius
and Maxwell.
• Jour. fir Math. Iv (1858), p. 25; Helmholtz' Wiss. Abh. i, p. 101; translated
Phil. Mag. xxxiii (1867), p. 485
245
electricity; for instance, the theorem that are-entrant vortex-

filament is equivalent to a uniform distribution of doubletsover any
surface bounded by it, correspondsto Ampere's theorem of the
equivalenceof electric currents and magneticshells.
In his memoir of 1855, Maxwell had not attempted to con-
struct a mechanical model of electrodynamic actions, but had
expressedhis intention of doing so. 'By a careful study,' he wrote,l
, of the laws of elasticsolids,and of the motions of viscousfluids, I
hope to discover a method of forming a mechanical conceptionof
this electrotonicstateadaptedto generalreasoning'; and in a foot-
note he referred to the effort which Thomson had already madein
this direction. Six yearselapsed,however,before anything further
on the subject was published. In the meantime, Maxwell had
become (1856-60) Professorof Natural Philosophy in Marischal
College,Aberdeen,2and then (1860-5) in King's College,London,3
In the London chair he had opportunities of personalcontact with
Faraday, whom he had long reverenced. It is true that in 1857
he wrote to Forbesthat he was' by no meansas yet a convert to
the views which Faraday maintained,' but in 1858 he wrote of
Faraday as ' the nucleusof everything electric since1830.' Faraday
had now concluded the Experimental Researches,and was living in
retirement at Hampton Court; but his thoughtsfrequently recurred
to the great problem which he had brought so near to solution.
It appears from his note book that in 1857· he was speculating
whether the velocity of propagation of magnetic action is of the
sameorder as that of light, and whether it is affected by the sus-
ceptibility to induction of the bodies through which the action is
transmitted.
1 Maxwell's Scientific Papers, i, p. 188
I His reason for leaving Trinity for the Aberdeen chair was the long summer vacation
as it then obtained in the Scottish universities, which enabled him to spend half the year
at Glenlair.
One of his students in Aberdeen, David Gill, afterwards Her Majesty's Astronomer at
the Cape of Good Hope, wrote long afterwards; • In those days a professor was little
better than a schoolmaster-and Maxwell was not a good schoolmaster; only some four
or five of us, in a class of seventy or eighty, got much out of him. We used to remain
with him for a couple of hours after lectures, till his terrible wife came and dragged him
away to a miserable dinner at three o'clock in the afternoon. By himself he was the most
delightful and sympathetic of beings--often dreaming and suddenly awakening-then he
spoke of what he had been thinking. Much of it we could not understand at the time,
and some of it we afterwards remembered and understood.'
There is good authority for the statement that the • terrible wife' wanted Maxwell to
live as a country gentleman-hunting, shooting and fishing-and that she was rude to
his scientific friends.
I In 1860 the two Colleges at Aberdeen were Wlited, and Maxwell lost his chair at
Marischal College.
• BenceJones's Life of Faraday, ii, p. 379
246
MAXWELL
The answer to this question was furnished in 1861-2, when

Maxwell fulfilled his promise of devising a mechanical conception
of the electromagneticfield.l
In the interval since the publication of his previous memoir
Maxwell had become convinced by Thomson's arguments that
magnetismis in its nature rotatory. 'The transferenceof electro-
lytes in fixed directions by the electric current, and the rotation of
polarisedlight in fixed directions by magnetic force, are,' he wrote,
'the facts the consideration of which has induced me to regard
magnetismas a phenomenonof rotation, and electric currents as
phenomenaof translation.' This conceptionof magnetismhe brought
into connectionwith Faraday'sidea, that tubesof force tend to con-
tract longitudinally and to expand laterally. Such a tendencymay
be attributed to centrifugal force, if it be assumedthat each tube of
force containsfluid which is in rotation about the axis of the tube.
Accordingly Maxwell supposedthat, in any magnetic field, the
medium is in rotation about the lines of magneticforce; each unit
tube of force may for the present-bepictured asan isolatedvortex.
The energy of the motion per unit volume is proportional to
I-'-H2,where I-'-denotes the density of the medium, and H denotes
the linear velocity at the circumferenceof each vortex. But, as we
have seen,2Thomson had already shown that the energy of any
magneticfield, whetherproducedby magnetsor by electric currents,
IS
where the integration is taken over all space,and where I-'-denotes

the magnetic permeability, and H the magnetic force. It was
thereforenatural to identify the density of the medium at any place
with the magnetic permeability, and the circumferential velocity
of the vorticeswith the magneticforce.
But an objection to the proposed analogy now presentsitself.
Since two neighbouring vortices rotate in the same direction, the
particles in the circumferenceof one vortex must be moving in the
oppositedirection to the particlescontiguousto them in the circum-
ferenceof the adjacent vortex; and it seems,therefore, as if the
motion would be discontinuous. Maxwell escapedfrom this diffi-
culty by imitating a well-known mechanical arrangement. When
it is desiredthat two wheelsshould revolve in the samesense,an
, idle' wheel is insertedbetweenthem so asto be in gear with both.
1 Phil. Mag. xxi (1861), pp. 161, 281, 338; xxiii (1862), pp. 12, 85; Maxwell's
Scientific Papers, i, p. 451 I cr. pp. 222,224
247
The model of the electromagneticfield to which Maxwell arrived

by the introduction of this device greatly resemblesthat proposed
by Bernoulli in 1736.1 He supposeda layer of particles, acting as
idle wheels, to be interposed between each vortex and the next,
and to roll without sliding on the vortices; so that each vortex
tendsto makethe neighbouringvorticesrevolvein the samedirection
as itself. The particles were supposedto be not otherwise con-
strained, so that the velocity of the centre of any particle would
be the mean of the circumferential velocitiesof the vorticesbetween
which it is placed. This condition yields (in suitable units) the
analytical equation A - -
TITS - cur1H ,
where the vector s denotesthe flux of the particles, so that its
x componentsz denotesthe quantity of particles transferredin unit
time acrossunit area perpendicular to the x direction. On com-
paring this equation with that which representsOersted'sdiscovery,
it is seenthat the flux s of the movableparticlesinterposedbetween
neighbouring vorticesis the analogueof the electric current.
It will be noticed that in Maxwell's model the relation between
electric current and magneticforce is securedby a connectionwhich
is not of a dynamical, but of a purely kinematical character. The
above equation simply expressesthe existence of certain non-
holonomic constraintswithin the system.
If from any causethe rotatory velocity of someof the cellular
vortices is altered, the disturbance will be propagated from that
part of the model to all other parts, by the mutual action of the
particles and vortices. This action is determined, as Maxwell
showed,by the relation
oH
p. N = - curl E
which connectsE, the force exertedon a unit quantity of particles

at any place in consequenceof the tangential action of the vortices,
with a::, the rate of changeof velocity of the neighbouringvortices.
It will be observedthat this equation is not kinematical but dynam-
ical. On comparing it with the electromagneticequations
curl a = p.H,
{ Induced electromotiveforce = - ~7,
it is seen that E must be interpreted electromagneticallyas the
1 cr. p. 95
248
MAXWELL
induced electromotive force. Thus the motion of the particles

constitutes an electric current, the tangential force with which
they are pressedby the matter of the vortex cellsconstituteselectro-
motive force, and the pressureof the particles on each other may
be taken to correspondto the tensionor potential of the electricity.
The mechanismmust next be extended so as to take account
of the phenomena of electrostatics. For this purpose Maxwell
assumedthat the particles, when they are displaced from their
equilibrium position in any direction, exert a tangential action on
the elastic substanceof the cells; and that this gives rise to a dis-
tortion of the cells, which in turn calls into play a force arising
from their elasticity, equal and opposite to the force which urges
the particlesaway from the equilibrium position. When the exciting
force is removed, the cells recover their form, and the electricity
returns to its former position. The state of the medium, in which
the electric particles are displacedin a definite direction, is assumed
to representan electrostatic field. Such a displacement does not
itself constitute a current, becausewhen it has attained a certain
value it remains constant; but the variations of displacementare
to be regarded as currents, in the positive or negative direction
accordingasthe displacementis increasingor diminishing.
The conception of the electrostatic state as a displacement of
something from its equilibrium position was not altogether new,
althoughit had not beenpreviouslypresentedin this form. Thomson,
aswe have seen,had comparedelectric force to the displacementin
an elastic solid; and Faraday, who had likened the particles of a
ponderabledielectric to small conductorsembeddedin an insulating
medium,l had supposedthat when the dielectric is subjectedto an
electrostaticfield, there is a displacementof electric charge on each
of the small conductors. The motion of these charges, when the
field is varied, is equivalent to an electric current; and it was
from this precedent that Maxwell derived the principle, which
became of cardinal importance in his theory, that variations of
displacementare to be counted as currents. But in adopting the
idea, he altogether transformed it; for Faraday's conception of
displacement was applicable only to ponderable dielectrics, and
was in fact introduced solely in order to explain why the specific
inductive capacity of such dielectrics is different from that of free
aether; whereasaccording to Maxwell there is displacementwher-
everthereis electricforce,whether material bodiesare presentor not.
The differencebetweenthe conceptionsof Faraday and Maxwell
1 cf. p. 187
249
in this respect may be illustrated by an analogy drawn from the

theory of magnetism. When a pieceof iron is placed in a magnetic
field, there is induced in it a magnetic distribution, say of intensity
I; this induced magnetisation exists only within the iron, being
zero in the free aether outside. The vector r may be comparedto
the polarisation or displacement, which according to Faraday is
induced in dielectrics by an electric field; and the electric current
constituted by the variation of this polarisation is then analogous
to oI/ot. But the entity which was called by Maxwell the electric
displacement in the dielectric is analogous not to I, but to the
magnetic induction B; the Maxwellian displacementcurrent corre-
spondsto oB/at, and may thereforehave a value different from zero
evenin free aether.
It may be remarked in passingthat the term displacement, which
was thus introduced, and which has been retained in the later
developmentof the theory, is perhapsnot well chosen; what in the
early modelsof the aetherwasrepresentedasan actual displacement,
was in later investigations conceived of as a change of structure
rather than of position in the elementsof the aether.
Maxwell's model is of such cardinal importance that it may be
well to recapitulate its principles, in the form in which he described
them in a letter to W. Thomson of date 10 December1861.
, I supposethat the " magnetic medium" is divided into small
portions or cells, the divisions or cell walls being composedof a
singlestratumof sphericalparticles,theseparticlesbeing" electricity."
The substanceof the cells I supposeto be highly elastic,both with
respectto compressionand distortion; and I supposethe connection
between the cells and the particles in the cell walls to be such that
there is perfect rolling without slipping between them and that
they act on eachother tangentially.
'I then find that if the cells are set in rotation, the medium
exertsa stressequivalent to a hydrostatic pressurecombined with a
longitudinal tensionalong the lines of axesof rotation.
, If there be two similar systems,the first a systemof magnets,
electric currents and bodiescapable of magnetic induction, and the
second composedof cells and cell walls, the density of the cells
everywhereproportional to the capacity for magnetic induction at
the correspondingpoint of the other, and the magnitude and direc-
tion of the cells proportional to the magneticforce, then
, I All the mechanical magnetic forces in the one systemwill be

proportional to forcesin the other arising from centrifugal force.
25°
MAXWELL
'2 All the electric currentsin the one systemwill be proportional

to currents of the particles forming the cell walls in the other.
'3 All the electromotiveforcesin the one system,whether arising
from changesof position of magnetsor currents,or from motions
of conductors,or from changesof intensity of magnetsor currents
will be proportional to forces urging the particles of the cell
walls arising from the tangential action of the rotating cells when
their velocity is increasingor diminishing.
'4 If in a non-conducting body the mutual pressureof the particles
of the cell walls (which correspondsto electric tension) diminishes
in a given direction, the particles will be urged in that direction
by their mutual pressure,but will be restrainedby their connection
with the substanceof the cells. They will, therefore, produce
strain in the cells till the elasticity called forth balances the
tendencyof the particles to move. Thus there will be a displace-
ment of particlesproportional to the electromotiveforce,and when
this forceis removed,the particleswill recoverfrom displacement.'
Maxwell supposedthe electromotive force acting on the electric
particles to be connected with the displacement D which accom-
paniesit, by an equation of the form
1
D = A_
-rrrC 1
2 E,
where C1 denotesa constant which dependson the elastic properties

of the cells. The displacementcurrent aD must now be inserted in
at
the relation which connectsthe current with the magnetic force;
and thus we obtain the equation
curl H = 47rS,
where the vector S, which is called the total current, is the sum of the
convectioncurrent's and the displacementcurrent ~~. By perform-
at
ing the operation diu on both sidesof this equation, it is seenthat
the total current is a circuital vector. In the model, the total
current is representedby the total motion of the rolling particles;
and this is conditioned by the rotations of the vortices in such a
way as to imposethe kinematic relation
div S = O.
Having obtained the equation of motion of his systemof vortices
and particles, Maxwell proceeded to determine the rate of propa-
251
gation of disturbances through it. He considered in particular

the casein which the substancerepresentedis a dielectric, so that
the conduction current is zero. If, moreover, the constant p. be
supposedto have the value unity, the equations may be written
div H = O.
C1
1
curl H = ~~,
1
Eliminating E, we see that H satisfiesthe equations
1
- curlE =-.
aH
at
divH = 0,
{ aatH
2
2
_
- C1
1r/2H
v •
But these are precisely the equations which the light vector
satisfiesin a medium in which the velocity of propagation is Cl :
it follows that disturbances are propagated through the model by
waveswhich are similar to wavesoflight, the magnetic (and similarly
the electric) vector being in the wave-front. For a plane-polarised
wave propagated parallel to the axis of z, the equations reduce to
aHy aEx aHx _ BEy aEy _ aHx aEx aHy
---az - at' az - at' - at'
I I _
- C1 Bi- = 7ft, C1 az -
whence we have
c1Hy = Ex, - c1Hx = Ey ;
these equations show that the electric and magnetic vectors are at
right anglesto each other.
The question now arises as to the magnitude of the constant
Cl.2 This may be detennined by comparing different expressions
for the energy of an electrostatic field. The work done by an
electromotive force E in producing a displacementD is
JD o
E. an or lED
per unit volume, sinceE is proportional to D. But if it be assumed
that the energy of an electrostatic field is resident in the dielectric,
the amount of energy per unit volume may be calculated by con-
For if a denoteany vector,we haveidentically
1
V2a +
grad diva curl curl a = o. +
• For criticismson the procedureby which Maxwell determinedthe velocity of
propagationof disturbance,cf. P. Duhem, us Thiories Electriquesde J. Clerk Maxwell,
Paris,1902.
MAXWELL
sidering the mechanical force required in order to increase the

distancebetweenthe platesof a condenser,soas to enlargethe field
comprisedbetweenthem. The result is that the energy per unit
volume of the dielectric is EE'2/87T, where E denotes the specific
inductive capacity of the dielectric and E' denotesthe electric force,
measuredin termsof the electrostaticunit; if E denotesthe electric
force expressedin terms of the electrodynamic units used in the
presentinvestigation,we haveE = cE', wherec denotesthe constant
which 1 occurs in transformations of this kind. The energy is
thereforeEE2/87Tc2 per unit volume. Comparingthis with the expres-
sionfor the energyin termsof E and D, we have
D = EE/4TTcl,
and therefore the constant Cl has the value ce-'. Thus the result
is obtained that the velocity of propagation of disturbances in
Maxwell's medium is cE-i, where E denotesthe specific inductive
capacityand c denotesthe velocity for which Kohlrausch and Weber
had found 2 the value 3·1 X 1010 cm./sec.
Now by this time the velocity of light was known, not only
from the astronomical observationsof aberration and of Jupiter's
satellites, but also by direct terrestrial experiments. In 1849
Hippolyte Louis Fizeau 3 had determined it by rotating a toothed
wheel so rapidly that a beam of light transmitted through the
gap between two teeth and reflected back from a mirror was
eclipsedby one of the teeth on its return journey. The velocity
of light was calculated from the dimensionsand angular velocity
of the wheel and the distance of the mirror; the result being
3·15 X 1010 cm./sec.'
1 ef. pp. 202, 232 • ef. p. 232
• ComptesRendus, xxix (1849), p. 90. A determination made by Cornu in 1874 was
on this principle.
• A different experimental method was employed in 1862 by Leon Foucault (Comptes
Rendus, Iv, pp. SOl, 792); in this a ray from an origin 0 was reflected by a revolving
mirror M to a fixed mirror, and so reflected back to M, and again to O. It is evident
that the returning ray MO must be deviated by twice the angle through which M turns
while the light passesfrom M to the fixed mirror and back. The value thus obtained by
Foucault for the velocity of light was 2·98 X 1010 cm.{sec. Subsequent determinations
by Michelson in 1879 (Ast. Papers of the Amer. Ephemeris, i), and by Newcomb in 1882
(ibid. ii) depended on the same principle.
As was shown afterwards by Lord Rayleigh (Nature, xxiv, p. 382; xxv, p. 52) and by
Gibbs (Nature, xxxiii, p. 582), the value obtained for the velocity of light by the methods
of Fizeau and Foucault represents the group-velocity, not the wave-velocity; the eclipses of
Jupiter's satellites also give the group-velocity, and so does the value deduced from the
coefficient of aberration; cf. Ehrenfest, Ann. d. Phys. xxxiii (1910), p. 1,571,and Rayleigh,
Phil. Mag. xxii (1911), p. 130. In a non-dispersive medium the group-velocity coincides
with the wave-velocity.
The velocity of light in dispersive media was directly investigated by Michelson in
1883-4, with results in accordance with theory.
253
Maxwell was impressed,as Kirchhoff had been before him, by

the close agreementbetween the electric ratio c and the velocity
of light 1; and having demonstratedthat the propagationof electric
disturbanceresemblesthat of light, he did not hesitateto assertthe
identity of the two phenomena. 'We can scarcelyavoid the infer-
ence,'he said, , that light consistsin the transverseundulationsof the
samemedium which is the causeof electric and magneticphenom-
ena.' Thus was answeredthe questionwhich Priestleyhad asked
almost exactly a hundred years before2: 'Is there any electric
fluid sui generisat all, distinct from the aether? '
The presenceof the dielectric constant ~ in the expression(E-t,
which Maxwell had obtained for the velocity of propagation of
electromagneticdisturbances,suggesteda further test of the identity
of thesedisturbanceswith light: for the velocity oflight in a medium
is known to be inversely proportional to the refractive index of the
medium, and thereforethe refractive index should be, accordingto
the theory, proportional to the squareroot of the specificinductive
capacity. At the time, however, Maxwell did not examinewhether
this relation wasconfirmed by experiment.
In what has preceded, the magnetic permeability p. has been
supposedto have the value unity. If this is not the case,the velocity
of propagation of disturbancemay be shown, by the sameanalysis,
to be CE'P.-t; so that it is diminishedwhen p. is greater than unity,
i.e. in paramagnetic bodies. This inference had been anticipated
by Faraday: 'Nor is it likely,' he wrote,S' that the paramagnetic
body oxygen can exist in the air and not retard the transmissionof
the magnetism.'
It was inevitable that a theory so novel and so capaciousasthat
of Maxwell should involve conceptionswhich his contemporaries
understoodwith difficulty and acceptedwith reluctance. Of these
the most difficult and unacceptablewas the principle that the total
current is alwaysa circuital vector; or, asit is generallyexpressed,
that ' all currents are closed.' According to the older electricians,
a current which is employedin charging a condenseris not closed,
but terminates at the coatingsof the condenser,where chargesare
accumulating. Maxwell, on the other hand, taught that the
dielectric between the coatings is the seat of a process-the dis-
placementcurrent-which is proportional to the rate of increaseof the
1 He had' worked out the formulae in the country, before seeing Weber's result.'
cf. Campbell and Garnett's Life of Maxwell, p. 244.
• Priestley's History, p. 488
• Faraday's laboratory notebook for 1857 ; cf. BenceJones's Lift of Faraday, ii, p. 380
254
MAXWELL
electric force in the dielectric; and that this processproducesthe

samemagnetic effectsas a true current, and forms, so to speak, a
continuation, through the dielectric, of the charging current, so
that the latter may be regardedasflowing in a closedcircuit.l
Another characteristic feature of Maxwell's theory is the con-
ception-for which, as we have seen,he was largely indebted to
Faraday and Thomson-that magnetic energy is the kinetic energy
of a medium occupying the whole of space,and that electric energy
is the energy of strain of the same medium. By this conception
electromagnetictheory was brought into such closeparallelism with
the elastic-solidtheoriesof the aether, that it was bound to issuein
an electromagnetictheory of light.
Maxwell's views were presentedin a more developedform in a
memoir entitled 'A Dynamical Theory of the Electromagnetic
Field,' which was read to the Royal Society in 18642; in this the
architectureof his systemwas displayed, stripped of the scaffolding
by aid of which it had beenfirst erected.
As the equations employed were for the most part the same
as had been set forth in the previous investigation, they need only
be briefly recapitulated. We shall measureall electric scalarsand
vectors in the electrostatic systemof units, and we shall measure
all magnetic scalars and vectors in the electromagnetic system.3
The magnetic induction fLH, being a circuital vector, may be
expressedin terms of a vector-potential A by the equation
fLH = curl A.
The electric displacementD is connectedwith the volume-density
pof free electric chargeby the electrostaticequation
div D = 4-rrp.
The principle of conservationof electricity yields the equation
div s = - 8p/ot,
where s denotesthe conduction current.
The law of induction of currents-namely, that the total electro-
motive force in any circuit is proportional to the rate of decrease
1 The existence of the displacement current in space free from matter or electric
charge seemedvery questionable to Maxwell's contemporaries, and its postulation was a
stumbling-block to the acceptance of his theory. In the twentieth century it was seen
that the term representing the displacement current is required in order that the equations
of the electromagnetic field should be relativistically invariant, so Maxwell's innovation
was really a relativity-correction.
• Phil. Trans, clv (1865), p. 459; Maxwell's Scient. Papers, i, p. 526
I cf. p. 202
255
of the number of lines of magnetic induction which passthrough

it-may be written 1 aH
- curl E =c I-'at ;
from which it follows that the electric force E must be expressible

in the form
loA
E = - -e --at grad '/',
./. +
where !f denotes some scalar function. The quantItIes A and !f
which occur in this equation are not asyet completelydeterminate;
for the equation by which A is defined in terms of the magnetic
induction specifiesonly the circuital part of A; and as the irrota-
tional part of A is thus indeterminate, it is evident that !f also must
be indeterminate. Maxwell decided the matter by assuming1
A to be a circuital vector; thus
div A = 0,
and therefore div E = - 'i;l2!f,
from which equation it is evident that !f representsthe electro-
static potential.
The principle which is peculiar to Maxwell's theory must now
be introduced. Currents of conduction are not the only kind of
currents; even in the older theory of Faraday, Thomson and
Mossotti, it had beenassumedthat electric chargesare setin motion
in the particles of a dielectric when the dielectric is subjectedto an
electric field; and the predecessorsof Maxwell would not have
refusedto admit that the motion of thesechargesis in somesensea
current. Suppose,then, that S denotesthe total current which is
capable of generating a magnetic field; since the integral of the
magnetic force round any curve is proportional to the electric
current which flows through the gap enclosedby the curve, we have
in the units aready specified
e curl H = 4rrS.
In order to determine S, we may consider the caseof a condenser
whose coatings are supplied with electricity by a conduction
current s per unit area of coating. If ± a denote the surface-
density of electric chargeon the coatings,we have
s = aa/at, and D = 4rra,
1 This is the effect of the introduction of (F', G', H') in § 98 of the memoir; cf. also
Maxwell's Treatise on Electricity alld Magnetism, § 616.
256
MAXWELL
where D denotesthe magnitude of the electric displacementD in

the dielectric between the coatings; so 4rrs = oD. But since the
at
total current is to be circuital, its value in the dielectric must be
the sameas the value s which it has in the rest of the circuit; that
is, thecurrent in the dielectric hasthe value~ a~. We shall assume
that the current in dielectrics always has this value, so that in the
general equations the total current must be understood to be
5 + ~ oD.
4rr at
The above equations, together with those which expressthe
proportionality of E to D in insulators, and to 5 in conductors
(D = £E when £ denotes the specific inductive capacity, and
5 = KE when K denotestheohmicconductivity), constituted Maxwell's
systemfor a field formed by isotropic bodieswhich are not in motion.
When the magnetic field is due entirely to currents (including both
conduction currents and displacement currents), so that there is
no magnetisation,we have
V'2A = - curl curl A =- curl H
4-rrS
=- c'
so that the vector-potential is connectedwith the total current by
an equation of the same form as that which connects the scalar
potential with the density of electric charge. To these potentials
Maxwell inclined to attribute a physical significance; he supposed
ifJ to be analogousto a pressuresubsistingin the massof particles
in his model, and A to be the measureof the electrotonicstate. The
two functions are, however, of merely analytical interest, and do
not correspondto physical entities. For let two oppositely charged
conductors,placed closeto each other, give rise to an electrostatic
field throughout all space. In such a field the vector-potential A is
everywherezero, while the scalar potential ifJ has a definite value
at every point. Now let these conductors discharge each other;
the electrostatic force at any point of space remains unchanged
until the point in question is reached by a wave of disturbance,
which is propagatedoutwardsfrom the conductorswith the velocity
of light, and which annihilates the field as it passesover it. But
this order of eventsis not reflected in the behaviour of Maxwell's
functions ifJ and A; for at the instant of disc:harge,ifJ is everywhere
257
annihilated, and A suddenly acquires a finite value throughout all

space.
As the potentials do not possessany physical significance, it is
desirable to remove them from the equations. This was afterwards
done by Maxwell himself, who 1 in 1868 proposed to base the
electromagnetictheory of light solely on the equations
c curl H = 4-rrS,
- c curl E = at
8B
'
together with the equationswhich define S in terms of E, and B in
terms ofH.
After tracing to the action of the surrounding medium both the
magnetic and the electric attractions and repulsions, and showing
that they are proportional to the inverse square of the distance,
Maxwell inquired 2 whether the attraction of gravitation, which
follows the samelaw of the distance, may not also be traceable to
the action of a surrounding medium.
Gravitation differs from magnetism and electricity in this:
that the bodiesconcernedare all of the samekind, instead of being
of oppositesigns,like magnetic poles or electrified bodies,and that
the force between thesebodiesis an attraction and not a repulsion,
as is the casebetween like electric and magnetic bodies. If, then,
gravitational potential energy is regarded aslocated in the rr.edium,
the energy per c.c. at any place must be
a - f:1R!
where a and f:1 are positive constantsand R is the gravitational force

per gram at the place. If we assume(as everybody in Maxwell's
day assumed)that energy is essentially positive, then the constant
a must have a value greater than f:1R 2, where R is the greatestvalue
of the gravitational force at any place in the universe: and hence
at any place where the gravitational force vanishes,the intrinsic
energy in the medium must have an enormouslygreat value. 'As
I am unable to understand in what way a medium can possess such
properties,' said Maxwell, ' I cannot go any further in this direction
in searchingfor the causeof gravitation.'
The memoir of 1864contained an extensionof the equationsto
the caseof bodies in motion; the considerationof which naturally
1 Phil. Trans. clviii (1868), p. 643; Maxwell's Scientific Papers, ii, p. 125
• In § 32 of the memoir of 1864
MAXWELL
revivesthe questionasto whether the aetheris in any degreecarried

along with a body which moves through it. Maxwell did not
formulate any expressdoctrine on this subject; but his customwas
to treat matter as if it were merely a modification of the aether,
distinguishedonly by alteredvaluesof suchconstantsasthe magnetic
permeability and the specificinductive capacity; so that his theory
may be said to involve the assumptionthat matter and aether move
together. In deriving the equationswhich are applicable to moving
bodies,he made use of Faraday's principle that the electromotive
force induced in a body dependsonly on the relative motion of the
body and the lines of magnetic force, whether one or the other is
in motion absolutely. From this principle it may be inferred that
the equation which determinesthe electric force 1 in terms of the
potentials,in the caseof a body which is moving with velocity w, is
1 1 a)l
E = C [w . ~H] - cat + grad if·
Maxwell thought that the scalar quantity if in this equation repre-
sented the electrostatic potential; but the researchesof other
investigators2 have indicated that it representsthe sum of the
electrostaticpotential and the quantity (lfc) ()l. w).
The electromagnetictheory of light was moreover extended in
this memoir so as to account for the optical properties of crystals.
For this purposeMaxwell assumedthat in crystalsthe valuesof the
coefficientsof electric and magnetic induction dependon direction,
sothat the equation
~H = curl)l
is replacedby
(~lH.~, ~IHy, ~3H:) = curl )l ;
and similarly the equation E = DIE
is replacedby E = (c1IDx, cl2Dy, csID:).
The other equationsare the sameasin isotropic media; sothat the
propagationof disturbanceis readily seento dependon the equation
alHx alHy aIH:)
( ~l at l '~I~'~S at l
= - cl curl {Cll (curl H)x, Cll (curl H)y, csl (curl H)z}.
Now, if ~1) ~I, ~s are supposedequal to each other, this equa-
1 It may be here remarked that later writers distinguished between the electric
force in a moving body and the electric force in the aether through which the body is
moving, and that E in the present equation correspondsto the former of thesevectors.
• Helmholtz, Jour. jUr Math. !xxviii (1874), p. 309; H. W. Watson, Phil. Mag. (5),
xxv (1888), p. 271
(823) 259
tion is the same as the equation of motion of MacCullagh's

aether in crystalline media,1 the magneticforce H correspondingto
MacCullagh's elastic displacement; and we may therefore im-
mediately infer that Maxwell's electromagneticequations yield a
satisfactorytheory of the propagation of light in crystals,provided
it is assumedthat the magneticpermeability is (for optical purposes)
the samein all directions,and provided the plane of polarisation is
identified with the plane which containsthe magneticvector. It is
readily shown that the direction of the ray is at right anglesto the
magneticvector and the electric force, and that the wave-frontis the
plane of the magneticvector and the electric displacement.2
Mter this Maxwell proceededto investigatethe propagation of
light in metals. The difference between metals and dielectrics, so
far as electricity is concerned,is that the former are conductors;
and it was thereforenatural to seekthe causeof the optical prop-
erties of metals in their ohmic conductivity. This idea at once
suggesteda physical reasonfor the opacity of metals-namely, that
within a metal the energy of the light vibrations is convertedinto
Joulean heat in the same way as the energy of ordinary electric
currents.
The equationsof the electromagneticfield in the metal may be
written c curI H = '"1'7T-S ,A
-ccurlE
aH
=-,
at
S=s + -aD aE
at = KE + -%c -,at
E
2
where K denotesthe ohmic conductivity; whenceit is seenthat the

electric force satisfiesthe equation
a- 2E + ,1_ aE -_ C 2t"7!E
€
at2 at
'"1'7TK- V •
This is of the sameform as the correspondingequation in the

elastic-solidtheory 3; and, like it, furnishesa satisfactorygeneral
explanation of metallic reflection. It is indeed correct in all details,
1 cr. p. 144 et sqq.
• In the memoir of 1864 Maxwell left open the choice between the above theory and
that which is obtained by assuming that in crystals the specific inductive capacity is (for
optical purposes) the same in all directions, while the magnetic permeability is aeolotropic.
In the latter case the plane of polarisation must be identified with the plane which
contains the electric displacement. Nine years later, in his Treatise (§ 794), Maxwell
definitely adopted the former alternative.
• cr. p. 163
260
MAXWELL
SO long asthe period of the disturbanceis not too short-i.e. so long

asthe light-wavesconsideredbelong to the extremeinfra-red region
of the spectrum; but if we attempt to apply the theory to the case
of ordinary light, we are confronted by the difficulty which Lord
Rayleigh indicated in the elastic-solidtheory,1 and which attends
all attemptsto explain the peculiar propertiesof metalsby inserting
a viscousterm in the equation. The difficulty is that, in order to
account for the properties of ideal silver, we must supposethe
.
coeffiClent 0f ,FE
- .
negatlve-t h'at IS, t he d'IeI'ectnc constant 0f t he
ot 2
metal must be negative,which would imply instability of electrical

equilibrium in the metal. The problem, as we have already
remarked,2was solved only when its relation to the theory of dis-
persionwasrightly understood.
At this time important developmentswere in progressin the
last-namedsubject. Sincethe time of Fresnel,theoriesof dispersion
had proceeded3 from the assumptionthat the radii of action of
the particlesof luminiferous media are so large asto be comparable
with the wave-lengthof light. It was generally supposedthat the
aether is loaded by the moleculesof ponderable matter, and that
the amount of dispersiondependson the ratio of the wave-lengthto
the distance between adjacent molecules. This hypothesis was,
however,seento be inadequate,when, in 1862, F. P. Leroux' found
that a prism filled with the vapour of iodine refracted the red
rays to a greater degree than the blue rays; for in all theories
which depend on the assumptionof a coarse-grainedluminiferous
medium, the refractive index increaseswith the frequency of the
light.
Leroux's phenomenon,to which the name anomalous dispersion
was given,5 was shown by later investigators8 to be generally
associatedwith 'surface colour,' i.e. the property of brilliantly
reflecting incident light of some particular frequency. Such an
associationseemedto indicate that the dispersiveproperty of a
substanceis intimately connectedwith a certain frequencyof vibra-
tion which is peculiar to that substance,and which, when it happens
1 cf. p. 164. cr. also Rayleigh, Phil. Mag. (5) xii (1881), p. 81, and H. A. Lorentz,
Over de Theorie de Terugkaatsing, Arnhem, 1875.
I cf. p. 164 I cr.p. 165
• ComptesRendus,lv (1862), p. 126. In 1870 C. Christiansen (AM. d. Phys. cxli, p. 479 ;
cx1iii, p. 250) obsen·ed a similar effect in a solution of fuchsin.
• Fox Talb'ot had discovered anomalous dispersion many years previously, but had
not published his work.
• Especially by Kundt, in a seriesf)f papers in the A1UIOlend. Phys., from vol. cxlii (1871)
onwards.
to fall within the limits of the visible spectrum, is apparent in the

surface colour. This idea of a frequency of vibration peculiar to
each kind of ponderable matter had already appeared in another
connection. Sir John Herschel1 first noticed that sulphateof quinine
showsa blue colour when light is incident on it under certain circum-
stances; later it was found that many other substances,if placed in
a dark room and exposedto invisible radiations beyond the violet
end of the visible spectrum,emit a bluish or greenishlight. To this
phenomenonthe namefluorescence was given by Stokesin 1852,2in a
famous paper disclosingits true nature. He remarked: 'Nothing
seemsmore natural than to supposethat the incident vibrations of
the luminiferous aether produce vibratory movements among the
ultimate moleculesof sensitivesubstances,and that the molecules
in turn, swinging on their own account, produce vibrations in the
luminiferous aether, and thus cause the sensation of light. The
periodic times of thesevibrations depend on the periods in which
the molecules are disposedto swing, not upon the periodic time
of the incident vibrations.'
The principle here introduced, of considering the moleculesas
dynamical systemswhich possessnatural free periods, and which
interact with the incident vibrations, lies at the basisof all the later
nineteenth-century theoriesof dispersion. The earliest of thesewas
devisedby Maxwell, who, in the Cambridge Mathematical Tripos
for 1869,3published the resultsof the following investigation:
A model ofa dispersivemedium may be constituted by embed-
ding systemswhich representthe atoms of ponderable matter in a
medium which representsthe aether. We may picture each atom .•
as composedof a single massiveparticle supported symmetrically
by springs from the interior face of a masslessspherical shell; if
the shell be fixed, the particle will be capableof executingvibrations
about the centre of the sphere,the effect of the springsbeing equiv-
alent to a force on the particle proportional to its distancefrom the
centre. The atoms thus constituted may be supposedto occupy
small spherical cavities in the aether, the outer shell of each atom
being in contact with the aether at all points and partaking in its
motion. An immensenumber of atoms is supposedto exist in each
unit volume of the dispersivemedium, sothat the medium asa whole
is fine-grained.
1 Phil. Trans. (1845),p. 143; Proc. R. S. v (1845),p. 547
• Phil. Trans. (1852),p. 463; Stokes'sColl. Papers, iii, p. 267
I Cambridge Calendar, 1869; republishedby Lord Rayleigh, Phil. Mag. xlviii (1899),
p.151. cr. alsoRayleigh, Phil. Mag. xxxiii (1917),p. 496

, This illustration is due to W. Thomson.
262
MAXWELL
Supposethat the potential energy of strain of free aether per

unit volume is
where 7] denotesthe displacement and E an elastic constant; so

that the equation of wave-propagationin free aether is
017] 017]
P ot' = E ox"
where p denotesthe aetherealdensity.
Then if a denotethe massof the atomic particlesin unit volume,
(7] +~) the total displacementof an atomic particle at the place
x at time t, ap2~ the attractive force, it is evident that for the cum-
pound medium the kinetic energyper unit volume is
and the potential energyper unit volume is

IE (87J/ox)' + !apI~I.
The equationsof motion, derived by the processusual in dynamics,
+ a (°at' + ot2
02~) -
are 1 1
017] 7] 0 1]
P otl E ax' = 0,
1 otl + o'~)
1
a(
0
otl + apI'
7]
= o.
Soeliminating ~ by operating with (I + p~ jj~:)on the first equation
and substitutingfrom the second,we have
(a)1 + p ?2J 0"TJ I 041] e 041]

l
+P pz oxljjtl
I
atl -- e oxl I ot4 - =0.
Either of the terms at'

0 41] 041]
or ox'jji"i would be sufficIent to produce
•
dispersion. Consider the propagation, through the medium thus

constituted,of vibrations whosefrequencyis n, and whosevelocity of
propagation in the medium is v; so that 7] and , are harmonic
functions of n(t - x/v). Substituting thesevaluesin the differential
equa.tions,we obtain
1 p ap'
V' = E E (p' - n')" +
Now, p/E has the value l/e 2, where e denotesthe velocity of light
263
in free aether; and c/v is the refractive index fL of the medium for
vibrations of frequencyn. So the equation, which may be written
ap
fLs = 1 + p (p _ n 2) ,
determinesthe refractive index of the substancefor vibrations of any

frequency n. The sameformula was independently obtained from
similar considerationsthree yearslater by W. Sellmeier.1
If the oscillationsare very slow, the incident light being in the
extremeinfra-red part of the spectrum;n is small, and the equation
gives approximately fL2 = (p + a) /p; for such oscillations, each
atomic particle and its shell move together as a rigid body, so that
the effect is the same as if the aether were simply loaded by the
massesof the atomic particles, its rigidity remaining unaltered.
The dispersionof light within the limits of the visible spectrum
is for most substancescontrolled by a natural frequency p which
correspondsto a vibration beyond the violet end of the visible
spectrum; so that, n being smaller than p, we may expand the
fraction in the formula of dispersion,and obtain the equation
2 4
fL2 = 1 + ap ( 1 + n + l'n + . .. ,
p2
)
which resemblesthe formula of dispersion in Cauchy's theory 2 ;

indeed, we may say that Cauchy's formula is the expansion of
Maxwell's formula in a series which, as it convergesonly when
n hasvalueswithin a limited range, fails to representthe phenomena
outside that range.
The theory as given above is defective in that it becomesmean-
ingless when the frequency n of the incident light is equal to the
frequency p of the free vibrations of the atoms. This defect may
be remedied by supposingthat the motion of an atomic particle
relative to the shellin which it is containedis opposedby a dissipative
force varying asthe relative velocity; sucha forcesufficesto prevent
the forced vibration from becoming indefinitely great as the period
of the incident light approachesthe period of free vibration of the
atoms; its introduction is justified by the fact that vibrations in
this part of the spectrum suffer absorption in passingthrough the
medium. When the incident vibration is not in the same region
of the spectrum as the free vibration, the absorption is not of much
importance, and may be neglected.
It is shown by the spectroscopethat the atomic systemswhich
1 Ann. d. Phys. cx1v (1872), pp. 399, 520; cx1vii (1872), pp. 386, 525. cf. also
Helmholtz, Ann. d. Phys. cliv (1875), p. 582 I cr. p. 166.
264
MAXWELL
emit and absorb radiation in actual bodies possessmore than one

distinct characteristic frequency. The theory already given may,
however, readily be extended1 to the case in which the atoms
have several natural frequencies of vibration; we have only to
supposethat the external massless rigid shell is connectedby springs
to an interior massiverigid shell, and that this again is connected
by springsto another massiveshell inside it, and so on. The corre-
spondingextensionof the equation for the refractive index is
P.
1 _ 1_
- P 11 C-n
1
2
+P CI
1 1
+ ...,
I-n
where Pt, P2 ... denote the frequenciesof the natural periods of
vibration of the atom.
The validity of the Maxwell-Sellmeier formula of dispersionwas
strikingly confirmed by experimental researchesin the closing years
of the nineteenth century.2 In 1897 Rubens 3 showed that the
formula representscloselythe refractive indice~of sylvin (potassium
chloride) and rock salt, with respect to light and radiant heat of
wave-lengthsbetween 4,240 Angstrom units and 223,000 A. The
constantsin the formula being known from this comparison,it was
possibleto predict thedispersionfor radiationsof still lowerfrequency;
and it wasfound that the squareof the refractive index should have
a negative value (indicating complete reflection) for wave-lengths
370,000 to 550,000 A. in the caseof rock salt, and for wave-lengths
450,000 to 670,000 A. in the caseof sylvin. This inference was
verified experimentally in the following year.
It may seemstrangethat Maxwell, having successfullyemployed
his electromagnetic theory to explain the propagation of light in
isotropic media, in crystals, and in metals, should have omitted to
apply it to the problem of reflection and refraction. .This is all the
more surprising, as the study of the optics of crystals had already
revealed a close analogy between the electromagnetic theory and
MacCullagh's elastic-solidtheory; and in order to explain reflection
and refraction electromagnetically,nothing more wasnecessarythan
to transcribe MacCullagh's investigation of the sameproblem, inter-
preting oe (the time-flux of the displacementof MacCullagh's aether)
ot
as the magnetic force, and curl e as the electric displacement. As
in MacCullagh's theory the differencebetweenthe contiguousmedia
1 This subjectwasdevelopedby Lord Kelvin in the"Baltimore Lectures.
I A discussionof the whole subjectin the light of sixty years'subsequentwork was
givenby C. G. Darwin, Proc. R. S. cxlvi (1934),p. 17.
I Ann. d. Phys. Ix (1897), p. 454
is representedby a difference of their elastic constants, so in the

electromagnetictheory it may be representedby a differencein their
specific inductive capacities. From a letter which Maxwell wrote
to Stokesin 1864,and which hasbeenpreserved,!it appearsthat the
problem of reflection and refraction was engaging Maxwell's atten-
tion at the time when he was preparing his Royal Society memoir
on the electromagneticfield; but he was not able to satisfyhimself
regarding the conditions which should be satisfied at the interface
between the media. He seemsto have been in doubt which of the
rival elastic-solidtheoriesto take asa pattern; and it is not unlikely
that he was led astray by relying too much on the analogy between
the electric displacement and an elastic displacement.2 For in the
elastic-solid theory all three componentsof the displacement must
be continuous acrossthe interface between two contiguous media;
but Maxwell found that it was impossibleto explain reflection and
refraction if all three componentsof the electric displacementwere
supposedto be continuous acrossthe interface; and, unwilling to
give up' the analogy which had hitherto guided him aright, yet
unable to disprove 3 the Greenian conditions at bounding surfaces,
he seemsto have laid asidethe problem until somenew light should
dawn upon it.
This was not the only difficulty which besetthe electromagnetic
theory. The theoretical conclusion, that the specific inductive
capacity of a medium should be equal to the squareof its refractive
index with respect to waves of long period, was not as yet sub-
stantiated by experiment; and the theory of displacementcurrents,
on which everything else depended, was unfavourably received by
the most distinguished of Maxwell's contemporaries. Helmholtz
indeed ultimately accepted it, but only after many years; and
W. Thomson (Kelvin) seemsnever to have thoroughly believed it to
the end of his long life. In 1888he referred to it as a ' curious and
ingenious, but not wholly tenable hypothesis,' 4 and proposed5 to
replace it by an extension of the older potential theories. In 1893
he wrote a preface for the English edition of Hertz's Electric Waves,
1 Stokes's Scientific CorrespOIldence,ii, pp. 25, 26
• It must be remembered that Maxwell pictured the electric displacement as a real

displacement of a medium. 'My theory of electrical forces,' he wrote, ' is that they are
called into play in insulating media by slight electric displacements, which put certain
small portions of the medium into a state of distortion, which, being resisted by the
elasticity of the medium, produces an electro-motive force.' (Campbell and Garnett's
Life of Maxwell, p. 244.)
• The letter to Stokes already mentioned appears to indicate that Maxwell for a time
doubted the correctness of Green's conditions.
• Nature, xxxviii (1888), p. 571 • Brit. Assoc. Report (1888), p. 567
266
MAXWELL
and there appeared to accept' magnetic waves.' But in 1896 he

. had som~inclination 1 to speculatethat alterations of electrostatic
force due to rapidly changing electrification are propagated by
condensationalwavesin the luminiferous aether; and in the same
year FitzGerald wrote to Heaviside2: 'I tried hard to knock
82e
down Lord K. with t12€ = kiL - as a sufficient answerto all his
8t2
complications as proving that every action was propagated with
velocity = (kiLtl, but I had to resort to an elaborate discussionof
the simple two spherescaseto get him to understand the electro-
magnetic point of view, and though he had nothing to reply, he is
not yetconvinced,nor doeshe, I think, evenyet understandMaxwell's
notion of displacement currents being accompanied by magnetic
force. I tried to get him to seethat his own investigation of the
penetration of alternating currents into conductors was only the
viscousmotion analogue of light propagation, but he shied at it
like a horseat a heap of stoneswhich he is accustomedin another
form to usefor riding over. He will always return to it that these
equations of light propagation are only thoseof transversevibrations
in an elastic solid, and his last giving up of everything was, I think,
partly due to my reminding him that that would require for a con-
tinuous current that the parts of the solid should be capable of
continuous rotation.' In September 1898 Oliver Lodge wrote to
Heaviside: 'Kelvin doesn'tevenbelievein Maxwell's light pressure.
He saysthat part is all wrong.' In 1904 he admitted 3 that a bar
magnetrotating about an axis at right angles to its length is equi-
valent to a lamp emitting light of period equal to the period of the
rotation, but gave his final judgment in the sentence4: 'The so-
called electromagnetictheory of light has not helped us hitherto.'
Thomson appearsto have basedhis ideas of the propagation of
electric disturbance on the casewhich had first becomefamiliar to
him-that of the transmissionof signalsalong a wire. He clung to
the older view that in such a disturbance the wire is the actual
medium of transmission; whereasin Maxwell's theory the function
of the wire is merely to guide the disturbance, which is resident in
the surrounding dielectric.
This opinion that conductors are the media of propagation of
electric disturbancewas entertained also by Ludwig Lorenz (1829-
1891) of Copenhagen,who independently developed an electro-
1 cf. Bottomley,in Nature, liii (1896),p. 268; Kelvin, ibid., p. 316; J. Willard Gibbs,
ibid., p. 509
I Letter of II June 1896 I Baltimore LeetUTes,
p. 376 • ibid., preface,p. 7
267
magnetic theory of light 1 a few years after the publication of

Maxwell's memoirs. The procedurewhich Lorenz followed wasthat
which Riemann had suggested2 in 1858-namely, to modify the
acceptedformulae of electrodynamicsby introducing terms which,
though too small to be appreciable in ordinary laboratory experi-
ments, would be capable of accounting for the propagation of
electrical effectsthrough spacewith a finite velocity. We have seen
that in Neumann'stheory the electric force E wasdeterminedby the
equation 1 aa (1)
E = grad 1> at' -c
where 1> denotesthe electrostaticpotential definedby the equation
1> = IH (p'/r) dx'd;Y'dz',
p' being the density of electric charge at the point (x', y', z'), and
where a denotesthe vector-potential, defined by the equation
a= HH (s'/r) dx'dy'dz',
s' being the conduction current at (x', y', z'). We supposethe

specific inductive capacity and the magnetic permeability to be
everywhereunity.
Lorenz proposedto replace theseby the equations
1> = IH {p'(t - r/e)/r} dx'dy'dz',
a = ~HI {s'(t - rle)/r} dx'dy'dz' ;
the change consistsin replacing the values which p' and s' have
at the instant t by those which they have at the instant (t - ric),
which is the instant at which a disturbance travelling with velocity
c must leave the place (x', y', z') in order to arrive at the place
(x,y, z) at the instant t. Thus the valuesof the potentialsat (x,y, z)
at any instant t would, according to Lorenz's theory, dependon the
electric stateat the point (x',y', z') at the previousinstant (t - ric) :
as if the potentials were propagatedoutwards from the chargesand
currents with velocity c. The functions 1> and a formed in this way
are generally known as the retardedpotentials.
1 Oversigt over det K. danske Vid. Selskaps Forhandlinger (1867), p. 26; Ann. d. Phys.
cxxxi (1867), p. 243; Phil. Mag. xxxiv (1867), p. 287
2 cf. p. 240. Riemann'smemoir was,however,publishedonly in the sameyear (1867)
asLorenz's.
268
MAXWELL
The equations by which e/> and a have been defined are equi-
valentto the equations 1 a2e/> (2)
VIe/> - - l - l = - %p
c at '
1 ala (3)
= - %s/c,
VIa - --
c2 atl
while the equation of conservationof electricity,
gives1
div s + =0 :r
d'Iva + -~
1 ae/> 0
= . (4)
c ut
From equations(1), (2), (4), we may readily derive the equation
div E = %p ; (I)
and from (1), (3), (4), we have
1 oE· %s
cur1 H=--+- (II)
c at c '
where H or curl a denotesthe magnetic force; while from (1) we
have aH
c curl E = - at" (III)
1 For we have
~- oaz_ 0 fffsz' -
ox -- ox r
the' d!>' ' dz'
= fH{sz' ~G) + ~a~;'~ (-D} the' dy' dz'
=- fff{ , ox' (1)

Sz _. 0
r
+ 1osz' 0 ( r) jI. the' d!>' ' dz.'
-
r ot ox'
~
c - - - -
Integrating the first of these integrals by parts, we have
coax
Ox
= fff'!' {o:.::.'- oSz'~ (- !:)} the' dy' tk,'
r ox' ot ox' c
= JH t [:::] the' dy' tk,'
where the square brackets imply that the differentiation is to be taken partially with
respect to x' only in so far as it occurs explicitly, ignoring its implicit occurrence in the
argument(t - ~).
Hf~{[~::J+ [~~] + [~:~']}

We have therefore
~diva = the' dy' tk,'
- - Jf f ~~; dx' dy' dz', by the equation of conservation of electricity
- - ~, which is the required equation (4).
269
The equations (I), (II), (III) are, however, the fundamental

equationsof Maxwell's theory; and thereforethe theory ofL. Lorenz
is practically equivalent to that of Maxwell, so far as concernsthe
propagation of electromagnetic disturbances through free aether.
Lorenz himself, however, doesnot appear to have clearly perceived
this; for in his memoir he postulated the presenceof conducting
matter throughout space, and was consequently led to equations
resembling those which Maxwell had given for the propagation of
light in metals. Observing that his equations representedperiodic
electric currents at right angles to the direction of propagation of
the disturbance, he suggestedthat all luminous vibrations might be
constituted by electric currents, and hencethat there was' no longer
any reason for maintaining the hypothesisof an aether, since we
can admit that spacecontainssufficient ponderablematter to enable
the disturbance to be propagated.'
Lorenz was unable to derive from his equationsany explanation
of the existenceof refractive indices, and his theory lacks the rich
physical suggestiveness of Maxwell's; the value of his memoir lies
chiefly in the introduction of the retarded potentials. It may be
remarked in passing that Lorenz's retarded potentials are not
identical with Maxwell's scalar-and vector-potentials; for Lorenz's
a is not a circuital vector, and Lorenz's 4> is not, like Maxwell's, the
electrostaticpotential, but dependson the positionsoccupiedby the
chargesat certain previous instants.
For some years no progresswas made either with Maxwell's
theory or with Lorenz's. Meanwhile, Maxwell had in 1865resigned
his chair at King's College,and had retired to his estateof Glenlair
in Kirkcudbrightshire, where he occupied himself in writing a
connected account of electrIcal theory. In 1871 he returned to
Cambridge as Professorof Experimental Physics; and two years
later published his Treatise on Electrici~y and Magnetism.
In this celebratedwork is comprehendedalmost every branch of
electric and magnetic theory; but the intention of the writer was
to discussthe whole as far as possiblefrom a single point of view,
namely, that of Faraday; so that little or no account was given of
the hypotheseswhich had been propounded in the two preceding
decadesby the great German electricians. So far as Maxwell's
purposewasto disseminatethe ideasof Faraday, it wasundoubtedly
fulfilled; but the Treatise was lesssuccessfulwhen consideredasthe
exposition of its author's own views. The doctrines peculiar to
Maxwell-the existenceof displacement currents, and of electro-
magnetic vibrations identical with light-were not introduced in
270
MAXWELL
the first volume, or in the first half of the secondvolume; and the
accountwhich wasgiven of them was scarcelymore complete, and
wasperhapslessattractive, than that which had been furnished in
the original memoirs.
Somematterswere,however,discussedmore fully in the Treatise
than in Maxwell's previous writings; and among these was the
questionof stressin the electromagneticfield.
It will be remembered1 that Faraday, when studying the
curvature of lines of force in electrostatic fields, had noticed an
apparent tendencyof adjacent lines to repel each other, as if each
tube of force were inherently disposedto distend laterally; and
that in addition to this repellent or diverging force in the transverse
direction,he supposedan attractive or contractile forceto be exerted
at right anglesto it, that is to say, in the direction of the lines of
force.
Of the existenceof thesepressuresand tensions Maxwell was
fully persuaded; and he determinedanalytical expressionssuitable
to representthem. The tension along the lines of force must be
supposedto maintain the ponderomotive force which acts on the
conductor on which the lines of force terminate; and it may
thereforebe measured(in the systemof units w,eare now using) by
the force which is exertedon unit area of the conductor, i.e. EEz/87T
or DE/87T. The pressureat right anglesto the lines of force must
then be determinedso as to satisfythe condition that the aether is
to be in equilibrium.
For this purpose,considera thin shell of aetherincluded between
two equipotential surfaces. The equilibrium of the portion of this
shell which is intercepted by a tube of force requires (as in the
theory of the equilibrium ofliquid films) that the resultant force per
unit area due to the above-mentionednormal tensionson its two
facesshall have the value T(l/pt + l/pz), where Pt and pz denote
the principal radii of curvature of the shell at the place, and where
T denotesthe lateral stressacrossunit length of the surfaceof the
shell,T being analogousto the surface-tensionof a liquid film.
Now, if t denotethe thicknessof the shell, the area intercepted
on the secondfaceby the tube of force bearsto the area intercepted
on the first face the ratio (Pl + t) (pz + t) /PlPZ; and by the funda-
mental property of tubes of force, D and E vary inversely as the
cross-sectionof the tube, so the total force on the secondface will
bear to that on the first face the ratio
PtPl/(Pl + t) (PI + t),
1 cr. p. 187
271
or approximately
(l-!"PI _i);
Ps
the resultant force per unit area along the outward normal IS
therefore I
- 817DE . t .(l/pI lips), +
and so we have I
T = - 817DE . t ;
or the pressureat right anglesto the linesof forceis I/817 DE per unit
area-that is, it is numerically equal to the tension along the lines
of force.
The principal stressesin the medium being thus determined,it
readily follows that the stressacrossany plane, to which the unit
vector N is normal, is 1
I I
4-; (D • N)E - 87T(D • E)N.
Maxwell obtained 2 a similar formula for the caseof magnetic

fields; the ponderomotive forces on magnetisedmatter and on
conductorscarrying currents may be accountedfor by assuminga
stressin the medium, the stressacrossthe plane N being represented
by the vector I I
47T(B • N)H - 817(B • H)N.
This, like the corresponding electrostatic formula, representsa

tension acrossplanes perpendicular to the lines of force, and a
pressureacrossplanesparallel to them.
It may be remarked that Maxwell made no distinction between
stressin the material dielectric and stressin the aether: indeed,so
long as it was supposedthat material bodieswhen displacedcarry
the contained aether along with them, no distinction was possible.
In the modifications of Maxwell's theory which were developed
many years afterwards by his followers, stressescorrespondingto
thoseintroduced by Maxwell were assignedto the aether,asdistinct
from ponderablematter; and it was assumedthat the'only stresses
set up in material bodiesby the electromagnetic'field are produced
indirectly; they may be calculated by the methodsof the theory of
elasticity, from a knowledgeof the ponderomotiveforcesexertedon
the electric chargesconnectedwith the bodies.
Another remark suggestedby Maxwell's theory of stressin the
1 cr. W. H. Bragg, Phil. Mag. xxxiv (1892), p. 18
• Maxwell's Treatise on Electricity and ltilagnetism, § 643

272
MAXWELL
medium is that he consideredthe question from the purely statical

point of view. He determinedthe stresssothat it might produce the
required forces on ponderable bodies, and be self-equilibrating in
free aether. But 1 if the electric and magnetic phenomenaare not
really statical, but are kinetic in their nature, the stressor pressure
neednot be self-equilibrating. This may be illustrated by reference
to the hydrodynamical modelsof the aether shortly to be described,
in which perforated solids are immersed in a moving liquid: the
ponderomotiveforcesexerted on the solidsby the liquid correspond
to those which act on conductors carrying currents in a magnetic
field, and yet there is no stressin the medium beyond the pressure
of the liquid.
Among the problems to which Maxwell applied his theory of
stressin the medium was one which had engagedthe attention of
many generationsof his predecessors. The adherents of the cor-
pusculartheory of light in the eighteenthcentury believedthat their
hypothesiswould be decisively confirmed if it could be shown that
rays of light possess momentum: to determine the matter, several
investigatorsdirectedpowerful beamsof light on delicatelysuspended
bodies,and looked for evidencesof a pressuredue to the impulse of
the corpuscles. Such an experiment was performed in 1708 by
Homberg,2 who imagined that he actually obtained the effect in
question; but Mairan and Du Fay in the middle of the century,
having repeatedhis opemtions,failed to confirm his conclusion.S
The subject was afterwards taken up by Michell, who' some
yearsago,' wrote Priestley' in 1772,' endeavouredto ascertainthe
momentumof light in a much more accurate manner than thosein
which M. Homberg and M. Mairan had attempted it.' He exposed
a very thin and delicately suspendedcopper plate to the rays of the
sun concentratedby a mirror, and observeda deflection. He was
not satisfied that the effect of the heating of the air had been
altogetherexcluded, but' there seemsto be no doubt,' in Priestley's
opinion, ' but that the motion above mentioned is to be ascribedto
the impulse of the rays oflight.' A similar experiment wasmade by
A. Bennet,5who directed the light from the focusof a large lens on
writing-paper delicately suspendedin an exhausted receiver, but
, could not perceiveany motion distinguishablefrom the effects of
heat. Perhaps,' he concluded, 'sensible heat and light may not
be causedby the"influx or rectilineal projections of fine particles,
1 cr. V.
Bjerknes, Phil. Mag. ix (1905), p. 491 • Histoire de l'Acad. (1708), p. 21
• J. J. de Mairan, Trait! de l'Aurore borlale, p. 370 ' History of Vision, i, p. 387
• Phil. Trans. 1xxxii (1792), p. 81
273
but by the vibrations made in .the universaly diffused caloric or

matter of heat,or fluid oflight.' Thus Bennet,and after him Young,!
regardedthe non-appearanceof light repulsionin this experimentas
an argument in favour of the undulatory systemof light. 'For,'
wrote Young, 'granting the utmost imaginable subtility of the
corpusclesof light, their effectsmight naturally be expectedto bear
someproportion to the effectsof the much lessrapid motionsof the
electrical fluid, which are so very easily perceptible, even in their
weakeststates.'
This attitude is all the more remarkable, becauseEuler many
yearsbefore had expressedthe opinion that light pressuremight be
expectedjust asreasonablyon the undulatory ason the corpuscular
hypothesis. 'Just as,' he wrote,2 'a vehement sound excitesnot
only a vibratory motion in the particles of the air, but there is also
observeda real movement of the small particles of dust which are
suspendedtherein, it is not to be doubted but that the vibratory
motion set up by the light causesa similar effect.' Euler not only
inferred the existence of light pressure, but even (adopting a
suggestionof Kepler's) accountedfor the tails of cometsby suppos-
ing that the solar rays, impinging on the atmosphereof a comet,
drive off from it the more subtle of its particles.
The questionwas examinedby Maxwell 3 from the point of view
of the electromagnetic theory of light, which readily furnishes
reasonsfor the existenceof light pressure. For supposethat light
falls on a metallic reflecting surfaceat perpendicularincidence. The
light may beregardedasconstitutedof a rapidly alternating magnetic
field; and this must induce electric currentsin the surfacelayersof
the metal. But a metal carrying currents in a magnetic field is
acted on by a ponderomotiveforce, which is at right anglesto both
the magneticforce and the direction of the current, and is therefore,
in the presentcase,normal to the reflecting surface: this pondero-
motive force is the light pressure. Thus, according to Maxwell's
theory, light pressureis only an extendedcaseof effectswhich may
readily be produced in the laboratory.
The magnitude of the light pressurewas deduced by Maxwell
from his theory of stressesin the medium. We have seenthat the
stressacrossa plane whoseunit-normal is N is representedby the
vector
1 1
41r (D • N)E - 8':;: (D • E)N + 41r1 (B • N)H 1
- 877 (B • H)N.
1 ibid. xcii (1802), p. 12 I Histoire de l'Acad. de Berlin, ii (1748), p. 117
• Maxwell's Treatise on Electricity and Magnetism, § 792
274
MAXWELL
Kow, supposethat a plane wave is incident perpendicularly on a

perfectly reflecting metallic sheet: this sheet must support the
mechanicalstresswhich existsat its boundary in the aether. Owing
to the presenceof the reflectedwave,D is zero at the surface; and
D is perpendicular to N, so (D. N) vanishes. Thus the stressis a
pressureof magnitude 1j87T (D. H) normal to the surface: that is,
the light pressureis equal to the density of the aetherealenergy in
the region immediately outside the metal. This was Maxwell's
result.
This conclusionhas been reached on the assumption that the
light is incident normally to the reflecting surface. If, on the other
hand, the surfaceis placed in an enclosurecompletely surrounded
by a radiating shell, so that radiation falls on it from all directions,
it may be shown that the light pressureis measuredby one-third
of the densityof aetherealenergy.
A different way of inferring the necessityfor light pressurewas
indicated in 1876 by Adolfo Bartolil (1851-96) of Bologna, who
showedthat, when radiant energy is transported from a cold body
to a hot oneby meansof a moving mirror, the secondlaw of thermo-
dynamicswould be violated unlessa pressurewere exerted on the
mirror by the light.
In the year in which Maxwell's treatise was published, Sir
William Crookes2 obtained experimental evidence of a pressure
accompanyingthe incidence of light; but this was soon found to
be due to thermal effects; and the existenceof a true light pressure
was not confirmed experimentally 3 until 1899. Since then the
subjecthasbeenconsiderablydeveloped,especiallyin regard to the
part played by the pressureof radiation in cosmical physics. In
connectionwith radiation pressure,it may be mentioned that there
is a pressureof radiation againstthe source-a recoil producedby the
emissionoflight. This wasinvestigatedin' 1909by J.H. Poyntingand
Guy Barlow,4who employedthin slipsof material which wereheated
by incident radiation. Slips blackenedon both sidesexperiencea
pressureequal to the energy-densityP of the incident radiation, since
the emitted radiation from the front and back have equal recoil
pressures. Slips blackenedon the incident sideand brightly silvered
1 Bartoli, Sopra i movimenti prodotli daUa luce e dal calore e sopra il radiometro di Crookes,
Firenze, 1876. Also Nuovo Cimento('), xv (1884),p. 193,and Exner's Rep. 'xxi (1885),
p.198.
• Phil. Trans. clxiv (1874),p. 501. The radiometerwasdiscoveredin 1875.
, P. Lebedeff,Archives des SciencesPhys. et Nat. (4), viii (1899),p. 184; Ann. d. Phys.
vi (1901),p. 433; E. F. Nicholsand G. F. Hull, Phys. Rev. xiii (1901),p. 293; Astrop/!ys.
Jour. xvii (1903),p. 315; W. Gerlachand Alice Golsen,Zs.j. Phys. xv (1923),p. 1
• Brit. Ass. Rep. (1909),p. 385
(823) 275
on the other experiencea pressureiP, the excessbeing due to the

radiation which is emitted according to the cosinelaw by the front
sideonly of theplate. Plateswhich arebright on both sidesexperience
a pressure2P, since they do not become heated, but reflect the
radiation.
Another matter which received attention in Maxwell's Treatise
Wagthe influence of a magnetic field on the propagation of light
in material substances. We have already seen1 that the theory of
magneticvorticeshad its origin in W. Thomson'sspeculationson this
phenomenon; and Maxwell in his memoir of 1861-2had attempted
by the help of that theory to arrive at someexplanation of it. The
more complete investigation which is given in the Treatise is based
on the samegeneralassumptions,namely,that in a medium subjected
to a magnetic field there exist concealedvortical motions, the axes
of the vortices being in the direction of the lines of magneticforce ;
and that waves of light passing through the medium disturb the
vortices,which thereuponreact dynamically on the luminousmotion,
and soaffect its velocity of propagation.
The manner of this dynamical interaction must now be more
closely examined. Maxwell supposedthat the magnetic vortices
are affected by the light wavesin the sameway as vortex filaments
in a liquid would be affected by any other coexistingmotion in the
liquid. The latter problem had beenalreadydiscussedin HelmholtZ'
great memoir on vortex motion; adopting Helmholtz' results,
Maxwell assumed for the additional term introduced into the
magnetic force by the displacementof the vortices the value oe/oB,
where e denotes the displacement of the medium (i.e. the light
vector), and the operator %() denotesHx%x+ Hy%y + Hz%z,
H denoting the imposedmagneticfield. Thus the luminous motion,
by disturbing the vortices, gives rise to an electric current in the
medium, proportional to curl oe/oB.
Maxwell further assumedthat the current thus produced inter-
acts dynamically with the luminous motion in such a manner that
the kinetic energy of the medium contains a term proportional to
the scalar product of Oe and curloe/oB. The total kinetic energy
at
of the medium may therefore be written
!p (~~r+
t U (~~ • curl oe/oo)),
wherep denotesthe densityof the medium, and u denotesa constant
1 cr. p. 245
276
MAXWELL
which measuresthe capacity of the medium to rotate the plane of

polarisationof light in a magneticfield.
The equation of motion may now be derived as in the elastic-
solid theoriesof light: it is
ole 0'
p ot2 = nV2e - (T ot88 curl e.
When the light is transmitted in the direction of the lines of

force, and the axisof x is taken parallel to this direction, the equation
reducesto { a2e" _ o'e" 03ez
p otl - n ox' + aH 8t8x2t
ole: ole: 83e" •
P ifiZ = n ox' - aH atox' ,
and theseequations, as we have seen,! furnish an explanation of

Faraday'sphenomenon.
It may be remarked that the term
ia (~; . curl oe/(8)

in the kinetic energy may by partial integration be transformed
into a term
!a (curl e • oe/(8), 2
togetherwith surfaceterms; or, again, into
- !a ( curl e • ~~e8)'
together with surface terms. These different forms all yield the
sameequation of motion for the medium; but, owing to the dif-
ferencesin the surfaceterms, they yield different conditions at the
boundary of the medium, and consequentlygive rise to different
theoriesof reflection.
The assumptionsinvolved in Maxwell's treatment of the magnetic
rotation of light were such as might scarcely be justified in them-
selves; but since the discussionas a whole proceededfrom sound
dynamical principles, and its conclusionswere in harmony with
experimentalresults,it wasfitted to lead to the more perfectexplana-
tions which were afterwardsdevisedby his successors.At the time
of Maxwell's death, which happenedin 1879,before he had com-
I cr. p. 192
• This form was suggestedby FitzGerald six yearslater, Phil. Trans. clxxi (1880),p.
691 ; FitzGerald'sScientift Writings,p. 45
277
pleted his forty-ninth year, much yet remained to be done both

in this and in the other investigations with which his name is
associated; and the energiesof the next generation were largely
spent in extending and refining that conception of electrical and
optical phenomena whoseorigin is correctly indicated in its name
of Maxwell's Theory.
Chapter IX
MODELS OF THE AETHER
THE early attempts of W. Thomson and Maxwell to representthe

electric medium by mechanical models opened up a new field of
research, to which investigators were attracted as much by its
intrinsic fascination as by the importance of the serviceswhich it
promised to render to physical theory. There seemed to be a
possibility of accounting for electric, magnetic and gravitational
forcesby the action of an intervening aether.
Of the modelsto which referencehas already been made, some
-such as those described in Thomson's memoir 1 of 1847 and
Maxwell's memoir 2 of 1861-2-attribute a linear character to
electric force and electric current, and a rotatory character to
magnetism; others-such asthat devisedby Maxwell in 18553 and
afterwards amplified by Helmholtz 4-regard magnetic force as a
linear, and electric current asa rotatory phenomenon. This distinc-
tion furnishesa natural classificationof models into two principal
groups.
Even within the limits of the former group diversity has already
become apparent; for in Maxwell's analogy of 1861-2, a con-
tinuous vortical motion is supposedto be in progressabout the lines
of magnetic induction; whereasin Thomson's analogy the vector-
potential was likened to the displacement in an elastic solid, so
that the magnetic induction at any point would be represented
by the twist of an elementof volume of the solid from its equilibrium
position; or, in symbols,
1 oe
a = e, E =- cat' B = curl e,
where a denotesthe vector-potential, E the electric force, B the
magneticinduction and e the elastic displacement.
Thomson's original memoir concluded with a notice of his
intention to resumethe discussionin another communication. His
purpose was fulfilled only in 1890, when 6 he showed that in his
model a linear current could be representedby a piece of endless
1 cr. p. 242 • cr. p. 247 • cr. p. 242 • cr. p. 245
• Kelvin's Math. and Phys. Papers, iii, p. 436
279
cord, of the same quality as the solid and embedded in it, if a

tangential force were applied to the cord uniformly all round the
circuit. The forces so applied tangentially produce a tangential
drag on the surrounding solid; and the rotatory displacementthus
causedis everywhereproportional to the magneticvector.
In order to representthe effectof varying permeability, Thomson
abandonedthe ordinary type of elastic solid, and replaced it by an
aether of MacCullagh's type; that is to say,an ideal incompressible
substance,having no rigidity of the ordinary kind (i.e. elastic resis-
tance to changeof shape),but capableof resistingabsoluterotation
-a property to which the name gyrostatic rigidifY was given. The
rotation of the solid representing the magnetic induction, and the
coefficient of gyrostatic rigidity being inversely proportional to the
permeability, the normal component of magnetic induction will be
continuousacrossan interface, as it should be.l
We have seen above that in models of this kind the electric
force is represented by the translatory velocity of the medium.
It might therefore be expected that a strong electric field would
perceptibly effect the velocity of propagation of light; and that
this does not appear to be the case,2 is an argument against the
validity of the scheme.
We now turn to the alternative conception, in which electric
phenomenaare regarded as rotatory, and magnetic force is repre-
sented by the linear velocity of the medium; in symbols,
~D = curl e,
{ H= oe
ot'
whereD denotesthe electricdisplacement,H the magneticforceand e
the displacementof the medium. In Maxwell's memoir of 1855,
and in most of the succeedingwritings for many years,attention was
directed chiefly to magnetic fields of a steady, or at any rate non-
oscillatory, character; in such fields, the motion of the particles of
the medium is continuously progressive; and it was consequently
natural to supposethe medium to be fluid.
Maxwell himself, as we have seen,3afterwards abandonedthis
conception in favour of that which representsmagnetic phenomena
as rotatory. 'According to Ampere and all his followers,' he wrote
1 Thomson inclined to believe (Papers, iii, p. 465) that light might be correctly
representedby the vibratory motion of sucha solid.

I Wilberforce,Trans. Camb.Phil. Soc.xiv (1887),p. 170; Lodge,Phil. Trans.clxxxix
(1897),p. 149 I cr. p. 247
280
in 1870,1 electric currents are regarded as a speciesof translation,

C
and magneticforce as dependingon rotation. I am constrainedto

agreewith this view, becausethe electric current is associatedwith
electrolysis,and other undoubted instances of translation, while
magnetismis associatedwith the rotation of the plane of polarisation
of light.' But the other analogy was felt to be too valuable to be
altogether discarded,especiallywhen in 1858Helmholtz extended
it 2 by showing that if magnetic induction is compared to fluid
velocity, then electric currents correspondto vortex filaments in the
fluid. Two years afterwards Kirchhoff 3 developedit further. If
the analogyhasany dynamical (asopposedto a merely kinematical)
value, it is evident that the ponderomotiveforcesbetweenmetallic
rings carrying electric currents should be similar to the pondero-
motive forcesbetweenthe samerings when they are immersed in
an infinite incompressiblefluid; the motion of the fluid being
such that its circulation through the aperture of each ring is pro-
portional to the strengthof the electric current in the corresponding
ring. In order to decide the question, Kirchhoff attempted, and
solved,the hydrodynamicalproblem of the motion of two thin, rigid
rings in an incompressiblefrictionlessfluid, the fluid motion being
irrotational; and found that the forces between the rings are
numerically equal to those which the rings would exert on each
other if they were traversedby electric currents proportional to the
circulations.
There is, however, an important difference between the two
cases,which was subsequentlydiscussedby W. Thomson, who
pursuedthe analogyin severalmemoirs.' In order to representthe
magneticfield by a conservativedynamical system,we shall suppose
that it is produced by a number of rings of perfectly conducting
material, in which electriccurrentsare circulating; the surrounding
medium beingfree aether. Now any perfectly conducting body acts
as an impenetrable barrier to lines of magnetic force; for, as
Maxwell showed,5when a perfectconductor is placedin a magnetic
field, electric currents are induced on its surfacein such a way as
to makethe total magn~ticforce zero throughou,tthe interio! of the
.conductor.6 Lines of force are thus deflected by the body in the
, Proc. Land. Math. Soc. iii (1870), p. 224; MaxweII's Scientific Papers, ii, p. 263
• cf. p. 245
• Berlin Monatsb. (1869), p. 881 ; .Joumalfiir Math. 1xxi (1869), p. 263; Kirchhoff's
Gesamm.Abhandl., p. 404. cr. also C. Neumann, Leipzig Berichte, xliv (1892), p. 86.
• Thomson's Papers in Elect. QlzdMag., §§ 573, 733, 751 (1870-2)
• Maxwell's Treatise on Elect. and Mag., § 654
• For this reason W. Thomson caIIed a perfect conductor an ideal extremediamagnlti& •
.- 281
sameway as the lines of flow of an incompressiblefluid would be

deflectedby an obstacleof the sameform, or as the lines of flow of
electric current in a uniform conducting masswould be deflectedby
the introduction of a body of this form and of infinite resistance. If,
then, for simplicity we consider two perfectly conducting rings
carrying currents, thoselines of force which are initially linked with
a ring cannot escapefrom their entanglement,and new linescannot
becomeinvolved in it. This implies that the total number of lines
of magnetic force which passthrough the aperture of each ring is
invariable. If the coefficientsof self and mutual induction of the
rings are denoted by L1, ~, Ln, the electrokinetic energy of the
systemmay be representedby
T = i (L1q11 + 2 L 12 qlq2 + L2q21),

where ql, q2 denotethe strengthsof the currents; and the condition
that the number of lines of force linked with each circuit is to be
invariable givesthe equations
L1ql + L q. = constant,
12
L12ql + L2q2 = constant.

It is evident that, when the systemis consideredfrom the point
of view of generaldynamics,the electric currents must be regarded
asgeneralisedvelocities,and the quantities
(L1ql +L 12 q2) and (L12ql + L.q.)

asmomenta. The electromagneticponderomotiveforceon the rings
tending to increaseany co-ordinate x is aT lax. In the analogous
hydrodynamical system,the fluid velocity correspondsto the mag-
netic force; and thereforethe circulation through eachring (which
is defined to be the integral Jvds, taken round a path linked once
with the ring) correspondskinematically to the electric current;
and the flux of fluid through each ring correspondsto the number
of lines of magnetic force which passthrough the aperture of the
ring. But in the hydrodynarnical problem the circulationsplay the
part of generalisedmomenta; while the fluxes of fluid through
the rings play the part of generalisedvelocities. The kinetic energy
may indeed be expressedin the form
K = ! (N1Kll + 2N12KIKI + NaKII),
where K1) K2, denote the circulations (so that K1 and K2 are pro-
portional respectivelyto ql and q2), and N1> Nn, N2 dependon the
282
positionsof the rings; but this is the Hamiltonian (as opposedto

the Lagrangian) form of the energy-function,! and the pondero-
motive force on the rings tending to increaseany co-ordinate x is
- oK/ox. SinceoK/ox is equal to aT/ox, we seethat the pondero-
motive forceson the rings in any position in the hydrodynamical
systemare equal, but opposite, to the ponderomotive forces on the
rings in the electric system.
The reasonfor the differencebetweenthe two casesmay readily
be understood. The rings cannot cut through the lines of magnetic
force in the one system,but they can cut through the stream line:>
in the other; consequentlythe flux of fluid through the rings is not
invariable when the rings are moved, the invariants in the hydro-
dynamical systembeing the circulations. If a thin ring, for which
the circulation is zero, is introduced into the fluid, it will experience
no ponderomotive forces; but if a ring initially carrying no
current is introduced into a magnetic field, it will experience
ponderomotiveforces,owing to the electric currentsinduced in it by
its motion.
Imperfect though the analogy is, it is not without interest. A
bar magnet, being equivalent to a current circulating in a wire
wound round it, may be compared(as W. Thomsonremarked) to a
straight tube immersedin a perfect fluid, the fluid entering at one
end and flowing out by the other, sothat the particles of fluid follow
the lines of magnetic force. If two such tubes are presentedwith
like endsto each other, they attract; with unlike ends,they repel.
The forces are thus diametrically oppositein direction to those of
magnets; but in other respectsthe laws of mutual action between
thesetubesand betweenmagnetsare preciselythe same.I
1 cf. Whittaker, Ana(ytical Dynamics, § 109
• The mathematical analysis in this case is very simple. A narrow tube through
which water is flowing may be regarded as equivalent to a source at one end of the tube
and a sink at the other; and the problem may therefore be reduced to the consideration
of sinksin an unlimi ted fluid. If there are two sinks in such a fluid, of strengths m and m',
the velocity-potential is
ml' + m'I",
where, and " denote distance from the sinks. The kinetic energy per unit volume of
the fluid is
i{{~(~+-'i:)}" + t~(~+ ~:)r+ {~(~+ ?')}"].

where p denotes the density of the fluid; whence it is easily seen that the total energy of
the fluid, when the two sinks are at a distance I apart, exceedsthe total energy when they
are at an infinite distance apart by an amount
pmm' JJJ {~(~) ~ (.!) + oy'0'"

ax' ax,'
~ (!)~(.!) + o{.,o{.r'
~ (!)~(~)}dx dy k,
283
Thomson, moreover, investigated 1 the ponderomotive forces

which act between two solid bodies immersed in a fluid, when one
of the bodies is constrained to perform small oscillations. If, for
example, a small sphere immersed in an incompressible fluid is
compelled to oscillate along the line which joins its centre to that of
a much larger sphere,which is free, the free spherewill be attracted
if it is denserthan the fluid; while if it is lessdensethan the fluid,
it will be repelled or attracted according as the ratio of its distance
from the vibrator to its radius is greater or lessthan a certain quantity
depending on the ratio of its density to the density of the fluid.
Systems of this kind were afterwards extensively investigated by
C. A. Bjerknes.2 Bjerknes showed that two spheres which are
immersed in an incompressiblefluid, and which pulsate (i.e. change
in volume) regularly, exert on each other (by the mediation of the
fluid) an attraction, determined by the inverse square law, if the
pulsations are consordant; and exert on each other a repulsion,
determined likewise by the inverse square law, if the phasesof the
pulsations differ by half a period. If the phasesdiffer by a quarter
period, there is no action. It is necessaryto supposethat the medium
is incompressible,so that all pulsations are propagated instantane-
ously; otherwise attractions would change to repulsions and vice
versa at distances greater than a quarter wave-length.3 If the
1 cf. F. Guthrie, Phil. Mag. xli (1871), p. 405, specially the letter of Thomson to
Guthrie dated 24 Nov. 1870,printed at p. 427..' i
• Repertorium d. Mathematik i (Leipzig, 1877), p. 268; Gollinger Nachmhten (1876),

p. 245; ComptesRendus, Ixxxiv (1877), p. 1375; cf. Nature, xxiv (1881), p. 360; cf. also
H. Witte, Ann. d. Phys. xxx (1909), p. 337, with a reply by V. Bjerknes, ibid. xxxi (1910),
p. 312, and a further note by Witte, ibid. xxxii (1910), p. 382.
• On the mathematical theory of the force between two pulsating spheresin a fluid,
d. W. M. Hicks, Proc. Camb. Phil. Soc. iii (1879), p. 276; iv (1880), p. 29.
the integration being taken throughout the whole volume of the fluid, except two small
H~,~G)
spheress, s', surrounding the sinks. By Green's theorem, this expressionreducesat once to
pmm' dS,
where the integration is taken over sand s', and n denotes the interior normal to s or s'.
7; H~(P
The integral taken over s' vanishes; evaluating the remaining integral, we have
• .' -p dS, . or 47rpmm'll. . .
The energy of the fluid is therefore greater when sinks of strengths m, m' are at a mutual
distance I than when sinks of the same strengths are at infinite distance apart by an
amount 47rpmm'll. Since, in the case of the tubes, the quantities m correspond to the
fluxes of fluid, this expression corresponds to the Lagrangian form of the kinetic energy ;
and therefore the force tending to increase the co-ordinate x of one of the sinks is
(alax) (47rpmm'Il). Whence it is seenthat the like ends of two tubes attract, and the unlike
ends repel, according to the inverse square law.
284
spheres,instead of pulsating, oscillate to and fro in straight lines

about their meanpositions,the forcesbetweenthem are proportional
in magnitude and the samein direction, but opposite in sign, to
thosewhich act betweentwo magnetsoriented along the directions
of oscillation.1
The resultsobtainedby Bjerkneswereextendedby A. H. Leahy 2
to the caseof two spherespulsating in an elastic medium; the
wave-lengthof the disturbancebeing supposedlarge in comparison
with the distancebetween the spheres. For this systemBjerknes'
resultsare reversed,the law being now that of attraction in the case
of unlike phases,and of repulsion in the caseof like phases; the
intensity is as before proportional to the inverse square of the
distance.
The same author afterwards discussed3 the oscillations which
may be produced in an elasticmedium by the displacement,in the
direction of the tangent to the cross-section,of the surfacesof tubes
of small sectionalarea: the tubes either forming closedcurves, or
extendingindefinitely in both directions. The direction and circum-
stancesof the motion are in general analogousto ordinary vortex
motions in an incompressiblefluid; and it was shown by Leahy
that, if the period of the oscillation be suchthat the wavesproduced
are long comparedwith ordinary finite distances,the displacement
due to the tangential disturbancesis proportional to the velocity due
to vortex rings of the sameform as the tubular surfaces. One of
these 'oscillatory twists,' as the tubular surfacesmay be called,
producesa displacementwhich is analogousto the magnetic force
due to a current flowing in a curve coincident with the tube; the
strength of the current being proportional to b2w sin pt, where b
denotesthe radius of the twist, and w sin pt its angular displacement.
If the field of vibration is exploredby a rectilineal twist of the same
period as that of the vibration, the twist will experiencea force at
right anglesto the plane containing the twist and the directi on of the
displacement which would exist if the twist were removed; if
the displacementof the medium be representedby F sin pt, and the
angular displacementof the twist by w sin pt, the magnitude of
the force is proportional to the vector-product of F (in the direction
of the displacement)and w (in the direction of the axis of the twist).
1 A theory of gravitation has been based by Korn on the assumption that gravitating
particles resemble slightly compressible spheres immersed in an incompressible perfect

fluid; the spheres execute pulsations, whose intensity corresponds to the mass of the
gravitating paTtic1es,and thus forces of the Newtonian kind are produced between them.
cr. Korn, Eine Theorie der Gravitation und der elect. Erscheinungen,Berlin, 1898.
• Trans. Camb. Phil. Soc. xiv (1884), p. 45
• Trans. Camb. Phil. Soc. xiv (1885), p. 188
285
A model of magnetic action may evidently be constructedon the

basisof theseresults. A bar magnet must be regarded as vibrating
tangentially, the direction of vibration being parallel to the axis of
the body. A cylindrical body carrying a current will have its surface
alsovibrating tangentially; but in this casethe direction of vibration
will be perpendicular to the axis of the cylinder. A statically electri-
fied body, on the other hand, may, asfollows from the sameauthor's
earlier work, be regarded as analogous to a body whose surface
vibratesin the normal direction. Larmor 1 suggestedthat gravitation
between electrons might be explained if the electrons are vacuous
and have free periodsof vibration in the fluid aether, the explanation
following the hydrodynamical pulsatory theory of Bjerknes.
We have now discussedmodels in which the magnetic force is
represented as the velocity in a liquid, and others in which it is
represented as the displacement in an elastic solid. Some years
before the date of Leahy's memoir, George Francis FitzGerald
(1851-1901) 2 had instituted a comparison between magnetic force
and the velocity in a quasi-elasticsolid of the type first devisedby
MacCullagh.3 Of all possible types of aether, this alone could
propagate waves having the properties of waves of light, and an
analogy with the electromagnetic theory of light is at once evident
when it is noticed that the electromagneticequation
lc oD
at
= curlH
is satisfiedidentically by the values
~D = curl e,
{ H = ae
ot'
where e denotesany vector; and that, on substituting thesevalues
in the other electromagneticequation,
aH
- c curl (DjE) - at'
ot + c
02e
£ ----
2
2 curl curl e = 0
which is no other than the equation of motion of MacCullagh's

1 Phil. Trans. c1xxxvi(1895), p. 697
• Phil. Trans. clxxi (1880), p. 691 (presentedOctober 1878)j FitzGerald's Scientijie
Writings, p. 45 •cr. p. 142
286
aether,l the specific inductive capacity € corresponding to the

reciprocal of MacCullagh's constant of elasticity. In the analogy
thus constituted, electric displacementcorrespondsto the twist of
the elementsof volume of the aether; and electric charge must
evidently be representedas an intrinsic rotational strain. Thus it
appearedpossibleto extend the capacitiesof this aether so as to
representnot only optical phenomena,but all kinds of magneticand
electric interaction. Mechanical models of the electromagnetic
field, basedon FitzGerald's analogy, were afterwards studied by
A. Sommerfeld,zby R. Reiff3 and by Sir J. Larmor.fo The last-
namedauthor 5 supposedthe electric charge to exist in the form of
discreteelectrons,for the creationof which he suggestedthe following
ideal processe: A filament of aether, terminating at two nuclei, is
supposedto be removed,and circulatory motion is imparted to the
walls of the channel so formed, at each point of its length, so as to
producethroughout the medium a rotational strain. When this has
beenaccomplished,the channel is to be filled up again with aether,
which is to be madecontinuouswith its walls. When the constraint
is removedfrom the walls of the channel, the circulation imposed
on them proceedsto undo itself, until this tendencyis balanced by
the elasticresistanceof the aether with which the channel has been
filled up; thus finally the systemassumesa state of equilibrium in
which the nuclei, which correspondto a positive and a negative
electron,aresurroundedby intrinsic rotational strain. On this view,
electrons,and henceall material bodiesbuilt up from them, are of
the nature of structuresin the aether, involving an atmosphereof
aetherealstrain all around them; action at a distanceis abolished.
, All that is needed,'said Larmor,7 , is a postulate of free mobility
of the nucleusthrough the aether. This is definitely hypothetical,
but it is not an unreasonablepostulate, becausea rotational aether
has the properties of a perfect fluid medium except where differ-
entially rotational motions are concerned,and so would not react
on the motion of any structure moving through it except after the
mannerof an apparent changeof inertia.' 'If,' he said elsewhere,8
, the electron is a mere passivepole-nucleus of beknottednessin
someway-in the aether,conditioned and controlled entirely by the
1 cf. p. 142 • Ann. d. Phys. xlvi (1892), p. 139

• Reiff, Eiasticitat und Eiektmillit, Freiburg, 1893
• Phil. Trans. clxxxv (1893), p. 719
• In a supplement,of date August 1894, to his above-citedmemoir of 1893
• Phil. Trans. clxxxv (1894), p. 810; cxc (1897), p. 210; Larmor, Aether and Matter
(1900), p. 326
• Brit. Ass. Rep. (1900), p. 626 • Phil. Mag. vii (1904-), p. 621
287
aether around it, just as a vortex ring is conditioned by the fluid in

which it subsistsand is also carried along thereby, then, as in the
familiar hydrodynamics of vortices, the motion of the aether deter-
mines the motion of the entirely passiveelectrons,and the idea of
force acting betweenthem and the aether is dispensedwith.'
Models in which magneticforce is representedby the velocity of
an aether are not, however, securefrom objection. It is necessary
to supposethat the aether is capable of flowing like a perfect fluid
in irrotational motion (which would correspondto a steady mag-
netic field), and that it is at the sametime endowedwith the power
(which is requisite for the explanation of electric phenomena) of
resisting the rotation of any element of volume.1 But when the
aether moves irrotationally in the fashion which correspondsto a
steady magnetic field, eachelementof volume acquiresafter a finite
time a rotatory displacementfrom its original orientation, in con-
sequenceof the motion; and it might therefore be expectedthat
the quasi-elastic power of resisting rotation would be called into
play-i.e. that a steady magnetic field would develop electric
phenomena.2
A further objection to all modelsin which magneticforce corre-
sponds to velocity is that a strong magnetic field, being in such
modelsrepresentedby a steadydrift of the aether,might be expected
to influence the velocity of propagation of light. The existenceof
such an effect appears,however,to be disprovedby the experiments
of Sir Oliver Lodge 3; at any rate, unlessit is assumedthat the
aether has an inertia at leastof the sameorder of magnitude asthat
of ponderable matter, in which casethe motion might be too slow
to be measurable.
Again, the evidencein favour of the rotatory as opposedto the
linear character of magneticphenomenahad perhaps,on the whole,
beenstrengthenedsinceThomson originally basedhis conclusionon
the magnetic rotation of light. This brings us to the consideration
of an experimental discovery.
1 Larmor (loc. cit.) suggestedthe analogy of a liquid filled with magnetic molecules
under the action of an external magnetic field.
It has often been objected to the mathematical conception of a perfect fluid that it
contains no safeguard against slipping between adjacent layers, so that there is no
justification for the usual assumption that the motion of a perfect fluid is continuous.
Larmor remarked that a rotational elasticity, such as is attributed to the medium a hove
considered, furnishes precisely such a safeguard; and that without some property of this
kind a continuous frictionless fluid cannot be imagined.
I Larmor proposed to avoid this by assuming that the rotation which is resisted by an
element of volume of the aether is the vector sum of the series of differential rotations
which it has experienced.
• Phil. Trans.c1xxxix (1897), p. 149
288
In 1879 Edwin H. Hall,! at that time a student at Baltimore,

repeating an experiment which had been previously suggestedby
H. A. Rowland, obtained a new action of a magneticfield on electric
currents. A strip of gold leaf mounted on glass,forming part of an
electric circuit through which a current was passing, was placed
betweenthe polesof an electromagnet,the plane of the strip being
perpendicular to the lines of magnetic force. The two poles of a
sensitivegalvanometerwere then placed in connectionwith different
parts of the strip, until two points at the samepotential were found.
When the magnetic field was created or destroyed,a deflection of
the galvanometerneedle was observed,indicating a change in the
relative potential of the two poles. It was thus shown that the
magneticfield producesin the strip of gold leaf a new electromotive
force, at right anglesto the primary electromotiveforce and to the
magneticforce, and proportional to the product of theseforces.2
From the physical point of view we may therefore regard Hall's
effect as an additional electromotiveforce generatedby the action
of the magneticfield on the current; or alternatively we may regard
it as a modification of the ohmic resistanceof the metal, such as
would be producedif the moleculesof the metal assumeda helicoidal
structure about the lines of magnetic force. From the latter point
of view, all that is neededis to modify Ohm's law
S =kE
(where S denotes electric current, k specific conductivity and E
electric force) so that it takesthe form
S=kE+h[E.H]
where H denotes the imposed magnetic force, and h denotes a
constant on which the magnitude of Hall's phenomenondepends.
It is a curious circumstance that the occurrence, in the case of
magnetisedbodies,of an additional term in Ohm's law, formed from
a vector product of E and H, had been expresslysuggestedin
Maxwell's Treatise 3: although Maxwell had not indicated the
possibilityof realising it by Hall's experiment.
An interesting application of Hall's discovery was made in the
1 Am. Jour. Math. ii (1879), p. 287; Am. J. Sci. xix (1880), p. 200, and xx, p. 161 ;
Phil. Mag. ix (1880), p. 225, and x, p. 301
• Oliver Lodge in the 1870's had worked experimentally on the flow of electricity in
a metallic sheet, and had come very near to discovering the Hall effect; but he read in
Maxwell's Electricityand Magnetism(vol. ii, § 501): 'It must be carefully remembered,
that the mechanical force which urges a conductor carrying a current across the lines of
magnetic force acts, not on the electric current, but on the conductor which carries it ' ;
and this deterred him from making the crucial test.
• § 303. cr. Hopkinson, Phil. Mag. x (1880), p. 430
289
sameyear by Boltzmann,l who remarked that it offered a prospect

of determining the absolute velocity of the electric chargeswhich
carry the current in the strip. For if it is supposedthat only one
kind (vitreous or resinous)of electricity is in motion, the force on
one of the chargestending to drive it to one side of the strip will be
proportional to the vector product of its velocity and the magnetic
intensity. Assuming that Hall's phenomenonis a consequenceof
this tendencyof chargesto move to one sideof the strip, it is evident
that the velocity in question must be proportional to the magnitude
of the Hall electromotive force due to a unit magnetic field. On
the basis of this reasoning, A. von Ettingshausen2 found for the
current sent by one or two Daniell's cells through a gold strip a
velocity of the order of 0.1 em. per second. It is clear, however,
that, if the current consistsof both vitreous and resinouschargesin
motion in opposite directions, Boltzmann's argument fails; for the
two kinds of electricity would give oppositedirections to the current
in Hall's phenomenon.
In the year following his discovery,Hall 3 extendedhis researches
in another direction, by investigating whether a magnetic field
disturbs the distribution of equipotential lines in a dielectric which
is in an electric field; but no effect could be observed.~ Such an
effect, indeed,5was not to be expectedon theoretical grounds; for
when, in a material system, all the velocities are reversed, the
motion is reversed,it being understood that, in the applic,.tion of
this theorem to electrical theory, an electrostatic state is to be
regardedas one of .est, and a current as a phenomenonof motion;
and if such a reversalbe performed in the presentsystem,the poles
of the electromagnet are exchanged, while in the dielectric no
changetakesplace.
We must now considerthe bearing of Hall's effecton the question
as to whether magnetism is a rotatory or a linear phenomenon.6
If magnetismbe linear, electric currents must be rotatory; and if
Hall's phenomenonbe supposedto take place in a horizontal strip
of metal, the magnetic force being directed vertically upwards, and
the primary current flowing horizontally from north to south, the
only geometrical entities involved are the vertical direction and a
1 Wien Anz. xvii (1880), p. 12; Phil. Mag. ix (1880), p. 307
I Ann. d. Phys. xi (1880), pp. 432, 1044
I Amer. Jour. Sci. xx (1880), p. 164
• In 1885~ E. van Aubel (Bull. de l'Acad. Roy. de Belgique [3], x, p. 609; xii, p. 280)
rep"ated the investigation in an improved form, and confirmed the result that a magnetic
field has no influence on the electrostatic polarisation of dielectric.s.
• H. A. Lorentz, Arch. Neerl. xix (1884), p. 123
• cr. F. Kolacek, Ann. d. Phys. Iv (1895), p. 503
29°
rotation in the east-and-westvertical plane; and theseareindifferent

with respectto a rotation in the north-and-south vertical plane, so
that there is nothing in the physical circumstancesof the systemto
determine in which direction the secondary current shall flow.
The hypothesisthat magnetism is linear appears therefore to be
inconsistentwith the existenceof Hall's effect.! There are, how-
ever, someconsiderationswhich may be urged on the other side.
Hall's effect, like the magnetic rotation of light, takesplace only in
ponderablebodies,not in free aether; and its direction is sometimes
in one sense,sometimesin the other, according to the nature of the
substance. It may therefore be doubted whether thesephenomena
are not of a secondarycharacter, and the argument basedon them
invalid. Moreover, as FitzGerald remarked,2the magnetic lines of
force associatedwith a systemof currents are circuital and have no
open ends,making it difficult to imagine how alteration of rotation
inside them could be produced.
Of the various attempts to represent electric and magnetic
phenomenaby the motions and strains of a continuous medium,
noneof thosehitherto consideredhasbeenfound freefrom objection.3
Beforeproceedingto considermodelswhich are not constituted by a
continuousmedium, mention must be made of a suggestionoffered
by Riemann in his lectures<I of 1861. Riemann remarked that the
scalar-potentialc/> and vector-potential a, correspondingto his own
law of force betweenelectrons,satisfythe equation
oc/>
ot
+ c diva = 0.
'
an equationwhich, aswe haveseen,is satisfiedalsoby the potentials
of L. Lorenz.5 This appearedto Riemann to indicate that c/> might
representthe densityof an aether,of which ca representsthe velocity.
It will be observedthat on this hypothesisthe electric and magnetic
forcescorrespondto second derivatesof the displacement-a circum-
stancewhich makesit somewhatdifficult to assimilatethe energy
possessed by the electromagneticfield to the energyof the model.
We must now proceed to consider those models in which the
1 Further evidence in favour of the hypothesis that it is the electric phenomena which
are linear is furnished by the fact that pyro-electric effects (the production of electric
polarisation by warming) occur in acentric crystals, and only in such. cr. M. Abraham,
Encyklnpadie der math. Wiss. iv (2), p. 43.
• cr. Larmor, Phil. Trans. clxxxv (1894), p. 780
• cr. H. Witte, Ueber den gegenwiirtigen Stand tier Frage ruuh einer mechallischen ErklaTUng der
elektrischen Erscheinungen, Berlin, 1906
• Published after his death by K. Hattendorff, under the title Schwere, ElektMliit,
und Magnetismus (1875), p. 330
• cr. p. 269
(823) 2g1
AETHER ANn ELECTRICITY
aether is representedas composedof more than one kind of con·

stituent: of these Maxwell's model of 1861-2, formed of vortices
and rolling particles, may be taken as the type. Another device
of the sameclasswas describedin 1885by FitzGerald 1; this was
constituted of a number of wheels, free to rotate on axes fixed
perpendicularly in a plane board; the axes were fixed at the
intersections of two systemsof perpendicular lines; and each
wheel was gearedto each of its four neighboursby an indiarubber
band. Thus all the wheelscould rotate without any straining of the
system,provided they all had the same angular velocity; but if
someof the wheelswererevolving fasterthan others,the indiarubber
bandswould becomestrained. It is evident that the wheelsin this
model play the same part as the vortices in Maxwell's model of
1861-2: their rotation is the analogueof magnetic force; and a
region in which the massesof the wheelsare large correspondsto
a region of high magnetic permeability. The indiarubber bands
of FitzGerald's model correspondto the medium in which Maxwell's
vortices were embedded; and a strain on the bands represents
dielectric polarisation, the line joining the tight and slack sidesof
any band being the direction of displacement. A body whose
specificinductive capacity is large would be representedby a region
in which the elasticity of the bands is feeble. Lastly, conduction
may be representedby a slipping of the bandson the wheels.
Such a model is capable of transmitting vibrations analogous
to those of light. For if any group of wheelsbe suddenly set in
rotation, those in the neighbourhood will be prevented by their
inertia from immediately sharing in the motion; but presentlythe
rotation will be communicatedto the adjacent wheels,which will
transmit it to their neighbours; and so a wave of motion will be
propagated through the medium. The motion constituting the
wave is readily seento be directed in the plane of the wave,i.e. the
vibration is transverse. The axesof rotation of the wheelsare at
right anglesto the direction of propagation of the wave, and the
direction of polarisation of the bandsis at right anglesto both these
directions.
The elasticbands may be replacedby lines of governorballs 2 :
if this bedone,the energyof the systemis entirely of the kinetic type.3
1 &i.ent. Proc. Roy. Dublin Soc. iv (1885), p. 407; Phil. Mag. xix (1885),p. 438;
FitzGerald'sScitntjft Writings, pp. 142,157
I FitzGerald'sScientjft Writings, p. 271
I It is of coursepossibleto devisemodelsof this classin which the rotation may be
interpretedas having the electricinsteadof the magneticcharacter. Sucha modelwas

proposedby Boltzmann, VorleSWlgenUb" Maxwell's The0ri4 (Leipzig 1891-93),ii.
292
Modelsof typesdifferent from the foregoinghave beenproposed,

which involve the idea of vortex motion. One of the greatestachieve-
ments of Helmholtz was his discovery1 in 1858 that vortex rings
in a perfect fluid are types of motion which possesspermanent
individuality throughout all changes,and cannot be destroyed,so
that they may be regarded as combining and interacting with each
other,althougheachof them consistsofa motion pervadingthe whole
of the fluid. The energy of the fluid can be expressedin terms of
,the positionsand strengthsof the vortices, and from a knowledge
of thesecharacteristicsit is possibleto determine the future course
of the system. As Helmholtz showed, vortex filaments interact
mechanicallylike linear electriccircuits (but with attraction replaced
by repulsionand vice versa),the strengthsof the vorticescorrespond-
ing to the current strengths,and the velocity of the fluid in the neigh-
bourhood of the filaments corresponding to the magnetic force ;
magneticpoleswould be representedby sourcesand sinksin the fluid.
The individuality of vortices suggesteda connection with the
atomic theory of matter. As we have seen,2the atomic hypothesisof
Leucippus and Democritus had been adopted by Gassendi,whose
viewsin turn profoundly influenced Robert Boyle. To Boyle must
be given the chief credit for introducing the notion of chemical
element, which became more precise when Antoine-Laurent
Lavoisier (1743-94) found that compoundswhich have a definite
individuality contain the chemicalelementsin constantproportions,
and that the total massconcernedin a chemicalreaction is the same
after it as beforeit. On the physical side, Newton 3 showedthat a
gas would obey Boyle's law of compressionif it were composedof
particles, which repel each other with forces reciprocally pro-
portional to the distances between their centres; and Daniel
Bernoulli 4 (1700-82) in 1738 gave a different explanationof Boyle's
law, in which the gaswassupposedto be constituted of small elastic
sphereswhich collide with each other-a kinetic theory of gasesin
fact. These researches,chemical and physical, culminated in the
work of John Dalton (1766-1844), who proved that to everyfj
chemical elementcould be assigneda definite number, representing

the relative weight of an atom of the element.'
1 Jour.fiir Math. Iv (1858), p. 25 • cr. pp. 12, 13 • Principia, ii, prop. 23

• Hydrodynamica (Strasbourg,1738), § 10. See also Hooke's LecturesDe potentia
restitutiva, London, 1678
• In a paper read before the ManchesterPhilosophicalSociety in 1803 (Mem.
ManchesterPhil. Soc.i [1805], p. 271), and in his New Systemof ChemicalPhilosophy (1808)
• cr. Kurd Lasswitz,Geschichtetier Atomistik vom Millelalter bis Newton, Hamburg, 1890,
P. Kirchberger,La IMom aIomiqw, SOlI histoire et oRIndiveloppnneru,Paris,1930
293
The earliest attempts to build up a general physical theory on

the basisof vortex motion were made in 1867by William Thomson
(Kelvin), and were suggestedby a display of smoke rings which
he happened to seein the lecture room of his friend, P. G. Tait,
in Edinburgh University. He used vortices in the first place to
illustrate the properties of ponderable matter rather than of the
luminiferous medium, and pointed out that if 1 the atomsof matter
are constituted of vortex rings in a perfect fluid, the conservation
of matter may be immediately explained. The mutual interactions
of atoms may be illustrated by the behaviour of smokerings, which
after approaching eachother closelyare observedto rebound: and
the spectroscopicproperties of matter may be referred to the pos-
sessionby vortex rings of free periodsof vibration.2
A further recommendation of the vortex-atom hypothesiswas
that it seemedto throw somelight on the mode of propagation of
gravity, which since Laplace's investigations on the subject had
been believed to be instantaneous. For the actions between two
vortex rings in a perfect liquid are not propagatedfrom one to the
other: they are due to the fact that eachvortex ring is accompanied
by motion throughout the liquid, so that each may be regardedas
extending over the whole of space, and involving all the others;
and the effects are simultaneous everywhere. In the seventeenth
century, Michell and Boscovichhad suggestedthat an atom should
not be supposedto have a definite size but should be conceivedof
as completelypenetrable, and extending throughout all space; and
Faraday had adopted this view in his Thoughts on Ray Vibrations3
of 1846.
In the nineteenth century it was objected to the vortex theory
that thevirtual inertia of avortex ring increasesasits energyincreases;
1 Phil. Mag. xxxiv (1867), p. 15; Proc.R. S. Edin. vi, (1869) p. 94

• An attempt was made in 1883 by J. J. Thomson (Phil. Mag. xv [1883], p. 427) to
explain the phenomena of the electric discharge through gasesin terms of the theory of
vortex atoms. The atoms of a gas were represented as vortex rings: the combination of
two atoms to form a molecule would then correspond to the union or pairing of two
vortex rings, a phenomenon which was well known experimentally and which Thomson
himself had investigated mathematically. Now supposethat there is a quantity of gas in
an electric field. The electric field he represented by a distribution of velocity in the
medium whose vortex motion constitutes the atoms of the gas; the disturbance due to
this distribution of velocity will causesome of the paired vortex rings to dissociate, and so
a mathematical theory of the phenomenon of conduction in gasesbecomespossible. He
worked out the conditions for the disruptive discharge, and showed that the intensity of
the field necessaryto produce discharge should diminish when the gas is rarefied. This
investigation is notable inasmuch asit was the beginning of thoseresearcheson conduction
in rarefied gases which occupied so much of J.J. Thomson's life and led him to his
groeatestdiscovery, that of the electron.
• Exp. &s. iii, p. 447
this fact, however, fits in with later discoveriesregarding the con-

nection of massand energy, and would therefore now be regarded
ashighly favourable.
It is, however, doubtful whether vortex atoms would be stable.
'It now seemsto me certain,' wrote W. Thomson 1 (Kelvin) in
1905, 'that if any motion be given within a finite portion of an
infinite incompressibleliquid, originally at rest, its fate is necessarily
dissipationto infinite distanceswith infinitely small velocitiesevery-
where; while the total kinetic energy remains constant. After
many years of failure to prove that the motion in the ordinary
HelmholtZ circular ring is stable, I came to the conclusionthat it is
essentiallyunstable, and that its fate must be to becomedissipated
asnow described.'
The vortex-atom hypothesisis not the only way in which the
theory of vortex motion has been applied to the construction of
modelsof the aether. It was shownin 1880by W. Thomson 2 that
in certain circumstancesa massof fluid can exist in a statein which
portions in rotational and irrotational motion are finely mixed
together,so that on a large scalethe massis homogeneous,having
within any sensiblevolume an equal amount of vortex motion in all
directions. To a fluid having such a type of motion he gave the
namevortex sponge.
Five years later, FitzGerald 3 discussedthe suitability of the
vortex spongeasa model of the aether. Sincevorticity in a perfect
fluid cannot be createdor destroyed,the modification of the system
which is to be analogousto an electric field must be a polarised
state of the vortex motion, and light must be represented by a
communicationof this polarisedmotion from one part of the medium
to another. Many distinct types of polarisation may readily be
imagined: for instance,if the turbulent motion were constituted of
vortex rings, thesemight be in motion parallel to definite lines or
planes; or if it were constituted of long vortex filaments, the fila-
mentsmight be bent spirally about axesparallel to a given direction.
The energyof any polarisedstateof vortex motion would be greater
than that of the unpolarised state; so that if the motion of matter
had the effect of reducing the polarisation, there would be forces
tending to produce that motion. Since the forces due to a small
vortex vary inversely as a high power of the distance from it, it
seemsprobable that in the caseof two infinite planes,separatedby
a regionof polarisedvortex motion, the forcesdue to the polarisation
1 PrIX. R. S. Edin. xxv (1905), p. 565 I Brit. Assoc. Rep. (1880), p. 473
• Scimt. Proc. Roy. Dublin Soc. iv (1885), p. 407; Scimtijie Writings of FitzGerald, p. 154
295
betweenthe planeswould dependon the polarisation,but not on the

mutual distance of the planes-a property which is characteristic
of plane distributions whose elements attract according to the
Newtonian law.
It is possibleto conceivepolarisedforms of vortex motion which
are steady so far as the interior of the medium is concerned,
but which tend to yield up their energy in producing motion
of its boundary-a, property parallel to that of the aether, which,
though itself in equilibrium tends to move objects immersed
in it.
In the sameyear Hicks 1 discussedthe possibility of transmitting
waves through a medium consistingof an incompressiblefluid in
which small vortex rings are closely packed together. The wave-
length of the disturbance was supposedlarge in comparison with
the dimensionsand mutual distancesof the rings; and the trans-
latory motion of the latter was supposedto be so slow that very
many wavescan passover anyone before it has much changedits
position. Such a medium would probably act as a fluid for larger
motions. The vibration in the wave-front might be either swinging
oscillationsof a ring about a diameter, or transversevibrations of the
ring, or apertural vibrations; vibrations normal to the plane of
the ring appear to be impossible. Hicks determined in each case
the velocity of translation, in terms of the radius of the rings, the
distance of their planes,and their cyclic constant.
The greatestadvancein the vortex-spongetheory of the aether
was made in 1887,when W. Thomson 2 showedthat the equation
of propagation of laminar disturbancesin a vortex sponge is the
same as the equation of propagation of luminous vibrations in
the aether. The demonstration, which in the circumstancescan
scarcelybe expectedto be either very simple or very rigorous, is as
follows:
Let (u, v, w) denote the componentsof velocity, and p the pres-
sure, at the point (x,y, z) in an incompressiblefluid. Let the initial
motion be supposedto consist of a laminar motion {f(y), 0, O},
superposedon a homogeneous,isotropic and fine-grained distribu-
tion (uo', Vo, wo): so that at the origin of time the velocity is
{f(y) + uo', Vo, wo}: it is desiredto find a functionf(y, t) such
that at any time t the velocity shall be {f (y, t) +
u', v, w}, where
u', v, w, are quantities of which everyaveragetakenover a sufficiently
large spaceis zero.
1 Brit. Assoc. Rep. (1885), p. 930
I Phil. Mag. xxiv (1887), p. 34-2; Kelvin's Math. and Phys. Papers, iv, p. 308
296
Substituting thesevalues of the componentsof velocity in the

equationof motion
au au - v -au - w -all -- ap
.- = - u-
at ax Oy az ax'
thereresults
OL(y,J} + au' = _ f(y t) au' _ v af(y, t) _ u' {I,l _ v au~
at at ax' Oy (Ix ay
au' ap
-w---
az ax·
Take now the xz-averagesof both members. The quantities au'lat,
au'fax, v, apfax have zero averages; sothe equation takesthe form
OJ(y, t) = _ A (u' ~t!.+ v au' + w aU')
at ax Oy az '
if the symbolA is usedto indicate that the xz-averageis to be taken
of the quantity following. Moreover, the incompressibility of the
fluid is expressedby the equation
au' + av + aw =0
ax Oy az '
whence
o=- A (u' au' + u' av + u' aw).
ax Oy az
When this is added to the preceding equation, the first and third
pairs of terms of the secondmember vanish, since the x averageof
any derivate aQfax vanishesif Q, is finite for infinitely great values
of x; and the equation thus becomes
OJ(y, t) __ A a (u'v)
at - Oy'
(1)
From this it is seenthat if the turbulent motion were to remain

continually isotropic as at the beginning, f (y, t) would constantly
retain its critical value fey). In order to examine the deviation
from isotropy, we shall determine Aa(u'v)fat, which may be done in
the following way: multiplying the u- and v-equationsof motion
by v, u' respectively,and adding, we have· -
11 gfJyJl + a (u'v) = _ fey t) a (u'v) _ Vi Y(y t) _ u' a (u'v)
at at ' ax Oy . ax
a (u'v)
-l1---W--
a (u'v) -v-ap -u-, ap
Oy az ax Oy.
Taking the xz-averageof this, we observethat the first term of the
297
first member disappears,since Av is zero, and the first term of the

secondmember disappears,since Ao(u'v)fox is zero. Denoting by
!R 2 the averagevalue of us, v2 or wS, so that R may be called the
average velocity of the turbulent motion, the equation becomes
~ {Au'v} = - iR.lf(y,t) - Q
h
were
ot Oy'
Q. = A {u' ~(u'v) + v o(u'v) + w o(u'v) + v op + u' .0P}.

oX oy oz ox ay
Let p be written (p' +
w), wherep' denotesthe value which p would
have iff were zero. The equationsof motion immediately give
_
"V r
II> = (OU)I
ox
+ (OV)I
oy
+ (OW)I
oz
+ 2 ov ow + 2 Ow ~ . 2 ou ov.
oz oy ox oz + Oy ox '
and on subtracting the forms which this equation takesin the two
cases,we have
which, when the turbulent motion is fine-grained, so that f (y, t) is

sensiblyconstant over rangeswithin which u', v, w passthrough all
their values,may be written
w = _ 2 If(y, t) "V-I ov.

ay ox
Moreover, we have
O -_ A { u , o(u'v) + o(u'v) + o(u'v) + op' + ' op') •

ox v oy W oz v ox tl By f '
for positive and negative values of u', v, ware equally probable;
and therefore the value of the secondmember of this equation is
doubled by adding to itself what it becomeswhen for u', 0, w we
substitute - u', - v, - w; which (as may be seenby inspection
of the above equation in "V2p) does not change the value of p'.
Comparing this equation with that which determinesthe value of
Q., we have
Q. = A .ow ow)(v ox + u' ay ,
or substituting for w,
Q.= _ 2 of(y, t) A
ay
(0 ox~ + u' -!)
oy
"V-I au.
ax
2g8
The isotropy with respectto x and z givesthe equation
2A (Vo ix + uo' ;y) V-I ~:o A {vo (::. = + ;;1)

+ ( Uo, ox a) a;a}
a + Wo az 'n-I
v Do'
But by integration by parts we obtain the equation
ax + Wo oz0)
A ( Uo, 0 0
ov·
'n-I
v Do -
_
-
A (ouaxo' + &
owo)-~
0 'n-I
v vo,
•
and by the condition of incompressibility the secondmember may

be written
A. (ovo/i)y). (o/i)y). V-IVO' or - Avo (Ol/i)yl) V-IV.;
sowe have
(01 1
Ol)}_
Qo = - ~
Of(y, t)
A
{
Vo ox! + OZI0 - i)y. ViVO'
On accountof the isotropy, we may write i for

1 1 1
0
( oxl -
0
oyl + oz. ( )
V-I;
so
Qo = _ RI Of(y, t) •
9 oy ,
and, therefore,
.~ {A(u'v) } = _ 2RI{0f(Y, t)}
ot 1'=0) 9 oy '=0.
The deviation from isotropy shownby this equation is very small,
becauseof the smallnessof Of(J, t)/oy. The equation is therefore
not restricted to the initial valuesof the two members,for we may
neglectan infinitesimal deviation from (2/9) R 2 in the first factor of
the secondmember, in consideration of the smallnessof the second
factor. Hencefor all valuesof t we have the equation
!.
at
A (u'v) = _ 2
"D"
R. Of(y, t)
oy'
which, in combination with (I), yields the result
01 2RI 01
otil (y, t) = 9" oyll (y, t) ;
the form of this equation showsthat laminar disturbances are propagated
299
AETHER AND ELECTRICrrY
through the vortex sponge in the same manner as waves of distortion in a

homogeneouselastic solid.
The question of the stability of the turbulent motion remained
undecided; and at the time Thomson seemsto have thought it
likely that the motion would suffer diffusion. But two yearslater 1
he showedthat stability was ensuredat any rate when spaceis filled
with a set of approximately straight hollow vortex filaments. Fitz-
Gerald 2 subsequentlydetermined the energy per unit volume in a
turbulent liquid which is transmitting laminar waves. Writing for
brevity
(2/9) RI = VI, fCy, t) = P, and A (u'v) = y,
the equationsare ap Qy Qy ap
at = - CJy'
and - = - VI-
at CJy
If the quantity
is integrated throughout space,and the variations of the integral

with respectto time are determined, it is found that
a~IH ~dxdydz = IH (p ~; + ~I Z)dxtfytk
=IH( P ~; - y ~;) dx tfy dz.
Integrating by parts the secondterm under the integral, and omitting

the superficial terms (which may be at infinity, or whereverenergy
entersthe spaceunder consideration),we have
-:e HI ~ dx tfy tk = IH p (~; + ~) dx tfy dz = O.

Hence it appearsthat the quantity ~, which is of the dimensionsof
energy, must be proportional to the energy per unit volume of the
medium. SinceP correspondsto one of the two quantities,electric
and magneticforce, while y correspondsto the other, the expression
HP2 +
y2fV2) for 1: showsthat the energy per c.c. of the wave is
the.sum of the squaresof the electric and magneticforces,in suitable
units; a result which showsthat there is a pronouncedsimilarity
between the dynamics of a vortex spongeand of Maxwell's elastic
aether.
A definite vortex-spongemodel of the aether was describedby
1 Proc. Roy. Irish Acad. (3), i (1889), p. 340; Kelvin's Math. and Phys. Papers, iv, p. 202
I Brit. Assoc. Rep. (1899), p. 632; FitzGerald's Scientific Writings, p. 484, see also
pp. 254, 472
300
Hicks in his presidential addressto the mathematical sectionof the

British Association in 1895.1 In this the small motions whose
function is to confer the quasi-rigidity were not completely chaotic,
but were disposedsystematically. The medium was supposedto be
constitutedof cubical elementsof fluid, eachcontaining a rotational
circulation completein itself. In any element, the motion closeto
the central vertical diameter of the element is vertically upwards ;
the fluid which is thus carried to the upper part of the elementflows
outwardsover the top, down the sidesand up the centre again. In
eachof the six adjoining elementsthe motion is similar to this, but
in the reversedirection. The rotational motion in the elements
conferson them the power of resistingdistortion, so that wavesmay
be propagated through the medium as through an elastic solid;
but the rotations are without effect on irrotational motions of the
fluid, provided the velocities in the irrotational motion are slow
comparedwith the velocity of propagation of distortional vibrations.
A different model was describedfour yearslater by FitzGerald.s
Sincethe distribution of velocity of a fluid in the neighbourhoodof
a vortex filament is the sameas the distribution of magnetic force
around a wire of identical form carrying an electric current, it is
evident that the fluid has more energy when the filament has the
form of a helix than when it is straight; soif spacewere filled with
vortices, whose axes were all parallel to a given direction, there
would be an increase in the energy per unit volume when the
vorticeswere bent into a spiral form; and this could be measured
by the squareof a vector-say, E-which may be supposedparallel
to this direction.
If now a single spiral vortex is surrounded by parallel straight
ones,the latter will not remain straight, but will be bent by the
action of their spiral neighbour. The transferenceof spirality may
be specifiedby a vector H, which will be distributed in circlesround
the spiral vortex; its magnitude will depend on the rate at which
spirality is being lost by the original spiral, and can be taken such
that its square is equal to the mean energy of this new motion.
The vectorsE and H will then representthe electric and magnetic
vectors; the vortex spiralsrepresentingtubesof electric force.
FitzGerald'sspirality is essentiallysimilar to the laminar motion
investigatedby Lord Kelvin, sinceit involvesa flow in the direction
of the axis of the spiral, and sucha flow cannot take place along the
1 Brit. Assoc. Rep. (1895),p. 595

• Proc. Roy. Dublin Soc., IX (Dec. 1899), p. 50; FitzGerald's Scientific Writings,
p.472
3°1
AETHER AND ELECTRICrrY
direction of a vortex filament without a spiral deformation of a

filament.
Other vortex analogues have been devised for electrostatical
systems. One such, which was describedin 1888by W. M. Hicks,!
dependson the circumstance that if two bodies in contact in an
infinite fluid are separatedfrom eachother, and if there be a vortex
filament which terminateson the bodies,there will be formed at the
point where they separatea hollow vortex filament 1 stretchingfrom
one to the other, with rotation equal and opposite to that of the
original filament. As the bodiesare moved apart, the hollow vortex
may, through failure of stability, dissociateinto a number of smaller
ones; and if the resulting number be very large, they will ultimately
take up a position of stable equilibrium. The two setsof filaments
-the original filaments and their hollow companions-will be
intermingled, and each will distribute itself according to the same
law as the lines of force betweenthe two bodieswhich are equally
and oppositelyelectrified.
Since the pressureinside a hollow vortex is zero, the portion of
the surfaceon which it abuts experiencesa diminution of pressure;
the two bodies are therefore attracted. Moreover, as the two
bodies separatefarther, the distribution of the filaments being the
same as that of lines of electric force, the diminution of pressure
for each line is the same at all distances,and therefore the force
between the two bodies follows the samelaw as the force between
two bodiesequally and oppositelyelectrified. It may be shownthat
the effect of the original filaments is similar, the diminution of
pressurebeing half aslarge again asfor the hollow vortices.
If another surfacewere brought into the presenceof the others,
those of the filaments which encounter it would break off and
rearrangethemselvessothat eachpart of a brokenfilament terminates
on the p.ew body. This analogy thus gives a complete account
of electrostatic actions both quantitatively and qualitatively; the
electric charge on a body correspondsto the number of ends of
filaments abutting on it, the sign being determined by the direction
of rotation of the filament asviewed from the body.
A magneticfield may be supposedto be producedby the motion
of the vortex filaments through the stationary aether, the magnetic
force being at right angles to the filament and to its direction of
1 Brit. Assoc. Rep. (1888), p. 577

• A hollow vortex is a cyclic motion existing in a fluid without the presence of any
actual rotational filaments. On the general theory cf. Hicks, Phil. Trans. clxxv (1883),
p. 161 ; dxxvi (1885), p. 725; cxcii (1898), p. 33.
302
motion. Electrostatic and magnetic fields thus correspondto states

of motion in the medium, in which, however, there is no bodily
flow; for the two kinds of filament produce circulation in opposite
directions.
It is possiblethat hollow vorticesarebetter adaptedthan ordinary
vortex filamentsfor the construction of modelsof the aether. Such,
at any rate, wasthe opinion of Thomson (Kelvin) in his later years.}
The analytical difficulties of the subjectare formidable, and progress
was consequentlyslow; but among the many mechanical schemes
which weredevisedin the nineteenth century to representelectrical
and optical phenomena, none possesses greater interest than that
which pictures the aether as a vortex sponge.
Towards the closeof the nineteenth century, chiefly under the
influence of Larmor, it came to be generally recognisedthat the
aetheris an immaterial medium, sui generis, not composedof identi-
fiable elementshaving definite locations in absolute space. The
older view had supposed' the pressuresand thrusts of the engineer,
and the strains and stressesin the material structures by which he
transmits them from one place to another, to be the archetype of
the processes by which all mechanicaleffect is transmitted in nature.
This doctrine implies an expectation that we may ultimately dis-
cover something analogousto structure in the celestial spaces,by
meansof which the transmissionof physical effect will be brought
into line with the transmission of mechanical effect by material
framework.' 2 Larmor urged on the contrary that ' we should not
be tempted towards explaining the simple group of relations which
have been found to define the activity of the aether by treating
them as mechanical consequencesof concealed structure in that
medium; we should rather rest satisfied with having attained to
their exact dynamical correlation, just as geometry explores or
correlates,without explaining, the descriptive and metric properties
of space.' This point of view enabled Larmor's theory to withstand
subsequentcriticisms based on the principle of relativity, which
shatteredpractically all rival conceptsof the aether.3
1 Proc. RlJy. Irish Acad, i (30 Nov. 1889),p. 340; Kelvin's Math. and Plrys. Papers, iv,
p. 202. 'Rotational vortex cores,'he wrote, ' must be absolutelydiscarded; and we

musthavenothingbut irrotational revolutionand vacuouscores.'
I Larmor, Brit. Ass. Rep. (1900),p. 618
I Modelsof the aetherwerediscussed, in the light of the knowledgeavailablein 1908,

by H. Witte, Ann. d. Phys. xxvi (1908),p. 235.
Chapter X
THE FOLLOWERS OF MAXWELL
THE most notable imperfection in the electromagnetic theory of

light, as presentedin Maxwell's original memoirs,was the absence
of any explanation of reflection and refraction. Beforethe publica-
tion of Maxwell's Treatise, however, a method of supplying the
omission was indicated by Helmholtz.l The principles on which
the explanation depends are that the normal componentsof the
electric displacement D and the magnetic induction B, and the
tangential components of the electric force E and the magnetic
force H, are to be continuous acrossthe interface at which the
reflection takesplace; the optical differencebetweenthe contiguous
bodiesbeing representedby a differencein their dielectric constants,
and the electric vector being assumedto be at right angles to the
plane of polarisation.2 The analysisrequired is a meretranscription
of MacCullagh's theory of reflection,3 if the derivateof MacCullagh's
displacemente with respectto the time be interpretedasthe magnetic
force,p.curl e asthe electric force, and curl e asthe electricdisplace-
ment. The mathematical details of the solution were not given by
Helmholtz himself, but were supplied a few years later in the
inaugural dissertationof H. A. Lorentz."
In the yearsimmediately following the publication of Maxwell's
Treatise, a certain amount of evidencein favour of his theory was
furnished by experiment. That an electric field is closelyconcerned
with the propagation of light was demonstratedin 1875,whenJohn
Kerr I> showed that dielectrics subjected to powerful electrostatic
forceacquirethe property of doublerefraction,their optical behaviour
being similar to that of uniaxal crystalswhoseaxesare directedalong
the lines of force.
Other researchesundertaken at this time had a more direct
1 Jour.fur Math. !xxii (1870), p. 68n
I Helmholtz (Ioc. cit.) pointed out that if the optical difference between the media
were assumed to be due to a difference in their magnetic permeabilities, it would be
necessary to suppose the magnetic vector at right angles to the plane of polarisation in
order to obtain Fresnel's sine and tangent formulae of reflection.
I cf. pp. 137, 138, 142-4
• Zeitsehriftfur Math. u. Phys.xxii (1877), pp. 1,205: 'Over de theorie der terugkaat-
sing en breking van het licht,' Arnhem, 1875. Lorentz' work was based on Helmholtz'
equations, but remains substantially unchanged when Maxwell's formulae are substituted.
• Phil. Mag. i (1875), pp. 337,446; viii (1879), pp. 85, 229; xiii (Hl82), pp. 153,248
304
bearingon the questionsat issuebetweenthe hypothesisof Maxwell

and the older potential theories. In 1875-6 Helmholtz 1 and his
pupil Schiller 2 attempted to discriminate between the various
doctrinesand formulae relative to unclosedcircuits by performing a
crucial experiment.
It was agreed in all theories that a ring-shaped magnet, which
returnsinto itself so asto have no poles,can exert no ponderomotive
force on other magnetsor on closedelectric currents. Helmholtz 3
had, however,shownin 1873 that according to the potential theories
such a magnet would exert a ponderomotive force on an unclosed
current. The matter was tested by suspendinga magnetisedsteel
ring by a long fibre in a closedmetallic case,near which wasplaced
a terminal of a Holtz machine. No ponderomotive force could be
observedwhen the machine was put in action so as to produce a
brush discharge from the terminal; from which it was inferred
that the potential theoriesdo not correctly representthe phenomena,
at least when displacementcurrents and convection currents (such
as that of the electricity carried by the electrically repelled air from
the terminal) are not taken into account.
Helmholtz also found that when a conductor was rotated in a
magnetic field of force symmetrical about the axis of rotation, a
differenceof potential was induced between the axial and circum-
ferential parts, which was shown by the resulting electrification.
The researchesof Helmholtz and Schiller brought into promi-
nence the question as to the effects produced by the translatory
motion of electric charges. That the convection of electricity is
equivalentto a current had beensuggestedlong beforeby Faraday.'
, If,' he wrote in 1838, , a ball be electrified positively in the middle
of a room and be then moved in any direction, effectswill be pro-
duced as if a current in the samedirectio~ had existed.' Maxwell
in his Treatise 6 endorsedthe' supposition' that' a moving electrified
body is equivalent to an electric current.' To decide the matter
a new experiment inspired by HelmholtZ was performed by
H. A. Rowland 6 in 1876. The electrified body in Rowland's dis-
position was a disk of ebonite, coated with gold leaf and capable
of turning rapidly round a vertical axis betweentwo fixed plates of
1 Monatsbtrichte d. Acad. d. Berlin (1875),p. 400; Ann. d. Phys. elviii (1876),p. 87
t Ann. d. Phys. elix (1876),pp. 456, 537; elx (1877),p. 333
• The valuable memoirs by Helmholtz in Jour.flir Math. Ixxii (1870), p. 57; lxxv
(1873),p. 35; !xxviii (1874),p. 273, to which referencehasalready beenmade,contain
a full discussionof the variouspollSibilitiesof the potential theories.
• Exper. Res., § 1644 • §§ 768-70
• Monatsbtrichte d. Akad. d. Berlin (1876),p. 211; Ann. d. Phys. elviii (1876),p. 487 ;
AnnaZesde Chim. et de Phys. xii (1877),p. 119
305
glass,each gilt on one side. The gilt facesof the plates could be
earthed, while the ebonite disk received electricity from a point
placed near its edge; each coating of the disk thus formed a con-
denser with the plate nearest to it. An astatic needle was placed
above the upper condenserplate, nearly over the edgeof the disk;
and when the disk was rotated a magnetic field was found to be
produced. This experiment, which has since been repeatedunder
improved conditions by Rowland and Hutchinson,l H. Pender,2
Eichenwald,3E. P. Adams' and H. Penderand V. Cn~mieu5 shows
that the' convectioncurrent' produced by the rotation of a charged
disk, when the other ends of the lines of force are on an earthed
stationary plate parallel to it, produces the same magnetic field
asan ordinary conduction current flowing in a circuit which coincides
with the path of the convectioncurrent. When two disksforming a
condenserare rotated together, the magnetic action is the sum of
the magnetic actions of each of the disks separately. It appears,
therefore, that electric chargescling to the matter of a conductor
and move with it, sofar asRowland's phenomenonis concerned.
The first examination of the matter from the point of view of
Maxwell's theory was undertaken by J. J. Thomson (1856-1940),8
in 1881. If an electrostatically charged body is in motion, the
change in the location of the charge must produce a continuous
alteration of theelectricfield at anypoint in the surroundingmedium;
or, in the languageof Maxwell's theory, there must be displacement
currents in the medium. I t was to thesedisplacementcurrents that
Thomson, in his original investigation, attributed the magnetic
effects of moving charges. The particular system which he con-
sidered was that formed by a charged spherical conductor, moving
uniformly in a straight line. It was assumedthat the distribution
of electricity remains uniform over the surface during the motion,
and that the electric field in any position of the sphereis the same
as if the spherewere at rest; theseassumptionsare true so long as
quantities of order (vfe) 2 are neglected,where v denotesthe velocity
of the sphereand e the velocity oflight.
Thomson's method was to determine the displacementcurrents
in the spaceoutside the spherefrom the known valuesof the electric
field, and then to calculate the vector potential due to thesedisplace-
ment currents by meansof the formula
A = Iff (S'fT) dx' dy' dz',
1 Phil. Mag. xxvii (1889), p. 445 • ibid. ii (1901), p. 179; v (1903), p. 34
• Ann. d. Phys.xi (1901), p. I • Amer. Jour. Sci. xii (1901), p. 155
• Jour. dePhys.ii (1903), p. 641 • Phil. Mag. xi (1881), p. 229
306
'rHE FOLLOWERS OF MAXWELL
where8' denotesthe displacementcurrent at (x',y' z'). The magnetic

field wasthen determined by the equation
H = curlA.
A defect in this investigation was pointed out by FitzGerald,

who, in a short but most valuable note,l published a few months
afterwards, observed that the displacement currents of Thomson
do not satisfy the circuital condition. This is most simply seenby
considering the case in which the system consistsof two parallel
plates forming a condenser; if one of the plates is fixed, and the
other plate is moved towardsit, the electric field is annihilated in the
space over which the moving plate travels; this destruction of
electric displacement constitutes a displacement current, which,
consideredalone, is evidently not a closedcurrent. The defect, as
FitzGerald showed,may be immediately removed by assumingthat
a moving charge itself is to be counted as a current element; the
total current, thus composedof the displacement currents and the
convectioncurrent, is circuital. Making this correction, FitzGerald
found that the magnetic force due to a sphere of charge e moving
with velocity v along the axis of z is curl (0, 0, ev/r)-a formula
which shows that the displacement currents have no resultant
magnetic effect, since the term ev/r would be obtained from the
convectioncurrent alone.
The expressionsobtained by Thomson and FitzGerald were
correct only to the first order of the small quantity vic. The effect
of including terms of higher order was considered in 1888-9 by
Oliver Heaviside,2whosesolution may be derived in the following
manner:
Suppose that a charged system is in motion with uniform
velocity v parallel to the axis of z; the total current consistsof the
displacementcurrent _!.. aE
417 at
where E denotesthe electric force, and
the convection current pV where p denotes the volume-density of
electricity. So the equation which connects magnetic force with
electric current may be written (taking, as usual, electrostatic units
for E and electromagneticunits for H)
at
eE
= c curl H - 4rrpv.
1 Proc. Roy. Dublin Soc. iii (Nov. 1881), p. 250; FitzGerald's Scientiji& Writings,
p. 102
• Eric/rician, 23 Nov. 1888; Phil. Mag. xxvii (1889), p. 324
(823) 307
Eliminating E betweenthis and the equation
ecurl E = - - ,
aH
at
and remembering that H is here circuital, we have
1 alH 41T
- -- - VIH = -- curl pv.
el at l e
If, therefore, a vector-potential a be defined by the equation
1 cia 47T
- l - l - Via =-pv
e at e'
the magnetic force will be the curl of a; and from the equation for
a it is evident that the componentsa% and a" are zero, and that a%is
to be determined from the equation
ii%le' - Vias = 47T pv.
c
Now, let (x, y, {) denote co-ordinatesrelative to axeswhich are
parallel to the axes (x,y, z), and which move with the charged
bodies; then a% is a function of (x,y, ') only; so we have
a a a a
a.?; = 0" and at =- v a, ;
and the preceding equation is readily seento be equivalent to
ala% ala% ala%
cPv,
41T
axl + ayl + a'11 = -
where '1denotes (1 - v2/e2) -
equation, with '1 !,.
But this is simply Poisson's
substituted for z; so the solution may be tran-
scribedfrom the known solution of Poisson'sequation: it is
ea%-
_fff { ('1 _ '/)1
p' v' dx' tfy' d' /
+ (x _ X')2 + (y _ y')1 P'
the integrations being taken over all the spacein which there are
moving charges; or
- fff
ea%- {(' _ ")1 + (1 _
p' v' dx' tfy' d"
v21e2)(x -X')2 +
(1 _ v21e2) (y _ y')2p·
If the moving systemconsistsof a single chargee at the point, = 0,
this gives ev
ea% = r (1 - Vi sinl 0le2 r;'
where sin20 = (Xl + y2)lr'.
308
THE FOLLOVf.ERS OF ~VVELL
It is readily seenthat the lines of magnetic force due to the

moving point-charge are circles whose centres are on the line of
motion, the magnitude of the magneticforce being
ev (1 - v'Je') sin 8
er' (1 - v'sin' 8Je')i"
The electricforce is radial, its magnitude being
e(l - v'Je')
r2 (1 - v'sin' 8Je')!'
The fact that the electric vector due to a moving point-chargeis
everywhereradial led Heavisideto conclude that the samesolution
is applicable when the charge is distributed over a perfectly con-
ducting spherewhosecentre is at the point, the only changebeing
that E and H would now vanish inside the sphere. This inference
was subsequentlyfound by G. F. C. Searle to be incorrect; a
distribution of electric chargeon a moving spherecould in fact not
be in equilibrium if the electric force were radial, sincethere would
then be nothing to balance the mechanical force exerted on the
moving charge (which is equivalent to a current) by the magnetic
field. The moving systemwhich gives rise to the samefield as a
moving point-chargeis not a sphere,but an oblate spheroid whose
polar axis (which is in the direction of motion) bearsto its equatorial
axisthe ratio (1 - v2Ie2)! : 1,1
It was moreovershown by W. B. Morton 2 that in the caseof
the moving electrifiedsphere,the surface-densityis unaltered by the
motion, but the lines of force no longer leave the surfaceperpen-
dicularly.
The energyof the field surrounding a chargedsphereis greater
when the sphereis in motion than when it is at rest. To determine
the additional energy quantitatively (retaining only the lowest
significantpowersof vic), we have only to integrate, throughout the
spaceoutsidethe sphere,the expressionH2/87T, which representsthe
electrokineticenergyper unit volume: the result is e2v2/3ac2, where
e denotesthe charge,v the velocity and a the radius of the sphere.
It is evident from this result that the vvork required to be done
in order to communicatea given velocity to the sphereis greater
when the sphereis charged than when it is uncharged; that is to
1 cr. Searle,Phil. Trans. clxxxvii (1896),p. 675; and Phil. Mag. xliv (1897),p. 329.
On the theory of the moovingelectrifiedsphere,cr. alsoJ.J. Thomson, lUcent Researchesin

Elect. and Mag., p. 16; O. Heaviside,Electrical Papers, i, p. 446,and ii, p. 514; Euctromag.
Theory, i, p. 269; A. Schuster,Phil. Mag. xliii (1897),p. 1.
I Phil. Mag. xli (1896),p.488
309
say,the virtual massof the sphereis increasedby an amount 2e2/3ac2,

owing to the presenceof the charge. This may be regarded as
arising from the self-induction of the convection current which is
formed when the charge is set in motion. I t was suggestedby
J. Larmor 1 and by W. Wien 2 that the inertia of ordinary ponder-
able matter may ultimately prove to be of this nature, the atoms
being constituted of systemsof electrons.3 It may, however, be
remarked that this view of the origin of massis not altogethercon-
sistent with the principle that the electron is an indivisible entity.
For the so-calledself-induction of the sphericalelectronis really the
mutual induction of the convection-currentsproducedby the elements
of electric charge which are distributed over its surface; and the
calculation of this quantity presupposesthe divisibility of the total
charge into elementscapable of acting severally in all respectsas
ordinary electric charges; a property which appearsscarcelycon-
sistentwith the supposedfundamental nature of the electron.
In his (1889) paper in the Philosophical Magazine on the moving
charge, Heaviside gave for the first time the formula for the
mechanicalforceon an electric chargewhich is moving in a magnetic
field, namely-
charge X the vector product of the velocity and the magnetic
induction.
J. J.
Thomson had given it ashalf this amount in 1881.'
After the first attempt of J. J. Thomson to determine the field
produced by a moving electrified sphere,the mathematicaldevelop-
ment of Maxwell's theory proceededrapidly. The problemswhich
admit of solution in terms of known functions are naturally thosein
which the conducting surfacesinvolved have simple geometrical
forms-planes, spheresand cylinders.5
A result which was obtained by Horace Lamb,6 when investi-
gating electrical motions in a spherical conductor, led to interesting
consequences.Lamb found that if a spherical conductor is placed
in a rapidly alternating field, the induced currentsare almostentirely
1 Phil. Trans. c1xxxvi(1895),p. 697 • Arch. Nterl (3), v (1900),p. 96
• Experimentalevidencewhich wasat the time thought to support the view that the
inertia of electronsis purely electromagneticwasafterwardsfurnishedby W. Kaufmann,
Gott. Nach. (1901),p. 143; (1902),p. 291.
• Phil. Mag. xi (1881),p. 227
• cf., e.g., C. Niven, Phil. Trans. clxxii (1881),p. 307; H. Lamb, Phil. Trans. clxxiv
(1883),p. 519; J.J. Thomson,Proc. Lond. Math. Soc. xv (1884),p. 197; H. A. Rowland,
Phil. Mag. xvii (1884),p. 413; J.J.Thomson,Proc. Lond. Math. Soc. xvii (1886),p. 310;
xix (1888), p. 520; and many investigationsof Oliver Heaviside, collected in his
Electrical Papers.
• Loc. cit.
310
confinedto a superficiallayer ; and his result wasshortly afterwards

generalisedby Oliver Heaviside,l who showedthat whatever be the
form of a conductor rapidly alternating currents do not penetrate
far into its substance.2 The reasonfor this may be readily under-
stood: it is virtually an application of the principle 3 that a perfect
conductor is impenetrable to magnetic lines of force. No perfect
conductor is known to exist; but f. if the alternations of magnetic
force to which a good conductor suchas copper is exposedare very
rapid, the conductor has not time (so to speak) to display the
imperfection of its conductivity, and the magnetic field is therefore
unable to extendfar below the surface.
The same conclusion may be reached by different reasoning.5
When the alternations of the current are very rapid, the ohmic
resistanceceasesto playa dominant part, and the ordinary equations
connectingelectromotiveforce, induction and current are equivalent
to the conditionsthat the currentsshall be so distributed asto make
the electrokinetic or magnetic energy a minimum. Consider now
the caseof a single straight wire of circular cross-section. The
magneticenergyin the spaceoutside the wire is the samewhatever
be the distribution of current in the cross-section(so long as it is
symmetricalabout the centre), sinceit is the sameas if the current
were flowing along the central axis; so the condition is that the
magnetic energy in the wire shall be a minimum; and this is
obviouslysatisfiedwhen the current is concentratedin the superficial
layer, sincethen the magnetic force is zero in the substanceof the
WIre.
In spiteof the advanceswhich were effectedby Maxwell and his
earliest followers in the theory of electric oscillations, the gulf
between the classicalelectrodynamicsand the theory of light was
not yet completely bridged. For in all the casesconsideredin the
former science,energy is merely exchangedbetweenone body and
another, remaining within the limits of a given system; while in
optics the energy travels freely through space,unattached to any
material body. The first discoveryof a more complete connection
betweenthe two theorieswas made by FitzGerald, who argued that
if the unification which had been indicated by Maxwell is valid, it
ought to be possibleto generateradiant energy by purely electrical
1 Electrician, xiv (10Jan. 1885) p. 178
I The mathematical theory was given by Lord Rayleigh, Phil. Mag. xxi (1886), p. 381.
cr. Maxwell's Treatise, § 689.
• cr. p. 281
• As was first remarked by Lord Rayleigh, Phil. Mag. xiii (1882), p. 344
• cr.j. Stefan, Wiener Sit,omgsber. xcix (1890), p. 319; Ann. d. Phys. xli (1890), p. 400
31l
means; 'it seemshighly probable,' he said on 5 May 1882,1

'that the energy of varying currents is in part radiated into
space and so lost to us.' In 18832 he described methods by
which the radiant energycould be produced.3
FitzGerald's system is what has since become known as the
magnetic oscillator: it consistsof a small circuit, in which the strength
of the current is varied according to the simple periodic law. The
circuit will be supposedto be a circle of small area S, whosecentre
is the origin and whoseplane is the plane of xy; and the surround-
ing medium will be supposedto be free aether. The current may be
taken to be (in e.-s. units) of strength cA cos (2 t/T), so that the 7T
moment of the equivalent magnet (in e.-m. units) is SA cos (2 t/T). 7T
Now in the older electrodynamics,the vector-potentialdue to a mag-

netic moleculeof (vector) moment M placed at the origin is (I/47r)
curl (M/r), where r denotesdistancefrom the origin. The vector-
potential due to FitzGerald's magneticoscillator would thereforebe
(I/47r) curl K, whereK denotesa vector parallel to the axisof z, and
of magnitude (I/r) SA cos(2 t/T). 7TThe changewhich is involved
in replacing the assumptionsof the older electrodynamicsby those
of Maxwell's theory is in the presentcaseequivalent( to retarding
the potential; so that the vector-potential a due to the oscillator is
(I /47T) curl K where K is still directed parallel to the axis of z, and
is of magnitude
K =
SA
r27T (
cos T t - c •
r)
The electricforceE at any point of spaceis - } aa, and the magnetic
cat
forceH is curl a: sothat thesequantities may be calculatedwithout
difficulty. The electric energyper unit volume is E2/87T; perform-
ing the calculations, it is found that the value of this quantity
averagedover a period of the oscillation and alsoaveragedover the
surfaceof a sphereof radius r is
The part of this which is radiated is evidently that which IS pro-
I Trans. Roy. Dublin Soc. i (1883), p. 325

I Trans. Roy. Dublin Soc. iii (1883) p. 57; FitzGerald's ScientifU Writings, p. 122
• According to O. J. Lodge, Journal of the Wireless Society of London, iii (1922), p. 77,
FitzGerald in 1879 read a paper on the ' Impossibility' of producing electric waves, but
he struck out the' 1m ' afterwarcL..
• cf. pp. 268-9
312
portional to the inversesquareof the distance,!sothe averagevalue

of the radiant energyof electric type at distancer from the oscillator
is 21T8A2S2j3c4r2T4 per unit volume. The radiant energy of mag-
netic type may be calculatedin a similar way, and is found to have
the same value; so the total radiant energy at distance r is
41T3A2S2j3c4r2T4 per unit volume; and thereforethe energyradiated
in unit time is 161T4A2S2j3c3T4. This is small, unlessthe frequency
is very high; so that ordinary alternating currents would give no
appreciable radiation. FitzGerald, however, at the Southport
meeting of the British Association,in the sameyear 2 indicated a
methodby which the difficulty of obtaining currents of sufficiently
high frequency might be overcome: this was, to employ the
alternating currents which are produced when a condenseris dis-
charged; in fact, the oscillatory dischargeof a Leyden jar. He
remarked,however, that the difficulty would be in detecting such
waveswhen they were produced.
The FitzGerald oscillator constructedon this principle is closely
akin to the radiator afterwards developed with such successby
Hertz; the only differenceis that in FitzGerald's arrangementthe
condenseris usedmerely as the store of energy (its plates being so
close together that the electrostatic field due to the charges is
practically confined to the space between them), and the actual
sourceof radiation is the alternating magneticfield dueto the circular
loop of wire; while in Hertz's arrangement the loop of wire is
abolished,the condenserplates are at somedistanceapart and the
sourceof radiation is the alternating electrostaticfield due to their
charges.
In the study of electrical radiation, valuable help is afforded by
a generaltheorem on the transfer of energy in the electromagnetic
field, which was discoveredin 1884 by John Henry Poynting,Sand
independently,almost at the sametime, by Heaviside.4 We have
seenthat the older writers on electric currents recognisedtha1an
electric current is associatedwith the transport of energyfro91 one
place (e.g. the voltaic cell which maintains the current) to another
(e.g. an electric motor which is worked by the current) ; but they
supposedthe energyto be conveyedby the current itself within the
1 The other term, which is neglected, is very small comparedto the term retained,at
greatdistancesfrom the origin; it is what would be obtainedif the effectsof induction
of the displacementcurrentswereneglected: i.e. it is the energyof theforceddisplacement
currentswhich are produceddirectly by the variation of the primary current, and which
originatethe radiating displacementcurrents.
• Brit. Assoc. Rep., 1883,p. 404; FitzGerald'sScientific Writings, p. 129
• Phil. Trans. clxxv (1884),p. 343
• Electrician xiv (10Jan. and 21 Feb. 1885),pp. 178,306
313
wire, in much the sameway asdynamical energyis carried by water

flowing in a pipe; whereasin Maxwell's theory, the storehouseand
vehicle of energy is the dielectric medium surrounding the wire.
What Poynting achievedwas to showthat the flux of energyat any
place might be expressedby a simpleformula in termsof the electric
and magneticforcesat the place.
Denoting as usual by E the electric force, by D the electric
displacement,Dy-l{ the magnetic force and by B the magnetic
induction, the energystoredin unit volume of the medium is 1
(1/87T) ED + (1/87T) BH ;
so the increaseof this in unit time is (sincein isotropic media D is
proportional to E, and B is proportional to H)
or
where S denotesthe total current, and s the current of conduction;

or (in virtue of the fundamental electromagneticequations)
- (E • s) + (c/47T) (E. curl H) - (c/47T) (H. curl E),
or - (E. s) - (c/47T) div [E. H].
Now (E. s) is the amount of electric energy transformedinto heat
per unit volume per second; and therefore the quantity - (c/4:7T)
div [E . H] must representthe depositof energyin unit volume per
seconddue to the streamingof energy; which showsthat the flux
of energy is representedby the vector (c/47T) [E. HJ.1 This is
Poynting'stheorem: that the flux of energy at any place is represented by
the vector product of the electric and magnetic forces, multiplied by C/47T.2
1 cf. pp. 222, 224, 253. As usual, we use electrostatic units for D and E, and electro-
magnetic units for Band H.
I Of course any circuital vector may be added. A. McAulay (Phil. Trans. clxxxiii
[1892], p. 685) and H. M. Macdonald (Electric Waves, p. 72) propounded forms which
differ from Poynting's by a non-circuital vector. K. Birkeland (Ann. d. Phys. Iii [1894],
p. 357), showed that Poynting's expression for the energy flux is the only one which is a
function of the electric and magnetic field-strengths only.
I The analogue of Poynting's theorem in the theory of the vibrations of an isotropic
elastic solid may be easily obtained; for from the equation of motion of an elastic solid,
p ~e2'= -
at
(k + 4n/3) grad div e- n curl curl e,
it follows that
!. at + I
a, Up (ae)2 (k + ~n) (div e)" + In (curl el"} = - div W,
314
In the special caseof the field which surrounds a straight wire

carrying a continuouscurrent, the lines of magneticforce are circles
round the axisof the wire, while the lines of electricforce are directed
along the wire; henceenergy must be flowing in the medium in a
direction at right anglesto the axis of the wire. A current in any
conductor may therefore be regarded as consistingessentiallyof a
convergenceof electric and magnetic energyfrom the medium upon
the conductor, and its transformation there into other forms.
This associationof a current with motions at right anglesto the
wire in which it flows doubtless suggestedto Poynting the con-
ceptionsof a memoir which he published 1 in the following year.
When an electric current flowing in a straight wire is gradually
increasedin strength from zero, the surrounding space becomes
filled with lines of magnetic force, which have the form of circles
round the axis of the wire. Poynting, adopting Faraday's idea of
the physicalreality of lines of force, assumedthat theselines of force
arrive at their placesby moving outwards from the wire; so that
the magnetic field grows by a continual emissionfrom the wire of
lines of force, which enlargeand spreadout like the circular ripples
from the place where a stoneis dropped into a pond. The electro-
motive force which is associatedwith a changing magneticfield was
now attributed directly to the motion of the lines of force, so that
whereverelectromotiveforce is produced by changein the magnetic
field, or by motion of matter through the field, the electric intensity
is equal to the number of tubesof magneticforce intersectedby unit
length in unit time.
A similar conceptionwasintroduced in regard to lines of electric
force. It wasassumedthat any changein the total electric induction
through a curve is causedby the passageof tubes of force in or out
acrossthe boundary; so that whenever magnetomotive force is
1 Phil. Trans. clxxvi (1885),p. 277
whereW denotesthe vector
- (k + 4n/3) dive. '!! + n [[curl

at e. '!!J ;
at
and sincethe expressionwhich is differentiatedwith respectto t representsthe sumof the
kinetic and potentialenergiesper unit volumeof the solid (savefor termswhich give only
surface-integrals),
it is seenthat W is the analogueof the Poyntingvector. cr.:
K. Pearson,Mess. of Math. xix (1889),p. 31
W. Wien, Ann. d. Phys. xlv (1892),p. 685
O. Heaviside,Electrician, xxvii (3 July 1891)and xxix (29July 1892),reprinted in
Electromagnetic Theory, pp. 76, 247
O. Heaviside,Phil. Trans. clxxxiii (1892),p. 426
G. Mie, Wien Sitzunsb. cvii (1898),Abth. 2a, p. 1113
L. Donati, Bologna Mem. (5), vii (1899),p. 633
315
produced by change in the electric field, or by motion of matter

through the field, the magnetomotive force is proportional to the
number of tubes of electric force intersectedby unit length in unit
time.
Poynting, moreover,assumedthat when a steadycurrent C flows
in a straight wire, C tubesof electric force closein upon the wire in
unit time, and are there dissolved,their energy appearing as heat.
H E denote the magnitude of the electric force, the energyof each
tube per unit length is t E, so the amount of energybrought to the
wire is tCE per unit length per unit time. This is, however, only
half the energy actually transformed into heat in the wire; so
Poynting further assumedthat E tubes of magneticforce also move
in per unit length per unit time, and finally disappearby contraction
to infinitely small rings. This motion accountsfor the existenceof
the electric field; and since each tube (which is a closed ring)
contains energy of amount tc, the disappearanceof the tubes
account; for the remaining tCE units of energy dissipatedin the
WIre.
The theory of moving tubes of force was extensivelydeveloped
by Sir J. J. Thomson.l Of the two kinds of tubes-magnetic and
electric-which had been introduced by Faraday and used by
Poynting, ThOInsonresolvedto discard the former and employ only
the latter. This wasa distinct departurefrom Faraday'sconceptions,
in which, as we have seen,great significancewas attached to the
physical reality of the magnetic lines; but Thomson justified his
choice by inferencesdrawn from the phenomena of electric con-
duction in liquids and gases. As will appear subsequently,these
phenomena indicate that molecular structure is closely connected
with tubes of electrostaticforce-perhaps much more closely than
with tubes of magnetic force; and Thomson therefore decided to
regard magnetismas the secondaryeffect, and to ascribemagnetic
fields, not to the presenceof magnetic tubes, but to the motion of
electric tubes. In order to account for the fact that magneticfields
may occur without any manifestation of electric force, he assumed
that tubes exist in great numbers everywherein space,either in the
form of closedcircuits or elsetenninating on atoms,and that electric
force is only perceivedwhen the tubeshave a greater tendencyto lie
in one direction than in another. In a steady magnetic field the
positive and negative tubes might be conceived to be moving in
oppositedirections with equal velocities.
1 Phil. Mag. xxxi (1891), p. 149; Thomson'sRecentResearches
in Elect. and Mag. (1893),
chap. i
A beam of light might, from this point of view, be regarded

simply as a group of tubes of force which are moving with the
velocity of light at right anglesto their own length. Such a con-
ception almostamountsto a return to the corpusculartheory; but
sincethe tubeshavedefinite directionsperpendicular to the direction
of propagation, there would now be no difficulty in explaining
polarisation.
The energyaccompanyingall electric and magneticphenomena
was supposedby Thomson to be ultimately kinetic energy of the
aether; the electric part of it being representedby rotation of the
aether inside and about the tubes, and the magnetic part being
the energyof the additional disturbanceset up in the aether by the
movementof the tubes. The inertia of this latter motion he regarded
asthe causeof induced electromotiveforce.
There was, however, one phenomenonof the electromagnetic
field as yet unexplainedin terms of theseconceptions-namely, the
ponderomotiveforce which is exerted by the field on a conductor
carrying an electric current. Now any ponderomotiveforce consists
in a transferof mechanicalmomentumfrom the agent which exerts
the force to the body which experiencesit; and it occurred to
Thomsonthat the ponderomotiveforcesof the electromagneticfield
might be explained if the moving tubes of force, which enter a
conductorcarrying a current and are there dissolved,were supposed
to possess mechanicalmomentum,which could be yielded up to the
conductor. It is readily seenthat suchmomentummust be directed
at right anglesto the tube and to the magneticinduction-a result
which suggeststhat the momentum stored in unit volume of the
aethermay be proportional to the vector product of the electric and
magneticvectors.
For this conjecture reasonsof a more definite kind may be
given.1 We have already seen2 that the ponderomotiveforceson
material bodiesin the electromagneticfield may be accountedfor
by Maxwell's suppositionthat acrossany plane in the aether whose
unit normal is N, there is a stressrepresentedby
I I
PN = 4; (D . N) E - 87T (D. E) N + 4rrI (B. N) H - I
87T (B. H) N.
So long asthe field is steady(i.e. electrostaticor magnetostatic)the

1 The hypothesisthat the aetheris a storehouseof mechanicalmomentum,which was
first advancedby J. J. Thomson (Recent Researchesin Elect. and Mag. (18931,p. 13), was
afterwardsdevelopedby H. Poincare (Archives Netrl. (2], v [1900], p. 252), and by
M. Abraham (Gott, Nach. [1902], p. 20).
• cr. p. 272
resultant of the stressesacting on any elementof volume of the aether

is zero, so that the element is in equilibrium. But when the field is
variable, this is no longer the case. The resultant stresson the aether
contained within a surfaceS is
integrated over the surface; transforming this into a volume

integral, the term (1/47r) (D. N) E gives a term (1/4rr) div D . E
+ (1/41T) (D. V) E, where V denotesthe vector operator (%x, %y,
%z); and the first of these terms vanishes,since D is a circuital
vector; the term - (l/81T) (D. E) N givesin the volume integral a
term (I /81T) grad (D . E) ; and the magnetictermsgive similar results.
So the resultant force on unit volume of the aether is
I
47T(D . V) E + 81TI grad (D . E) + 47T(B.
I I
V) E + 81T grad (B . H)
which may be written

I
41T [curl E . D] + 41T1 [curl H . B] ;
or, by virtue of the fundamental equationsfor dielectrics,
4:C[ - 0:. D J+4;c[0~ . B J, or 4~c(0/ot)[D.B].

This result compelsus to adopt one of three alternatives: either
to modify the theory so as to reduce to zero the resultant force on an
element of free aether-this expedient has not met with general
favour; 1 or to assumethat the force in question setsthe aether in
motion-this alternative was chosen by Helmholtz,2 but is incon-
sistent with the theory of the aether which was generally received
in the closing years of the century; or lastly, with Thomson,3to
acceptthe principle that the aether is itself the vehicle of mechanical
momentum, of amount (1/4rrc) [D. B] per unit volume.·
Maxwell's theory was now being developed in ways which
could scarcelyhave been anticipated by its author. But although
every year added something to the superstructure,the foundations
remained much as Maxwell had laid them; the doubtful argument
by which he had sought to justify the introduction of displacement
1 It was, however,adoptedby G. T. Walker, Aberration and the Eleetromagrutic Field,
Camb., 1900.
• Berlin Sitzungsberiehte (1893),p. 649; Ann. d. Phys. liii (1894),p. 135. Helmholtz
supposedthe aetherto behaveasa frictionlessincompressiblefluid.
3 Loc. cit.
, As usual,we usee.-s.units for D and E, and e.-m.units for B and H.
318
currents was still all that was offered in their defence. In 1884,
however,the theory wasestablished1 on a different basisby Heinrich
Hertz (1857-94).
Hertz, the sonof a senator,had originally beenintended for the
professionof an architect; but when studying engineeringas part
of his course,he felt the attraction of pure science,and in 1880 he
took a doctorate in physicsat Berlin. He was so fortunate as to
attract the attention of Helmholtz, who in the sameyear appointed
him his assistant; in 1883 he becamea privat-docent at Kiel, and it
washere that he beganthe researcheswhich will now be described.
The train of Hertz's ideasresemblesthat by which Ampere, on
hearing of Oersted'sdiscoveryof the magnetic field produced by
electriccurrents,inferred that electriccurrentsshouldexert pondero-
motiveforceson eachother. Ampere argued that a current, being
competent to originate a magnetic field, must be equivalent to a
magnetin other respects; and thereforethat currents,like magnets,
shouldexhibit forcesof mutual attraction and repulsion.
Ampere'sreasoningrestson the assumptionthat the magnetic
field producedby a current is in all respectsof the samenature as
that produced by a magnet; in other words, that only one kind
of magneticforce exists. This principle of the ' unity of magnetic
force' Hertz now proposed to supplement by asserting that the
electricforce generatedby a changing magneticfield is identical in
nature with the electric force due to electrostatic charges; this
secondprinciple he called the ' unity of electric force.' 1I
Suppose,then, that a system of electric currents s exists in
otherwise empty space. According to the older theory, these
currents give rise to a vector-potential at> equal to Pot (l/c)s 3 ;
and the magnetic force H1 is the curl of a1; while the electric
force ElJ at any point in the field, producedby the variation of the
. lOat
currents,IS - - -- •
c at
It is now assumedthat the electric force so produced is indis-
tinguishable from the electric force which would be set up by
electrostatic charges, and therefore that the system of varying
currents exerts ponderomotive forces on electrostatic charges;
the principle of action and reaction then requires that electrostatic
1 Ann. d. Phys. xxiii (1884), p. 84: English versionin Hertz's Miscellaneous Papers,
translatedby D. E. Jonesand G. A. Schott,p. 273
I That an electric field is producedby a varying magneticfield was shownexperi-
mentallyby Lodgein 1889,by observingthe motion of a gold leaf in a varying magnetic

field j and by A. Righi, Nuovo Cimento,v (1901),p. 233.
I a = PQtfJ is usedto denotethe solutionof the equation Via + 4-rrfJ = o.
3 9
1
chargesshould exert ponderomotiveforces on a systemof varying

currents, and consequently(again appealing to the principle of the
unity of electric force) that two systemsof varying currents should
exert on eachother ponderomotiveforcesdue to the variations.
But just as Helmholtz,I by aid of the principle of conservation
of energy,deducedthe existenceof an electromotiveforce of induc-
tion from the existenceof the ponderomotiveforcesbetweenelectric
currents (i.e. variable electric systems),so from the existenceof
ponderomotive forces between variable systemsof currents (Le.
variable magneticsystems)we may infer that variations in the rate
of changeof a variable magneticsystemgive riseto inducedmagnetic
forces in the surrounding space. The analytical formulae which
determine theseforces will be of the same kind as in the electric
case; sothat the induced magneticforceH' is given by an equation
of the form
where hI' which is analogousto the vector-potential in the electric

case,is a circuital vector whose curl is the electric force E1 of the
variable magneticsystem. The value of hI is therefore (I j'hr) curl
Pot ~: so we have
1 a·
H' = - %c. ot2 curl Pot a1•
This must be added to HI. Writing ~ for the sum, HI + H', we

seethat ~ is the curl of a2, where
1 al
a. = a1 - %cl otl Pot al ;
and the electric force E2 will then be _! o~.

c at
This systemis not, however, final; for we must now perform
the processagain with these improved values of the electric and
magnetic forcesand the vector-potential; and so we obtain for the
magneticforce the value curl aa, and for the electric force the value
laaa
-- - where
c at'
1 a·
aa = a1 - 4~ --;-t Pot a.
TTC 0 2
1 ~ 1 ~
= al - %c. at. Pot a1 + (%c2)lot' Pot Pot al_
l cf. p. 218
320
This processmust again be repeated indefinitely; so finally we

obtain for the magneticforce H the value curl a, and for the electric
lOa
forceE the value - - - , where
C ot
1 01 1 oe
a = al - 4rrcZ otZ Pot al + (47TCZ).ote Pot Pot al
1 0'
- (47TC ott Pot Pot Pot a)
Z)
3 +.
It is evident that the quantity a thus defined satisfiesthe equation
VIa = Vial
o· a,
+ Ci1 otl
1 81 4rr
or VZa - - - Z a = - - S
c ot i c·
This equation may be written
•
curl H =c1 8t
oE 4rr
+ cs
W
'l he equatIOns
hlet . HIE
= cur a, = _ !oa
C
.
ot gIve
10H
curl E = - C--.ot
Theseare, however, the fundamental equationsof Maxwell's theory
in the form given in his memoir of 1868.1
That Hertz's deduction is ingenious and interesting will readily
be admitted. That it is conclusive may scarcely be claimed, for
the argument of Helmholtz regarding the induction of currents is
not altogether satisfactory; and Hertz, in following his master, is
on no surer ground.
In the courseof a discussion2 on the validity of Hertz's assump-
tions, which followed the publication of his paper, E. Aulinger 3
brought to light a contradiction betweenthe principles of the unity
of electric and of magnetic force and the electrodynamicsof Weber.
Consideran electrostatically charged hollow sphere,in the interior
of which is a wire carrying a variable current. According to Weber's
theory, the spherewould exert a turning couple on the wire; but
according to Hertz's principles, no action would be exerted, since
charging the spheremakesno difference to either the electric or the
1 cr. p. 258
• Lorberg, Ann. d. Ph)s. xxvii (1886),p. 666; xxxi (1887),p. 131. Boltzmann, ibid.
xxix (1886),p. 598
• Ann. d. Ph)s. xxvii (1886),p. 119
321
magneticforce in its interior. The experimentthus suggestedwould

be a crucial test of the correctnessof Weber's theory; it has the
advantageof requiring nothing but closedcurrents and electrostatic
chargesat rest; but the quantities to be observedwould be on the
limits of observationalaccuracy.
Mter his attempt to justify the Maxwellian equations on
theoretical grounds, Hertz (who in 1885 had becomeProfessorof
Physics at Karlsruhe) turned his attention to the possibility of
verifying them by direct experiment. His interest in the matter
had first been aroused some years previously, when the Berlin
Academy proposedas a prize subject' To establishexperimentally
a relation between electromagneticactions and the polarisation of
dielectrics.' Helmholtz suggestedto Hertz that heshouldattempt the
solution; but at the time Hertz sawno way of bringing phenomena
of this kind within the limits of observation. From this time forward,
however, the idea of electric oscillations was continually presentto
his mind; and in the spring of 1886he noticed an effect 1 which
formed the starting-point of his later researches. When an open
circuit was formed of a piece of copper wire, bent into the form of
a rectangle, so that the endsof the wire were separatedonly by a
short air-gap, and when this open circuit was connectedby a wire
with any point of a circuit through which the spark dischargeof an
induction coil was taking place, it was found that a spark passed
in the air-gap of the open circuit. This wasexplainedby Su~)posing
that the change of potential, which is propagated along the con-
necting wire from the induction coil, reachesone end of the open
circuit before it reachesthe other, so that a spark passesbetween
them; and the phenomenontherefore was regarded as indicating
a finite velocity of propagation of electric potential along wires.2
Continuing his experiments, Hertz 3 found that a spark could
be induced in the open or second""rycircuit evenwhen it wasnot in
metallic connection with the primary circuit in which the electric
1 Ann. d. Phys. xxxi (1887), p. 421 ; Hertz's Electric Waves, translated by D. E. Jones,
p. 29
• Unknown to Hertz, the transmission of electric waves along wires had been observed
in 1870 by Wilhelm von Bezold, Miinchen Sitzungsberichte,i (1870), p. 1I3; Phil. Mag. xl
(1870), p. 42. • If,' he wrote at the conclusion of a seriesof experiments, • electrical waves
be sent into a wire insulated at the end, they will be reflected at that end. The phenomena
which accompany this processin alternating discharges appear to owe their origin to the
interference of the advancing and reflected waves,' and, • an electric discharge travels with
the same rapidity in wires of equal length, without reference to the materials of which
these wires are made.'
The subject was investigated by O. J. Lodge and A. P. Chattock at almost the same
time as Hertz's experiments were being c.arried out; mention was made of their researches
at the meeting of the British Association in 1888. • loc. cit.
322
oscillationswere generated; and he rightly interpreted the phenom-

enon by showing that the secondarycircuit was of such dimensions
asto makethe free period of electric oscillationsin it nearly equal to
the period of the oscillationsin the primary circuit; the disturbance
which passedfrom one circuit to the other by induction would
consequentlybe greatly intensified in the secondarycircuit by the
effectof resonance.
The discoverythat sparksmay be produced in the air-gap of a
secondarycircuit, provided it has the dimensionsproper for reson-
ance,wasof great importance: for it supplieda method of detecting
electricaleffectsin air at a distancefrom the primary disturbance; a
suitabledetectorwasin fact all that wasneededin order to observethe
propagationof electricwavesin freespace,and therebydecisivelytest
the Maxwellian theory. To this work Hertz now addressedhimself.!
Unknown to him, he had been anticipated, about sevenyears
earlier, by David Edward Hughes (1830-1900), who had shown
that signals sent from a spark transmitter could be detected at
distances up to five hundred yards by a microphone contact
(essentiallywhat was later called a coherer), to which was connected
a telephonein which the signalswere heard. He claimed (rightly)
that the signalswere transmitted by electric wavesin the air. In
1879 and 1880 he demonstrated these experiments before the
Presidentof the Royal Society (Spottiswoode),Sir George Stokes
and Mr W. H. Preece,the Electrician to the PostOffice; unfortun-
ately they inclined to the opinion that the effectsmight be explained
by ordinary electromagneticinduction; and Hughes, discouraged,
did not print any accountof his work until long afterwards,2so that
the priority of publication belongsto Hertz.
The oscillator or primary source of the disturbancesstudied by
Hertz may be constructedof two sheetsof metal in the sameplane,
each sheet carrying a stiff wire which projects towards the other
sheetand terminates in a knob; the sheetsare to be excited by
connectingthem to the terminals of an induction coil. The sheets
may be regardedasthe two coatingsof a modified Leydenjar, with
air as the dielectric between them; the electric field is extended
throughout the air, instead of being confined to the narrow space
1 Sir Oliver Lodgewasabout this time independentlystudyingelectricoscillationsin
air in connectionwith the theory of lightning-conductors; cr. Lodge, Phil. Mag. xxvi
(1888),p. 217.
• In The Electrician, xliii (5 May 1899),p. 40
Solongbeforeas1842,joseph Henry, of Washington,had noticedthat the' inductive'
effectsof the Leydenjar dischargecould be observedat considerabledistances,and had
evensuggested a comparisonwith· a sparkfrom flint and steelin the caseof light.'
(823) 323
AETHER AND ELECTRlCrrY
betweenthe coatings,asin the ordinary Leydenjar. Such a disposi-

tion ensuresthat the systemshall lose a large part of its energy by
radiation at each oscillation.
As in the jar discharge,! the electricity surgesfrom one sheetto
the other, with a period proportional to (CL)I, where C denotesthe
electrostatic capacity of the systemformed by the two sheets,and
L denotesthe self-induction of the connection. The capacity and
induction should be made as small as possiblein order to make the
period small. The detector used by Hertz was that already
described, namely, a wire bent into an incompletely closedcurve,
and of suchdimensionsthat its free period of oscillation wasthe same
asthat of the primary oscillation, so that resonancemight take place.
Towards the end of the year 1887, when studying the sparks
induced in the resonatingcircuit by the primary disturbance,Hertz
noticed 2 that the phenomenawere distinctly modified when a large
massof an insulating substancewasbrought into the neighbourhood
of the apparatus; thus confirming the principle that the changing
electric polarisation which is produced when an alternating electric
forceactsonadielectriciscapableof displayingelectromagneticeffects.
Early in the following year (1888) Hertz determined to verify
Maxwell's theory directly by showing that electromagneticactions
are propagated in air with a finite velocity. 3 For this purpose he
transmitted the disturbance from the primary oscillator by two
different paths, viz. through the air and along a wire; and having
exposedthe detector to the joint influence of the two partial dis-
turbances, he observedinterference between them. In this way he
found the ratio of the velocity of electric wavesin air to their velocity
when conductedby wires; and the latter velocity he determinedby
observingthe distancebetweenthe nodesof stationary wavesin the
wire, and calculating the period of the primary oscillation. The
velocity of propagation of electric disturbancesin air wasin this way
shown to be finite and of the sameorder as the velocity of light.'
1 cr. p. 227 I Ann. d. Phys. xxxiv, p. 373; Electm Waves (English edn.), p. 95
I Ann. d. Phys. xxxiv (1888), p. 551 ; Electm Waves (English edn.), p. 107
• Hert,,'s experiments gave the value 45/28 for the ratio of the velocity of electric
waves in air to the velocity of electric waves conducted by the wires, and 2 X 1010 em.
per sec. for the latter velocity. These numbers were afterwards found to be open to
objection; Poincare (ComptesRendus,cxi [1890], p. 322) showed that the period calculated
by Hertz was v'2 X the true period, which would make the velocity of propagation in
air equal to that of light X v'2. Ernst Lecher (Wiener Berichte, 8 May 1890; Phil. Mag.
xxx [1890], p. 128), experimenting on the velocity of propagation of electric vibrations
in wires, found instead of Hertz's 2 X 1010 cm. per sec. a value within 2 per cent of
the velocity of light. E. Sarasin and L. de la Rive at Geneva (Archives desSc. Phys. xxix,
1893, pp. 358,441) finally proved that the velocities of propagation in air and along wires
are equal.
Later in 1888 Hertz 1 showed that electric waves in air are

reflected at the surface of a wall; standing waves may thus be
produced, and interference may be obtained between direct and
reflectedbeamstravelling in the samedirection.
The theoreticalanalysisof the disturbanceemitted by a Hertzian
radiator according to Maxwell's theory was given by Hertz in the
following year.2
The effectsof the oscillator are chiefly determined by the free
electric charges which, alternately appearing at the two sides,
generatean electric field by their presenceand a magneticfield by
their motion. In eachoscillation, asthe chargeson the polesof the
radiator increasefrom zero, lines of electric force, having their ends
on thesepoles,move outwards into the surrounding space. When
the chargeson the polesattain their greatestvalues,the lines cease
to issueoutwards, and the existing lines begin to retreat inwards
towardsthe poles; but the outer lines of force contract in such a
way that their upper and lower parts touch each other at some
distancefrom the radiator, and the remoter portion of eachof these
linesthus takesthe form of a loop; and when the rest of the line of
force retreats inwards towards the radiator, this loop becomes
detachedand is propagatedoutwardsasradiation. In this way the
radiator emits a seriesof whirl rings, which as they move grow
thinner and wider; at a distance,the disturbanceis approximately
a plane wave, the oppositesidesof the ring representingthe two
phasesof the wave. When one of theserings has becomedetached
from the radiator, the energy contained may subsequently be
regardedastravelling outwardswith it.
To discussthe problem analytically 3 we take the axis of the
radiator asaxisof z, and the centreof the spark gap asorigin. The
field may be regarded as due to an electric doublet formed of a
positive and an equal negative charge, displacedfrom each other
along the axis of the vibrator, and of moment
Ae-It sin (27TCtJ"),
the factor e-pt being insertedto representthe damping.
The simplest method of proceeding, which was suggestedby
FitzGerald,' is to form the retardedpotentialsc/> and a ofL. Lorenz.s
1 Ann. d. Phys. xxxiv (1888),p. 610; Electric Waves (Englishedn.), p. 124
2 ibid. xxxvi (1889),p. I; Electric Waves (Englishedn.), p. 137
2 ef. Karl Pearson and A. Lee, Phil. Trans. cxciii (1899),p. 165
• Brit. Assoc. Rep., Leeds (1890),p. 755
• ef. p. 268. They had beenemployedin the meantimeby Helmholtz, and by Lord
Rayleighin his Theory of Sound. The useof retarded potentialswas alsorecommended
in the following year by Poincare,ComptesRendus, cxiii (1891),p. 515.
325
Theseare determined in terms of the chargesand their velocitiesby

the equations A. 1: (e)t-rlc 1: 1 ( a;:;)
'I' = -r-' as = cr e at t-rlc
whenceit is readily shown that in the presentcase
ep = - of/a;:;, ca = (0,0, of/at),
where Ae-I<t-rlc) • 217
F = r SIll J:: (ct - r).
The electric and magnetic forces are then determined by the
equations
1 aa
E = grad 'I'
A. - - -
cat' H = curl a.
It is found that the electric force may be regarded as compounded

of a force ep2, parallel to the axis of the vibrator and depending at
any instant only on the distancefrom the vibrator, together with a
force epl sin () acting in the meridian plane perpendicular to the
radius from the centre, where epl dependsat any instant only on the
distancefrom the vibrator, and ()denotesthe anglewhich the radius
makeswith the axis of the oscillator. At points on the axis, and in
the equatorial plane, the electric force is parallel to the axis. At a
great distancefrom the oscillator, ep2" is small compared with epl' so
the wave is purely transverse. The magnetic force is directed along
circles whose centres are on the axis of the radiator; and its
magnitude may be represented in the form epa sin (), where 1>a
dependsonly on rand t; at great distancesfrom the radiator, cepa
is approximately equal to epl'
If the activity of the oscillator be supposedto be continually
.maintained, so that there is no damping, we may replacep by zero,
and may proceed as in the case of the magnetic oscillator 1 to
determine the amount of energy radiated. The mean outward flow
of energy per unit time is found to be !cA2(217j,\)'; from which it
is seenthat the rate of lossof energyby radiation increasesgreatly as
the wave-length decreases.
The action of an electrical vibrator may be studied by the aid of
mechanicalmodels. In one of these,devisedby Larmor,2 the aether
is representedby an' incompressibleelastic solid, in which are two
cavities, correspondingto the conductorsof the vibrator, filled with
incompressible fluid of negligible inertia. The electric force is
represented by the displacement of the solid. For such rapid
alternations as are here considered, the metallic poles behave as
1 cr. p. 312 • Proc. Camb. Phil. Soc.vii (1891), p. 165
326
perfectconductors; and the tangential componentsof electric force

at their surfacesare zero. This condition may be satisfiedin the
modelby supposingthe lining of eachcavity to be of flexible sheet-
metal,soasto be incapableof tangential displacement; the normal
displacementof the lining then correspondsto the surface-densityof
electricchargeon the conductor.
In order to obtain oscillationsin the solid resemblingthoseof an
electricvibrator, we may supposethat the two cavitieshave the form
of semicircular tubes forming the two halves of a complete circle.
Each tube is enlargedat each of its ends,so as to presenta front of
considerablearea to the correspondingfront at the end of the other
tube. Thus at eachend of one diameter of the circle there is a pair
of opposingfronts, which are separatedfrom each other by a ~hin
sheetof the elasticsolid.
The disturbancemay be originated by forcing an excessof liquid
into one of the enlarged endsof one of the cavities. This involves
displacingthe thin sheetof elasticsolid, which separatesit from the
opposingfront of the other cavity, and thus causinga corresponding
deficiency of liquid in the enlarged end behind this front. The
liquid will then surge backwards and forwards in each cavity
betweenits enlargedends; and, the motion being communicatedto
the elasticsolid, vibrations will be generatedresemblingthosewhich
are producedin the aether by a Hertzian oscillator.
In the latter part of the year 1888 the researchesof Hertz 1
yielded more completeevidenceof the similarity oCelectricwavesto
light. It wasshownthat the part of the radiation from an oscillator
which was transmitted through an opening in a screenwas propa-
gated in a straight line, with diffraction effects. Of the other
propertiesof light, polarisation existedin the original radiation, as
was evident from the manner in which it was produced; and
polarisation in other directions was obtained by passingthe waves
through a grating of parallel metallic wires; the component of the
electric force parallel to the wires was absorbed, so that in the
transmitted beam the electric vibration was at right angles to the
wires. This effect obviously resembledthe polarisation of ordinary
light by a plate of tourmaline. Refraction was obtained by passing
the radiation through prismsof hard pitch.2
1 Ann. d. Phys. xxxvi (1889), p. 769; Electric Waves (English edn.), p. 172
• O.J. Lodge andJ. L. Howard in the same year showed that electric radiation might
be refracted and concentrated by means oflarge lenses; cf. Phil. Mag. xxvii (1889), p. 48.
Other demonstrations of optical properties possessedby electromagnetic waves were
given by A. Righi (Bologna Ace. Sci. Mem. iv [1894], p. 487), and A. Garbasso (Annalen d.
Phys. !iii [1894], p. 534, and Atti Ace. Torino, xxx [1895], p. 442).
327
As Larmor remarked,1 ' the discoveriesof Hertz left no further

room for doubt that the physical schemeof Maxwell ... constituted
a real formulation of the underlying unity in physical dynamics.'
The old question as to whether the light vector is in, or at right
anglesto, the plane of polarisation 2 now presenteditself in a new
aspect. The wave-front of an electric wave contains two vectors,
the electric and magnetic, which are at right anglesto each other.
Which of theseis in the plane of polarisation? The answer was
first given theoretically by Heaviside,3by working out the formulae
for the ratio of the reflected to the incident wave, and then experi-
mentally by FitzGerald and Trouton,' who found on reflecting
Hertzian waves from a wall of masonry that no reflection was
obtained at the polarising angle when the vibrator wasin the plane
of reflection. The inference from this is that the magnetic vector
is in the plane of polarisation of the electric wave, and the electric
vector is at right anglesto the plane of polarisation. An interesting
development followed in 1890, when O. Wiener 5 succeededin
photographing stationary wavesof light. The stationary waveswere
obtained by the composition of a beam incident on a mirror with
the reflected beam, and were photographedon a thin film of trans-
parent collodion, placed closeto the mirror and slightly inclined to
it. If the beam usedin such an experiment is plane-polarised,and
is incident at an angle of 45°, the stationary vector is evidently that
perpendicular to the plane of incidence; but Wiener found that
under theseconditions the effect was obtained only when the light
waspolarised in the plane of incidence; sothat the chemicalactivity
must be associatedwith the vector perpendicular to the plane of
polarisation-i.e. the electric vector.
In 1890 and the following years appearedseveralmemoirs on
the fundamental equationsof electromagnetictheory. Hertz, after
presenting 6 the generalcontentof Maxwell's theoryfor bodiesat rest,
extended7 the equationsto material bodiesin motion in the field.
1 Brit. Ass. Rep. (1900), p. 617 I cf. p. 154 et sqq.
I Phil. Mag. xxv (Feb. 1888), p. 130 (in Note A) • Nature, xxxix (1889), p. 391
• Ann. d. Phys. xl (1890), p. 203. cf. a controversy regarding the results: Comptes
Rendus,cxii (1891), pp. 186,325,329,365,383,456; and Ann. d. Phys. xli (1890), p. 154 ;
xliii (1891), p. 177; xlviii (1893), p. 119
• Gott. Nach. (1890), p.106; Ann. d. Phys. xl (1890), p. 577; Electric Waves(English edn.),
p. 195. In this memoir Hertz advocated the form of the equations which Maxwell had
used in his paper of 1868 (cf. supra, p. 258) in preference to the earlier form, which involved
the scalar and vector potentials.
7 Ann. d. Phys. xli (1890), p. 369; Electric Waves (English edn.), p. 241
The prepagation of light through a moving dielectric had been discussedpreviously,
on the basis of Maxwell's equations for moving bodies, by J. J. Thomson (Phil. Mag. ix
[1880], p. 284; Proc. Camb. Phil. Soc. v [1385], p. 250).
328
In a really comprehensiveand correct theory, asHertz remarked,

a distinction should be drawn between the quantities which specify
the state of the aether at every point, and those which specify the
stateof the ponderablematter entangled with it. This anticipation
has beenfulfilled by later investigators; but Hertz consideredthat
the time wasnot ripe for sucha completetheory, and preferred, like
Maxwell, to assumethat the state of the compound system-matter
plus aether-ean be specified in the same way when the matter
movesas when it is at rest; or, as Hertz himself expressedit, that
'the aether contained within ponderable bodies moves with
them.'
Maxwell's own hypothesis with regard to moving systems1
amountedmerely to a modification in the equation
!aB = _ curl E
c at '
which representsthe law that the electromotive force in a closed
circuit is measuredby the rate of decreasein the number of lines
of magnetic induction which passthrough the circuit. This law is
true whether the circuit is at rest or in motion; but in the latter
case,the E in the equation must be taken to be the electromotive
force in a stationary circuit whose position momentarily coincides
with that of the moving circuit; and since an electromotive force
(lfe) [w. B] is generatedin matter by its motion with velocity w in
a magneticfield B, we seethat E is connectedwith the electromotive
force E' in the moving ponderablebody by the equation
E' = E + c~[w. B],
so that the equation of electromagneticinduction In the moving

body is
~~= - c curl E' +
curl [w. BJ.
Maxwell made no change in the other electromagneticequations,

which thereforeretained the customaryforms
aD
D =€E', div D = %p, c curl H = 7ft + %s,
Hertz, however, impressedby the duality of electric and magnetic

phenomena,modified the last of theseequationsby assumingthat
1 cr. p. 258-9
329
a magnetic force (lfc) [D. w] is generated in a dielectric which

moves with velocity w in an electric field; such a force would
be the magnetic analogueof the electromotiveforce of induction.
A term involving curl [D. w] is then introduced into the last
equation.
The theory of Hertz resembles in many respects that of
Heaviside,l who likewiseinsistedmuch on the duplex nature of the
electromagneticfield, and was in consequencedisposedto accept
the term involving curl [D. w] in the equationsof moving media.
Heaviside recognisedmore clearly than his predecessorsthe dis-
tinction betweenthe force E', which determinesthe flux D, and the
force E, whose curl representsthe electric current; and, in con-
formity with his principle of duality, he made a similar distinction
betweenthe magnetic force H', which determinesthe flux B, and
the force H, whosecurl representsthe' magnetic current.' This
distinction, as Heavisideshowed,is of importance when the system
is acted on by 'impressed forces,' such as voltaic electromotive
forcesor permanent magnetisation; theselatter must be included
in E' -and H', since they help to give rise to the fluxesD and B;
but they must not be included in E and H, sincetheir curls are not
electric or magneticcurrents; so that in generalwe have
E' = E + e, H' =H + h,
wheree and h denotethe impressedforces.
Developingthe theory by the aid of theseconceptions,Heaviside
was led to make a further modification. An impressedforce is best
definedin termsof the energywhich it communicatesto the system;
thus, if e be an impressedelectric force, the energycommunicated
to unit volume of the electromagneticsystemin unit time is the
scalar product of e and the electric current. In order that this
equation may be true, it is necessaryto regard the electric current
in a moving medium as composed of the conduction current,
displacementcurrent, convection current, and also of the term in
curl [D. w], whose presencein the equation we have already
noticed. This may be called the current of dielectric convection. Thus
the total current is
1 aD
S = 47r 8t + s + pw + 47r1 curl [D . w],
1 Heaviside'sgeneraltheory waspublishedin a series of papersin the Electrician, from
1885onwards. His earlier work was republishedin his Electrical Papers (2 vols., 1892),
and his ElectromagneticTheory (2 vols, 1894). Mention may bespeciallymadeof a memoir
in Phil. Trans. clxxxiii (1892),p. 423.
330
where pw denotesthe convection current; and the equation con-

necting current with magneticforce is
c curl (H' - ho) = 4-rrS,
where ho denotes the impressedmagnetic forces other than that
induced by motion of the medium.
We must now considerthe advanceswhich were effectedduring
the period following the publication of Maxwell's Treatise in some
of the specialproblemsof electricity and optics.
We have seen1 that Maxwell accountedfor the rotation of the
plane of polarisation of light in a medium subjectedto a magnetic
field K by adding to the kinetic energy of the aether, which is
representedby lp (oe/ot)2, a term la (~. curl oe/oO), where u is a
magneto-opticconstantcharacteristicof the substancethrough which
the light is transmitted,and 0/00 standsfor Kx%x Ky%y +
Kz%;;;. +
This theory was developed further in 1879 by FitzGerald,2 who
brought it into closer connection with the electromagnetictheory
of light by identifying the curl of the displacemente of the aethereal
particles with the electric displacement; the derivate of e with
respectto the time then correspondsto the magnetic force. Being
thus in possessionof a definitely electromagnetic theory of the
magnetic rotation of light, FitzGerald proceeded to extend it so
asto takeaccountofa closelyrelatedphenomenon. In l876J. Kerr 3
had shownexperimentallythat whenplane-polarisedlight is regularly
reflectedfrom either pole of an iron electromagnet,the reflectedray
hasa componentpolarisedin a plane at right anglesto the ordinary
reflectedray. Shortly after this discoveryhad been made known,
FitzGerald 4. had proposedto explain it by meansof the sameterm
in the equationsas that which accountsfor the magnetic rotation of
light in transparent bodies. His argument was that if the incident
plane-polarisedray be resolvedinto two rays circularly polarised in
oppositesenses,the refractive index will have different values for
these two rays, and hence the intensities after reflection will be
different; so that on recompounding them, two plane-polarised
rays will be obtained--one polarised in the plane of incidence, and
the other polarisedat right anglesto it.
The analytical discussionof Kerr's phenomenon, which was
given by FitzGerald in his memoir of 1879, was based on these
1 cr. p. 276
• Phil. Trans.clxxi (1879),p. 691; FitzGerald's ScientiftWritings, p. 45
• Phil. Mag. (5), iii (1877),p. 321
• Proc.R. S. xxv (1877), p. 447 ; FitzGerald's Seientift Writings, p. 9
331
ideas; the mostessentialfeaturesof the phenomenonwereexplained,

but the investigationwasin somerespectsimperfect.l
A new and fruitful conception was introduced in 1879-80,
when H. A. Rowland 2 suggesteda connectionbetweenthe magnetic
rotation of light and the phenomenonwhich had been discovered
by his pupil Hall.3 Hall's effect may be regarded as a rotation of
conduction currents under the influence of a magnetic field; and
if it be assumedthat displacementcurrentsin dielectricsare rotated
in the sameway, the Faraday effect may evidently be explained.
Consideringthe matter from the analytical point of view, the Hall
effect may be representedby the addition of a term k [K . S] to the
electromotiveforce, where K denotesthe impressedmagneticforce,
and S denotesthe current; so Rowland assumedthat in dielectrics
there is an additional term in the electric force, proportional to
[K . °0~J. i.e. proportional to the rate of increaseof [K . D]. Now
it is universally true that the total electric force round a circuit is
proportional to the rate of decreaseof the total magneticinduction
through the circuit; so the total magnetic induction through the
circuit must contain a term proportional to the integral of [K . D]
taken round the circuit; and therefore the magnetic induction at
any point must contain a term proportional to curl [K.D]. We
may thereforewrite B = H +
u curl [K . D],
where u denotes a constant. But if this be combined with the

customaryelectromagneticequations
1 oD
curl H = cat' curl E =- c1fit,
oB
D = £E,
and all the vectors except B be eliminated (K being treated as a
constant), we obtain the equation
02B
otl = (elf£) VIB + eu curl (0IB/ot08),
1 ef. Larmor's remarks in his ' Report on the Action of Magnetism on Light,' Brit.
Assoc. Rep., 1893; and his editorial comments in FitzGerald's Scientific Writings. Larmor
traced to its source an inconsistency in the equations by which FitzGerald had represented
the boundary conditions at an interface between the media. FitzGerald had indeed made
the mistake, similar to that which was so often made by the earlier writers on the elastic-
solid theory of light, of forgetting that when a medium is assumed to be incompressible,
the condition of incompressibility must be introduced into the variational equation of
motion (as was done, supra, p. 157). Larmor showed that when this correction was made,
new terms (resembling the terms in p, supra, p. 157) made their appearance; and the
inconsistency in the equations was thus removed.
• Amer. Jour. Math. ii, p. 354; iii, p. 89; Phil. Mag. xi (1881), p. 254
• ef. p. 289
332
where NoB stands for (Kx%x Ky%y +

Kz%z); and this is+
identical with the equation which Maxwell had given 1 for the
motion of the aether in magnetised media. It follows that the
assumptionsof Maxwell and of Rowland, different th,oughthey are
physically, lead to the sameanalytical equations-at any rate sofar
asconcernspropagation through a homogeneousmedium.
The connections of Hall's phenomenon with the magnetic
rotation of light, and with the reflection of light from magnetised
metals,were extensivelystudied 2 in the yearsfollowing the publica-
tion of Rowland's memoir; but it was not until the modern theory
of electronshad been developed that a satisfactory representation
of the molecular processesinvolved in magneto-optic phenomena
wasattained.
1 cr, p. 277
• The theory of Basset(Phil. Trans. clxxxii (1891], p. 371) was, like Rowland's, based
on the idea of extending Hall's phenomenon to dielectric media. An objection to this
theory was that the tangential component of the electromotive force was not continuous
across the interface between a magnetised and an unmagnetised medium; but Basset
subsequently overcame this difficulty (Nature, Iii [1895], p. 618; liii [1895], p. 130;
Amer. Jour. Math. xix [1897], p. 60)-the effect analogous to Hall's being introduced into
the equation connecting electric displacement with electric force, so that the equation
took the form
E = (If<) D+ u [ K. 8~J
Basset,in 1893 (Proc. Camh. Phil. Soc. viii, p. 68), derived analytical expressionswhich
represent Kerr's magneto-optic phenomenon by substituting a wmplex quantity for the
refractive index in the formulae applicable to transparent magnetised substances.
The magnetic rotation of light and Kerr's phenomenon have been investigated also by
R. T. Glazebrook, Phil. Mag. xi (1881), p. 397; by J. J. Thomson, Recent Researches,
p. 482; by D. A. Goldhammer, Ann. d. Phys. xlvi (1892), p. 71 ; xlvii (1892), p. 345 ;
xlviii (1893), p. 740; I (1893), p. 772; by P. Drude, Ann. d. Phys. xlvi (1892), p. 353 ;
xlviii (1893), p. 122; xlix (1893), p. 690; Iii (1894), p. 496; by C. H. Wind, Vers/agen
Kon. Akad. Amsterdam, 29 Sept 1894; by Reiff, Ann. d. Phys. Ivii (1896), p. 281; by
J. G. Leathem, Phil. Trans. cxc (1897), p. 89; Trans. Camb. Phil. Soc. xvii (1898), p. 16 ;
and by W. Voigt in many memoirs, and in his treatise, Magneto- und Elektro-optik.
Larmor's report presented to the British Association in 1893 has been already mentioned.
In most of the later theories the equations of propagation of light in magnetised metals
are derived from the two fundamental electro-magnetic equations
c curl H = 41TS, - c curl E = 8H

-- ;
at
the total current S being assumed to consist of a part (the displacement-current) pro-
portional to~, a part (the conduction current) proportional to E, and a part proportional
at
to the vector product of E and the magnetisation.
Various mechanical models of media in which magru:to-optic phenomena take place
have beendevised at different times. W. Thomson (Proc. Lond. Math. Soc.vi (1875], p.190)
investigated the propagation of waves of displacement along a stretched chain whose links
contain rotating fly-wheels: cr. also Larmor, Proc. Lond. Math. Soc. xxi (1890), p. 423 ;
xxiii (1891), p. 127; F. Hasenohrl, Wien Sit;;ungsberichte, cvii, 2a (1898), p. 1015 ;
W. Thomson (Kelvin), Phil. Mag. xlviii (1899), p. 236, and Baltimore Lectures; and
FitzGerald, Electrician, 4 Aug. 1899 = FitzGerald's Scienti,% Writings, p. 481.
333
The allied phenomenonof rotary polarisation in naturally active

bodies was investigated in 1892 by Goldhammer.l It will be
remembered2 that in the elastic-solid theory of Boussinesq,the
rotation of the plane of polarisation of sugar solutionshad been
representedby substituting the equation
e' = Ae + B curl e
in place of the usual equation
e' = Ae.
Goldhammer now proposed to represent rotatory power In the
electromagnetictheory by substituting the equation
E = (If€) D + kcurl D,
in place of the customaryequation
E = (If€) D :
the constantk being a measureof the natural rotatory power of the
substanceconcerned. The remaining equationsare asusual,
oD oH
c curl H = at' - ccurl E =-
ot
Eliminating Hand E, we have
02D
ot =
2 (Cl/E) V2D + kc l
VI curl D.
For a plane wave which is propagatedparallel to the axis of x, this

equation reducesto
olD.z = t! ()2D~ _ kcl o3D~
otl E ox l ox 3 ,
{ 02Dz =~ 02Dz + kc2. 03D" •

ot' E ox2 ox 3 ,
and, as MacCullagh had shown in 1836,3these equations are

competentto representthe rotation of the plane of polarisation.
In the closingyearsof the nineteenthcentury, the generaltheory
of aether and electricity assumeda new form. But beforediscussing
the memoirs in which the new conception was unfolded, we shall
considerthe progresswhich had been made sincethe middle of the
century in the study of conduction in liquid and gaseousmedia.
1 Jour. de Phys. (3), i, pp. 205, 345 I cf. p. 168 I cf. p. 159
334
ChapterXI
CONDUCTION IN SOLUTIONS AND GASES, FROM
FARADAY TO THE DISCOVERY OF THE ELECTRON
THEhypothesiswhich Grothussand Davy had advanced1 to explain

the decompositionof electrolytes was open to serious objection in
morethan onerespect. Since the electric force was supposedfirst to
dissociatethe moleculesof the electrolyte into ions, and afterwards
to setthem in motion toward the electrodes,it would seemreasonable
to expect that doubling the electric force would double both the
dissociationof the moleculesand the velocity of the ions, and would
therefore quadruple the electrolysis-an inference which is not
verified by observation. Moreover it might be expected, on
Grothuss' theory, that some definite magnitude of electromotive
force would be requisite for the dissociation,and that no electrolysis
at all would take place when the electromotiveforce was below this
value, which again is contrary to experience.
A way of escapefrom these difficulties was first indicated, in
1850,by Alex. Williamson,2who suggestedthat in compound liquids
decompositionsand recombinationsof the moleculesare continually
taking place throughout the whole massof the liquid, quite inde-
pendently of the application of an external electric force. An atom
of one element in the compound is thus paired now with one and
now with another atom of another element, and in the intervals
betweenthesealliancesthe atom may be regarded as entirely free.
In 1857this idea wasmade by R. Clausius,3of Zurich, the basisof a
theory of electrolysis. According to it, the electromotive force
emanatingfrom the electrodesdoesnot effect the dissociationof the
electrolyteinto ions, since a degreeof dissociationsufficient for the
purpose already exists in consequenceof the perpetual mutability
of the moleculesof the electrolyte. Clausiusassumedthat theseions
are in oppositeelectric conditions; the applied electric force there-
fore causesa general drift of all the ions of one kind towards the
anode, and of all the ions of the other kind towards the cathode.
These opposite motions of the two kinds of ions constitute the
galvanic current in the liquid.
1 cr.
p. 76
• Phil. Mag. xxxvii (1850),p. 350; Liebig's Annalen d. Chern. u. Pharm. Ixxvii (1851),
p.37 • Ann. d. Phys. ci (1857), p. 338; Phil. Mag. xv (1858),p. 94
335
CONDUCTION IN SOLUTIONS AND GASES
The merits of the Williamson-Clausiushypothesiswere not fully

recognisedfor many years; but it becamethe foundation of that
theory of electrolysiswhich was generally acceptedat the end of the
century.
Meanwhile another aspectof electrolysiswasreceivingattention.
It had long been known that the passageof a current through an
electrolytic solution is attended not only by the appearanceof the
products of decompositionat the electrodes,but also by changesof
relative strength in different parts of the solution itself. Thus in the
electrolysisof a solution of copper sulphate, with copper electrodes,
in which copper is dissolvedoff the anode and depositedon the
cathode,it is found that the concentration of the solution diminishes
near the cathode,and increasesnear the anode. Someexperiments
on the subject were made by Faraday 1 in 1835; and in 1844-
it was further investigated by J. F. Daniell and W. A. Miller, 2
who explained it by assertingthat the cation and anion have not
(ashad previduslybeensupposed)the samefacility of moving to their
respectiveelectrodes; but that in many casesthe cation appears
to movebut little, while the transport is effectedchiefly by the anion.
This idea was adopted by W. Hittorf (1824-1914), of Munster,
who, in the years 1853 to 1859, published 3 a seriesof memoirson
the migration of the ions. Let the velocity of the anions in the
solution be to the velocity of the cationsin the ratio u : u. Then
it is easilyseenthat if (u +
u) moleculesof the electrolyteare decom-
posedby the current, and yielded up as ions at the electrodes,u of
these moleculeswill have been taken from the fluid on the side of
the cathode,and u of them from the fluid on the sideof the anode.
By measuring the concentration of the liquid round the electrodes
after the passageof a current, Hittorf determined the ratio ulu in a
large number of casesof electrolysis."
The theory of ionic movementswas advanceda further stageby
F. W. Kohlrausch 5 (1840-1910), ofWurzburg. Kohlrauschshowed
that although the ohmic specific conductivity k of a solution
diminishesindefinitely asthe strength of the solution is reduced,yet
the ratio kim, where m denotesthe number of gramme-equivalents6
1 Exp. Res. §§ 525-30
I Phil. Trans. (1844), p. 1 ; cf. also PouilIet, Campus Rendus, xx (1845), p. 1544
• Ann. d. Phys.lxxxix (1853), p. 177 ; xcviii (1856), p. 1; ciii (1858), p. 1; cvi ('859),
pp. 337,513
• The ratio v/(u+ v) was termed by Hittorf the transport number of the anion.
• Ann. d. Phys. vi (1879), pp. I, 145. The chief results had been communicated to the
Academy of Gottingen in 1876 and 1877.
• A gramme-equivalent means a mass of the salt whose weight in grammes is the
molecular weight divided by the valency of the ions.
336
FROM FARADAY TO THE ELECTRON
of salt per unit volume, tends to a definite limit when the solution
is indefinitely dilute. This limiting value may be denoted by A.
He further showedthat A may be expressedasthe sum of two parts,
oneof which dependson the cation, but is independentof the nature
of the anion; while the other dependson the anion, but not on the
cation-a fact which may be explained by supposingthat, in very
dilute solutions,the two ionsmoveindependentlyunder the influence
of the electric force. Let u and v denote the velocitiesof the cation
and anion respectively,when the potential differenceper cm. in the
solution is unity; then the total current carried through a cube of
unit volume is mE(u + v), where E denotes the electric charge
carried by onegramme-equivalentofion.l Thus mE(u + v) = total
current = k = m?\, or ?\ = E(u + v). The determination of v/u by
the method of Hittorf, and of (u +
v) by the method of Kohlrausch,
made it possibleto calculate the absolute velocities of drift of the
ionsfrom experimentaldata.
Meanwhile, important advancesin voltaic theory were being
effectedin connectionwith a different classof investigations.
Supposethat two mercury electrodesare placed in a solution of
acidulated water, and that a difference of potential, insufficient to
produce continuous decompositionof the water, is set up between
the electrodesby an external agency. Initially a slight electric
current-the polarising current,2rs-observed; but after a short time
it ceases; and after its cessationthe state of the systemis one of
electricalequilibrium. It is evident that the polarising current must
in someway have setup in the cell an electromotiveforce equal and
opposite to the external difference of potential; and it is also
evident that the seat of this electromotive force must be at the
electrodes,which are now said to be polarised.
An abrupt fall of electric potential at an interface between two
media, such as the mercury and the solution in the present case,
requiresthat there should be a field of electric force, of considerable
intensity, within a thin stratum at the interface; and this must owe
its existenceto the presenceof electric charges. Since there is no
electricfield outsidethe thin stratum, there must be asmuch vitreous
as resinouselectricity present; but the vitreous chargesmust pre-
ponderateon one side of the stratum, and the resinouschargeson
the other side; so that the systemas a whole resemblesthe two
coatingsof a condenserwith the intervening dielectric. In the case
of the polarisedmercury cathodein acidulated water, there must be
on the electrodeitself a negativecharge; the surfaceof this electrode
1 i.e. E is 96,580coulombs • cr. p. 76
337
in the polarised state may be supposedto be either mercury, or

mercury coveredwith a layer of hydrogen. In the solution adjacent
to the electrode,there must be an excessof cationsand a deficiency
of anions,so asto constitute the other layer of the condenser; these
cations may be either mercury cations dissolvedfrom the electrode,
or the hydrogen cationsof the solution.
It wasshownin 1870 by Cromwell FleetwoodVarley (1828-1883) 1
that a mercury cathode,thus polarised in acidulatedwater,showsa
tendencyto adopt a definite superficialform, asif the surface-tension
at the interface betweenthe mercury and the solution were in some
way dependent on the electric conditions. The matter was more
fully investigated in 1873 by a young French physicist, then pre-
paring for his inaugural thesis,Gabriel Lippmann.2 In Lippmann's
instrumental disposition, which is called a capillary electrometer,
mercury electrodesare immersed in acidulated water; the anode
Ho has a large surface,while the cathode H hasa variable surfaceS
small in comparison. When the external electromotive force is
applied, it is easilyseenthat the fall of potential at the large electrode
is only slightly affected, while the fall of potential at the small
electrode is altered by polarisation by an amount practically equal
to the external electromotive force. Lippmann found that the
constant of capillarity of the interface at the small electrodewas a
function of the external electromotive force, and therefore of the
differenceof potential betweenthe mercury and the electrolyte.
Let V denotethe external electromotiveforce; we may without
lossof generality assumethe potential of Ho to be zero, so that the
potential of H is - V. The state of the systemmay be varied by
altering either V or S; we assumethat these alterations may be
performed independently, reversibly and isothermally and that the
state of the large electrodeHo is not altered thereby. Let de denote
the quantity of electricity which passesthrough the cell from Ho to
H, when the state of the systemis thus varied; then if E denotethe
available energy of the system,and y the surface-tensionat H, we
have
dE = ydS Vde, +
y being measuredby the work required to increasethe surfacewhen
no electricity flows through the circuit.
In order that equilibrium may be re-establishedbetween the
electrodeand the solution when the fall of potential at the cathode
1 Phil. Trans. c1xi (1871), p. 129

• ComptesRendus, !xxvi (1873), p. 1407; Phil. Mag. xlvii (1874), p. 281 ; Ann. de Chim.
e( de Phys. v (1875), p. 494; xii (1877), p. 265
338
is altered,it will be necessarynot only that somehydrogen cations

shouldcomeout of the solution and be depositedon the electrode,
yielding up their charges,but also that there should be changesin
the clusteringof the chargedionsof hydrogen,mercury and sulphion
in the layer of the solution immediately adjacent to the electrode.
Each of thesecircumstancesnecessitates a flow of electricity in the
outer circuit; in the onecaseto neutralisethe chargesof the cations
deposited,and in the other caseto increasethe surfacedensity of
electricchargeon the electrode,which forms the oppositesheetof the
quasi-condenser.Let Sf(V) denotethe total quantity of electricity
which hasthus flowed in the circuit when the external electromotive
forcehasattained the value V. Then evidently
de = d { Sf (V) } ;
so
dE = {y +V f(V) } dS + VSf' (V) dV.
Sincethis expressionmust be an exact differential, we have
doy
dV +f(V) = 0;
so that - dy/dV is equal to that flux of electricity per unit of new
surfaceformed, which will maintain the surfacein a constant con-
dition (V being constant) when it is extended. Integrating the
previousequation, we have
E=s{ y-V/~-}.
Lippmann found that when the external electromotive force was
applied, the surface tension increased at first, until, when the
external electromotive force amounted to about one volt, the
surfacetensionattained a maximum value, after which it diminished.
He found that d2y/dV2 was sensiblyindependent of V, so that the
curve which representsthe relation betwe~ny and V is a parabola.1
The theory so far is more or lessindependentof assumptionsas
to what actually takesplace at the electrode; on this latter question
many conflicting views have been put forward. In 1878 Josiah
Willard Gibbs,2of Yale (1839-1903), discussedthe problem on the
suppositionthat the polarising current is simply an ordinary electro-
lytic conduction current, which causesa liberation of hydrogen
from the ionic form at the cathode. If this be so, the amount of
electricity which passesthrough the cell in any displacementmust
1 Lippman, ComplesReruius,xcv (1882), p. 686
• Trans. COM. Acad. iii (1876-8), pp. 108,343; Gibbs'sScientiM Papers, i, p. 55
(823) 339
be proportional to the quantity of hydrogen which is yielded up to

the electrodein the displacement; sothat dy j dV mustbeproportional
to the amount of hydrogendepositedper unit area of the electrode.1
A different view of the physical conditions at the polarised
electrodewas taken by Helmholtz 2 who assumedthat the ions of
hydrogenwhich are brought to the cathodeby the polarisingcurrent
do not give up their chargesthere, but remain in the vicinity of the
electrode,and form onefaceof a quasi-condenser of which the other
faceis the electrodeitself.3 If a denotethe surfacedensity of electricity
on either face of this quasi-condenser,we have, therefore,
de =- d(Sa); so a = dyjdV
This equation showsthat when dyfdV is zero-i.e. when the surface-
tensionis a maximum-a must be zero; that is to say, there must
be no differenceof potential betweenthe mercury and the electro-
lyte. The external electromotive force is then balanced entirely
by the discontinuity of potential at the other electrode Ho; and
thus a method is suggestedof measuringthe latter discontinuity of
potential. All previous measurementsof differencesof potential
had involved the employment of more than one interface; and it
was not known how the measureddifferenceof potential should be
distributed among these interfaces; so that the suggestionof a
means of measuring single differencesof potential was a distinct
advance, even though the hypotheseson which the method was
basedwere somewhatinsecure.
A further consequencededucedby Helmholtz from this theory
leadsto a secondmethod of determining the differenceof potential
betweenmercury and an electrolyte. If a mercury surfaceis rapidly
extending, and electricity is not rapidly transferred through the
electrolyte, the electric surface-densityin the double layer must
rapidly decrease,sincethe samequantity of electricity is being dis-
tributed over an increasingarea. Thus it may be inferred that a
rapidly extending mercury surfacein an electrolyte is at the same
potential asthe electrolyte.
This conception is realised in the dropping electrode, in which a
1 This is embodied in equation (690) of Gibbs's memoir.
• Berlin Monatsber. (1881), p. 945; Wiss. Abh. i, p. 925; Ann. d. Phys. xvi. (1882),
p.31. cr. also Planck, Ann. d. Phys. xliv (1891), p. 385
• The conception of double layers of electricity at the surface of separation of two
bodies had been already applied by Helmholtz to explain various other phenomena-
e.g. the Volta contact difference of potential of two metals, frictional electricity and
, electric endosmose,' or the transport of fluid which occurs when an electric current is
passed through two conducting liquids separated by a porous barrier. cr. Helmholtz,
Berlin Monatsberichte, 27 Feb. 1879; Ann. d. Phys. vii (1879), p. 337; Helmholtz, Wiss.
Abh. i, p. 855
340
jet of mercury, falling from a reservoir into an electrolytic solution,

is so adjusted that it breaks into drops when the jet touches the
solution. According to Helmholtz' conclusionthere is no difference
of potential between the drops and the electrolyte; and therefore
the difference of potential between the electrolyte and a layer of
mercury underlying it in the samevesselis equal to the difference of
potential between this layer of mercury and the mercury in the
upper reservoir,which differenceis a measurablequantity.
I t will be seenthat according to the theoriesboth of Gibbs and
of Helmholtz, and indeed according to all other theories on the
subject,! dyfdV is zero for an electrode whose surface is rapidly
increasing-e.g. a dropping electrode; that is to say, the difference
of potential betweenan ordinary mercury electrodeand the electro-
lyte, when the surface-tensionhas its maximum value, is equal to
the difference of potential between a dropping electrode and the
sameelectrolyte. This result has been experimentally verified by
variousinvestigators,who haveshownthat the applied electro-motive
forcewhenthe surface-tensionhasits maximum value in the capillary
electrometer,is equal to the electromotive force of a cell having
as electrodesa large mercury electrodeand a dropping electrode.
Another memoir which belongsto the sameperiod of HelmholtZ'
1 e.g. that of E. Warburg, Ann. d. Phys. xli (1890), p. I. In this it is assumedthat the
electrolytic solution near the electrodes originally contains a salt of mercury in solution.
When the external electromotive force is applied, a conduction current passesthrough the
electrolyte, which in the body of the electrolyte is carried by the acid and hydrogen ions.
At the cathode, Warburg supposed that the hydrogen ions react with the salt of mercury,
reducing it to metallic mercury, which is deposited on the electrode. Thus a considerable
change in concentration of the salt of mercury is caused at the cathode. At the anode,
the acid ions carrying the current attack the mercury of the electrode, and thus increase
the local concentration of mercuric salt; but on account of the size of the anode this
increase is trivial and may be neglected.
Warburg thus supposed that the electromotive force of the polarised cell is really that
of a concentration cell, depending on the different concentrations of mercuric salt at the
electrodes. He found dy/dY to be equal to the amount of mercuric salt at the cathode per
unit area of cathode, divided by the electrochemical equivalent of mercury. The
equation previously obtained is thus presented in a new physical interpretation.
Warburg connected the increase of the surface-tension with the fact that the surface-
tension between mercury and a solution always increaseswhen the concentration of the
solution is diminished. His theory, of course, leads to no conclusion regarding the absolute
potential difference between the mercury and the solution, as HelmholtZ' does.
At an electrode whose surface is rapidly increasing-e.g. a dropping electrode--
Warburg supposedthat the surface-density of mercuric salt tends to zero, so dy/dV is zero.
The explanation of dropping electrodes favoured by Nernst, Beilage zu den Ann. d. Phys.
Iviii (1896), p. 1,wasthat the difference of potential corresponding to the equilibrium between
the mercury and the electrolyte is instantaneously established; but that ions are with-
drawn from the solution in order to form the double layer necessaryfor this, and that
theseions are carried down with the drops of mercury, until the upper layer of the solution
is so much impoverished that the double layer can no longer be formed. The impoverish-
ment of the upper layer of the solution was actually observed by W. Palmaer, Zeitseh. Phys.
Chern. xxv (1898), p. 265; xxviii (1899), p. 257; xxxvi (1901), p. 664.
341
career,and which hasled to important developments,wasconcerned

with a special classof voltaic cells. The most usual type of cell is
that in which the positive electrodeis composedof a different metal
from the negative electrode, and the evolution of energy depends
on the difference in the chemical affinities of thesemetals for the
liquids in the cell. But in the classof cells now consideredI by
Helmholtz, the two electrodesare composedof the samemetal (say,
copper); and the liquid (say,solution of copper sulphate) is more
concentrated in the neighbourhood of one electrode than in the
neighbourhoodof the other. When the cell is in operation, the salt
passesfrom the placesof high concentration to the placesof low
concentration, so as to equaliseits distribution; and this process
is accompaniedby the flow of a current in the outer circuit between
the electrodes. Such cells had been studied experimentally by
JamesMosera short time previously 2 to Helmholtz' investigation.
The activity of the cell is due to the fact that the available
energy of a solution dependson its concentration; the molecules
of salt, in passingfrom a high to a low concentration, are there-
fore capableof supplying energy,just asa compressedgasis capable
of supplying energywhen its degreeof compressionis reduced. To
examine the matter quantitatively, let nf(n/V) denote the term in
the available energy of a solution, which is due to the dissolution
of n gramme-moleculesof salt in a volume Y of pure solvent; the
function f will of course depend also on the temperature. Then
when dn gramme-moleculesof solvent are evaporated from the
solution,the decreasein the availableenergyof the systemis evidently
equal to the available energy of dn gramme-moleculesof liquid
solvent, less the available energy of dn gramme-moleculesof the
vapour of the solvent, together with nf (n/Y) lessnf {n/(Y - vdn)},
where v denotesthe volume of one gramme-moleculeof the liquid.
But this decreasein available energymustbe equalto the mechanical
work supplied to the external world, which is dn. PI (V' - v), if
PI denote the vapour pressureof the solution at the temperaturein
question, and v' denote the volume of one gramme-moleculeof
vapour. We have therefore
dn . PI (v' - v) = - available energy of dn gramme-moleculesof
solventvapour
+ available energy of dn gramme-moleculesof
liquid solvent
+ nf(n/Y) - nf{n/(Y - vdn)}.
1 Berlin Monatsber. (1877), p. 713; Phil. Mag. (5), v (1878), p. 348; reprinted with
a~ditions in Ann. d. Phys. iii (1878),p. 201 I Ann. d. Phys. iii (1878),p. 216
342
Subtracting from this the equation obtained by making n zero,

we have
dn. (PI - Po) (v' - v) = nf(nfV) - nf{ n/(V - vdn) },
where Po denotesthe vapour pressureof the pure solvent at the
temperaturein question; so that
(PI - Po) (v' - v) = - (nl/VI)f'(n/V) v.
Now, it is known that when a salt is dissolvedin water, the vapour
pressureis loweredin proportional to the concentrationof the salt-
at any rate when the concentrationis small: in fact, by a law due to
Raoult, (Po - Pt)/Po is approximately equal to nv/V; so that the
previousequation becomes
PoV (v' - v) = nf'(n/V).
Neglectingv in comparisonwith v', and making useof the equation
of stateof perfect gases(namely,
Pov' = RT
where T denotes the absolute temperature, and R denotes the
constantof the equation of state),we have
f'(n/V) = RTV/n,
and therefore
f(n/V) = RT log (n/V).
Thus in the availableenergyof one gramme-moleculeof a dissolved
salt, the term which dependson the concentration is proportional
to the logarithm of the concentration; and hence,if in a concen-
tration cell one gramme-moleculeof the salt passesfrom a high
concentrationC2 at one electrodeto a low concentration C1 at the
other electrode,its available energy is thereby diminished by an
amount proportional to log (C2/Cl). The energy which thus dis-
appearsis given up by the systemin the form of electrical work;
and thereforethe electromotiveforce of the concentrationcell must
be proportional to log (C2/Cl). The theory of solutions and their
vapour pressurewasnot at the time sufficiently developedto enable
Helmholtz to determinepreciselythe coefficientof log (C2/C1) in the
expression.l
An important advancein the theory of solutionswas effectedin
1 The formula given by Helmholtz was that the electromotive force of the cell is
equal to b (1 - n) v log (cllcl), where cl and cl denote the concentrations of the solution
at the electrodes, v denotes the volume of one gramme of vapour in equilibrium with the
water at the temperature in question, n denotes the transport number for the cation
(Hittorf's lin), and b denotes q x the lowering of vapour pressure when one gramme-
equivalent of salt is dissolved in q grammes of water, where q denotes a large number.
343
1887, by a young Swedishphysicist,SvanteArrhenius (1859-1927).1

Interpreting the properties discoveredby Kohlrausch 2 in the light
of the ideas of Williamson and Clausiusregarding the spontaneous
dissociation of electrolytes, Arrhenius inferred that in very dilute
solutionsthe electrolyte is completely dissociatedinto ions, but that
in more concentratedsolutionsthe salt is lesscompletelydissociated;
and that asin all solutionsthe transport of electricity in the solution
is effected solely by the movement of ions, the equivalent conduc-
tivity 3 must be proportional to the fraction which expressesthe
degreeof ionisation. By aid of theseconceptionsit becamepossible
to estimatethe dissociationquantitatively, and to constructa general
theory of electrolytes.
Contemporary physicistsand chemistsfound it difficult at first
to believethat a salt existsin dilute solution only in the form of ions,
e.g. that the sodium and chlorine exist separatelyand independently
in a solution of common salt. But there is a certain amount of
chemical evidencein favour of Arrhenius' conception. For instance,
the testsin chemical analysisare really testsfor the ions; iron in
the form of a ferrocyanide, and chlorine in the form of a chlorate,
do not respond to the characteristic tests for iron and chlorine
respectively,which are really the testsfor the iron and chlorine ions.
( The general acceptanceof Arrhenius' views was hastenedby
the advocacyof W. Ostwald, who brought to light further evidence
in their favour. For instance,all permanganatesin dilute solution
show the same purple colour; and Ostwald considered their
absorptionspectrato be identical'; this identity is easilyaccounted
for on Arrhenius' theory, by supposingthat the spectrumin question
is that of the anion which correspondsto the acid radicle. The blue
colour which is observedin dilute solutions of copper salts, even
when the strong solution is not blue, may in the same way be
ascribed to a blue copper cation. A striking instanceof the same
kind is afforded by ferric sulphocyanide; here the strong solution
showsa deep red colour, due to the salt itself; but on dilution the
colour disappears,the ions being colourless.
If it be granted that ions can have any kind of permanent
existencein a salt solution, it may be shownfrom thermodynamical
1 Zeitschrift fUr phys. Chern. i (1887), p. 631. Previous investigations, in which the
theory was to some extent foreshadowed, were published in Bihang till Svenska Vet. Ak.
Forh. viii (1884), nos. 13 and 14.
I cf. p. 336
I i.e. the ohmic specific conductivity of the solution divided by the number of gramme-
equivalents of salt per unit volume
• Examination of the spectra with higher dispersion does not altogether confirm this
conclusion.
344
considerationsthat the degree of dissociation must increase as the

dilution increases,and that at infinite dilution there must be com-
plete dissociation. For the available energy of a dilute solution of
volume V, containing nl gramme-moleculesof one substance, n2
gramme-moleculesof another, and so on, is (as may be shown by
an obviousextensionof the reasoningalready employed in connec-
tion with concentration cells) 1
J:,n,cp, (T)
, + R TJ:,n,
, log (n,[V) + the available energy
possessed by the solvent before the introduction of the solutes,
where cp, (T) depends on T and on the nature of the Tth solute,
but not on V, and R denotes the constant which occurs in the
equationof stateof perfectgases. When the systemis in equilibrium,
the proportions of the reacting substanceswill be so adjusted that
the available energy has a stationary value for small virtual altera-
tionsSnl, Sn2, • • • of the proportions; and therefore
o = J:,Snr • cpr (T) + RTJ:,Sn,. log (nr/V) + RTJ:,Sn,.
Applying this to the caseof an electrolytein which the disappearance
of one molecule of salt (indicated by the suffix I) gives rise to one
cation (indicated by the suffix 2) and one anion (indicated by the
suffix 3)' we have Snl = - Sn2 = - Sna; so the equation becomes
o = cpdT) - cpa(T)- CPa (T) + RT log (nlV/n2na) - RT,
or n1V /n2na = a function of T only.
Since in a neutral solution the number of anions is equal to the
number of cations,this equation may be written
n33 = Vn1 X a function of T only;
it showsthat when V is very large (so that the solution is very
dilute), ~ is very large compared with nl; that is to say, the salt
tendstowards a state of complete dissociation.
The ideas of Arrhenius contributed to the successof Walther
Nernst 2 (1864-1941) in perfecting Helmholtz' theory of concentra-
tion cells,and representingtheir mechanismin a much more definite
fashionthan had beendone heretofore.
In an electrolytic solution let the drift velocity of the cations
under unit electric force be u, and that of the anions be v, so that
the fraction u/(u + v) of the current is transported by the cations,
1 cr. pp. 342-3
• Zeitschr.fiir phys. Chern.ii (1888), p. 613; iv (1889), p. 129; Berlin Sitzungsberichte
(1889), p. 83; Ann. d. Phys. xlv (1892), p. 360. cr. also Max Planck, Ann. d. Phys. xxxix
(1890)1..p,161; xl (1890), p. 561
345
and the fraction v/(u + v) by the anions. If the concentration of

the solution be C1 at oneelectrode,and Cz at the other, it followsfrom
the formula previously found for the available energy that one
gramme-ion of cations,in moving from one electrodeto the other,
is capable of yielding up an amount 1 RT log (C2/Cl) of energy;
while one gramme-ion of anions going in the opposite direction
must absorb the same amount of energy. The total quantity of
work furnished when one gramme-moleculeof salt is transferred
from concentration C2 to concentration C1 is therefore
u - v RT 10 ~.
u + v g C1
The quantity of electric charge which passesin the circuit when
one gramme-moleculeof the salt is transferred is proportional to
the valencyv of the ions,and the work furnishedis proportional to the
product of this charge and the electromotive force E of the cell ;
so that in suitable units we have
E = RT u - v 10 c~.
v u +
v g C1
A typical concentration cell to which this formula may be
applied may be constituted in the following way: let a quantity
of zinc amalgam, in which the concentration of zinc is C1, be in
contact with a dilute solution of zinc sulphate, and let this in turn
be in contact with a quantity of zinc amalgam of concentration C2"
When the two massesof amalgam are connectedby a conducting
wire outside the cell, an electric current flows in the wire from the
weak to the strong amalgam,2while zinc cations passthrough the
solution from the strong amalgam to the weak. The electromotive
force of such a cell, in which the current may be supposedto be
carried solelyby cations,is
Not content with the derivation of the electromotiveforce from

considerationsof energy, Nernst proceeded to supply a definite
mechanical conception of the processof conduction in electrolytes.
The ions are impelled by the electric force associatedwith the
gradient of potential in the electrolyte. But this is not the only
1 The correct law of dependence of the available energy on the temperature was by
this time known.

• It will hardly be necessaryto remark that this supposed direction of the current is
purely conventional.
forcewhich actson them; for, sincetheir availableenergydecreases

asthe concentrationdecreases, there must be a forCl~assistingevery
processby which the concentrationis decreased. The matter may
be illustrated by the analogyof a gascompressedin a cylinder fitted
with a piston; the availableenergyof the gasdecreases asits degree
of compressiondecreases; and therefore that movement of the
piston which tends to decreasethe compressionis assistedby a
force-the 'pressure' of the gas on the piston. Similarly, if a
solution were contained within a cylinder fitted with a piston
which is permeableto the pure solvent but not to the solute, and
if the whole were immersedin pure solvent, the available energy
of the systemwould be decreasedif the piston were to move out-
wardsso as to admit more solventinto the solution; and therefore
this movement of the piston would be assistedby a force-the
, osmoticpressureof the solution' asit is called.1
Consider,then, the caseof a single electrolyte supposedto be
perfectly dissociated; its state will be supposedto be the sameat
all points of any plane at right anglesto the axis of x. Let v denote
the valency of the ions, and V the electric potential at any point.
Since2 the available energy of a given quantity of a substancein
very dilute solution dependson the concentration in exactly the
sameway as the available energy of a given quantity of a perfect
gasdependson its density,it foilows that the osmoticpressurep for
eachion is determinedin termsof the concentrationand temperature
by the equation of stateof perfectgases
Mp = RTe,
where M denotesthe molecular weight of the salt, and e the mass
of salt per unit volume.
Considerthe cations contained in a parallelepiped at the place
x, whosecross-sectionis of unit area and whoselength is dx. The
mechanicalforce acting on them due to the electricfield is - (vefM)
dVfdx . dx, and the mechanical force on them due to the osmotic
pressureis - dpfdx . dx. If u denote the velocity of drift of the
cationsin a field of unit electric force, the total amount of charge
which would be transferredby cations acrossunit area in unit time
under the influence of the electric forcesalone would be - (uvefM)
dVf dx; so,under the influence of both forces,it is
_.!!...v...e
M dx
(dV + RT de)
cvdx'
I cf. van't Hoff, SvenskaVet.-Ak. Handlingar xxi (1886), no. 17; Zeitsehrijtfir Phys.
Chern. i (1887),p. 481
I As followsfrom the expression obtained, supra, p. 343
347 ~
Similarly, if v denote the velocity of drift of the anions in a unit

electric field, the charge transferred acrossunit area in unit time
by the anionsis
VVC(_
dV RT~) +
M dx cvdx·
We have therefore,if the total current be denotedby i,
I
•
= - (u+ v) -MdxdVvC
- - (u -
RT dc
v) --
M dx'
or
_ tJ'!'! dx= Mdx i + u - v RT t!! dx
dx (u + v)vc u +v vc dx .
The first term on the right evidently representsthe product of the
current into the ohmic resistanceof the parallelepiped dx, while
the second term representsthe internal electromotive force of the
parallelepiped. It follows that if r denotethe specificresistance,we
must have u +
v = M/ TVC,
in agreement with Kohlrausch's equation 1; while by integrating
the expressionfor the internal electromotiveforceof the parallelepiped
dx, we obtain for the electromotive force of a cell whose activity
dependson the transferenceof electrolyte between the concentra-
tions Cl and (;2' the value
u-vRTj!lf£dx
u+v v cdx'
or u - v RT I c.
u + v -;- og~,
in agreementwith the result already obtained.
It may be remarked that although the current arising from a
concentrationcell which is kept at a constant temperatureis capable
of performing work, yet this work is provided, not by any diminution
in the total internal energy of the cell, but by the abstraction of
thermal energy from neighbouring bodies. This indeed (as may
be seenby referenceto W. Thomson'sgeneral equation of available
energy) 2 must be the casewith any systemwhoseavailable energy
is exactly proportional to the absolutetemperature.
The advanceswhich were effected in the last quarter of the
nineteenth century in regard to the conduction of electricity through
liquids, considerablethough theseadvanceswere, may be regarded
as the natural developmentof a theory which had long been before
the world. It was otherwisewith the kindred problem of the con-
• cf. p. 337 • cf. p. 217
348
duction of electricity through gases; for although many generations

of philosophershad studied the remarkable effects which are pre-
sentedby the passageof a current through a rarefied gas,it was not
until recent times that a satisfactory theory of the phenomena was
discovered.
Some of the electricians of the earlier part of the eighteenth
century performed experiments in vacuous spaces; in particular,
Hauksbee1 in 1705 observeda luminosity when glassis rubbed in
rarefied air. But the first investigatorsof the continuous discharge
through a rarefied gas seem to have been Gottfried Heinrich
Grummert (17l9-76?) in 1744,2and Watson,Swho, by means of
an electricalmachine,senta current through an exhaustedglasstube
three feet long and three inchesin diameter. 'It was,' he wrote, , a
most delightful spectacle,when the room was darkened, to seethe
electricity in its passage: to be able to observenot, as in the open
air, its brushesor pencils of rays an inch or two in length, but here
the coruscationswere of the whole length of the tube between the
plates,that is to say,thirty-two inches.' Its appearancehe described
asbeing on different occasions' of a bright silver hue; resembling
very much the most lively coruscationsof the aurora borealis'; and
'forming a continued arch of lambent flame.' His theoretical
explanation was that the electricity 'is seen, without any preter-
natural force, pushing itself on through the vacuum by its own
elasticity, in order to maintain the equilibrium in the machine '-a
conceptionwhich follows naturally from the combination of Watson's
one-fluid theorywith the prevalent doctrineof electrical atmospheres.
4.
A different explanation was put forward by Nollet, who per-

formed electrical experimentsin rarefied air at about the sametime
as Watson,Sand saw in them a striking confirmation of his own
hypothesis of efflux and afflux of electric matter. 6 According to
Nollet, the particles of the effluent stream collide with thoseof the
affluent streamwhich is moving in the oppositedirection; and being
thus violently shaken,are excited to the point of emitting light.
Almost a century elapsedbefore anything more was discovered
regarding the dischargein vacuous spaces. But in 1838 Faraday,?
while passing a current from the electrical machine between two
brassrods in rarefied air, noticed that the purple haze or stream of
1 Phil. Trans. xxiv (1705), p. 2165; Fra. Hauksbee,Physico-Mechanical Experiments,
London, 1709
I ef. J. H. Winkler, die Starke d. elektr. Kraft d. Wassersin gliisernen Gefiissen, Lips., 1746
I Phil. Trans. xlv (1748),p. 93; xlvii (1752),p. 362
• cr. ch. ii • Nollet, Recherchessur l' Electricitt, 1749,troisiemediscoun

• ef. p. 44 • Phil. Trans. (1838),p. 125; Exp. Res. i, § 1526
349
light which proceededfrom the positive pole stoppedshort beforeit

arrived at the negative rod. The negative rod, which was itself
coveredwith a continuousglow, wasthus separatedfrom the purple
column by a narrow dark space: to this, in honour of its discoverer,
the nameFaraday's dark spacehasgenerallybeengiven by subsequent
writers.
That vitreous and resinouselectricity give rise to different types
of dischargehad long beenknown; and indeed, as we have seen,l
it wasthe study of thesedifferencesthat led Franklin to identify the
electricity of glasswith the superfluity of fluid, and the electricity of
amber with the deficiencyof it. But phenomenaof this classare in
general much more complex than might be supposedfrom the
appearancewhich they presentat a first examination; and the value
of Faraday's discovery of the negative glow and dark spacelay
chiefly in the simple and definite character of thesefeaturesof the
discharge,which indicated them as promising subjectsfor further
research. Faraday himself felt the importance of investigationsin
this direction. 'The resultsconnectedwith the different conditions
of positiveand negativedischarge,'he wrote,2, will havea far greater
influence on the philosophy of electrical sciencethan we at present
imagine.'
Twenty more years, however, passedbefore another notable
advancewasmade. That a subjectsofull of promiseshouldprogress
so slowly may appear strange; but one reasonat any rate is to be
found in the incapacity of the air-pumps then in useto rarefy gases
to the degreerequired for effectivestudy of the negativeglow. The
invention of Geissler'smercurial air-pump in 1855 did much to
removethis difficulty; and it wasin Geissler'sexhaustedtubesthat
Julius Plucker,3 of Bonn, studied the dischargethree yearslater.
It had been shown by Sir Humphry Davy in 1821 4 that one
form of electric discharge-namely, the arc betweencarbon poles-
is deflected when a magnet is brought near to it. PlUcker now
performed a similar experiment with the vacuum discharge,and
observeda similar deflection. But the most interestingof his results
were obtained by examining the behaviour of the negativeglow in
the magnetic field; when the negative electrodewas reducedto a
single point, the whole of the negative light becameconcentrated
along the line of magneticforcepassingthrough this point. In other
1 cr.p. 47 • Exp. Res.i, § 1523

, Ann. d. Phys. ciii (1858), pp. 88, 151; civ (1858), pp. 113,622; cv (1858), p. 67;
cvii (1859), p. 77; Phil. Mag. xvi (1858), pp. 119,408; xviii (1859), pp. 1,7
« Phil. Trans. cxi (1821), p. 425
350
words,the negative glow disposeditself as if it were constituted of

flexible chainsof iron filings attached at one end to the cathode.
Plucker noticed that when the cathode was of platinum, small
particles were torn off it and deposited on the walls of the glass
bulb. 'It is most natural,' he wrote, , to imagine that the magnetic
light is formed by the incandescenceof theseplatinum particles as
they are torn from the negative electrode.' He likewise observed
that during the dischargethe walls of the tube, near the cathode,
glowedwith a phosphorescentlight, and remarked that the position
of this light wasaltered when the magneticfield waschanged. This
led to another discovery; for in 1869 PlUcker'spupil, W. Hittorf,l
having placed a solid body between a point-cathode and the phos-
phorescentlight, was surprisedto find that a shadowwas cast. He
rightly inferred from this that the negative glow is formed of rays
(glow rays,Glimmstrahlen, as he called them) which proceedfrom the
cathodein straight lines, and which causethe phosphorescence when
they strike the walls of the tube.
Hittorf's observationwas amplified in 1876 by Eugen Goldstein 2
(1850-1930), who found that distinct shadowswere cast, not only
when the cathode was a single point, but also when it formed an
extendedsurface,provided the shadow-throwing object was placed
closeto it. This clearly showedthat the cathoderays (a term now for
the first time introduced) are not emitted indiscriminately in all
directions, but that each portion of the cathode surfaceemits rays
which are practically confined to a single direction; and Goldstein
found this direction to be normal to the surface. In this respecthis
discoveryestablishedan important distinction between the manner
in which cathode rays are emitted from an electrode and that in
which light is emitted from an incandescentsurface.
The questionasto the nature of the cathoderays attracted much
attention during the next two decades. In the year following
Hittorf's investigation,Cromwell Varley 3 put forward the hypothesis
that the rays are composedof' attenuated particles of matter, pro-
jected from the negativepole by electricity' ; and that it is in virtue
of their negative chargesthat these particles are influenced by a
magneticfield.4.
During someyearsfollowing this, the propertiesof highly rarefied
1 Ann. d. Phys. cxxxvi (1869),pp. 1, 197; translated, Annales de Chimie, xvi (1869),
p.487
• Berlin Monatsberichte (1876),p. 279 • Proc. R. S. xix (1871),p. 236
• Priestleyin 1766had shownthat a current of electrifiedair flowsfrom the pointsof
bodieswhich are electrified either vitreously or resinously; d. Priestley's History of
Electricity, p. 591.
351
gases were investigated by Sir William Crookes (1832-1919).

Influenced, doubtless, by the ideas which were developedin con-
nection with his discovery of the radiometer, Crookes,llike Varley,
proposed to regard the cathode rays as a molecular torrent: he
supposedthe moleculesof the residual gas,coming into contact with
the cathode, to acquire from it a resinouscharge, and immediately
to fly off normally to the surface,by reasonof the mutual repulsion
exerted by similarly electrified bodies. Carrying the exhaustionto
a higher degree, Crookeswas enabled to study a dark spacewhich
under such circumstances appears between the cathode and the
cathode glow; and to show that at the highest rarefactions this
dark space(which hassincebeengenerally known by his name, and
which must not be confused with the dark space discovered by
Faraday) enlargesuntil the whole tube is occupied by it. He sug-
gestedthat the thicknessof the dark spacemay be a measureof the
mean length of free path of the molecules. 'The extra velocity,' he
wrote, , with which the moleculesrebound from the excited negative
pole keeps back the more slowly moving molecules which are
advancing towards that pole. The conflict occurs at the boundary
of the dark space,where the luminous margin bears witnessto the
energy of the collisions.' 2 Thus according to Crookes the dark
spaceis dark and the glow bright becausethere are collisionsin the
latter and not in the former. The fluorescenceor phosphorescence
on the walls of the tube he attributed to the impact of the particles
on the glass.
Crookes spoke of the cathode rays as an 'ultra-gaseous' or
'fourth state' of matter. These expressionshave led some later
writer:; to ascribeto him the enunciation or prediction of a hypothesis
regarding the nature of the particles projected from the cathode,
which arose some years afterwards, and which we shall presently
describe; but it is clear from Crookes'smemoirs that he conceived
the particles of the cathode rays to be ordinary gaseousmolecules,
carrying electric charges; and by , a new stateof matter' he under-
stood simply a state in which the free path is so long that collisions
may be disregarded.
Crookesfound that two adjacentpencilsof cathoderays appeared
to repel each other. At the time this was regarded as a direct con-
firmation of the hypothesisthat the rays are streamsof electrically
charged particles; but it was shown later that the deflection of the
rays must be assignedto causesother than mutual repulsion.
1 Phil. Trans. elxx (1879), pp. 135,641 ; Phil. Mag. vii (1879), p. 57
• Phil. Mag. vii (1879), p. 57
352
How admirably the molecular-torrent theory accounts for the

deviation of the cathoderays by a magnetic field was shown by the
calculationsof Eduard Riecke (1845-1915) in 1881.l If the axis of
z be taken parallel to the magneticforce H, the equationsof motion
of a particle of massm, chargee, and velocity (u, v, w) are
emdu/dt = evH, emdv/dt = - euH, emdw/dt = O.
The last equation shows that the component of velocity of the
particle parallel to the magnetic force is constant; the other
equationsgive
u = A sin (eHt/em), v = A cos (eHt/em),
showingthat the projection of the path on a plane at right angles
to the magnetic force is a circle. Thus, in a magnetic field the
particles of the molecular torrent describespiral paths whose axes
are the lines of magnetic force.
But the hypothesis of Varley and Crookes was before long
involved in difficulties. Tait 2 in 1880 remarked that if the particles
are movingwith greatvelocities,theperiodsof the luminousvibrations
receivedfrom them should be affected to a measurableextent in
accordancewith Doppler'sprinciple (for which c£ p. 368). Tait tried
to obtain this effect, but without success. It could, however, be
argued that if, asCrookessupposed,the particles becomeluminous
only when they have collided with other particles,and have thereby
lost part of their velocity, the phenomenonin questionis not to be
expected.
The alternative to the molecular-torrent theory was to suppose
that the cathoderadiation is a disturbanceof the aether. This view
was maintained by severalphysicists,3and notably by Hertz,4.who
rejectedVarley's hypothesiswhen he found experimentally that the
rays did not appear to produce any external electric or magnetic
force, and were apparently not affectedby an electrostaticfield. It
was, however, pointed out by FitzGerald 6 (in a review of Hertz's
Miscellaneous Papers) that external spaceis probably screenedfrom
the effectsof the rays by other electric actions which take place in
the dischargetube.
FitzGerald has left on record that at the British Association
meeting in 1896,Lenard and Bjerknesargued in favour of the idea
1 GOtt. Nach., 2 Feb. 1881; reprinted, Ann. d. Phys. xiii (1881),p. 191
2 Proc. R. S. Edin. x (1880),p. 430
3 e.g.E. Wiedemann,Ann. d. Phys. x (1880),p. 202; translated,Phil. Mag. x (1880),
p. 357. E. Goldstein,Ann. d. Phys. xii (1881),p. 249 • Ann. d. Phys. xix (1883),p. 782
• Nature, Iv (5 Nov. 1896),p. 6; FitzGerald'sScientific Writings, p. 433
353
that cathode rays are an aether propagation of some kind inde-

pendent of matter, while all the English physicistssupported the
projectile hypothesis. The great difficulty in the former theory was
that the rays are deflectedby a magnet, and in the latter that they
are capable of passingthrough films of metal which are so thick
as to be quite opaque to ordinary light; 1 it seemedinconceivable
that particles of matter should not be stoppedby even the thinnest
gold leaf. At the time of Hertz's experimentson the subject,an
attempt to obviate this difficulty was made by J. J. Thomson,2who
suggestedthat the metallic film when bombardedby the rays might
itself acquire the property of emitting chargedparticles,so that the
rays which were observedon the further side need not have passed
through the film. It was Thomson who ultimately found the true
explanation; but this dependedin part on another order of ideas,
whoseintroduction and developmentmust now be traced.
The tendency,which was now general,to abandonthe electron
theory of Weber in favour of Maxwell's theory involved certain
changesin the conceptionsof electriccharge. In the theoryof Weber,
electric phenomenawere attributed to the agencyof stationary or
moving charges,which could most readily be pictured as having
a discreteand atom-like existence. The conceptionof displacement,
on the other hand, which lay at the root of the Maxwellian theory,
wasmore in harmony with the representationof electricity assome-
thing of a continuous nature; and as Maxwell's views m·~twith
increasing acceptance, the atomistic hypothesis seemedto have
entered on a period of decay. Its revival was due largely to the
advocacyof Helmholtz,3 who, in a lecture deliveredto the Chemical
Society of London in 1881,pointed out' that it was thoroughly in
accord with the ideasof Faraday,S on which Maxwell's theory was
founded. 'If,' he said,, we acceptthe hypothesisthat the elementary
substancesare composedof atoms,we cannot avoid concludingthat
electricity also, positive as well as negative,is divided into definite
elementaryportions which behavelike atomsof electricity.'
When the conduction of electricity is consideredin the light of
this hypothesis, it seemsalmost inevitable to conclude that the
1 The penetrating power of the rays had been noticed by Hittorf, and byE. Wiedemann
and H. Ebert (Sitzber. d. phys.-med.Soc.zu Erlangen, 11 Dec. 1891). It was investigated more
thoroughly by Hertz (Ann. d. Phys. xlv [1892], p. 28) and by his pupil Philipp Lenard of
Bonn (Ann. d. Phys. Ii [1894], p. 225; Iii [1894], p. 23) who conducted a series of experi-
ments on cathode rays which had passedout of the discharge tube through a thin window
of aluminium.
2 J. J. Thomson, RecentResearches,p. 126
2 cf. also G. Johnstone Stoney, Phil. Mag. xi (May 1881), p. 381

• Jour. Chem. Soc. xxxix (1881), p. 277 • cf. p. 180
354
processis of much the samecharacter in gasesas in electrolytes;

and beforelong this view was actively maintained. It had indeed
long been known that a compound gas might be decomposedby
the electric discharge; and that in somecasesthe constituentsare
liberated at the electrodesin such a way as to suggestan analogy
with electrolysis. The questionhad beenstudiedin 1861by Adolphe
Perrot, who examined1 the gasesliberated by the passageof the
electric spark through steam. He found that while the product of
this action wasa detonating mixture of hydrogen and oxygen,there
wasa decidedpreponderanceof hydrogenat one pole and of oxygen
at the other.
The analogy of gaseousconduction to electrolysiswas applied
by W. Giese,2of Berlin, in 1882, in order to explain the conductivity
of the hot gasesof flames,which had been discoveredin the course
of the experimentsof the Academiadel Cimento of Florencein 1667.
, It is assumed,'he wrote, 'that in electrolytes, even before the
application of an external electromotive force, there are present
atomsor atomicgroups-the ions,asthey arecalled-which originate
when the moleculesdissociate; by thesethe passageof electricity
through the liquid is effected, for they are set in motion by the
electric field and carry their chargeswith them. We shall now
extend this hypothesisby assumingthat in gasesalso the property
of conductivity is due to the presenceof ions. Such ions may be
supposedto exist in small numbers in all gasesat the ordinary
temperature and pressure; and as the temperature rises their
numberswill increase.'
Ideas similar to this were presentedin a general theory of the
dischargein rarefied gases,which was devisedtwo years later by
Arthur Schuster(1851-1934), of Manchester.3 Schusterremarked
that when hot liquids are maintained at a high potential, the
vapours which rise from them are found to be entirely free from
electrification; from which he inferred that a moleculestriking an
electrifiedsurfacein its rapid motion cannot carry away any part of
the charge,and that one molecule cannot communicateelectricity
to another in an encounterin which both moleculesremain intact.
Thus he wasled to the conclusionthat dissociation,or ionisation, as
it cameto be called later, of the gaseousmoleculesis necessaryfor
the passageof electricity through gases."
1 Annalesde Chimie (3), lxi, p. 161 • Ann. d. Phys. xvii (1882),pp. 1, 236, 519
• Proc. R. S. xxxvii (1884), p. 317
• In the caseof an elementarygas,this would imply dissociationof the moleculeinto
two atomschemicallyalike, but oppositelycharged; in electrolysisthe dissociationis into
two chemicallyunlike iODlJ.
(828)
355
Schusteradvocatedthe charged-particletheory of cathoderays,

and by extending and interpreting an experiment of Hittorf's was
able to adducestrong evidencein its favour. He placedthe positive
and negative electrodesso close to each other that at very low
pressuresthe Crookes dark space extended from the cathode to
beyond the anode. In these circumstancesit was found that the
dischargefrom the positive electrode always passedto the nearest
point of the inner boundary of the Crookesdark space-which, of
course,was in the oppositedirection to the cathode. Thus, in the
neighbourhood of the positive discharge,the current was flowing
in two oppositedirections at closelyadjoining places; which could
scarcelyhappen unlessthe current in one direction were carried by
particles moving againstthe lines of force by virtue of their inertia.
Continuing his researches,Schuster1 showed in 1887 that a
steady electric current may be obtained in air betweenelectrodes
whose difference of potential is but small, provided that an inde-
pendent current is maintained in the samevessel; that is to say, a
continuous discharge produces in the air such a condition that
conduction occurs with the smallest electromotive forces. This
effect he explained by aid of the hypothesispreviously advanced;
the ionsproducedby the main dischargebecomediffusedthroughout
the vessel,and, coming under the influenceof the field setup by the
auxiliary electrodes,drift so as to carry a current between the
latter.
A discoveryrelated to this wasmadein the sameyear by Hertz,2
in the courseof the celebratedresearches3 which have beenalready
mentioned. Happening to notice that the passageof one spark is
facilitated by the passageof another spark in its neighbourhood,he
followed up the observation,and found the phenomenonto be due
to the agency of ultra-violet light emitted by the latter spark. It
appearedin fact that the distanceacrosswhich an electric sparkcan
passin air is greatly increasedwhen light of very short wave-length
is allowed to fall on the spark gap. It was soon found 4, that the
effective light is that which falls on the negative electrodeof the
gap; and Wilhelm Hallwachs 6 extendedthe discoveryby showing
that when a sheetof metal is negatively electrified and exposedto
1 Prot:. R. S.xlii (1887),p. 371. Hittorfhad discoveredthat very small electromotive
forcesaresufficientto causea dischargeacrossa spacethroughwhich thecathoderadiation

is passing.
• Berlin Ber. (1887), p. 487; Ann. d. Plzys.xxxi (1887), p. 983; Electric Waves
(Englishedn.), p. 63
• ef. p. 322
• By E. Wiedemannand H. Ebert, Ann. d. Phys.xxxiii (1888),p. 241
• Ann. d. Phys.xxxiii (1888),p. 301
ultra-violet light, the adjacent air is thrown into a state which

permits the chargeto leak rapidly away.l
E. Rutherford showed2 that when ultra-violet light falls on a
metal plate, the ions produced in the air around are all negative;
and he devisedan ingeniousarrangementfor finding their velocities.
The great advancesin the theory of this photo-electric effect, asit came
to be called, washowevermade only in the twentieth century.
Interest was now thoroughly aroused in the problem of con-
ductivity in gases; and it was generally felt that the best hope of
divining the nature of the processlay in studying the dischargeat
high rarefactions. 'If a first step towards understanding the rela-
tions between aether and ponderable matter is to be made,' said
Lord Kelvin in 1893,3' it seemsto me that the mosthopeful founda-
tion for it is knowledgederived from experimentson electricity in
high vacuum.'
Within the two following yearsconsiderableprogresswaseffected
in this direction. J. J. Thomson,4 by a rotating-mirror method,
succeededin measuringthe velocity of the cathode rays, finding it
to be 5 1·9 X 107 cm./sec.; a value so much smaller than that of
the velocity of light that it was scarcelypossibleto conceiveof the
rays as vibrations of the aether. A further blow was dealt at the
latter hypothesiswhen Jean Perrin,6 having received the rays in a
metallic cylinder, found that the cylinder became charged with
resinouselectricity. Thomson showed that when the rays were
deviated by a magnet in such a way that they could no longer
enter the cylinder, it no longer acquired a charge. This appearedto
demonstratethat the rays transport negativeelectricity.
With cathoderays is closelyconnectedanother type of radiation,
which was discoveredon 8 November 1895, by Wilhelm Konrad
Rontgen (1845-1923) of Wtirzburg.7 The discoveryseemsto have
originated in an accident: Rontgen was experimenting with a
Crookestube, which was completely enclosedin an opaque shield
of black cardboard. He noticed however that when a current was
passedthrough the tube, a piece of paper painted with barium
1 Elsterand Geitel publisheda number of paperson this phenomenonin 1889-95,

Ann. d. Phys. xxxviii (1889),pp. 40,497; xxxix (1890),p. 332; xli (1890),pp. 161,166;
Iii (1894),p. 433; Iv (1895),p. 684
• Proc. Camb. Phil. Soc. ix (1898),.p.401 • Proc. R. S. liv (1893),p. 389
• Phil. Mag. xxxviii (1894),p. 358
• The valuefoundby the sameinvestigatorin 1897wasmuch larger than this.
• Compus Rendus, cxxi (1895),p. 1130. Hertz had previouslytried an experimentof
the samekind, but without success.
, Si/;:.ungsber.der Wur;:.burger Physikal.-Medic. Gesellsehaft, 28 Dec. 1895; reprinted,
Ann. d. Phys. !xiv (1898),pp. 1, 12; translated,Noturl, !iii (23Jan. 1896),p. 274
357
platino-cyanide, which happenedto be lying on the bench, showed

fluorescence. Experimentssuggestedby this proved that radiation,
capableof affectingphotographic platesand of causingfluorescence
in certain substances,is emitted by tubes in which the electric dis-
charge is passing; and that the radiation proceedsfrom the place
where the cathoderays strike the glasswalls of the tube. The X-ray's,
as they were called by their discoverer,were propagatedin straight
lines, and could neither be refracted by any of the substanceswhich
refract light, not deviated from their courseby a magnetic field;
they were moreover able to pass with little absorptions through
many substanceswhich are opaque to ordinary and ultra-violet
light-a property of which considerableuse was at once made in
surgery; Schusterhasrecordedthat for sometime after the discovery
his own laboratory at Manchesterwas crowded with medical men
bringing patients who were believedto have needlesin variousparts
of their bodies.1
The nature of the new radiation wasthe subjectof much specula-
tion. Its discoverersuggestedthat it might prove to representthe
long-sought-for longitudinal vibrations of the aether; while other
writers advocated the rival claims of aethereal vortices, infra-red
light and' sifted' cathoderays. The hypothesiswhich subsequently
obtained general acceptancewas first propounded by Schuster2 in
the month following the publication of Rontgen'sresearches.It is,.
that the X-rays are transversevibrations of the aether,of exceedingly
small wave-length; ultra-ultra-violet light, in fact. A suggestion
which was put forward later in the year by E. Wiechert 3 and Sir
George Stokes,4 to the effect that the rays are pulsesgeneratedin
the aether when the glassof the dischargetube is bombardedby the
cathode particles, is not really distinct from Schuster'shypothesis;
for ordinary white light likewise consistsof pulses,as Gouy 6 had
shown; and the essentialfeature which distinguishesthe Rontgen
pulsesis that the harmonic vibrations into which they can be resolved
by Fourier's analysisare of very short period.
1 The discovery of X-rays was narrowly missed by several physicists; in the caseof
Hertz and Lenard, becauseof failure to distinguish them from cathode rays. An Oxford
physicist, Frederick Smith, having found that photographic plates kept in a box near a
Crookes tube were liable to be fogged, told his assistant to keep them in another
place!
• Nature, liii (23 Jan. 1896), p. 268. FitzGerald independently made the samesugges-
tion in a letter to O. J. Lodge, printed in the Electrician, xxxvii, p. 372.
• Ann. d. Phys. lix (1896), p. 321
• Nature,liv (3 Sept. 1896), p. 427; Proc. Camb. Phil. Soc. ix (1896), p. 215; Mem.
Manchester Lit. & Phil. Soc. xli (1896-7), No. 15
• Jour. de Phys. v (1886), p. 354
The rapidity of the vibrations explainsthe failure of all attempts

to refract the X-rays. For in the formula
(Jp2
/L2 = 1 + P (p2 _ n2)
of the Maxwell-Sellmeiertheory,! n denotesthe frequency,and sois
in this caseextremelylarge; whencewe havevery nearly
/L2 = 1,
i.e. the refractiveindex of all substancesfor the X-rays is unity. In
fact, the vibrations alternate too rapidly to have an effect on the
sluggishsystemswhich are concernedin refraction.
One of the most important properties of X-rays was dis-
covered,shortly after the rays themselveshad becomeknown, by
J.J. Thomson,2who announcedthat when they passthrough a gas
(or, ashe thought at that time, evena liquid or a solid) they render
it conducting. This he attributed, in accordancewith the ionic
theory of conduction, to ' a kind of electrolysis,the moleculebeing
split up, or nearly split up, by the Rontgenrays'; and this led him
to speakof the gasasbeing ionised.
The conductivity produced in gasesby this meanswas at once
investigated3 more closely. It was found that a gas which had
acquiredconducting power by exposureto X-rays lost this quality
when forced through a plug of glass-wool; whenceit was inferred
that the structurein virtue of which the gasconductsis of so coarse
a characterthat it is unable to survive the passagethrough the fine
poresof the plug. The conductivity wasalsofound to be destroyed
whenan electriccurrent waspassedthrough the gas-a phenomenon
for which a parallel may be found in electrolysis. For if the ions
were removed from an electrolytic solution by the passageof a
current, the solution would ceaseto conduct as soon as sufficient
electricity had passedto removethem all; and it may be supposed
that the conductingagentswhich are producedin a gasby exposure
to X-rays are likewise abstractedfrom it when they are employed
to transportcharges.
The sameidea may be applied to explain another property of
gasesexposedto X-rays. The strength of the current through the
gas dependsboth on the intensity of the radiation and also on
the electromotiveforce; but if the former factor be constant,and the
electromotiveforce be increased, the current does not increase
1 cr. p. 264 I Nature liii (27 Feb. 1896), p. 391
• J.J.Thomson and E. Rutherford, Phil. Mag. xlii (1896), p. 392
359
indefinitely, but tends to attain a certain' saturation' value. The

existenceof this saturation value is evidently due to the inability of
the electromotiveforce to do more than to remove the ions as fast
asthey are produced by the rays.
Meanwhile other evidencewas accumulating to show that the
conductivity produced in gasesby X-rays is of the samenature as
the conductivity of the gasesfrom flames and from the path of a
discharge,to which the theory of Gieseand Schusterhad already
been applied. One proof of this identity was supplied by observa-
tions of the condensationof water vapour into clouds. It had been
noticed long before by John Aitken 1 that gasesrising from flames
causeprecipitation of the aqueousvapour from a saturated gas;
and Robert von Helmholtz 2 had found that gasesthrough which an
electric dischargehas been passedpossessthe same property. It
was now shown by C. T. R. Wilson (b. 1869),S working in the
CavendishLaboratory at Cambridge, that the sameis true of gases
which have been exposedto X-rays. The explanation furnished by
the ionic theory is that in all three casesthe gascontainsionswhich
act as centresof condensationfor the vapour.
During the year which followed their discovery,the X-rays were
so thoroughly examined that at the end of that period they were
almost better understood than the cathode rays from which they
derived their origin. But the obscurity in which this subject had
beensolong involved was now to be dispelled.
Lecturing at the Royal Institution on 30 April 1897,J. J. Thom-
son 4 advanceda new suggestionto reconcile the molecular-torrent
hypothesiswith Lenard's observationsof the passageof cathoderays
through material bodies. 'We seefrom Lenard's table,' he said,
, that a cathode ray can travel through air at atmosphericpressure
a distance of about half a centimetre before the brightnessof the
phosphorescence falls to about half its original value. Now the mean
free path of the moleculeof air at this pressureis about 10-5 em., and
if a moleculeof air were projected it would losehalf its momentum
in a spacecomparablewith the meanfree path. Even if we suppose
that it is not the same molecule that is carried, the effect of the
obliquity of the collisionswould reducethe momentum to half in a
short multiple of that path.
, Thus, from Lenard's experimentson the absorption of the rays
outside the tube, it follows on the hypothesisthat the cathode rays
1 Trans. R. S. Edin. xxx (1880), p. 337 I Ann. d. Phys.xxxii (1887), p. 1
• Proc.R. S. lix (19 Mar. 1896), p. 338; Phil. Trans. c1xxxix (1897), p. 265
• RoyalInst. Proc.xv (1897), p. 419
360
arechargedparticlesmoving with high velocitiesthat the sizeof the

carriers must be small compared with the dimensionsof ordinary
atomsor molecules.l We seeon this hypothesiswhy the magnetic
deflectionis the sameinside the tube whatever be the nature of
the gas,for the carriers are the samewhatever gas be used. The
assumptionof a state of matter more finely subdivided than the
atom of an elementis a somewhatstartling one; but a hypothesis
that would involve somewhatsimilar consequences-viz. that the
so-calledelementsare compoundsof someprimordial element-has
beenput forward from time to time by various chemists.'
In order to explain this last remark, it must be recalled that
from the very beginningsof physics,amongstthe Milesian philoso-
phers in the sixth century B.C., the doctrine had been maintained
that all things are but different forms of one single primary and
universal substance. Thales conjectured that it might be water,
Anaximenesthat it might be air, and Anaximander that it was a
boundlessindefinite something--'Toa7T€£pOv--out of which every-
thing is generatedand to which everythingreturns. Modern science
hasentertainedthe sameidea, in many different forms: aswe have
seen,Lord Kelvin suggestedthat all atoms might be vortex rings in
an infinite liquid, which would be precisely Anaximander's TO
o.7Tf.'POV; but when J. J. Thomson discoveredthat the cathode-ray
particles were the same,whatever atoms they were torn out of, his
mind recurred chiefly to the hypothesisput forward in 1815-16 by
William Prout (1785-1850), namely, that the atomic weights of the
elements are whole-number multiples of the atomic weight of
hydrogen, and that, in fact, hydrogen is the universal primordial
substance. 2 Thomson now conjecturedthat Prout had been wrong
in selectinghydrogen, and that the fundamental place in nature

belongedrightly to the cathode-rayparticles, or ' corpuscles' as he
called them, which werevery much smaller than the hydrogenatom
and were now proved to be c0ntained in the atoms of all elements.
1 In 1896 Lorentz, in commenting on Zeeman's discovery (seep. 411), had suggested
that the electric atoms concerned in it had massesabout n?o-n- of the hydrogen atom.
Thomson himself said that the idea of the particles in cathode rays being much lessmassive
than atoms first arose in his mind when he noticed how great was the deflection of the
rays by a magnet. A similar suggestion was made, and a value of the ratio of charge to
mass obtained, by E. Wiechert, Sehriften d. physik.-iicon. Gesellseh.<:u Konigsberg xxxviii
(Jan. 1897.), p. [3J
I Prout's paperswere published in Annals of Philosophy,vi (1815), p. 321, and vii (1816),
p. 111. The author's name was not attached to the papers, but was revealed a few
months later by the editor (Annals of Philosophy,vii [1816J, p. 343.) Prout's own words are,
• If the views we have ventured to advance be correct, we may almost consider the
."pcf,'TT/ v>'7] of the ancients to be realised in hydrogen.' On the subsequent fortunes of
Prout's hypothesis, cf. J. Kendall, Proc. R. S. Edin. !xiii (1950), p. 1.
361
This idea seemsto have fascinated him at the time, as well it

might.
Later researches-most of them carried out or inspired by
Thomson himself-showed the necessityof modifying this opinion
as to the place of the cathode-ray corpusclesin the schemeof the
universe. Sincethey carry chargesof negativeelectricity, while the
atom as a whole is neutral in respectof charge,it follows that some
part of the atom must be positively charged. At this time Thomson
seemsto have inclined to the idea that positive electricity was,so to
speak,disembodied,existing asa kind of cloud within the atom, and
not attached to any material corpuscles,so that the massof the
atom was derived entirely from the negative corpusclesof which it
wascomposed. As we shall see,this conceptionwaslater discarded.
Thomson'sRoyal Institution lecture drew from FitzGerald 1 the
suggestionthat ' we are dealing with free electronsin thesecathode
rays '-a remark the point of which will becomemore evidentwhen
we come to considerthe direction in which the Maxwellian theory
wasbeing developedat this time.
Shortly afterwards Thomson himself published an account2 of
experimentsin which the only outstandingobjectionsto the charged-
particle theory were removed. The chief of these was Hertz's
failure to deflect the cathode rays by an electrostaticfield. Hertz
had causedthe rays to travel betweenparallel platesof metal main-
tained at different potentials; but Thomson now showedthat in
thesecircumstancesthe rays generateions in the rarefied gas,which
settle on the plates, and annul the electric force in the intervening
space. By carrying the exhaustion to a much higher degree,he
removedthis sourceof confusion,and obtainedthe expecteddeflection
of the rays.
The electrostaticand magnetic deflectionstaken togethersuffice
to determinethe ratio of the massof a cathodeparticle to the charge
which it carries. For the equation of motion of the particle is
dlr
m dtl = eE + ce [v . H],
where r denotesthe vector from the origin to the position of the

particle; E and H denotethe electric and magneticforcesmeasured
in e.-s. and e.-m. units respectively; e the charge in e.-s.units, m
the mass, and v the velocity of the particle. By observing the
circumstancesin which the force eE,due to the electric field, exactly
balancesthe force e[ v . H]f c, due to the magneticfield, it is possible
1 Electrician, xxxix (21 May 1897), p. 103 I Phil. Mag. xliv (1897), p. 298
362
to determinev; v wasfound to be 15,000 kilometresper second;

and it is readily seenfrom the aboveequation that a measurement
of the deflectionin the magneticfield suppliesa relation betweenv
and mle; so both v and mle may be determined. Thomson found
the value of mle to be independentof the nature of the rarefied gas;
its amount was 10-7 (grammes/electromagneticunits of charge)
which is only about the thousandthpart of the value of mle for the
hydrogenatom in electrolysis.1 If the chargewere supposedto be
of the sameorder of magnitude as that on an electrolytic ion, it
would be necessaryto conclude that the particle whosemasswas
thus measuredis much smallerthan the atom, and aswe have seen,
the conjecture was entertained that it is the primordial unit or
corpuscleof which all atomsare ultimately composed.2
The nature of the resinouslychargedcorpuscleswhich constitute
cathode rays being thus far determined, it became of interest to
inquire whether correspondingbodies existed carrying chargesof
vitreous electricity-a question to which at any rate a provisional
answerwasgiven by W. Wien (1864-1928),3 of Aachen,in the same
year. More than a decadepreviouslyE. Goldstein4. had shownthat
when the cathodeof a dischargetube is perforated, radiation of a
certain type passesoutward through the perforationsinto the part
of the tube behind the cathode. To this radiation he had given the
namecanal rays. Wien now showedthat the canal rays are formed
of positively chargedparticles, obtaining a value of mle immensely
larger than Thomsonhad obtainedfor the cathoderays, and indeed
of the sameorder of magnitudeas the correspondingratio in elec-
trolysis. It wasfound moreoverthat whereasthe negativelycharged
particles are the same, from whatever source they are derived,
the positivelychargedparticles are different from eachother.
The disparity thus revealed betweenthe corpusclesof cathode
rays and the positiveions of Goldstein'srays excited great interest;
it seemedto offer a prospectof explaining the curious differences
between the relations of vitreous and of resinous electricity to
ponderablematter. Thesephenomenahad beenstudied by many
previous investigators; in particular Schuster,5in the Bakerian
lecture of 1890, had remarkedthat' if the law of impact is different
I A later and more accuratevalue for mle is 0·568 X 10-7 (grammes/electromagnetic
units of charge),which is about nlrr the correspondingratio for the hydrogenatom.

S The value of mle for cathode rays was determined also in the same year by
W. Kaufmann, Ann. d. Phys. lxi, p. 544.

I Verhandl. tier physik. Gesells. zu Berlin, xvi (1897),p. 165; Ann. d. Phys. !xv (1898),
p.440. SeealsoP. Ewers,Ann. d. Phys. !xix (1899),p. 167

• Berlin Sitvmgsber. (1886),p. 691 • Proc. R. S. xlvii (1890),p. 526
363
betweenthe moleculesof the gasand the positive and negative ions

respectively,it follows that the rate of diffusion of the two setsof
ions will in general be different,' and had inferred from his theory
of the dischargethat' the negativeions diffuse more rapidly.' This
inference was confirmed in 1898 by John Zeleny,! who showedthat
of the ions produced in air by exposureto X-rays, the positive are
decidedly lessmobile than the negative.
The magnitude of the electric charge on the ions of gaseswas
not known with any degreeof certainty until 1897-8, when a plan
for determining it was successfullyexecutedby J. J. Thomson and
his pupil J. S. E. Townsend (b. 1868).2 The principles on which
this celebrated investigation was based are very ingenious. By
measuring the current in a gas which is exposedto Rontgen rays
and subjected to a known electromotive force, it is possible to
determine the value of the product nev, where n denotesthe number
of ions in unit volume of the gas,e the charge on an ion, and v the
mean velocity of the positive and negative ions under the electro-
motive force. As v had been already determined,3 the experiment
led to a determination of ne; so if n could be found, the value of e
might be deduced.
The method employed by Thomson to determine n was founded
on the discovery, to which we have already referred, that when
X-rays passthrough dust-free air, saturated with aqueousvapour,
the ions act as nuclei around which the water condenses,so that a
cloud is produced by sucha degreeof saturation aswould ordinarily
be incapable of producing condensation. The sizeof the drops was
calculated from measurementsof the rate at which the cloud sank;
and, by comparing this estimatewith the measurementof the mass
of water deposited,the number of drops was determined, and hence
the number n of ions. The value of e consequentlydeduced was
found to be independentof the nature of the gasin which the ions
were produced, being approximately the samein hydrogenasin air,
and being in both casesthe sameas for the charge carried by the
hydrogen ion in electrolysis.4
Even in the latter part of 1899 most physicists did not fully
believe in Thomson's ideas. At the British Associationmeeting at
Dover in Septemberof that year, he read a paper On the existenceof
massessmaller than the atoms 5 to a deeply interestedbut not altogether
1 Townsend, Proc.Camb.Phil. Soc.ix (Feb. 1897), p. 244; Phil. Mag. xlv (Feb. 1898),
p. 125; J. J. Thomson, Phil. Mag. xlvi (1898), p. 120
• Phil. Mag. xlvi (1898), p. 528 • By E. Rutherford, Phil. Mag. xliv (1897), p. 422
• Townsend, Phil. Trans. cxciii (May 1899), p. 129
I Printed under a different title in Phil. Mag. (5), xlviii (1899), p. 547
364
convincedaudience. He first referred to his 1897determination of

the ratio mfe for the corpuscleswhich constitute cathode rays; this
he now supplementedby measurementsof mfe and e for the negative
electrification discharged by ultra-violet light,l and also of mfe for
the negative electrification produced by an incandescent carbon
filamentin an atmosphereofhydrogen.2 It wasfound that the value
of mfe in the caseof the ultra-violet light,3 and also in that of the
carbonfilament, is the sameasfor the cathoderays; and that in the
caseof the ultra-violet light, e is the samein magnitude asthe charge
carried by the hydrogen atom in the electrolysis of solutions. He
concludedthat negative electrification, though it may be produced
by very different means,is made up of units each having an electric
charge of about 6 X 10-10 electrostatic units, equal to the positive
chargecarried by the hydrogen atom in electrolysis. In gasesat low
pressure,theseunits of negative electricity are alwaysassociatedwith
carriers or corpusclesof a definite mass,which is exceedinglysmall,
being only about 1·4 X 10-3of that of the hydrogen ion. A certain
number of thesecorpusclesare normally presentin every atom, and
the ionisation of a gas consistsin the detachment of one corpuscle
from the gas atom.
Since the publication of Thomson's papers, these general con-
clusionshave been abundantly confirmed. It is now certain that
electric charge existsin discrete units, vitreous and resinous,each
of magnitude4·80 X 10-10 electrostaticunits or 1·6 X 10-19 coulombs.
For the corpusclesof negative charge, the name electrons, which had
beencoinedsomeyearsearlier by G.Johnstone Stoney,4soonbecame
universal. Each ion, whether in an electrolytic liquid or in a gas,
carriesone (or an integral number) of thesecharges. An electrolytic
ion or a positive gaseousion also contains one or more atoms of
matter; but amongthe negativeions in a gas,someare not attached
to atoms,but are, in fact, free electrons.
That electric currents cannot be continuous, but must consistof
discreteelectric chargesseparatedby free aether, had been insisted
on in earlier papers by Larmor.1i He inferred this from the experi-
ments of Rowland 1 (the creation of a magnetic field by a rotating
electrified conductor) and of Helmholtz 2 (the creation of an electric
1 cf. p. 356 • cf. p. 426
• This wasannouncedin a l~tter to Rutherford of date 23July 1899. Valuesof mle
and v for the electronsemitted by a conductorunder the influenceof ultra-violet light
werealsofound by P. Lenard, Wien Sit;:;ungsb. cviii, abth. 2a (Oct. 1899),p. 1649; and
Ann. d. P~ys.ii (1900),p. 359.
• Phil. Mag. I) xxxviii (1894),p. 418
I Phil. Trans. (A), clxxxv (1894),p. 719; clxxxvi (1895),p. 695. SeealsoAether and
Matter, p. 337,and Phil. Mag. vii (1904),p. 621

365
potential-difference by the rotation of a conductor in a magnetic

field) and the phenomenon of unipolar induction (in which an
electromotive force is induced when a magnet revolves round its
axis of symmetrythrough its own field offorce). For example,in the
last-namedexample,he assertedthat the electronsin the magnet,as
they are moved acrossthe magnetic field of the aether, are subject
to forces proportional to the component magnetic intensity in the
meridian plane, which produce electrical separationby drifting the
positive ions towards the axis and in the direction of the length of
the magnet one way, and the negativeions the oppositeway.
The discovery of the electron, and the innumerable researches
associatedwith it, establishedThomson's reputation as the first of
living experimentalphysicists; and it wasin this aspectthat he was
generally regarded during the latter half of his life. But his con-
tributions to theoretical or mathematical physicswere scarcelyless
important than the fruits of his researchesin the laboratory. Perhaps
his greatestachievementin pure theory was the discovery(to which
we have already referred 3) of electromagneticmomentum: namely,
that in an electromagneticfield there is storedin every unit volume
an amount of mechanical momentum, proportional to the vector
product of the electric and magnetic vectors. This principle is
necessaryfor the construction of a quantity which will be described
later under the name of the energy tensor; and without it, neither
relativity theory nor quantum-electrodynamicscould havedeveloped.
During the whole of his tenure of the Cavendishchair, Thomson
was the active head and inspirer of a great researchschool. The
institution of the statusof' advancedstudent' at Cambridgein 1895
led to a great increasein the numbers of young graduatesof other
universities who came to work in his laboratory; with the con-
sequencethat in a few yearsnearly all the important chairsof physics
in the British Empire were filled by his disciples.
1 Seepp. 305-6 • Seep. 305 • ef. p. 317
366
Chapter XII
CLASSICAL RADIATION-THEORY
THE fact that a flame becomesintensely yellow when common salt

is addedto the burning substancehasdoubtlessbeenknown from the
earliesttimes. A somewhatvague suggestionthat the phenomenon
might be amenable to a quantitative treatment, and so brought
within the domain of science,was made in the eighteenth century
by Thomas Melvill,l who remarked that the yellow colour appeared
to beofa definite' degreeof refrangibility.' The problemofradiation
was approachedfrom another direction in 1802 by William Hyde
Wollaston 2 (1766-1828), who found that the spectrum of sunlight
was crossedby sevendark lines, perpendicular to its length. This
discovery,however,passedwithout notice until it was made again in
a much ampler fashion by the Bavarian, Joseph von Fraunhofer 3
(1787-1826), who presentedto the Munich Academy in 1814-15 a
map of the solar spectrum showing a multitude of dark lines, the
chief of which he distinguishedby letters of the alphabet. Later,
by use of a diffraction grating, he measured" the wave-lengths
correspondingto the strongestlines: the wave-lengthof the D line
for instancewasfound to be (in modern units 6) 5887·7 A 6: and the
wave-lengthof the yellow light of many flames (which was not yet
definitely known to be due to the presenceof sodium) was found to
be equal to this.
In 1826 W. H. Fox Talbot 7 (1800-77) examined with a prism
the light from flamesplaced beforea slit and showedthat, ashe said,
, a glanceat the prismatic spectrumof a flame may showit to cont:>.in
substances,which it would otherwise require a laborious chemical
1 Physical and Literary Essays, ii (Edinburgh, 1752)
I Phil. Trans. xcii (1802), p. 365. Wollaston's lines A, B,f, g, D, E seem to correspond
respectively to Fraunhofer's B, D, b, F, G, H.
I Ann. d. Phys. Ivi (1817), p. 264
• Denlcschr.Akad. d. Wiss. .(;U Miinchen, viii (1821-2), p. 1

I The unit of wave-length now generally used is the Angstrom or tenth-metre, named
after the Swedish physicist, Anders Jons Angstrom (1814-74) : it is 10-10 metres, or one
ten-millionth of a millimetre. The visible spectrum is from about 8,000 to 3,900
Angstroms; the ultra-violet rays which can be studied in air extend to about 1,850 A.
Other units occasionally used are the 1-'1-', which is 10 A., and the I-' or micron, which is
1,000/-11-',or nnHJ millimetre.
• The D line is really double, its components having wave-lengths 5,896·2 A. and
5,890·2 A.
, Edin. Jour. Sci. v (1826), p. 77
analysisto detect.' In particular, the red given to a flame by a salt

of strontium and the red due to a salt of lithium may be at once
distinguishedfrom each other in this way. Fox Talbot's paper may
be regarded asbeginning the application of spectroscopyto chemical
analysis.l
In 1833 William Hallows Miller 2 (1801-80), continuing a study
of the absorption of light by coloured substanceswhich had been
begun by Sir John Herschel 3 (1792-1871), examined the spectrum
of sunlight which had been transmitted through various gases,and
found dark lines in it, evidently due to absorption: whence the
opinion became general that the Fraunhofer lines were due to
absorption in the outer layers of the sun.'
The next important contribution to spectroscopy was the
enunciation by Christian Doppler 5 (1803-53) of a principle now
generally known by his name, namely, that if a sourceof light be in
motion relative to an observer, the period of the wavesas received
by the observer will be different from the period of the waves as
emitted by the source. To prove this, let T be the original period
of the wavesemitted by a sourcetravelling towards the observerwith
a relative velocity v, supposedsmall comparedwith the light velocity
V in the intervening medium. Then in one second the sourcede-
scribesa distancev, and emits 1IT waves. The first of the waveswill
passthe terminus of this distance v at a time vlY after the beginning
of the second, and the last of them will pass the terminus at the
end of the second. So 1IT waves pass the terminus in (1 - vlY)
seconds,or the period of the waves as they pass the terminus is
T(l - vIY).
A corresponding alteration of period is produced if light is
reflectedfrom a moving surface. Supposethat a plane wave of light,
in a medium in which its velocity is Cl, is incident at an angle i on a
surface parallel to the plane of xy, which is moving parallel to the
axis of z with velocity v. Then it is easily shown that the increment
of period causedby the reflection is ST, where
ST 2v cos i
T
From 1849onwards much attention was paid to the interdepen-

1 cr. his later work, Phil. Mag. (II iv (1834), p. 112; (I) vii (1835), p. 113; (I) ix
(1836),p. I
I Phil. Mag. II), ii (1833),p. 381
I Trans. R. S. Edin. ix (1823),p. 445
• This is assertedexplicitly in Herschel'sTreatise on Astronomy, 1833

• Abh. Kongl. bohm. Gesellsch.ii (1842),p. 465
368
dencebetweenthe light emitted by a radiating body, and the absorp-

tion of light by the samebody. Leon Foucault,! working with the
electric arc, had found in its light a bright yellow radiation whose
wave-lengthwas exactly that of Fraunhofer's dark D line. He dis-
coveredthat if a brighter radiator giving a continuousspectrumwas
placedbehind the arc, sothat the spectrumof this radiator wasseen
through the arc, then the D line appeareddark, just as in the solar
spectrum. Thus, as he pointed out, the arc emits the D line, but
absorbsit when it is emitted by the other radiator. Foucault did not
however attribute the D line to sodium, or extend his observation
to other lines, or draw any theoretical conclusions.
Sometime prior to the summerof 1852,2at Cambridge, Stokes,
in a conversation with William Thomson (Kelvin), referred to
Fraunhofer's discovery of the coincidence in wave-length between
the double dark line D of the solar spectrum and the double bright
linewhichwasseenin thespectraof many flames,andto averyrigorous
experimental test of this coincidence which had been made by
ProfessorW. A. Miller,3 which showed it to be accurate to an
astounding degreeof minuteness: and to the fact that the strong
yellow light given out when salt is thrown on burning spirit consists
almost solely of light correspondingto this double bright line. In
a letter to Thomson of date 24 Feb. 18544 Stokes mentioned
observationsof his own, that the bright line D was absent from a
candle-flamewhen the wick was snuffed clean, so as not to project
into the luminous envelope,and from an alcohol flame, when the
spirit was burned in a watch glass; which indicated that the sub-
stanceresponsiblefor the D line was contained not in the alcohol,
but in the wick. In 1855 Foucault went to London to receive the
Copley Medal of the Royal Society, and describedhis experiments
with the arc to Stokes, who thereupon communicated them to
Thomson. 'In conversationwith Thomson,' said Stokes,5'I ex-
plained the connectionof the bright and dark lines by the analogyof
a set of piano stringstuned to the samenote, which if struck would
give out that note, and would also be ready to sound out, to take it
up in fact, if it were soundedin air. This would imply absorption of
1 Foucault's results were communicated to the Societe Philomathique on 7 Jan. 1849,
and printed in L'/nstitut of 7 Feb. 1849, but they became known generally only when they
were reprinted in Annates de Chimie et de Physique (3) Iviii (1860), p. 476.
• The date is fixed by the fact that Thomson never was in Cambridge from the
summer of 1852 until 1866; and the conversation was certainly in Cambridge: Stokes's
Math. and P~vs. Papers, iv, p. 374.
• This work does not appear to have been published by Miller, but see Phil. Mag.
xxvii (1845), p 81 • Printed in Stokes's Math. and Phys. Papers, iv, p. 368
• Stokes's Merrwir and Sci. Correspondence,ii (Cambridge, 1907), p. 75
369
the aerial vibrations. I told Thomson,' he added,1 'I believed

there wasvapour of sodium in the sun's atmosphere.'
Thomson seized eagerly on these two ideas-the dynamical
explanation of the dark lines and the application to solar chemistry
-and introduced them into his professoriallecturesin the University
of Glasgow.2 In a letter to Stokesof 2 March 1854,3he suggested
that other vapours than sodium might be found in the atmospheres
of sun and starsby searchingfor substancesproducing,in the spectra
of flames, bright lines coinciding with dark lines of the solar and
stellar spectraother than the D line.4
In 1853 Angstrom presentedto the SwedishAcademya memoir 5
in which he assertedthat an incandescentgas emits luminous rays
of the samerefrangibility as thosewhich it absorbsat the ordinary
temperature-which, though not quite accurate,indicatesthe extent
to which the mutual relations of radiation and absorptionwere now
being studied.
In 1858 the question was approached by a Scottish physicist,
Balfour Stewart (1828-87), from a wholly different standpoint,
which must now be explained.
In the present chapter, we have spoken as if rays of mono-
chromatic light existed. This, however, is only an approximation
to the truth: a ray of light may be nearly monochromatic, but,
strictly speaking,it must always include constituentswhosewave-
lengths differ from each other, however narrow may be the range
of the spectrum within which they are comprised. In this respect,
optics differs from acoustics,sincethere is nothing impossiblein the
notion of a sound of a single definite frequency. When we are
considering the radiation emitted by, say, a red-hot poker, we must
recognisethat it is composedof a continuous aggregateof con-
stituents, whose wave-lengths extend over a wide range of the
spectrum.
At the end of the eighteenth and beginning of the nineteenth
century, Pierre Prevost6 (1751-1839), of Geneva,remarked that a
1 ibid., p. 83
• This must have been soon after Stokes's first conversation with him, for he says,
•I had begun teaching it to my class in the session 1852-3, as an old student's notebook
which I have shows' (Stokes, Math. and Phys. Papers, iv, p. 374).
I Printed in Stokes's Math. and Phys. Papers, iv, p. 369
• cf. also Thomson's presidential address to the British Association in 1871 (B. A.
Rep. 1871, p. Ixxxiv) ; and Rayleigh's obituary of Stokes (Royal SocieryTear Book, 1904 ;
Rayleigh's ScientijU;Papers, v, p. 173)
• Stockh.Akad. Handl. 1852-3, p. 229; Phil. Mag. (4) ix (1855), p. 327; Ann. d. Phys.
xciv (1855), p. 141
I Recherches physico-mlcaniquesSIlT la chaleur,Geneva, 1792 ; Essaisur Ie caloriquerayonnant,
Geneva, 1809
370
red-hot body, by emitting radiation, tends to becomecooler: it is,

however, possible to imagine that a number of such bodies, so
placedthat they intercept the whole of each other's radiations, exist
in a steady state, in which they all remain permanently at some
definite temperature. In such a system,each body radiatesjust as
if the other bodieswere not present; but it receivesradiation from
them, and, as Prevostpointed out, it must receiveexactly as much
heat asit radiates. This becameknown asPrevost'sLaw of Exchanges.
Prevost'sline of thought wasnow followed up by Balfour Stewart,1
who observedthat a plate of rock saIt is much lessdiathermanousfor
rays of radiant heat emitted by a massof the samesubstanceheated
to 100°C., than for rays emitted by any other body at the same
temperature; considerationsbasedon this led him to the conclusion
that the radiating power of every kind of substanceis equal to its absorbing
power, for every kind of ray of radiant heat. The extension to rays of
light was obvious. Stokes, who had not thought of extending
Prevost'slaw of exchangesfrom radiation as a whole to radiation of
each particular refrangibility by itself, had not realised that a soda
flame which emits the bright D line must on that very account absorb
light of the samerefrangibility. 2
A year and a half after Balfour Stewart's paper, Kirchhoff 3
of Heidelberg, by an independent investigation, was led to the
samediscovery,and showedhow it could be applied to the chemistry
of the sun and stars.4 His treatment was more complete than that
of any of his predecessors, and made a much greater impressionat
the time on the scientific world.
We can define the coefficientof absorption A of a body for radiation
of a given wave-lengthpropagatedin a given direction by sendinga
beam of rays of approximately this wave-length in this direction
on to the body: then A is defined to be the ratio of the radiation
absorbedby the body (measuredasenergy)to the incident radiation.
We definea perfectabsorber,or black body, asa body whosecoefficient
of absorption is unity for all typesof radiation.
1 Trans. R. S. Edin. xxii (Mar. 1858),p. I ; cf. Rayleigh, Phil. Mag. i (1901),p. 98
I Stokes,Mem. and Sci. Corresponderu:e,
ii, p. 84
I Berlin MOl1alsber. (Oct. 1859),p. 662; Phil. Mag. xix (1860),p. 193; Ann. d. Phys.
cix (1860),p. 148
Berlin Monalsber. (Dec. 1859),p. 783; Ann. d. Phys. cix (1860),p. 275
Phil. Mag. xxi (1861),p. 240; Chern. News, iii (1861),p. 115
Berlin Abhandl. (1861[Phys.]),p. 63; (1862[Phys.J),p. 227; Ann. d. Phys. cxviii (1863),
p. 94; cr. W. Ritchie, Ann. d. Phys. xxviii (1933),p. 378
A proof of the Stewart-Kirchhoff theorem simpler than Kirchhoff's was given by
E. Pringsheim,Verh. deutsch.Phys. Ces. iii (1901),p. 81 .
• To him belongsthe credit of having been the first to seekfor, and find, metals
other than sodiumin the sun by spectroscopic methods.
(823) 371
Considera body C inside the cavity formed by a hollow perfectly

absorbing shell B, the whole being at the sametemperature. Since
heat-energyis lost as fast asit is gained by every part of the system
in this equilibrium state, the total energy radiated by C must be
equal to that part of the radiant energyproceedingfrom B which is
absorbedby C. Now let B be moved about relative to C, or even
replacedby another shell at the sametemperature. The amount of
energy radiated by C is unaffected, and therefore the part of the
radiant energy proceedingfrom B which is absorbedby Cremains
constant when B varies. But the absorbing coefficient may vary
in any manner from point to point of C and from one wave-length
to another, so this constancyrequires that the energy of radiation
of any particular wave-lengthincident on a surface-elementexposed
anywhere within a hollow black shell must be independent of the
nature and position of the shell, and must depend only on its
temperature.
Now consider the interchange of radiation, of wave-lengthsin
+
the range from A to A dA, between surface-elementsdy on C
and dfJ on B. The energypassingfrom dj3 to dy is eo, where this is
the energy of radiation of this type which would be exchanged
betweendy and dj3 if both C and B were black. The energypassing
from dy to dj3 consistsof two parts, namely, first, the radiation of this
type which is actually emitted by dy (which is not now supposed
to be necessarilyblack) and incident on dj3, and which we shall
denote bye, and secondly,the radiation originally proceedingfrom
all parts of B which is reflected or scatteredby dy in such a way
asto fall on dfJ. We shall denotethis byj, so
eo = e +j.
But a systemis possiblewhich differs from this systemonly in that
every ray takesthe samepath but in the oppositedirection: if we
considerthis system,we seethatfis equal to the part of the radiation
emitted by dfJ which is reflectedor scatteredat dy so asto fall some-
where on B: that is, f = eo - eoa.
Hence
or e
a = eo·
This is the law of Balfour Stewart and Kirchhoff, that for a given
type of radiation (specifiedby its wave-length,direction of propaga-
tion and state of polarisation) the emissivepower of an} body at a given
372
temperature,divided by its coefficientof absorption, is the samefor all bodies,

and is equal to the emissivepower of a black body at that temperature.1
We can now proceed to a definition of the intrinsic brightness,
or specificintensity of radiation of a surface. Consider the energy
radiated within the range of wave-lengths~ to A +d~, in the interval
dt of time, and contained within a narrow cone of solid angle dw,
by an area dS of surface, when the direction of the axis of the cone
makesan angle i with the normal to the surface. It will evidently
be proportional to d~, dt, dS, cos i, and dw, and we may therefore
write it H dt dS cosidw dA
where H will be called the intrinsic brightness of the surface with
respectto radiations in the range ~ to ~ +
d~. A simple integration
showsthat the total energy radiated per secondby unit area in this
range of wave-lengthsis 7TH.
Ifwe imagine a closedc::avity,with walls at a uniform tempera-
ture T, and occupied by a medium of refractive index unity, then
the cavity will be filled with radiation: the energyper c.c. represented
by radiations in the range ~ to ~ +
d~ can be representedby
F(A, T) dA
whereF(A, T) is independent of the nature of the substancecomposing
the walls. We can now connect F ('\, T) with the intrinsic brightness
of the walls, supposed black. Within the range ~ to '\+d'\, the
element dS of surface emits energy H cosi dSdw d'\ per secondinto a
cone of solid angle dw whose direction makes an angle i with the
to dS. This travels with velocity c, sothe energy due to dSper normal
unit length of the cone is (lfc) H cosi dS dw d,\. Thus this energy
per c.c. at distancer along the cone is
~ H cosidSdA.
cr
The energyper c.c. at this point due to the entire radiating surface
is therefore 1 rr cosi 1 rr
c
H dA)) 7 dS or H dA)) dQ c
(where dQ is the element of solid angle subtended at the point)
or 4rr H dA.
c
Thus we have
4rr
F(A, T) = - H.
c
For more rigorousproofsof the law, seeG. C. Evans (Proc. Amer. Acad. xlvi [1910],
1
p. 97), and D. Hilbert (Phys.Zs. xiii [1912], p. 1056; xiv [1913], p. 592).
373
Kirchhoff showed that the state of radiation, thus described as

existingwithin a hollow chamberwhosewalls are at temperatureT,
can be excited if the walls are perfectly reflecting (so that they do
not radiate on their own account), provided that a body (which may
be of any kind, so long as it has some emissivepower for every
wave-length) is placed somewhereinside the chamber and main-
tained at temperature T.
The problem confronting physicists after the discoveries of
Balfour Stewart and Kirchhoff was to find the function F(.\, T).l
The first step towards this was made in 1879 by the Austrian Josef
Stefan 2 (1835-93), who found empirically that the energy per c.c.
of the total radiation, that is JF(.\,T) d.\, was proportional to the
fourth power of the absolute temperature. Five years later Boltz-
mann 3 proved this theoretically, by combining the SecondLaw of
Thermodynamics with the theory of the pressure of radiation.
Let u = JF(.\, T) d.\ be the total aethereal energy per c.c. in a
hollow chamber whose walls are at temperature T. Then as we
have seen,"the pressureof the radiation on a surface-elementdS
exposedto this radiation is PdS where P = tu. Now regard the
aether within a hollow extensible vesselat temperature T as the
, working substance' of a heat engine. The total internal energy
is U = uv, where v is the volume of the vessel; and the element
of mechanical work performed on the external world (by radiation
pressure)is dA = Pdv or tudv. Now if Q be the heat absorbedfrom
the external world,
d$ or dU t dA is a perfect differential.
so ~{d(uv)+~udv} or ~{v~~dT+;udv}
is a perfect differential. Therefore
o (v du) 0 (4 u ) 1 du 4 du 4 u
ov TaT = oT" 31" or TdT = 3TdT -3P
or du 4u
ilT =-1'-
whence u = aT" where a is a constant,
which is Stefan'sformula.
1 The assertion that' all bodies become red-hot at the same temperature,' which may
be regarded as the germ from which the recognition of the function F (,\, T) ultimately
developed, was first made by Thomas Wedgwood, Phil. Trans. !xxxii (1792), p. 270.
• Wien Ber. lxxix, abt. ii (1879), p. 391
• Ann. d. Phys. xxii (1884), pp. 31, 291 ; cr. also B. Galitzine, Ann. d. Phys. xlvii (1892),
p. 479 • cr. p. 275
374
Meanwhile, progress was being made in understanding the

line-spectraof the chemicalelements. In 1871 G.JohnstoneStoney1
of Dublin expressedthe view that the best preparation for studying
the structure of thesespectrawould be to map them on a scaleof
wave-numbers,i.e. reciprocals of wave-lenths.2 This suggestion
was carried out by another Dublin worker, Walter Noel Hartley 3
(1846-1913), who showedthat in the caseof, for example,the triplets
which occurin the spectrumof zinc, the intervals betweenthe three
componentsof a triplet are the samefor all the triplets, when these
intervals are expressedin terms of wave-numbers. This was the
first discovery of a connection between lines in a spectrum that
could be expressedby a numerical law ; it pointed to the addition
and subtraction of wave-numbersas being the key to the analysis
of spectra. In the sameyear two Cambridge men, GeorgeDowning
Liveing (1827-1924) and James Dewar (1842-1923) noticed ( that
certain setsof lines in the spectraof the alkalis and alkaline earths
evidently formed series. 'This relation,' they said [of the relations
between the lines], 'manifests itself in three ways: first, by the
repetition of similar groupsof lines; secondly,by a law of sequence
in distance, producing a diminishing distance between successive
repetitionsof the samegroup asthey decreasein wave-length; and
thirdly, a law of sequenceas regards quality, an alternation of
sharper and more diffuse groups, with a gradually increasing
diffusenessand diminishing intensity of all the related groups as
the wave-length diminishes. The first relationship has long since
been noticed in the case of the sodium lines, which recur in
pairs.'
In 1885 Alfred Cornu 6 (1841-1902) observed sequencesof
doublets in the ultra-violet spectra of thallium and aluminium,
which, from the similarity of their behaviour with regard to
reversibility, he concluded to be the successivemembers of a
senes.
The time was now ripe for the discovery of a numerical law
connectingthe different lines of a series. This was made in 1885 by
1 Rep.Brit. Ass. (1871), notes and abstracts, p. 42

• A wave-number is generally defined as 10',\-1 where ,\ is a wavelength expressedin
Angstroms; so that it representsthe number of waves in one em. in air.
• Jour. Chern.Soc.xliii (1883), p. 390
• Phil. Trans. c1xxiv (1883), p. 187 (at pp. 213, 214)
I Comptesrendus,e (1885), p. 1181
375
the SwissphysicistJohann Jakob Balmer I (1825-98), who gavethe

formula ml
A = Ao 2 4 (m = 3, 4, 5, 6, ... )
m -
for thewave-lengthsof thewell-known hydrogenlines
Ha = "6563, Hp = ;\4861, Hy = ;\4340, Ha = ;\4102, .••
"0
where is a constant. If this formula is expressedin wave-numbers
it becomes
n = no
)
(I - :2
(m = 3,4,5, ••. )
In 1890 Heinrich Kayser (1853-1940) and Carl Runge (1856-

1927) gave II the names Principal Series, First Subordinate Series and
SecondSubordinate Series, to the seriesof doubletsin the spectraof the
alkali metals which had been recognisedby Liveing and Dewar.
In the caseof sodium, the principal seriesconsistsof the D lines, a
doublet at "3303 and "3302, and further lines in the ultra-violet.
Thesedoublets are sharp and easily reversed. The first subordinate
series consists of strong and diffuse doublets, while the second
subordinate seriesconsistsof fainter but sharper doublets. Each of
theseseriesconvergesto a limit or ' head' at its ultra-violet end.
In the same year Johannes Robert Rydberg 3 of Lund (1854-
1919) gave a new general formula for series,which was capableof
representingall the seriesin the spectraof the alkalis, aswell asthe
hydrogen series,namely R
n = no - (m P)I +
where n is the wave-number, no and p are constants,m takes in
successionthe values of the positive whole numbers, and R is a
constant commonto all series and to all elements,having a value which
he gave as R = 109721.6cm-l.
R is called Rydberg's constant. Balmer's formula for the hydrogen
seriesthus becomes n =R (i _ ~I)'
In the Principal Series of the alkali metals, the no of Rydberg's
formula has the samevalue for the two lines of a doublet, but p has
different values,sayPI and PII, for the lessand more refrangiblelines
1 Basel Verh.vii (1885), pp. 548, 750
• Ann. d. Phys.xli (Sept. 1890), p. 302; Abhandl.Berlin Akad., 1890 (Anh.), 66 pp.
• Phil. Mag. (6', xxix (Apr. 1890), p. 331 ; Kongl. Sv. Velensk.Ak. Bandl. xxxii (1890),
DO. 11
respectively. For the first or diffuse subordinate series,P has the

samevalue (say d) for the two lines of any doublet; and similarly
for the second or sharp subordinate series,P has the same value
(say s) for the two lines of any doublet. The convergencewave-
number no has the samevalue for the more refrangible members of
the doublets in the two subordinate series. Thus we can write:
Principal series :
R
lessrefrangible lines of doublets, n = no - (m + PI)'
more refrangible lines of doublets, n=n ----- R
o (m + P,)B"
First or dijJuse subordinate series:
, R
lessrefrangible lines of doublets, n = no - (m + d)'
more refrangible lines of doublets, n = no
"- (m +R d)' •
Secondor sharp subordinate series :
R
lessrefrangible lines of doublets, n = no' -----
(m + s)P
more refrangible lines of doublets, n = no" - (m ~ s) ,.
In 1896Rydberg discovered1 that, with this notation,
, R " R
no = (2 + PI)" no = (2 + PI)' ;
and with a similar evaluation of no which followed, we have finally:
Principal series :
lessrefrangible lines of doublets,
R R
n = (1 + s). - (m + PI)" m = 2, 3, 4, 5, •••
more refrangible lines of doublets,
R R
n = (1 + s)' - (m + p,)_' m = 2,3,4,5, •••
First subordinate series :
R R
n = (2 +
P,)_ - (m d)1' m + = 3,4,5, •••
R R
n = (2 + PI) , - (m + d) " m = 3, 4, 5,. • •
1 Ann. d. Phys. Iviii (1896), p. 674. The same law was found independently by
A. Schuster,Nature Iv (Jan. 1897), pp. 200, 223
377
Secondsubordinate series :
R R
n = (2 + PI)I (m + S)I' m = 2,3,4,5, •••
R R
n = (2 + Pl)1 - (m + S)2' m = 2, 3, 4, 5,
It wasperceivedby Rydberg 1 and by Walther Ritz 2 (1878-1909)
that the wave-numbersof all the lines of the principal seriesand the
two subordinateseriesare differencesof numberstaken from the sets
1 1 1 1
(1 + s)" (2 + s)" (3 + s)" (4 + S)I' ..•
1 1 1
(2 + Pl)" (3 + Pl)" (4 + Pl)!'
1 1 1
(2 + p,)I' (3 + Pz)" (4+P2)1'
1 1
(3 + d)I' (4 +
d)2'
and they accordingly assertedthe principle (generally known as the
combination-principle) that to the spectrum of arry element there corresponds
a set of ' terms,' such that the wave-numbersof the lines of the spectrum are
the dijferencesof theseterms taken in pairs. It wassoonfound that in the
sodium spectrum there are lines not belonging to any of the three
series above-mentioned, whose wave-numbers are differences of
terms taken from the aboveset: such as the lines (in the infra-red)
R R R R
and
(2 + S)I (3 + Pi)1 (2 + s) I (3 + PI)I
and the lines (in the infra-red)
R R R R
and
(3 + Pi)1 - (4 + d)1 (3 + PI)I (4 + d)1
Meanwhile another serieshad been discoveredin the spectraof the
alkalis. In 1904 F. A. Saunders8 found five new lines in the infra-
1 Rapp. pres. au congresinto (Paris, 1900),ii, p. 200
I Phys. <s, ix (1908),p. 521; Ges. Werke, p. 141
Astrophys. Jour. xxviii (1908),p. 237; Ges. Werke, p. 163. Many of the predictionsmade
possibleby Ritz's theory were verified by F. Paschen,Ann. d. Phys. xxvii (1908),p. 537;
xxix (1909),p. 625.
I Astrophys. Jour. xx (1904),p. 188; xxviii (1908), p. 71
378
redspectrumof caesium; and in 1907it wasshownby A. Bergmann1

that these belong to a serieswhich has been called the 'F' or
fundamental series,and which existsin the spectra of all the alkalis
(where it is in the infra-red) and many other elements(where it is
not necessarilyin the infra-red). Its head or limit has the wave-
number Rj (3 + d) 2,and its lines involve a new seriesof terms
111
(4 +
f)I' (5 f)I' + ----
(6 f)I' .•• +
I t was noticed that under normal circumstances the terms
involving s combine only with the PI or P2 terms, that the PI and Pt
termscombineonly with sand d terms,the d termsonly with P terms
andfterms. On this account,the setsof terms are usually arranged
in the order s, P, d, f . .. With the aid of powerful spectroscopes,
it was found that the d terms and f terms are doubled like the
P terms.
Let us now return to consider the radiation emitted by a black
body at the absolute temperature T; as before, we denote by
F('\,T)d'\ the energy per c.c.3 representedby radiations of wave-
lengthsbetween,\ and A + d'\.
In 1893 W. Wien,2 in studying the form of the function
F('\,T) made an important discovery,which will now be described.
Supposethat a perfectly reflectingsphereis contracting, its radius
at time t being r. Consider a beam of light incident on the inner
surfaceof the sphereat an angle i. At eachreflection we have,3 if T
be the period of the light, 8T 2 cosi dr
T c dt
wherec is the velocity of light.
The length of path betweentwo impacts is 2r cosi, sothe interval
of time between two impacts is 2r cos ijc. So if dT denotes the
dt
changein T in one second,we have
2r .dT
8T = C cos t dt'
IdT
whence T dt rldrdt' or T = Constant X r.
Thus as the spherecontracts, the period of the light also contracts,
so asalwaysto remain proportional to the radius of the sphere.
1 Dissertation, lena, 1907. cf. E. Riecke, Phys. .('s. ix (1908),p. 241
• Berlin Sitl;. (9 Feb. 1893),p. 55; Ann. d. Phy,,".Iii (1894), p. 132. A more careful
proof wasgiven by E. Buckingham,Phil. Mag. xxiii (1912),p. 920.
• cf. p. 368
379
AETHER AND ELECTRIcrrY
Moreover if £ denotethe amount of energyin onewave-lengthfor

this particular radiation, we have in the sameway
s£ 2 cos i dr 2r cos i d£
;- = - -c- dt and S£ edt'
so 1 d£ 1 dr
~ dt = - ,.dt' or £T = Constant.
Thus the total amount of radiant energy in the spherevaries as 1fT,
so the radiant energy per c.c. varies as Ifr4. for the wave whose
shrinkage we are following, and therefore for all the waves. Now
supposethat the sphereis initially filled with radiation of all wave-
lengths soas to be in equilibrium with a black body at somedefinite
temperature T. Since the total energy per c.c. under thesecircum-
stancesvaries as T4.,we seethat the total radiant energy within the
sphere at any instant is that appropriate to equilibrium with a
radiating enclosureat temperature varying as 1fT. We have not yet
shown that the partition of energy amongthe different wave-lengths
which is reached by the compressionis one which would be in
equilibrium with a radiating enclosure. This however follows from
the fact that if it were not so, the radiation of certain wave-lengths
would have an energy correspondingto a higher temperature than
the radiation in certain other wave-lengths; and by introducing
bodies capable of absorbing theserespectivewave-lengthsonly, we
should be able to obtain differences of temperature from which
mechanical work could be obtained in contravention of the Second
Law of Thermodynamics.
It follows that if, in a distribution of radiant energy which is in
equilibrium with a radiating enclosureat temperatureT, we multiply
the wave-lengthsby k (= T2fT1) and multiply the energy per c.c. in
each wave-length by k-4., we shall obtain a distribution of radiant
energy which will be in equilibrium with a radiating enclosureat
temperature k-1T.
Thus k-& F (A, T) dA = F (kA, k-1T) d (kA)
whencewe seethat F (.:\,T) d.:\ must be of theform T"4>(T.:\)fiA, which
in view of later developmentswe may write T.:\-44>(TA) d.:\, where 4>
is afunction of its argument T.:\only. This is Wien's theorem.
Henceit follows that if .:\m denotesthe wave-lengthcorresponding
to the maximum ordinate of a graph in which energyof radiation is
plotted against wave-length,we must have
.:\mT = Constant,
so .:\tn is inverselyproportional to the temperature. This is known as
Wien's displacementlaw.
380
Three years later Wien 1 investigated the form of the function

~(T>..). He remarked that the law, that in an exhaustedvesselthe
radiation is the sameasthat from a black body at the sametempera-
ture as the walls of the vessel,holds also if the radiating body be a
gaswhich is shut offfrom the vacuousspacebya transparentwindow,
and from the exterior by reflecting walls. We postulate, therefore,
a gas which possesses a finite absorptive power for radiation of all
wave-lengths,and which is radiating under purely thermal excite-
ment. We suppose that Maxwell's law of the distribution of
velocities holds,2 so the number of molecules whose velocity lies
betweenv and v +
dv is proportional to
-vI/a"
vIe dv,
where a denotesa constant,which is connectedwith the mean square
of velocity by the equation _ 3
V! -- -2 at .
The absolute temperature is therefore proportional to a2• We

assumethat each molecule sendsout vibrations whosewave-length
and intensity dependonly on the velocity of the molecule.
As the wave-length >..of the radiation emitted by any molecule
is a function of v, it follows that v is a function of >... The intensity
F (A,T) of the radiation whose wave-length lies between >..and
>..+ d>" is therefore proportional to the number of molecules that
send out radiations of this period, and also to a function of the .
velocity v (i.e. a function of >..). Thus f(>')
F(A, T) =g (A) e--1'

where g and f denote two unknown functions. But we know from
Wien's previouswork that
F(A, T) = A_6~(TA).
Combining theseresults,we have b
F(A, T) = C A-6 e->.T,

which is Wien's law of radiation. Experimental determinations were
found to agree with this formula in the region of high frequencies,
but not for long waves. At the other end of the graph of F(>..,T)-
1 Ann. d. Phys. Iviii (1896),p. 662; Phil. Mag. xliii (1897),p. 214
" It may be mentionedthat an attempt to determinethe law of radiation of a black

body by assumingthat Maxwell's law of the distnbution of velocitiesholds for its
molecules,had beenmadeby Vladimir A. Michelson,Journal de Physique(2), vi (1887),
p.467.
the region oflow frequencies-a different formula wasfound to hold,

the derivation of which must now be explained.
When severaldifferent gases(which may be initially at different
temperatures) are mixed, the mixture becomesultimately of the
same temperature throughout. Since the temperature is pro-
portional to the averagekinetic energyof a molecule,we seethat the
systemtends to assumea state in which the averagekinetic energy
of a moleculeis the samewhether the moleculebe light or massive:
the inferiority in massof the lighter moleculesis compensatedby
their moving with greater velocities. This is a particular instance
of what is known asthe theoremof equipartition of energy, which asserts
that in a state of statistical equilibrium at absolute temperatureT, the total
energy of any systemwhich obeysthe laws of dynamics, is partitioned in such
a way that every degreeof freedom possesseson the average the same kinetic
energy, namely tkT, where k is known as Boltzmann's constant. Thus for
an atom of a monatomic gas, which has three degreesof freedom,
the mean kinetic energy at temperature T is ikT. The theorem is
due to Maxwell 1 and Boltzmann.2
Some very familiar phenomenacan be explained by reference
to the law of equipartition of energy. For instance,if a hollow rigid
vesselis filled with air, and an arbitrary disturbanceis setup in the
air, this disturbance will in time become dissipated into random
motionsof the molecules,suchasare studied in the kinetic theory of
gases. The reasonis that any motion in the air can be regardedasa
superpositionof its various typesof free vibration. Now the number
of types of free vibration is very great, and thosetypeswhich are of
extremely short wave-length vastly outnumber the types of longer
wave-lengths. Each type of vibration countsasa degreeof freedom
in the equipartition theorem, and the theorem therefore predicts
that after a time the energyof the disturbancewill be practically all
in the vibrations of very short wave-length,i.e. it will have become
degraded into small irregular disturbances. The random motion
of the moleculescan be analysedby Fourier's theorem into waves,
and the energyrepresentedby wavesof lengthsA to ;\ + d;\ is found
to be 47TkT;\-4d;\ per unit volume.3 The thermal motion of the
moleculesof a gas is, in fact, essentiallysound of very short wave-
lengths.
1 Phil. Mag. xix (Jan. 1860),p. 19; xx (July 1860),p. 21; Trans. Comb. Phil. Sac. xii
(1878),p. 547
• Wien Sitzungsber. Iviii (1868(Abth. 2]), p. 517; lxiii (1871(Abth. 2]), p. 397
• This is half the value found below for the aether,the differencebeing due to the
fact that vibrationsin a gasare longitudinal, and in the aetherare transverse,eachstate
of polarisationof a wavecountingasa separatedegreeof freedom.
382
There has however been much difference of opinion regarding

the rangeof validity of the doctrine of equipartition of energy. Lord
Kelvin 1 believed that he had disproved the theorem altogether by
a ' decisivetestcase.' The questionwastaken up by Lord Rayleigh,2
who showedthat Kelvin's reasoningwas inconclusive; although he
recognisedthat there were difficulties in applying the theorem to
actual gases.'In the caseof argonand helium and mercury vapour'
hesaid,, the ratio of specificheats(1'67) limits the degreesoffreedom
of each moleculeto the three required for translatory motion. The
value (1'4) applicable to the principal diatomic gasesgives room
for the three kinds of translation and for two kinds of rotation.
Nothing is left for rotation round the line joining the atoms,
nor for relative motion of the atoms in this line. Even if we
regard the atoms as mere points, whose rotation means nothing,
there must still exist energy of the last-mentioned kind, and its
amount (according to the law) should not be inferior.' The
discussionwas continued by Poincare,3G. H. Bryan" and others.6
In 1900 Lord Rayleigh,6 assuming the truth of the theorem of
equipartition of energy, derived from it a formula for the energy of
black-body radiation. It wasobtained asfollows:
In the free vibrations of the aether inside a cubical enclosure
whoseedge is of length t, each mode of vibration correspondsto
valuesof the componentsof the electric and magneticforcesinvolving
a factor of the form
c?s (PTTX) c.os(flTTY) c.os(rTTZ)
sm t sm t sm t
wherep, q, r are whole numbers. The frequency v of this vibration is
c
v = 'Ii (p2 + q2 + r2)+,
where c denotesthe velocity of light. We have to find how many
modesof vibration correspondto the range v to v +
dv of frequency.
Let whole-numbervaluesof (p, q, r) be regarded as the co-ordinates
of a point; the whole systemof points constitutesa cubic array of
volume-density unity. Let R denote the distance of the point
1 Proc. R. S. vol. L (1891),p. 85; Phil. Mag. xxxiii (May 1892),p. 466
• Phil. Mag. xxxiii (April 1892),p. 356; Phil. Mag. xlix (Jan. 1900),p. 98
• Revuegeneraledes Sciences,July 1894
• ' Report on Thermodynamics,'Brit. Ass. Rep. (1894),p. 64; Proc. Comb. Phil. Soc.
viii (1895),p. 250
• W. Peddie,Proc. R. S. Edin. xxvi (1905-6), p. 130j J. H. Jeans,Phil. Mag. xvii
(1909),p. 229 j J. Larmor, Brit. Ass. Rep. (1913),p. 385
• Phil. Mag. xlix (1900),p. 539; Nature, !xxii (May 1905),p. 54 and (July 1905)
p. 243. cr.J. H. Jeans,Phil. Mag. (6), x (July 1905),p. 91
383
(p, q, r) from the origin, soR 2 = p2 + q2 + r2.

The numberof points
in the shell for which R rangesfrom R to R + dR is equal to the
included volume (sincep, q, r are positive, we take only the positive
octant), which is 17TR2dR.
Hence the number of setsof values of p, q, r correspondingto the

range v to v + dv is
i (21)3
7TC v 2
dv
or in terms of A (where v = cfA)

47T13'A-4d'A.
To obtain the number of modesthat are to be counted in applying
the theorem of the equipartition of energy, we must double this
number, since the vibrations are transverse,and double it again,
since the whole energy is (on the average)double the electric or
magnetic energy taken alone. Thus the number of modesto be
counted is
At temperature T, each mode possesses

the average energy ikT.
This givesfor the total energy
87T!3kT'A -4d>...
so the energyof radiation per c.c. in the range of wave-lengthsfrom
A to A + dA is F(i\, T) = 87TkTi\ -4di\
where k is Boltzmann's constant. This is Rayleigh'sformula. k has
the value 1·38 X 10-18 erg. (degree)-l.
It is obvious that according to this formula the radiant energy
would tend to run entirely into wavesof the highestfrequency,i.e.
of infinitesimal wave-length. This is due fundamentally to the fact
that the aether is regarded as a continuum, and has therefore an
infinite number of degreesof freedom. Rayleigh'sformula therefore
cannot representthe true law of black-body radiation. It is, how-
ever, asymptotically correct for the long waves, as was shown
experimentally by H. Rubens and F. Kurlbaum in 1900-1,1 and
it was deduced from the electron theory by Lorentz,2 who calcu-
lated by help of that theory the emissivepower and absorptioncoeffi-
cient of a thin sheetof metal on which radiation of great wave-length
is incident.
The subsequenthistory of the radiation problem must be post-
1 Berlin Sitzungsb. xli (Oct. 1900),p. 929; Ann. d. Phys. iv (Apr. 1901),p. 649
• Proc. Arnst. Acad. v (1902-3), p. 666
384
poned to a later chapter. It may be mentioned that Lord Kelvin

refusedto the last to acceptthe equipartition theorem. Referring to
a remark of Rayleigh's-' What would appear to be wanted is some
escapefrom the destructive simplicity of the general conclusion'-
he said: 'The simplestway of arriving at this desired result is to
deny the conclusion; and so, in the beginning of the twentieth
century, to losesight of a cloud which has obscuredthe brilliance of
the molecular theory of heat and light during the last quarter of the
nineteenth century.' 1
I Kelvin, Baltirrwre Lectures (1904), p. 527
Chapter XIII
CLASSICAL THEORY IN THE AGE OF LORENTZ
THE attempts of Maxwell 1 and of Hertz 2 to extend the theory of

the electromagnetic field to the casein which ponderable bodies
are in motion had not been altogether successful. Neither writer
had taken account of any motion of the material particles relative
to the aether entangledwith them, so that in both investigationsthe
moving bodieswere regardedsimply ashomogeneousportions of the
medium which fills all space,distinguishedonly by specialvaluesof
the electricand magneticconstants. Suchan assumptionis evidently
inconsistent with the admirable theory by which Fresnel3 had
explained the optical behaviour of moving transparent bodies;
it was therefore not surprising that writers subsequentto Hertz
should have proposedto replace his equationsby othersdesignedto
agree with Fresnel's formulae. Before discussingthese, however,
it may be well to review briefly the evidencefor and against the
motion of the aether in and adjacent to moving ponderablebodies,
asit appearedin the last decadeof the nineteenthcentury.
The phenomenaof aberration had been explained by Young'
on the assumptionthat the aether around bodies is unaffected by
their motion. But it was shown by StokesII in 1845that this is not
the only possibleexplanation. For supposethat the motion of the
earth communicates motion to the neighbouring portions of the
aether; this may be regardedassuperposedon the vibratory motion
which the aethereal particles have when transmitting light: the
orientation of the wave-frontsof the light will consequentlyin general
be altered; and the direction in which a heavenly body is seen,
being normal to the wave-frontswill thereby be affected. But if the
aetherealmotion is irrotational, sothat the elementsof the aetherdo
not rotate, it is easily seenthat the direction of propagation of the
light in spaceis unaffected; the luminous disturbanceis still propa-
gated in straight lines from the star, while the normal to the
wave-front at any point deviatesfrom this line of propagationby the
small angle u/e, where u denotesthe component of the aethereal
velocity at the point, resolvedat right anglesto theline of propagation,
and e denotesthe velocity of light. If it be supposedthat the aether
1 cf. p. 259 • cf. p. 329 • cf. p. 108 • cf. p. 108
• Phil. Mag. xxvii (1845), p. 9; xxviii (1846), p. 76; xxix (1846), p. 6
386
THE AGE OF LORENTZ
near the earth is at rest relatively to the earth's surface,the star will
appearto be displaced towards the direction in which the earth is
moving, through an angle measuredby the ratio of the velocity of
the earth to the velocity of light, multiplied by the sineof the angle
betweenthe direction of the earth's motion and the line joining the
earth and star. This is preciselythe law of aberration.
An objection to Stokes's theory was pointed out by several
writers, amongstothersby H. A. Lorentz.l This is, that the irrota-
tional motion of an incompressiblefluid is completely determinate
whenthe normal componentof the velocity at its boundary is given;
so that if the aether were supposedto have the samenormal com-
ponentof velocity asthe earth,it would not havethe sametangential
componentof velocity. It follows that no motion will in general
exist which satisfiesStokes'sconditions; and the difficulty is not
solvedin any very satisfactoryfashion by either of the suggestions
which havebeenproposedto meetit. One of theseis to supposethat
the moving earth does generate a rotational disturbance, which
however, being radiated away with the velocity of light, doesnot
affect the steadier irrotational motion; the other, which was
advancedby Planck 2 in 1899,is that the two conditions of Stokes's
theory-namely, that the motion of the aether is to be irrotational
and that at the earth's surfaceits velocity is to be the sameas that
of the earth-may both be satisfiedif the aether is supposedto be
compressiblein accordancewith Boyle'slaw, and subjectto gravity,
so that round the earth it is compressedlike the atmosphere; the
velocity of light being supposedindependentof the condensationof
the aether.
Lorentz,a in calling attention to the defectsof Stokes'stheory,
proposedto combine the ideas of Stokesand Fresnel,by assuming
that the aethernear the earth is moving irrotationally (asin Stokes's
theory), but that at the surfaceof the earth the aetherealvelocity
is not necessarilythe sameas that of ponderable matter, and that
(as in Fresnel's theory) a material body imparts the fraction
(/,,2 _ l)fjL2 of its own motion to the aether withi:l it. Fresnel's
theory is a particular caseof this new theory, being derived from it
by supposingthe velocity potential to be zero.
Aberration is by no means the only astronomical phenom-
enon which dependson the velocity of propagation o( light; we
have indeed seen4 that this velocity was originally determined by
1 Archives NierI, xxi (1896), p. 103
• cr. Lorentz, Proc. Amst. Acad. (English edn.), i (1899),p. 443
• Archives NierI. xxi (1886),p. 103; cr. also ZittinsgsversI. Kon. Ak. Amsterdam (1897-8),
p. 266 • cr. p. 22
(823) 387
CLASSICAL THEORY IN
observing the retardation of the eclipsesof Jupiter's satellites. It

was remarked by Maxwell 1 in 1879 that these eclipsesfurnish,
theoretically at least, a means of determining the velocity of the
solar systemrelative to the aether. For if the distancefrom the
eclipsedsatelliteto the earth be divided by the observedretardation
in time of the eclipse,the quotient representsthe velocity ofpropaga-
tion of light in this direction, relative to the solar system; and this
will differ from the velocity of propagation of light relative to the
aether by the component, in this direction, of the sun's velocity
relative to the aether. By taking observationswhen Jupiter is in
different signs of the zodiac, it should therefore be possible to
determine the sun's velocity relative to the aether, or at least that
componentof it which lies in the ecliptic.
The sameprinciples may be applied to the discussionof other
astronomicalphenomena. Thus the minimum of a variable star of
the Algol type will be retarded or acceleratedby an interval of time
which is found by dividing the projection of the radius from the sun
to the earth on the direction from the sun to the Algol variable by
the velocity, relative to the solar system,of propagation of light
from the variable; and thus the latter quantity may be deduced
from observationsof the retardation.2
Another instance in which the time taken by light to crossan
orbit influencesan observablequantity is affordedby the astronomy
of double stars. Savary 3 long ago remarked that when the plane
of the orbit of a double star is not at right anglesto the line of sight,
an inequality in the apparent motion must be causedby the circum-
stancethat the light from the remoter star hasthe longerjourney to
make. Yvon Villarceau 4 showedthat the effect might be repre-
sentedby a constant alteration of the elliptic elementsof the orbit
(which alteration is of course beyond detection), together with a
periodic inequality, which may be completely specified by the
following statement: the apparent co-ordinatesof one star relative
to the other have the valueswhich in the absenceof this effect they
would have at an earlier or later instant, differing from the actual
time by the amount
m1 - m2 Z
---
m1 +m -, c
2•
where ml and m1 denote the massesof the stars, c the velocity of

1 Proc. R. S. xxx (1880),p. 108
I The velocity of light was found from observationsof Algol, by C. V. L. Charlier,
Ofversigt af K. Vet.-Ak. Forhandl. xlvi (1889),p. 523. I Conn. des Temps, 1830
• Additions Ii la Connaissancedes Temps, 1878; an improved deductionwasgiven by

H. Seeliger,Sit<;ungsberichted. K. Ak. <;u Miinchen, xix (1889),p. 19.
388
THE AGE OF LORENTZ
light, and z the actual distanceof the two starsfrom each other at
the time when the light wasemitted, resolvedalong the line of sight.
In the existing state of double-star astronomy, this effect would be
maskedby errors of observation.
Villarceau alsoexaminedthe consequences of supposingthat the
velocity of light dependson the velocity of the sourceby which it is
emitted. If, for instance,the velocity of light from a star occulted
by the moon were lessthan the velocity of light reflected by the
moon, then the apparent position of the lunar disk would be more
advancedin its movementthan that of the star, so that at emersion
the star would first appear at somedistanceoutside the lunar disk,
and at immersion the star would be projected on the interior of the
disk at the instant of its disappearance. The amount by which the
image of the star could encroachon that of the disk on this account
could not be somuch as0"'71 ; encroachmentto the extent of more
than 1" has been observed,but is evidently to be attributed for the
mostpart to other causes.1
Among the consequences of the finite velocity of propagation of
light which are of importance in astronomy,a leading place must be
assignedto the principle enunciated by Doppler,2 that the motion
of a sourceof light relative to an observermodifies the period of the
disturbancewhich is receivedby him. The phenomenonresembles
the depressionof the pitch of a note when the source of sound is
receding from the observer. In either case, the period of the
vibrations perceivedby the observeris ( +
v)/e X the natural period,
wherev denotesthe velocity of separationof the sourceand observ:er,
and e denotesthe velocity of propagation of the disturbance. If, for
example,the velocity of separationis equal to the orbital velocity of
the earth, the D lines of sodium in the spectrum of the sourcewill
be displacedtowards the red, ascompared with lines derived from a
terrestrial sodium flame, by about one-tenth of the distancebetween
them. The application of this principle to the determination of the
relative velocity of starsin the line of sight, which hasproved of great
servicein astrophysicalresearch,wassuggestedby Fizeau in 1848.3
Passingnow from the astronomicalobservatory,we must examine
the information which has been gained in the physical laboratory
regarding the effect of the earth's motion on optical phenomena.
We have already 4 referred to the investigationsby which the truth
1 The question of the possible dependence of the velocity of light on the velocity of
the sourcewill be discussedin Volume II of this work, in connection with Ritz's' ballistic'
hypothesis. • cf. p. 368
• An apparatus for demonstrating the Doppler-Fizeau effect in the laboratory was
constructed by Belopolsky, Aslrophys. Jour., xiii (1901), p. 15. • cr.pp. 110-11
389
CLASSICAL THEORY IN
of Fresnel'sformula was tested. An experiment of a different type

was suggestedin 1852 by Fizeau,l who remarked that, unless the
aether is carried along by the earth, the radiation emitted by a
terrestrial sourceshould have different intensitiesin different direc-
tions. It was, however, shown long afterwards by Lorentz 2 that
such an experiment would not be expectedon theoretical grounds
to yield a positive result; the amount of radiant energy imparted
to an absorbing body is independent of the earth's motion. A few
yearslater Fizeau investigated3 another possibleeffect. If a beam
of polarisedlight is sentobliquely through a glassplate, the azimuth
of polarisation is altered to an extent which depends,amongstother
things, on the refractive index of the glass. Fizeau performed this
experiment with sunlight, the light being sent through the glassin
the direction of the terrestrial motion, and in the oppositedirection;
the readings seemedto differ in the two cases,but on account of
experimental difficulties the result was indecisive.
. Someyearslater, the effect of the earth's motion on the rotation
of the plane of polarisation of light propagated along the axis of a
quartz crystalwasinvestigatedby Mascart.4 The resultwasnegative,
Mascart stating that the rotation could not have been altered by
more than the If40,OOOth part when the orientation of the apparatus
wasreversedfrom that of the terrestrial motion to the oppositedirec-
tion. This wasafterwardsconfirmed by Lord Rayleigh,6 who found
that the alteration, ifit existed,could not amount to IflOO,OOOth part.
P. G. Tait in 1882 suggested6 that' if the aether be in motion
relative to the earth, the absolutedeviationsoflines in the diffraction
spectrumshould be different in different azimuths.'
As Maxwell pointed out in the letter of 1879 already referredto,7
in terrestrial methodsof determining the velocity of light the ray is
made to retrace its path, so that any velocity which the earth might
possesswith respect to the luminiferous medium would affect the
time of the double passageonly by an amount proportional to the
square of the constant of aberration.s In 1881, however, Albert
Abraham Michelson 9 (1852-1931) remarked that the effect, though
1 Ann. d. Phys. xcii (1854),p. 652 • Proc. Arnst. Acad. (Englishedn.), iv (1902),p. 678
• Annales de Chim. (3), !xviii (1860),p. 129; Ann. d. Phys. cxiv (1861),p. 554
• Annales de l'Ee. Norm. (2), i (1872),p. 157 • Phil. Mag. iv (1902),p. 215
• Tait's Life and Scientific Work (ed. C. G. Knott), p. 92
7 Proc. R. S. xxx (1880),p. 108
• The constantof aberrationis the ratio of the earth'sorbital velocity to the velocity
of light; cr. supra, p. 95.
• Amer. Jour. Sci. xxii (1881), p. 20. His method was afterwardsimproved; cf.
Michelson and E. W. Morley, Amer. Jour. Sci. xxxiv (1887),p. 333; Phil. Mag. xxiv
(1887),p. 449. SeealsoE. W. Morley and D. C. Miller, Phil. Mag. viii (1904),p. 753,
and ix (1905),p. 680.
390
THE AGE OF LORENTZ
of the secondorder, shouldbe manifestedby a measurabledifference

between the times for rays describing equal paths parallel and
perpendicular respectivelyto the direction of the earth's motion.
He producedinterferencefringes betweentwo pencilsof light which
had traversedpaths perpendicular to each other; but when the
apparatuswas rotated through a right angle, so that the difference
would be reversed,the expecteddisplacementof the fringes could
not be perceived. This result was regarded by Michelson himself
asa vindication of Stokes'stheory,1in which the aetherin the neigh-
bourhood of the earth is supposedto be set in motion. Lorentz,2
however,showedthat the quantity to be measuredhad only half the
valuesupposedby Michelson,and suggestedthat the negativeresult
of the experiment might be explained by that combination of
Fresnel's and Stokes'stheories which was developed in his own
memoir 8; since, if the velocity of the aether near the earth were,
say, half the earth's velocity, the displacement of Michelson's
fringeswould be insensible.
A sequelto the experiment of Michelson and Morley was per-
formed in 1897,when Michelson' attempted to determine by ex-
perimentwhetherthe relative motion of earth and aethervarieswith
the vertical height abovethe terrestrial surface. No result, however,
could be obtained to indicate that the velocity of light dependson
the distancefrom the centreof the earth; and Michelsonconcluded
that if there were no choicebut betweenthe theoriesof Fresneland
Stokes,it would be necessaryto adopt the latter, and to supposethat
the earth's influence on the aether extendsto many thousandkilo-
metresaboveits surface. By this time, however,aswill subsequently
appear,a different explanation wasat hand.
Meanwhile the perplexity of the subject was increased by
experimentalresultswhich pointed in the oppositedirection to that
of Michelson. In 1892Sir Oliver Lodge 6 observedthe interference
betweenthe two portions of a bifurcated beam of light, which were
madeto travel in oppositedirectionsround a closedpath in the space
betweentwo rapidly rotating steeldisks. The observationsshowed
that the velocity of light is not affectedby the motion of adjacent
matter to the extent of 1/200th part of the velocity of the matter.
Continuing his investigations Lodge 8 strongly magnetised the
1 cf. p. 386
• Arch. Nierl. xxi (1886), p. 103. On the Michelson-Morley experiment cf. also Hicks,
Phil. Mag. iii (1902), p. 9 • cf. p. 387
• Amer. Jour. Sci. (4), iii (1897), p. 475 • Phil. Trans. clxxxiv (1893), p. 727
• ibid. clxxxix (1897), p. 149. Seealso Lodge's' Silvanus Thompson Memorial Lecture:
Journal qf the Ron/genSocie!)'.xviii, July 1922.
391
CLASSICAL THEORY IN
moving matter (iron in this experiment), sothat the light waspropa-

gated across a moving magnetic field; and electrified it so that
the path of the beamslay in a moving electrostaticfield; but in no
casewas the velocity of the light appreciably affected.
We must now trace the stepsby which theoretical physicistsnot
only arrived at a solution of the apparent contradictions furnished
by experimentswith moving bodies, but so extendedthe domain of
electrical sciencethat it becamenecessaryto enlargethe conceptions
of spaceand time to contain it.
The first memoir in which the new conceptionswere unfolded
was published by Hendrik Antoon Lorentz 1 (1853-1928) in 1892.
The theory of Lorentz was, like those of Weber, Riemann and
Clausius,2a theory of electrons; that is to say, all electrodyna-
mical phenomena were ascribed to the agency of moving electric
charges, which were supposedin a magnetic field to experience
forces proportional to their velocities, and to communicate these
forces to the ponderable matter with which they might be
associated. 3
In spite of the fact that the earlier theoriesof electronshad failed

to fulfil the expectationsof their authors, the assumptionthat all
electric and magnetic phenomenaare due to the presenceor motion
of individual electric chargeswasone to which physicistswere at this
time disposedto give a favourable consideration; for, as we have
seen,4evidence of the atomic nature of electricity was now con-
tributed by the study of the conduction of electricity through liquids
and gases. Moreover, the discoveriesof Hertz 5 had shown that a
moleculewhich is emitting light mustcontain somesystemresembling
a Hertzian vibrator; and the essentialprocessin a Hertzian vibrator
is the oscillation of electricity to and fro. Lorentz himself from the
outset of his career 8 had supposedthe interaction of ponderable
1 Archives NlerI. xxv (1892), p. 363 : the theory is given in ch. iv, pp. 432
et sqq.
• ef. pp. 201, 206, 234
• Some writers have inclined to use the term' electron theory' as if it were specially
connected with Sir Joseph Thomson's justly celebrated discovery (cf. p. 364, supra) that
all negative electrons have equal charges. But Thomson's discovery, though undoubtedly
of the greatest importance as a guide to the structure of the universe, has hitherto exercised
but little influence on general electromagnetic theory. The reason for this is that in
theoretical investigations it is customary to denote the changes of electrons by symbols,
el' ea, .•• ; and the equality or non-equality of these makes no difference to the
equations. To take an illustration from celestial mechanics, it would clearly make no
difference in the general equations of the planetary theory if the massesof the planets
happened to be all equal.
• cf. ch. xi • ef. pp. 322-8
• Verll. d. Ak. v. Wetensehappm,Amsterdam, DeeI xviii, 1878
392
THE AGE OF LORENTZ
matter with the electric field to be effectedby the agencyof electric

chargesassociatedwith the material atoms.
The principal differenceby which the theory now advancedby
Lorentz is distinguishedfrom the theoriesof Weber, Riemann, and
Clausiusand from Lorentz' own earlier work, lies in the conception
which is entertainedof the propagationof influencefrom oneelectron
to another. In the older writings, the electronswere assumedto be
capableof acting on eachother at a distance,with forcesdepending
on their charges,mutual distancesand velocities; in the present
memoir, on the other hand, the electronswere supposedto interact
not directly with each other, but with the medium in which they
were embedded. To this medium were ascribed the properties
characteristicof the aetherin Maxwell's theory.
The only respect in which Lorentz' medium differed from
Maxwell's was in regard to the effects of the motion of bodies.
Impressedby the success of Fresnel'sbeautiful theory of the propaga-
tion oflight in moving transparentsubstances, 1 Lorentz designedhis
equationsso as to accord with that theory, and showedthat this

might be doneby drawing a distinction betweenmatter and aether,
and assumingthat a moving ponderablebody cannot communicate
its motion to the aether which surroundsit, or even to the aether
which is entangledin its own particles; sothat no part of the aether
can be in motion relative to any other part. Such an aether is
simply space endowed with certain dynamical properties; its
introduction was the most characteristicand most valuable feature
of Lorentz' theory, which differs completely from, for example,the
theory of Hertz so far as concernsthe electrodynamicsof bodies in
motion.
The general plan of Lorentz' investigation was, therefore,
to reduce all the complicated cases of electromagnetic action
to one simple and fundamental case,in which the field contains
only free aether with solitary electronsdispersedin it; the theory
which he adopted in this fundamental case was a combination
of Clausius' theory of electricity with Maxwell's theory of the
aether.
Supposethat e (x,y, z) and e'(x',y', z') are two electrons. In the
theory of Clausius,2the kinetic potential of their mutual action is
(in the electrostaticsystemof units)
ee' (~
r c2
ax ax' +! Oy?l + ~az az' _
at at c at at c at at
2 l
I) '.
I cr. pp. 108 et sqq. I cr. p. 234
393
CLASSICAL THEORY IN
SO when any number of electronsare present,the part of the kinetic

potential which concernsanyone of them-say, e-may be written
Le = e (!c ax atox + c!.ay atOy + ~c ax ~z

at
- tP)
where a and tP denotepotential functions, definedby the equations
a = HJp~;' dx'dy'dz', tP = HJ~ dx'dy'dz' ;
p denoting the volume-densityof electric chargein electrostaticunits,

and v its velocity, and the integration being taken over all space.
We shall now reject Clausius' assumption that electrons act
instantaneouslyat a distance,and replaceit by the assumptionthat
they act on each other only through the mediation of an aether
which fills all space,and satisfiesMaxwell's equations. This modi-
fication may be effectedin Clausius' theory without difficulty; for,
as we have seen,l if the state of Maxwell's aether at any point is
defined by the electric vector d and magneticvector h,2 thesevectors
may be expressedin terms of potentials a and tP by the equations
1 oa
d = grad ,J.. - - - h = curl a ;
't' cot'
and the functions a and tP may in turn be expressedin terms of the

electric chargesby the equations
a = Iff {(pvx)'lcr} dx'dy'dz', tP = ffI {(p) ,Ir} dx' dy' dz',
where the bars indicate that the values of (pvx)' and (p)' refer to
the instant (t - r/e).3 Comparing theseformulae with thosegiven
above for Clausius' potentials, we seethat the only change which
it is necessaryto make in Clausius' theory is that of retarding the
potentials in the way indicated by L. Lorenz.4
It wasshownby T. Levi-Civita 6 that when a and tP are defined
by theseintegrals they satisfythe equation
aax + oay + oaz + ~p._ 0

ox Oy oz cat - •
1 cr. pp. 255-6

• We shall usethe small lettersd and h in placeofE an6.H, when we are concerned
with Lorentz' fundamental case,in which the systemconsistssolelyof free aether and
isolatedelectrons.
• Theseintegralsfor a and q,- had beengivenin theyearbeforeLorentz' workof 1892,
by H. Poincare,ComptesRendus,cxiii (1891), p. SIS.
o cr. p. 268 • Nuovo Cimento (0), vi (1897), p. 93
394
THE AGE OF LORENTZ
It had further been shown so long ago as 1858by Riemann 1 that a

function rP defined by the above integral satisfiesthe equation
a:rP a:rP alrP 1 alrP .
axl + fryl + azl - Ci fiil- = - %p ,
but Riemann considered only electric forces, and therefore did not
give the equation
~a a~ a~
axl + fryl + azl -
1
Ci
a~
atl = - 4,
w
pv.
The electric and magnetic forces,thus defined in terms of the position

and motion of the charges,satisfy the Maxwellian equations I
div d = %p,
div h = 0,
] ah
curl d = ---
c at'
1 ad %
curl h = C at +-cPv•
The theory of Lorentz is basedon thesefour aethereal equations
of Maxwell, together with the equation which determines the
ponderomotive force on a charged particle; this, which we shall
now derive, is the contribution furnished by Clausius' theory.
The Lagrangian equations of motion of the electron e are
d(~)
ift a(~) - ax
aL
= 0,
and two similar equations,where L denotesthe total kinetic potential

due to all causes,electric and mechanical. The ponderomotive force
exerted on the electron by the electromagnetic field has for its
x-component
where
or
In a paperpublishedafter his death,Ann. d. Phys. (0), xi (1867),p. 237

1
Theseequations,with the terms in pv, were first proposedby FitzGerald at the
I
Southportmeetingof theBritishAssociationin 1883. cr. Heaviside,Elect. Papers,ii, p. 508

395
CLASSICAL THEORY IN
which, since
reducesto
e (_ aeP_1. aax) +~(}z (aas _ aax) +~;y (aay _ aax)
ax c at c at ax az c at ax iJy'
or
edx + -e,at(oy
-- hs - aat- zhy
)
'
so that the force in question is
ed +~
c
[v. bJ,
in agreementwith the formula obtained by Heaviside 1 in 1889for

the ponderomotive force on an electrified corpuscle of charge e
moving with velocity v in a field defined by the electric force d and
magnetic force b.2
The rate ofloss of energyby radiation from a chargee having an
acceleration r was discoveredby Larmor 3 in 1897: it is
~~
3 ,3 r·
(the charge being measuredas usual in electrostatic units). This
formula has beenvery widely applied in physicsand astrophysics.
In Lorentz' fundamental case,which hasbeenexamined,account
wastaken only of the ultimate constituentsof which the universewas
supposedto be composed,namely, corpusclesand the aether. We
must now seehow to build up from thesethe more complex systems
which are directly presentedto our experience.
The electromagnetic field in ponderable bodies, which to our
sensesappears in general to vary continuously, would present a
different aspectif we were able to discern molecular structure; we
should then perceive the individual electronsby which the field is
produced, and the rapid fluctuations of electric and magnetic force
between them. As it is, the values furnished by our instruments
represent averagestaken over volumes which, though they appear
small to us, are large compared with molecular dimensions.' We
1cr. p. 310
• The derivation of the complete set of equationsfor the electromagneticfield,
including this ponderomotiveequation,from the Principleof LeastAction, wasgivenby
K. Schwarzschild,Golt. Nach. (1903),p. 126.
• Phil. Mag. xliv (1897),p. 503
• Theseprincipleshad beenenunciated,and to someextentdeveloped,by J. Willard
Gibbs in 1882-3: Amer.Jour. Sci. xxiii, pp. 262, 460; xxv, p. 107j Gibbs'sScitnt~
Papers, ii, pp. 182,195,211.
396
THE AGE OF LORENTZ
shall denotean averagevalue of this kind by a bar placed over the

correspondingsymbol.
Lorentz supposedthat the phenomena of electrostatic charge
and of conduction currents are due to the presenceor motion of
simple electronssuch as have been consideredabove. The part of
p arisingfrom theseis the measurabledensityof electrostaticcharge;
this we'Shalldenoteby Pl' If w denotethe velocity of the ponderable
matter, and if the velocity v of the electronsbe written w + u, then
the quantity p;, sofar asit arisesfrom electronsof this type, may be
written PlW + Pli. The former of theseterms representsthe con-
vection current, and the latter the conduction current.
Considernext the phenomenaof dielectrics. Following Faraday,
Thomson and Mossotti,1 Lorentz supposed that each dielectric
moleculecontainscorpusclescharged vitreously and also corpuscles
charged resinously. These in the absenceof an external field are
so arranged as to neutralise each other's electric fields outside the
molecule. For simplicity we may supposethat in each molecule
only onecorpuscle,of chargee, is capableof being displacedfrom its
position; it follows from what has been assumedthat the other
corpusclesin the molecule exert the sameelectrostatic action as a
charge- e situated at the original position of this corpuscle. Thus
if e is displacedto an adjacentposition, the entire moleculebecomes
equivalentto an electric doublet, whosemoment is measuredby the
product of e and the displacement of e. The molecules in unit
volume,taken together,will in this way give rise to a (vector) electric
momentper unit volume, P, which may be comparedto the (vector)
intensity of magnetisationin Poisson'stheory of magnetism.2 As
in that theory we may replace the doublet distribution P of the
scalar quantity p by a volume distribution of P, determined by the
equation 3
p = - divP.
This representsthe part of P due to the dielectric molecules.
Moreover,the scalarquantity pWz hasalsoa doublet distribution,
to which the sametheorem may be applied; the averagevalue of
the part of pWz, due to dielectric molecules,is therefore determined
by the equation
PWz = - div (wzP) = - wz div P - (P . V) WZ,

or pw = - div P. w - (P • V) W.
1 cr. pp. 187, 188 I cr. p. 62-5
I We assumeall transitionsgradual, so as to avoid surfacedistributions.

397
CLASSICAL THEORY IN
We have now to find that part of pu

which is due to dielectric
molecules. For a single doublet of moment p we have, by
differentiation,
Iff pu dx dy dz = dp/dt,
where the integration is taken throughout the molecule; so that
Iff pu fix dy dz = (d/dt) (VP),
where the integration is taken throughout a volume V, which
enclosesa large number of molecules,but which is small compared
with measurablequantities; and this equation may be written
- 1 d
pu = V dt (VP).
Now, if aP refers to differentiation at a fixed point of space (as

at
opposedto a differentiation which accompaniesthe moving body),
we have
(d/dt) P = (a/at) P (w. V) P, +
and (d/dt) V = V div w ;
so that
-
pu =
ap
7ft + (w. V) P + dlv. w . P
= ~~ + curl [P . w] + div P • w + (P • V) w,
and therefore _ _ ap
pu + pw = at + curl [P . w].
This equation determines the part of pv which arises from the
dielectric molecules.
The general equations of the aether thus become, when the
averagingprocessis performed,
diva = 4"7rPl - 4"7r div P, div h = 0,
1 ali
curld=---
c at' {COnvectioncurrent + conductioncurrent}
- ad ap
+at+
4"7r
curlh=(l/c) at+7 cur1[P.w]
In order to assimilatetheseto the ordinary electromagneticequations,
we must evidently write
d = E, the electric force ;
E + 4"7r P = D, the electric induction ;
Ii = H, the magnetic vector.
398
THE AGE OF LORENTZ
The equationsthen become(writing p for Pl, as there is no longer

any needto usethe subscript),
1 aH
div D = 47T P, - curlE =--
c at'
divH = 0, curlH
417
= -
c
s,
where
S = conductioncurrent + convectioncurrent +~ a~ + curl [P •w].
The term .!.. aD
417 at in S evidently representsthe displacement-
current of Maxwell; and the term curl [P . w] will be recognised
as a modified form of the term curl [D. w], which was first intro-
duced into the equationsby Hertz.1 It will be rememberedthat
Hertz supposedthis term to representthe generationof a magnetic
forcewithin a dielectric which is in motion in an electric field; and
that Heaviside,2by adducing considerationsrelative to the energy,
showedthat the term ought to beregardedaspart of the total current,
and inferred from its existencethat a dielectric which movesin an
electric field is the seat of an electric current, which produces a
magnetic field in the surrounding space. The modification intro-
duced by Lorentz consistedin replacing D by 47TP in the vector
product; this implied that the moving dielectr~cdoesnot carry along
the aetherealdisplacement,which is representedby the term E in
D, but only carries along the chargeswhich existat oppositeendsof
the moleculesof the ponderabledielectric, and which are represented
by the term P. The part of the total current representedby the term
curl [P . w] is generallycalled the current of dielectric convection.
That a magneticfield is producedwhen an unchargeddielectric
is in motion at right anglesto the linesof force of a constantelectro-
static field had been shown experimentally in 1888 by Rontgen.s
His experiment consistedin rotating a dielectric disk between the
plates of a condenser; a magnetic field was produced, equivalent
to that which would be produced by the rotation of the ' fictitious
charges' on the two facesof the dielectric, i.e., chargeswhich bear
the samerelation to the dielectric polarisation that Poisson'sequi-
valentsurface-densityof magnetism' bearsto magneticpolarisation.
If U denotethe differenceof potential betweenthe oppositecoatings
of the condenser,and € the specificinductive capacityof the dielectric,
the surface-densityof electric chargeon the coatingsis proportional
1 cr. p.329 I cr. p. 330
I Ann. d. Phys. xxxv (1888),p. 264; xl (1890),p. 93 • cr. p. 64
399
CLASSICAL THEORY IN
to ± £U, and the fictitious chargeon the surfacesof the dielectric is

proportional to =f (£ - I) U. I t is evident from this that if a plane
condenseris chargedto a given differenceof potential, and is rotated
in its own plane, the magnetic field produced is proportional to £
if (as in Rowland's experiment 1) the coatings are rotated while the
dielectric remains at rest, but is in the opposite direction, and is
proportional to (£ - I) if (asin Rontgen'sexperiment) the dielectric
is rotated while the coatings remain at rest. If the coatings and
dielectric are rotated together, the magnetic action (being the sum
of these)should be independentof £-a conclusionwhich wasverified
later by Eichenwald.2
Hitherto we have taken no account of the possiblemagnetisation
of the ponderable body. This would modify the equations in the
usual manner,3 so that they finally take the form
div D = 47rp, (I)
div B = 0, (II)
47r
curl H =- S, (III)
c
laB
- curl E = -~, (IV)
c vt
where S denotes the total current formed of the displacement
current, the convection current, the conduction current and the
current of dielectric convection. Moreover, since
- I ad
we have
S = pv + 41T at'
div S = div + (1/41T) div (ad/at)
pv
= div pv + ap/ at,
which vanishesby virtue of the principle of conservationof electricity.

Thus div S = 0, (V)
or the total current is a circuital vector. Equations (I) to (V) are

the fundamental equationsof Lorentz' theory of electrons.
1 cr. p. 305
I Ann. d. Phys. xi (1903), p. 421 ; xiii (1904-), p. 919. Eichenwald performed other
experiments of a similar character, e.g. he observed the magnetic field due to the changes
of polarisation in a dielectric which was moved in a non-homogeneous electric field.
• It is possible to construct a purely electronic theory of magnetisation, a magnetic
molecule being supposed to contain electrons in orbital revolution. It then appears that
the vector which represents the average value of h is not H, but B.
400
THE AGE OF LORENTZ
We have now to consider the relation by which the polarisation

P of dielectricsis determined. If the dielectric is moving with velocity
w, the ponderomotive force on unit electric charge moving with it
is (asin all theories.)1 I
E' = E C [w. B]. + (1)
In order to connectP with E', it is necessaryto consider the motion

of the corpuscles. Let e denote the charge and m the mass of a
corpuscle, (g, 7], ') its displacementfrom its position of equilibrium,
k2 (g, 7], ') the restitutive force which retains it in the vicinity of this
point; then the equations of motion of the corpuscleare
m d2~
dt2 + k2C = eEz,'
!>
and similar equationsin '1] and t When the corpuscleis setin motion
by light offrequency n passingthrough the medium, the displacements
and forceswill be periodic functions of nt-say,
C _ A ,,/v-=-1 E' - E eM/V -1.
s- e , "'- 0
Substituting thesevaluesin the equationsof motion, we obtain

A (k2 - mn2) = eEo, and therefore ~ (kl - mn2) = eE~.
Thus, if N denote the number of polarisable molecules per unit
volume, the polarisation is determined by the equation
P = Ne (~, 7], ') = Ne2 E'/(k2 - mn2).
In the particular casein which the dielectric is at rest, this equation
g1Ves
But, as we have seen,2 D bears to E the ratio fL2, where fL denotes

the refractive index of the dielectric; and therefore the refractive
index is determined in terms of the frequency by the equation
fL' = 1 + 47Te Nj(kI 2 - mn2).
This formula is equivalent to that which Maxwell and Sellmeier 3
had derived from the elastic-solid theory. Though superficially
different, the derivations are alike in their essentialfeature, which is
the assumptionthat the moleculesof the dielectric contain systems
which possessfree periods of vibration, and which respond to the
oscillations of the incident light. The formula may be derived on
electromagneticprinciples without any explicit referenceto electrons;
1 cr. p. 329 I cr. p. 253 I cf. p. 264
401
CLASSICAL THEORY IN
all that is necessaryis to assumethat the dielectric polarisation has

a free period ofvibration.1
When the luminous vibrations are very slow, so that n is small,
,.,,2reducesto the dielectric constant € 2; sothat the theory of Lorentz
leadsto the expression € = 1 41TNe2/k2 +
for the specific inductive capacity in terms of the number and
circumstancesof the electrons.3
Returning now to the casein which the dielectric is supposedto
be in motion, the equation for the polarisation may be written
41TP = (,.,,2 - 1) E', whereE' is given by (1) ; (2)
from this equation, Fresnel'sformula for the velocity of light in a
moving dielectric may be deduced. For, let the axis of z be taken
parallel to the direction of motion of the dielectric, which is supposed
to be alsothe direction of propagation of the light; and, considering
a plane-polarised wave, take the axis of x parallel to the electric
vector, so that the magnetic vector must be parallel to the axis of y.
Then equation (III) above becomes
_ ally _ ! aD% + 417 ~ aP % •
oz - c at c az '
equation (IV) becomes(assumingB equal to ll, asis alwaysthe case
in optics), aE% 1 ally
- az =,8[.
The equation which definesthe electric induction gives
D% = E% + 4?TP% ;
1 A theory of dispersion, which, so far as its physical assumptions and results are
concerned, resembles that described above, was published in the same year (1892) by
Helmholtz (Berl. Ber. [1892], p. 1093; Ann. d. Phys. xlviii [1893], pp. 389, 723). In this, as
in Lorentz' theory, the incident light is supposed to excite sympathetic vibrations in the
electric doublets which exist in the molecules of transparent bodies. Helmholtz'
equations were, however, derived in a different way from those of Lorentz, being deduced
from the Principle of Least Action. The final result is, as in Lorentz' theory, represented
(when the effect of damping is neglected) by the Maxwell-Sellmeier formula. Helmholtz'
theory was developed further by Reiff (Ann. d. Phys. Iv [1895], p. 82).
In a theory of dispersion given by Planck (Bert. Ber. [1902], p. 470), the damping of the
oscillations is assumed to be due to the loss of energy by radiation: so that no new
constant is required in order to expressit.
Lorentz, in his lectures on the Theory of Electrons (Leipzig, 1909, p. 141), suggested
that the dissipative term in the equations of motion of dielectric electrons might be
ascribed to the destruction of the regular vibrations of the electrons within a molecule by
the collisions of the molecule with other molecules.
Some interesting references to the ideas of Hertz on the electromagnetic explanation
of dispersion will be found in a memoir by Drude (Ann. d. Phys. (6), i [1900], p. 437).
• cf. p. 254 • cf. p. 188
024
THE AGE OF LORENTZ
and equations(1) and (2) give
41rPs= (p.I - 1) (Es - !£ Hy).
c
Eliminating Ds, Ps and Hy, we have
c
1alEs
OZI =
02Es
ifiI +
(Ip. - 1) (0at + w ozo)IE •
s,
or, neglectingwl/el,
alEs _ [!:I '02Es + 2W(p.1 - 1) alEs 1
OZI - cl otl el at oz'
Substituting Es = en (t-z/V)yi, so that V denotes the velocity of
light in the moving dielectric with respectto the fixed aether,we have
cl = p.IVI - 2w (p.I - 1) V,
or (neglectingw 2 /e2)
p.1 -1
V=-+
p.
C
--w
p.1 ,
which is the formula of Fresnel.2 The hypothesisof Fresnel, that

a ponderable body in motion carries with it the excessof aether
which it contains as compared with spacefree from matter, is thus
seento be transformedin Lorentz' theory into the suppositionthat
the polarised moleculesof the dielectric, like so many small con-
densers,increasethe dielectric constant, and that it is (so to speak)
this augmentation of the dielectric constant which travels with the
moving matter. One evident objection to Fresnel'stheory, namely,
that it required the relative velocity of aether and matter to be
differentfor light of different colours,is thusremoved; for the theory
of Lorentz only requires that the dielectric constant should have
different valuesfor light of different colours,and of this a satisfactory
explanationis provided by the theory of dispersion.
The correctnessof Lorentz' hypothesis,as opposed to that of
Hertz (in which the whole of the contained aether was supposedto
be transported with the moving body), was afterwards confirmed
by various experiments. In 1901R. Blondlot 3 drove a current of
air through a magneticfield, at right anglesto the lines of magnetic
force. The air current was made to passbetween the faces of a
condenser,which were connectedby a wire, so asto be at the same
potential. An electromotiveforce E' would be produced in the air
1 This equationwasfirst given asa result of the theory of electronsby Lorentz in the
lastchapterof his memoir of 1892,Arch. Nltrl. xxv, p. 525. It wasalsogiven by Larmor,

Phil. Trans. cIxxxv (1894),p. 821.
I cf. p. 110 • Compus Rmdus, cxxxiii (1901),p. 778
(823) 403
CLASSICAL THEORY IN
by its motion in the magnetic field; and, according to the theory

of Hertz, this should produce an electricinduction D of amountEE'
(where E denotesthe specificinductive capacity of the air, which is
practically unity); so that, according to Hertz, the facesof the
condensershould becomecharged. According to Lorentz' theory,
on the other hand, the electric induction D is determined by the
equation
D =E +
(E - 1) E'
where E denotesthe electric force on a chargeat rest, which is zero

in the presentcase. Thus, accordingto Lorentz' theory, the charges
on the faceswould have only (E - 1)IE of the values which they
would havein Hertz' theory; that is, they would bepractically zero.
The result of Blondlot's experiment was in favour of the theory of
Lorentz.
An experiment of a similar character wasperformedin 1905by
H. A. Wilson.l In this, the spacebetween the inner and outer
coatings of a cylindrical condenserwas filled with the dielectric
ebonite. When the coatingsof such a condenserare maintained at
a definite differenceof potential, chargesare inducedon them; and
if the condenserbe rotated on its axisin a magneticfield whoselines
of force are parallel to the axis, thesechargeswill be altered, owing
to the additional polarisation which is produced in the dielectric
moleculesby their motion in the magnetic field. As before, the
value of the additional charge according to the theory of Lorentz
is (E - 1)IE times its value as calculated by the theory of Hertz.
The result of Wilson's experimentswas, like that of Blondlot's, in
favour of Lorentz.
The reconciliation of the electromagnetictheory with Fresnel's
law of the propagation of light in moving bodies was a distinct
advance. But the theory of the motionlessaetherwashamperedby
one difficulty: it was, in its original form, incompetent to explain
the negative result of the experiment of Michelson and Morley.2
The adjustment of theory to observation in this particular was
achievedby meansof a remarkable hypothesiswhich must now be
introduced.
In the issueof Nature for 16June 1892,3Lodge mentioned that
FitzGerald had communicated to him a ne-w.suggestionfor over-
coming the difficulty. This was, to supposethat the dimensionsof
material bodiesare slightly altered when they are in motion relative
< Phil. Trans. cciv (1905), p. 121 I cr. p. 390

• Nature, xlvi (1892), p. 165
404
THE AGE OF LORENTZ
to the aether. Five monthsafterwards,this hypothesisof FitzGerald's

was adopted by Lorentz, in a communication to the Amsterdam
Academy1; after which it won favour in a gradually widening circle,
until eventually it came to be generally taken as the basis of all
theoreticalinvestigationson the motion of ponderablebodiesthrough
the aether.
Let us first see how it explains Michelson's result. On the
suppositionthat the aether is motionless,one of the two portions
into which the original beam of light is divided should accomplish
its journey in a time lessthan the other by w21/c3, where w denotes
the velocity of the earth, c the velocity of light and I the length of
eacharm. This would be exactly compensatedif the arm which is
pointed in the direction of the terrestrial motion were shorter than
the other by an amount w21/2c2; as would be the caseif the linear
dimensionsof moving bodieswere alwayscontractedin the direction
of their motion in the ratio of (1 - w2/2c2) to unity. This is
FitzGerald'shypothesisof contraction. Sincefor the earth the ratio
w/c is only 30 km./sec.
300,000km./sec.'
the fraction W2/C2 is only one hundred-millionth.

Several further contributions to the theory of electrons in a
motionlessaetherwere madein a short treatise 2 which waspublished
by Lorentz in 1895. One of theserelated to the explanation of an
experimental result obtained some years previously by Th. des
Coudres,3 of Leipzig. Des Coudres had observed the mutual
inductanceof coils in different circumstancesof inclination of their
common axis to the direction of the earth's motion, but had been
unable to detect any effect dependingon the orientation. Lorentz
now showedthat this could be explained by considerationssimilar
to thosewhich Budde and FitzGerald' had advancedin a similar
case; a conductor carrying a constant electric current and moving
with the earth would exert a force on electric chargesat relative
rest in its vicinity, were it not that this force induceson the surface
of the conductor itself a compensatingelectrostatic charge, whose
action annuls the expectedeffect.
Another result first published in the Versuch 6 was that Fresnel's
1 Verslagend. Kon. Ak. van Wetenschappen, I, r892-3 (26 Nov. 1892),p. 74-
t Versucheiner Theone der electrischenund optischenErscheinungenin bewegtenKorpem, von
H. A. Lorentz; Leiden, E. J. Brill. It wasreprinted by Teubner, of Leipzig, in 1906.
a Ann. d. Phys. xxxviii (1889),p. 73 • ct. p. 235
• Versuch,p. 101: Lorentz, Theory of Electrons (1909),pp. 191,316
405
CLASSICAL THEORY IN
formula for the propagation of light in a moving medium requires

correction when the medium is dispersive: namely, the formula
for the velocity should read,
£. +
fL
( 1 _.! - fL~dfL
fLl dA
)w.
Subsequentexperimentsby Zeeman I verified the correctedformula
with an accuracythat leavesnothing to be desired.
The most satisfactory method of discussingthe influence of the
terrestrial motion on electrical phenomenais to transform the funda-
mental equationsof the aether and electronsto axesmoving with the
earth. Taking the axis of x parallel to the direction of the earth's
motion, and denoting the velocity of the earth by w, we write
x = Xl + wt, Y = Yh Z = Zh
sothat (xI'YI> ZI) denoteco-ordinatesreferredto axesmoving with the
earth. Lorentz completedthe changeof co-ordinatesby introducing
in place of the variable t a ' local time' tl> defined by the equation
t = tl wxl/el•+
It is also necessaryto introduce, in place of d and h, the electric
and magnetic forcesrelative to the moving axes: theseare I
dl = d + e![w. h]
hl = h + (lIe) [d. w] ;
and in place of the velocity v of an electron referred to the original
fixed axes,we must introduce its velocity VI relative to the moving
axes,which is given by the equation
Vl = V - w.
The fundamental equations of the aether and electrons,referred

to the original axes,are
1 ah
div d = %p, curl d = - - ~,
e CIt
1 ad %
divh = 0, curl h = - !;1
e CIt e
pV, +-
1
F=d+-[v.h],
e
where F denotesthe ponderomotive force on a particle carrying a
unit charge.
1 Proc. Arnst. Ac. xvii (1914), p. 445 : xvii 1915), p. 398
I As usual, we take electrostatic units for d and electromagnetic units for b; ef.
p.202
THE AGE OF LORENTZ
By direct transformationfrom the original to the new variables

it isfoundthat, whenquantitiesof order w2/e2 and wv/e2 areneglected,
theseequationstake the form
div 1 d1 = %p,
Since these have the same form as the original equations, it

follows that when terms depending on the square of the constant
of aberration are neglected,all electrical phenomenamay be ex-
pressedwith referenceto axesmoving with the earth by the same
equationsasif the axeswere at rest relative to the aether.
In the last chapter of the Versueh Lorentz dicussedthoseexperi-
mental resultswhich were as yet unexplained by the theory of the
motionlessaether. That the terrestrial motion exerts no influence
on the rotation of the plane of polarisation in quartz 1 might be
explainedby supposingthat two independenteffects,which are both
due to the earth's motion, cancel each other; but Lorentz left the
question undecided. Five years later Larmor 2 criticised this
investigation,and arrived at the conclusionthat there should be no
first order effect; but Lorentz 3 afterwardsmaintained his position
againstLarmor's criticism.
Although the physical conceptions of Lorentz had from the
beginning included that of atomic electric charges,the analytical
equationshad hitherto involved p, the volume-density of electric
charge; that is, they had been conformed to the hypothesisof a
continuousdistribution of electricity in space. It might hastily be
supposedthat in order to obtain an analytical theory of electrons,
nothing more would be required than to modify the formulae by
writing e (the chargeof an electron) in place of p dxdyd;;. That this
is not the casewas shown & a few yearsafter the publication of the
Versueh.
1 cf. p. 390 • Larmor, Aether and Matter, 1900
I Proc. Amsterdam Acad. (Englishedn.), iv (1902),p. 669
• A. Lienard,L' &lairage Ilect. xvi (1898),pp. 5, 53, 106; E. Wiechert,Arch. Nlerl. (2),
v (1900),p. 549. For the solution of many special problemsin which the Lienard
potentials are involved, d. G. A. Schott, Elutromagnetw Radiation (Cambridge Uni-
versityPress,1912).
CLASSICAL THEORY IN
Consider,for example,the formula for the scalarpotential at any

point in the aether.
where the bar indicatesthat the quantity underneathit is to haveits

retarded value.I
This integral, in which the integration is extended over all
elementsof space,must be transformed before the integration can
be taken to extend over moving elementsof charge. Let de' denote
the sum of the electric chargeswhich are accountedfor under the
heading of the volume-element dx' dy'dz' in the above integral.
This quantity de' is not identical with p'dx'dy'dz'. For, to take the
simplestcase,supposethat it is required to computethe value of the
potential-function for the origin at the time t, and that the chargeis
receding from the origin along the axis of x with velocity u. The
chargewhich is to be ascribedto any position x is the chargewhich
occupiesthat positionat the instant t - x/e ; sothat whenthe reckon-
ing is made according to intervals of space,it is necessaryto reckon
within a segment (x2 - Xl) not the electricity which at anyone
instant, occupies that segment, but the electricity which at the
instant (t - x2/e) occupiesa segment(X2 - X'l)' where X'l denotes
the point from which the electricity streamsto Xl in the interval
betweenthe instants(t - Xi/C) and (t - Xl/C). We haveevidently
Xl - X'l = U (XI - xl)/e, or XI - x\ = (XI - Xl) (1 + ufe).
For this casewe should thereforehave
de' = XI -
XI -
X'l
Xl
p' dx'dy'dz' = (1 +~)p' dx'dy'dz'.
C
In the generalcase,it is only necessaryto replaceu by the com-

ponent of velocity of the electric chargein the direction of the radius
vector from the point at which the potential is to be computed. This
componentmay be written v cos(v. r), wherer is measuredpositively
from the point in questionto the charge,and v denotesthe velocity
of the charge. Thus
ede' = {c + cos(v • r)} p'
1)
"
-dx' dy' dz',
and therefore I ee
q, (x,y, z, t) = }:;a rv cos v, r + -- (--)
where the summation is extendedover all the chargesin the field,
1 cr. p. 268
• Still usingelectrostaticunits for d and electromagneticunits for h
408
THE AGE OF LORENTZ
and the bars over the letters imply that the position of the charge
consideredis that which it occupied at the instant t - ric. In the
same way the vector-potential may be shown to have the value
given by
a:& ()x,y, Z, t = 1: -
CT
ev:&
T v cos V, T
+ -- r -)
and two similar equations.
Any electromagneticfield is thus expressedin terms of the four
functions 4>, a:&,ar, az (the scalar-potential and the three components
of the vector-potential), and theseare given by the above formulae
in termsof the positionsand velocitiesof the electronswhich generate
the field. It was however shown in 1904by E. T. Whittaker 1 that
only two functions are actually necessary(in place of the four),
namely, functions F and G defined by the equations
'f +~'- Z
F (x,y, z, t) = ! 1:e log .•
T-Z-Z
(_' )'
G (x,y, z, t) = - !t
. x' - x +
i (j' - y)
1:e log I' -x - t.(_'
y -y )
where the summation is taken over all the electronsin the field, and
where x'(t), y'(t), z'(t) denote the position of the electron at the
instant t, and x'(t) signifiesthe value of x' at an instant such that a
light-signal sent from the electron at this instant reachesthe point
(x,y, z) at the instant t, so that x' = x'(t - ric). Thus i',}', are z'
known functions of (x,y, z, t) when the motions of the electronsare
known, and 2
r = (x' - X)I + ()' - y)1 + (~'- Z)I.
The electric vector (d:&, dr, dz) and the magnetic vector (h:&, h", h.)
are then given by the formulae
02F 1 olG 1 olF olG
d:& = ox8Z + c (;)yot' h:& = C (;)yot - oxoz
olF 1 olG 1 olF olG
dy = (;)yoz--c oxot' h" = - C oxot -ayoZ
olF 1 olF olG olG
d. = a,ii -Ci otl' h. = oxl + qyl'
It will be noted that F and G are defined in terms of the positions
of the electronsalone, and do not explicitly involve their velocities.
Since in the above formulae for d and h an interchange of
electric and magneticquantities correspondsto a changeof G into F
1 p,.oc. Lond. Math. Soc.(2). i (1904), p. 367
409
CLASSICAL THEORY IN
and of F into G, it is clear that the two functions F and G exhibit

the duality which is characteristicof electromagnetictheory: thus
an electrostaticfield can be describedby F alone, and a magneto-
static field by G alone; again, if the field consistsof a plane
wave of light, then the functions F and G correspondrespectivelyto
two plane-polarised components into which it can be resolved.
Since there are an infinite number of ways of resolving a plane
wave of light into two plane-polarisedcomponents,it is natural to
expectthat, correspondingto any given electromagneticfield, there
should be an infinite number of pairs of functions F and G capable
of describingit, their differencefrom each other dependingon the
choiceof the axesof co-ordinates-as is in fact the case. Thus there
is a physical reasonwhy any particular pair of functions F and G
should be specially related to one co-ordinate, and cannot be
describedby formulae symmetricallyrelatedto the threeco-ordinates
(x,y, z).
From what has beensaid, it will be evident that, in the closing
years of the nineteenth century, electrical investigation was chiefly
concernedwith systemsin motion. The theory of electronswas,
however, applied with successin other directions, and notably to
the explanation of a new experimentaldiscovery.
The last recorded observationof Faraday 1 was an attempt to
detect changesin the period, or in the state of polarisation, of the
light emitted by a sodium flame, when the flame was placed in a
strong magneticfield. No result wasobtained; but the conviction
that an effect of this nature remained to be discoveredwas felt by
many of his successors.Tait 2 examinedthe influenceof a magnetic
field on the selectiveabsorption of light; impelled thereto, as he
explained,by theoreticalconsiderations. For from the phenomenon
of magneticrotation it may be inferred 3 that rayscircularly polarised
in opposite sensesare propagated with different velocities in the
magnetisedmedium; and thereforeif only thoserays are absorbed
which have a certain definite wave-lengthin the medium, the period
of the ray absorbedfrom a beam of circularly polarisedwhite light
will not be the samewhen the polarisation is right-handed aswhen
it is left-handed. 'Thus,' wrote Tait, 'what was originally a single
dark absorption line might becomea double line.'
The effect anticipated under different forms by Faraday and
Tait was discoveredtowards the end of 1896, by a young Dutch
physicist, who was at the time working in Prof. Kamerlingh Onnes'
1 BenceJones'sLife of Faraday, ii, p. 449 I Proc. R. S. Edin. ix (1875), p. 118
• cf. p. 159
410
THE AGE OF LORENTZ
laboratory at Leiden, P. Zeeman (1865-1943).1 Repeating Fara-

day's procedure,he placed a sodium flame betweenthe polesof an
electromagnet,and observeda widening of the D lines in the spec-
trum when the magnetisingcurrent wasapplied.
A theoretical explanation of the phenomenonwas immediately
furnished to Zeeman by Lorentz.2 The radiation was supposedto
be emitted by electrons which describe orbits within the sodium
atoms. If e denotethe charge (in electrostaticunits) of an electron
of massm, the ponderomotiveforce which actson it by virtue of the
external magneticfield is (elc) [~: •K J. where K denotesthe mag-
netic force in electromagneticunits and r denotesthe displacement
of the electronfrom its position of equilibrium; and therefore,if the
forcewhich restrainsthe electron in its orbit be K2r, the equation of
motion of the electronis
d1r e
m dtS +
K1r = C [r. K.]
The motion of the electron may (as is shown in treatises on
dynamics) be representedby the superpositionof certain particular
solutionscalled principal oscillations, whosedistinguishing property.is
that they areperiodic in the time. In order to determinethe principal
oscillations,we write roentv' - 1for r, where ro denotesa vector which
is independent of the time, and n denotes the frequency of the
principal oscillation; substituting in the equation, we have
(K1 - mnl) ro = en
c
Y=l [ro• K.]
This equation may be satisfiedeither (1) if ro is parallel to K, m

which caseit reducesto
K1 - mnl = 0,
so that n has the value Km - t, or (2) if ro is at right angles to K,
in which caseby squaring both sidesof the equation we obtain the
result el
(K1 - mnl)1 = -.nIK',
c
which givesfor n the approximate values Km- t ±eKj2mc.
1 Verslagender Akad. v. Wet. Ie Amsterdam, v (31 Oct. 1896),p. 181,(28 Nov. 1896),
p. 242; vi (29 May 1897),p. 13, (26June 1897)p. 99, (30 Oct. 1897)p. 260; Phil.
Mag. 10), xliii (1897),p. 226
I Phil. Mag. xliii (1897),p. 232; Ann. d. Phys. lxiii (1897),p. 278. It shouldbe
mentionedthat beforethe effectwasdiscoveredexperimentally,Larmor had examined

the questiontheoretically: the electronwas not yet discovered,and Larmor did not
lupposethat anythingsmallerthan an atom could radiate; sincethe effect depended
on the ratio of chargeto mass,Larmor foundfor it a valuenearly2,000timestoo small,
andconcludedthat it would be too smallto be oblerved.
411
CLASSICAL THEORY IN
When there is no external magnetic field, so that K is zero, the

three valuesof n which have beenobtained all reduceto Km-l, which
representsthe frequency of vibration of the emitted light beforethe
magnetic field is applied. When the field is applied, this single
frequencyis replacedby the threefrequenciesKm-l, Km-l + eKf2me,
Km-l - eKf2me; that is to say, the single line in the spectrum is
replaced by three lines close together. The apparatus used by
Zeeman in his earliest experimentswas not of sufficient power to
exhibit this triplication distinctly, and the effect was therefore
describedat first asa widening of the spectrallines.
We have seenabove that the principal oscillation of the electron
corresponding to the frequency Km-l is performed in a direction
parallel to the magnetic force K. I t will therefore give rise to
radiation resembling that of a Hertzian vibrator, and the electric
vector of the radiation will be parallel to the lines of force of the
external magnetic field. It follows that when the light receivedin
the spectroscopeis that which has been emitted in a direction at
right anglesto the magnetic field, this constituent (which is repre-
sentedby the middle line of the triplet in the spectrum)will appear
polarised in a plane at right anglesto the field; but when the light
received in the spectroscopeis that which has been emitted in the
direction of the magnetic force, this constituent will be absent.
We have also seenthat the principal oscillationsof the electron
correspondingto the frequenciesKm-l ± eKf2me are performed in a
plane at right anglesto the magneticfield K. In order to determine
the nature of these two principal oscillations,we observethat it is
possiblefor the electron to describea circular orbit in this plane, if
the radius of the orbit be suitably chosen; for in a circular motion the
forcesK2r and (efe) [~ . K ] would be directed towardsthe centreof
the circle; and it would therefore be necessaryonly to adjust the
radius so that thesefurnish the exact amount of centripetal forcere-
quired. Such a motion, being periodic, would be a principal oscilla-
tion. Moreover,sincethe force (efe) [~ . K ] changessignwhen the
senseof the movementin the circle is reversed,it is evident that there
are two suchcircular orbits, correspondingto the two senses in which
the electron may circulate; thesemust, therefore,be no other than
the two principal oscillationsof frequenciesKm-l ± eKf2mc. Whenthe
light received in the spectroscopeis that which has been emitted
in a direction at right anglesto the external magneticfield, the circles
are seenedgewise,and the light appearspolarisedin a plane parallel
412
THE AGE OF LORENTZ
to the field; but when the light examined is that which has been
emitted in a direction parallel to the external magnetic force, the
radiations of frequencies Km-l ± eK/2me are seen to be circularly
polarisedin oppositesenses. All thesetheoretical conclusionshave
beenverified by observation.
It wasfound by Cornu 1 and by C. G. W. Konig I that the more
refrangible component (i.e., the one whose period is shorter than
that of the original radiation) has its circular vibration in the same
senseas the current in the electromagnet. From this it may be
inferred that the vibration must be due to a resinously charged
electron; for let the magnetisingcurrent and the electron be sup-
posedto circulate round the axisof z in the direction in which a right-
handed screw must turn in order to progress along the positive
direction of the axis of z; then the magnetic force is directed
positively along the axis of z, and, in order that the force on the
electron may be directed inward to the axis of z (so as to shorten
the period), the chargeon the electron must be negative.
The value of elm for this negative electron may be determined
by measurementof the separation between the componentsof the
triplet in a magneticfield of known strength; for, as we have seen,
the differenceof the frequenciesof the outer componentsis eKlme.
The valuesof elm thus determinedagreedwell with the estimations3
of elm for the corpusclesof cathoderays.
Towards the end of 1897the concordanceof theory and observa-
tion for the Zeemaneffectseemedto beperfect,the number of verified
predictions, both qualitative and quantitative, being very great.
Yet only fourteen months after Zeeman'sfirst publication, a new
experimentaldiscoverywas made which led ultimately to the aban-
donment of the Lorentz explanation.
The author was Thomas Preston, who on 22 December 1897
read a paper to the Royal Dublin Society,"in which he announced
that the spectrallines >..4722 of zinc and >..4800 of cadmium became
quadruplets in a strong magnetic field.1i In the following year
A. Cornu II obtained a quadruplet (obtained by a splitting of the
central compoment of the triplet) as the Zeeman pattern for the
sodium line D1, and a sextuplet (obtained by a splitting of all three
componentsof the triplet) for Dli• Similar resultsweresoonobtained
by many other workers: in 1900C. Runge and F. Paschen7 found
1 ComptesRnu!us, cxxv (1897),p. 555 • Ann. d. Phys. lxii (1897), p. 240
• a. p. 363 • Trans. Roy. Dublin Soc. vi (1898),p. 385
I Later he found that the quadrupletscan be resolvedinto sextupletsby the splitting
of the sidelinesinto doublets.

• Compus rendus, cxxvi (1898),p. 181 ' Phys. ~s. i (4 Aug. 1900),p. 480
413
CLASSICAL THEORY IN
that the green mercury line .\5461 is split into 11 components,and

the blue mercury line ,\4359 into 8 components. Each component
of the Zeeman triplet, in fact, consistsof groups formed of several
sub-components. Spectral lines of the samesubstancemay be, and
generally are, very differently affected; but 1 lines which belong
to the sameserieshave the sametype of resolution; and correspond-
ing lines in the spectraof different elementsof the samefamily (e.g.
the alkalis or alkaline earths) are resolvedin the sameway. It will
appearlater that the correct explanationof thesephenomenadepends
on considerationswhich belong to quantum theory.2
The phenomenon discovered by Zeeman is closely related to
the Faraday magnetic rotation of the plane of polarisationof light. 3
The connection was first indicated by FitzGerald in a discussionat
a meeting of the Royal Societyon 20January 1898,at which a paper
on the Zeeman effect was read by T. Preston. In the Faraday
effect, a beam of polarised light is transmitted through a medium,
which is subjectedto a magnetic field, parallel to the lines of force
of the field, and it is found that the plane of polarisation is rotated.
The beam of plane-polarised light may be regarded as the super-
position of two beams which are circularly polarised in opposite
senses; and the rotation of the plane of polarisationis then expressed
by the statement that the two circularly polarised beams have
different velocitiesof propagation in the medium. Now the Zeeman
effect showsthat if we considera radiation which, in the absenceof
a magnetic field, is representedby a single spectral line, and if a
magnetic field is set up and the radiation is observedin a direction
parallel to the field, then the spectral line is replaced by two
lines, representing radiations circularly polarised in opposite
senses,of frequencies respectively a little greater and a little
less than the original line. The relation discovered by Balfour
Stewart and Kirchhoff· then leads us to infer that what is true of
emissionis true alsoof absorption,and that thereforeif, for example,
a sodium flame is placed in a strong magnetic field, the radiations
it will absorbwill not be of the same frequencyas in the absence
of the field, but will be rays circularly polarised in opposite
senses,of frequences respectively below and above the original
frequency. This may be called the inverse Zeeman effect; but
accordingto all theoriesof dispersion,the velocity of propagation of
a ray of light in a transparent medium is controlled in part by the
1 T. Preston,Proc. R. S. !xiii (1898),p. 26
• Lorentz attempted to explain them by the classicalelectron theory in Ph)s. -<s. i
(1899),p. 39
• cr. p. 190 • cr. p. 371
THE AGE OF LORENTZ
position of the absorption line of the medium.1 We therefore con-

clude that for circularly polarised rays of light propagated in a
transparent medium parallel to the lines of magnetic force, the
velocity of propagationwill be slightly different for rayswith the two
oppositesensesof circular polarisation, sincethey will be controlled
by different absorption lines; and this, as we have just seen, is
preciselythe Faraday effect.
Later in the same year D. A. Goldhammer I (1860-1922) of
Kasan and Woldemar Voigt 3 (1850-1919) of Gottingen carried
FitzGerald's idea further, by consideringwhat happenswhen, for
example,a sodiumflame placedin a magneticfield is traversedby a
ray of plane-polarisedlight, of frequencynearly the sameas that of
the D radiation, and propagated in a direction perpendicular to the
linesof magneticforce; they showedthat in this casethe propagation
would dependon the orientation of the plane of polarisation, the ray
being doubly refracted,the direction of the lines of force playing the
part of optical axis of a uniaxal crystal. This prediction was con-
firmed experimentallyby Voigt and Wiechert.f, Voigt gavea mathe-
matical treatment covering the combined theories of the Faraday
and inverse Zeeman effects, and not depending on the theory of
electrons. He remarked that the magnetic rotation should be very
largefor light whoseperiod is nearly the sameas that corresponding
to an absorption line; a large rotation is in fact observed when
plane-polarised light, whose frequency differs but little from the
frequenciesof the D lines, is passedthrough sodium vapour in a
direction parallel to the lines of magneticforce.Ii
A different way of connecting the Faraday and Zeeman effects
by generalreasoningwasgiven by Larmor in 1899.e This depended
on a theoremwhich he had proved in 1897 7 to the effect that when
in a material moleculethere existsan independentlyvibrating group
of ionsor electrons,for all of which the ratio elm of electric chargeto
inertia is the same,then the influence of a magnetic field H on the
1 cr. p. 265
• Paperread at Kiev, 10Sept. 1898,printed Ann. d. Phys. Ixvii (1899),p. 696
• Giitt. Nach. (26 Nov. 1898),p. 355; Ann. d. Phys. !xvii (1899),p. 345; Phys. Zs. i
(1899),p. 138
• Goldhammerhad conjecturedthe existenceof such a phenomenonas far back as
1887,but no experimentalconfirmation was at that time obtained. cf. alsoA. Cotton,
ComptesRendus,cxxviii (1899),p. 294; H. Becquerel,ComptesRendus,cxxviii (1899),p. 145;
J. Geest,Arch. Nlerl. (2), x (1905),p. 291
• The phenomenonwasfirst observedby D. Macalusoand O. M. Corbino, Comptes
Rendus, cxxvii (1898),p. 548; Nuovo Cimento (tl, viii (1898),p. 257; ix (1899), p. 381 ;
Rend. Linen (6), vii (2) (1898), p. 293. cr. P. Zeeman, Proc. Amsterdam Acad. v (1902),
p.41 ; J.J. Hallo, Arch. Nlerl. 1'1, x (1905),p. 148
• Proc. Camb. Phil. Soc. x (6 Mar. 1899),p. 181
, Phil. Mag. (6), xliv (Dec. 1897),p. 503
415
CLASSICAL THEORY IN
motionsof the group is preciselythe sameas that of a rotation with

angular velocity w, equal to eHf2me (in electrostaticunits for e and
electromagneticunits for H), imposedon the group around the axis
of the field, on the hypothesisthat the extraneousforcesacting on the
ions are symmetricalwith respectto this axis. (This is known as
the Larmor precession.) Now, n being the angular velocity of the
displacementvector in a train of circularly polarisedwavestraversing
the medium along the axis, the stateof synchronousvibration which
it excites in the molecules will have exactly the same formal
relation to this train when the magneticfield is off, asit would have
to a train with the very slightly different angular velocity n ± w
when the magneticfield is on, the sign being different, accordingas
the train is right-handed or left-handed. But changeof this angular
velocity n meanschangeof period of the light; thusthe propagation
of a circularly polarisedwave-train, when the field is on, is identical
with that of the samewave-train when the period is altered by its
being carried round with angular velocity ± wand there is no
influencing magneticfield. This is the Faraday effect.!
Yet another way of formulating the connection between the
Zeemanand Faraday effectsis to say that the moleculesof material
bodies contain electric systemswhich possessnatural periods of
vibration, the simplestexample of such a systembeing an electron
which is attracted to a fixed centre. Zeeman'seffect representsthe
influenceof an external magneticfield on thefree oscillationsof these
electric systems,while Faraday's effect representsthe influence of
the external magneticfield on theforeed oscillationswhich the systems
perform under the stimulus of incident light.
The optical propertiesof metalsmay be explained, accordingto
the theory of electrons,by a slight extensionof the analysiswhich
applies to the propagation of light in transparent substances. It is,
in fact, only necessaryto supposethat someof the electronsin metals
are free instead of being bound to the molecules: a supposition
which may be embodiedin the equationsby assumingthat an electric
force E givesrise to a polarisation P, where
the term in a representsthe effect of the inertia of the electrons;

the term in fJ representstheir ohmic drift; and the term in 'Y repre-
sents the effect of the restitutive forces where these exist. This
1 Seealso H. Becquerel, CamptesRendus,cxxv (1897), p. 679
416
THE AGE OF LORENTZ
equation is to be combined with the customary electromagnetic

equations
I aE
curl H = -c -at -c -aP
+ 4-rr at'
18H
- curl E = - -.
c at
In discussingthe propagation of light through the metal, we may
for convenience suppose that the beam is plane-polarised and
propagatedparallel to the axisof z, the electric vector being parallel
to the axis of x. Thus the equationsof motion reduce to
a Ex
2 I 2Ex a 47T a Px
2
-aZI' = CI W + &I' W '
{ a l Px ilPx
Ex = aifP + f3 7ft + yPx.
For Ex and Px we may substitute exponential functions of
nv=1 (t - ZI-"/c),
where n denotesthe frequency of the light, and I-" the quasi-index of
refraction of the metal: the equations then give at once
(I-"I - I) (- ani + f3nv - I + y) = 4-rr.
Writing v (I - K v-I) for 1-", so that v is inversely proportional
1;0the velocity of light in the medium, and K denotesthe coefficient
of absorption, and equating separatelythe real and imaginary parts
of the equation, we obtain
4-rr (y - an2)
+ f3 n + (y _
2 2 _ •
v (1 -- K ) - I 2 l an2)1 ,
{
27Tf3n
+ (y
I
v K = f3lnl _ an2)1.
When the wave-length of the light is very large, the inertia
representedby the constant a has but little influence, and the
equations reduce to those of Maxwell's original theory 1 of the
propagation of light in metals. The formulae were experimentally
confirmedfor this caseby the researchesofE. Hagen and H. Rubens 2
with infra-red light; a relation being thus establishedbetween the
ohmic conductivity of a metal and its optical properties with respect
to light of great wave-length.
1 cr.
p. 260
• Berlin SilzungsbeT. (1903), pp. 269, 410; Ann. d. Phys. xi (1903), p. 873; Phil. Mag.
vii (1904), p. 157
CLASSICAL THEORY IN
When, however, the luminous vibrations are performed more

rapidly, the effect of the inertia becomespredominant; and if the
constantsof the metal are suchthat, for a certain rangeof valuesofn,
y2K is small, while y2 (1 - K2) is negative, it is evident that, for this
range of values of n, y will be small and K large, i.e. the properties
of the metal will approach those of ideal silver.l Finally, for in-
definitely great values of n, y2K is small and v2 (1 - K2) is nearly
unity, so that y tends to unity and K to zero: an approximation to
theseconditions is realisedin the X-rays. 2
In the last years of the nineteenth century, attempts were made
to form more definite conceptionsregarding the behaviour of electrons
within metals. It will be remembered that the original theory of
electronshad beenproposedby Weber 3 for the purposeof explaining
the phenomenaof electric currents in metallic wires. Weber, how-
ever, made but little progresstowards an electric theory of metals;
for being concerned chiefly with magneto-electric induction and
electromagneticponderomotive force, he scarcelybrought the metal
into the discussionat all, except in the assumption that electrons
with chargesof oppositesignstravel with equal and oppositeveloci-
ties relative to its substance. The more comprehensiveschemeof
his successorshalf a century afterwards aimed at connecting in a
unified theory all the known electrical properties of metals, such as
the conduction of currents according to Ohm's law, the thermo-
electric effects of Seebeck,Peltier and W. Thomson, the gl.lvano-
magnetic effect of Hall, and other phenomena which will be
mentioned subsequently.
The later investigators, indeed, ranged beyond the group of
purely electrical properties, and sought by aid of the theory of
electronsto explain the conduction of heat. The principal ground
on which this extension was justified was an experimental result
obtained in 1853 by G. Wiedemann and R. Franz,' who found
that at any temperature the ratio of the thermal conductivity of a
body to its ohmic conductivity is approximately the same for all
metals, and that the value of this ratio is proportional to the absolute
temperature. In fact, the conductivity of a pure metal for heat is
almost independent of the temperature; while the electric con-
ductivity varies in inverse proportion to the absolute temperature,
1 cr. p. 162
I Modelsillustrating the selectivereflectionand absorptionof light by metallic bodies
and by gaseswerediscussedby H. Lamb, Mem. and Proc. MaTl£hesterLit. and Phil. Soc. xlii
(1898), p. 1; Proc. Lond. Math. Soc. xxxii (1900), p. 11; Trans. Camb. Phil. Soc. xviii
(1900),p. 348.
• cr. p. 201 • Ann. d. Phys. lxxxix (1853),p. 497
418
THE AGE OF LORENTZ
SO that a pure metal asit approaches the absolutezeroof temperature

tendsto assumethe characterof a perfect conductor. That the two
conductivitiesare closelyrelated was shown to be highly probable
by the experimentsof Tait, in which piecesof the samemetal were
found to exhibit variations in ohmic conductivity exactly parallel to
variations in their thermal conductivity.
The attempt to explain the electrical and thermal properties of
metalsby aid of the theory of electronsrestson the assumptionthat
conductionin metalsis more or lesssimilar to conduction in electro-
lytes; at any rate, that positiveand negativechargesdrift in opposite
directionsthrough the substanceof the conductorunder the influence
of an electric field. It was remarked in 1888by J. J. Thomson,!
who must be regardedas the founder of the modern theory, that the
differenceswhich are perceived between metallic and electrolytic
conduction may be referred to special features in the two cases,
which do not affect their general resemblance. In electrolytesthe
carriersare provided only by the salt, which is dispersedthroughout
a largeinert massof solvent; whereasin metalsit may be supposed
that every molecule is capable of furnishing carriers. Thomson,
therefore,proposedto regard the current in metals as a seriesof
intermittent discharges,causedby the rearrangement of the con-
stituentsof molecularsystems-a conceptionsimilar to that by which
Grothuss2 had pictured conduction in electrolytes. This view
would, ashe showed,lead to a generalexplanationof the connection
betweenthermal and electrical conductivities.
Most of the writers on metallic conductionin the latter part of the
nineteenth century preferred to take the hypothesisof Arrhenius 3
rather than that of Grothussas a pattern; and thereforesupposed
the intersticesbetweenthe moleculesof the metal to be at all times
swarmingwith electricchargesin rapid motion. In 1898E. Riecke'
effectedan important advanceby examiningthe consequences of the
assumptionthat the averagevelocity of this random motion of the
chargesis nearly proportional to the square root of the absolute
temperatureT. P. Drude 5 in 1900replacedthis by the moredefinite
assumptionthat the kinetic energyof eachmoving chargeis equal to
the averagekinetic energyof a moleculeof a perfect gasat the same
temperature,and may thereforebe expressedin the form qT, where
q denotesa universal constant.
a J.J. Thomson,Applualions of Dynamus /() Physics and Chnnistry (1888), p. 296. cr. also
Giese,Ann. d. Phys. xxxvii (1889),p. 576 • cr. p. 76 • cr. p. 344
• Coli. Nach. (1898), pp. 48, 137; Ann. d. Phys. !xvi (1898), pp. 353, 545, 1199;
ii (1900),p. 835
• Ann. d. Phys. (4) i (1900),p. 566; iii (1900),p. 369; vii (1902),p. 687
(828) 419
CLASSICAL THEORY IN
In the sameyear J. J. Thomson 1 remarked that it would accord

with the conclusionsdrawn from the study of ionisation in gasesto
supposethat the vitreous and resinouschargesplay different parts
in the processof conduction. The resinouschargesmaybe conceived
of as carried by simple negative corpusclesor electrons,such as
constitute the cathode rays; they may be supposedto move about
freely in the intersticesbetweenthe atomsof the metal. The vitreous
charges,on the other hand, may be regardedasmore or lessfixed in
attachment to the metallic atoms. According to this view the trans-
port of electricity is due almostentirely to the motion of the negative
charges.
An experiment which was performed at this time by Riecke 2
lent somesupport to Thomson'shypothesis. A cylinder of aluminium
was inserted between two cylinders of copper in a circuit, and a
current was passedfor such a time that the amount of copper
depositedin an electrolytic arrangement would have amounted to
over a kilogramme. The weight of each of the three cylinders,
however, showedno measurablechange; from which it appeared
unlikely that metallic conduction is accompaniedby the transport of
metallic ions.
Many years later, this view was still more convincingly sub-
stantiated when R. C. Tolman and T. D. Stewart 3 gave a fairly
direct experimentalproof that electric currentsin metalsare carried
by , free' negative electrons. A coil of wire was rotated at a high
speedand suddenlybrought to rest: a pulse of electric current was
produced,arising from the momentumof the electronsin the rotating
coil, which causedthem to continue in motion after the rotation of
the coil had ceased; and a quantitative discussionyielded a general
confirmation of Thomson'shypothesis.
The ideas of Thomson, Riecke and Drude were combined by
Lorentz 4. in an investigation which, as it is the most complete,will
here be given asthe representativeof all of them. More than twenty
yearslater, it wasreplacedby a theory of metalsdependingon a new
type of statistics; but at the time it representeda great advanceon
everything that had precededit.
It is supposedthat the atoms of the metal are fixed, and that in
the interstices between them a large number of resinouselectrons
are in rapid motion. The mutual collisionsof the electronsare dis-
regarded, so that their collisions with the fixed atoms alone come
1 Rapports prls. au Congresde Ph)'sique, iii (Paris, 1900), p. 138
• Phys. Zs. iii (1901), p. 639 • Phys. Rev. viii (1916), p. 97

• Amsterdam Prot. (English edn.), vii (1904-5), pp. 438, 585, 684
4 20
THE AGE OF LORENTZ
under consideration; theseare regarded as analogousto collisIons
betweenmoving and fixed elasticspheres.
The flow of heat and electricity in the metal is supposedto take
place in a direction parallel to the axis of x, so that the metal is in
the samecondition at all points of any plane perpendicular to this
direction; and the flow is supposedto be steady,sothat the state of
the systemis independentof the time.
Considera slab of thicknessdx and of unit area; and suppose
that the number of electrons in this slab whose x-componentsof
velocity lie between u and u +
du, whosey-components of velocity
lie between v and v + dv, and whosez-componentsof velocity lie
betweenwand w +
dw, is
f(u, v, w, x) dx du dv dw.
One of theseelectrons,supposingit to escapecbllision, will in

the interval of time dt travel from (x, y, z) to (x + udt, y + vdt,
z + wdt); and its x component of velocity will at the end of the
interval be increased by an amount eEdtlm, if m and e denote its
massand charge,and E denotesthe electric force. Supposethat the
number of electronslost to this group by collisionsin the interval dt
is a dx du dv dw dt, and that the number added to the group by
collisionsin the sameinterval is b dx du dv dw dt. Then we have
f(u, v, w, x) + (b - a) dt =f(u + eE dtlm, v, w, x + udt),

and therefore
b _ a = eE Of + u of.
m au ax
Now, the law of distribution of velocitieswhich Maxwell postu-
lated for the moleculesof a perfect gas at rest is expressedby the
equation
where N denotesthe number of moving corpusclesin unit volume,

, denotesthe resultant velocity of a corpuscle (so that = u2 + v2 ,2
+ w2), and a denotesa constantwhich specifiesthe averageintensity
of agitation, and consequentlythe temperature. It is assumedthat
the law of distribution of velocities among the electronsin a metal
isnearly of this form; but a term must be addedin order to represent
the general drifting of the electronsparallel to the axis of x. The
421
CLASSICAL THEORY IN
simplestassumptionthat can be made regarding this term is that it is

of the form .
u X a functiOn of r only;
we shall, therefore, write
-t -~
f= N7T a-3e ,.' + ux (r).
The value of x(r) may now be determined from the equation
h - a = ~~of
m ou
+ u ax
?1.
'
for on the left-hand side, the Maxwellian term
,.'
7T
-t a-'N e- eli
would give a zero result, since b is equal to a in Maxwell's system;

thus b - a must depend solely on the term ux(r) ; and an examina-
tion of the circumstancesof a collision, in the manner of the kinetic
theory of gases,shows that (b - a) must have the form - urx(r)/l,
where l denotesa constant which is closely related to the mean free
path of the electrons. In the terms on the right-hand side of the
equation, on the other hand, Maxwell's term givesa result different
from zero; and in comparison with this we may neglect the terms
which arise from ux(r). Thus we have
_ urX(r)
l
= (eE~OU + u~)~e
m ox l-
-~, 7T a8
or
and thus the law of distribution of velocities is determined.

The electric current i is determined by the equation
i = e HJzif(u, v, w) du dv dw,
where the integration is extended over all possible values of the

components of velocity of the electrons. The Maxwellian term in
f(u, v, w) furnishesno contribution to this integral, so we have
i =e JJJu 8
X (r) du dv dw.
422
THE AGE OF LORENTZ
When the integration is performed, this formula becomes
or
The coefficient of i in this equation must evidently represent the

ohmic specific resistanceof the metal; so if y denote the specific
conductivity, we have
4lel N
y = 31T!m~·
Let the equation be next applied to the caseof two metalsA and
B in contact at the sametemperatureT, forming an open circuit in
which there is no conduction of heat or electricity (so that i and
daJdx are zero). Integrating the equation
E _!!: al dN
- 2e N dx
acrossthejunction of the metals,we have
D·IscontmUlty
., f . I . . mall NB
0 potentla at Junctlon = 2C og N ;
A
or since !ma!l, which representsthe average kinetic energy of an

electron, is by Drude's assumption equal to qT, where q denotes
a universalconstant,we have
D·IscontmUlty
.. f . I . . 2 q TINog N~'
0 potentIa at JunctIOn= 3"C
B
This may be interpreted as the differenceof potential connected

with the Peltier 1 effect at the junction of two metals; the product
of the differenceof potential and the current measuresthe evolution
of heat at the junction. The Peltier discontinuity of potential is of
the order of a thousandth of a volt, and must be distinguishedfrom
Volta's contact difference of potential, which is generally much
larger, and which, as it presumably dependson the relation of the
metals to the medium in which they are immersed, is beyond the
scopeof the presentinvestigation.
d. p. 237
423
CLASSICAL THEORY IN
Returning to the general equations,we observethat the flux of

energy W is parallel to the axis of x, and is given by the equation
W = !mH~ur1"(u, u, w.) du dudw,
wherethe integration is again extendedover all possiblevaluesof the

componentsof velocity; performing the integration, we have
W = 2ml (2ea NE _ al dN _ 3Na' da) .

37T' m dx dx '
or, substituting for E from the equation already found,

mal. 4ml N Ida
W = -e- % - 37T~ a dx·
Consider now the case in which there is conduction of heat

without conduction of electricity. The flux of energy will in this
casebe given by the equation
dT
W = - K dx'
where K denotesthe thermal conductivity of the metal expressed

in suitable units; or
3ma da
W= -K.2q- dx·
If it be assumedthat the conduction of heat in metalsis effectedby
motion of the electrons,this expressionmay be comparedwith the
preceding; thus we have 8
K = 97T-UaqN ;
and comparing this with the formula already found for the electric
conductivity, we have
y
~=~T('l)1
9 e'
an equation which showsthat the ratio of the thermal to the electric
conductivity is of the form T X a constantwhich is the samefor all
metals. This result accordswith the law of Wiedemannand Franz.
Moreover, the value of q is known from the kinetic theory of
.gases; and the value of e had been determined by J. J. Thomson1
and his followers; substituting thesevaluesin the formula for K/Y,
a fair agreement is obtained with the values of K/Y determined
experimentally.
I ef. p. 365
424
THE AGE OF LORENTZ
It was remarked by J.J. Thomson that if, as is postulated in the

above theory, a metal contains a great number of free electrons
in temperature equilibrium with the atoms, the specific heat of the
metal must depend largely on the energy required in order to raise
the temperature of the electrons. Thomson considered that the
observedspecificheatsof metals are smaller than is compatible with
the theory, and was thus led to investigate 1 the consequencesof his
original hypothesis:II regarding the motion of the electrons, which
differsfrom the onejust describedin much the sameway asGrothuss'
theory of electrolysis differs from Arrhenius'. Each electron was
now supposedto be free only for a very short time, from the moment
when it is liberated by the dissociation of an atom to the moment
when it collides with, and is absorbed by, a different atom. The
atoms were conceived to be paired in doublets, one pole of each
doublet being negatively,and the other positively, electrified. Under
the influence of an external electric field the doublets orient them-
selvesparallel to the electric force, and the electrons which are
ejectedfrom their negative polesgive rise to a current predominantly
in this direction. The electric conductivity of the metal may thus
be calculated. In order to comprise the conduction of heat in his
theory. Thomson assumedthat the kinetic energy with which an
electronleavesan atom is proportional to the absolutetemperature;
sothat if onepart of the metal is hotter than another, the temperature
will be equalisedby the interchange of corpuscles. This theory, like
the other, leads to a rational explanation of the law of Wiedemann
and Franz.
The theory of electronsin metals has received support from the
study of another phenomenon. It was known to the philosophers
of the eighteenth century a that the air near an incandescentmetal
acquires the power of conducting electricity. 'Let the end of a
poker,' wrote Canton,' ' when red-hot, be brought but for a moment
within three or four inchesof a small electrified body, and its electrical
power will be almost, if not entirely, destroyed.'
The subject continued to attract attention at intervals 5; and as
the processof conduction in gasescame to be better understood, the
1 J. J. Thomson,The Corpuscular Theory tif Malter, London, 1907 • cr. p.419
• Perhapsfirst to Du Fay in 1733 • Phil. Trans. Iii (1762),p. 457
• cf. E. Becquerel,Annales de Chimie, xxxix (1853),p. 355; Guthrie, Phil. Mag. xlvi
(1873),p. 254; alsovarious memoirsby Elster and GeiteI in the Ann. d. Phys. from
1882onwards,particularly Ann. d. Phys. xxxviii (1889),p. 315. The phenomenonis very
noticeable,as Edisonshowed(Engineering [12 Dec. 1884],p. 553), when a filament of
carbon is heatedto incandescence in a rarefied gas. It was later found that ions are
emittedwhen magnesia,or any of the oxidesof the alkaline earth metals,is heatedto a
dull red heat.
CLASSICAL THEORY IN
conductivity produced in the neighbourhoodof incandescentmetals

was attributed to the emissionof electrically charged particles by
the metals. But it wasnot until the developmentof J.J. Thomson's
theory of ionisation in gasesthat notable advanceswere made. In
1899,Thomson 1 determined the ratio of the charge to the massof
the resinouslycharged ions emitted by a hot filament of carbon in
rarefied hydrogen, by observingtheir deflectionin a magneticfield.
The value obtained for the ratio was nearly the sameas that which
he had found for the corpusclesof cathoderays; whencehe concluded
that the negative ions emitted by the hot carbon were negative
electrons.
The correspondinginvestigation 2 for the positive leak from hot
bodies yielded the information that the massof the positive ions is
of the sameorder of magnitudeasthe massof material atoms. There
are reasonsfor believing that theseions are producedfrom gaswhich
hasbeen absorbedby the superficial layer of the metal.3
If, when a hot metal is emitting ions in a rarefied gas,an electro-
motive force be establishedbetween the metal and a neighbouring
electrode,either the positive or the negativeions are urged towards
the electrodeby the electric field, and a current is thus transmitted
through the intervening space. When the metal is at a higher
potential than the electrode,the current is carried by the vitreously
chargedions; when the electrodeis at the higher potential, by those
with resinouscharges. In either case,it is found that when the
electromotive force is increasedindefinitely, the current does not
increase indefinitely likewise, but acquires a certain 'saturation'
value. The obvious explanation of this is that the supply of ions
available for carrying the current is limited.
When the temperatureof the metal is high, the ions emitted are
mainly negative; and it is found & that in thesecircumstances,when
the surrounding gas is rarefied, the saturation current is almost
independentof the nature of the gasor of its pressure. The leak of
resinouselectricity from a metallic surface in a rarefied gas must
therefore depend only on the temperature and on the nature of the
metal; and it wasshownby O. W. Richardson6 that the dependence
1 Phil. Mag. xlviii (1899),p. 547
I J . .1. Thomson,Proc. Camb. Phil. Soc. xv (1909),p. 64; O. W. Richardson,Phil. Mag.
xvi (1908),p. 740
I cf. Richardson,Phil. Trans. ccvii (1906),p. I
• cr.J. A. McClelland, Proc. Camb. Phil. Soc. x (1899),p. 241; xi (1901),p. 296. On
the resultsobtainedwhen the gasis hydrogen,cr. H. A. Wilson,Phil. Trans. ccii (1903),
p. 243; ccviii (1908),p. 247; and O. W. Richardson,Phil. Trans. ccvii (1906),p. I
• Proc. Comb. Phil. Soc. xi (1902),p. 286; Phil. Trans. cci (1903),p. 497. cr. also
H. A. Wilson,Phil. Trans. ceii (1903),p. 243
426
THE AGE OF LORENTZ
on the temperature might be expressedby an equation of the form

i = Are -T, "
wherei denotesthe saturation current per unit areaof surface(which
is proportional to the number of ions emitted in unit time), T denotes
the absolute temperature and A and bare constants.l From 1902
to 1912this equation was acceptedas the basisof the theory of the
subject,for which the name thermionics, suggestedby Richardson,is
now in generaluse.
In order to account for thesephenomena,Richardson 2 adopted
the hypothesiswhich had previously been proposed8 for the ex-
planation of metallic conductivity; namely, that a metal is to be
regarded as a sponge-like structure of comparatively large fixed
positive ions and molecules, in the interstices of which negative
electronsare in rapid motion. He assumedthat the free electrons
have the sameenergy as is attributed in the kinetic theory of gases
to gasmoleculesat the sametemperatureasthe metal (an assumption
which, as will be explained in the secondvolume of this work, was
subsequentlyabandoned). Since the electrons do not all escape
freely at the surface, he postulated a superficial discontinuity of
potential, sufficient to restrain most of them. Thus, let N denote
the number of free electronsin unit volume of the metal; then in
a parallelepipedwhoseheight measuredat right anglesto the surface
is dx, and whosebaseis of unit area, the number of electronswhose
x-componentsof velocity are comprisedbetweenu and u du is +
m denoting the massof an electron, T the absolutetemperature, and

q the universal constantpreviously introduced.
Now, an electron whosex-component of velocity is u will arrive
at the interface within an interval dt of time, provided that at the
beginningof this interval it is within a distance u dt of the interface.
So the number of electrons whose x-components of velocity are
comprisedbetween u and u +
du which arrive at unit area of the
interfacein the interval dt is
-,,'
7T-ia.-1Ne asu du dt.
If the work which an electron must perform in order to escape
1 The samelaw appliesto the emission from other bodies,e.g.heatedalkaline earths,
and to the emissionof positiveions-at any rate when a steadystateof emissionhasbeen
reachedin a gaswhich is at a definite pressure.
• Phil. TrQlU. cci (1903), p. 497 • d. p. ·U9 et sqq.
427
, AETHER AND ELECTRICITY
through the surface layer be denoted by cP, the number of electrons

emitted by unit area of metal in unit time is therefore
u' - 24>
ft",,,'=4>
cc
1r-ta-1Ne-;Zu du, or -}r-\N ae ma'
The current issuing from unit area of the hot metal is thus
_ 3</>_
or NE. (qT/31rm)te 2qT,
where E denotes the charge on an electron. This expression, being

of the form
agrees with the equation which Richardson had derived from the
experimental measures; and the comparison furnishes the value of
the superficial discontinuity of potential which is implied in the
existence of cP.l
The formula of 1902 for the connection between the saturation
current and the temperature was subsequently abandoned; but an
account of this development must be postponed to the secondvolume
of the present work.
1 This discontinuity of potential wasfound to be 2·45 volts for sodium,4·1 volts for
platinum and 6·1 volts for carbon.

A History of The Theories of Aether & Electricity 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A History of The Theories of Aether & Electricity 1

Uploaded by

Copyright:

Available Formats

A History of the Theories of

SIR EDMUND WHITTAKER

HARPER TORCHBOOKS / The ScienceLibrary

HARPER f5 BROTHERS, NEW YORK

C0?7i..t· ,L-~ I , ~I'n~

Davy's Chemical Theory of the Pile 75

IV THE LUMINIFEROUS MEDIUM FROM BRADLEY TO FRESNEL

V THE AETHER AS AN ELASTIC SOLID

VII THE MATHEMATICAL ELECTRICIANS OF THE MIDDLE OF THE

The Researchesof Hittorf and Kohlrausch 336

XII CLASSICAL RADIATION-THEORY

XIII CLASSICAL THEORY IN THE AGE OF LORENTZ

THE earliest science,that of the ancient Greeks,had its origin in

• Opusc.lx (ed. 1571),in Boet. de Trinit., quo 5, art. 1

mentedby further experiments. They abandoned8t Thomast view

1 T. Harper, The Metaphysicof the &hool (London, 1879-84), i, p. 41

first attempted a general reconstruction of ideas regarding the

there should be an interrogation of Nature by meansof carefully

had little or no experienceof mechanismsthat were self-contained

in 1691),whichare printed in V. Cousin'sFragmentsphilosophiques,tomeii, p. 155.

in this science,when no principles but those of mechanicswere

The generalexplanationoflight on theseprincipleswasamplified

1 Diop/rique, Discou,s second

also that the component of velocity parallel to the interface 15

Up by the Cartesianslong after the death of their master. Fermat

1 Epist. xlii, written at Toulouse in August 1657to Monsieur de la Chambre;

of Copernicusand Galileo, who re-introduced the doctrine of the

1 The imprimatur of Viscount Brouncker,P.R.S., is dated 23 Nov. 1664.

• Boyle'sWorks (ed. 1772),i, p. 742 • Hooke'sPosthumousWorks, p. 186

(1613-63), a Jesuit, who had given to the phenomenonthe name

it necessarilyfollows, that all the parts of theseSpheresundulated

trajected from a very remote point of a luminous body through an

• To the Royal Society, 15 February 1671-72 • Micrographia, p. 64

an oblique and confus'dpulseoflight, whosestrongestpart precedes,

1 The astronomer John Flamsteed explained his objections to it in a letter to Collin.

of date 17 April 1672, printed in Rigaud's Correspondence

discovered,asweshallpresentlysee,by Huygens,and first interpreted

in prindple from Descartes'proof.

phosphorescentsubstancesdo not necessarilyemit the samekind of

each other, and was acceleratedin the contrary case. Roemer

one, since beamsof light proceedingin directions inclined to each

inclined to AB at the same angle as AC. This gives the law of

The law of refraction is establishedby similar reasoning,or the

The matter had arrived at this stagewhen it was taken up by

interspacesaround thesespheroids,and which servesto bind them

againstthe theory 3 which wasput forward a century later by Young

which the waveadvances.

The work of Newton was of such unparalleled brilliance that

In Germany, Leibnitz described the Newtonian formula as a

1 S. Clarke, Papm which passedbetweenMr Leibnit~ and Dr Clarke (1717),p. 265

Cotes's principle recommends that inter-phenomena should be

THE ancientswere acquainted with the curious properties possessed

A. Crichton Mitchell, Terrestrial Magnetism and AtmosphericElectricity, xxxvii (1932),105;

disposition became evident; they formed circles which girdled the

ledging the discoveriesof Gilbert, attempted to bring the subject

1 It should be said for Cabeo, however, that in his commentary on Aristotle's

contains various humours or kinds of moisture-phlegm, blood,

takenin its metaphoricalmeaningfor nicety of distinction.'

substanceof the body, he concluded that corpusclescan penetrate

Heat' consistsin ' a various, vehement, and intestine commotion

Somewhatlater it was found that the heating effectsof the rays

1 Scheele in 1777 supposed c~loric to be a compound of oxygen and phlogiston, and

We must now return to 'sGravesande'sbook. The phenomena

o;~+ ~+ ~ = _ %p,