The Tiger and The Shark Empirical Roots of Wave-Particle Dualism

The tiger and the shark
Empirical roots of wave-particle dualism

It is like a struggle between a tiger and a shark,
each is supreme in his own element,
but helpless in that of the other.
J. J. Thomson, 1925
Empirical roots of wave-particle dualism
Bruce R. Wheaton
with a foreword by Thomas S. Kuhn
CAMBRIDGE
UNIVERSITY PRESS
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo
Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521250986
Cambridge University Press 1983
This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 1983

First paperback edition 1991
Reprinted 1992
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data

Wheaton, Bruce R.
The tiger and the shark.
Bibliography: p.
Includes index.
1. Wave-particle duality - History. 2. Radiation - His-
tory. 3. X-rays - History. 4. Gamma rays - History.
I. Title.
QC476.W38W45 1983 539 82-22069
ISBN 978-0-521-25098-6 hardback

ISBN 978-0-521-35892-7 paperback
Transferred to digital printing 2008

For Ariana and Geoffrey;
may they transcend the
indeterminism of their elders
CONTENTS
Foreword by Thomas S. Kuhn page ix

Preface xv
Notes on sources xxi
1 Introduction i
Part I The introduction of temporal discontinuity, 1896-1905
2 The electromagnetic impulse hypothesis of x-rays 15
3 The analogy between y-rays and x-rays 49
Part II Ionization and the recognition of paradox, 1906-1910
4 Secondary rays: British attempts to retain mechanism 71
5 The appeal in Germany to the quantum theory 104
Part III Seeking an electrodynamic solution, 1907-1912
6 Localized energy in spreading impulses 135
7 Problems with visible light 168
Part IV Interference of x-rays and the corroboration of paradox,
1912-1922
8 Origins of x-ray spectroscopy 199
9 Quantum transformation experiments 233
Part V The conceptual origins of wave-particle dualism,
1921-1925
10 Synthesis of matter and light 263
Epilogue: The tiger and the shark 302
Bibliography 309
Index 347
FOREWORD
The first three decades of the twentieth century embraced two

great changes in physical theory: relativity, both general and spe-
cial, and quantum theory, both the old and the new. Among the
innovations that constituted these theories, none was more diffi-
cult to accept and assimilate than Einstein's suggestion that light
displays particulate properties, especially at high frequencies. To-
gether with the related recognition of the wavelike properties of
material particles, Einstein's light-particle theory has proved a
fundamental constituent of modern physics, perhaps the single
feature that most sharply distinguishes it from the generally New-
tonian physics of the preceding three hundred years. But for nearly
twenty years after Einstein's proposal in 1905, the concept of
light-particles was almost everywhere rejected. Even R. A. Milli-
kan, who in 1914-16 provided the first unequivocal evidence for
Einstein's surprising law of photoelectric emission, continued,
equally unequivocally, to disdain the light-particle hypothesis
from which that law had been derived. Only after 1923, when
Compton and Debye independently used the light-particle hy-
pothesis to explain the shift in frequency of scattered x-rays, did
more than a very few isolated physicists begin to take seriously the
idea that electromagnetic radiation often behaves like particles.
That, in outline, is the way that historians of physics have
recently been telling the story of the light-particle hypothesis, and
with respect to Einstein it remains very nearly the way the story
should be told. But, as Bruce R. Wheaton amply demonstrates in
the pages that follow, Einstein's was only one approach to con-
ceiving radiation as particulate. A second, far less well known, was
associated with observations on x-rays and y-rays, both discovered
during the decade before Einstein's hypothesis was enunciated and
neither unequivocally identified with light for another decade. By
x Foreword
1900, five years after their discovery, x-rays were almost every-
where assumed to be impulses, formed when fast-moving elec-
trons are decelerated at the target electrode of an x-ray tube and
thereafter propagated radially from the target through the electro-
magnetic ether. This in itself, as Dr. Wheaton shows, had a signifi-
cant and heretofore unappreciated effect on the early accommo-
dation of classical ideas of radiant energy with the new quantum
theory. That x-ray energy is transported by impulses required
physicists to take explicitly into account the number of pulses that
pass as well as their individual energies. This set the stage for a
reinterpretation of radiant intensity, one that increasingly di-
verged from the classical dependence of electromagnetic effect on
amplitude toward one in which the temporal impulse width, or
later the frequency, was the determining factor.
A similar impulse explanation was widely applied by 1905 to
y-rays as well, with acceleration of an electron ejected from an
atom replacing deceleration at the target of an x-ray tube. Sup-
ported by existing theories and suggested by a variety of experi-
ences, these explanations were nevertheless difficult to reconcile
with two recognized sets of observations on the ionization of gases
by the new rays. Those observations were in turn the source of a
continuing series of suggestions, often independent both of each
other and of Einstein, that x-rays and y-rays be viewed as particu-
late.
The first set of observations, readily made but not of great force
by themselves, gave rise to what Dr. Wheaton calls the paradox of
quantity. Passing through a containerfilledwith gas, an advancing
impulse should affect all gas atoms equally, but in practice it
ionizes only a minuscule portion of them. How can the pulse
ionize any atoms if it does not ionize all or most of them? A second
set of observations, far harder to establish but also far more
difficult to explain away, constitute Dr. Wheaton's paradox of
quality. An electron torn from an atom in the process of ionization
by x-rays carries with it energy of the same order of magnitude as
that of the electron decelerated in forming the original impulse. It
is as though, at the moment of ionization, all the energy in an
isotropically expanding x-ray pulse were suddenly concentrated at
the point on the pulse's spherical surface where ionization was to
take place. Experimental data from which these paradoxes might
have been identified were available in 1900, and more accumu-
Foreword xi
lated steadily thereafter. By 1908, at least two well-known experi-
mental physicists, William Henry Bragg and Johannes Stark, had
based corpuscular theories of radiation upon them. But such
evidence of corpuscular properties impressed few physicists until
two additional conditions were fulfilled. Less paradoxical explana-
tions first had to be set aside, for example a variety of hypotheses
that derived the energy of an ejected electron not from radiation
but from the interior of the atom, radiation serving only to "trig-
ger" emission. In addition, evidence for the paradoxes had to be
refined, primarily, Dr. Wheaton shows, by the detailed study of
x-ray absorption spectra and the development of techniques for
/?-ray spectroscopy.
With these new data in hand, it became possible to strike precise
balances between the energy lost at the target by the cathrode-ray
electron in an x-ray tube, the orbital energies of the interior
electrons of the target atom, and the energy hv (and thus the
frequency v) of the resulting x-ray. The last of those energies could,
in turn, be shown to be equal to the sum of the kinetic energy of an
electron ejected during ionization and its orbital energy prior to
ejection. In the laboratories where these techniques were devel-
oped and these energy balances observed, it was impossible not to
notice that, with respect to spatial localization and energetic rela-
tions, radiation behaved precisely like particles.
Those conditions were not, however, fulfilled before 1920, and
then in only a very few places. One of these few was the private
laboratory set up in his Paris home by the French nobleman,
Maurice de Broglie. By 1921 de Broglie's laboratory had mapped
the x-ray energy levels of many atoms, and during that year an
improved /?-ray spectrometer was used to study the velocities of the
photoelectrons ejected by x-rays of known frequency. Reporting
the results to the Third Solvay Congress in April of that year, de
Broglie insisted that the radiation "must be corpuscular, or, if it is
undulatory, its energy must be concentrated in points on the
surface of the wave" (p. 270). The following year, with additional
results in hand, de Broglie wrote that the phenomena "present
facts in such a way that their qualities are sometimes described in
terms of the wave theory, sometimes in terms of the [particulate]
emission theory" (p. 277).
Remarks of that sort, both from de Broglie and others, prepared
the way for a reevaluation of Einstein's work even before the
xii Foreword
particulate interpretation of the Compton effect in 1923. Further-
more, the first major contribution to that reevaluation was made
starting in 1922 by Maurice's younger brother, Louis de Broglie.
Louis had been trained by France's leading theoretical physicist,
Paul Langevin, a man who knew and greatly admired Einstein; he
had also worked closely with Maurice on the interpretation of
laboratory results. Louis de Broglie was strategically placed to
bring together the long-separate theoretical and experimental ap-
proaches to the concept of light-particles. In the process he took
major steps toward the emergence in 1926 of Schrodinger's wave
mechanics.
From Wheaton's perspective, the independence of Schro-
dinger's wave mechanics from the slightly earlier matrix me-
chanics of Heisenberg can be explained by the different emphasis
each investigator placed on the paradoxes of free radiation. To
Heisenberg, a product of Bohr's school, atomic structure held the
center of interest. To Schrodinger, as before for Einstein and de
Broglie, the nature of light and its behavior in traversing matter
were the most pressing problems. Although theoretical research on
both atoms and light utilized some of the same empirical data -
most notably from x-ray spectroscopy - Einstein and de Broglie
were virtually alone before 1923 in directing it toward a resolution
of the nature of radiation. The development of two independent
but ultimately equivalent forms of the new quantum mechanics
thus arose in part from the lack of interest in the nature of light
shown by most theoretical physicists in the decade before 1923.
The preceding remarks suggest only some main themes of Dr.
Wheaton's narrative. The following pages supply others together
with the rich circumstantial detail, the accounts of false starts and
local triumphs that make history of science a story about people.
The fascination of that detail, much of it here recorded for the first
time, is for me the greatest of the rewards that Wheaton's text
offers readers, but it also qualitatively alters a standard picture of
the development of quantum theory. By emphasizing experiments
and the people who perform them, it redresses the emphasis on
abstract theory characteristic of most previous historical literature
on quantum theory. Simultaneously, by emphasizing x- and y-ra-
diation, it redresses that literature's special emphasis on optical
spectra. These two changes may, in turn, permit a third. Dr.
Wheaton's is the first account of quantum theory known to me in
Foreword xiii
which Louis de Broglie appears less as a surprising intruder than as
a person with just the background required to play the role for
which he is known.
THOMAS S. KUHN
PREFACE
Leopards break into the temple and drink dry the

sacrificial pitchers; this occurs again and again until it
can be predicted, and it becomes part of the ceremony.
F. Kafka, 1935
The first years of this century witnessed the final rejection of

determinism in physical theory; there is no more compelling
example of this than the synthesis forged in the early 1920s be-
tween theories of matter and theories of light. The insight of Louis
de Broglie that led to the most complete formulation of wave-
particle dualism was the last act in a series of preliminary attempts
by physicists to resolve paradoxes that had arisen in theories of
radiation following the discovery of x-rays. Historians have not
directed sufficient attention either to radiation theory or to experi-
mental studies of recent physics. In the case at hand, significant
empirical data were recognized to challenge classical radiation
theory long before the theory was successfully modified to agree
with them. The gradual recognition, based on these experimental
results, that internal consistency is unattainable by electrome-
chanical interpretations of radiation forms the subject of this book.
This study grew out of concerns first raised while I was a student
of physics. The inadequacy of most textbook discussions of histori-
cal and epistemological issues led me back to the original papers
and then to the literature on history of science. I was fortunate
to be introduced to the latter by John Heilbron, whose critical
approach and demand for clarity in expression showed how insight
can be won in historical analysis of scientific thought. When I
undertook a study of the photoelectric effect - the liberation of
electrons from metals by ultraviolet light1 - I found that many of
the textbook accounts that had confused me were demonstrably
false when viewed historically.2
1
Wheaton, Photoelectric effect (1971). Abbreviations used in the footnotes and
the bibliography are identified in the notes on sources.
2
For example, see Eisberg, Fundamentals (1961), 76-81; Jammer, Conceptual
development (1966), 35.
xvi Preface
After an invaluable exposure to the modern descendant of the
European culture within which these issues developed, I was again
fortunate to have the opportunity to work on the more general
historical problem of wave-particle dualism with Thomas Kuhn
as guide. When I extended my original concerns to include high-
frequency radiations, I found that the experimental evidence
amassed by 1911 concerning x-rays played a far more significant
role in preparing physicists to accept dualistic theory than did
evidence regarding ordinary light. My doctoral dissertation
formed the second stage in the development of this book.3 In it I
clarify the extent to which the then standard impulse interpreta-
tion of x-rays led to an implicit tension of a kind destined to bring
reformulation of classical concepts regarding the distribution of
energy in all forms of radiation in the 1920s.
But the story does not end there, and in this book I bring the
discussion to its historical conclusion. The results of attempts to
force consistency on electromechanical interpretations of radia-
tion rebounded to affect also the theory of matter. While reconsid-
ering Einstein's lightquantum hypothesis in 1921, Louis de Broglie
tried to find a theory of light that would combine the macroscopi-
cally incompatible representations of wave and particle. He based
this work on the solid foundation provided by his elder brother's
experimental corroboration of Einstein's photoelectric law for
x-rays. The result in 1923 was a hypothetical synthesis of matter
and light: Each was to be considered to possess both particle and
wave attributes that make their presence felt to a greater or lesser
extent according to experimental conditions. The successful elabo-
ration of this remarkable hypothesis signaled the end of strictly
deterministic representations of both matter and of light.
The reader may wonder why the name of Albert Einstein does
not figure more heavily in this account of evolving understanding
that neither a wave nor a particle characterization of radiation is
alone sufficient. After all, Einstein introduced the lightquantum
hypothesis in 1905 and early on recognized that a synthetic theory
was needed. The reasons are straightforward. Einstein's revolu-
tionary hypothesis had almost no followers and very little influ-
ence before 1921, and the growth of its acceptance by others marks
the limit of our direct concerns. His lightquantum was not initially
intended to apply either to x-rays or to y-rays, but rather only to
3
Wheaton, Nature ofx- and gamma rays (1978).
Preface xvii
visible and ultraviolet light. Before 1911 it was not clear to most
physicists that x-rays and visible light are common species of
radiation. Consequently the experimental evidence that had accu-
mulated by that time that x-rays and y-rays do transfer energy
in individual units did not justify Einstein's lightquantum in the
eyes of most physicists. The difficulties of any corpuscular theory
in explaining interference properties of radiation were too formi-
dable. And for almost a decade following Niels Bohr's quantum
theory of the atom in 1913, most mathematical physicists showed
little interest in the paradoxes that complicated a consistent elec-
tromechanical interpretation of free radiation.
Thus, our story developed quite independently of the imagina-
tive and prescient statistical treatment of light of which Einstein
was the chief architect. Einstein's own developing realization of
the need for duality in radiation theory has been discussed histori-
cally by Martin Klein.4 Here we analyze other physicists' efforts to
wrestle with closely related issues, discuss their rejection of and
eventual accommodation to dualistic ideas in physical theory, and
trace the way in which this development came to alter ideas not
simply about radiation and matter but about humans' ability to
construct consistent models of physical phenomena based on their
experiences in the macroscopic world of human senses.
The present narrative evolved out of an extensive search in the
physics literature of the period, all significant results of which
appear in the notes. Documents in several manuscript collections,
some in archives and others privately held, were invaluable for
reconstructing contemporary insights and opinions, for support-
ing evidence, and for opening new areas for research. In particular,
I benefited from examination of the Archive for History of Quan-
tum Physics,5 the Bohr archive, the Cherwell papers, the Einstein
archive, Paul Langevin's papers, the Lorentz papers, holdings of
the Nobel archives, the archives of the Paris Academy of
Sciences, the Rutherford papers, the Schwarzschild papers, and
the archives of the Solvay Institute. Full identifications and loca-
tions of these and other collections are given in the notes on
sources. I am grateful to these and other repositories for the kind
access granted me, and to the holders of literary rights for permis-
sion to quote documents. I was greatly assisted in my research by
4
Klein, Natural philosopher, 3 (1964), 3-49.
5
Described in Kuhn, Heilbron, Forman, and Allen, Sources (1967).
xviii Preface
concurrent work on the Inventory of sources for history of twen-
tieth-century physics, being led to several important and formerly
untapped collections.6 The immense quantity of data the ISHTCP
has compiled and their as yet incomplete description encourage
further mining of their riches.
An undertaking of this magnitude is possible only with the help
of colleagues and of prior work in the field.71 am pleased here to
acknowledge debts of both sorts. I am grateful for the comments of
several anonymous reviewers of the typescript before publication.
I have already mentioned the importance of Martin Klein's work
on the early work of Einstein; other studies by him are mentioned
in the Notes. The quantum interpretation of the Compton effect in
1922 was an event of great significance; Roger Stuewer's detailed
review of the evolution of Compton's approach to it made it
unnecessary to recount that episode in depth here.8 Although
Stuewer's and my interpretations differ on the importance of the
electromagnetic impulse hypothesis of x-rays and about the renais-
sance of interest in the lightquantum, his study is a significant
addition to the meager literature on the development of experi-
mental physics in this century.9
I have been fortunate, professionally and personally, in my own
introduction to history of science to have had as guides both
Thomas Kuhn and John Heilbron. I have benefited greatly from
discussions with both on many aspects of physical theory. Profes-
sor Kuhn's reassessment of the conceptual origins of quantum
theory has influenced this study in many ways.10 Only one who has
had the benefit of his insightful criticism and sympathetic ear can
understand how much of a debt I owe him. His conceptualization
of scientific advance encourages historical analysis within which
the humanistic aspects of science are no longer submerged in
mythic objectivity. We are all in his debt for that insight.
6
This inventory identifies the location, author, recipient, and approximate date
of half a million letters from or to physicists active between 1896 and 1952.
It will soon be published in microfiche/book form. Contact Springer Verlag,
New York for information on ordering.
7
For background on the material basis of national style in physics, see Forman,
Heilbron, and Weart, HSPS, 5 (1975), 1-185.
8
Stuewer, Compton effect (1975).
9
For the neglect of experimental physics, see Heilbron and Wheaton, Literature
(1981).
10
Kuhn, Black-body theory (1978).
Preface xix
I would be pleased if this study of radiation theory were taken as
complementary to John Heilbron's analysis of concurrent devel-
opments in atomic theory.11 Matter and light are the two perceiv-
able manifestations of nature. Their interaction provided the evi-
dence that showed that electromechanical representations of
nature would fail. But although the classical-quantum compro-
mise inherent in the Bohr theory of the atom acted to delay
recognition of failure for the electromechanical atom until the
early 1920s, the paradoxical behavior of radiation was evident to
some as early as 1908. The problems that led to the rejection of
deterministic physical theory presented themselves early, and
most clearly, in attempts to bring consistency to the theory of
radiation.
BRUCE R. WHEATON
11
Heilbron, Atomic structure (1964); contribution in Twentieth century physics
(1977X40-108.
NOTES ON SOURCES
PUBLISHED SOURCES
All works cited in short form in the footnotes will be found in the
bibliography under the author's name, but the footnote references
are intended to be sufficient to lay hands on the work itself. In
general, and following Library of Congress practice, names of
journals are listed under the name of the issuing organization.
Because many are cited frequently I have used the following
abbreviations:
AHES Archive for History of Exact Sciences
AJP American Journal of Physics
AJS American Journal of Science
Amsterdam Koninklijke Akademie van Wetenschappen te Am-
sterdam. Wis- en Natuurkundige Afdeeling
AP Annalen der Physik
ASPN Geneve Bibliotheque universelle. Archives des sciences phy-
sique et naturelle, Geneve
AusAAS Australasian Association for the Advancement of
Science
BAAS British Association for the Advancement of Science
Berlin Mb, Sb Konigliche preussische Akademie der Wissenschaf-
ten, Berlin. Physikalisch-mathematische Klasse,
Monatsberichte, Sitzungsberichte
BJHS British Journal for the History of Science
Bologna Reale Accademia della Scienze delFIstituto di Bologna
CR Academie des Sciences, Paris. Comptes rendus hebdo-
madaires des seances
DSB Dictionary of Scientific Biography, 14 vols., C. C. Gil-
lispie, ed. (New York, 1970-6)
XX11 Notes on sources
GDNA Gesellschaft Deutscher Naturforscher und Arzte
Gottingen Konigliche Gesellschaft der Wissenschaften zu Got-
tingen
Halle Naturforschende Gesellschaft zu Halle
Heidelberg Sb Heidelberger Akademie der Wissenschaften, Sitz-
ungsberichte (listed by Abhandlung number)
HSPS Historical Studies in the Physical Sciences
JFI Franklin Institute, Journal
JHI Journal of the History of Ideas
JP Journal de physique et le radium
JRE Jahrbuch der Radioaktivitdt und Elektronik
Konigsberg Physikalisch- okonomisch Gesellschaft zu Konigsberg
Lincei Reale Accademia dei Lincei, Rome
Manchester MP Manchester Literary and Philosophical Society,
Memoires and Proceedings
Munchen Sb Konigliche Bayerische Akademie der Wissenschaften
zu Munchen, Sitzungsberichte
NAS National Academy of Sciences, Washington, D.C.
NNGC Nederlandsch natuur- en geneeskundig congres
NRC National Research Council, Washington, D.C.
PCPS Cambridge Philosophical Society, Proceedings
PM Philosophical Magazine
PPSL Physical Society of London, Proceedings
PR Physical Review
PRI Royal Institution, Proceedings
PRS RSL, Proceedings
PTRS RSL, Philosophical Transactions
PZ Physikalische Zeitschrift
RGS Revue general des sciences pures et appliquees
RDS Sci trans Royal Dublin Society, Scientific Transactions
RSL Royal Society of London
SFP PVRC Societe Frangaise de Physique, Proces-verbaux et re-
sume des communications
SHPS Studies in History and Philosophy of Science
TCPS Cambridge Philosophical Society, Transactions
TPRRSSA Royal Society of South Australia, Transactions and
Proceedings and Report
VDNA GDNA, Verhandlungen
VDpG Deutsche physikalische Gesellschaft, Verhandlungen
(usually bound with Berichte)
VI Victoria Institute
WAS Washington Academy of Science
Wien Sb Kaiserlich-Konigliche Akademie der Wissenschaften,
Notes on sources xxiii
Wien. Mathematisch-naturwissenschaftlichen Klasse,
Sitzungsberichte
Wiirzburg Sb Physikalisch-medizinische Gesellschaft zu Wiirzburg,
Sitzungsberichte
ZP Zeitschrift fur Physik
Zurich Physikalische Gesellschaft zu Zurich
UNPUBLISHED SOURCES
A work of this kind depends on both published and unpublished
records of contemporary opinion and results. We are fortunate
that the subject at hand developed after the typewriter was in-
vented and before the telephone was widely used. Carbon copies of
typed letters double the chance that they will be retained, and the
lack of telephone use encourages creation of a written record.
In my research I used many sources of unpublished documents,
and I am indebted to several people for permission to quote from
them: Victor de Pange for permission to quote Maurice de Broglie,
the American Friends of the Hebrew University to quote Albert
Einstein, Mme. Luce Langevin to quote Paul Langevin, Dr. Bent
Nagel to quote transactions of the Nobel Prize committee, and
Ruth Braunizer for permission to quote Erwin Schrodinger. I am
grateful to Professor Louis de Broglie, who graciously set aside
several hours for a discussion with me about this research. Of
course I, not he, am answerable for the interpretation given here.
I am no less grateful to many other libraries and archives in
Europe and America for making their collections of correspon-
dence available to me. Consulting them was essential for back-
ground information, for leads to new sources, and to ensure that
the general picture I painted was true to life. Many of the letters
cited here are quoted in part or whole in other published works.
References appear in the microfiche index An inventory of pub-
lished letters to and from physicists, 1900-1950, listed in the
bibliography under Wheaton and Heilbron (1982). Many letter
collections related to modern physics are now described in detail in
the Inventory of sources for history of twentieth-century physics (a
database soon to be published in microfiche/book form by Sprin-
ger Verlag) including all that are noted below, save the Nobel
archives.
xxiv Notes on sources
AHQP. Archive for history of quantum physics. Berkeley, Copenhagen,
Minneapolis, New York, Philadelphia, and Rome. Microfilm
references follow the form "JC, y" where x is reel number, y is
section number. For a description, see Kuhn, Heilbron, Forman,
and Allen, Sources (1967). Sources marked with a *are available
on microfilm at several of these locations
Bohr (Niels) archive.* Niels Bohr Institutet, Copenhagen, Denmark
L. de Broglie dossier. Archives of the Academie des Sciences, Paris
Cherwell (F. A. Lindemann) papers. Nuffield College Library, Oxford,
England
Ehrenfest (Paul) archive.* Museum Boerhaave, Leiden, Netherlands
Einstein (Albert) papers. Institute for Advanced Study, Princeton, New
Jersey
Langevin (Paul) papers. Formerly privately held in Paris, now at "Fonds
des ressources historique Langevin," Ecole superieur de physique
et chemie de la ville de Paris, France.
Lorentz (Hendrik) papers.* Algemeen Rijksarchief, the Hague, Nether-
lands
Nobel archives. Nobelstifting, K. Vetenskapsakademiens Nobelkommit-
teer, Stockholm, Sweden
Planck (Max) papers.* Staatsbibliothek Preussischer Kulturbesitz, Berlin
Richardson (Owen) papers.* Humanities Research Center, University of
Texas at Austin
Rutherford (Ernest) papers. Cambridge University Library, Cambridge,
England
Schrodinger (Erwin) papers.* Privately held in Alpbach, Austria
Schwarzschild (Karl) papers. Niedersachsische Staats- und Universi-
tatsbibliothek, Gottingen
Solvay (Institute) archives. Archives de 1'Universite Libre, Brussels, Bel-
gium
Sommerfeld (Arnold) papers. Sondersammlung, Deutsches Museum,
Munich
1
Introduction
For all practical purposes the wave theory of light is a

certainty.1
This is the story of a radical change in humans'concept of light. In

1896 most physicists were convinced that light consisted of wave
disturbances in a medium, the electromagnetic aether. The energy
transported in radiation was thought to propagate spherically
outward from its source and to spread out over successively larger
volumes in space. Thirty years later, a remarkably different con-
cept of light prevailed. Physicists then took seriously the evidence
collected over the preceding two decades showing that the energy
of radiation does not spread in space. Under certain conditions,
light behaves like a stream of particles.
Our subject is more than the story of a shift from one theoretical
explanation of radiation to another. This reconsideration of the
nature of light was a significant event in our scientific understand-
ing of the world. To resolve the paradox that faced them, physicists
rejected the venerable Platonic dictum that the microscopic realm
recapitulates the macroscopic; that laws generalized from the
behavior of objects in the perceivable world may be applied to the
imperceivable one. By 1927 physicists had assigned to all forms of
radiation a curious amalgam of wave and particle behavior. Waves
spread energy over larger and larger volumes of space; particles do
not. Reconciliation of these conflicting properties was possible
only through appeal to an ontology that transcended mechanical
incompatibility.
Nature was declared to be only imperfectly rationalizable in
terms of human experience with macroscopic interactions. The
programmatic goal formulated in the seventeenth century to re-
1
Hertz, GDNA Tageblatt (1890), 144.
2 Introduction
duce all physical phenomena to consistent mechanical representa-
tions was here recognized to be unattainable. This was realized in
part because of paradoxes found in the behavior of radiation,
particularly x-rays. This study is therefore intended to clarify the
developing understanding of x-rays and related phenomena early
in the twentieth century. But it is physicists' growing awareness
that electromechanical explanations of radiation would not work
in principle that forms the real subject of this book.
STRUCTURE OF THE ARGUMENT

In the earliest period of our concern, analogies were drawn be-
tween known properties of light and the growing empirical evi-
dence about new forms of radiation. Shortly after the discovery of
x-rays, it was proposed that the new rays were impulses, not
periodic waves, propagating in the aether. Empirical corrobora-
tion followed, matched by increasing acceptance; 1907 marks the
high-water mark of the pulse hypothesis. Contemporary studies on
radioactivity contributed a new example of an electromagnetic
impulse: When the properties of the y-rays were closely investi-
gated, they came increasingly to be explained in the same way as
were x-rays. X-rays were the spherically propagating impulses due
to the deceleration of cathode-ray electrons; y-rays were the spher-
ically propagating impulses caused by /?-ray electrons accelerating
from a disintegrating atom.
The second period brought the first clear attempts to formulate a
consistent theoretical picture of the process of ionization by x-rays
and y-rays. Problems were soon recognized when seemingly con-
tradictory aspects of x-ray behavior were discovered. Increasing
discussion threw doubt on the validity of any electromagnetic
explanation of x-rays. J. J. Thomson and W. H. Bragg in England,
and Wilhelm Wien and Johannes Stark in Germany, each aban-
doned critical aspects of the impulse hypothesis. The energy that
x-rays pass on to atoms was more than could be expected from any
spreading impulse. Moreover, only a few atoms passed over by
x-rays emitted an electron; and some physicists asked whether the
energy of x-radiation was not propagated only into narrow solid
angles centered on the source.
The paradoxical behavior of x-rays provoked two different kinds
of investigations in the third period. There were experimental
Introduction 3
studies of x-rays, light, and y-rays to determine the angular extent
of the effective region that acts on individual atoms. The results
indicated that the radiations do not propagate their energy iso-
tropically, but instead concentrate it in a cone of limited extent. At
the same time, theoretical attempts were made to explain the
restricted solid angle in a way that was compatible with electro-
magnetic theory. The strongest defender of impulse x-rays was
Arnold Sommerfeld, who found explanations for some of the
difficulties. But Sommerfeld managed this only by stretching the
impulse model to extremes, and by 1912 other problems had
arisen. It had become clear that x-rays and light are common forms
of radiation; consequently, the paradoxes that plagued x-rays
applied just as forcefully to ordinary light. The photoelectric effect
in particular became a serious problem after 1911, and several
unsuccessful attempts were made to find a classical explanation for
it. The basic question was: How can periodic light waves appar-
ently concentrate their entire quantity, or quantum, of energy on
an individual electron? Sommerfeld had no answer.
The fourth period began with a new discovery that seemed to
corroborate both sides in the debates of the preceding years. In
1912 it was found that x-rays can interfere. On the one hand, x-ray
spectroscopy provided unquestionable evidence that some x-rays
possess a periodic structure and are therefore no different from
ordinary light. X-rays from a given metal produce a characteristic
pattern of lines in the spectroscope that is entirely equivalent to
optical emission spectra. On the other hand, the new spectroscopy
supplied precisely the tool needed to confirm the suspicion that
x-rays transfer energy to matter only in discrete units. The quan-
tity of energy that they pass to an electron was found to be roughly
equivalent to the total quantum of energy in the radiation. This
seemed to leave no alternative but to accept the fact that radiation
concentrates its energy in specific directions and does not spread
that energy to any extent in space. The same techniques that by
1914 showed x-rays and y-rays to be bona fide waves soon pro-
vided the strongest evidence that they, and ordinary light as well,
consist of streams of hypothetical lightquanta, localized particle-
like concentrations of energy in proportion to frequency or im-
pulse width.
The evidence corroborating the failure in principle of a classical
representation of radiation had only begun to appear when Niels
4 Introduction
Bohr proposed his quantum theory of the atom. The lightquantum
hypothesis then won almost no support among those whose major
goal was the construction of an internally consistent mathematical
representation of nature. Throughout most of the decade
1911-21, mathematical physicists simply avoided the issues that
complicated understanding of free radiation. The Bohr theory of
the atom provided a promising field of investigation that diverted
attention from the paradoxes of radiation. During World War I,
interest in these problems was maintained almost exclusively by
empiricists, those whose views were derived from experiments on
the absorption of radiation. Some of these physicists accepted the
lightquantum as a valuable heuristic concept before 1920, but
none of them inquired deeply into the structure of radiation itself.
The fifth period began in 1921 when influential members of this
small group of empirical physicists began to call for a serious
reconsideration of the problematic status of light. They based their
concern on the successful corroboration of the quantum relation
for the x-ray and y-ray photoelectric effects. In an environment of
increasing acceptance of Einstein's lightquantum, two compelling
results were achieved. First, in 1922 Arthur Compton showed that
x-rays conserve linear momentum in their interaction with elec-
trons. Were the quantum nature of x-rays assumed, remarkable
harmony could be brought to the assembled experimental evi-
dence about such interactions. Second, in 1923 Louis de Broglie
forged a connection between the properties of light and those of
atoms. Each was to have both wave and particle characteristics;
only through the interaction of both properties could the true
behavior of either matter or radiation be demonstrated. The final
reconciliation of the problems of radiation occurred through ap-
peal to indeterministic theories. Soon thereafter, powerful mathe-
matical techniques were applied by Schrodinger and others to
provide a full quantum-mechanical solution to problems with
both radiation theory and atomic theory.
The polarization between mathematical and empirical ap-
proaches to problems about the nature of free radiation ended at
just the time that the quantum theory of the atom confronted
equally serious problems. Two separate strains of thought, one
concerned with atoms and the other with light, were reunited after
a decade of unprofitable isolation. It is no coincidence that this
recombination of matter theory and radiation theory produced
Introduction 5
two essentially independent formulations of the new quantum
mechanics. One was addressed to the problems of the atom; the
other was a development from ideas already formulated to inter-
pret the absorption of light.
BRITISH MECHANISTIC PHYSICS

Most British physicists in the late nineteenth century interpreted
nature as if the laws of ordinary mechanics were valid on the
microscopic level. Many thought that the electromagnetic aether
could be represented in terms of mechanical models. Their ap-
proach was based on the conviction that the laws governing the
interaction of perceived objects could be transferred bodily to the
actions of molecules and atoms. For some, like Maxwell, this was
no slavish dependence on physical models, but rather a conviction
that, however accurate the mathematical representation of nature
might be, the creative impulses of physicists are best served by an
appeal to mechanical representations.2 The British proclivity to
seek mechanical models in physics played an essential role in the
early recognition that an electromechanical representation is in-
herently impossible for light.
For our purposes, the most significant example of this attitude in
British physics concerns the study of electric discharges through
vacuum. It had been known since midcentury that reproducible
displays of light and shadow accompany the discharge of electricity
through a tube partially evacuated of air. After the improvement of
vacuum pumps in the 1860s, it was possible to study the effects of
the discharge after most of the gas had been removed and the
normal discharge pattern had entirely disappeared. Under these
conditions, some sort of disturbance still traverses the tube. Al-
though the responsible "rays" themselves are invisible, they create
a bright glow where they strike the glass walls of the discharge tube.
Because they seem to come exclusively from the negative elec-
trode, or cathode, they were called cathode rays.3
One should not be misled by the denotation "rays." To most
British physicists, cathode rays were a stream of "charged material
molecules." This interpretation had been offered first by William
2
Kargon, JHI, 30 (1969), 423-36.
3
Cathode rays were discovered by Plucker, AP, 103 (1858), 88-106; named by
Goldstein, Berlin Mb (1888), 82-124.
6 Introduction
Crookes in 1879 after he showed that the beam is deflected by a
magnet, can exert force on an object placed within the tube, and
casts sharply defined shadows even when the emitting cathode
surface is quite large.4 The last observation argued against the
contention that the rays are emitted isotropically from the cath-
ode, as light would be. There were variations on this theme: Some
thought that the particles were complete molecules that had accu-
mulated a net charge; others thought they were the negatively
charged products of molecular dissociation. But the majority of
British physicists agreed that the cathode rays were material.
Significant research was based on this interpretation. Crookes
claimed that he had discovered a "fourth state of matter" that was
neither solid, liquid, nor gas because of the extremely low density
of the particles. In 1884 Arthur Schuster used the magnetic devia-
tion of the cathode beam to estimate the ratio of charge to mass
(e/m) of the presumed particles.5 As is well known, Joseph John
Thomson soon found this ratio to be the same for the beam
regardless of the nature of the emitting cathode and the residual gas
in the tube. The value of e/m that he found was larger by a factor of
a thousand than that of the smallest known atom, and Thomson
claimed in 1897 that the cathode beam particles are parts of atoms
previously undetected.6 Thomson's particle is today called the
electron.
Of most significance in this investigation is Thomson's and
Schuster's conviction that the laws of electrostatic deflection, de-
rived from the behavior of macroscopic objects, should necessarily
apply to electrons. They were rewarded in this belief by the discov-
ery of new phenomena. Very soon thereafter, Thomson, in partic-
ular, encountered perplexing inconsistencies when he tried to
apply the same representational standards to the behavior of
radiation.
THE ELECTROMAGNETIC WORLD VIEW

Many German physicists regarded the mechanistic bias of British
physicists as superficial and potentially misleading. They consid-
ered the dependence on mechanical models to be a limitation on
4
Crookes, PTRS, 170 (1879), 135-64. Crookes' lecture "On radiant matter" to
the BAAS does not appear in the 1879 Report but was reprinted privately the
same year.
5
Schuster, PRS, 57(1890), 526-59.
6
Thomson, PM, 44(1897), 293-316. See also Topper, AHES, 7(1971)393-410.
Introduction 7
the mind rather than an aid to creative thought. The origin of this
deeply ingrained cultural tendency in German physics has not
been fully explained. One source of influence is assuredly the
prevalent atmosphere of idealistic philosophy to which most turn-
of-the-century German-speaking physicists were introduced in
their formative or university education. Also, a more abstract
interpretation of natural law was common in German-speaking
countries, one that placed less emphasis on material interpretation
of phenomena than was true in Britain.
In part because of this predisposition, German physicists did not
accept British theories about cathode rays. Most of them described
the rays as a form of light that traverses the tube.7 The earliest
justification for this view was the belief that only light could
stimulate fluorescence in matter like that of the discharge tube
walls. The magnetic deflection of the beam could be attributed to
an abrupt alteration of the refractive index in the residual gas near
the magnet. Heinrich Hertz provided the strongest evidence for
the German view in 1883 when he was unable to deflect the
cathode-ray beam in an electrostatic potential. He also demon-
strated that the beam travels straight ahead even if the course of the
electrical discharge curves to the side.8
There were therefore very good experimental reasons for Ger-
man physicists to accept the radiant interpretation of cathode rays.
Philipp Lenard, for example, based a Nobel-prize-winning series
of researches on the aetherial view of cathode rays. He exploited
the fact that the rays could pass right through thin metal foils.9 This
too supported the aetherial explanation; solid matter stops streams
of molecules, but some forms of matter are transparent to visible
or electric waves. Even after Thomson identified the electron by
measuring its e/m ratio, some German physicists - both experi-
mentalists and mathematical theorists - persisted in attempts to
explain its behavior entirely in terms of the aether.
German mathematical physicists were not deterred by an inabil-
ity to represent physical concepts mechanically. A school of
thought at this time in German-speaking countries sought to
reformulate ideas of matter solely in terms offerees in the aether.10
7
Lehmann, Lichterscheinungen (1898), 518-47.
8
Hertz, AP, 79(1883X782-816.
9
Lenard, AP, 51 (1894), 225-68.
10
On continental rejection of mechanical models, see Klein, Centaurus, 17
(1972), 58-82. Joseph Larmor in Britain held views similar to those of many
Germans; see his Aether and matter (1900).
8 Introduction
It was a form of electromagnetic reductionism and was heavily
influenced by Kantian principles of epistemology. In a famous talk
"On the relations between light and electricity" in 1889, Heinrich
Hertz conjured up a vision of this way of thinking. Like the
pre-Socratic monists who attempted to find the single principle
underlying all apparent phenomena, Hertz claimed that physicists
were confronted with the question of "whether everything that is,
is fashioned from the aether." The aether offers understanding not
only of the imponderables, he said, "but even of the nature of
matter itself and its most intimate properties, weight and mass."11
The forces that define our perception of mass and solidity were to
be given higher ontological significance than was the idea of
matter.
Not all German-influenced physicists agreed with the fullest
formulation of this viewpoint. By the mid-1890s, some success had
attended a modified version of the reductionist program. This was
the electron theory of the eminent Dutch physicist H. A. Lorentz.12
According to Lorentz, all interaction between matter and the
aether is mediated solely by then-hypothetical electrons. Electrons
act like particles, but their mass and momentum are only reflec-
tions of the effect that a center of electric force has on the contin-
uum of force that constitutes the aether. The mechanical proper-
ties of electrons are due to the reaction between their inherent
charge and the surrounding electric and magnetic field. Lorentz'
detailed theory was able to account for several perplexing observa-
tions, including the failure of interference experiments to find
evidence for the relative motion of the earth and the aether.
But other physicists interpreted the issues Hertz raised in an
extreme form. Some felt that they stood "on the threshold of a new
epoch in physics," one in which suitably defined discontinuities in
the continuous electromagnetic field would entirely replace me-
chanical concepts.13 Emil Wiechert and Wilhelm Wien, for exam-
ple, felt that the very concept of electron could be reduced to the
idea of a form or structure in the electromagnetic medium.14 Those
who subscribed to this all-encompassing reductionist program
11
12
Hirosige, HSPS, 1 (1969), 151-209.
13
Quincke, GDNA Tageblatt (1890), 149.
14
Wiechert, Konigsberg Schriften, 35 (1894), 4n- Wien, Archives Neerlan-
daises (2), 5 (1900), 96-107. Abraham, Elektrizitdt, 2 (1905), i47ff.
Introduction 9
described their goal as the "electromagnetic world picture." To
them, only the electromagnetic aether existed. It was all that was
necessary to explain the phenomena of the world.15
THE WAVE THEORY OF LIGHT

Throughout the second half of the nineteenth century, light was
thought to be a wave propagating in an all-pervading medium. Its
properties - diffraction, interference, and polarization - con-
vinced its students that visible monochromatic light is a periodic
transverse oscillation with a wavelength between about 3 x io~5
and 7 x icr5 cm. The origin and development of optical spectros-
copy were based on this interpretation; the assignment of a nu-
merical wavelength to each pure color of light brought great
advantages in the analysis of terrestrial and astronomical matter.
During the nineteenth century, radiations on both sides of the
visible spectrum were discovered: the infrared and the ultraviolet
or chemical rays, so-called because of their ability to provoke
chemical reactions. Like visible light, infrared and ultraviolet light
were understood to be periodic transverse waves.
Yet lacking was a physical understanding of the medium of
which the undulations constitute light. This omnipresent
medium - the aether - had to have seemingly contradictory prop-
erties. It had to be extremely rigid to sustain the high-frequency
vibrations of light, and yet be sufficiently yielding to allow matter,
such as the planets, to pass through it freely. Physicists had at-
tempted to formulate a consistent representation of an elastic solid
with such properties. Even if they had succeeded, they would have
faced grave problems. The aberration of starlight indicated that the
earth moves relative to the aether during its yearly revolution
about the sun, but interferometric experiments failed to corrobor-
ate this effect. Moreover, if the aether were strictly analogous to an
elastic solid, then a longitudinal wave traveling at a greater velocity
than the transverse should have been observed. It had not been.
15
McCormmach, his, 61 (1970), 459-97. Hertz himself rejected the electromag-
netic force ontology. His Mechanik (1894) attributes the forces to insensible
particles. McCormmach, DSB, 6, 340-50. For representative statements
about the electromagnetic view during our general period of concern, see the
proposition by Marx, Grenzen in der Natur (1908), and the antithesis by Kunz,
Theoretische Physik (1907).
10 Introduction
Maxwell's theoretical formulation of electricity and magnetism
in the 1860s offered partial answers to some of the puzzles about
the luminiferous medium.16 He proposed that light is a transverse
wave traveling in a continuous medium whose stresses and strains
constitute electric and magnetic forces. Like jiggles moving along a
rope, oscillations perpendicular to the direction of motion of the
wave should occur along the force vectors established by electrified
bodies. Maxwell showed that the velocity of these waves would be
close to that found for electrical signals in a wire. Even before it
received experimental corroboration, the theoretical interpreta-
tion that light is a periodic electromagnetic wave provided support
for the wave theory of light.
George Gabriel Stokes described in 1884 the "views respecting
the nature of light which are at present held, I might say almost
universally, in the scientific world."17 He meant the wave theory of
light. He regarded the idea that light is a stream of particles shot
from the luminous source as "altogether exploded." "No one who
has studied the subject can doubt," he said, that "light really
consists of a change of state propagated from point to point in a
medium."18 The colors of thin films and diffraction phenomena
show, he said, that periodicity is "naturally, almost inevitably,
involved in the fundamental conception."19
In Germany, Hermann von Helmholtz began his lectures on the
electromagnetic theory with an affirmation of the wave theory.
"Hardly a doubt remained about the superiority of the wave
theory compared to the emission theory" after polarization was
explained by the hypothesis of transverse waves, he claimed; the
emission theory is "hardly mentioned anymore."20 Gustav Kirch-
hoff based his theory of diffraction in 1882, as Stokes had done
before, on the assumption that "light is a transverse undulation in
the aether."21 The leading French mathematical physicist of the
time, Henri Poincare, wrote in 1888 that "of all theories in physics,
the least imperfect is that of light based on the work of Fresnel and
his successors."22
In 1888 Heinrich Hertz completed proofs that "electric waves"
16
Maxwell, PM, 21 (1861), 161-75, 338-48; 23 (1862), 12-24, 85-95.
17
Stokes, Nature of light (1884), viii.
18
Ibid., 25.
19
Ibid., 32.
20
von Helmholtz, Vorlesungen (1897), 3.
21
Kirchhoff, Berlin Mb (1882), 641.
22
Poincare, Theorie mathematique de la lumiere, 1 (1889), 1.
Introduction 11
radiate from sparks and obey the same laws that govern the
behavior of light. They can be refracted and reflected; their wave-
length can be measured; their polarization indicates that they are
transverse waves; and they propagate at the velocity of light.23
Maxwell's theory, proposed as an explanation of optics, had
achieved a "brilliant victory."24 The inconsistent properties of the
mechanical aether, it seemed, might disappear entirely when rein-
terpreted on the basis of the electromagnetic continuum supposed
by Maxwell's theory.
After Hertz' experiments, physicists' conviction about the wave
theory of light became virtual certainty. Hertz provided the mod-
ern electromagnetic context for the wave hypothesis and spoke
forcibly to German physicists who had found the concept of a
mechanical aether alien to their sensibilities. By 1890 most physi-
cists thought that the wave theory of light was as close to certainty
as physics could make it. This theory had assumed a status equal to
that of the theory of gravitation as an example of a fully corrobor-
ated theory in classical physics.
The differences between the British and German approaches to
electromagnetic theory did not lead directly to a difference in their
respective interpretations of light. Both British and German physi-
cists were convinced that monochromatic light obeys all the laws
governing periodic waves. But they did differ on how light waves
may be analyzed. Many influential German physicists thought
that electromagnetic forces, extending over great distances in
space, can, through an appropriate twist or discontinuity, give rise
to apparently localized properties in the aether. In Germany, more
widely than in Britain, the idea was accepted that a properly
selected set of sinusoidal oscillations, if added together, will pro-
duce a single localized force. The willingness of some German
physicists to combine, or their inability to separate, periodic waves
and temporally or spatially localized events in the aether was a
crucial advantage early in the century when the behavior of radia-
tion was first interpreted on the basis of quantum principles. And
national distinctions did play a significant role in the various
hypotheses raised after 1895 when word came from Wurzburg
about x-rays, an entirely "new type of radiation" that could pass
right through matter itself.
23
Hertz, AP, 34 (1888), 551-69; Berlin Sb (1888), 197-210; AP, 36 (1889),
769-83.
24
PART I
The introduction of temporal discontinuity
1896-1905
The electromagnetic impulse hypothesis
of x-rays
The Rontgen emanation consists of a vast succession of

independent pulses . . .l
Between 1898 and 1912, a majority of physicists thought that

x-rays were impulses propagating through the electromagnetic
field. Only the extremely large number of pulses gave the x-ray
beam its seeming continuity. Although this hypothesis was com-
patible with the wave theory of light, it was a special case of that
theory. Impulses are not ordinary waves. Although they propagate
spherically outward from their source, pulses are not periodic
oscillations. The energy in an impulse is temporally but not spa-
tially localized. It is contained within an ever-expanding shell, but
the shell's radial thickness remains constant and small. Along the
circumference, energy is distributed uniformly. But radially, from
front to back so to speak, electromagnetic energy rises quickly
from zero and drops back just as rapidly. When it passes a point in
space, an impulse exerts only a single push or a single push - pull. A
pulse collides, rather than resonates, with an atom.
In their temporal discontinuity, impulses differ decisively from
their periodic-wave cousins. Light has an intrinsic oscillatory char-
acter that allows it to interfere; the superposition of two beams of
coherent monochromatic light produces alternate regions of con-
structive and destructive interaction, the well-known interference
fringes. A pulse has no oscillatory structure. Its interference prop-
erties are qualitatively different from those of light. A truly mono-
chromatic light wave must extend infinitely in time; if it does not,
an intrinsic ambiguity arises in the definition of its frequency. A
pulse is restricted in temporal extent, and the very concept of
1
Stokes, PCPS, 9 (1896), 216.
16 The introduction of temporal discontinuity, 1896 -1905
frequency cannot readily be applied. Recognition of these differ-
ences varied according to national styles in research and from
physicist to physicist, as we shall see in this chapter.
In a larger context, it should be remembered that the hypothesis
of x-ray impulses competed for acceptance with many claims for
"new rays." In the first decade of x-ray research, physicists were
called upon to pass judgment on numerous explanations for radi-
ant interaction. Some of the newly proposed entities - x-rays,
cathode rays, canal rays, and a-, /?-, and y-rays, plus a few radioac-
tive emanations and a bewildering variety of decay products-were
later recognized to be legitimate. Others - discharge rays, radioac-
tinic rays, black light, N-rays, Moser rays, spathofluoric rays, and
the peculiar light of glowworms - were, at best, special cases of the
others and, at worst, totally spurious.2 No clear distinction could
be drawn between the phenomena themselves and the many
explanations proposed, nor could one be certain about which new
observations were valid at all. In this circumstance, it is hardly
surprising that there was early confusion over which of the new
rays were particle streams and which were waves in an aether.
Impulses came to occupy a position midway between these ex-
tremes. A pulse is a spherical wave, but it collides with an atom as if
it were a particle. Part of the early appeal of the impulse hypothesis
was this ability, chameleonlike, to express the characteristics of
both particles and waves.
THE NEW RAYS

Within six months of Wilhelm Conrad Rontgen's announcement
of his discovery of x-rays late in 1895, several proposals were heard
about the physical nature of the "new type of ray."3 Suggestions
included discharged and uncharged particles,4 vortices in the
2
Nye, HSPS, 11 (1980), 125-56.
3
Rontgen, Wurzburg Sb (1895), 132-41; Wurzburg Sb (1896), 11 -16; Berlin Sb
(1897), 576-92. Reprinted in AP, 64 (1898), 1-11, 12-17, and 18-37. Future
references follow the form "Rontgen I, II, or III," with page references to the AP
reprinting. The most useful bibliography is Klickstein, Wilhelm Conrad Ront-
gen (1966). Gocht, Rontgen-Literatur (1911 -25), concentrates on medical and
technical issues.
4
The many claims may be found in Phillips, Bibliography (1897). The most
influential claimant for uncharged electrons was Bernhard Walter, AP, 66
(1898), 74-81.
The electromagnetic impulse hypothesis of x-rays 17
5 6 7
aether, and acoustical or gravitational waves of high frequency.
Within a short time, however, the choices were narrowed to three,
each a form of electromagnetic disturbance. The first theory,
popular in Germany before 1900, stated that x-rays were ordinary
light of extremely high frequency. This hyperultraviolet light con-
sisted of periodic transverse oscillations in the electromagnetic
field. The second idea, suggested by Rontgen himself, raised the
hope that x-rays were the long-sought condensational, or longitu-
dinal, aether wave. The third identified x-rays as temporally dis-
continuous transverse impulses in the aether.
From the start, most physicists took Rontgen's rays to be a form
of electromagnetic wave. In support of this claim, Rontgen pro-
vided only three pieces of evidence: (1) The rays, like light, propa-
gate in straight lines. (2) The rays are undeflected by magnetic or
electric influence. (3) In contrast to the behavior of cathode rays,
x-rays are not absorbed by matter in strict proportion to the
density of the absorber.8 On the other hand, there was strong
evidence that x-rays do not behave like ordinary light rays: Ront-
gen showed that they suffer no measurable reflection, refraction,
interference, or polarization. How then can one account for the
consensus view that x-rays are waves? The answer lay in the form
of the evidence. The overarching exhibit, the technological fait
accompli that irrevocably cemented Rontgen's name to a phenom-
enon seen by others prior to 1895, was the photographs. Photogra-
phy by ordinary light was new enough in 1896 to hold the interest
of many physicists. But here was evidence of new light, not visible
to the eye, that exposed a photographic plate nonetheless. Techni-
cal claims were open to interpretation, but the images of a compass
needle with graduations seen through its metal case, of nails in the
door frame seen in a photograph taken outside a closed room
containing the discharge tube, and, eeriest of all, of the bones
inside a living hand were accepted at face value. A new form of
light had been recorded.9
5
Michelson, AJS, 1 (1896), 312-14; Nature, 54 (1896), 66-7.
6
Edison, Electrical engineer, 21 (1896), 353-4.
7
Lodge, Electrician, 36 (1896), 438-40, 471-3.
8
Rontgen I, pp. 3, 5, 6, 8, 9, and 10.
9
"For me the photography was a means to an end but [the press] made it the
most important thing." Rontgen to L. Zehnder, February 1896. Zehnder, W. C.
Rontgen (1935), 39-44. For a description of the photographs that Rontgen sent,
see Schuster, British medical journal (18 January 1896), 172-3.
To explain the lack of characteristic radiative behavior, Rontgen
proposed tentatively that the rays were longitudinal aether waves,
long anticipated on the basis of the elastic solid model of the
aether.10 Rontgen's proposal did not spring de novo from his
imagination. Indeed, his complete lack of supporting evidence
suggests that the source of his confidence lay outside the context of
his own investigations. One need not seek far to find it. Both
Gustav KirchhofFand Lord Kelvin had published formulations of
electromagnetism that predicted the existence of longitudinal as
well as transverse electromagnetic waves.11 Hermann von Helm-
holtz and Heinrich Hertz had both toyed with the idea that
cathode rays were compression waves,12 and Gustav Jaumann had
developed this concept in a series of papers that predate Rontgen's
work.13 The problems that had complicated a compression-wave
model of cathode rays seemed to disappear entirely when x-rays
became the subject.
So Rontgen's hypothesis of longitudinal-wave x-rays struck
sympathetic chords in representatives of his generation. Ludwig
Boltzmann, who had recently analyzed the mechanical properties
of the electromagnetic medium, supported Rontgen's idea.14 He
thought the x-rays might be longitudinal waves of sufficiently short
period to induce fluorescence but of long enough wavelength to
pass easily through matter.15 An acoustical analogy is helpful here:
Long-wavelength sound is less readily absorbed by matter than
short-wavelength sound. Henri Poincare agreed that the lack of
refraction cast doubt on any transverse-wave model.16 Gustav
10
Rontgen I, p. u .
11
[W. Thomson], Baltimore lectures (1904), 45, 141. Kirchhoff, Berlin Mb
(1882), 641-69; Vorlesungen, 2(1891), 79-97.
12
Hertz wrote Helmholtz in 1883 suggesting that cathode rays might be longitu-
dinal waves. Helmholtz replied that he had once thought that cathode rays
were electromagnetic waves, likely longitudinal, that propagate from plotz-
lichen Stosses, molecular collisions at the cathode. Helmholtz to Hertz, 29 July
1883. Hertz, Gesammelte Werke, 3 (1895), xxv-vi.
13
Jaumann, Wien Sb, 104:11a (1895), 747-92.
14
Boltzmann, Elektro-Techniker, 14 (1896), 385-9; Electrician, 36 (1896), 449.
The earlier study: AP, 48 (1893), 78-99.
15
The velocity of the compression wave was expected to be greater than that of
the transverse wave. Thus, longitudinal waves might possess high frequencies
and long wavelengths simultaneously. Both properties help to explain the rays'
extraordinary ability to pass through matter.
16
Poincare, RGS, 7(1896X55.
Jaumann, who had developed a longitudinal hypothesis for the
cathode rays, simply added x-rays to the category.17 Oliver Lodge
"hoped" that the rays were longitudinal oscillations, but he did not
think they were electromagnetic. He thought that x-rays might be
the gravitational analog to electromagnetic waves: As light is the
wave whereby electric and magnetic forces are transmitted, so
x-rays might be the wave of gravitational force. Their frequency,
he thought, might be as much greater than that of visible light as
the frequency of light is above that of acoustical vibration.18
Lord Kelvin took the compression-wave hypothesis most seri-
ously. He quickly recognized that partial corroboration of his
elastic-solid model of the aether might be forthcoming. He had
recently expressed concern that the compression wave had not yet
been observed, and had suggested other experimental tests.19
"There are such waves," he had claimed, "and I believe that the
unknown condensational velocity is the velocity of propagation of
electrostatic force."20 Rontgen's rays appeared to hold out new
hope, and Kelvin actively encouraged experiments on them.
The critical experiment to disprove Rontgen would have been a
demonstration of polarization. But throughout the period
1896-1904, polarization of x-rays remained problematic. Many
physicists claimed its demonstration. Karnozhitskiy and Golitsyn
in St. Petersburg reported in March 1896 that the transmission of
x-rays through tourmaline depends on the relative orientation of
the sheets.21 But the discovery of radioactivity cast doubt on the
purity of the x-ray beam that had been used, and little weight was
given to the result. J. J. Thomson, in an early experiment, failed to
find any effect of tourmaline on x-rays.22 Some took the lack of
polarization evidence as support of the longitudinal hypothesis.
Opinion soon turned against Rontgen's suggestion even without
positive evidence. Arthur Schuster pointed out that even trans-
verse waves of sufficiently high frequency would suffer little refrac-
17
Jaumann, Wien Sb, 106:11a (1897), 533-50.
18
Lodge, Electrician, 36 (1896), 438-40.
19
[W. Thomson], PRS, 59 (1896), 270-3.
20
[W. Thomson], Baltimore lectures (1904), 141.
21
Reported in Nature, 53 (1896), 528. Stokes, Manchester MP, 41 (1897), no. 15,
p. 2. Gifford, Nature, 54 (1896), 172.
22
J. J. Thomson, Nature, 53 (1896), 391-2.
20 The introduction of temporal discontinuity, 1896-1905
tion, diffraction, or reflection.23 The principal motive of the longi-
tudinal wave thus removed, Ockham's razor, which slices away
multiple hypotheses where one will suffice, more than compen-
sated for the lack of definite evidence about polarization. By the
fall of 1896, transverse waves had become "the surviving hypoth-
esis."24
Major support for the transverse wave came from Henri Bec-
querel's discovery of radioactivity in 1896, for Becquerel had
claimed to find polarization in his uranium rays.25 At the end of
1896, most physicists agreed with Thomson when he described the
Becquerel rays as resembling "the Rontgen rays more closely than
any kind of light known hitherto."26 The Becquerel rays seemed to
be intermediate between light and x-rays. This view "greatly
strengthened the position . . . that the x-rays were of the nature
of ultra-ultraviolet light."27 Because the Becquerel rays reportedly
showed reflection and refraction, the possibility of these properties
was more easily projected onto x-rays. Lodge concluded, late in
1896, that "it has become almost certain that the x-rays are simply
an extraordinary extension of the spectrum," that is, periodic
transverse waves of high frequency.28
THE BRITISH VIEW OF ELECTROMAGNETIC

IMPULSES
X-rays are produced when a cathode ray beam strikes the metal
anode or the walls inside an evacuated electrical discharge tube. In
Britain, most researchers were convinced that cathode rays con-
sisted of projected charged particles. It was quickly recognized that
each particle, stopped by the collision, should give rise to a distur-
23
Schuster, Nature, 53 (1896), 268. That longitudinal waves were not required to
explain the penetrability of x-rays was recognized immediately in Britain. See
Thomson to Lodge, 19 January 1896, in [R. J. Strutt], Life of Thomson (1942),
65-6; and FitzGerald to Lodge, 24 January 1896, in Lodge, Electrician, 37
(1896), 370-3-
24
Lodge, ibid.
25
Becquerel, CR, 122 (1896), 762-7; and earlier studies, ibid., 559-64 and
689-94. For a useful discussion of Becquerel's early work, see Malley, Hyper-
phosphorescence (1976).
26
The Rede Lecture, delivered 10 June 1896, Nature, 54 (1896), 304. See also J. J.
Thomson, BAAS Report (1896), 7O3ff.
27
Lodge, Electrician, 57(1896), 370.
28
Ibid, 371.
bance in the equilibrium of the surrounding electromagnetic field.
This dislocation would propagate spherically outward from the
impact site and affect charged particles as it passed over them. This
impulse hypothesis of x-rays was discussed informally in Britain
very early in 1896.29 Late that year, as the material interpretation of
cathode rays neared experimental corroboration, the electromag-
netic impulse hypothesis of x-rays began to exercise greater influ-
ence.
In the English-speaking world, the pulse hypothesis was pro-
pounded by Sir George Gabriel Stokes, then in his forty-seventh
year as Lucasian Professor of Mathematics at Cambridge.30 Stokes
had not believed Rontgen's compression-wave hypothesis; he
cited George Green's arguments against the necessary existence of
compression waves.31 At first, he had taken x-rays to be high-
frequency transverse waves.32 But in March 1896 he recognized
that even this "supposition is by no means exempt from diffi-
culty."33 In July he hinted at the impulse hypothesis. Speaking as
president to the Victoria Institute, Stokes cited Becquerel's claims
for polarization of uranium rays as if they confirmed the transverse
nature of x-rays.34 He ignored the fact that Becquerel had, on three
separate occasions, pointed out that his new rays were not x-rays.
Stokes had other reasons to support a transverse model of x-rays,
reasons that he did not yet specify in detail.
Kelvin was dubious. One of the chief difficulties with any peri-
odic wave model was how extremely small the wavelength had to
be to explain how x-rays pass through matter. The French physicist
L. Georges Gouy had determined an upper limit on the wave-
length of 0.005 /*m> o r l0/ that of green light.35 Stokes and Kelvin
both doubted that any mechanical process could generate oscilla-
tions at the rate of 70 thousand million million (7 x 1016) per
second. Thinking that Stokes was proposing periodic transverse
29
Schuster, Nature, 53 (1896), 268. Lodge, Electrician, 37 (1896), 370-3.
30
For general information on Stokes, see Parkinson, DSB, 13 (1976), 74-9.
31
Green held that the aether is incompressible and that the longitudinal wave has
no physical existence. See TCPS, 7(1839), 2-24, 121-40.
32
Stokes to S. P. Thompson, 29 February 1896. Stokes, Memoir, 2 (1907), 495-6.
33
Stokes to S. P. Thompson, 2 March 1896. Stokes, Memoir, 2 (1907), 496.
34
Stokes, VI Journal of transactions, 30 (1898), 13-25; Nature, 54 (1896),
427-30. The Victoria Institute was a Christian apologist organization, dedi-
cated to countering "oppositiones falsi nominis scientiae," founded in 1866.
35
Gouy, CR, 123 (1896), 43-4. Precht, AP, 61 (1897), 350-62.
waves of these high frequencies, and still believing that x-rays are
compression waves, Kelvin simultaneously praised and gently
chided the seventy-seven-year-old Stokes:36
In physical science my touchstone as to truth is, "what does Sir George
Stokes think of it?" When I hear that he declares for transverse vibrations,
and when I hear the strong reasons he has put before us for this conclu-
sion, I myself am very strongly fortified indeed in accepting it. I do not,
however, forget that it is put before us not as absolutely demonstrated but
as his present opinion . . . he leaves our minds open to the possibility of
other explanations - indeed he has tantalized us very much with the idea
of the possibility of another explanation to which he has alluded.
Stokes committed himself to the impulse hypothesis in No-
vember in a talk to the Cambridge Philosophical Society.37 Ac-
cording to Maxwell's theory, he said, individual impacting cath-
ode-ray particles should give rise to a "vast succession of
independent pulses [in the aether] analogous to the 'hedge-fire' of a
regiment of soldiers."38 Stokes thought that the impulse due to a
colliding electron might very well be as short as Gouy's estimated
wavelength.
Stokes, the leading British authority on optical diffraction, was
thus led to a new conception of refraction - one that, he said,
"fairly startled" him atfirst.In his Wilde lecture to the Manchester
Literary and Philosophical Society in July 1897, he worked out the
details.39 Molecules of ponderable matter respond by resonance to
passing waves. Energy so extracted serves to slow the beam, and
the resulting interference of Huygens' wavelets refracts the ray.
Aperiodic pulses, on the other hand, can never reinforce the initial
push-pull; they lack the requisite oscillation. Therefore, Stokes
concluded, x-ray impulses could pass through matter without
36
Kelvin's remarks following Stokes' address, VI Journal of transactions, 30
(1898X25-7.
37
Stokes, PCPS, 9 (1896), 215-16. Stokes also gave "a beautifully clear and
animated exposition of his theory of Rontgen rays" at the 1896 meeting of the
British Association for the Advancement of Science, according to J. J. Thom-
son, Recollections (1936), 409. That Stokes spoke is reported in Nature, 54
(1896), 565; no details are given in the 1896 BAAS Report.
38
Stokes, PCPS, 9 (1896), 216.
39
Stokes, Manchester MP, 41 (1897), no. 15. Readers familiar with the so-called
wave theory of Christiaan Huygens may recognize that his theory was also one
of impulses. The drawings in Traite de la lumiere (Leiden, 1678), for example,
on p. 17, that appear to be snapshots of periodic wave crests are, in fact,
multiple exposures of the progress of an individual impulse.
propagation
Figure 2.1. G. G. Stokes' doubly directed electromagnetic impulse.
appreciable loss of energy, and consequently suffer little refraction

or reflection.
The impulse hypothesis also showed why x-ray diffraction had
not been observed: The effect was very small. To show this, Stokes
assumed that the pulse was doubly directed; that is, the total
positive excursion of impulse amplitude from equilibrium is
matched by the negative. Figure 2.1 shows the waveform schemat-
ically. The net integral of amplitude taken across the width of the
pulse, or over the time of passage, is zero.40 Thus, argued Stokes,
the net transverse impulse radiated toward a point P from that
portion of the pulse, shown in Figure 2.2, that lies far from the line
joining the source O to P, will be vanishingly small. Seen from an
oblique angle greater than </), the positive and negative displace-
ments will arrive at P almost simultaneously. Only a limited
portion of the pulse, delimited by angle </) in Figure 2.2, can make a
sensible contribution at P.
To fix the size of this active region Stokes employed standard
diffraction analysis, considering separately the effect at P due to all
segments of the impulse that lie within arcs of constant radius r
centered on P.41 (1) If r > PS, then either the transverse displace-
40
I have drawn a fairly symmetrical doubly directed impulse, but as long as the
area under the positive excursion equals that under the negative one, Stokes'
criterion is satisfied. The reasons why another pulse shape was incompatible
with the "dynamic theory of diffraction" became clear in 1912, as we shall see
in Chapter 8.
41
Stokes, TCPS, 9 (1849), 1-62. The final equation of section 22 is the general
relation for the transverse displacement at P due to an initial disturbance at 0.
Several small-angle approximations and symmetry about the azimuthal angle
reduce the complex double integral to a simple expression proportional to the
transverse displacement in each differential element dr d6.
Figure 2.2. Diffraction effect of Stokes' impulse.
ment is everywhere zero (i.e., to the left of, the right of, or fully
within the impulse) or its effect arises from the integral across the
full impulse. But according to Stokes' assumption, this integral
vanishes; viewed from the side, the net contribution at P will then
be zero. (2) If r < PQ, no part of the pulse can affect P. (3) The only
contribution at P comes from the lenticular shaded region for
PS > r > PQ. The total transverse displacement at P is just the
sum of the contributions from this region, attenuated in the ratio
f Of + r). Stokes showed that this region was quite small. For a
pulse of width icr6 in., and for reasonable values of r and f of 4 in.,
the radius of the active zone is only 0.002 in. The complete de-
crease in pulse intensity to the side of the shadow occurs in only
1/250 in., below the threshold of contemporary photographic
resolution.
J. J. THOMPSON ON PULSES
Joseph John Thomson, Cavendish Professor of Physics at Cam-
bridge University, carried out the detailed calculation of Stokes'
hypothesis. Thomson was an expert on Maxwell's theory. Further-
more, he was actively collecting proofs that cathode rays are
subatomic particles, having recently proclaimed them to be such.
The impulse hypothesis of x-rays, a corollary to the particulate
interpretation of cathode rays, was bound to attract his interest.
Thomson had considered the longitudinal wave hypothesis in an
early study.42 But he quickly recognized the merits of Stokes' idea
and, in December 1897, put it to a test.43
An electron is the source of isotropically diverging electric lines
of force. If the electron moves at a velocity that is small compared
to that of light, this set of lines remains symmetric as it moves
along with and centered on the electron. The accompanying mag-
netic field is arranged in circles lying in successive planes perpen-
dicular to the direction of electron motion. If the electron stops
abruptly, it takes some small time for that fact to be broadcast to
the surrounding field. The propagation of the dislocation outward
along the field lines constitutes the electromagnetic impulse. Fig-
ure 2.3 illustrates the growth of the spherical impulse at succes-
sive equal intervals of time after the electron stops.44 The center
marked "x" indicates the position the electron would have occu-
pied had it not stopped. The field outside the impulse remains
centered on "x," whereas the field inside diverges from the station-
ary electron.
Using Figure 2.4, one may easily calculate the ratio of tangential
to radial electric force. At time t after the impact, it will be equal to
the ratio
NP' vt sin 6
{2A)
NN7 J {2A)
where 6 is the thickness of the pulse and v is the original velocity of

the electron. The normal electric force in the radial direction is
e/4nr\ where e is the electronic charge. Thus, the total transverse
electric force amplitude is given by
AE ?A
(2.2)
Anrd c
where c = r/t is the velocity of wave transmission along the field
42
J. J. Thomson, PCPS, 9 (1896), 49-61.
43
J. J. Thomson, PM, 45 (1898), 172-83. Thomson mentioned in his Princeton
lectures in October 1896 that electromagnetic impulses should arise from
cathode ray impacts. Discharge of electricity (1898), 192-3.
44
Heaviside, PM, 27 (1889), 324-39; PM, 44 (1897), 503-12.
t=3
Figure 2.3. Electricfieldaround an electron t units of time after it stops

moving to the right. X marks the position the electron would have
occupied.
Figure 2.4. J. J. Thomson's derivation of radial to transverse electric

force for an impulse. [Based on J. J. Thomson, Conduction ofelectricity
(1903), 537.]
lines.45 The total magnetic force is given by ^ncAE) or
AH = v sin 6 (2.3)
ro
The point is that the transverse effect of the impulse drops only as
i/r, whereas the radial force drops as i/r2. Thus, the electrody-
namic effect of the impulse decreases less rapidly than does the
normal radial electric force, and will predominate at great dis-
tances from the electron.
In 1903, Thomson calculated the energy residing in one impulse.
In a given direction 6 from the direction of electron motion, the
energy density is
sin
J A 2 = ___ (2.4)
4TI ^" 4^ rS
AH is the magnitude of the magnetic force, and ft is the magnetic
permeability of the medium (in a vacuum, // = 1). Integrating this
expression over the volume of the expanding spherical shell gives
the total energy carried by the impulse:
2JU
F = (2.5)
^total 3S
45
In 1897 Thomson published only hisfindingson the magnetic force; those on
the electric force followed in 1903.
time
-5/2c +5/2c
Temporal width = 5/c

spatial width = 5
Figure 2.5. Singly directed square impulse as a function of time.
Note, in particular, that the energy is inversely proportional to the

impulse width.46 The narrower the pulse, the greater the energy
carried. Note also that Thomson's impulse is quite different in
form from Stokes' (Fig. 2.1). Expressed as a function of time,
Thomson's impulse is like that shown in Figure 2.5. The integral
taken across this pulse does not vanish. Stokes had stated that the
"dynamical theory" of diffraction would not allow a singly di-
rected pulse.47 Thomson was more concerned with the physical
constraint that the deceleration of source electrons occurs in one
direction only. He could see no way to justify a doubly directed
impulse, and never illustrated it as I have in Figure 2.5.
46
J. J. Thomson, Conduction of electricity (1903), 540.
sin 2 0 ,
And1 J j r2
Integrating over a hemisphere:
"* 2nr sin3 6
And1 Jof
*J 0
r1
f sisin3 Ode
rn
Jo
Since
-7T/2 2
s i n 3 M ? =
=
3' 3 J
47
Stokes, Papers, 5 (1905), 265-6.
By 1899, Kelvin too had accepted the impulse hypothesis. His
original concern had been that an extremely high-frequency longi-
tudinal wave was necessary to explain the low absorption of x-rays.
He analyzed the analogous case of oscillations in a mechanical
system to chart the relative amplitudes of the transverse and
longitudinal waves when the wavelength was about equal to the
molecular spacing in matter.48 He concluded that the amplitude of
the condensational wave might drop to negligible values for fre-
quencies above io12 per second.49 No explicit denial of the longitu-
dinal wave can be found in Kelvin's writings, but it is difficult not
to interpret his several attempts in 1899 to calculate the theoretical
properties of single, detached impulses as tacit approval of the
impulse hypothesis of x-rays.50
CONTINENTAL RESPONSE TO X-RAY

DIFFRACTION
Outside of Britain, reception of the impulse hypothesis was less
favorable and occurred more slowly. This resistance may be partly
explained by two views particularly characteristic of contemporary
German physics: Cathode rays were widely thought to be periodic
waves, not particles, and discontinuities in the aether could be
derived from the superposition of periodic waves. Cathode rays
induce fluorescence in several materials. This capability was
widely thought at the time to be restricted to light.51 When it was
discovered that x-rays originate from the fluorescent spot on the
tube, the analogy seemed complete - x-rays are periodic trans-
verse waves, like cathode rays, but of slightly lower frequency to
accord with Stokes' law of fluorescence. Some continental physi-
cists saw differences in the properties of x-rays and cathode rays as
a matter of degree only.52
There were, of course, some German physicists who rejected
Rontgen's longitudinal waves. But in the early years, most of these
48
[W. Thomson], PM, 46 (1898), 494-500; BAAS Report (1898), 783-7.
49
[W. Thomson], Baltimore lectures (1904), 159; BAAS Report (1898), 784.
50
[W. Thomson], PM, 47 (1899), 179-91, 480-93; 48 (1899), 227-36, 388-93.
51
There was other evidence against the material interpretation of cathode rays.
See Chapter 1; Hertz, AP, 79(1883), 782-816; Lenard, AP, 57 (1894), 225-68.
52
The most outspoken proponent of this interpretation was Philipp Lenard. See
Wheaton, HSPS, 9 (1978), 299-323.
preferred periodic transverse waves to aperiodic impulses. Conti-
nental texts often did not mention the impulse hypothesis.53 Ernst
Mach let it be known that he thought very little of the hypothesis of
"isolated impulses." He insisted that x-rays were ordinary light of
exceedingly short wavelength in order to explain the sharp-edge
shadows they cast.54 Yet partly on the basis of evidence for diffrac-
tion, physicists by 1900 would favor impulse x-rays, even on the
Continent.
Acceptance of the impulse hypothesis on the Continent was
hastened by increasing tolerance for a material interpretation of
cathode rays after 1897. Thomson's measurement of e/m for the
cathode beam was soon corroborated by Philipp Lenard and
Walther Kaufmann.55 By that time, there were strong theoretical
reasons to accept the result. The theory of the electron, in which
the charged particle was conceived by Lorentz and others as an
intermediary between ponderable matter and the electromagnetic
field, had borne fruit. With it Lorentz had given explanations of
Fizeau's experiment, of anomalous dispersion, of the Zeeman
effect, of Faraday's experiments on the rotation of the plane of
polarization in light, and of the Michelson-Morley experiment.56
The combination of theoretical significance and experimental
demonstration was hard to deny. By 1899 the localized concentra-
tion of electric charge, if not the material nature of the electron,
had become a respectable microreality, and its acceptance was
growing. When cathode rays, satisfying all the criteria of charged
particles, collide with atoms of the anticathode, sharply defined
electromagnetic dislocations should radiate into the surrounding
field. The question was only whether the impulses so produced
could be identified as x-rays.
It is indicative of the relation between these views that an early
electron theorist in Germany, Emil Wiechert, had been the first to
53
Examples include Donath, Einrichtung zur . . . Roentgen-strahlen (1899),
and Guillaume, Rayons x (1896), both of which claimed to discuss current
ideas about the nature of x-rays.
54
Mach, Zeitschrift fur Elektrotechnik, 7 (1896), 259-61.
55
The most important measurements were those of Kaufmann and Aschkinass,
AP, 62 (1897), 588-95; Lenard, AP, 64 (1898), 279-89.
56
Lorentz, Amsterdam Verslag, 7 (1899), 507-22. On the successes of electron
theory, see Hirosige, HSPS, 1 (1969), 151-209; McCormmach, his, 61 (1970),
459-97-
57
suggest the impulse hypothesis. In a talk to the Konigsberger
physikalisch-okonomische Gesellschaft in May 1896, six months
before Stokes' first public suggestion of impulses, Wiechert pro-
posed, but did not develop, his hypothesis oijdhen Lichtstossey or
irregular impulses of light.58 He was still equivocal; x-rays might be
impulses or high-frequency transverse waves. In either case, he
attributed their low absorption by matter to the inability of mole-
cules to react to the extremely rapid changes in the electromagnetic
aether.
In February 1899, two Dutch physicists announced that x-rays
could be diffracted.59 By itself the photograph submitted to support
this claim would not have been convincing. A rash of false claims
of diffraction had hardened most physicists, and the photograph
that C. H. Wind and H. Haga had taken was extremely difficult to
assess (see Figure 2.6). Working with his institute director at
Groningen, Wind exposed a wrapped plate to the x-rays passing
through a wedge-shaped slit, narrowing from 14 to 2 jum in width.
Exposures of up to 200 hours, extending over 10 days, were
required; the Erschutterungsfreiheit provided by the institute's
free-standing pier was essential. The investigators claimed that
they discerned evidence at the 9-jum width (marked - in the
figure) for broadening of the image. This they interpreted as
diffraction of periodic waves of approximately 1 A in length, io~4
times the wavelength of visible light, and a factor 100 times smaller
than Gouy's estimate. They provided very little photometric anal-
ysis, and the only justification for their claim was that Wind had
done this study after a careful analysis of most of the preceding
false evidence. He showed that prior claims for diffraction fringes
had resulted from optical illusions.60 Haga and Wind, unlike their
fringe-seeking predecessors, based their argument on evidence for
simple broadening of the image. They explained away all other
57
Wiechert, Konigsberg Schriften, 37 (1896), 1-48; AP, 59 (1896), 283-323,
especially 321-2.
58
"Irregular light blows." There is no question that Wiechert used jdhe in the
temporal sense, and he specified a sense for Stoss that made it clear that he
meant individual impulses.
59
Haga and Wind, Amsterdam Verslag, 7(1899), 387-8, 500-7.
60
Wind, Amsterdam Verslag, 7 (1899), 12-19. Among other results called into
question were those of Fomm, Munchen Sb, 26 (1896), 283-6.
Figure 2.6. Haga and Wind's photograph of x-ray diffraction by a slit.
Left: Exposure 5, collimation 25 /an, distance from slit to photoplate
1 cm, exposure 29 hr. No discernible diffraction. Right: Exposure 2,
collimation 14 /zm, distance from slit to photoplate 75 cm, exposure
100 hr. Broadening of the image appears at approximately 8 to 9 jum.
The dark region at the lower right is the image of a reference hole drilled
in one of the 0.5-mm-thick platinum plates that form the slit. It is
elongated in the right-hand exposure because of angular extension of
the x-ray source. [Haga and Wind, Amsterdam Verslag, 7(1899), facing
502.]
possible sources for the broadening and concluded that it must
have arisen from diffraction of x-rays.
SPECTRAL DECOMPOSITION OF AN IMPULSE

Wind and Haga's photograph was taken as proof of x-ray diffrac-
tion because it provided material for a new treatment of diffraction
proposed by Arnold Sommerfeld, then Privatdozent at Gottingen.
Sommerfeld had already shown himself to be an expert on diffrac-
tion; his Habilitationsschrift at Gottingen had been an attempt to
rectify the faults of KirchhofFs treatment of the subject, which, in
SommerfekTs view, failed "to satisfy the demands of mathemati-
cal precision."61 Nor was he unfamiliar with the impulse concept
of x-rays. As a student in Konigsberg, he had studied with Wie-
chert, whom he characterized as a "profound mathematical-
physical thinker."62 Indeed, Wiechert soon became the model of a
physicist for Sommerfeld, and they collaborated on, among other
things, a harmonic analyzer for complex waveforms in the years
immediately preceding Wiecherf s suggestion of impulse x-rays.63
Only a few years after leaving Konigsberg, Sommerfeld claimed
that the "Wiechert-Stokes" hypothesis of discontinuous x-ray
impulses followed from the "almost certain concept of the nature
of cathode rays according to which they consist of ejected electri-
cally charged particles of small mass."64 But the pulse hypothesis
presented a problem for diffraction analysis, one that Sommerfeld
thought would provide an ideal opportunity to test the flexibility of
his new rigorous theory.
Sommerfeld recognized that diffraction theory would fail for
single impulses. He was referring to the square pulse pictured in
Figure 2.5. This is the same pulse that Thomson had assumed. Like
Thomson, Sommerfeld ignored Stokes' warning that the integral
over the impulse must vanish. The result was that when Sommer-
feld decomposed the impulse into its Fourier component spec-
trum, he found that the components of greatest amplitude were
61
Sommerfeld, Gottingen Nachrichten (1894), 341; Deutsche Mathematiker-
Vereinigung Jahresbericht, 4 (1895), ll2~4\ Mathematische Annalen, 47
(1896X317-74.
62
Sommerfeld, Geist undGestalt, 2 (1959), 100-9.
63
Sommerfeld and Wiechert, Deutsche Mathematiker-Vereinigung Katalog
(1892), 214-21.
64
Sommerfeld, Zeitschrift fur Mathematik und Physik, ^6(1901), 11-12.
those for long, actually infinite, wavelengths. "If [however] we
regard the impulse as a whole, and not as if decomposed into single
sine waves," Sommerfeld argued, "all results become clearer."65
A Fourier series assigns a specific amplitude and a definite phase
relation to each member of a set of sinusoidal oscillations. The
frequency of each wave is a different integral multiple of a funda-
mental frequency. If the various waves are then superimposed,
physically added together, the resulting complex wave is fully
determined. It repeats its vibration at a rate equal to that of the
fundamental frequency. The Fourier series expansion is a straight-
forward mathematical method used to reduce any repeating com-
plex waveform to its component set of sinusoidal vibrations.
Complex waveforms that are not repeated are more difficult to
decompose, but conceptually are no different. The required set of
component sine waves is no longer limited to harmonic multiples
of a fundamental frequency. There is no fundamental frequency.
Instead, a Fourier integral must be taken over a continuous range
of frequency values. What results is, in fact, a spectrum; each value
of the frequency is associated with a definite amplitude. If this
extremely large number of sine waves, not harmonically related to
one another, is superimposed, the initial single impulse is ob-
tained. Everywhere else, the component waves cancel each other
out.
The general Fourier integral is given by the expression
f
J c
f(t)e~iwt dt (2.6)
where/(0 is the known functional dependence of the waveform on

time. In the case of the square impulse pictured in Figure. 2.5, the
contribution to the integral from the regions to the left and right of
the impulse is zero; hence the integral reduces to
r+s/2c . , A sin cot +S/2c sin coS/2c\
AEe~lcot dt = E = 2Af (2.7)
J-d/2c 0) S/2c (D -
If we now plot the absolute value of the integral as a function of the
positive values of frequency co, we obtain Figure 2.7. The height of
the function at any frequency gives the amplitude that must be
65
Sommerfeld, comments to Wien following VDNA (1900), part 2, first half,
p. 24.
Figure 2.7. Fourier spectrum of square impulse.
assigned to that frequency in order, by superposition of all waves,

to re-create the original square impulse. As may be seen, the
maximum contribution occurs at zero frequency, and the ampli-
tude for a wavelength corresponding to impulse width S is zero.66
SOMMERFELD S DIFFRACTION ANALYSIS

Sommerfeld characteristically presented his results in two ver-
sions: one for mathematicians interested in rigor, and the other for
physicists interested in x-rays.67 He spoke to the physicists through
the pages of the Physikalische Zeitschrift in 1899, and in person at
the seventy-second Naturforscherversammlung the following
year.68 As with any consideration of diffraction, Sommerfeld had
to contend with the original disturbance in interaction with its
scattered self. The planar impulse is assumed to approach the
screen shown in Figure 2.8 from the left, attenuating in amplitude
as it traverses space. The impulse scatters from the edge of the
66
Sommerfeld used the standard symbol for wavelength A to characterize the
impulse width, although he stated quite explicitly that no periodicity is to be
expected in the impulses. I will use the neutral symbol S throughout this study
to indicate spatial impulse width.
67
Sommerfeld's initial statement of the failure of classical diffraction theory
begins with two statements of the problem, one physikalisch ausgedruckt and
the other mathematisch ausgedruckt.
68
Sommerfeld, PZ, 1 (1899), 105-11; PZ, 2 (1900), 55-60.
screen
impulse
Figure 2.8. Sommerfeld's diagram of x-ray diffraction by an edge.
screen, and by Huygens' principle a second cylindrical impulse

results, centered on point O. Sommerfeld split it into the three
sections shown. In region I, the net disturbance consists of the
scattered impulse and the original impulse reflected from the
screen; the boundary between regions I and II is the edge of the
reflected impulse. Given the lack of evidence that x-rays are
reflected, Sommerfeld took that contribution to be zero. Region
III, in the shadow of the screen, contains only the cylindrical
scattered impulse, whereas in region II one must add to this the
undeviated original planar impulse.
Sommerfeld uncovered one of the more dramatic corollaries of
the fact that impulses are not periodic waves: No diffraction fringes
result. The periodic structure in ordinary light that produces the
fringes is not present. In effect, one obtains only the envelope of the
classical diffraction pattern: a monotonic decrease in intensity to
the side of the shadow. Sommerfeld's three-dimensional represen-
tation of the resulting distribution of x-ray intensity is shown in
Figure 2.9. The maximum intensity in the forward direction peaks
at one-half the amplitude of the unscattered original impulse. The
smaller the ratio of pulse width to the distance r from the screen,
the more quickly this intensity drops off to the sides. This is shown
The electromagnetic impulse hypothesis ofx-rays 37
Figure 2.9. Distribution of diffracted impulse x-ray intensity. [Sommer-

feld,/>Z, 7(1899), 108.]
Figure 2.10. Change in diffracted intensity with impulse width. [Based

on Sommerfeld, PZ, 1 (1899), 109.]
in Figure 2. io. The smaller the pulse width, the less the discernible
diffraction.
To relate the calculation to the results from Groningen, Som-
merfeld had to introduce approximations. He needed to convert
electromagnetic amplitude to a measure of photographic blacken-
ing. Then again, Wind and Haga's data came from slit diffraction;
the introduction of the second edge, a problem Sommerfeld had
not solved completely in his 1896 study, introduced error. As an
approximation, he took the total broadening of the image as due
half to each side. The result was a calculated x-ray pulse width of
33 A, compared to the Dutch estimate of a i-A periodic wave-
length.
In the second phase of his study, Sommerfeld the physicist
retracted what Sommerfeld the mathematician had done. After
some discussion with Wind on sources of error in the photograph,
he revised his calculation.69 This time he reworked the problem
explicitly by taking both edges of the slit into account. "Refraction
from a slit," he said, "often turns out very different from refraction
at an edge."70 He now found that the "dark core" of the wedge-slit
image, his interpretive sketch of which is shown in Figure 2.11,
would cease at a slit width of \f$Sr/io. Wind's best estimate was
that this occurred at 9 //m; Sommerfeld derived a pulse with a
spatial width d= 1.3 A and a temporal width of 4 x io~19 sec. He
thought this a "much more trustworthy" estimate. To his friend,
the astrophysicist Karl Schwarzschild, an expert on photochemical
response, Sommerfeld wrote, "with such a small [pulse width] it is
easy to understand why x-rays are neither absorbed nor refracted.
Molecules simply cannot follow such short impacts."71
Sommerfeld went on to propose that a continuum of radiation
waveforms exists, one that extends from the purely periodic sine
waves of monochromatic light to the temporally localized aperi-
odic impulses of x-rays. There would generally be a mixture of
pulse widths in an x-ray beam. He had proven to his satisfaction
that x-ray impulses are aperiodic; the lack of characteristic diffrac-
tion fringes was evidence for their intrinsic temporal discontin-
uity.72 Thus, Sommerfeld had a physical if not a formal reason to
reject the consideration of an impulse as if broken into its Fourier
spectrum, as others were suggesting. Wind, for example, was
surprised to have his photograph successfully analyzed in terms of
aperiodic pulses; recall that he had first claimed the diffraction
69
Details of Wind's criticism exist in his several letters to Sommerfeld between
April and August 1900. Sommerfeld papers; AHQP 34, 14.
70
Sommerfeld, PZ, 2 (1900), 55.
71
Sommerfeld to Karl Schwarzschild, 24 September 1900. Schwarzschild papers.
72
Discussion following Wind's presentation to the seventy-second Naturfor-
scherversammlung, PZ, 2(1900), 297-8.
9 junn
Figure 2.11. Schematic distribution of x-ray impulse intensity after

diffraction by a slit. Width exaggerated 100-fold. [Sommerfeld, PZ, 2
(1900), 293.]
photo as evidence of periodic-wave x-rays. He responded to Som-

merfeld's careful analysis that the "wavelength" of i A should be
interpreted as the high-amplitude component of the Fourier series
that represents the impulse.73
Sommerfeld would have none of it. Wind's claim could not be
true for the impulses Sommerfeld had treated; their maximum
Fourier component fell at much longer wavelengths, and their
contribution at the wavelength comparable to the pulse width was
zero. At the conclusion of his second study of pulse diffraction, the
one for the mathematicians, Sommerfeld contrasted his method to
that of Wind. "One cannot speak of the superiority of one or the
73
Wind, AP, 68 (1899), 896-901; PZ, 2 (1900), 189-96.
other method," he remarked; "each has its strengths and weak-
nesses." Wind's method offered "great generality and adaptability
to experimental conditions"; Sommerfeld's provided "conceptual
clarity and precision of results."74 For Sommerfeld, the difference
between an impulse and its Fourier representation had assumed a
physical as well as a formal significance. There was never any
question in his mind that mathematically the two representations
were equivalent. But with the idea that a continuum exists extend-
ing from periodic light to aperiodic x-rays, Sommerfeld had intro-
duced what would become an influential conceptual framework,
one that emphasized the distinction between x-ray pulses and
ordinary light.
THE NATURE OF WHITE LIGHT

Whether a discontinuous impulse is better represented as an iso-
lated event or as a superposition of harmonic oscillations had been
discussed before; it was clear that differences of opinion cut across
national boundaries. The question was, which waveform best
represents polychromatic white light? L. Georges Gouy had con-
cluded in 1886 that there is no sense in claiming that white light is a
repeating wave.75 It is the superposition of all colors of light, and as
such it cannot have any well-defined wavelength. He thought that
the actual waveform consisted of irregular impulses in the aether.
It had been claimed that interference experiments give a measure
of the degree of periodicity of white light: A beam is split and
reunited after one component travels a great number of wave-
lengths. The resulting interference fringes, it was supposed, prove
that the light repeats its oscillation even after these long intervals.
Gouy maintained that the regularity is introduced entirely by the
spectral analyzer. A diffraction grating separates out of the light
only the "different simple movements" that until then had a
purely analytical existence.76 Gouy insisted that we should think
only of the total electromagnetic disturbance, the mouvements
lumineux, which arrive at the eye. The pure frequencies are only
analytical tools out of which we fashion the concept of the physical
disturbance.
74
Sommerfeld, Zeitschrift fur Mathematik und Physik, 46 (1901), 9711.
75
Gouy, JP, 5 (1886), 354-62. Righi, JP, 2 (1883), 437-46.
76
Gouy, ibid., 362.
The light we see from most objects is a mixture of pure frequen-
cies. Green light, for example, is a superposition of many frequen-
cies of light, which are predominantly but not exclusively greenish
to the eye. By adding each pure frequency, with its proper ampli-
tude and phase, to all the others, we arrive at the complex distur-
bance that is presented to the eye. According to Fourier analysis,
the complex wave so constructed can have virtually any functional
form. If it is a periodically repeating disturbance, it is repeated less
rapidly than is any one of the sine waves of which it is composed.
Its periodic frequency is the least common factor of its component
frequencies.77
Polychromatic light contains a virtual continuum of frequen-
cies. As one constructs the complex wave, the least common factor
of the frequencies quickly drops to an exceedingly small value.
Consequently, the period of the resultant complex wave grows
very long; for all practical purposes, it becomes infinite. It is, in a
sense, the human attempt to assign quantitative measures that
creates the conceptual problem. It is, in fact, just as much of a
problem for the representation of green light, but because white
light cannot be assigned a perceived frequency, the problem can-
not be avoided. If we insist that there is a one-to-one relation
between numbers and pure frequencies of light, then the colors
added to make white light can no more share a commensurate
wavelength than can all points on the number line share a com-
mon factor. The "frequency" of the resultant wave is zero, and
there is no meaning to the concept of regularity. As Lord Rayleigh
said, it is truly "nonsense."78
But in order for any wave to have a well-defined frequency, it
must also have a long temporal extension. A piece of a sine wave
does not possess a well-defined frequency; only a pure oscillation
that extends to infinity in space can strictly be called monochro-
matic. It is not simply the rate at which a wave repeats, but also the
actual number of oscillations it undergoes, that define a mono-
77
Counting frequencies limits the apparent discontinuity. Assigning each pure
color a number implies that some small commensurate frequency unit exists
that relates any two frequencies. In this sense all frequencies are "harmonic,"
although the fundamental is exceedingly small. In principle, unless frequencies
are irrationals, a totally isolated impulse cannot arise, according to Fourier
analysis.
78
[J. W . Strutt], PM, 27 (1889), 463; Encyclopaedia Britannica, 24 (1888),
421-59-
chromatic beam. Passing an impulse through an analyzing appa-
ratus sufficiently extends it so as to make interference of its compo-
nent wavelengths possible. In 1894 Arthur Schuster showed that
both a diffraction grating and a prism will impose their own
periodicity when stimulated by a discontinuous disturbance such
as white light. "There is no distinction between regular and irregu-
lar light," he said, "beyond that which is brought out by the
distribution of intensity in the spectrum."79
Following Stokes' suggestion of x-ray impulses in 1897, similar
issues were raised in regard to x-rays. Lord Rayleigh, an expert on
acoustics and Fourier analysis, asked in a note to Nature why
Stokes maintained that x-ray pulses were physically different from
periodic waves.80 One need only ascribe to them an extraordinarily
long repetition period, say on the order of days; they can then be
reduced to simple harmonic sums. G. Johnstone Stoney pro-
ceeded to do just that, and concluded that this "represents what
really takes place in nature."81 A year later, Stokes wrote to Ray-
leigh, asking in disbelief if he had been serious in his assertion that
individual pulses were broken into harmonic waves by the action
of a spectroscope.82 Rayleigh replied, yes; "whatever be the distur-
bance incident upon a grating, the light sent off in a particular
direction will have been made regular."83
The repeated objections and continued discussion of this ques-
tion are puzzling to the modern reader. Mathematically, the two
representations of an impulse are equivalent. What was at stake
was the physical, conceptual vision. Does it in fact mean anything
to think of pulses as the resultant of an extremely large number of
waves extending much farther in space than the impulse alone?
Emmanuel Carvallo tried to show in 1900 that impulses cannot
physically be represented by Fourier expansion.84 Max Planck, still
seeking a complete physical interpretation of the new constant he
had introduced in his thermodynamic interpretation of cavity
radiation, saw a parallel in 1902 to his own work.85 He argued that a
79
Schuster, PM, 57(1894), 509-45. Larmor, PM, 70(1905), 574-84.
80
[J. W. Strutt], Nature, 57 (1898), 607.
81
Stoney, PM, 45 (1898), 532-6; 46 (1889), 253-4.
82
Stokes to Rayleigh, 31 July 1899. Stokes, Memoir, 2 (1907), 123-4.
83
Rayleigh to Stokes, 10 August 1899 and 12 February 1900. Stokes, Memoir, 2
(1907), 124-5.
84
Carvallo, CR, 130(1900), 79-82,130-2. Gouy,C7?, 750(1900), 241-4,560-2.
85
Planck, AP, 7 (1902), 390-400.
spectroscope could no more separate the energy of a single partial
oscillation from white light than could experiments isolate the
kinetic energy of a single gas molecule. Joseph Larmor suggested in
1905 that white light differed from x-rays in that the "molecular
shocks" of white light "would be regarded as having some sort of
statistical order in their distribution," unlike x-ray impulses.86
The differences of opinion about the physical nature of pulses
symbolized an essential ambiguity that gave electromagnetic im-
pulses their intrinsic adaptability. The impulses that Stokes had
proposed, and that Sommerfeld imagined to pass through a
wedge-shaped slit, had no periodic properties. They did not occur
in regular succession; if they had, then diffraction fringes should
have been observed. Impulses are waves. But these waves transport
energy in much the same way as do particles. In part due to this
flexibility, in part due to electron theory, and in part due to
Sommerfeld's successful diffraction analysis, the hypothesis of
impulse x-rays was soon as widely accepted on the Continent as it
was in Britain. Early discussions, like that in the 1897 edition of
Wiillner's Lehrbuch der Experimentalphysik, either neglect the
pulse hypothesis entirely or make only implicit concession to it.
Wiillner described x-rays as "undoubtedly a process in the aether,
as are the cathode rays," then ranked them as ordinary light but
with a frequency higher than that of ultraviolet light.87 But by 1902,
Lorentz described x-rays as "extremely short violent electromag-
netic disturbances of the aether,"88 and by 1903 the impulse hy-
pothesis had virtually everywhere supplanted the view that x-rays
are periodic waves.89
C. G. BARKLA AND X-RAY POLARIZATION

Even in the absence of any demonstration of x-ray polarization,
there was very little doubt after 1898 that the form of an x-ray
aether disturbance is transverse. In 1904 this long-anticipated
experimental result was finally demonstrated in Britain, adding
indirect support to the pulse hypothesis. There was never evidence
86
Larmor's editorial comment, Stokes, Papers, 5 (1905), 27211.
87
Wiillner, Lehrbuch der Experimentalphysik, 3 (1897), 1300-02.
88
Lorentz, untitled Nobel lecture. Nobel 1902 (1905), separately paginated, p. 4.
89
Among examples of the pedagogical acceptance of impulse x-rays is
Khvol'son, Lehrbuch der Physik, 2 (1904), 176.
that x-rays can be reflected. However, early studies had turned up
what was called a diffuse reflection of x-rays from certain solids.90
Rontgen had mentioned in his third and final x-ray study that a
fluorescent screen might still glow even when shielded from the
direct x-ray beam.91 He concluded that x-rays can stimulate other
x-rays in matter; the primary x-ray beam gives rise to secondary
x-rays. Sagnac showed in 1898 that the secondary x-rays from
irradiated metals differ from the primary by being more easily
absorbed.92 Further, their penetrability seems to depend, for the
same primary beam, on the type of material used to produce them.
Sagnac based his research on a modified periodic-wave hypoth-
esis of x-rays.93 There was some question of whether the impulse
hypothesis was also able to explain the result. In the first edition of
his book Conduction of electricity through gases, J. J. Thomson
answered in the affirmative. Secondary impulses should radiate
from electrons in the solid. Each primary x-ray pulse gives rise to a
large number of secondary pulses, one from each charged particle
passed over. Not only might ionic forces in the solid constrain the
individual secondary pulses, but the multiplicity of secondaries
would be expressed as a semiperiodic wave with a length equal to
the spacing of the radiating charged particles. This, Thomson
suggested, could cause as much difference as existed "between the
sharp crack of lightning and the prolonged roll of thunder."94
Nevertheless Thomson gave no quantitative argument to show
why the penetrability of x-ray impulses would be less in the
secondary rays. This problem was especially difficult in the case of
irradiated gases, in which ionic bonding would not cause impulse
degradation and in which the radiating electrons are spaced rela-
tively far apart.
Enter Charles Glover Barkla. Schooled in Liverpool by Oliver
Lodge, in 1899 Barkla went to Cambridge, where he heard lectures
by Stokes and learned experimental physics from Thomson. Not
surprisingly, he became an early advocate of the pulse hypothesis
and, before leaving Cambridge in 1903, had begun studies on
secondary x-rays. His first experiments dulled Sagnac's criticism;
90
[W. Thomson], PRS, 59(1896), 332-3. J. J. Thomson, PCPS, 9(1898), 393-7.
91
Rontgen III, p. 18.
92
Sagnac, CR, 125 (1897), 942-4; 126 (1898), 467-70, 521-3.
93
Sagnac, Theses (1900), 163.
94
J. J. Thomson, Conduction of electricity (1903), 270.
Oscillations at A
only in plane normal
to beam 1.
Oscillations at B
only perpendicular
to beams 1 and 2
Beam 3 is plane- As seen at A

polarized.
As seen at B
Figure 2.12. Schematic diagram of Barkla's demonstration of x-ray

polarization.
the penetrability of secondary x-rays produced in five gases was,

within io percent, the same as that of the primary. He concluded
further that "there is obviously a proportionality between the
intensity of secondary radiation and the density of the gas."95 This
supported the impulse concept in Barkla's view. The greater the
number of electrons shaken in an atom by the primary pulse, the
greater the strength of the expected secondary pulse.
In the course of these studies, Barkla found that a weak tertiary
x-radiation could be discerned when the secondary x-rays were
directed against some metals. This led to a test for x-ray polariza-
tion. A colleague at Liverpool, Lionel Wilberforce, suggested that
this tertiary beam should be plane-polarized if the initial impulses
were transverse. The electrons in scattering substance A can be
displaced by the primary beam in all directions except that of the
primary beam itself. This follows immediately from the transverse
nature of the pulses. Those secondary rays that propagate at right
angles to the primary beam should then be able to excite displace-
ment of electrons in scatterer B only in a direction perpendicular
to both primary beam i and secondary beam 2 (see Figure 2.12).
The resulting tertiary beam is thus plane-polarized and should
95
Barkla, PM, 5 (1903), 696.
show a distinct intensity minimum when viewed in a direction
orthogonal to both beams i and 2.
Barkla found that the tertiary beam was too weak for the experi-
ments. But he quickly realized that the secondary beam should
show the same effect, because the monodirectional cathode ray
beam has already partially polarized the primary x-rays. Thomson
had pointed out in 1898 that the unidirectional cathode beam
caused the pulse-producing deceleration to occur in only one
direction.96 He had predicted a factor difference of 2 in x-ray
intensity radiated in the line of the cathode beam compared to that
seen at right angles to the beam. Experiment disproved the predic-
tion, and Thomson attributed thisfindingto multiple scattering of
cathode ray electrons as they collide with atoms of the anticathode.
Measuring the more completely polarized secondary beam gave
Barkla the experimental edge, and he was rewarded with success in
1904.97 He rotated the x-ray tube itself about the axis defined by the
primary x-ray beam 1 and found a 15 percent difference in the
ionizing power of secondary x-rays scattered from paper. The
result sparked no controversy; as we have seen, there was little
opposition to the transverse nature of x-rays by 1904. Although the
result did not in fact distinguish between periodic waves and
aperiodic transverse impulses, it was widely accepted as further
support for the impulse hypothesis.98
ERICH MARX MEASURES THE VELOCITY

OF X-RAYS
The high point of the seventy-seventh Naturforscherversammlung
in Meran was the presentation of the first convincing attempt to
measure the velocity of x-rays. It had been tried before, notably by
Rene Blondlot, but the results were ambiguous.99 The method
outlined by Erich Marx at Meran seemed to answer all criticism.100
96
J. J. Thomson, PM, 45 (1898), 172-83.
97
Barkla, Nature, 69 (1904), 463; PTRS, 204A (1905), 467-79. Two years later,
Barkla succeeded in showing polarization of tertiary x-rays. PRS, 77A (1906),
247-55.
98
See, for example, Thomson, Conduction of electricity, 2nd ed. (1906), 322.
99
Blondlot, CR, 135 (1902), 666-70, 721-4, 763-6, 1293-5. An early unsuc-
cessful attempt was made by Brunhes, CR, 130 (1900), 127-30.
100
Marx, VDpG, 7 (1905), 302-21; PZ, 6 (1905), 768-77; expanded in AP, 20
(1906), 677-722. The use of electric waves as a comparison had been made for
the velocity of cathode rays by Des Coudres, VDNA, 14 (1895), 86-8; 16
(1897), 157-62. See also Wiechert, AP, 69 (1899), 739-66.
-S
Fl
Figure 2.13. Marx' apparatus to measure x-ray velocity. [Franck and

Pohl, VDpG, 70(1908), 118.]
His apparatus is shown in Figure 2.13. A high negative potential on

the curved cathode of x-ray tube R causes a burst of cathode rays
that impacts at the anode. X-rays are released and travel to elec-
trode P of the lower tube. At the same instant, a negative electric
impulse is induced in wire B, which is attached directly to P
through the sliding contact on the right.
By 1905 it was well known that x-rays will stimulate the release
of electrons from metals. If P is negatively charged when the x-rays
arrive, electrons will be accelerated to the Faraday cage F and
thence to the electrometer. This condition is met only if the
negative electrical impulse from the wire arrives at P simultane-
ously with the x-rays. Sliding the shunt wire on the right changes
the time required for the electric impulse to arrive at P. Once he
had optimized the shunt for maximum electrometer response,
Marx found that the length of wire B had to change in step with
any alteration of the distance between the discharge tubes. Marx
asserted that the x-ray velocity was, within 5 percent, the same
as that of the electrical impulse in the wire. It was a common mis-
conception at the time that electrical signals travel through a wire
at the velocity of light. For short, uninsulated wires the agree-
ment is very close. Thus, Marx concluded that x-rays also prop-
agate at that velocity. His result was not without opponents, but
they were few in number and of little influence.101 A strong piece of
evidence supporting any aether-wave hypothesis of x-rays had
been provided. Although this result, like Barkla's, favored neither
the pulse nor the periodic-wave hypothesis, it was widely taken as
yet more evidence for the former.
By 1905 the impulse hypothesis had achieved virtually unani-
mous support. There was still a current of thought in Germany
that interpreted pulses as if they were constructed of periodic
waves. In Chapter 5, we shall see that this was an important
influence on early quantum treatments of x-rays in Germany. But
in one form or another, the impulse hypothesis of x-rays achieved
supremacy in the period 1903-5. The views of H. A. Lorentz, who
was widely respected as a consummate theoretical physicist, may
be taken to represent those of most scientists. He taught in 1907
that x-rays are "an irregular succession of impacts, each of which
persists for a much shorter [time] than does one oscillation of the
farthest ultraviolet light yet observed."102
101
The most persistent critic was Bernhard Walter, who had a material interpre-
tation of x-rays. Fortschritte aufdem Gebiete der Rontgenstrahlen, 9 (1905),
223-5. After 1908, Walter's criticism began to find support; see Chapter 6.
102
Lorentz, Lehrbuch, 2 (1907), 168.
The analogy between y-rays and x-rays
The y-rays are very penetrating Rontgen rays [produced

by] the expulsion of the /? . . . particle.1
The discovery of x-rays reduced much natural reluctance on the

part of physicists to report other discoveries. The most significant
new claim, one recognized within weeks of Rontgen's announce-
ment, was the discovery of radioactivity by Henri Becquerel.2
Certain minerals were found to emit rays spontaneously. Not even
the electrical potential difference necessary for the production of
cathode rays and x-rays was required. No human artifice or in-
tervention was needed to produce the Becquerel rays; this was
always their most significant and unique property. For this reason,
it was soon widely believed that the rays were the residue of
spontaneously disintegrating atoms.
However distinct the physical origin of Becquerel rays was from
that of x-rays, telling similarities between the two new emanations
were quickly recognized. Both could penetrate opaque paper to
expose a photographic plate. Like x-rays, Becquerel rays were first
discovered in this way. They seemed to be another invisible radia-
tion, rendered perceivable only through the superhuman sensitiv-
ity of photographic plates. It was soon shown that x-rays and
Becquerel rays both induce electrical conductivity in gases through
which they pass. In fact, as we have seen, Becquerel's rays were first
1
Rutherford, Nature, 69 (1904), 436-7.
2
Becquerel, CR, 122 (1896), 501-3. Fortschritte der Physik, the abstracting
journal of the Deutsche physikalische Gesellschaft, first listed research on
Becquerel rays under "Optics" in the section for "phosphorescence and fluore-
scence." A new subheading in optics for Becquerelstrahlen was added in 1900.
Rontgenstrahlen were, from 1896 on, listed as a separate subhead under "elec-
tricity and magnetism." Another reorganization in 1903 put Becquerel rays,
together with cathode and x-rays, under "electricity and magnetism."
thought to fill the gap that separated x-rays from ordinary light.
Becquerel claimed that they could be refracted and polarized;
although soon disproven, these claims served to solidify the view
that Becquerel rays resembled light more closely than did x-rays.
For these reasons, Becquerel rays were early on identified as a
form of radiation in the aether. Questions were never raised about
the generic terms rays, Strahlen, and rayons used to describe them.
Minerals that emitted them were called radioactive by the Curies
in 1898; radio, from the Latin radix or radius, was selected to
emphasize the isotropy of emission and to distinguish it from the
ambiguous rayon, which in French means both ray and radius?
Ernest Rutherford began his lifetime research on radioactive
phenomena with a study of the conductivity that uranium rays
induce in gases. He agreed that the causative agent was likely a true
radiation, a view unaltered by his disproof of Becquerel's claims
for refraction and polarization.4 In the course of his first investiga-
tions, he found that a sheet of paper or metal foil reduced the
ionizing effect of the rays considerably, but additional layers
brought no comparable decline. He concluded that two species of
ray, one easily absorbed, the other more penetrating, existed side
by side; the first he called a to distinguish it from the more
penetrating ^-component. He then compared the penetrating
/?-rays to x-rays incident on a metal and described the a-rays in an
analogy to the diffuse, and less penetrating, "reflected" x-rays.5
The analogy of x-rays and Becquerel rays influenced other
discoveries as well. On the heels of Sagnac's demonstration in 1898
that x-rays stimulate the release of electrons from metals, the
Curies searched for the same behavior in Becquerel rays. They
found that a form of radioactivity could be induced in matter; an
ionizing activity was detected that, unlike the secondary emission
of electrons, persisted even after the primary irradiation ceased.6
But this difference appeared to be a consequence of the spontane-
ous nature of radioactive decay; it did not alter the opinion that the
uranium rays were an unusual form of light.
3
P. and [M] Curie, CR, 727(1898), 175-8.
4
Rutherford, PM, 47(1899), 109-63.
5
The "diffusely reflected" x-rays were largely secondary electrons, but this was
not known at the time. See Chapter 4.
6
P. and [M.] Curie, CR, 729(1898), 714-16.
The analogy between y-rays and x-rays 51
TRANSFORMING WAVES INTO PARTICLES
It came, therefore, as a surprise to many when, in 1899, Becquerel
rays were first deflected by a magnet. They bent in precisely the
manner expected of negatively charged particles.7 Soon afterward,
it was demonstrated that the ray trajectories also bend in an
electrostatic potential.8 Part of the credit for this recognition must
go to the new and more intense sources of radioactivity, polonium
and radium, provided by the Curies in 1898. Credit also must go to
the widespread interest, particularly in Germany, in techniques for
the study of cathode rays by magnetic deflection. In any event, the
nature of the Becquerel rays was now in dispute. As one reviewer
put it, "not content with attributing to Becquerel rays all the
properties of x-rays, physicists and chemists have found that these
rays also possess properties hitherto attributed to cathode rays
alone."9
Pierre Curie showed that Rutherford's division of Becquerel
rays into a- and ^-components was also expressed in their mag-
netic deflection. Only the /^-component seemed to bend. The
a-rays traveled on, apparently unaffected.10 A photoplate, exposed
to uranium rays that had passed through a magnetic field, thus
showed two exposed regions. Rutherford's division of Becquerel
rays into components offered the first hope of systematic study of
uniform properties. No longer were the seemingly variable proper-
ties of the rays approachable only through the heroic chemical
processes of the Curies, Friedrich Giesel, and others. Around 1899,
an effluorescence of research interest occurred in this field, which
was increasingly referred to as radioactivity.11 Figure 3.1 shows the
number of papers on related topics for the early years cited in an
authoritative review by Max Ikle.12
7
Three independent demonstrations were given. The earliest was by Friedrich
Giesel at Braunschweig, AP, 69 (1899), 834-7, following strong hints by Julius
Elster and Hans Geitel. See also S. Meyer and von Schweidler, PZ, 1 (1899),
90-1, 113-14; CR, 729(1899), 996-1001.
8
Dorn, Halle Abhandlungen, 22 (1901), 39-43, 47-50.
9
Stewart, PR, 11 (1900), 155.
10
P. Curie, CR, 130(1900), 73-6.
11
Becquerel in Guillaume and Poincare, Congres international, 3 (1900), 47 - 78.
Crookes, PRS, 66 (1900), 409-23.
12
Ikle, JRE, 1 (1904), 413-42 is more complete than Wilson, Nature, 70 (1904),
241-2.
250-

200- 1
150- I
100-
/*
50 H
f f \i 1 f < 1 1
1896 97 98 99 1900 01 02 03
Figure 3.1. Number of studies of radioactivity, 1896- 1903.
By 1900 it was clear that the /?-rays were high-speed electrons.

Measurement of their charge-to-mass ratio (e/m) yielded values on
the order of io7 emu/g. Their velocities were found to be on the
order of io10 cm/sec, higher even than those attained by cathode
rays.13 Reaction to the identification was mixed. Rutherford and
Thomson both found it difficult to reconcile the great penetrability
of the /?-rays with the known properties of the lower-speed corpus-
cles.14 R. J. Strutt, the future fourth Lord Rayleigh, undertook a
critical reappraisal of x-rays in 1900 to detect a comparable mag-
netic deflection. He was not able to find one.15
Although by 1900 the nature of /?-rays had been brought to light,
the same was not true for a-rays. Since a-rays seemed to be
unaffected by a magnetic field, studies were limited to charting
their rapid absorption by matter. Pierre Curie showed in 1900 that
any a-ray from radium and polonium lost its ability to ionize after
passing through a layer of air whose thickness was characteristic of
the radioactive atomic source, that is, independent of the chemical
13
Becquerel, CR, 130 (1900), 372-6, 809-15. See also Kaufmann, Gottingen
Nachrichten (1901), 143-55.
14
[R. J. Strutt], Life of Thomson (1942), 132-33.
15
R. J. Strutt, PRS, 66 (1900), 75-9.
16
compound in which the element was framed. With a qualitative
limit thus placed on the penetrability of the undeviated a-rays, the
stage was set for the discovery of a third component of Becquerel
rays - one that, although not deviated by a magnet, was nonethe-
less more highly penetrating than the /?-rays.
HIGHLY PENETRATING RAYS

Paul Villard, a physicist at the Ecole Normale Superieur in Paris,
was a dedicated investigator of cathode rays.17 His interest had
been sparked by study of the chemical effects of both cathode rays
and x-rays. Convinced that cathode beams do not actually pass
through metal foils, as had been claimed by Hertz and Lenard in
Germany, Villard fabricated the implausible hypothesis that cath-
ode rays incident on metal foil stimulate the release of other
electrons from the back side of the foil.18 To test this quaint idea, he
developed a research technique of firing cathode beams along the
plane of a wrapped photographic plate. In this way, one obtained a
record of the path followed by the beam both before and after it
collided with the metal foil he oriented at right angles to the plate.
This rather recondite research took on new significance in 1900
when Villard added the /?-rays from radium chloride to his stock of
cathode beam sources. He scattered the beam off a 0.3-mm-thick
aluminum sheet and obtained a photograph that he interpreted as
showing a "refracted" secondary beam. This, he claimed, had
emerged from the back side of the aluminum in a direction normal
to the surface of the metal regardless of the angle between the foil
and the primary cathode beam. This extravagant but spurious
finding was accompanied by more. "I almost always observed," he
reported to the Paris Academy in April 1900, "that a straight beam
was superimposed on the refracted beam . . . I think it due to the
presence of non-deviable rays of lower absorptivity than those
described by M. Curie."19
Villard corroborated his claim by showing that the new pene-
trating rays proceed in their straight course even while a magnetic
16
P. Curie, CR, 130 (1900), 73-6.
17
Villard, Rayons cathodiques (1900); and contribution in H. Abraham and
Langevin, Ions (1905), 1013-28.
18
Villard, CR, 130 (1900), 1010-12.
19
Ibid., 1012.
field perpendicular to the plate curves the /?-beam to the side.
Thus, the straight rays could not be the ^-component. Nor could
they be the ocomponent; far from being stopped by a few centi-
meters of air, they passed right through the metal foil, not to
mention the 18-cm depth of edge-on photographic emulsion. Soon
the find was corroborated by Becquerel in the context of denounc-
ing Villard's wild claim that /?-rays are refracted by the foil.20
Villard himself was somewhat less than pleased about the whole
matter because the new rays interfered with his experiments on
pure cathode beams. Before he abandoned radium as a cathode
beam source, he tried to isolate the properties of what he called the
"x-rays from radium," using a strong source borrowed from the
Curies for the purpose. He showed that the rays passed through
thicknesses of glass and lead that would stop all ^-particles (and, of
course, all a-rays too). Further, they would expose more photo
plates piled in a stack than would either the a- or the ^-compo-
nent.21
Villard was not alone in his lack of interest in the new rays; very
little effort was expended on their study by anyone in thefirstfew
years. One reason was that the discovery came in the midst of
attempts to obtain accurate e/m measurements of the ^-particle. In
addition, the new component was exceedingly weak. Using ioni-
zation techniques, Curie had not discerned its existence in his
careful studies just prior to Villard's photographic experiment.
Rutherford estimated that the new undeviable rays from radium
had only i percent of the ionizing effect of the accompanying
/?-rays and only o.oi percent of that of the a-rays. Finally, there
were only two known sources of the rays: thorium and radium.
Thorium produced them at a low rate, in keeping with its generally
lower activity compared to radium. To study the nondeviable,
penetrating rays, one therefore had to have radium, a commodity
to which only a handful of physicists had access in the years
1900-5. Although Becquerel, for example, made repeated refer-
ences to the existence of the nondeviable rays between 1901 and
1903, he studied their properties only once.22 He showed a curious
effect: They darkened a photographic plate more strongly on the
20
Becquerel, CR, 130 (1900), 1154-7.
21
Villard, CR, 130 (1900), 1178-9.
22
Becquerel, CR, 132 (1901), 1286-9; 134 (1902), 208-11; 136 (1903), 431-4.
23
half covered by a metal sheet. The problem in studying the
penetrating rays, he remarked, was "the difficulty one has in
stopping them [at all]."24
The real topic of research in radioactivity, one that overshad-
owed any questions about the new rays found by Villard, was the
continuing effort to find clues to the nature of orays. Many
physicists now doubted that these were rays because of their lower
penetrability relative to the ^-particles. The case for waves had
always been weakest for a-rays. "The theory that [both compo-
nents in] Becquerel rays consist of extremely short ultraviolet light
waves is now almost abandoned," wrote Oscar Stewart in 1900.25 A
great effort was made, particularly by Rutherford, to demonstrate
magnetic deflection of the a-component. In late 1902, he suc-
ceeded.26 Previously, Rutherford had looked briefly at the as yet
unnamed "highly penetrating" rays from radium. He stopped all
a- and /?-rays with 1 cm of lead, and then tested what remained in
the very strong magnetic field he was preparing for his assault on
the a-rays. He found no deflection of the third component.27
Rutherford had completely changed his view of the nature of
radioactive products. By 1902 he was convinced that all three
components were streams of particles. The ionization produced in
gases by the undeviated rays showed, he claimed, that they "have a
closer resemblance to kathode than to Rontgen rays."28 The
rate at which they formed ions in a gas was closer to the rate that
had been found for cathode rays and undifferentiated Becquerel
rays by J. J. Thomson and by Strutt.29 Rutherford pointed out the
possibility that harder x-rays, as penetrating as the new y-rays,
might match their ionizing properties more closely. But he sug-
gested explicitly that the penetrating y-rays were electrons; they did
not react noticeably to a magnet, he said, because their velocity
was essentially equal to that of light itself.
23
Becquerel, CR, 752(i90i), 371-3.
24
Becquerel, CR, 133 (1901), 710.
25
Stewart, PR, 77(1900), 175.
26
Rutherford, PM, 5 (1903), 177-87. For a detailed discussion, see Heilbron,
AHES, 4(1968), 247-307.
27
Rutherford, Nature, 66 (1902), 318-19.
28
Ibid.
29
J. J. Thomson, PCPS, 10 (1898), 10-14. R. J. Strutt, PTRS, 196A (1901),
507-27.
In 1902 the penetrating rays were given the name y following the
informal but universally accepted taxonomy established by Ruth-
erford.30 This action signaled the first sustained research interest in
y-rays. Marie Curie, in her thesis on radioactive substances pre-
sented in 1903, drew a conceptually crucial analogy between the
three components of radioactivity and the known products of
electrical discharge: a-rays can be compared to the positively
charged molecules called canal rays, /?-rays to cathode rays, and
y-rays to x-rays. In numerous comparisons between the last pair,
she implied strongly that y-rays arise from the impact of very-high-
velocity ^-particles upon the atoms within the sample of radium
itself.31 Her diagram, reproduced as Figure 3.2, symbolized the
tripartite division of decay products by their relative deviation in a
magnetic field. Becquerel was convinced and Rutherford was
influenced by Curie's presentation.32 In a talk to the London
Physical Society in June, Rutherford reversed his position yet
again, agreeing that y-rays were "probably" equivalent to x-rays.33
In his talk to the British Association that September, however,
Rutherford demurred; insufficient knowledge, he said, prevented a
definite conclusion.34
Not everyone agreed with the triumvirate of Rutherford, Curie,
and Becquerel. R. J. Strutt recognized only belatedly that his 1901
investigation of the ionizing power of a-rays had been confounded
by the presence of y-rays. Accordingly, he returned to the subject.35
He thought Curie's interpretation of the y-rays tempting, but his
own data indicated otherwise. He measured the ionizing strength
of the rays by the time needed for a given electrical charge to
develop or disperse in his electrometer. His table (Figure 3.3)
shows clearly that the relative ionization produced in several gases
by y-rays agreed closely with that produced by a- and /?-rays,
approximately in proportion to the density of the gas. Even the
hardest x-rays, those produced by the greatest potential differ-
ences, failed to mimic y-rays. He suspected that y-rays were parti-
30
Trenn, Isis, 67(1976), 61-75.
31
M. Curie, Theses (1903), 48, 80, 83, 138.
32
Becquerel, CR, 7^6(1903), 1517-22.
33
Rutherford, Nature, 68 (1903), 163.
34
Nature, 68 (1903X 610.
35
R. J. Strutt, PRS, 72 (1903), 208-10.
Figure 3.2. Marie Curie's schematic diagram of differential magnetic

deflection of radioactive emanations. [Theses (1903), 49.]
cles. Strutt believed that electromagnetic impulses would arise

from the /^-particle impacts, but he thought them simply too weak
to detect. The net/?-flux from radium, he pointed out, is some nine
orders of magnitude less than the expected cathode ray flux in a
typical laboratory electric discharge. Although he offered no em-
pirical evidence for the claim, Strutt explained the lack of mag-
netic deflection and extraordinary penetrability by suggesting that
the y-particles had lost their charge. By late 1903, discounting the
freewheeling speculations of the aged Lord Kelvin,36 three views
were competing for the description of y-rays: short transverse x-ray
impulses, neutral material particles, and extremely high-velocity
electrons.
36
Kelvin thought the "simple prima facie view" was that y-rays are uncharged
free atoms of radium. BAAS Report (1903), 535-7; PM, 7 (1904), 220-2.
Kelative Ionisations.
Relative 1 onidation.
Gas. Relative
density. Rout gen
a rays. /3 rays. 7 rajs- rays.
Hydrogen o -or>93 0 220 0*157 0-1C9 0114

.Air 1-00 1*00 100 1-00 TdO
111 116 1 -21 117 1-.VJ
Carbon dioxide 1 53 1-54 1-57 1 53 i-r>o
Cyanogen 1-hG 1-94 1 -8f> 1 71 1 -or,
Sulphur dioxide 2-ld 2 04 2-31 2*13 7-07
Chloroform 4-32 4 4t 4 -Si) 4-8S 310
Metliyl iodide 5 05 3-51 5-lS 4 M) 72 0
Carbon tetrachloride. . .. 5 31 5 34 5*83 5-67 45 3
Figure 3.3. R. J. Strutfs data on relative ionization of gases by a-, /?-, y-,
and x-rays. [PRS, 72 (1903), 209.]
CORROBORATING THE y-RAY-X-RAY ANALOGY

J. J. Thomson thought that increasing the density of radioactive
matter might increase its rate of decay.37 He made this suggestion
in the course of an enlightening discussion of atomic disintegra-
tion as the mechanism of radioactive decay. The longevity of the
decay from radium, he proposed, might be explained if only a
small fraction of the total number of atoms decay at any instant.
The potent ones, he claimed, were those with abnormally high
kinetic energy. A decaying atom might raise its closest neighbors to
potent energy if the material was packed closely enough. The
suggestion appealed to Rutherford, who had moved meanwhile
from England to McGill University in Canada. But when Ruther-
ford did some tests, he found that Thomson's predicted increase in
activity with sample density did not occur.38
A reader of Nature, where Rutherford published this result,
pointed out that radium y-rays could not arise from self-bombard-
ment.39 He alluded to Stokes' opinion that an impulse similar to an
x-ray should arise not only where a charged particle is stopped but
also where it starts. Might not the y-rays be due to the acceleration
37
J. J. Thomson, Nature, 67(1903), 601-2.
38
39
Ashworth, Nature, 69 (1904), 295.
of a- and ^-particles out of the disintegrating atom? Rutherford
agreed, and early in 1904 undertook a critical reevaluation of the
mechanism proposed by Marie Curie for ^-induced emission of
y-rays. He encouraged his elder colleague, A. S. Eve, to analyze
carefully the data Strutt had amassed on the relative ionization of
various radioactive species. The following month, Eve reported
that the rate of ionization produced by the most highly penetrating
x-rays was much closer than previously suspected to that of the
y-rays.40 (See Figure 3.4 for a comparison of the old and new data.)
Rutherford stepped in immediately to strengthen the new hy-
pothesis by identifying y-rays as the electromagnetic impulse pro-
duced when ^-particle electrons exit at very high velocity from a
disintegrating atom. "All the experimental evidence so far ob-
tained," he claimed, "is now in agreement with the view that the
y-rays are very penetrating Rontgen rays which have their source
in the atom of the radioactive substance at the moment of the
expulsion of the /? or kathodic particle."41 "I have found," he went
on to explain, "that the y-rays from radium always accompany the
/?-rays and are alway proportional in amount to them . . . Active
products which give a-rays and no /?-rays do not give rise to y-rays.
The /? and y-rays appear always to go together and are present in
the same proportion."42
If y-rays are highly penetrating x-ray impulses, they should carry
no charge. Supporting evidence for this claim was offered on the
heels of the Eve-Rutherford data. James McClelland, a former
Cavendish student working in Dublin, passed y-rays through an
insulated cylinder filled with lead shot. The cylinder was long
enough so that all the rays were absorbed. With due care to
eliminate the effect of ionization in the surrounding air, McClel-
land showed that no net charge was deposited in the cylinder. "The
weight of evidence is certainly in favor of the y-rays being similar to
Rontgen rays," he concluded.43
McClelland asked why the y-rays should be absorbed by gases in
approximate proportion to the gas density. For a- and /^-particles,
indeed for any particle, the law follows directly from the simple
assumption that the particle is smaller than the atoms of the target
40
Eve, Nature, 69 (1904), 436; PM, 8 (1904), 610-18.
41
42
Rutherford, Radioactivity (1904), 144-5.
43
McClelland, RDS Sci trans, 8 (1904), 99-108; PM, 8 (1904), 76.
Gas. Relative Previous Present

Density. values. values.
H 07 11* 42
Air 10 10 10
HOS 12 6 9
SO., 22 8 23
Chloroform 43 32 46
Carbon Tetrachloride 53 45 4*9
Methyl Iodide 50 72 135
* Strutt 11; J. J. Thomson 33; Eutherford -5; Perrin 026.
Figure 3.4. Eve's data revising knowledge of ionization by hard x-rays.

[PM, 8 (1904), 613.]
and therefore interacts not with entire atoms but with their consti-
tuent electrons. Equal volumes of gases contain equal numbers of
molecules, so that the gas density is proportional to the molecular
weight. According to Thomson's influential ideas about atomic
structure, the number of electrons in an atom is proportional to its
atomic weight; thus, it is also proportional to the density of the gas.
A particle of subatomic dimension that passes through the gas will
lose energy in proportion to the number of electrons it encounters.
Therefore, particle absorption should be proportional to the den-
sity of the gas.
McClelland was uneasy, however, about whether spreading
waves could be absorbed by gases in the same proportion. "It
would seem difficult to explain such a law," he said.44 He did not
have in mind the difference one might imagine to exist between the
absorption of periodic waves and impulses by bound electrons.45
Rather, McClelland was echoing reservations that his former
teacher, J. J. Thomson, had recently expressed about the isotropic
44
McClelland, PM, 8 (1904), 77.
45
For a periodic wave, not all electrons have natural frequencies sufficiently
close to resonance to be appreciably excited. But a sharp impulse will excite all
electrons. An organ tone transfers energy only to the closely tuned strings of a
harp; a pistol shot excites them all.
character of x-ray pulses; McClelland tried to find justification for
Thomson's nascent ideas about the structure of the aether. If a
spherical impulse or periodic wave spreads over areas that are large
with respect to the size of the atoms, it will no longer, McClelland
asserted, interact with constituent electrons. Rather, it must be
absorbed by entire atoms, or groupings of atoms. Since the num-
ber of molecules per unit volume of a gas is constant for the same
pressure and is independent of the gas density, McClelland
doubted that spherical impulses could be absorbed in proportion
to the gas density. Nor was he alone in his concern. William Henry
Bragg in Australia raised the same issue more forcefully at the
same time. But Thomson's and Bragg's important views will be
considered in the next chapter.
Eve's discovery that hard x-rays follow the same law of absorp-
tion in gases as do particles permanently tied the fortunes of the
y-ray to those of the x-ray from 1904 on. The results appeared only
in a footnote in Rutherford's book Radioactivity, published that
year.46 They brought an end to the cautious balancing of three
alternative models for the y-rays thatfilledthe text itself. But in the
second edition of the book, published the following year, Ruther-
ford's discussion of y-rays is extensively simplified and reorganized
around the newly justified impulse hypothesis. The idea of impulse
y-rays was widely circulated in Rutherford's book, which was
quickly recognized as a working digest of essential research in
radioactivity and was published in several editions over the next
decade.
FRIEDRICH PASCHEN AND THE

FAST-ELECTRON HYPOTHESIS
At just the time that McClelland showed that y-rays possess no

electrical charge, L. C. H. Friedrich Paschen, noted Ordinarius at
Tubingen, published what he took to be evidence to the contrary.
He had enclosed a sample of radium bromide in a lead box
sufficiently thick to retain all a- and ^-particles, isolated the box in
a vacuum, and found that it developed a spontaneous positive
charge.47 Since the only things able to escape from the box were
46
Rutherford, Radioactivity (1904), 14411.
47
Paschen, AP, 14 (1904), 164-71. Paschen was not the first to note the self-
charging capability of radium. See M. Curie, Theses (1903), and Dorn, PZ, 4
(1903X507-8.
y-rays, he concluded that they must carry a negative charge.
Paschen was convinced that y-rays were ordinary electrons travel-
ing at speeds greater than even the 2.8 x io10 cm/sec of some
^-particles. The rays had shown no magnetic deflection, he said,
only because a strong enough field had not been developed. He
cited recent theoretical discussions that proposed a "unit of elec-
tricity that moves at the velocity of light," and thought that he had
found it in y-rays.48
An elegant but unsound second experiment further confirmed
Paschen's view. He used an intense magnetic field to force the
/?-particles from radium into spiral paths so tight that they could
not reach a surrounding lead ring. The ring, shown in cross section
in Figure 3.5, is marked a; the radium sample is at b. Electrons
from the sample are forced into paths of circular projection in the
plane of the magnet pole faces TV and S. Only the y-rays could reach
the ring, the a-rays having been stopped by the container sur-
rounding the sample. Paschen found that a negative potential
developed on the ring, and claimed that it was caused by a negative
charge deposited there by the y-rays.49 Paschen's clever experimen-
tal technique was an extension of prior work by Becquerel and
H. Starke.50 It was later developed into what we now call p-ray
spectroscopy. But here it led Paschen astray.
Paschen's conclusion was clouded by his total neglect of second-
ary electrons; he had the misapprehension that only cathode rays
could stimulate secondary electrons in matter.51 He utterly failed
to take into account the possibility that y-rays, like x-rays, might do
the same. In his first experiment, secondary electrons might be
emitted from the outside of the lead shield, leaving a net positive
charge; in the second, electrons released from the surrounding
vacuum vessel walls could easily make their way back to the lead
ring. The reaction was quick; both McClelland and Eve immedi-
ately pointed out Paschen's error.52 A few of Paschen's results
repeated work done in 1899 and 1900 by Becquerel. The latter took
an unusual step for a Frenchman and sent a detailed review of his
48
Paschen, AP, 74(1904), 171.
49
Ibid., 389-405.
50
Becquerel in Guillaume and Poincare, Congres international, 3 (1900), 63-9.
Starke, VDpG, 5(1903), 14-22.
51
Paschen, PZ, 5 (1904), 502-4.
52
McClelland, RDS Sci trans, 8 (1904), 99-108. Eve, Nature, 70 (1904), 454.
Figure 3.5. Diagram of Paschen's experiment that made him think that
y-rays deposit negative charge. [AP, 14 (1904), 390.]
work to the editors of the Physikalische Zeitschrift for translation

into German in the hope that it would "show what the state has
been for over a year of the issues that Herr Paschen has taken up
again."53
Before Paschen's mistake became clear to him, another of his
papers appeared; this one contained an estimate of the velocity of
the presumed electronic y-rays. The lack of deflection in a strong
magnetic field he took as evidence that their e/m must be less than
io3 emu/g (in comparison to io7 for the /^-particle).54 He cited
53
Becquerel, PZ, 5 (1904), 561-3. This was the first of only a few papers
Becquerel ever published in German.
54
Paschen, PZ, 5 (1904), 563-8.
Abraham's theory and Kaufmann's experiments in order to ex-
ploit the idea that the mass of an electron increases as its velocity
approaches c. A mass increase by a factor of io4 would explain the
discrepancy. Paschen concluded that his y-ray electrons might
travel within io" 40000 of the velocity of light! The Lorentz trans-
formation predicts only a velocity within 5 parts in a billion (5 x
io" 9 ) of c.
To support these claims, Paschen turned to measurements of
the heat evolved in radioactive decay. If the kinetic energy of each
y-particle is, in fact, as large as that of 40 hydrogen atoms moving
very close to the velocity of light, the y-component should carry
away most of the energy. A widely accepted estimate of energy
emission held that 1 g of radium evolves 98.5 cal/hr, not counting
the y-rays.55 So, Paschen attempted the nearly impossible: He
tried to determine by ice calorimetry the total heat produced by
radium enclosed in so much lead that even the y-rays were trap-
ped. It was by pure chance that he got any result at all; thermal
equilibrium in the massive lead shield could not occur as quickly
as calorimetry requires. But miraculously, he found a result, 226
cal/g-hr, that corroborated his conclusion: The y-component
alone radiates more than half of the total energy of radium. Citing
data assembled by Wilhelm Wien on /3-particles, Paschen con-
cluded that each /3-particle carries only i/74th the energy of a
single y-electron. Thus, he concluded, "the y-rays cannot be the
Rontgen effect of the /3-rays" because the latter do not have
enough energy.56
Rutherford had previously studied the heating effect of the
various components of radioactive decay in radium.57 He had
completed a short review of the subject just before Paschen's
claims appeared.58 In light of his experience, Paschen's results
seemed simply fantastic. "This result was so unexpected," he
wrote, "and was of so much importance in considering the nature
of the y-rays, that we decided to verify the experiments."59 Before
Rutherford could put more than a preliminary result in print,
Paschen himself disavowed the issue. He had found that he could
not duplicate the calorimetric results during the winter months,
55
Precht, VDpG, 6 (1904), 101-3. Runge and Precht, Berlin Sb (1903), 783-6.
56
Paschen, PZ, 5 (1904), 567.
57
Rutherford and McClung, PZ, 2 (1900), 53-5. For a description of Ruther-
ford's research effort, see Heilbron in Bunge and Shea, Rutherford (1979),
42-73.
58
Rutherford, AusAAS Report (1904), 86-91.
59
Rutherford and Barnes, PM, 9 (1905), 623.
which was clear evidence that his experiments hid systematic
errors.60 Rutherford's experiments showed that the y-rays carried
only a small percentage of the total heat from radium. Most of the
energy resided in the a particles. "No appreciable increase in
heating effect was brought about by absorbing the y-rays," he
reported, and the result was corroborated later that year by Knut
Angstrom.61
In the meantime, Paschen's claim that an electrical charge is
deposited by y-rays was destroyed by Thomson, developing
McClelland's earlier work. Thomson tested the relative discharge
rate of two brass cylinders, both hollow, closed, and electrically
isolated in a vacuum.62 One was empty, the other filled with lead
shot. Each was irradiated along its axis with the y-rays from radium
bromide. Secondary electrons stimulated inside the metal are
quickly reabsorbed. Thus, if the y-rays can deposit a charge,
whether positive or negative, the increase of charge should always
be greater for the lead-filled cylinder. If, on the other hand, the
y-rays are uncharged, and affect electrical potential only by the
stimulation of secondary electrons at the surface of the metal, then
the discharge rate of the empty cylinder should be greater. Both
ends of the empty cylinder emit secondary electrons, but only the
near end of the lead-filled cylinder radiates them because virtually
all of the y-rays are absorbed in passing through the lead. In every
case, Thomson found that the discharge rate was greater for the
empty cylinder, and it always indicated a loss of negative charge.
This showed that Paschen's results were entirely due to secondary
electrons, and Paschen acknowledged this fact.
THE STATUS OF THE y-RAY-X-RAY ANALOGY
In the years after 1905, y-rays took a permanent place as the

radiation component of radioactivity decay. Indeed, the verifica-
tion of the threefold identification of radioactive decay products
with products of electric discharge brought the final severing of ties
between the two fields. The editors of the Annalen der Physik
removed the newer field bodily from the section Elektricitdtslehre
60
Paschen, PZ, 6(1905), 97.
61
Rutherford and Barnes, PM, 9 (1905), 626. Angstrom, PZ, 6 (1905), 685-8.
62
J. J. Thomson, PCPS, 13 (1905), 121-3.
in their classifactory scheme, elevating Radioaktivitdt to a status
equivalent to that of electricity.63
R. J. Strutt, who had presented the strongest opposition to the
y-ray- x-ray analogy in Britain, conceded the case in 1904. He
described the nature of the y-rays as "one of the most obscure
questions in connection with radioactivity," but concluded his
survey with the remark that y-rays were "probably x-rays."64 In the
second edition of Radioactivity, Rutherford rehearsed the argu-
ments against Paschen's interpretation and concluded that "the
weight of the evidence, both experimental and theoretical, at
present supports the view that the y-rays are of the same nature as
the x-rays but of a more penetrating type."65 In Britain, where the
impulse hypothesis for x-rays had been accepted most firmly, the
extension of the hypothesis to include the y-rays was complete in
1904.
In Germany there was greater diversity of opinion. Paschen's
debacle is indicative of this fact. Aside from his eminent position at
Tubingen, Paschen had held an international reputation since
1895 for his work in optical and infrared spectroscopy. That he was
simultaneously ignorant of possible secondary electrons, and will-
ing to exhibit that ignorance in defense of a corpuscular hypothesis
of y-rays, is symptomatic of the greater resistance in Germany to
the facile acceptance of the y-ray-x-ray analogy. It also indicates
the difficulty of isolating the purely wavelike properties of the
penetrating y-rays. For many in Germany, the requisite hard
evidence had simply been lacking before the reaction to Paschen.
Even after the barrage of evidence from Britain, occasional voices
of opposition were heard. Otto Wigger, for example, refused to
adopt the analogy in his 1905 review of y-ray absorption.66 But by
then his was a minority view, even in Germany.
There was no serious resistance to the identification of y-rays
with x-rays following Paschen's experience. The reaction to Pas-
chen's advocacy showed beyond doubt that negatively charged
particles are released from metals by the action of y-rays. Bec-
querel had noted in 1901 that a wrapped photoplate exposed to
63
See the tables of contents of the Beibldtter zu den Annalen der Physik for the
relevant years.
64
R. J. Strutt, Becquerel rays (1904), 86.
65
Rutherford, Radioactivity, 2nd ed. (1905), 186.
66
Wigger, JRE, 2 (1905), 391-433-
y-rays was more heavily exposed if the plate was covered by a metal
sheet.67 He suggested that this result was probably caused by the
stimulated emission from the metal of some photoreactive ray that
was more easily absorbed by the emulsion than were the y-rays; it
was clear now that these secondary emissions were electrons. Eve
thought in 1904 that, like x-rays, y-rays stimulate both negatively
charged particles and secondary y-rays.68 McClelland jumped to
the conclusion that the particles are ^-electrons, "the mass and the
charge being no doubt the same," although he had no evidence to
corroborate the claim.69 Indeed, no serious study to determine the
e/m ratio for the secondary particles was undertaken until 1906.70
The result corroborated McClelland's claim, but by then it was
widely expected. X-rays stimulate secondary electrons; so should
y-rays.
The y-ray-x-ray analogy was settled by 1905. Shortly thereafter,
continental leaders such as Lorentz in theoretical physics and
Walther Nernst in physical chemistry accepted the analogy as an
accurate pedagogical classification, if not an explanation, of
y-rays.71 There was a marked decline in research on y-rays between
1905 and 1908, and this too may be explained by the orthodox
analogy. As soon as one could unambiguously associate y-rays
with the impulse model of x-rays, they lost allure as a research
topic. Experimental study of either ray offered data on both, and
x-rays were considerably easier to study.
The strength of the y-ray-x-ray analogy held quite generally
until the early 1920s. Albeit based more on similitude than on hard
data, it remained a rock in the sea of uncertainty that would
surround attempts to specify the precise nature of both forms of
radiation in the intervening years. What came under discussion
first in 1907 and increasingly thereafter was no longer the similarity
of x-rays to y-rays, but rather the impossibility of describing either
in any electromechanically consistent manner.
67
Becquerel, CR, 132 (1901), 371-3.
68
Eve, PM, 8 (1904), 669-85.
69
McClelland, RDS Sci trans, 8 (1904), 169-82; PM, 9 (1905), 230-43.
70
Allen, PR, 23 (1906), 65-94.
71
Lorentz, Lehrbuch, 2 (1907), 587. Nernst, Theoretische Chemie, 5th ed. (1907),
410. Rutherford inserted a new chapter on y-rays in the second edition of
Radioactivity (1905).
PART II
Ionization and the recognition of paradox
1906-1910
Secondary rays: British attempts to
retain mechanism
There is . . . a reasonable argument that the y and

x rays are also material.1
When x-rays or y-rays strike atoms, electrons and secondary x-rays

are emitted. Experiments to sort out the differing properties of the
two secondary components produced new and perplexing obser-
vations in the first decade of this century. The most perplexing
problems concerned the effect of the rays in producing secondary
electrons. First, x-rays and y-rays ionize only a small fraction of the
total number of gas molecules through which they pass. Spheri-
cally expanding pulses should affect all molecules equally; mani-
festly, they do not. Second, the velocity that x-rays impart to
electrons is many orders of magnitude higher than one would
expect to come from a spreading wave. The energy in the new
radiations seemed to be bound up in spatially localized packages,
available to an electron in toto.
British physicists responded to these paradoxes with attempts to
revise, rather than replace, classical electromechanics. J. J. Thom-
son suggested that old ideas about the microscopic structure of the
aether might have to be reformulated. William Henry Bragg con-
cluded that x-rays and y-rays are not impulses at all but rather
neutral material particles. Bragg and Thomson tried at first to find
explanations using models based on human experience with ma-
chines. Each sought a resolution within the context of classical
mechanics; each ultimately failed. But the discussion of basic
concepts that resulted enriched physics in two ways. Materially, it
stimulated Charles Barkla to identify a new form of secondary
x-ray, one that is homogeneous in penetrating power and charac-
1
W. H. Bragg, TPRRSSA, 31 (1907), 97.
72 Ionization and the recognition ofparadox, 1906-1910
teristic of the atom in which it is stimulated. Conceptually, the
discussion directed concern to the difficulty of formulating any
theory of x-rays and y-rays that answered the dual paradoxes
encountered in the study of their ionizing power, as we shall see in
this chapter.
THE ENERGY OF SECONDARY ELECTRONS

Early attempts to determine the energetic properties of x-rays
focused on the heat they produce in irradiated bodies. Rontgen
had predicted but not investigated the effect.2 An unfairly forgot-
ten physicist, Ernst Dorn, professor at Halle, irradiated metal foils
enclosed within one of two connected glass vessels in 1897. The
enclosed air was heated, the pressure rose, and the expansion could
be read by the movement of afluidplug in the tube that connected
the vessels. By comparing the effect caused by a known electric
current, Dorn derived approximate values for heating by x-rays.3
Ernest Rutherford recognized the importance of the heating
effect to his earliest studies on radioactivity. In his first year in
Canada, he found a capable assistant in Robert McClung. To-
gether they measured the change in electrical resistivity of a plati-
num strip bolometer heated by x-rays.4 They then measured the
fraction of ionizing power lost by x-rays in traversing a unit
thickness of air. The two results taken together gave an estimate of
the total energy extracted from the beam by a known volume of
air. Measurement of the total charge released then allowed Ruth-
erford to calculate the approximate energy needed to ionize a
single nitrogen atom: 3.8 x io~10 erg. Their goal was to estimate
from the measured effect of radium on the same gas the total
energy available in the radioactive rays. It was this prior work that
gave Rutherford the instant conviction that Paschen's calorimetric
results must have been wrong.
All of these early experiments measured macroscopic effects, the
net result of a large number of interacting x-rays and secondary
products. None was capable of giving information on the detailed
energy balance of single x-ray impulses. Moreover, early investiga-
2
Rontgen I, p. 5.
3
Dorn, AP, 65(1897), 160-76.
Rutherford and McClung, PTRS, 196A (1901), 25-59.
British attempts to retain mechanism 73
tions were confounded by a mixture of two secondary effects; in
addition to electrons, bodies irradiated by x-rays also emit second-
ary x-rays. At first, both were referred to as diffuse reflection of
x-rays. The dual nature of these secondary effects was unrecog-
nized, but it complicated experimental study. One had to learn
first to distinguish the secondary x-rays from the particle flux, at
the beginning called secondary rays, that only later came univer-
sally to be identified as electrons.
During his initial researches, Rontgen discovered that x-rays
discharge electrified bodies, but he did not mention this in his first
paper. Immediately, a number of physicists lodged claims for
priority in showing that the rays neutralize electrically charged
objects.5 Rontgen's second paper was devoted to this effect.6 Most
people agreed that the action bore a remarkable similarity to the
photoelectric effect, which was at the time incompletely under-
stood. But the difference was even more telling: Whereas ultravio-
let light discharged only a negative electrical potential, x-rays
always acted, apparently oblivious to the sign of the initial charge.7
Only after Thomson's isolation of the electron in cathode rays in
1897 did the study of microscopic energy relations become feasi-
ble. The mixture of electrically active particles indiscriminately
called ions produced in gases by the passage of x-rays suggested
that ions, like the cathode rays, were mostly electrons.8 By 1900,
investigations on both sides of the English Channel had corrobo-
rated the fact that x-rays and ultraviolet light stimulate the release
of electrons from solids and gases.9 The difference lies in the
velocity with which each is emitted. Lenard had shown that a
typical velocity for photoelectrons was io8 cm/sec.10 In 1900 Dorn
magnetically deflected the electrons released by x-rays and inserted
the value of e/m known for cathode rays.11 The derived velocity
5
Among the claimants were J. J. Thomson, PRS, 59 (1896), 274-6; Righi,
Bologna Memorie, 6(1896), 231- 301; and Benoist and Hurmuzescu, CR, 122
(1896X235-6.
6
Rontgen II.
7
For contemporary understanding, see Wheaton, Photoelectric effect (1971).
8
J. J. Thomson, PM, 46 (1898), 528-45.
9
The major study besides Thomson's was by P. Curie and Sagnac, CR, 130
(1900), 1013-15.
10
Lenard, AP, 2 (1900), 368.
11
Dorn, Archives Neerlandaises (2), 5 (1900), 595-608. The original measure-
ments appeared in Dorn, Halle Abhandlungen, 22 (1900), 37-43.
exceeded io9 cm/sec, which was equivalent to the recently deter-
mined velocity of cathode rays.12
Dorn's result did not arouse the interest that in retrospect might
have been expected. He had shown that the velocity of secondary
electrons produced by x-rays is of the same order of magnitude as
that of the electrons whose impacts produce the x-rays in the first
place. This is surprising because the energy density in the x-ray
would be expected to decrease rapidly as the x-ray spreads spheri-
cally and traverses macroscopic distances. Nonetheless, the full
energy of the ray seemed to be concentrated on an individual
electron in the irradiated body. Dorn did not call attention to this
issue, nor is there reason to suspect that he recognized its peculiar-
ity. He admitted that irregularities in the output of his x-ray tube
might have clouded the result. The work was not published in a
major journal. Dorn himself was not considered to be a leading
experimentalist - competent, but not distinguished.
THE TRIGGERING HYPOTHESIS

By the time that Dorn's empirical result was widely recognized to
be accurate, another way of explaining it was available; thus, the
potential problem went unrecognized for another five years. Phi-
lipp Lenard completed important experiments on the photoelec-
tric effect in 1902.13 He uncovered the surprising fact that the
maximum velocity with which electrons are ejected by ultraviolet
light is entirely independent of the intensity of the light. This result
convinced him that there could be no transformation of light
energy into electron kinetic energy. Instead, he proposed that
electrons in an atom already possess their photoelectric velocity,
or the potential energy equivalent, by virtue of their membership
in the atomic system. The light only triggers the release of selected
electrons; it does not add energy to them. Until 1911 this triggering
12
Wiechert, AP, 69(1899), 739-66.
13
Lenard, AP, 8 (1902), 149-98. Despite claims to the contrary, even by histo-
rians, Lenard did not here demonstrate that photoresponse varies with light
frequency. He only claimed that it varies with the type of light used-arc light,
spark light, or the type of electrode metal. The word frequency appears no-
where in his study; wavelength appears only once, and in a different context.
Compare Jammer, Conceptual development (1966), 35; Kleinert and Schon-
beck, Gesnerus, 35 (1978), 318; Hendry, Annals of science, 37 (1980), 64.
hypothesis formed the basis of almost all physicists' understanding
of the photoelectric effect.14
The triggering hypothesis owed some of its appeal to its compati-
bility with contemporary hypotheses of atomic disintegration. It
seemed to promise a means of investigating the mysterious decay
of radioactive elements. Thomson adopted triggering as an expla-
nation of the photoelectric effect in 1905.15 Between 1905 and 1907,
his student, J. A. McClelland, devoted a series of studies to the
corroboration of the triggering hypothesis for the emission of
^-particles by y-rays. McClelland's chief weapon in this study was
the observed monotonic increase he found in the total ionizing
power of the secondary rays, including what he thought were
secondary y-rays, with the atomic weight of the scattering material.
He took this as evidence that the primary beam, a mixture of /?-
and y-rays from radium, could induce atomic disintegration in
matter.16 If the electrons were free before interacting with the rays,
McClelland explained, "it is difficult to see how such remarkable
relations should exist between the atomic weight and the intensity
of the secondary radiation."17
It was no great step to expand the capabilities of the triggering
hypothesis to include the excitation of electrons by x-rays. Wil-
helm Wien in Wiirzburg concluded in 1905 that the triggering
mode offered the only realistic explanation.18 Thomson publicly
announced his adoption of the triggering hypothesis for x-ray
ionization in 1906.19 His student, P. D. Innes, found direct sup-
porting evidence the following year; like the velocity of photoelec-
trons, the velocity of the secondary electrons from x-rays appears
to be fully independent of the intensity of the incident radiation.20
For a brief period in 1906, the influence of the triggering hypoth-
esis was at its height. It was widely thought to explain the observed
velocities of electrons ejected from matter by light, by x-rays, and
14
Wheaton, HSPS, 9(1978), 299-323.
15
J. J. Thomson, PM, 10 (1905), 584-90.
16
McClelland, RDS Sci trans, 9 (1905), 1-8, 9-26, 37-50. Righi also found
evidence of a correlation with atomic weight. Lincei Atti, 14 (1905), 556-9.
17
McClelland, RDS Sci trans, 9 (1905), 36.
18
Wien, AP, 18 (1905), 991-1107. See Chapter 5.
19
J. J. Thomson, PCPS, 13 (1906), 322-4; Corpuscular theory (1907), 320-1.
Some of the evidence that convinced Thomson came from Bumstead, PM, 11
(1906), 292-317.
20
Innes, PRS, 79A (1907), 442-62.
by y-rays. Indeed, for a while it seemed as if the study of these
velocities would reveal the distribution of electron energy within
the atom and shed light on the mysterious process of radioactive
disintegration. These were two of the most pressing fields of physi-
cal research in the first decade of the century.
THE PARADOX OF QUANTITY
J. J. Thomson had one serious reservation about the impulse

explanation of x-ray ionization. His early experimental studies had
convinced him that only a vanishingly small fraction of gas atoms
is ionized by x-rays.21 It is difficult to see how any spreading wave
could fail to affect all atoms in its path equally. Each x-ray impulse
spreads outward from its source. In any particular direction along
an x-ray "beam," the diverging pulse should disturb all atoms to
the same extent. Indeed, the lateral area of the x-ray beam is of
macroscopic dimension, and all atoms within the beam should be
affected the same way. Rontgen had shown in his third study that
the intensity of the x-ray beam did not decrease appreciably until
one reached very large angles from the normal to the anticath-
ode.22
Thomson had quickly exploited the discovery that x-rays can
induce conductivity in a sample of gas through which they pass.
The earliest means of measuring the ionizing strength of x-rays was
to determine the rate at which a given initial charge decayed in an
electrometer. In their early study of the process of ionization by
x-rays, Thomson and Rutherford had charted the course of the
induced conductivity by measuring charge per unit time as a
function of the increasing potential difference applied between two
electrodes in a gas (see Figure 4. i). X-rays produce ions in the gas;
when a small potential difference is applied, current immediately
begins to pass. As the applied potential increases, the current
increases too, but only up to a point. At values of a few hundred
volts per centimeter, the current settles down to a saturation value.
The discharge is a dynamic equilibrium. X-rays tear electrons
from atoms, but the electrons spontaneously recombine with
stripped atoms if given the chance. A high potential forces more
21
J. J. Thomson and Rutherford, PM, 42(1896), 392-407. See also Y)orn,AP, 63
(1897), 160-76.
22
Rontgen III, pp. 21-3.
Potential
Figure 4.1. Saturation of current produced by ionizing a gas.
free charges to reach the electrodes before recombination occurs.

The existence of a saturation current shows clearly that the total
number of ionization events is limited. If the space between the
electrodes is increased, the saturation current also increases be-
cause more gas atoms are now able to contribute electrons. It was
clear from electroscopic measurement that not all atoms are
treated equally by the x-ray beam; only a small fraction of gas
atoms is ionized. In 1896 Thomson had estimated it to be a
maximum of 1 in 1012 atoms.23
The puzzling fact that an x-ray beam ionizes only some of the
atoms over which it passes I shall call the paradox ofquantity. Note
that Thomson's discovery was partly due to his recent determina-
tion of the charge on the electron. Without knowledge of this unit
of electricity, it would not have been possible to count the number
of ionized atoms and learn that there were so few. The paradox of
quantity was created in the wake of the unitary electron.
In 1903 Thomson offered several potential explanations of why
x-rays ionize so few atoms. He raised the unsettling possibility
that different atoms of the same substance are, in some undeter-
mined way, different from each other. Perhaps, he suggested, the
x-rays only affect electrons already free of their parent atom. His
23
J. J. Thomson and Rutherford, PM, 42 (1896), 392-407.
78 lonization and the recognition ofparadox, 1906-1910
most realistic explanation was a modification of the triggering
hypothesis; it fixed attention on the kinetic energy of the target
atoms. Thomson suggested that ionization occurred only in atoms
with velocity greater than some threshold value.24 Perhaps only the
very small fraction of atoms at the high end of the Maxwell
velocity distribution was able to discard electrons. Were this the
case, heating the target should increase the yield of electrons.
Thomson enlisted the services of Robert McClung, recently ar-
rived from Montreal, to test the matter. In May 1903, McClung
reported that there was no measurable difference in the ionization
of air over a range of 140 to 2ooC.25 The paradox of quantity
remained unanswered.
Thomson was concerned, but not despairing, about this situa-
tion. He briefly revived ideas he had had for a decade about the
microscopic structure of the electromagnetic field. As early as
1893, Thomson had speculated that lines or "tubes" of electric
force might be more than just mathematical abstractions. If they
were thought to be finite in number, an electromagnetic theory of
light would result with some "characteristics of the Newtonian
emission theory."26 An oncoming pulse, traveling on the tubes,
would not be continuously distributed in space. In 1903 Thomson
described the pulse front in his Silliman Lectures at Yale as a
pattern of "bright specks" on a dark background.27 This view
directly explained the small number of ionized atoms; there was
only a limited number of tubes. Thomson submitted this opinion
largely in the spirit of medieval speculationes; for four years, he
hesitated to raise the radical hypothesis again. When he did so, it
was in response to more than simply the paradox of quantity. In
the intervening years a more formidable difficulty had arisen from
closer study of the ionizing energy, or "quality" of both x-rays and
y-rays.
THE QUALITY OF X-RAYS

In the first years of x-ray research, the very newness of the rays
made it difficult to know how best to characterize their strength.
24
J. J. Thomson, Conduction of electricity(1903), 255-8.
25
McClung, PCPS, 12 (1903), 191 - 8 .
26
J. J. Thomson, Notes (1893), 3; Conduction of electricity (1903), 258. See
McCormmach, BJHS, 3 (1967), 362-87.
27
J. J. Thomson, Electricity and matter (1904), 62ff.
There were differences of opinion about which attribute of the rays
should be quantified as a measure of x-ray quality. Rontgen had
tried to quantify the fluorescent activity of a detecting screen using
photometric techniques. He found "from three closely agreeing
measurements" that the inverse-square law seemed to hold for the
decrease in x-ray intensity.28 Rutherford too attempted to quantify
the brightness of x-ray fluorescence.29 But formidable problems
confronted the accurate measurement of fluorescent or photo-
graphic effectiveness of x-rays. One of the reasons for the dearth of
useful empirical work on x-rays on the Continent in the early years
was the widespread use there of photography to detect the rays.
Accurate photometry was still in its infancy.
In Britain, where until 1907 the most useful research on x-rays
was done, the approach was quite different. The method devised
by Thomson to measure the saturation current passing through an
irradiated gas was widely applied. For a given gas at a fixed pressure
and for a standard electrode separation, a fairly reproducible
measure of x-ray ionizing strength could be obtained. In turn, this
method made possible the calibration of a new, and ultimately
more direct means to measure x-ray quality. By using the electro-
meter to determine saturation current, one could measure the
relative intensity of the rays that passed through an absorbing sheet
and express this as a fraction of the incident intensity. One could
then fit the result to an anticipated exponential decay in the
effective intensity of the rays,
(4.1)
Here 70 is the original measure of what was called the intensity of
the rays; I will return in a moment to this interesting term. The
quantity d is the thickness of the absorbing sheet, and // is the
absorption coefficient of the rays relative to the absorbing sub-
stance. The coefficient// is a measure of the penetrating power of
the rays; large // implies low penetrating power. The value of //
measured in aluminum came soon to be accepted as a standard.
Each of these methods measured something distinct in the rays,
and it was unclear which, if any, ought to be adopted as the
standard means to specify x-ray intensity. Indeed, there was a
certain novelty in defining the intensity of a train of electromag-
28
Rontgen I, p. 7.
29
Rutherford and McClung, PTRS, 196A (1901), 25-59.
netic impulses. The accepted definition was clearly insufficient.
Classically, intensity expresses the energy flux in periodic waves.
Poynting had shown that this directed quantity has a magnitude
equal to the vector product of the amplitudes^ a n d ^ . Consider-
ing that both AE and AH decrease as i/r (cf. equations 2.2 and 2.3),
their product will decrease as i/r2, as expected.
But impulses, unlike periodic waves, require an additional state-
ment about quantity. To fix the net energy radiated in a given
direction, one needs to know the number of impulses as well as the
electromagnetic amplitude of each. As the number of periodic
radiating sources increases, their additional output can simply be
added to the varying net electromagnetic wave amplitude. But the
temporal discontinuity of impulses requires that their quantity be
explicitly taken into account in measuring intensity. Thus, inten-
sity as determined by Thomson's ionization methods included a
major new emphasis on the number of pulses radiated past a point,
a quantity that does not vary with the azimuthal angle or the
distance from the source. The result was to modulate the definition
of radiant intensity in a subtle and historically significant man-
ner.30
This change in emphasis, unrecognized then, was to have far-
reaching ramifications. One of its first influences was to give the
pulse width a role in determining the energy of the impulse. As
Thomson had shown (equation 2.5), the total energy carried by a
single pulse is inversely proportional to the pulse width. Thus, an
impulse x-ray of given width should transport a fixed quantity of
energy. Although it was widely anticipated from Thomson's analy-
sis that different pulse widths would be manifested as different
penetrating powers in the rays, there was no real evidence of either
a correlative or a causative relation. Sommerfeld's analysis of
diffraction was, at the time, the only direct experimental means to
determine x-ray pulse width.
The research of Barkla, above all, served to fix x-ray penetrabil-
ity in aluminum as the standard means to characterize x-ray
quality. From its introduction in 1903 until 1912, no other way
existed of measuring the temporal duration of x-ray impulses. To
the extent that a given beam could be assigned a single coefficient
of absorption, one could speak of the beam as homogeneous in
30
Wheaton, HSPS, 11 (1981), 367-90.
ionizing power. If a second thickness of absorber failed to remove
the same fraction of ionizing capability from the beam as the first,
the beam was declared inhomogeneous, and effort was devoted to
separating out the homogeneous components. This means of
characterizing x-rays gained a great deal from Rutherford's work
to separate the a- and /?-rays.
W. H. BRAGG AND THE PROCESS OF IONIZATION

William Henry Bragg was Elder Professor of Mathematics and
Physics at the University of Adelaide in Australia. He was born in
England and educated at Cambridge, but he left for the colony in
1885, a year after Thomson had assumed the Cavendish Professor-
ship.31 Bragg did very little research in physics until 1904 and then
undertook a critical analysis of the explanation of ionization
produced in gases by various forms of radiation. He took special
objection to what he thought was the unwarranted assumption
that the exponential law of absorption of radiation could be ap-
plied to ionization by particles. In the course of this study, he
convinced himself that y-rays could not possibly be the spreading
impulses that, after 1904, most European physicists took them
to be.
Bragg delivered his results early in 1904 as his presidential
address to the section on physics, astronomy, and mathematics of
the Australasian Association for the Advancement of Science.32
The talk is a testament to Bragg's remarkable physical intuition
and is rich in the use of mechanical analogies. His conclusions are
based on simple calculations using data culled from a wide, but not
exhaustive, knowledge of relevant experiments. Already discern-
ible is his tendency to overstate his position; his eloquence and
persuasiveness occasionally made the evidence in support of his
claims seem to be greater than it was. But he succeeded in giving
his audience a clear understanding of each step in the mechanics of
ion formation in gases.
Bragg's comments were directed largely to the ionization in-
duced by particles, but early in his talk he established an important
distinction between the properties of x-rays and particles. When
31
Forman, DSB, 2, (1973), 397-400.
32
W. H. Bragg, AusAAS Report, 10 (1904), 47-77. Bragg was not inactive before
1904; see Am A AS Report, 5(1891), 57-71; 4(1892), 31-47; 6(1895), 223-31.
electrons or a-particles traverse matter, they interact with the
individual electrons that comprise the atoms of the absorbing
body. Therefore, they transfer energy through collisions in propor-
tion to the number of electrons they encounter, and are thus
absorbed roughly in proportion to the density of matter. As
pointed out by Rontgen and demonstrated by Strutt, x-rays are
absorbed according to a different law, and they therefore consti-
tute, as Bragg said, an "exception to the rule that pene-
tration . . . depend[s] only on the density of the substance pene-
trated."33 Impulses or periodic waves in the aether, Bragg said,
would not interact with single electrons, but rather with the "mole-
cule as a whole." Consequently, x-rays transfer energy less readily
to matter than do electrons of comparable energy.
One of the significant contributions of Bragg's early study was
his recognition that, strictly speaking, the exponential law of ab-
sorption does not hold for streams of particles. The exponential
law is a direct consequence of absorption in proportion to absorber
density; Bragg claimed that for particles this was valid only as a
first approximation. As a- and /^-particles slow down, their energy
is more easily lost because collisions with atoms become more
frequent. Therefore, a particle conforms to the exponential law
only in the high-velocity first part of its trajectory; toward the end
of its path, an accelerating increase in its absorption coefficient is
to be expected. Other observers had deluded themselves that the
exponential law holds because they experimented on beams con-
taining a fairly broad distribution of particle velocities.
Bragg therefore had particular reason to differentiate waves
from particles. Whatever the difference in the relationship of
absorption coefficient to absorber density, for homogeneous waves
the absorption coefficient should remain constant. For particles,
which are able to slow down, the coefficient increases with time.
When one is confronted with a new sort of ray, the characteristic
way that it is absorbed by matter offers a means to identify it as
particulate or aetherial. Bragg evaluated the y-rays according to
this condition, and although he could not demonstrate that their
absorption rate was constant, he could nonetheless see clearly that
even in the first approximation their absorption was substantially
different from that of periodic waves or impulses.
33
W. H. Bragg, AusAAS Report, 10 (1904), 7 2 - 3 .
REVERSING THE y-RAY-X-RAY ANALOGY
Bragg discussed Strutt's experiments on the ionization produced
by y-rays, and he emphasized that y-rays appear to follow the laws
governing the absorption of a- and ^-particles. They do not match
the x-ray ionization rate. Bragg accordingly came to the conclu-
sion that y-rays act on the individual electrons in atoms. Therefore,
in some sense, they must be localized in space. He was at first
reluctant to attribute to them a purely corpuscular nature because
of their great penetrability and because they seemed to suffer no
electric or magnetic deflection. But, he said, if they are waves,
"they are waves so small as to be unable to act [on] a whole
molecule or atom at once."34
Bragg's statement is remarkably obscure. A wave that, after
traversing macroscopic distances, can affect only a region smaller
than an atom is not a wave in the classical sense. Bragg was at a loss
to characterize it accurately. In his ambiguous statement, one
perceives an unconscious attempt to combine wave and particulate
characteristics. But soon he resolved his indecision by attributing
to y-rays all the essential properties of particles. Bragg therefore
reacted quite differently than did most European physicists to
Eve's discovery that hard x-rays also follow the absorption law
characteristic of particles. What Rutherford took as proof in 1904
that y-rays must be impulses like x-rays was interpreted by Bragg as
evidence that x-rays must have the same particulate qualities as
y-rays.
Bragg's recognition that the absorption coefficient of a particle
will increase along its trajectory opened the way to new studies of
a-particles. In the years 1904-7 he was able to study the range of
a-particles thanks to the gift of some radium bromide. The work
was done with the assistance of various students at Adelaide.
Richard D. Kleeman was the first and most important of the
collaborators.35 With his help, Bragg soon found that the total
distance in a substance or gas over which the ionizing effect of the
a-particles persists is directly related to the velocity with which the
particle left the disintegrating atom.
34
Ibid., 77. The passage actually reads "unable to act as a whole molecule," but
from the next sentence it is clear that this is a misprint.
35
These ten papers and one letter may be found under "Bragg, W. H." in the
indexes of PM, 8, 10, 11, 13, and 14 (1904-7).
In most of these experiments, Bragg adopted Thomson's means
to measure the ionizing strength of the a-beam: A standard poten-
tial difference is applied between electrodes enclosed in the irra-
diated gas. The potential is then increased until the current stimu-
lated by the beam achieves its fully saturated value. In 1906 Bragg
discerned that some source other than the incident a-particle
beam was ionizing the gas. He quickly dismissed the possibility
that the electrode potential alone could be the cause. "If it were,"
he explained, "there would be no saturation current"; potential
gradients substantially greater than the few hundred volts per
centimeter at which saturation occurs give ions sufficient energy to
ionize atoms themselves.36 An avalanche effect occurs, and the
current increases very rapidly. Rather, Bragg suspected that the
new ionizing emanation might be the x-rays expected to arise
when the charged a-particles strike gas molecules.
But herein lay a significant problem. The ionization caused by
a- or /^-particles should saturate. Each particle can interact with
only a limited number of molecules. But the x-rays produced by
collisions should radiate in all directions (in the orthodox view)
and the impulses should affect all molecules in their vicinity
equally. Since large numbers of atoms would therefore be affected,
the ionization current should not saturate. Bragg, like Thomson,
knew that the current does saturate at microscopically small levels.
It appeared to Bragg that far too few electrons were being released.
Thus, Bragg confronted the paradox of quantity, only hinted at in
his 1904 claim for "waves so small." It was reaffirmed in a particu-
larly clear operational form for Bragg, and he could not ignore it
without upsetting his ongoing a-particle studies.
Bragg suggested a possible explanation for the ionization due to
x-rays produced by a-rays. Perhaps, he thought, the secondary
effect of the x-ray is fully expended within a microscopically small
distance from its source. Within that short distance, the impulse
might retain sufficient energy to ionize an atom. But this explana-
tion clearly could not be complete. The current induced by ordi-
nary x-rays from a discharge tube also saturates. This occurs even
after the rays have traveled distances equal to billions of atomic
diameters. Even if one allowed the impulse to be severely restricted
in its spatial width, after traversing these long distances it would
36
W. H. Bragg, PM, 11 (1906), 623.
have diffracted to encompass regions much larger than atoms. The
ionization data indicated that soft x-rays might do this, but hard
x-rays and y-rays acted like particles: They did not appear to
diffuse their ionizing power in space at all. By 1906 it was clear to
Bragg that something was seriously amiss in the widely accepted
hypothesis of x-ray impulses, and he reaffirmed the extension to
x-rays of the corpuscular status he had assigned to y-rays in 1904.
THE PARADOX OF QUALITY
Bragg's research interests shifted in 1906 from a-particles to y-rays.

He soon came to recognize a new problem that the pulse hypoth-
esis of x-rays seemed incapable of explaining. It was the same issue
that the early researches of Dorn had raised but that had been
glossed over by the adoption of the triggering hypothesis in Eu-
rope. Bragg, however, was not in Europe. His relative isolation in
Australia gave him the courage to question the impulse hypothesis
of x-rays.
Bragg apparently had little firsthand knowledge of the photo-
electric effect. In 1904 the relevant evidence he cited came from
Townsend's study of ionization of gases by ultraviolet light, not
Lenard's work.37 The triggering hypothesis was not mentioned.
Although Bragg was certainly familiar with the triggering hypoth-
esis, his understanding derived from his work on radioactive
disintegration, and he was not convinced that triggering needed to
be extended to include ionization by x-rays. Nor did he take
seriously McClelland's claim that the triggering hypothesis also
explains the high velocity of secondary electrons stimulated by
y-rays. Bragg was convinced that y-rays are particles, and he sus-
pected as much for the hard x-rays. He had also become aware of
the high secondary electron velocities that Dorn measured after
irradiating matter with x-rays.38 The high velocity of secondary
electrons was easily understood if one assumed that x-rays and
y-rays are particles. The evidence that others used to show that the
triggering mechanism operated convinced Bragg that the rays
could not be aether pulses at all. Viewed this way, the evidence of
secondary electron velocity constituted a new paradox for the
pulse theory.
37
W. H. Bragg, AusAAS Report, 10 (1904), 47-77.
38
W. H. Bragg, TPRRSSA, 31 (1907), 95.
Since the difficulty deals with the quality, or energy content, of
x-rays and electrons, I shall call it the paradox ofquality. It is a test
of the microscopic validity of energy conservation. Each impact-
ing electron brings kinetic energy in an amount determined by its
velocity. That energy reappears in the form of a transverse impulse
that propagates, it would seem, isotropically outward. The impulse
eventually passes over an atom or molecule. One would expect
that the amount of pulse energy available to the cross-sectional
area of an atom is exceedingly small. The atom subtends an
insignificant angle when viewed across the macroscopic distances
over which the potency of the x-rays is known to persist. The pulse
apparently dissipates itself in all directions. How can the full
energy in the impulse be concentrated on the atom? Yet it seems to
be. The atom releases an electron with an energy virtually equal to
that of the electron whose impact alone created the impulse!
Bragg had already clearly perceived the distinction of quantity
and quality in 1904 when he discussed the energetics of particle
streams. He had objected that the simple phrase amount of [parti-
cle] radiation made no sense.39 He insisted that one had to consider
separately the energy transferred by the particles and the particle
flux itself. In 1904 he transferred this way of thinking to y-rays and
by 1906 was convinced that it also applied to x-rays.
One of the influences on Bragg in this critical period was the
work of his former colleague, Richard Kleeman. Kleeman left
Australia in 1906 for Emmanuel College, Cambridge. There he
began research work at the Cavendish Laboratory under J. J.
Thomson. His first experiments, perhaps still under the influence
of Bragg's concerns, treated an issue closely related to the conser-
vation of quality in y-ray interactions. He showed that the velocity
of secondary electrons produced by the combined action of /?- and
y-rays from a radioactive source far exceeded that due only to the
a-component.40 Interpreted by Bragg, this evidence was signifi-
cant.41 One knew that the largest part of the energy in radioactive
decay is carried by the a-particles. Nevertheless, even the small
fraction of energy carried by the y-component seemed to be con-
centrated on individual electrons.
In March 1907 Kleeman provided more evidence that the action
39
W. H. Bragg, AusAAS Report, 10 (1904), 69.
40
Kleeman, PM, 12 (1906), 273-97.
41
W. H. Bragg, TPRRSSA, 31 (1907), 96.
of the y-rays is concentrated on individual atoms. He showed that
the ionization produced by all three components of radioactive
decay appeared to be strictly an atomic, not a molecular, prop-
erty.42 One first measures the rate of ionization caused by the rays
in a gas consisting only of atoms A, and denotes the result by a;
that found in gas B is called b. Then the ionization rate produced
by the same radiation in a gas consisting of molecules AWBW is
approximately na + mb. To Bragg, Kleeman's result signified
again that all three component radioactive emanations, y in-
cluded, interact with the electrons that comprise the atoms, not
with entire molecules.43
In May 1907, Bragg submitted a paper outlining his new
thoughts to the Royal Society of South Australia. It was titled "A
comparison of some forms of electric radiation." In it Bragg
explained: "when x-rays were first investigated, and again when
y-rays were discussed, it was often suggested, in each case, that the
radiation might consist of material particles."44 The "often" ex-
pressed Bragg's boundless optimism rather than historical fact; the
great penetrating power of the rays had always stood in the way of a
corpuscular interpretation, and few had suggested it. "The diffi-
culty of accounting for the great penetration of these radiations on
the basis of a material interpretation," Bragg explained, "was quite
exaggerated, and even imaginary."45 And Bragg proposed an en-
tirely new view of x-rays and y-rays.
THE NEUTRAL-PAIR HYPOTHESIS

To replace the "overrated" impulse, Bragg proposed the neutral
pair. It was a dipole consisting of an electron coupled with a bit of
positive charge to make the whole electrically neutral.46 In 1907
42
Kleeman, PRS, 79A (1907), 220-33.
43
W. H. Bragg, TPRRSSA, 31 (1907), 96.
44
Ibid., 90.
45
Ibid., 90. Bragg implied that Rontgen had proposed a participate hypothesis of
x-rays, but this was a convenient falsehood. Rontgen had pointed out in
Rontgen III that x-rays are similar to cathode rays. Bragg's own corpuscular
model of cathode rays led him to project a similar interpretation of x-rays on
Rontgen, a view that the latter never held.
46
Andrade has suggested, in RSL Obituary notices, 4 (1943), 280, that Bragg's
idea of neutral pairs may have been derived from Lenard's dipole "dynamid,"
proposed in 1903. Bragg was aware of Lenard's hypothesis but commented
negatively on it in 1904.
88 Ionization and the recognition of paradox, 1906 -1910
Bragg selected the a-particle for the neutralizing role; later investi-
gations showed that its charge was too large, and other candidate
particles took its place. The easy penetration by the pair was
attributed to its electrical neutrality and consequent lack of inter-
action with matter. Some ionization events not only disturb the
gas atom but also disrupt the pair. When this happens, the heavy
a-particle moves off slowly, but the less massive electron continues
with essentially the same high velocity it possessed within the x-ray
pair.
At one blow, Bragg solved the paradoxes of both quantity and
quality. In his own words: 47
If the x-ray is an aether pulse, it is difficult to understand . . . why the
spreading pulse should only affect a few of the atoms passed over [in
ionization], why the secondary cathode rays are ejected with a velocity
which is independent of the intensity of the pulse which weakens as it
spreads, and why [the pulse] should be able to exercise ionising power
when its energy is distributed over so wide a surface as that of a sphere of,
say, ten or twenty feet radius.
Like the triggering hypothesis, Bragg's new hypothesis required no
transformation of radiant to mechanical energy. The energy of the
neutral pair is mechanical to begin with. The high velocity of the
electron exists before ionization of an atom. But instead of arising
from the internal dynamics of the atom, the electron is released to
move independently, where before it had constituted half of a pair.
Bragg had good reason to suggest that the high-velocity electrons
were not secondary products of ionization. Not just a-, /?-, and
y-rays from radioactive decay, but also their electrical analog
canal, cathode, and x-rays induce conductivity in gases. The vast
majority of charge-carrying particles produced in gases by these
radiations are electrons moving at about io 8 cm/sec. So ubiquitous
are these slow electrons that Bragg adopted the special name d-rays
applied to them by others. 48 "In all cases," Bragg said, "the bulk of
the ionization which the rays effect is of the same character. [It]
consists in the displacement of slow moving electrons, or S rays,
from the atoms of the gas." 49 In this sense, all six forms of radiation
47
W. H. Bragg, TPRRSSA, 31 (1907), 91-2.
48
J. J. Thomson, PCPS, 13 (1904), 49-54. This use of S has, of course, no
connection with my use of S to denote pulse width.
49
W. H. Bragg, TPRRSSA, 31 (1907), 79.
were "electric," and this is how Bragg intended the term in his title.
That the velocity is common to electrons released by all forms of
radiation convinced Bragg that it was an atomic property. The
slow electrons are the true product of atomic ionization, he
claimed; the high-speed electrons are more likely remnants of the
incident neutral-pair x-rays.
Bragg's study appeared in two parts in Australia. They were
published back-to-back but were separated by a month's careful
consideration. In the first part, where he discussed the evidence
regarding J-ray electrons, Bragg made it clear how seriously he
took Marx's demonstration in 1905 that x-rays travel with the
velocity of light. "If we attempt to explain the properties of the
X-rays on the supposition that it is a neutral pair," he complained,
"we meet with a difficulty which does not occur in the case of the
y-ray . . . Unless some way out is found [the hypothesis] must
remain simply an interesting comparison."50 During the month
between the two papers, Bragg realized that only slow electrons
would be susceptible to the detailed control exercised by the
electric oscillations in Marx's experiment. The very fast secondary
electrons would easily travel from the cathode P to the electro-
meter Fin the apparatus shown in Figure 2.13, even if the ambient
field momentarily acted to decelerate them.
In the second part of his study, called "The nature of Rontgen
rays," Bragg proposed that two entities inhere in any x-ray beam.
First, there are electromagnetic impulses that travel at the velocity
of light. Combined with these are the neutral pairs that travel at
somewhat lower velocity. Upon collision, the latter can break into
their constituent parts and release high-velocity electrons. The
pulses, Bragg suggested, might be those produced when the neutral
pairs themselves break away from the anticathode of the x-ray
tube.51 The actual velocity of the pairs could be anywhere within
the limits set by the minimum velocity required to penetrate
matter and the minimum velocity of the released free ^-particles.
Based on Thomson's results, he estimated the range as io8 to io9
cm/sec.
Bragg's hypothesis applied only to x-rays and y-rays. He never
50
W. H. Bragg, TPRRSSA, 31 (1907), 91-3.
51
Note the influence of Rutherford's hypothesis that y-rays are the impulses
produced when ^-electrons are ejected from a disintegrating atom.
intended it to apply to light. The ionizing action of x-rays, he
pointed out, "must be entirely different from that of ultra-violet
light."52 Although Einstein had proposed his lightquantum hy-
pothesis two years before, Bragg was unaware of it and would not
likely have seen the connection to his own ideas in any event.53
Indeed, in disclaiming any connection between x-rays and visible
light, Bragg betrayed the essentially conservative nature of his
appeal; it was intended as even more of a throwback to purely
mechanical theory than the impulse hypothesis. To Bragg, all six
types of radiation - a, /?, y, canal, cathode, and x - were particles,
although some were accompanied by unavoidable aetherial effects
such as impulses.
After satisfying himself that he could answer possible criticism
based on Marx's experiment, Bragg allowed the entire study to
appear in the Philosophical Magazine.5* It was a broadside attack
on the British pulse theory, which at the time was most strongly
supported by Charles Barkla. While Bragg had been developing his
novel hypothesis, Barkla's experiments had turned up new and
remarkably homogeneous secondary x-rays. In its fully interpreted
form, this evidence would prove to be virtually unanswerable by
Bragg. And it was Barkla who took on the responsibility of defend-
ing the impulse hypothesis from Bragg's assault.
HOMOGENEOUS X-RAYS
Barkla had shown in 1903 that x-rays scattered off light elements
reradiate with a quality roughly equivalent to that of the primary
rays. In this way, he sought corroboration of Thomson's simple
picture of the interaction of pulses and matter. In the course of his
polarization studies in 1904 and 1905, he extended his survey to
some of the heavier elements - calcium (Ca), iron (Fe), copper
(Cu), zinc (Zn), platinum (Pt), and lead (Pb) - and found that the
secondary x-rays are consistently less penetrating than the primary
ones.55 He invoked Thomson's "thunder and lightning" clause in
52
W. H. Bragg, TPRRSSA, 31 (1907), 96.
53
"When I first put forward the neutral pair theory I was ignorant of the work of
Einstein . . . I did not think of carry ing over the idea to the theory of light; on
the contrary, I had hopes of proving that no connection existed between the
two kinds of radiation." W. H. Bragg, Studies in radioactivity (1912), 192-3.
54
W. H. Bragg, PM, 14 (1907), 429-49.
55
Barkla, PTRS, 204A (1905), 467-79.
i"
In
g 70 11 \ Sb
\
1 so V
A
i
s
0 y 7 V
;
1
j. 0 10 20 30 40 50 60 70 80 90 100 110 ieO 130 140 150 160 170 180 190 2(fl 210
ATOMIC WEIGHT
Figure 4.2. Barkla's graph of periods in the penetrating power of x-rays

scattered off elements. [PM, 11 (1906), 820.]
explanation: Interionic forces in the solid are presumably responsi-
ble for lowering the penetrating strength of the secondary x-rays by
broadening the secondary impulses.
J. A. McClelland began a fruitful line of investigation in 1905 by
surveying a wide variety of elements for their response to the /?- and
y-rays from radium.56 He did this as part of his investigations,
mentioned earlier, to demonstrate that atomic disintegration is the
mechanism of secondary electron emission. He measured the
intensity of secondary electrons by the ionization rate they in-
duced in a gas sample; he was therefore measuring the number of
electrons released from different elements by the same combina-
tion of /?- and y-rays from radium. He discovered that the number
of secondary electrons increases monotonically with the increasing
atomic weight of the scattering substance. Moreover, the rate at
which secondary electrons increase with atomic weight appeared
to change from period to period among the elements. This was the
behavior that convinced McClelland that the secondary emission
must be an "atomic" property caused by a mechanism analogous
to the triggering hypothesis.
Barkla was encouraged by McClelland's work, and he tried to
discover a similar reaction between the penetrability of the second-
ary x-rays (as distinct from secondary electrons) produced in
matter by a primary x-ray beam. A survey in 1906 turned up the
rough relationship illustrated in Figure 4.2: Increases in the pene-
56
McClelland, RDS Sci trans, 9 (1905), 1-8.
60 80 100 ZOO 210
ATOMIC WEIGHT
Figure 4.3. J. J. Thomson and Kaye's graph of ionizing strength of

secondary electrons released from elements under x-ray bombardment.
[PCPS, 14(1906), 113.]
trating power correspond to the periodic properties of the ele-

ments.57 Barkla admitted that the instability of his primary x-ray
beam rendered the results somewhat unreliable, but he was firmly
convinced that he had found a means to determine the order of the
elements in a way independent of the atomic weight.
Thomson quickly saw the power of the new survey, and with
G. W. C. Kaye he studied the secondary electrons produced by
x-rays.58 They discovered that the ionizing strength of the second-
ary electrons increases monotonically with the atomic weight of
the scattering substance. Figure 4.3 illustrates their result. McClel-
land had found a similar behavior of electrons produced by /?- and
y-rays. Thomson found that, according to x-ray analysis, nickel
was improperly placed when ranked according to atomic weight in
the periodic table.
The next year, 1907, Kleeman carried the investigation into the
realm of y-rays and was careful to separate the primary gammas
from their accompanying betas. He then repeated McClelland's
experiment of two years before by measuring the relative ioniza-
57
Barkla, PM, 11 (1906), 812-28.
58
J. J. Thomson, PCPS, 14 (1906), 109-14.
13(10
12m
1HK)
AirJ
1000
J
I*
>n
ir
tot.
>
T 1
c
p
Ntt
14 ^ 4L' rntf S4 iS 111' ll'ti Hi"
Atomic Weip-ht.
Figure 4.4. Kleeman's graph of ionizing strength of all secondary radia-

tion from elements bombarded by y-rays. [PM, 14 (1907), 624.]
tion caused by the secondary electrons.59 His graph, Figure 4.4, is

remarkably similar to Thomson's, and Kleeman took this as yet
another proof that x-rays and y-rays are equivalent. Fully under
the Thomson influence by this time, he concluded in August 1907
that both radiations were "probably" electromagnetic impulses.60
In the meantime, Barkla enlisted the assistance of his Liverpool
physics demonstrator, Charles Sadler, in order to investigate fur-
ther the anomalous case of nickel found by Thomson. They
discovered that most of the secondary x-rays spent their ionizing
power after traversing a few layers of aluminum. But among those
that persisted was a component of relatively high intensity that
always lost the same fraction of its ionizing power in passing
through successive test thicknesses of aluminum.61 This compo-
nent followed the absorption law, equation 4.1, with a constant
value of//, clear evidence of homogeneous penetrating power. Not
only were some secondary rays homogeneous, but they showed a
59
Kleeman, PM, 74(1907X618-44.
60
Ibid., 643.
61
Barkla and Sadler, PM, 14 (1907), 408-22. Barkla, Nature, 75 (1907), 368.
particular transparency for the material of their origin. Secondary
rays from copper, for example, passed with lower than normal loss
through copper. Barkla called this a "property which appears to be
unknown among x-ray beams hitherto experimented upon."62
The announcement of this new sort of x-ray came in September
1907. But the full significance of Barkla's discovery only emerged
slowly over the next year as his investigation proceeded. Barkla
was encouraged to develop and extend his interpretation of the
result by the controversy that quickly developed over Bragg's
paper on neutral pairs. It appeared in England just a month after
Barkla's paper on homogeneous x-rays.
THE BRAGG-BARKLA CONTROVERSY

The conflict between Bragg and Barkla arose largely from the fact
that the protagonists approached the common domain of x-rays
from opposite directions. Bragg's view had developed out of de-
tailed studies of the properties of a-particles, the results of which
were first applied to y-rays and by extension to x-rays. Barkla's
research was concerned exclusively with x-rays, and not particu-
larly hard x-rays at that. The two had very little common ground
from the very start of their discussion. Barkla was dismayed that
Bragg could suggest what appeared to him to be so implausible an
idea as neutral pairs. Bragg frequently exploited the impassioned
tone in Barkla's defense of the impulse hypothesis. The debate was
carried on in the pages of Nature in 1908; the details have been
presented elsewhere, and I will discuss only those of direct concern
to the broader issues of our study.63
Bragg had attempted to reinterpret Barkla's experimental results
according to the neutral-pair hypothesis. He had suggested that the
penetrating power of a pair was, at least in part, determined by the
separation of its constituent charged particles, which in turn fixes
the dipole moment of the pair.64 Soft x-rays, produced at relatively
62
Barkla and Sadler, PM, 16 (1908), 560.
63
The interchange fills many letters to Nature, 77 and 78. Barkla's are virtually
all titled "The nature of x rays" and Bragg's "The nature of y and x rays." See
Stuewer, BJHS, 5 (1971), 258-81. Barkla had all the evidence except unpolar-
izability for characteristic x-rays before the debate with Bragg started. The
controversy led him to reinterpret data that already existed.
64
Bragg left ambiguous whether the speed of the pair or its moment determines
its penetrating power. He wished to attribute it solely to velocity, but had to
invoke the change in moment to explain Barkla's scattering results.
low discharge potentials, are poor penetrators because, said Bragg,
they have large moments. Hard x-rays do not ionize gas atoms
readily because they possess small moments. On striking a rela-
tively light atom able to absorb some of the collision energy by
rebounding, a neutral pair might simply change its course with no
change in dipole moment. But on striking a heavy atom, it could
suffer either an increase in moment or be "shattered altogether." If
shattered, it releases its captive high-speed electron. In this way
Bragg explained Barkla's discovery that x-rays scattered from light
atoms exhibit roughly the same penetrability as the primary rays,
whereas secondary x-rays from heavy atoms are less penetrating
than their primaries.
Bragg's dipoles also rotated, and the plane of rotation of the pair
was conserved. Bragg exploited this capability to explain Barkla's
polarization data, and it was here that the opening shots in the
debate werefired.65Bragg assumed that electric oscillations within
an atom occur only in afixedplane. An atom can absorb a neutral
pair, but only if the planes of rotation of the atom and of the pair
are parallel. The polarization of secondary x-ray beams followed
directly from this constraint between the two interacting planes of
rotation. An atom, Bragg said, can eject a pair only into the plane
of the atom's oscillation.
Barkla carried the consequences of Bragg's position further to
argue that an x-ray beam from an anticathode could consist only of
pairs rotating in planes parallel to the beam direction. The second-
ary radiation stimulated by this beam, when viewed at right angles
to the primary beam, could then consist only of pairs rotating in
the plane defined by both beams. If the atoms in the scattering
material are randomly oriented, the net flux of secondary neutral-
pair x-rays should be "proportional to the density of the lines of
longitude" (like those running from pole to pole in Figure 4.5), that
is, as the secant of the latitude. So, Barkla reasoned, the secondary
rays seen at A, only slightly off the beam axis, should be many
times more intense than those scattered at 900 to the beam, seen at
B. He estimated the difference to be about 8:1, then allowed for
possible mismatch between the planes of atom and pair, and
revised the ratio to 4:1.
On the other hand, Thomson's proposed mechanism for the
stimulation of secondary x-rays predicted that an unpolarized
65
Barkla, Nature, 76 (1907), 661-2.
Figure 4.5. Barkla's prediction of x-ray scattering according to Bragg's

neutral-pair hypothesis.
primary x-ray beam would produce electron displacements only in

the planes perpendicular to that beam. Secondary x-ray impulses
radiated by these electrons in the direction of the primary beam
would have the benefit of all electron accelerations as sources. But
impulses viewed from a direction normal to the primary beam are
caused only by the component of electron displacement that is
confined to the single plane normal to both directions. A mo-
ment's reflection will show that the intensity of secondary x-rays
directed at right angles to the primary beam is just half that
directed back along the primary beam. Barkla concluded that the
intensity distribution of secondary impulses should approximate
that shown in Figure 4.6, and that the ratio for observations at A
and B could not exceed 2:1.
Experiment showed that soft x-rays yield ratios approaching
1.85:1. For hard x-rays, the ratio was less. Barkla described this
result as "most conclusive evidence against the 'neutral pair'
hypothesis."66 But it scarcely supported the impulse hypothesis
any better for hard x-rays. The test also raised questions about the
recently discovered homogeneous secondary x-rays from heavy
atoms. When, in defense of the pulse hypothesis, Barkla extended
his analysis to heavy atom scatterers, he found that the ratio
66
Barkla, PM, 15 (1908), 288-96.
Figure 4.6. Barkla's expectation of scattered x-ray intensity according to

Thomson's hypothesis.
approached i: i. The homogeneous secondary x-rays are isotropi-

cally scattered, in accord with neither prediction!
Bragg was not at all deterred by Barkla's argument. He pointed
out that one was not entitled to assume isotropic scattering of pairs
into the plane defined by atomic vibration. To do so would require
detailed knowledge of atomic structure, "as to which it is scarcely
possible to do more than speculate."67 He responded with a "deci-
sive argument" based on his recent experiments with J. P. V.
Madsen on the intensity distribution of secondary rays following
irradiation by y-rays.68 If y-rays are transverse impulses, Bragg
argued, the secondary rays, whether material electrons or aetherial
impulses, should reradiate with equal probability in the forward
and backward directions. (Thomson had made a similar assump-
tion the previous year.69) Neutral pairs, on the other hand, would
be expected to conserve their incident momentum; secondary rays
should preferentially scatter in the forward direction.
Bragg and Madsen's experiments showed unmistakable evi-
dence for an asymmetric emission of secondary electrons. Bragg
used an ionization chamber (Figure 4.7) closed at the top and
bottom by double plates of lead (Pb) and aluminum (Al). Bragg
and Madsen showed that with the aluminum sheet against the
67
W. H. Bragg, Nature, 77(1908), 560.
68
W. H. Bragg, Nature, 77 (1908), 270-1. Bragg and Madsen, TPRRSSA, 32
(1908), 1-10; PM, 75(1908X663-75.
69
J. J. Thomson, Conduction of electricity, 2nd ed. (1906), 4056. Asymmetry had
been observed for scattered ^-electrons by MacKenzie, PM, 14 (1907), 176 - 87.
Figure 4.7. Apparatus used by W. H. Bragg and Madsen to demonstrate

asymmetric emission of secondary electrons after y-irradiation. [PM, 15
(1908), 665.]
chamber at the top, the ionization was slightly greater than when
the two top plates were interchanged. This was so regardless of the
order of plates on the bottom. Kleeman's finding that a metal
emits secondary rays in rough proportion to its atomic weight thus
implied that the net ionizing radiation scattered in the "through"
direction was greater for aluminum than for lead.
When the order of plates was changed on the bottom with no
change on the top, 44 percent larger currents occurred with lead
closest to the chamber. Thus, the number of rays scattered back,
against the direction of the primary beam, was greater for lead than
for aluminum. Since all four layers were too thin to absorb appre-
ciable y-rays, either aluminum or lead or both emitted more
secondaries in the forward compared to the backward direction.
This, they claimed, was "fatal" to any transverse impulse or wave
theory of y-rays.
Madsen soon claimed that what he took to be secondary y-rays
also are asymmetrically emitted from matter.70 Bragg completed
the quatrain with an investigation of the weaker effect in secondary
x-rays. There, deep within Barkla's own x-ray research domain,
70
Madsen, TPRRSSA, 32 (1908), 163-92; PM, 17 (1909), 423-48. For the
conditions under which Madsen carried on his researches, see his correspon-
dence with Bragg in Home, Historical records ofAustralian science, 5:2 (Nov.
1981), 1-29.
71
Bragg found the asymmetry he sought. But Bragg, always confi-
dent and somewhat parochial, had quite neglected the momentum
carried by an electromagnetic wave. The error was first mentioned
by Charlton Cooksey at Yale when he presented evidence of an
equivalent asymmetry in the distribution of secondary electrons
stimulated by x-rays.72 For the moment, nothing was made of the
oversight.
Bragg's commitment to the neutral-pair hypothesis led to a
second important discovery in 1908. In May he listed six properties
of the y-rays that, he claimed, were best interpreted in terms of the
neutral-pair view. One, on the transformation of y-rays, had pre-
viously been buried in raw experimental data: "the penetration
and therefore the speed of the /? radiation . . . produced [by
y-rays] increases with the penetration of the y radiation to which it
is due."73 All neutral pairs of equivalent penetrating power should
release electrons of equal speed. The velocity of the secondary
betas should be independent of the number of neutral pairs and
should depend only on their penetrability.
This result was difficult for any impulse hypothesis to explain. If
one adopted the triggering hypothesis, the velocity of secondary
electrons should depend on the type of scattering material. Bragg's
conviction led directly to a test, and he was vindicated.74 As with
the photoelectric effect, measured velocities of the secondary fi-
particles were independent not only of the intensity of the incident
y-radiation but also of the nature of the scattering material.
Bragg was convinced that the neutral-pair theory explained "all
the known properties of the y-rays much more simply and com-
pletely" than did the impulse hypothesis.75 He conveniently sum-
marized his position in a series of largely rhetorical questions
related to the possible origin of secondary betas: (1) both the
particle and the energy arise from the atom (in essence, the trigger-
71
W. H. Bragg and Glassen, TPRRSSA, 32 (1908), 301-10; PM, 17 (1909),
855-64.
72
Cooksey, Nature, 77 (1908), 509-10. Thomson had pointed out in 1907 that
x-ray impulses would carry momentum. In the "bright speck" hypothesis of
light, to which we shall return in Chapter 6, he claimed that x-rays would "have
all the properties of material particles." PCPS, 14 (1907), 424.
73
W. H. Bragg, Nature, 78 (1908), 271.
74
W. H. Bragg and Madsen, TPRRSSA, 32 (1908), 35-54; PM, 16 (1908),
918-39.
75
W. H. Bragg, Nature, 78 (1908), 271.
ing hypothesis); (2) the particle comes from the atom but its energy
is transformed from the incident radiation (the impulse or wave
hypothesis) and (3) both the electron and the energy come from the
primary y (the neutral-pair hypothesis). Of (1) he inquired, how
could the penetrability of the y-ray affect the energy of the ^-parti-
cle at all? Also, why does the energy not vary with the type of atom?
Finally, how could one explain the asymmetry in secondary emis-
sion, as if the y-rays found the "guns pointed in the direction in
which they are travelling themselves?"76 Explanation (2) hardly
fared better. To use one of Bragg's favorite expressions of later
years, it was as if a plank dropped from a height of 100 meters into
the sea sent out a circular impulse that, after spreading over
thousands of kilometers, concentrated its effect on another plank,
giving it sufficient energy to impel it 100 meters into the air! The
y-ray or x-ray somehow delivered its energy bundle to a single
electron. But replace the bundle by a neutral pair, he exclaimed,
"and the whole affair seems simple enough."77 A cathode-ray
electron, incident on the anticathode, picks up a neutralizing
positive charge and becomes an x-ray or y-ray. On collision it
breaks into its component parts, and the electron continues on
with close to its original velocity.
CHARACTERISTIC SECONDARY X-RAYS

Throughout the debate, Barkla stood firm by the x-ray evidence
and refused to be drawn into a discussion of more penetrating
radiations. He believed his position to be secure for soft x-rays.
Bragg's data were largely drawn from the behavior of the highly
penetrating y-rays. Each felt confident in his own domain. Barkla
never answered the paradoxes of ionization, and Bragg always
admitted the difficulty presented by the partial polarization of
x-rays.
In response to Bragg's second attack, Barkla retreated to his
strongest evidence: homogeneous x-rays scattered from matter. In
1908 he made it clear that there are two distinct types of x-rays.
"One," Barkla explained, is "a scattered radiation produced by the
motion of electrons controlled by the electric force in the primary
76
W. H. Bragg and Madsen, PM, 16 (1908), 934.
77
Ibid., 936.
Rontgen pulses," and is therefore inhomogeneous. These scattered
x-rays are polarizable; this is a consequence of the unidirectional
cathode beam that produced the primary x-rays. The other type of
x-ray is, according to Barkla, "a homogeneous radiation charac-
teristic of the element emitting it, and produced by the motion of
electrons uncontrolled by the electric force in the primary
pulses."78
The emphasis on the characteristic nature of the secondary
x-rays was new and was offered largely as a challenge to Bragg.
Barkla had found that these homogeneous rays characteristic of
heavy elements are emitted isotropically, and they show no incli-
nation to polarize. Bragg was hard-pressed to find any explanation
using neutral pairs for the characteristic nature of the secondary
rays. Neutral pairs are, by definition, independent of the atoms
that scatter them. The hypothesis denies detailed interaction be-
tween a pair and an atom; it consequently cannot explain how the
pairs scatter so differently depending on the weight of the atom.
The discovery of homogeneous x-rays forced careful study of
their properties; it was quickly recognized that the results of all
prior experiments might have been affected by their selective but
strong influence.79 They also provided a potential source of x-rays
of definite penetrating power needed for detailed analysis of x-ray
absorption. Anticipating that fine distinctions in penetrating
power might be discerned, Barkla discovered that the secondary
rays from some elements - silver (Ag), tin (Sn), antimony (Sb),
and iodine (I) - each possess two distinct components. He called
the less penetrating component A, the more penetrating one B,
and realized that the periodicity he had charted the year before
(Figure 4.2) was due to the overlap of the two species of secondary
x-ray (cf. Figure 4.8). Within two years he had renamed the species
L and K; he thought it "highly probable that series of radiations
both more absorbable and more penetrating exist," and he wanted
to leave room on both ends for additions.80
Unlike Bragg, Barkla could at least hint at a plausible mecha-
nism for the production of the characteristic secondary x-rays.
When a pulse disturbs an atom, electrons within it should oscillate
78
Barkla and Sadler, PM, 16 (1908), 576.
79
Barkla and Sadler, PM, 17 (1909), 739-60. Barkla and Nicol, Nature, 84
(1910), 139. Barkla, PCPS, 15 (1909), 257-69.
80
Barkla, PM, 22 (1911), 4o6n.
Atomic Weight of Radiator
Figure 4.8. Barkla's graph showing two components of different pene-

trating power in secondary x-rays. [PCPS, 15 (1909), 259.]
at frequencies determined by the atomic structure. The radiated

impulse that results can be thought of as compounded of a large
number of secondary partial oscillations. The net effect could be a
secondary impulse with a width determined entirely by the atom.
This particular pulse width is relatively intense because each atom
in the sample gives rise to it.
The most remarkable property of the new homogeneous sec-
ondary x-rays, only hinted at in earlier studies, became clear after
Barkla moved from Liverpool to London in 1909. His former
colleague, Charles Sadler, continued the investigation of precise
conditions for the stimulation of the characteristic x-rays.81 Using
the fine distinction in the penetrating power of the secondary
x-rays to excite selected tertiary characteristic x-rays, he corrobo-
rated in detail his suspicion that characteristic rays can be stimu-
lated only by an incident radiation of lower absorption coefficient
than the anticipated secondary. If the primary beam does not meet
this requirement, only the inhomogeneous scattered radiation is
detected.
Barkla's hypothesis for the origin of the characteristic secondary
x-ray impulses partially supported the uncorroborated association
of pulse width with penetrating power. The rays should have equal
pulse widths because all rays are produced from the same mixture
of inhomogeneous pulses by identical atoms. To Barkla and Sadler
this meant that each of the new homogeneous pulses carries the
81
Sadler, PM, 18(1909), 106-32.
same energy; x-ray intensity for homogeneous rays is then propor-
tional to the simple number of impulses. Here was a reaffirmation
of the reinterpretation of radiant intensity provided by the impulse
hypothesis. Bragg had defined the quality of neutral pairs in terms
of the dipole moment and the rotational velocity of a pair; the
quantity was simply thefluxof pairs. Many pulse theorists came to
define the intensity of an x-ray beam (whether composed of char-
acteristic rays or not) in terms of only the number of pulses passing
a point in space, the quality of each impulse as a function solely of
pulse width.
Seen in this light, the characteristic x-rays took on new signifi-
cance. Their stimulation required a primary beam of greater pene-
trating power than they themselves had. The necessary threshold
could now be expressed in terms of pulse width. This, in turn, was
analogous not to the amplitude of a periodic wave but to the
wavelength. The condition found by Sadler for the stimulation of
characteristic x-rays appeared to many, especially those outside
Britain, to be analogous to Stokes' law offluorescence.Just as light
of a given wavelength can be stimulated only by light of a shorter
wavelength, so characteristic x-rays can be stimulated only by
other x-rays of shorter pulse width. Some physicists began to speak
of the characteristic rays as fluorescent x-rays for this reason, but
Barkla resisted the identification before 1911. He admitted that
this "extension of Stokes' law of fluorescence" was "the most
difficult point to explain" about secondary x-ray pulses.82
Barkla's influence was at its peak in the years 1908-11. He was
confident that the sum of the evidence on homogeneous secondary
x-rays "strikingly verified Thomson's scattering theory" of
pulses.83 Thomson himself thought that the continued study of
characteristic x-rays was "of the utmost importance."84 Three of
Barkla's most comprehensive studies were translated into German
and published in the Jahrbuch der Radioaktivitdt und Elek-
tronik}5 The interpretation placed on his results by some Ger-
mans, most notably the editor of the Jahrbuch, Johannes Stark,
was quite different from Barkla's, as we shall see next.
82
Barkla, PCPS, 15 (1909), 269.
83
Ibid., 257.
84
J. J. Thomson, Encyclopaedia Britannica, 1 ith ed, 23 (1911), 695.
85
Barkla, JRE, 5 (1908), 246-324; 7(1910), 1 -15; 8 (1911), 471-88.
The appeal in Germany to the
quantum theory
The unit of light energy does not travel in all directions

away from an oscillator, but only in a single [direction.]1
German-speaking physicists arrived at an explanation for the

problems facing radiation theory different from that of their Brit-
ish colleagues. Moreover, their attempt to discover a consistent
theory formed a complete contrast to the British attempt. The
problems themselves were perceived differently. It would be an
exaggeration to say that the problems ever reached the same status
as paradoxes in Germany as they did in Britain. Physics as prac-
ticed in Germany was not as dependent on conceptual pictures,
and physicists there were much more willing to adopt formal
principles to solve the difficulties without demanding a consistent
physical interpretation.
By 1905 the very word mechanics meant to many Germans
something essentially different from its meaning in Britain. Influ-
ential voices had proclaimed that matter itself is only a construct of
the mind, fashioned out of the more ontologically significant
electromagnetic forces that giveriseto human perceptions of mass
and extension. The more influential Germans were not closely tied
to the logical requirements of mechanistic thought. When the
electromagnetic impulse hypothesis of x-rays was accepted in
Germany, it was interpreted by many as a further example of the
versatility of an electromechanical ontology. Electromagnetic im-
pulses were considered by some, most notably those who first tried
to make sense of x-ray behavior, as a form of wave with extended
properties in space. Consequently, as we shall see in this chapter,
the sharp distinction between x-rays and periodic light was never
made as strongly in Germany as it was in Britain.
1
Stark, Elementare Strahlung (1911), 262.
The appeal in Germany to the quantum theory 105
The most far-reaching proposal to reformulate classical ideas
about radiation was not concerned with x-rays at all. Albert Ein-
stein, an unknown patent clerk lacking advanced degrees and
totally outside the profession of academic physics, proposed in
1905 that ordinary light behaves as though it consists of a stream of
independent localized units of energy that he called lightquanta.
He was led to this revolutionary view by a statistical analysis of the
properties of an idealized sample of radiation in thermodynamic
equilibrium. The statistical mode of reasoning was not common in
physics at the time, and the resulting "heuristic" description of
lightquanta seemed to virtually all physicists to be simply unten-
able.
Before reading about Einstein's lightquantum, Wilhelm Wien
introduced a significant change in the way energy is reckoned in
electromagnetic impulses. Building on an analysis by Max Abra-
ham of the energy radiated from uniformly accelerated electrons,
Wien recognized that impulse width as well as pulse amplitude
varies with the angle from the direction of the source of accelera-
tion. Wien began to use the impulse width, rather than the classi-
cally orthodox field amplitude, as the parameter by which to
specify the energy content of the impulse.
Shortly thereafter, Wien and Johannes Stark each applied the
new quantum principle to x-rays. To do so, they had to describe
the impulse energy in terms of a frequency; this was possible
because of the previous revision Wien had suggested in the signifi-
cance of the pulse width. Arnold Sommerfeld came to the defense
of classical electrodynamics, but he also justified the nonclassical
use of pulse width tofixthe penetrating power of x-rays radiated in
a given direction. The impulse hypothesis of x-rays, as interpreted
in Germany, was a critical conceptual stepping stone between
classical and quantum definitions of the intensity and frequency of
radiation.2
PHYSIK OHNE DEN ATHER

There were two opposing theoretical programs in early-twentieth-
century Germany designed to resolve the inconsistencies con-
fronted in the interaction of aether and matter. The most widely
2
Wheaton, HSPS, 7/(1981), 367-90.
106 Ionization and the recognition ofparadox, 1906 -1910
accepted and developed was the electromagnetic view of nature in
its various forms. As mentioned in Chapter i, the most radical
framers of this program wished to dispense entirely with the
concept of matter and to reformulate mechanical properties solely
in terms of electrodynamics. The primacy of force was asserted;
many German physicists believed that progress was being made
when the gross concept of matter could be traced back along the
Kantian chain of perception to the forces that underlay it. Diamet-
rically opposed to this aether ontology was the theory of relativity
proposed by Albert Einstein in 1905.3 Einstein reformulated New-
tonian dynamics and resolved the inconsistencies that compli-
cated understanding of the aether-matter interaction by eliminat-
ing the aether altogether. Einstein's solution was ultimately the
more successful.
An integral part of Einstein's rejection of the medium for light
waves was his suggestion of the lightquantum hypothesis.4 He
began by questioning the baseless assumption that there must be a
strong conceptual distinction between light and matter - light
being thought to be waves spreading through a medium, matter to
consist of localized particles.5 His suggestion of the lightquantum
hypothesis arose from the close analogy he perceived between the
behavior of radiation and the behavior of a gas.6 His point is easily
understood today, although the statistical form of his argument
was unusual at the time.
Einstein showed that the entropy of an idealized sample of
radiation enclosed in a cavity of known volume could be obtained
directly from the energy density. In the mid-1890s, Wien had given
a possible form for energy density as a function of frequency and
temperature.7 Einstein recognized that Wien's law was valid only
for high-frequency light. But he used it to calculate the change in
entropy S to be expected in the radiation within the frequency
range v to v + dv if the occupied volume in the cavity was de-
creased from Vo to V:
^ (5.1)
3
McCormmach, HSPS, 2 (1970), ix-xx, 41-87. Miller, Relativity (1981).
4
Einstein, AP, 77(1905), 132-48.
5
See Klein, Natural philosopher, 2 (1963), 59-86; 3 (1964), 3-49.
6
Klein, Science, 757(1967), 509-16.
7
Wien, AP, 52(1894), 132-65; AP, 58 (1896), 662-9.
Here E is the total energy contained in the sample and b is a
constant.
According to Ludwig Boltzmann, the entropy change at con-
stant temperature in a gas of particles can be expressed as a
function of the gas constant R and Avogadro's number N. It is
S-SQ = ^ln W (5.2)
where W'\s the probability that a state of the particles of entropy S

will occur relative to the likelihood of a state of entropy So. Clearly,
the probability that, at a random instant, one molecule will be in a
sub volume V smaller than the cavity volume Vo is the ratio V/Vo.
The relative probability that n molecules will be found in volume
Kis thus W = (V/V0)n. Einstein therefore rewrote equation 5.2 as
Rn
S-So = In (V/Vo) (5.3)
Noting the similarity in form between equations 5.1 and 5.3,

Einstein remarked that "the entropy of a monochromatic radia-
tion of sufficiently low density follows the same laws of variation
with volume as does the entropy of an ideal gas or a dilute
solution."8 The application of Boltzmann's statistical formulation
of the second law of thermodynamics led directly to the surprising
result that the energy of the radiation might be quantized. Noting
that E/bv for the radiation corresponded to Rn/N for the gas,
Einstein concluded that monochromatic radiation "behaves in a
thermodynamic sense as if it consisted of mutually independent
energy quanta of magnitude [Rbv/N]."9
Einstein went on in this epochal paper to sketch the observable
consequences of his hypothesis. He pointed out that Stokes' law of
fluorescence, a relationship that had never been fully understood,
followed immediately from the fact that the energy of each light-
quantum is proportional to its frequency. Light of a given fre-
quency can never stimulate light of a greater frequency because of
simple conservation of energy. He also drew some verifiable con-
clusions about photochemical reactions. But for our purposes, the
prediction of the functional form of the photoelectric relation is
Einstein, AP, 77(1905), 139.

Ibid., 143.
the most significant consequence of the lightquantum. If light
consists of localized quanta of energy Einstein argued, an electron
in an atom will receive energy from only one lightquantum at a
time. Monochromatic light of frequency v can therefore grant
electrons only energy Rbv/N, or, as we express it today, hv, where h
is Planck's constant. If one supposes that some small part p of that
energy must be used to release the electron from the metal itself, all
electrons of charge e so released will be stopped by a decelerating
potential P, following the relation
Pe = hv-p (5.4)
At the time Einstein made this prediction, very little was known
about the precise form of the photoelectric relation. The triggering
hypothesis was gaining wide acceptance, and it effectively ren-
dered close experimental study of the relation unnecessary.10 Note
that Einstein's formulation also explains the observation that had
led Philipp Lenard to the triggering hypothesis in the first place.
The maximum velocity of photoelectrons must be independent of
the light intensity according to the lightquantum hypothesis; only
the frequency of the light determines how much energy any elec-
tron may receive. The intensity of light becomes, in Einstein's
view, equivalent to the total lightquantum flux: the number of
lightquanta that pass through a unit area in unit time. The size of
the photocurrent should vary with light intensity, but the stopping
potential should not. A full decade of uncertainty about the experi-
mental form of the photorelation was to ensue before Einstein's
predicted equation 5.4 was finally and completely verified.
There are two other aspects of Einstein's study that we should
note. First, Einstein used Wien's distribution function to make the
calculation; he did not use Planck's. The following year, Einstein
showed that Planck had erred in his derivation of the blackbody
law, and could have arrived at his new, successful relation only by
implicitly adopting the lightquantum hypothesis.11 Recent schol-
arship has suggested that Einstein's lightquantum theory owed
very little to Planck's work. It has been shown that Planck placed
no proscription on the energies of material oscillators in 1900; they
could possess and transfer energy in a continuum of values. Ein-
10
Wheaton, Photoelectric effect (1971); HSPS, 9 (1978), 299-323.
11
Einstein,^, 20(1906), 199-206.
stein's paper of 1905 therefore marks the origin of the quantum
theory as we understand it today.12
Second, Einstein's paper contains no reference to x-rays or to
y-rays. Einstein did note that the properties of light he wished to
emphasize were most pronounced at high frequency, where
Wien's law was valid. But he did not suggest, nor is he likely to
have believed, that similar considerations should be applied to
x-rays. X-rays and periodic light were then thought to be essen-
tially different forms of radiation. Viewed retrospectively, Ein-
stein's lightquantum hypothesis solved the dual paradox just then
being recognized for any classical theory of x-rays. But it also
introduced serious difficulties for radiation theory that most physi-
cists justifiably thought insurmountable.
Einstein's hypothesis of lightquanta was not taken seriously by
mathematically adept physicists for just over fifteen years. The
reasons are clear. It seemed to be an unnecessary rejection of the
highly verified classical theory of radiation. Most physicists re-
jected the idea as merely the resuscitation of the discredited emis-
sion theory of light. In the domain of visible or ultraviolet light, not
even a hint of the x-ray paradoxes then existed. How lightquanta
could possibly explain interference phenomena was always the
central objection. Einstein offered very little explanation; reconcil-
iation of the lightquantum with interference was a problem with
which he wrestled for the rest of his life. In 1921 he characterized
the conflict to his friend Paul Ehrenfest as something fully capable
of driving him to the madhouse.13
Although Einstein's theory of relativity was soon accepted by
leading physicists, his lightquantum hypothesis was not. Einstein's
own study of the problems of radiation had almost no influence on
the work done by others before the early 1920s. For this reason, we
shall not discuss it in further detail here. Some very useful histori-
cal discussions have been provided, but it is a story yet deserving
scrutiny.14
12
Kuhn, Black-body theory (1978), argues that Einstein and Ehrenfest first
quantized the energy of oscillators in 1905-6. For another view, see Klein,
AHES, 1 (1962), 459-79-
13
Einstein to Ehrenfest, 15 March [1921]. AHQP 1, 77.
14
Klein, Natural philosopher, 2 (1963), 59-86. McCormmach, HSPS, 2 (1970),
41-87.
WILHELM WIEN AND THE ENERGY OF X-RAYS
The first person to perceive the connection of Einstein's unortho-
dox lightquantum hypothesis with the nature of x-rays was Wil-
helm Wien, since 1900 ordentlicher professor of physics at the
University of Wlirzburg. Wien was a proponent of the electromag-
netic view of nature. He was used to thinking that the properties of
matter are artifacts of the continuous electromagnetic field, not
consequences of actual material particles. He believed that there
was great value in the idea that the electron mediates between the
events in the electromagnetic field that we call radiation and those
that we ascribe to atoms. Therefore, his concepts were closely
allied to the electron theory of H. A. Lorentz, in which interaction
between matter and radiation occurs only by means of the elec-
tron.15
But Wien deviated from Lorentz in two important respects.
First, Wien thought that the electron itself has no material exis-
tence; its apparent mechanical properties are due exclusively to
electrodynamic effects. Second, he believed that the electron is
compressible, and it was this idea that first directed his attention to
x-rays.16 He gave the electron a dipole moment that could change
with time. In 1904 he treated x-ray impulses as if they arose from
impacts of this deformable electron.17 He thought the impulse
hypothesis was the "most probable" explanation of x-rays but later
that year admitted that "theoretical conceptions of x-rays can in
no way be considered certain."18 He hypothesized that some
3 x io~13 ergs of energy should be radiated away in each electron
impact, and to test this prediction he turned to experiments on the
heat produced in material bodies by x irradiation.
The results appeared in a Festschrift dedicated to Adolph
Wiillner in 1905.19 Wien found by experiment that no more than
one part in a thousand of the heating capacity of the cathode rays
actually appears as x-ray energy. Joseph Larmor had shown in
1897 that the rate at which energy radiates from an accelerated
particle with charge e is:20
15
Hirosige, HSPS, 1 (1969), 151-209.
16
Wien, Archives Neerlandaises (2), 5 (1900), 96-107.
17
Wien, PZ, 5 (1904), 128-30.
18
Wien, JRE, 1 (1904), 215-20.
19
Wien, Festschrift Adolph Wiillner (1905), 1-14.
20
Larmor, PM, 44 (1897), 503-12.
(5 5)
T" '
where a is the mean acceleration. Wien used a more recent form
for the total energy derived by Max Abraham to take account of
what we would now classify as relativistic effects from a rigid
spherical electron accelerating over a time interval r:21
* 71 (5-6)
where k = V i v2/c2. Wien integrated this expression over the

time needed to decelerate the electron uniformly in a distance /.
The total emitted energy thus obtained had to be ôooth of the
kinetic energy of the electron, which in turn may be determined
directly from the discharge tube potential.
To turn this result into a determination of the pulse width, Wien
had to settle one more matter, and in so doing he made a signifi-
cant new observation. "The impulse width depends on the direc-
tion of the emitted wave," he pointed out.22 This is certainly true.
Along the line of electron motion, "the radiated electromagnetic
energy is zero." Classically, the energy radiated along the line of
acceleration is zero because the directed field amplitudes AE and
AH both drop to zero. But Wien implicitly ascribed the penetrating
power, intimately related to the energy content in the pulse, to the
pulse width, not the pulse amplitude. "Even for quite uniform
deceleration and completely equal electron velocities," Wien
pointed out, "the radiation will be inhomogeneous."23 The physi-
cal analog of the pulse width is the wavelength; although the
amplitude of a periodic wave varies with the azimuth, its wave-
length remains constant with the changing direction from which it
is viewed. The pulse width, on the other hand, varies according to
the azimuthal angle. Wien showed that at an angle 6 with respect to
the axis of electron motion, the pulse width is
d = /(cos 6 + 2c/v) (5.7)
21
M. Abraham, AP, 10 (1903), 156.
22
Wien, Festschrift Adolph Wullner (1905), 8.
23
Ibid., 8-9. Wien's unorthodox views reflect the relative lack of interest in x-ray
research in Germany at the time. Rontgen never seriously returned to the
study of x-ray properties, and a perusal of the papers given at the Naturfor-
scherversammlungen between 1900 and 1908 turns up none concerned with
the nature of x-rays apart from those by Marx.
From this relation, Wien found that in the direction perpendicular
to the electron acceleration, the pulse width S had the value
icr10 cm, about 3 percent of that calculated by Sommerfeld from
his analysis of diffraction.
After making this comparison, Wien turned from the produc-
tion of x-rays to an analysis of their absorption. He had expected
that all of the x-rays would "convert themselves to secondary
[cathode] rays."24 At first, he thought that he had shown this to be
true. But before his expanded study appeared in the Annalen der
Physik, he realized that something was amiss. "If one looks more
closely at the mechanism of this transformation," he pointed out,
"difficulties arise in seeing just where the high velocities of the
secondary electrons originate."25 Using the classical definition of
energy content in the pulse, that is, the product of electric and
magneticfieldamplitudes, Wien calculated the maximum velocity
that each x-ray impulse could produce in an electron. For a typical
x-ray tube potential, it was on the order of io~4 cm/sec; Dorn had
shown that typical x-rays release secondary electrons that move as
fast as io9 cm/sec. The electrons moved some thirteen orders of
magnitude faster than Wien expected! All the x-ray impulses over
a very long time interval, on the order of 0.01 sec, would have to
conspire to contribute their energy to the same electron. This
Wien could not believe, and he concluded, "it is therefore impossi-
ble that secondary [electrons] obtain their velocity through direct
acceleration by the x-ray waves."26 Wien had run full tilt into the
paradox of quality.
Wien found only temporary solace in the triggering hypothesis,
suggesting that the incident x-rays simply release electrons that
already possess high velocity because of their membership in the
atomic system. But Wien's allegiance to the triggering hypothesis
was not strong. The photoelectric effect required a set of electrons
in the atom with velocities on the order of io8 cm/sec; x-rays now
required another set of even greater velocity. Furthermore, Wien
believed that serious problems would arise for this explanation
once radiative losses were taken into account. To explain the high
secondary-electron velocities, an atom must contain electrons
moving at about io9 cm/sec. Wien imagined that an electron
24
Wien, Festschrift Adolph Wullner (1905), 7.
25
Wien,ylP, 75(1905), 1003-4.
26
Ibid., 1005.
revolves in the inverse-square potential surrounding a positive
charge; the orbit is of atomic radius icr 7 cm. An electron moving
5 x io7 cm/sec, slower even than electrons released by light in the
photoelectric effect, should, according to equation 5.6, radiate its
5 x io~13 erg at the rate of io"6 erg/sec. The total energy of the
electron is gone in io~7 sec!27 As early as 1905, Wien had concluded
that grave difficulties confronted the triggering hypothesis, not
simply for x-rays but for visible light as well. But he had no better
alternative to offer at the time.
Actually, the problem is not as severe as Wien implied. In the
article in which Larmor derived expression 5.5, he went on to
suggest that radiative losses might drop almost to zero for atomic
systems with more than one electron.28 This would be true as long
as the vector sum of all electron accelerations remains zero, a
situation most easily achieved when two electrons describe the
same circular orbit at opposite ends of a diameter. J. J. Thomson
developed this idea in 1903 for radiation from multielectron orbits,
finding that the energy radiated per electron drops by a factor of
roughly 1,000 for each additional electron in the ring when the
particles move at a velocity of 0.0 ic.29 Each of six electrons in a ring
radiates only 10"17 of the energy predicted by expressions 5.5 or
5.6. But of this work Wien was apparently unaware.
Wien was convinced very early that there is very little difference
between x-rays and periodic light. His use of the term Rontgen-
welle to describe x-ray pulses, as quoted above, indicates the close
connection in his mind of the two forms of radiation. At the same
time as he assumed the similarity of x-rays and light, most British
physicists were emphasizing their differences. Very likely, Wien's
conflation of periodic wave and discontinuous pulse was a product
of his experience in conjuring the localized properties of atoms out
of the continuous electromagnetic field. It is, of course, possible to
describe an impulse as the sum of many periodic waves, each of
which extends to infinity in space. In any event, the connection
Wien saw between x-rays and light, combined with his concern to
find an explanation for x-ray ionization, increased his interest in
any attempt to explain the photoelectric effect on some basis other
than the triggering hypothesis.
27
The radiated power is given on p. 1007 as io6 ergs/sec, but the sign of the
exponent is clearly a misprint.
28
Larmor, PM, 44 (1897), 503-12.
29
J. J. Thomson, PM, 6 (1903), 681.
QUANTUM IMPULSES
In the spring of 1907, Wien found what he had been looking for:
Einstein's unorthodox treatment of the photoelectric effect using
the lightquantum. In all likelihood, Wien discovered Einstein's
paper by himself. He was fully conversant with the literature on
cavity radiation, particularly Planck's work, and Einstein's paper
was closely related. Furthermore, Wien was aware of Einstein's
new relativity theory and recognized early the profound nature of
Einstein's thought. A direct influence may also have contributed to
Wien's awareness of Einstein's quantum studies. Wien's student,
Jakob Laub, completed a dissertation on secondary electrons early
in 1907.30 At about the same time, he prepared a seminar talk at
Wiirzburg on the theory of relativity.31 Wien was sufficiently
impressed to suggest that the younger man travel to Bern to meet
and learn from Einstein.32 This Laub did. After a short visit, he
returned in the spring of 1907 to Wiirzburg. The following year, he
went back to Bern to collaborate with Einstein.33 It is not at all
unlikely that Laub's close early contact with Einstein widened
Wien's appreciation for Einstein's work, even if Laub's enthusi-
asm for Einstein's lightquantum hypothesis provided nothing
new, and was not entirely acceptable, to Wien.
Wien'sfirstuse of the quantum hypothesis, the Planck theory, as
he called it, arose out of his interest in the problems of cavity
radiation. It was an attempt to extend Planck's quantum treat-
ment of heat radiation to the discrete optical emission spectrum.34
Wien sought the conditions under which emission lines might, like
cavity radiation, be expressed solely as a function of intensity and
frequency for a given temperature. But Wien soon recognized
another application of the quantum transformation rule, one that
he saw clearly only after reading about Einstein's treatment of the
photoelectric effect. It might hold the key to the difficulties yet
unresolved for x-ray ionization. "A simple generalization of the
Planck radiation theory to x-rays," he suggested in November
1907, "would not only explain, but demand, the production of the
30
Laub, AP, 23 (1907), 285-300.
31
Laub, AP, 23 (1907), 738-44; AP, 25 (1908), 175-84.
32
Seelig, Albert Einstein (1954), 85-6; Albert Einstein, Leben und Werk (i960),
i2off. For more on the Laub-Einstein connection, but only with regard to
relativity theory, see Pyenson, HSPS, 7 (1976), 83-123.
33
Einstein and Laub, AP, 26 (1908), 532-40; AP, 26 (1908), 541-50.
34
Wien, AP, 23 (1907), 415-38.
high velocities of secondary electrons. [It would] as well provide
yet another way to calculate the wavelength of the active x-rays."35
Note Wien's use of the word "wavelength" to characterize x-ray
impulses. The lack of distinction in his mind between impulses
and waves here yielded a benefit. The quantum transformation
relation requires that radiation be expressed in terms of a fre-
quency. An impulse has no frequency. But Wien could now
interpret the inverse pulse duration r as analogous to a frequency.
Thus, the circle was complete; by 1907 Wien characterized the
energy in x-rays not in terms of pulse amplitude but in terms of
pulse width alone. He set the kinetic energy of the electron mv2/i
equal to h/r and then derived S from the relation S = ex = ihc/mv2.
The result was S = 6.75 x io~9 cm, considerably closer than his
1905 estimate to Sommerfeld's derived value for the x-ray pulse
width. In support of his use of the quantum relation, Wien cited
the experiments by P. D. Innes showing that the maximum elec-
tron velocity stimulated by x-rays, like that stimulated by periodic
ultraviolet light, does not depend on the intensity of the x-rays but
only on their quality.36 But neither, in the first approximation, did
it seem to depend on the material from which electrons are ejected,
so Wien explained just where the analogy to light breaks down.37
Wien employed the quantum relation to achieve a specific
result; he was not particularly concerned to find a modellmdssig
interpretation for why the relation should be employed. The
quantum transformation relation offered quantitative success
without recourse to the suspect triggering hypothesis. Wien's ex-
tension to x-rays of Einstein's relation for visible light was made
possible by his close identification of periodic light with impulse
x-rays, an identification that followed in turn from Wien's accept-
ance of the electromagnetic view of nature. In December 1907, a
month after the new calculation of S, Wien declared in an impor-
tant review of radiation theory, "there seems to me to be no basis
for any essential distinction between x-rays and light."38
The quantum transformation relation was not synonymous
35
Wien, Gottingen Nachrichten (1907), 599. Laub soon repeated Wien's experi-
ment. AP, 26 (1908), 712-26.
36
Innes, PRS, 79A (1907), 442-62.
37
Wien neglected the energy required to release an electron from the metal after
it has left the atom. He felt that this was negligible relative to x-ray energies but
important for visible light.
38
Wien, Wurzburg 6(1907), 107.
with Einstein's lightquantum hypothesis. Wien accepted the
former but not the latter. One could not assume that the energy of
x-rays exists in "indivisible quanta [in the way that] matter consists
of atoms."39 His analysis had been directed solely to the produc-
tion of x-rays by electron impacts; he did not treat the conceptually
more demanding case of x-ray absorption. Only in the latter case
need one explain how an isotropic pulse can concentrate its total
energy on an individual electron.
JOHANNES STARK AND THE

QUANTUM HYPOTHESIS
Wien's identification of x-rays with light and his recourse to the

quantum relation were both developed and extended by Johannes
Stark. Like Wien, Stark did not distinguish x-ray impulses from
periodic waves. In a review article written in 1905 on electric
discharges, he repeatedly interchanged the terms wavelength and
pulse width, and used wavelength where he explicitly considered
the properties of discontinuous impulses.40 The mere equivalence
of these terms carried with it an implicit connection of x-ray
energy with frequency. It is a short step from this to the attribution
of a frequency to the x-ray, a frequency that measures x-ray energy.
Stark explained, freely mixing incompatible concepts, that among
the various pulse widths in an x-ray beam, "one definite wave-
length will dominate corresponding to the velocity of the stimulat-
ing cathode rays."41
Thus, Stark was already conceptually primed for the quantum
interpretation of x-rays when Wien dropped it in his lap. Wien's
first use of the quantum relation had been for spectral radiation
from positively charged ions, the so-called canal rays. Stark had
discovered in 1905 that the light radiated by these ions shows a
Doppler shift in frequency. It was Wien who had suggested that
canal rays were molecular ions, and his work on the subject was of
continuing interest to Stark.42 Moreover, Stark read widely in the
relevant literature because of his position as editor of the Jahrbuch
der Radioaktivitdt und Elektronik. When Wien applied the quan-
39
Ibid., 106.
40
Stark in Winkelmann, Handbuch, 4 (1905), 642-8.
41
Ibid, 643.
42
Wien, AP, 5 (1901), 421-35, 8 (1902), 244-66; 9 (1902), 660-4.
turn relation to canal ray spectra in May 1907, Stark could hardly
miss its importance. Wien suggested that secondary electrons
produced by cathode rays and canal rays might themselves stimu-
latefluorescence"in much the same manner as in that connection
presumed by Einstein [to hold] between the elementary quanta of
energy and the velocities of photoelectrons."43 The quantum rela-
tion, and explicit reference to Einstein's lightquantum, passed
from Wien to Stark.
In the earlier study, in which Wien calculated the pulse width of
x-rays, he pointed to two cases in which the quantum relation
might be profitably applied: the photoeffect and canal rays. Inde-
pendently of Wien, Stark recognized a third case: the ionization
produced by x-rays. In October, Stark published his first quantum
paper, a dilettante's attempt to connect a hypothetical unit of
positive charge and the electron with a synthesis of quantum and
relativistic concepts.44 Among the chaff, he inserted an irrelevant
but inspired footnote outlining a way to determine the minimum
"wavelength" of x-rays produced by electrons decelerating from a
known velocity. Electrons accelerated by potential P have energy
Pe; if they stop in a distance taken to be one-half the "wavelength"
of the spherical pulse that results, then Amin = ihc/Pe. For a typical
potential of 60,000 V, Amin is 6 X io~9 cm. As we have seen, Wien
reported the same result the following month and compared it
favorably to the value 5 X io~9 cm obtained in Haga and Wind's
second detailed study of the x-ray diffraction photograph.45
But unlike Wien, Stark went on to treat the inverse case of
ionization by x-rays. In extending the quantum treatment to the
absorption of radiation, he opened the door that led to the light-
quantum itself. The transfer of an electron's kinetic energy in
quanta to a spherical wave or impulse may have no classical
rationale, but it is conceptually understandable. How the full
quantum of energy contained in that spherical wave can be passed
back to an individual electron defies understanding; it makes
credible the spatial localization of the radiant energy proposed by
Einstein. The maximum velocity of the electrons stimulated by
light or x-rays would, in the quantum hypothesis, be independent
43
Wien, AP, 23 (1907), 433.
44
Stark, PZ, 8 (1907), 881-4. Stark persistently used X and referred to the
"wavelength" of x-ray impulses.
45
Haga and Wind, AP, JO (1903), 305-12.
of the radiation intensity and, Stark claimed, independent of the
chemical nature of the metal as well. Of course, Einstein had
already pointed this out for ultraviolet light; in a subsequent paper,
Stark gave Einstein credit for a "similar treatment" one that he
said he had read only after the fact.46 This may be true; all that
Stark really needed had been contained in Wien's paper. Although
Wien had cited Einstein's lightquantum paper, Stark's first plunge
into the quantum theory was quite possibly without direct knowl-
edge of that seldom cited source. Stark did not take into account
the possibility that energy might be lost in the removal of an
electron from the structure of the metal. This was likely a holdover
from his initiation to the subject by Wien, who made a similar
assumption. The work function is negligible in the case of x-ray
ionization, but not in the photoelectric effect.
Stark was certainly acquainted with Einstein. The two were
corresponding over a review article on relativity for Stark's jour-
nal. Einstein, perhaps protective of his ideas because of his lack of
academic position, or perhaps simply encouraged to hear a re-
spected physicist employ the quantum condition, shipped off
reprints of all his papers to Stark and wrote to express his happiness
that Stark had "become aware of the lightquantum issue."47
Stark eagerly exploited the quantum relation. Hisfirsttopic after
x-rays was his own discovery, the Doppler shift in canal-ray spec-
tra.48 The problem was to explain why a discrete line appeared at a
slightly higher frequency for each original, rather than a continu-
ous broadening of each line toward higher frequency. Presumably
a continuous range of ion velocities exists in the discharge tube
from zero up to some maximum. The lack of a corresponding
continuum of shifted frequencies cast doubt on the Doppler na-
ture of the effect. The quantum principle offered a possible expla-
nation for the exclusion of lines below a certain threshold. Al-
though later shown to be incorrect, Stark's quantum treatment
sparked wider interest in a quantum discussion of topics other
than cavity radiation and specific heats.49 Next, Stark used the
46
Stark, PZ, 8 (1907), 91411; 9 (1908), 767-73.
47
Einstein to Stark, 7 December 1908. Quoted in A. Hermann, Sudhoffs Ar-
chiv, 50(1966), 272.
48
Stark, PZ, 9(1908), 767-73.
49
Einstein to Stark, 2 December 1908. A. Hermann, Sudhoffs Archiv, 50
(1966), 276; Fruhgeschichte (1969), 88.
quantum hypothesis to calculate the high-frequency absorption
limit for the stimulation of molecular spectra.50 He soon went on
with a student to analyze the relation between photoelectric sensi-
tivity andfluorescencein some fifty-seven organic molecules.51
In the course of these studies, Stark returned repeatedly to the
close connection he perceived between ionization, the photoelec-
tric effect, andfluorescence.Even before recourse to the quantum
in 1907, he had expressed his confidence in the interconnection.52
With the success of the quantum principle demonstrated, his
confidence increased. "Absorption of light," he said, "has as its
consequence the ionization of the absorbing atom or molecule,
that is, the excitation of photoelectric cathode rays, accompanied
by afluorescenceconforming to Stokes' law."53 Stokes' law, which
stated that the frequency of incident light must equal or exceed
that of emitted fluorescent light, lacked a classical explanation. In
the spring of 1908, Stark expressed his appreciation of Einstein's
success in "illuminating another dark point offluorescencephe-
nomena" by giving the straightforward quantum theoretical basis
for Stokes' law. This prompted his request of Einstein in 1908 for
an elaboration of the subject, suitable for chemists' understanding,
to appear in the Jahrbuch.54
A review article by Charles Barkla, which spelled out the de-
tailed properties of characteristic x-rays for the first time in the
German language, appeared early in 1908 in Stark's Jahrbuch.55
Stark, as editor, had known of its content for some time. In that
paper, Barkla revealed that the characteristic secondary x-rays
possess precisely the trait that Stark had emphasized in his long
study of optical absorption: "properties of the chemical elements
that are periodic functions of atomic weight."56 Barkla emphasized
that "without exception" the stimulation of a homogeneous radia-
tion requires a primary beam of greater penetrability than that
produced. Further, the intensity of the primary beam could be
50
Stark, PZ, 9(1908), 85-94, 356-8.
51
Stark and Steubing, PZ, 9 (1908), 481-95; 661-9.
52
Stark, PZ, 5 (1907), 8 in.
53
Stark, PZ, 9(1908), 92.
54
Stark to Einstein, 11 February 1908. Hermann, Sudhoffs Archiv, 50 (1966),
272-3.
55
Barkla, JRE, 5 (1908), 264-324.
56
Stark, JRE, 5 (1908), 124-53.
altered almost at will without affecting the hardness of the sec-
ondary.57
Stark had already converted in his own mind x-ray "hardness"
or "penetrability" to frequency. He recognized immediately that
Barkla had found the x-ray analog to optical fluorescence. The
hardness threshold for the stimulation of characteristic x-rays was
a simple case of Stokes' law offluorescence.Stark was very early in
this realization; it was to be three years before Barkla would speak
of fluorescent x-rays as "an extension of Stokes' Law." Stark was
quick to ensure that his quantum view of characteristic x-rays
appeared in English in Nature.58
Stark was unusual among German physicists in his early recog-
nition of the problems surrounding the paradox of quantity. Be-
fore adopting the quantum relation, he puzzled over the "small
fraction of the total number of molecules" that are placed in a
"chemically different state" after absorbing light energy.59 Follow-
ing contemporary work by Lenard on fluorescence, Stark came to
the conclusion in 1908 that the effect of light is "localized in
definite atomic groups."60 Then Barkla raised the analogous issue
for x-rays. The primary impulse "passes over every single elec-
tron" in all atoms of the scattering substance, but "only a small
fraction of the total number of electron [radiation] sources pro-
duce the homogeneous [secondary] radiation."61 Not only did
Barkla provide the evidence for Stark's extension of fluorescence
to x-rays, he pointed out the single most difficult observation to be
explained in a theory of spreading pulses, whether treated with the
quantum or not.
SPATIALLY LOCALIZED X-RAYS

Stark was the first influential physicist after Einstein to take the
lightquantum seriously. He was led to this step from the analogy of
x-rays to light. Because he therefore was able to assign a wave-
length to the x-rays, he could apply the quantum relation. This
57
Barkla, JRE, 5 (1908), 229, 306, 308, 309, 317, 318, 319.
58
Stark, Nature, 77(1908), 320.
59
Stark, PZ, 5(1907), 8211.
60
Stark, JRE, 5(1908), 149.
61
Barkla, JRE, 5 (1908), 320, 323.
worked so well to explain observed frequency limits and secondary
electron velocities that little doubt remained in his mind about the
validity of the relation. Although Stark adopted the quantum
relation, he tended more than most German physicists to interpret
the elementary processes in terms of conceptual mechanisms.
Unlike most of his German colleagues, Stark took very seriously
the observation that x-ray ionization affects only a very small
fraction of atoms. Sometime before 1909, this forced him to adopt
the lightquantum as a physical reality.
Stark may have adopted the lightquantum by the time of his talk
to the Naturforscherversammlung held in Cologne in 1908. There
was a marked shift in his terminology. What had formerly been the
Planck'sche Elementargesetz, or Planck's law, became the Licht-
quantenhypothese. His talk at the meeting was entitled "New
observations on canal rays with respect to the lightquantum hy-
pothesis," and he considered the significant new case of spectral
lines radiated after ion collisions.62 He had undertaken the experi-
ments, he claimed, "as a proof of the lightquantum hypothesis,"
and expressed the hope that so suspect a motive might not dis-
credit his results. "This hypothesis is so unusual and so strongly
opposes the customary concept of emission and absorption of light
that the general reservations against it are perfectly understand-
able," he said.63 All he actually used in the study was the quantum
transformation relation, which by that time had been adopted by
Planck, Wien, Einstein, and others. That Stark so clearly recog-
nized the controversy into which he headed suggests strongly that
he had already accepted the spatially localized lightquantum.
If Stark had previously been ignorant of Einstein's work, he was
no longer. He cited all three relevant papers, doubtless assisted by
the reprints Einstein had sent him. He spoke freely about the
transfer of radiant energy in units of lightquanta, although such a
view clearly did not require accepting the lightquantum per se.64 In
an Anhang added to the printed version of his talk, he reviewed the
successes of the hypothesis, significantly crediting it solely to
Einstein. His playing down of Planck's contribution may have
62
Stark, PZ, 9(1908), 767-73.
63
Ibid., 768.
64
As Einstein is reputed to have put it: "Although beer is sold only in pint bottles,
it does not follow that it exists only in indivisible pint portions." Frank,
Einstein (1953), 71.
been due to his acceptance of Einstein's 1906 claim that Planck's
work was a consequence of the more basic lightquantum assump-
tion. There was at that time general misunderstanding ofjust what
Planck had done in 1900.65 Einstein, who had hoped to attend the
meeting in Cologne at which Stark spoke, decided instead to spend
his short vacation elsewhere. But he was naturally exceedingly
interested in Stark's remarks because he saw how close Stark had
come to accepting the "revolutionary" lightquantum.66
I suggest that Stark had already done so. The terminology of his
Cologne talk strongly suggests it; but radiation from canal rays -
light produced by mechanical impact - did not require spatially
localized rays in the same way as did absorption of light. Stark had
only to express his results in a manner that would allow him to
refer back to this work as consistent with lightquanta. Stark had
other reasons to pause before taking so bold a step. He knew how
controversial the topic was. And he was at that moment embark-
ing on a bit of political maneuvering to obtain a regular appoint-
ment at Aachen. For this he was counting on, and getting, strong
support from the former resident physicist, Arnold Sommerfeld.67
Sommerfeld was the foremost German proponent of the im-
pulse hypothesis for x-rays, precisely the object of Stark's quantum
assault. Immediately after he received the post, with Sommerfeld's
endorsement, Stark offered Einstein a position as his assistant.
Einstein wisely declined.68 The next month, Einstein published a
masterful review of the state of radiation theory and began to
formulate a statistical treatment of radiation based on large num-
bers of lightquanta.69 Stark wrote that the new view interested him
greatly.70 Four months later, Stark proclaimed publicly that he had
adopted the spatially localized lightquantum for x-rays.71
His paper was entitled "On x-rays and the atomic constitution of
radiation." The critical transition to the lightquantum was de-
65
Klein, AHES, 1 (1962), 459-79. Kuhn, Black-body theory (1978).
66
Einstein to Stark, 2 December 1908. A. Hermann, Sudhoffs Archiv, 50
(1966), 276.
67
Stark to Sommerfeld, 12 October 1908. A. Hermann, Centaurus, 12 (1968),
41-2.
68
Einstein to Stark, 6 April 1909. A. Hermann, Sudhoffs Archiv, 50 (1966), 278.
69
Einstein, PZ, 10 (1909), 185-93. See also the discussion between Einstein and
Ritz, ibid., 224-5, 323~4-
70
Stark to Einstein, 8 April 1909. A. Hermann, Sudhoffs Archiv, 50 (1966), 278.
71
Stark, PZ, 10(1909), 579-86.
signed to win converts directly from the impulse theory. First, he
calculated the mean time between electron impacts in the x-ray
tube; typically operating at 30,000 V and passing 3 X io~4 amps,
the tube produced 1014 collisions every second. On the other hand,
the accepted pulse width, io~8 cm, implied a mean pulse duration
of some io~19 sec. It was extremely unlikely, Stark claimed, that
two or more pulses would combine their effect on the same atom.
The impulses were separated by an interval io5 times longer than
the duration of any pulse. Note that Stark implicitly assumed that
ionization begins instantaneously when the radiation arrives; he
had explicitly claimed this before, but did not do so now.72
Stark sketched a vague mechanism that he thought might ex-
plain interference for visible lightquanta. Optical lightquanta are
radiated from "centers" in atoms that are generally far closer
together than the wavelength. Thus, a "space-time superposi-
tion" of lightquanta arises that has a "complex of specifically
ordered properties." It was not clear what Stark meant by this, and
it was likely that Stark himself did not know. In some unexplained
manner, the superposition of lightquanta would apparently direct
quanta toward regions of constructive interference and away from
regions of destructive interference. For wavelengths that are short
compared to the separation of the centers, on the order of io"8 cm
or less, the complex does not arise and the radiation acts like more
or less directed units.73 Stark concluded that the "oscillating en-
ergy" of each impulse must therefore be contained within a vol-
ume of space equal to (c/v)3. He no longer bothered to qualify his
use of the then standard frequency symbol to characterize x-rays,
while at the same time his argument depended on the rays' spatial
and temporal limitations.74 This article formally opened Stark's
attempt "as quickly as possible to bring the conflict between the
lightquantum hypothesis and its older opposing concepts from the
sphere of speculation and theoretical discussion to the firm ground
of experiment."75
72
Stark had recognized the significance of an immediate response in optical
fluorescence. He thought that his chemical interpretation explained "the fact
that the onset and cessation of fluorescence occurs almost at the instant the
illumination start and stops." Stark and Steubing, PZ, 9 (1908), 491.
73
Stark,/>Z, 9(1909), 582.
74
Ibid., 583. The standard symbol for frequency was n, not v.
75
Ibid.
>30*
Electron beam direction

Figure 5.1. Stark's graph of forward scattering of x-rays. [PZ, 70(1909),
911.]
The weapon he chose was the asymmetry of emitted x-rays. He

thought he had found an experimentum crucis. In November he
applied his substantial gifts as an experimentalist to a careful
demonstration that x-rays are not emitted isotropically from the
anticathode. He found that they show a distinct maximum inten-
sity in the forward direction. The unbroken line in Figure 5.1
records his corrected measures of photographic blackening pro-
duced by x-rays scattered from an aluminum plate at different
angles from the horizontal direction of the cathode beam toward
the right.76 For a typical cathode beam velocity of io9 cm/sec, the
maximum effect lies within a cone some 6o from the forward
direction. For once, Stark's knowledge of the foreign-language
literature failed him. Bragg had emphasized Cooksey's demonstra-
tion of the asymmetry in 1908 and earlier in 1909 had himself
repeated the result for x-rays.77
Stark claimed that the asymmetry of emission could not be
76
Stark, PZ, 70(1909), 902-13.
77
Cooksey, Nature, 77(1908), 509-10. W. H. Bragg and Glasson, TPRRSSA, 32
(1908), 301-10.
Figure 5.2. Stark's vector diagram of x-ray emission. [PZ, 10 (1909),

905.]
explained by what he called the aether wave hypothesis of x-rays.

By this he seems to have meant the view that each electron radiates
isotropically and that the energy density in the impulse decreases
as the square of the distance traveled. On the other hand, he
asserted, lightquanta carry momentum in quantity hv/c. Since the
electron beam is unidirectional, each lightquantum released on
impact should have a net forward component of momentum, as
indicated in the vector diagram, Figure 5.2. Here m, is the mass of
the electron, v, its velocity before impact, and v/ its velocity after
impact; m2 and v2' are the mass and the final velocity of the atom
that is struck; and hv/c is the "electromagnetic momentum" of the
lightquantum.
In addition to the variation in x-ray intensity, Stark claimed that
there should be a decrease in the penetrating power or, as he put it,
the frequency v, of the emitted x-rays as the angle from the forward
direction increases. Thus, he expected that the absorption coeffi-
cient of the rays should increase the farther one looked from the
forward direction. This experimental result he also demonstrated,
as shown in Figure 5.3 where, again, the beam direction is horizon-
tal toward the right. It was soon made clear to Stark that the
classical impulse theory was fully capable of explaining the azi-
muthal intensity variation of x-rays and the shift with the azimuth
in the hardness of x-rays. Arnold Sommerfeld pointed it out.
Sommerfeld felt that when a man as influential as Stark made a
mistake of such magnitude, the correction of the error should be
made public.78
78
Sommerfeld to Stark, 4 December 1909. Hermann, Centaurus, 12 (1968), 45.
90
60
Electron beam
direction
Figure 5.3. Stark's graph of variation of absorption index of x-rays

emitted at various angles to the direction of the incident cathode-ray
beam. [PZ, 10(1909), 912.]
THE SOMMERFELD-STARK EMBROGLIO

Arnold Sommerfeld, since 1907 ordentlicher professor of theoreti-
cal physics at Munich, had a personal stake in x-rays. His treat-
ment of diffraction by impulses in 1900 had convinced many in
Germany that x-rays were discontinuous impulses in an electro-
magnetic continuum. Although he had not written on x-rays since
then, at Munich he was in close contact with Rontgen, whose
students studied their properties. In fact, even before Stark's
claims for the lightquantum, Rontgen had asked Sommerfeld to
consider the asymmetry that his student, Bassler, had observed in
x-ray intensity.79 Moreover, Barkla's characteristic homogeneous
x-ray radiation appeared to corroborate Sommerfeld's early sug-
gestion that a continuum of x-rays exists from the periodic to the
aperiodic. Stark provided the opportunity for Sommerfeld to ex-
press considerations on which he had been reflecting for a "long
time."80
Before addressing Stark, Sommerfeld commented on the new
homogeneous x-rays to distinguish them explicitly from the old.81
The isotropic, unpolarized, homogeneous secondary x-rays, he
said, were likely due to emission of energy by electrons bound
79
Bassler, AP, 28 (1909), 808-84. Sommerfeld, Scientia, 51 (1932), 48. Other
important evidence was provided by Kaye, PCPS, 15 (1909), 269-72.
80
Sommerfeld, PZ, 70(1909), 969.
81
Ibid.
within the atom. He thereafter called these fluorescent rays. The
partially polarized, aperiodic transverse impulses created by elec-
tron impacts he called the Bremsanteil or braking part of the rays.
This division was not original; the British and Stark had made it
before.82 But although Sommerfeld allowed the homogeneous
fluorescent rays a wavelength, he categorically denied it to the
Bremsanteil "For this, as before," he said, "one would more
correctly speak of pulse width rather than wavelength."83 Although
Sommerfeld here first admitted that the Planck relation might be
needed to treat the fluorescent x-rays, he was certain that it had
"nothing to do with the Bremsanteil" The division between the
two types of x-rays seemed so strict to Wien that he quickly wrote
to be assured that Sommerfeld did not conceive of either type as
"nonelectromagnetic" in origin.84
Then Sommerfeld turned to the defense of the electromagnetic
theory of pulses. "Recent speculations on the lightquantum," he
said, "have tarnished our trust in the validity of electromagnetic
theory."85 To set the picture straight, he showed that the directed
field amplitudes AE and AH at a great distance from an electron,
stopped instantaneously from a velocity v, are both proportional to
(v/c) sin 6
(5.8)
1 (v/c) cos 6
Here 6 is the angle between the direction of the acceleration and
the direction of interest. For the more realistic case in which the
deceleration occurs over a finite time interval, he argued that the
field amplitudes become proportional to
Sin
(5.9)
[1 -(v/c) cos 6]3
Energy radiates instantaneously at a rate given by the product of
these vectors; the total energy radiated comes from integration
over the duration of the impulse.86 Sommerfeld found that the
total energy flux varies with 6 as
82
Stark had suggested that there was a distinction between impulse and fluores-
cent x-rays, PZ, 10 (1909), 579-86. He was one of the first physicists to point
up the close parallels between x-rays and light, their identity being a viewpoint
that, he said, was "gaining ground." See also Stark, PZ, 10 (1909), 614-23.
83
Sommerfeld, PZ, 70(1909), 970.
84
Wien to Sommerfeld, 27 December 1909. Sommerfeld papers; AHQP 34, 13.
85
Sommerfeld, PZ, 10 (1909), 976.
86
Poynting, PTRS, 175 (1884), 343-61.
70
Figure 5.4. Sommerfeld's graph of forward scattering expected of classi-

cal x-ray impulses. [PZ, 10 (1909), 972.]
sin2 6 / 1
S(6) = - 1 (5.10)
cos0 \[l - (v/c) cos 6]4
The maximum of this expression occurs at an angle with cosine
equal to v/c. The closer the initial velocity is to c, the greater is the
fraction of energy radiated toward the front. The resulting inten-
sity distribution is shown in Figure 5.4. Roman numerals denote
the initial electron velocities c/10, c/5, and c/3 in increasing order.
The maximum in the intensity at about 6 = 6o for case III agreed
well with Stark's measurement for an initial electron velocity
near c/3.
Sommerfeld commented that the asymmetry he had derived
classically could be visualized in a simple diagram. In Figure 5.5,
point / marks the beginning of the electron's deceleration. Point x
marks the position that the electron would have occupied had it
not slowed at all. The expanding impulse front is centered on /.
The trailing edge of the impulse emanates from point/ where the
deceleration ends. The resulting pulse width varies with the azi-
muthal angle 0, as shown exaggeratedly in the figure. The field
vectors are maximum where the density of the radial force lines is
greatest. This will occur in the plane perpendicular to the instanta-
Figure 5.5. The electric field in the vicinity of a rapidly decelerated

electron, showing the azimuthal variation in impulse width.
neous velocity vector of the electron. But it is not until after the
pulse has passed a point in space that the field, so to speak, knows
that any acceleration has occurred. The maximum energyfluxwill
always be directed through the region of the impulse that intersects
an imaginary plane perpendicular to the acceleration vector and
centered on the moving point x Consequently, the maximum
x-ray intensity will always be in the forward direction, and will
occur at smaller angles 6 the greater the initial electron velocity.
But the net energy flux is compounded of both the quantity of
pulses and the quality of each. Sommerfeld turned to Wien's
analysis of 1905 to point out something "that appears to be impor-
tant in connection with recent observations."87 Figure 5.5 directly
answers Stark's claim that the penetrating power of the x-rays
decreases for increasing azimuth. The impulse width increases
with angle 6 up to 1800. Thomson had shown that the energy
carried by a pulse varies as i/d; but he had not noticed that S will
87
Sommerfeld, PZ, 10 (1909), 975.
vary with the azimuthal angle around the pulse. In pointing this
out explicitly, Wien and Sommerfeld were providing a significant
reformulation of the definition of radiant intensity.88
If an electron oscillates periodically about a fixed point, it
radiates a periodic wave that varies in amplitude with the azimuth,
but does not vary in wavelength or in frequency. An impulse acts
differently. Both its amplitude and its pulse width vary with the
azimuth. The electromagnetic impulse hypothesis offered a con-
ceptual route, unavailable with periodic waves, whereby the prop-
erties associated classically with wave amplitude could be trans-
ferred to pulse duration or frequency. When it became clear that a
quantum theory of x-rays was needed, Sommerfeld's and Wien's
analyses justified assigning a significant role in defining the energy
content of impulse radiation to the impulse's duration rather than
to its amplitude.
Stark had placed himself in so indefensible a position that an
essential point he wished to convey was obscured. In describing the
radiation from a single electron, Stark had said that its energy
density far from the electron "has equal magnitude in all direc-
tions."89 Sommerfeld showed without question that Stark's claim
was simply wrong. Stark had confused the equality of energy in all
directions with the constancy of energy in a given direction. When
Sommerfeld pointed out his error, Stark retreated to a position
somewhat closer to his intent. What he now sought to explain with
the lightquantum was the fact that "the electromagnetic radiation
from a single electron does not spread to a volume that grows as the
square of the separation from the emission center."90 He now said
that what he had called the Atherwellenhypothese was not compat-
ible with Maxwell's electromagnetic equations.91 He seems to have
meant, in accordance with the successes of the quantum hypoth-
esis, that a true aether wave must be able to transfer energy in this
non-Maxwellian way.
Stark's position was no clearer to others. Einstein commented
privately that Stark had "once again produced pure dung."92
88
Wheaton, HSPS, 11 (1981), 367-90.
89
Stark, PZ, 10 (1909), 903.
90
Ibid.
91
Stark to Sommerfeld, 6 December 1909. Hermann, Centaurus, 12 (1968), 46.
92
"Stark hat wieder einmal gediegenen Mist produziert, Sommerfeld wohl die
Beweiskraft jenes Phanomens uberschatzt." Einstein to Laub, 16 March 1910.
From a transcription by Laub in the Einstein papers.
Sommerfeld was unable to fathom what Stark meant and con-
cluded that his caricature of classical electromagnetic theory had
only been a straw man set up especially for the kill. "Your aether-
wave hypothesis," he wrote, "is electromagnetically impossible
and has neither followers nor interest; it makes no sense [for you]
to oppose it by experiment."93 In the ensuing discussion, Stark
deleted all reference to his unfortunate phrase "equal magnitude
in all directions." Instead, he emphasized only that energy will
remain constant in a given direction.94 Sommerfeld knew perfectly
well that Stark had changed his stance, and called the later exposi-
tion an "artfully dissected transcription" of the original version.95
Stark'sfinalpaper on the x-ray asymmetry reported only empirical
findings that were no longer in doubt. No interpretation of the
results, which would have brought him back into the dispute, was
provided.96 In any given direction from the source of an x-ray
impulse, the electromagnetic amplitude decreases as the pulse
spreads over larger volumes of space; the pulse width likewise
changes. But the frequency of a periodic disturbance remains
constant in a given direction and is equal in all directions. Already
used to interpreting pulse duration as an intrinsic frequency for
x-rays, Stark saw its directed constancy as the key to the preserva-
tion of the energy units required by the quantum relation. Som-
merfeld's impulse did not fulfill that condition. When Stark tried
to point this out after the fact, it appeared only that he was
switching to a new argument to save face. That may have been the
case, but Stark had also raised an important question. Not only did
the electromagnetic theory have to show that energy is localized, it
also had to demonstrate how that energy stays localized as the
"wave" propagates.
In the first volume of a long exposition on the Principles of
atomic dynamics, published in the midst of the controversy with
Sommerfeld, Stark stated his general concern. He opposed what he
called the uncritical application of traditional continuum me-
chanics to the new field of elementary processes. "All objections
raised to date against the lightquantum hypothesis are based on
93
Sommerfeld to Stark, io December 1910. A. Hermann, Centaurus, 72 (1968),
48-9.
94
Stark, PZ, 11 (1910X30.
95
Sommerfeld, PZ, 11 (1910), 101.
96
Stark, PZ, 11 (1910), 107-12.
incorrect or unprovable assumptions," he said.97 He began to do
experiments on visible and ultraviolet light.98 Even stronger claims
for the lightquantum appeared in the second volume of Stark's
Atomic dynamics, entitled Elementary radiation, which appeared
in 1911. There Stark reformulated the lightquantum in terms more
suited to the temperament of German physicists. It should be
thought of as a "deformation in the energetic medium" that
remains localized.99 Stark could afford to come somewhat closer to
Sommerfeld's position because by 1911 it was clear that the energy
of x-rays is distributed far less isotropically than had been thought.
From this point on, the issue was no longer whether x-rays diverge
spherically from a source, but whether they diverge at all. And it
was far from clear that any electrodynamic theory, with or without
the new definition of frequency and intensity, could achieve the
degree of localization that ionization experiments implied the
energy of x-rays possessed.
97
Stark, Electrischen Quanten (1910), vii.
98
Stark, PZ, 11 (1910), 179-87.
99
Stark, Elementare Strahlung (1911), 283-4.
PART III
Seeking an electrodynamic solution
1907-1912
Localized energy in spreading impulses
On the basis of . . . pure undulatory theory, one is

led . . . to consequences that approximate the
Newtonian emission theory.1
The discussions between Bragg and Barkla in Britain, and between

Stark and Sommerfeld in Germany, encouraged attempts to de-
termine how much spatial concentration of radiant energy was
demanded by ionization, and how much was compatible with
classical electromagnetic theory. Mirroring the different percep-
tions of and approach to the problems of x-ray absorption in
Britain and in Germany, there were two parallel responses. J. J.
Thomson recognized that the paradoxes applied equally to all
forms of radiation. He influenced important experimental tests of
energy localization, and concluded that the classical ideal of sym-
metry of the electric field around an electron would have to be
abandoned. Arnold Sommerfeld took no refuge in the alteration of
the physical interpretation of Maxwell's theory. He pressed formal
electrodynamics to its extreme to show that the energy of y-rays
might be localized to an extent compatible with recent experi-
ments.
But neither Thomson's nor Sommerfeld's treatment addressed
the most compelling problem raised by Bragg and Stark. The
energy transferred from a wave to an electron does not seem to
change when the distance from the source changes. No matter how
severely limited in angular extent the energy of x-rays and y-rays
might be according to Thomson or Sommerfeld, that energy is still
diluted in strength as the rays traverse space. Neither of them
therefore confronted the issues in quite the same way as did Bragg,
who had never intended his neutral-pair hypothesis to apply to
ordinary light. As evidence increased in 1910 and 1911 that x-rays
1
Sommerfeld, Mtinchen Sb, 41 (1911), 4.
136 Seeking an electrodynamic solution, 1907-1912
share fundamental properties of light, Bragg began to call for a
truly dualistic interpretation of radiation, one in which wave and
particulate properties were to be combined in a radically nonclas-
sical fashion. To some, predominantly the British physicists, the
paradoxes of radiation had reached crisis proportion by 1911.
THE VICISSITUDES OF THE TRIGGERING

HYPOTHESIS
William Bragg had not been the only British physicist to express
reservations about the impulse hypothesis of x-rays in 1907. J. J.
Thomson, long uneasy about the paradox of quantity, rejected the
triggering hypothesis as an explanation of x-ray ionization in 1907,
as we shall see. For Thomson, the paradox of quality was held at
bay no longer. He instituted a systematic search for evidence on
the seemingly contradictory properties of all forms of radiation.
Unlike Bragg, who consistently excluded light from his concerns,
Thomson realized as early as 1907 that the problems involved any
spreading wave, periodic or not. It had become clear to him that
even for the photoelectric effect with visible light, the triggering
hypothesis was open to serious question.
In September 1907, Erich Ladenburg reported in Berlin that
ultraviolet light releases electrons with a continuum of velocities
by photoelectric action.2 He found that the maximum electron
velocity increased monotonically with the frequency of the light;
there seemed to be no resonance peaks. Although Ladenburg
interpreted his results as providing support for Lenard's triggering
hypothesis, Thomson saw it differently. If every incident fre-
quency of light can induce an electron to leave an atom with a
corresponding velocity, the triggering hypothesis demands that an
unreasonably large number of different electron motions must be
contained in an atom. Recall that in this view each electron
possesses a velocity in the atom that is uniquely related to its
rotational or vibrational frequency. If all frequencies of incident
light release electrons, it would seem that the photocathode atoms
contain a continuum of electron oscillation frequencies.
Thomson was wary. Was it possible that different atoms of the
same element have electrons of quite different frequency; that only
2
E. Ladenburg, VDpG, 9(1907), 504-14.
Localized energy in spreading impulses 137
the large number of atoms provided the apparent continuum
effect? This proposal was disagreeable to anyone familiar with
chemistry and conflicted with spectroscopic evidence for atomic
homogeneity. Could it possibly be that within a single atom so
large a number of vibrational or rotational frequencies are repre-
sented? To account for this, extremely large numbers of electrons
would have to reside in each atom.
Just the year before, Thomson had challenged the prevailing
notion that atoms contain large numbers of electrons.3 Based on
results of scattering experiments, Rutherford and Bragg had sup-
posed that most of the atomic weight of any atom consists of
electrons; a single atom of hydrogen would possess a swarm of
1,000 electrons. But Thomson, combining evidence from ^-parti-
cle absorption, x-ray scattering, and optical dispersion, revised the
estimate downward by a factor of io3. This left about as many
electrons per atom as the atomic weight in units of hydrogen
atoms.4 It was difficult to see how so few electrons could execute
mechanical oscillations of enough different frequencies to explain
Ladenburg's findings. The paradox of quality, eclipsed by the
triggering hypothesis since 1902, began to exert its influence on
Thomson.
The year before, Thomson had suggested to P. D. Innes that
some further tests be made on x-rays by analogy to the photoeffect.
Innes had found that the velocities of secondary electrons due to
x-rays are independent of the intensity of the x-rays and depend
only on their quality.5 The similarity to photoeffect behavior had
convinced Innes that the triggering mechanism should be ex-
tended to x-rays. And Thomson, like Wien before him, temporar-
ily adopted the triggering hypothesis for x-ray ionization. Now
Thomson realized that, whatever the theoretical difficulties for a
triggering explanation of the photoeffect, the same considerations
would apply to ionization by x-rays. Within a month of Laden-
burg's announcement, and just as Bragg's paper on neutral-pair
3
J. J. Thomson, PM, 11 (1906), 769-81.
4
Heilbron, AHES, 4 (1968), 269-74.
5
Innes used a variable spark gap to measure the x-ray tube potential and,
consequently, the x-ray quality. He found that changing the gap from 5.5 to
16 cm produced a 7 percent increase in the maximum velocity of released
electrons, PRS, 79A (1907), 442-62.
x-rays appeared in England, Thomson returned to the problem of
radiation.6
The problem was especially acute for x-rays. Whereas the
maximum velocity of electrons released by ultraviolet light is on
the order of io8 cm/sec, Innes had corroborated Dorn's figure of
several times io9 cm/sec for the velocity of x-ray electrons. The
energy differs by a factor of about io3; Thomson believed it to be
io4. Velocities of electrons released by y-rays are even greater. Even
if the atom encompassed the profusion of electron motions re-
quired by the triggering explanation of the photoelectric effect, an
entirely new set of motions, with energy io3 times greater, would
be required by the x-ray data. The y-rays needed an even more
energetic set. It was clear to Thomson that, at least for x-rays and
y-rays, the energy could not come from the atom but must be
transformed from the radiation itself. The double paradox of
quantity and quality convinced him in 1907 that the energy of the
new radiations comes in units, the "bright specks on a dark
background" that he had first discussed in 1903.7
Thomson had already calculated the size of the x-ray unit. In
1906 he had used the absurdity of the result to support the trigger-
ing hypothesis.8 The transverse electricfieldamplitude AE acts on a
charge e to induce a velocity measured to be at least io9 cm/sec.
The net change in electron velocity must be (AE e/m)dt where dt is
the time the pulse takes to pass the electron. The known value of
e/m implies that AEdt is approximately 60 g cm/emu/sec. But the
energy density in the expanding impulse at the position of the
ionized electron is
AE2S/4nc2 = AE2 dt/Anc (6.1)
Substituting 60 for AEdt, and putting an upper limit of io~5 cm for
<5, Thomson found that almost 1 cal/cm2 has to be provided by the
x-ray impulse.9 The results of his earliest analysis showed that even
in the direction of maximum radiant intensity, 6 = 900, a large
6
J. J. Thomson, PCPS, 14 (1907), 417-24. This answers McCormmach's query
on why Thomson should have revived his 1903 speculations on light at this
time. McCormmach, BJHS, 3 (1967), 372, 374.
7
J. J. Thomson, Electricity and matter (1904).
8
J. J. Thomson, PCPS, 13 (1906), 321. It was, he said, "an altogether inadmissi-
ble amount" of energy. The following year, he interpreted the result to indicate
the extreme coarseness of the distribution of energy "units" in x-rays.
9
J. J. Thomson, Conduction of electricity, 2nd ed. (1906), 320-1.
20
enough energy density occurs only within io~ cm of the origin of
the x-ray pulse.10 This is inside the decelerating electron itself!
In 1907 Thomson turned the argument around. If the energy
was bound up in units, he said, the spatial density of those units
needed only to be correspondingly low. For the faint but visible
light intensity of io"4 erg/cm2/sec, there would be only 1 unit/liter
of space - "exceedingly coarse," Thomson said.11 "As these units
possess momentum as well as energy they will have all the proper-
ties of material particles," he continued. "Thus we can readily
understand why many of the properties of the y-rays resemble
those of uncharged particles moving with high velocities."12
The implicit reference to Bragg's neutral-pair hypothesis was
clear enough. Thomson tested Bragg's hypotheses: Were the slow
electrons emitted by different causes of ionization actually of
uniform velocity? It appeared that they were. Thomson found that
the electrons released by cathode rays had about the same velocity
as did those elicited by canal rays. The velocity was independent of
the type of irradiated gas and of its pressure.13 Early in 1908,
Thomson concluded that this common effect could be explained if
"a doublet consisting of a corpuscle and a positively charged
particle" was removed from the atom.14 He soon agreed, at the
height of the Bragg-Barkla controversy, that Bragg's hypothesis
had much to offer, not just for x-rays but for y-rays as well.15
But Thomson did not believe that x-rays are particles. He sought
a solution within quasi-classical electromechanics. The coarse-
grained aether, he thought, would rescue the advantages of a wave
model of radiation and still explain the dual paradoxes. The energy
transmitted in x-ray and y-ray impulses is, in his interpretation,
concentrated in specific directions, so that few atoms are hit, and
those receive close to the full complement of radiant energy. He
expected to be able to explain the similar behavior of periodic light
waves in much the same way.
10
J. J. Thomson, PM, 46 (1898), 528-45.
11
J. J. Thomson, PCPS, 14 (1907), 423-
12
Ibid., 424.
13
Thomson, PCPS, 14 (1908), 541-5. Fuchtbauer, PZ, 7 (1906), 748-50.
14
Ibid., 545.
15
J. J. Thomson, PCPS, 14 (1908), 540. Just what the effect of this was on Barkla,
who saw himself as the defender of Thomson's views, is not clear. But the
message was limited to the small readership of the PCPS. Not until 1910 did
Thomson raise these issues in the widely distributed PM.
THOMSON'S RESTRUCTURED AETHER
In 19 io, after two years of silence on the nature of light, Thomson
announced that prior concepts about the microscopic structure of
the aether must be in error.16 According to his "coarse-grained
aether" hypothesis, radiant energy does not propagate spherically
from its source because the structure of electric field lines sur-
rounding each electron is microscopically discontinuous. Now he
went even further, suggesting that each electron is the seat of only a
single "tube" of electric force; its energy is radiated only into one
narrow cone. Macroscopic charged bodies appear to exert electric
force isotropically only because of the large number of electrons
they contain. Each of these electrons broadcasts its influence in
only one specified direction in space.
The impulse produced by a rapid displacement of the electron is
represented as a transverse kink in the electric field, as in Figure
6.1. The kink travels outward along the cone of solid angle Cl
defined by the tube. The transverse electric force is approximately
where v is the total velocity lost by the radiating electron and dt is

the time interval over which the deceleration occurs. The total
energy carried by a kink is therefore
(6 3)
Wi -
As the kink moves outward, the transverse displacement as well as
the angular width of the cone increase in proportion to the distance
r. Thus, the total energy of the kink remains constant. Thomson
concluded that "the resulting Rontgen radiation is concentrated
into small patches which possess momentum and energy . . . we
have a condition of things much more closely represented in many
respects by the old emission theory of light than by the wave
theory."17
One impediment to the extension of this idea to visible light was
the potential difficulty of explaining interference phenomena. In
16
J. J. Thomson, PM, 19 (1910), 301-13. Hints of Thomson's radical proposal
had been given in the BAAS Report (1909), 3-29; Nature, 81 (1909), 248-57.
17
J. J. Thomson, PM, 19 (1910), 308.
Figure 6.1. Thomson's concept of x-ray impulses as kinks in an electro-

magnetic tube of force.
1909 Thomson had encouraged his student, Geoffrey Taylor, to

seek interference effects in light of extremely low intensity. Per-
haps one might find evidence of the "grainy" structure of light if
the number of radiating sources was sufficiently low. In January
1909, Taylor reported the results of, among others, a test that
spanned 2,000 hr of illumination by a sub visible source.18 He
found clear evidence of ordinary interference fringes.
Thomson was reluctant in 1910 to extend his revision of the
electromagnetic field beyond x-rays and y-rays. He still held out
hope that the photoeffect would prove tractable for the triggering
hypothesis. This required an atom that would support a contin-
uum of electron frequencies. In 1910 Thomson offered such an
atomic model to the readers of the Philosophical Magazine.19
18
Taylor, PCPS, 15 (1909), 114-15.
19
J. J. Thomson, PM, 20(1910), 238-47.
Electrons follow orbits in the effective inverse-cube force sur-
rounding a dipole such as one of Bragg's neutral pairs. Thomson
showed that electrons describing orbits on a cone extending from
the negative end of the dipole possess kinetic energy in proportion
to their orbital frequency. Because an orbit may occur at any
distance from the dipole, a full continuum of frequencies awaits a
passing light wave. Electrons released by resonance fly off with an
energy proportional to the triggering frequency. Thomson even
showed that the constant of proportionality was tolerably close in
value to Planck's h.
Thomson was not receptive to the quantum hypothesis and set
no great store in a quantum of light. He compared his photoeffect
constant to Planck's not because it was required by the empirical
data then available but in an attempt to show how one might
explain so-called quantum effects on the basis of classical concepts.
His continued attempts to develop the idea after 1911 must be read
as opposition to the growing acceptance of the quantum theory,
which, he thought, achieved mere agreement with experimental
results without providing a conceptual explanation of the para-
doxes he perceived in the absorption of radiation.
STATISTICAL FLUCTUATIONS! CAMPBELL'S

TEST OF LIGHT
Thomson's concerns about light influenced a Cambridge student
to undertake a remarkable and significant test of its nature. Unlike
Thomson and virtually all of his compatriots, Norman Campbell
was convinced of the value of Einstein's lightquantum. His early
adoption of this unorthodox view was the result of two influences:
Thomson and Einstein. In 1907, as Thomson made the first tenta-
tive proposal that the unit hypothesis might apply even to ordinary
light, Campbell showed his detailed understanding of the para-
doxes that had led to that state of affairs. His book, Modern
electrical theory, published that year, reviewed the peculiar ioniz-
ing behavior of radiation with depth and originality.20 By 1909,
under the sway of Einstein's relativistic abandonment of the
aether, Campbell suggested that the "confusions" about light arose
from continued reliance on the outmoded concept of an aether.
The aether, he said, should be relegated "to the dust heap where
20
Campbell, Theory (1907), 221-9.
21
now 'phlogiston' and 'caloric' are mouldering." The new
"atomic" theories of radiation, free of aetherial taint, should be
adopted. "The trend of modern theory," he explained, "is every-
where to replace by discontinuity the continuity which was the
basis of the science of the last century."22
Casting about for a suitable formalism to treat discontinuous
events, Campbell selected the statistical methods used to describe
radioactive decay of individual atoms. Egon von Schweidler, a
Viennese expert in the field, had realized in 1905 that a statistical
interpretation of atomic decay offered a means to determine the
absolute number of decaying atoms. One needed only to measure
the size of the fluctuations from the ideal exponential law of decay.
At the First International Conference on Radiology and loniza-
tion held in Liege, Schweidler considered how the resultant ion-
ization current changes as Z, the number of undecayed atoms,
decreases.23 The fluctuation e in the current increases as Z gets
smaller. But what, precisely, is the form of this deviation from
an exponentially declining current?
The number, Z, of Zo initial atoms still undecayed after an
elapsed time interval t is given in the ideal case by
Z=Z o exp(-C0 (6.4)
where is a constant characteristic of the element used. This
relation is shown by the central dashed line in Figure 6.2. The
measured ionization current (mean value proportional to Z) will
fluctuate about the ideal case, as shown schematically by the
irregular line. But the limits of the fluctuation, shown by dotted
lines, should increase as the number of undecayed atoms falls.
Simply applying Bernoulli's calculation of expected error,
Schweidler found the mean value of the limit e expressed as a
fraction of Zo:
l-[l-exp(-C0H1/2
6 = (6.5)
[l-exp(-C0]Z o 1
If/ is taken to be small enough so that exp( CO 1 > a n adequate
approximation is:
e = l/yfz (6.6)
21
Campbell, PM, 19 (1910), 181-91.
22
Campbell, PCPS, 15 (1909), 117.
23
von Schweidler, Premier congres, Liege (1906), paginated section 3, pp. 1-3.
The time rate of change of Z is equal to - Z because d/dx exp(ax) = a
exp(ax).
time
Figure 6.2. Schematic representation of the limit of fluctuations of

actual radioactive decay about the ideal law of decay.
Schweidler thought that a measure of , from the ionization

current produced by the decaying element, would directly yield Z.
But the application was not as easy as it appeared. In 1905, errors in
methods used to obtain relative values of Z were larger than the
anticipated fluctuations. The method was applied in 1906 by
K. W. F. Kohlrausch to determine the number of a-particles
ejected from polonium;24 an experiment repeated the following
year by Edgar Meyer and Erich Regener in Berlin.25 Kohlrausch
had found the fluctuations to be larger than anticipated, and
Meyer located the difficulty in the slow response of the electro-
meter used to measure the ionization current. Hans Geiger at
Rutherford's laboratory in Manchester criticized Kohlrausch for
failing to ensure that his currents were at saturation, and went on
to develop a new means of counting individual a-particles. This
bypassed the problem by allowing individual a-particles to induce
electrical discharges large enough to be detected.26
Thomas H. Laby at Cambridge developed a sensitive quartz
fiber electrometer that he hoped would allow measurement of
24
Kohlrausch, Wien Sb, 115:11a (1906), 673-82.
25
E. Meyer and Regener, VDpG, 10 (1908), 1-13; AP, 25 (1908), 757-74.
26
Geiger, PM, 15 (1908), 539-47. Rutherford and Geiger, PRS, 81A (1908),
141-61.
rapid transient effects. He apparently tried to test the device in
1908 on the ionization produced in a gas by the y-rays from
radium.27 Conversations with Norman Campbell may have pro-
vided him with the technique he was seeking. Shortly thereafter,
Campbell began to apply Schweidler's method to the paradoxical
problems of light then under scrutiny by J. J. Thomson. First,
Campbell reformulated Schweidler's derivation using finite differ-
ences and combinatorial analysis, as befit its application to discon-
tinuous processes.28 But the real intent of his work became clear in
a two-part study of the photoelectric effect in 1909.29
Campbell saw in the fluctuation method a means to test the
dependence or independence of the light radiated in different
directions from a single point source. He split the light from a
Nernst lamp into two beams, each incident upon one of a matched
pair of sensitive photocells. He then balanced the two output
signals so that only the difference between them was observed. If
the light spreads spherically, one would expect that the "elemen-
tary" ionization events in the two cells would coincide and pro-
duce few differences. However, if the radiation in each beam is
directed, and there is no correlation of microscopic effect, then a
measurable fluctuation in the null current would arise.
The experiment failed because Campbell was unable to find a
light source simultaneously intense and steady enough. He care-
fully calibrated the quartz fiber electrometer and removed the
oxidized impurities from the photocells. He also developed a new
form of high resistance across which to read the electrometer
current, one that was itself free of fluctuations.30 But he required
high-intensity light. The only suitable source was a 50-watt Nernst
lamp, and, because it was run off the Cambridge ac mains, its
output varied by 5 percent, considerably greater than the fluctua-
27
Laby, PCPS, 15 (1909), 106-13. See also Laby and Burbidge, Nature, 87
(1911), 144; Campbell, PCPS, 15 (1909), 136.
28
Campbell, PCPS, 15 (1909), 117-36. By "atomic" theories of radiation,
Campbell had in mind "at least two:" Thomson's unit hypothesis and
"Planck's theory (as interpreted by Stark)," PM, 19 (1910), 190.
29
Campbell, PCPS, 15 (1909), 310-28, 513-25.
30
Others, including Meyer and Regener, typically used a "Bronson" resistance -
a gas cell in which low conductivity is induced by a source of ionizing
radioactive material. Campbell understandably criticized Meyer for introduc-
ing an additional source of fluctuation and replaced the Bronson resistance by
a liquid measure of xylol and ethanol, calibrated at 9 x io10 1.
tions of interest.31 A few months later he reported failure, for the
same reason, in a test on x-rays.32
Although Campbell reached no experimental conclusion, he
considerably advanced the analysis of the problem. The mean
square fluctuation in the number of electrons registered from a
single photocell is
~ (~ + ~tf)dt, (6.7)
where TV is the mean number of "light disturbances" (spherical
impulses or localized units) that appear in a unit of time.33 The
period of illumination is dt, and co is the number of electrons
released by each disturbance; today we would call co the quantum
efficiency. But co can vary from one ionization event to another,
and t] measures the instantaneous deviation of co from its mean
value characteristic of the cathode atoms, the light frequency, and
light intensity. Campbell fixed attention on the change in rf when
the illuminated area of the photocathode becomes larger and the
photocurrent consequently increases. The larger area serves to
integrate irregularities in electron receptiveness to the light over a
larger sample of atoms. According to the classical wave theory of
light, waves spread to encompass all electrons in the illuminated
surface, so that fluctuations in co should approach zero when larger
areas are illuminated. For the rival theory, this is manifestly not
the case. A directed light uniâffects electrons only along a specific
direction from the source; rf should remain roughly constant as
the cathode area increases.
If one increases the intensity of the light while holding the
illuminated area of the photocathode fixed, the current also
changes. Under all circumstances the current is proportional to
Nco, the total number of electrons released per unit time. For
31
The Nernst lamp was designed to compete with carbon-filament light bulbs. It
needed no surrounding vacuum and produced a higher ratio of illumination to
electrical power. Nernst sold the patent rights to his lamp to the Allgemeine
Elektrizitats Gesellschaft in the mid-1890s for a million marks, a good-sized
fortune at the time. But the zirconium rare-earth compound used as the
filament requires preheating to bring its internal resistance down to allow
steady operation, and the lamp was never produced commercially. See Men-
delssohn, Walther Nernst (1973), 44-6.
32
Campbell, PZ, 11 (1910), 826-33. _
33
Campbell, PCPS, 15 (1909), 314, shows rf rather than rf, but this is clearly a
misprint.
spreading waves, the number of impulses TV is independent of the
distance from light to cell. Thus, the decrease in current with light
intensity must be due to a corresponding decrease in the potency
of each event, co. Neglecting the variation in co, the mean square
fluctuation in ionization current is then proportional to co2. Since
the light intensity must be proportional to co, x2 should be propor-
tional to the square of the light intensity. But it was Campbell's
stated "hope" to find evidence against the "spherical wave theory"
of light.34 For the unit hypothesis, co remains constant; each
ionizing event, if it occurs at all, delivers the full unit of energy.
Thus, it must be N that changes with the intensity of the light. A
smaller solid angle of light containing fewer energy bundles will hit
the photocathode as the source is moved away. In this case,
fluctuations should depend directly on N, that is, in proportion to
the first power of the light intensity.
But Campbell's aims were obstructed again. Wheji he included
the anticipated_variation in co by putting back the rf term in the
expression for x2, both cases were reduced to simple proportion-
ality of fluctuation with light intensity. Having considerably ex-
tended the analysis, he was left with no more than a plausibility test
of his method. He derived an order-of-magnitude estimate for the
elusive quantity co, and compared his estimate to what he called
Planck's theory, but which was actually derived from Einstein.35
Einstein predicted that co = i under all circumstances. Abram
Ioffe's reworking of Ladenburg's experimental photoeffect data for
the energy of electrons released from zinc suggested that co = 3.36
Campbell s independent data on fluctuations in the current from
zinc-amalgam photocathodes indicated that co = 3.1, a factor 100
times lower than that expected from any spherical wave theory of
light.
ANISOTROPIC y-RAYS
At the time that Campbell completed his analysis of light in early

1910, Egon von Schweidler applied the same method in a proposed
test of the nature of y-rays. The discussion between Stark and
Sommerfeld had raised Schweidler's hopes of resolving the issue of
34
Campbell, PCPS, 15 (1910), 521.
35
Einstein, AP, 20(1906), 199-206.
36
Ioffe, AP, 24 (1907), 939-40.
energy localization. Einstein's statement at the eighty-first Natur-
forscherversammlung that statistical fluctuations might lie at the
heart of a reconciliation of wave behavior with lightquanta also
influenced Schweidler.37 After the talk, Heinrich Rubens suggested
a possible test of the idea. Like a- and ^-particles, x-rays and y-rays
might produce scintillations on a fluorescent screen.38 Schweidler
quickly published his own proposed test.39 If the effective solid
angle of y-rays allowed into an ionization chamber is reduced,
either by moving the radioactive source away or by closing a
diaphragm, the net ionization current will decrease. But if y-rays
are spherical impulses, the number of pulses entering the chamber
remains constant. Therefore, their effective ionizing power must
be reduced. Conversely, if the energy is localized, the ionizing
power of each unit must remain constant, and only the number of
units that pass into the chamber decreases with solid angle.
The analysis was in principle the same as Campbell's, giving
Campbell an excellent opportunity to criticize the proposal. He
pointed out that any valid analysis must include r\, the fluctuation
in the number of electrons released in each ionization event. And
Campbell was not alone in his criticism. Sommerfeld wrote to
object to the supposition that impulses can exist without an aether.
Planck suggested that discontinuities influorescencemight be due
as much to differing thresholds for emission as to spatial localiza-
tion of radiant energy. Bragg insisted that gas ionization must be
due in either case to intermediary particles, and Schweidler specifi-
cally introduced this possibility into his analysis.40
Schweidler's proposal was tested by Edgar Meyer, then a Privat-
dozent under Stark at Aachen.41 The hemispherical ionization
chamber between A and B in Figure 6.3 received y-rays from the
radium-bromide source at Ra. In trial 1 the entire hemisphere
was open to the rays, whereas in trial 2 the lead ring P limited the
effective solid angle as shown. The ratio of saturation currents,
iji2, is related differently to the ratio of the mean fluctuations of
these quantities depending on whether the y-ray energy is localized
or not.
37
Einstein, PZ, 10(1909), 817-25.
38
Ibid., 826.
39
von Schweidler, PZ, 11 (1910), 225-7.
40
von Schweidler, PZ, 77 (1910), 614-19.
41
E. Meyer, Berlin Sb, 32 (1910), 624-62. For related studies by Meyer, see JRE,
5 (1908), 423-50; 6 (1909), 242-5; VDpG, 12 (1910), 252-74.
Figure 6.3. Meyer's apparatus to test fluctuations in the ionization

current induced by y-rays. [JRE, 7(1910), 283.]
Reporting to the Berlin Academy in June 1910, Meyer distin-

guished between two possibilities, calling one isotropic and the
other anisotropic. The current / is always proportional to Ncoe (in
Campbell's notation), where e is the electronic charge. For the
isotropic hypothesis, where co varies, the ratio of fluctuations in the
current for cases 1 and 2 should be
ei2
that is, directly proportional to the ratio of the currents themselves.
But for the anisotropic case, TV varies while co remains constant.
Thus, the ratio of fluctuations is
Nxcoe \I~N~2 (6.9)
ei2
that is, proportional to the square root of the ratio of currents.
Meyer's data weakly indicated that the anisotropy was the more
likely result, the mean ratio of fluctuations in each of two cases
being at least 50 percent closer to the square root of the ratio of
currents than to the ratio itself.
Meyer did not deny that the impulse hypothesis might explain
his result. He did wonder aloud whether Sommerfeld's success in
finding asymmetry in x-rays to answer his Aachen colleague,
Stark, could also explain the degree of anisotropy he had now
found in y-rays. Early in 1911, Sommerfeld demonstrated that it
could.
SOMMERFELD'S y-RAY THEORY

Arnold Sommerfeld was not fully convinced that all of the issues
raised in the Stark affair had been, or could be, satisfactorily
answered by the impulse hypothesis. William Henry Bragg, whose
discussions with Barkla made him extraordinarily interested in the
Stark controversy, had written Sommerfeld early in 1910 to point
out that the degree of spatial localization required in the energy of
x-rays was considerably more than Sommerfeld's model could
achieve. He cautioned against "dimming the light [of the neutral-
pair hypothesis] which is guiding so many useful results." Al-
though Sommerfeld had little faith in the kinds of mechanical
considerations that moved Bragg, he nevertheless appealed to the
triggering hypothesis rather than to any transformation of radiant
energy to explain the extraordinarily high secondary electron
velocities produced by y-rays. 'The production of secondary [elec-
trons] is an intra-atomic process," he replied, one for which
present knowledge is limited.42 But in the theory of x-rays "we are
on familiar ground." Sommerfeld wanted to see how much locali-
zation of energy an impulse could possess classically; the obvious
case to try was the y-rays.
Early in 1911 he completed the first detailed analysis of impulse
y-rays.43 It was a pulse theory pushed to extremes to describe the
42
W. H. Bragg to Sommerfeld, 7 February 1910, and Sommerfeld to W. H.
Bragg, 8 February 1910. Stuewer, BJHS, 5 (1971), 269-70. Bragg's most
complete summary of the data in support of neutral-pair x-rays and y-rays had
just appered in German in Stark's JRE, 7 (1910), 348-86.
43
Sommerfeld, Miinchen Sb, 41 (1911), 1-60.
radiation produced by ^-electrons accelerated out of an atom at a
speed close to that of light. Sommerfeld had a variety of reasons to
undertake the study. He was influenced by Meyer's experiments
on the fluctuations in ionization current produced by y-rays.
Sommerfeld's analysis was explicitly intended to defuse the poten-
tial problem raised by Meyer's conclusion.
Another motive for Sommerfeld was the revival of doubts about
his analysis of diffraction of x-rays. Two young Berlin experimen-
talists began to claim that the evidence for x-ray diffraction and for
Marx's x-ray velocity determination was faulty. These were re-
vivals of concerns first voiced by one Bernhard Walter, who as
early as 1898 had thought that x-rays were cathode-ray electrons
that had lost their electric charge upon colliding with the anticath-
ode.44 Walter did competent experiments on x-rays and func-
tioned as the chief physicist-advisor on the medical uses of x-rays
in Germany.45 Although he published a series of articles in physics
journals critical of both the Haga-Wind photograph analysis and
Marx's velocity determination, his views had not been taken very
seriously by physicists.46 In 1908 that situation changed. Robert
Pohl and James Franck joined in Walter's arguments and forced
Marx to defend his x-ray velocity experiment again.47 Pohl and
Walter revived the diffraction criticism.48 All three insisted that
they were concerned only about systematic errors in experimental
technique. They did not object to the wave or impulse hypothesis
of x-rays but rather to the evidence on which the impulse hypoth-
esis had been established. But it is likely that their criticism of the
44
Walter, AP, 66 (1898), 74-81.
45
He published extensively in the Fortschritte auf dem Gebiete der Rontgen-
strahlen, a journal serving those with medical interest in x-rays. For a discus-
sion of Walter's role in setting early radiological standards, see Serwer, Radia-
tion protection (1976).
46
Walter, KDM4 (1901), part 2, p. 43',PZ, 3(1902), 137-40. Haga and Wind, AP,
10 (1903), 305-12. Walter, Fortschritte auf dem Gebiete der Rontgenstrahlen,
9(1906), 223-5.
47
Franck, VDpG, 70(1908), 117-36. Marx, VDpG, 70(1908), 137-56, 157-201.
Franck and Pohl, VDpG, 10 (1908), 489-94. Marx, AP, 28 (1909), 37-56,
153-74; 33 (1910), 1305-91. Franck and Pohl, AP, 34 (1911), 936-40. Marx
stopped the exchange with the question, "Are my experiments on the velocity
of x-rays by interference of electrical waves explainable?" AP, 35 (1911),
397-400.
48
Walter and Pohl, AP, 25 (1908), 715-24; 29 (1909), 331-54. Pohl, Physik der
Rontgenstrahlen (1912), 144.
experiments had been encouraged by the new theoretical objec-
tions raised by Stark in 1908 against an aether model of x-rays.
The chief motive behind Sommerfeld's analysis of y-rays was his
growing interest in the quantum transformation relation {not the
lightquantum). Late in 1910, H. A. Lorentz had questioned
whether the quantum hypothesis could explain absorption of
radiation at all without some additional hypothesis to limit the
spatial extent of radiant energy.49 In 1909 Sommerfeld had denied
that the quantum relation had anything to offer to the analysis of
impulse radiation; in 1911 he explicitly reversed that statement.50
The y-ray study was intended to answer Lorentz and was Sommer-
feld's first use of the quantum transformation relation. It was
intended to show how the quantum relation, formulated in terms
of frequency, could be extended to nonperiodic events such as the
ejection of an electron by a disintegrating atom.
The same asymmetry that had answered Stark's objections
regarding x-rays also served to define the localization of y-ray
impulses. The two cases are remarkably similar even though in one
case the electron is stopping and in the other it is starting. The
electron accelerates from rest at point / toward the right in Figure
6.4 and attains its final velocity at point / The resulting field
exhibits a variation in pulse width around the azimuth similar to
that of the decelerating electron in Figure 5.5. One qualitative
difference is that the quiescent and dynamic fields are inter-
changed; here the former, which Sommerfeld called the electro-
magnetic atmosphere, is the outerfield.51Another difference is that
once the electron reaches relativistic velocities, the dynamic field
within the inner sphere deviates from the schematic symmetry
shown in Figure 6.4. The field of a relativistic electron becomes
restricted to the region between two cones facing each other, each
with its vertex on the electron and symmetrically oriented about
the axis of acceleration.52 As the velocity approaches c the angle
49
Lorentz, PZ, 11 (1910), 1234-57. These matters will be discussed at length in
Chapter 7.
50
Sommerfeld, Miinchen Sb, 41 (1911), 43n.
51
Ibid., 4.
52
Heaviside, PM, 27 (1889), 324-39. J. J. Thomson had concluded in 1898 that
the electrodynamic effect of stopping an electron abruptly from a speed close
to that of light is twofold. The roughly planar impulse of the surrounding
dynamic field would continue as though the electron had not stopped. At the
same time, a spherical impulse would arise, centered on the arrested electron
and of essentially infinitesimal thickness. Thomson, PM, 45 (1898), 172-83.
Figure 6.4. The electric field in the vicinity of a ^-particle rapidly

accelerated to the right.
subtended by each cone approaches m steradians, so that the field

isfinallyrestricted to the plane that is perpendicular to the axis and
that passes through the moving electron. A schematic approxima-
tion to the actual field for an electron accelerated to 0.9c is pre-
sented in Figure 6.5. The electric force in the inner sphere acts only
within the trailing conical "bow wave" indicated by the vertical
lines; at each instant it is directed along these lines, always in, or
close to, the plane normal to the velocity vector. The broken
arrows indicate the direction of maximum energy flux.
Because the acceleration of the ^-electron is so much greater in
magnitude than even the deceleration of a cathode-ray electron,
the radiated energy is more localized in the forward direction for
the y-pulse than for the x-ray impulse. Cathode ray electrons move
at some io9 cm/sec; /^-electrons travel at 0.9 to 0.99 times the
velocity of light, some thirtyfold faster. To assess the difference in
effect, Sommerfeld started, as he had for x-rays, with Max Abra-
ham's reformulation of the Poynting equation.53 Energy passing
53
M. Abraham, Elektrizitat, 2 (1905), section 14, equations 74, 75, and 76a.
Figure 6.5. Schematic representation of the electric field in the vicinity

of an electron accelerated to relativistic velocity.
per unit time at a distance r from the point /, and at an angle 6 from
the acceleration axis as seen from point /, is given by
e2a2 sin2 6 (6.10)
\6n2c3r2 (1 - (v/c) cos Of
Here a is the magnitude of the acceleration and v is the final
velocity of the electron. The factor sin2 6 arises from the transverse
nature of the impulse; its amplitude falls to zero in the forward and
backward directions and presents a maximum at 900 to the axis.
This is true regardless of the time duration of the acceleration. The
modulating factor (1 (v/c) cos 9f produces the asymmetry in
the radiated energy. It arises from the classical interpretation of
wave intensity and explains the localization of y-ray energy in the
forward direction.
Figure 6.6. SommerfekTs calculated distribution of y-ray intensity as a

function of azimuthal angle. [Munchen Sb, 41 (1911), 11.] (The two
cases for v/c are not to the same scale.)
EXPANDING-RING y-RAYS
The energy is radiated into a hollow cone that is symmetric about
the axis of electron acceleration. Integrating equation 6. io over the
acceleration time produces the total energy radiated into the cone
defined by the angle 9:
sin20 1
(6.11)
cos 6 [(I (v/c) cos 6)4
For values of v close to c, and for small values of 6,
62
=s 16 eyeg = 16/06 (6.12)
-v/c + 62/2)
Figure 6.6 is a plot of S(9) for two cases v = o.gc and v = 0.99c. In
the former case 5max lies at about 150, and in the latter about 50,
from the axis of acceleration. The energy never quite rises above
zero in the forward direction but is the more tightly constrained
Figure 6.7. The ring-pulse y-ray.
around the axis the closer the final velocity is to c.54 Corrections for
nonlinearity in the acceleration and for the relativistic increase in
electron mass cause only a small broadening of the cone.
The net result is a ring-shaped concentration of energy. The
impulse is still a temporally discontinuous event, so that the energy
always lies at the intersection of the hollow cone with the expand-
ing spherical shell. Thus it forms a ring, as in Figure 6.7, that
expands in proportion to distance as it travels in the same direction
as its parent ^-electron. "One can see from this," Sommerfeld
continued hopefully, "that for very great velocities the energy
emission is almost corpuscular."55
"It is quite remarkable," Sommerfeld said, "that on the basis of
the Maxwell theory, that is, pure undulatory theory, one is led
under certain conditions to consequences that approximate the
Newtonian emission theory, and that appear to have more similar-
ity to Newtonian light projectiles than to Huygens' spherical
waves."56 In fact, Smax grows in proportion to (1 v/c)~3; thus,
energy radiated into the cone of maximum effect resulting from
acceleration to 0.99c is io3 times greater, as well as being more
localized, than that due to the slower 0.9c electron.
Thus far, the calculation was entirely classical. But to compare
results to experiment, Sommerfeld needed data, as yet unavail-
able, on the precise nature of the acceleration. He wanted to obtain
an expression for Efi/Ey, the ratio of total ^-particle energy to that
which appears in the resulting y-pulse. Sommerfeld expressed the
54
Sommerfeld calculated the extremum values of 6 where S dropped to 0. iSmax.
For v/c = 0.9, these limits are 12.5 to 470, for v/c = 0.99, they are 4 to 160.
55
Sommerfeld, Munchen Sb, 41 (1911), 13. A misprint on p. 10,1.12 puts /? for y.
56
Ibid., 4.
unknown acceleration in terms of the distance / over which it
occurs and obtained the result
EJEi-^ jx=W (6.13)
where me is the rest mass of the electron. It was to eliminate the

unknown distance / that he turned to the quantum theory. "The
preceding considerations are already so hypothetical," he ex-
plained, that it was reasonable to extend "the fundamental hy-
pothesis of the Planck radiation theory to [include] radioactive
emission."57 But how to formulate the problem when the interac-
tion is nonperiodic and so lacks a frequency? The answer emerged
from a reconsideration of the relationship between the energy and
the duration of discontinuous electromagnetic impulses.
An entire section of Sommerfeld's study was devoted to the
relation between pulse width and the azimuthal angle from which
the impulse is viewed. The pulse width is, he said, a "theoretical
measure of the hardness of the y-ray."58 This followed from the
analogy to the x-ray case, and, as was true for x-rays, rested on no
direct experimental evidence. The hardness of x-rays depends on
the potential of the discharge tube; the greater the kinetic energy
dissipated in the collision of an electron, the harder the x-ray that
results. For the y-ray case, the analogous condition concerns the
magnitude of the electron acceleration; the more rapidly the
^-electron reaches full velocity, and the greater that final velocity,
the harder the y-ray pulse. Like the x-ray impulse radiated from a
decelerated electron, the hardness of the y-impulse also varies with
azimuth. For Sommerfeld, by 1911 a direct connection had been
formed between the energy that contributes to an x-ray or y-ray
impulse and the pulse width that results.
To discuss the azimuthal variation Sommerfeld used the symbol
X for the pulse width, but I shall continue to use the more neutral 5.
The symbol X is traditionally applied to periodic waves and was
used by Sommerfeld "to remind us of the analogy of the [pulse
width] to the wavelength of periodic light."59 He calculated the
azimuthal variation in S for the y-pulse of Figure 6.6. The really
significant value, the massgebende Impulsbreite, was the one de-
57
Ibid., 24.
58
Ibid., 34.
59
Ibid., 35.
fined at the azimuthal angle #max, of maximum energy flow.60 Its
value is
Sm = ^ (1 + Vl-i> 2 /c 2 ) - /cos <9max (6.14)
Increasing the velocity of the final electron produces a smaller Sm,

and <5m vanishes completely when v = c. In general, decreasing the
value of 6 implies greater hardness; the hardness of both x-rays and
y-rays increases with the velocity of the stimulating electron.
From this perspective, Sommerfeld made the same conceptual
break with classical physics that Wien had made four years before.
"If we accept the hypothesis that these energy ratios are connected
to Planck's h, we can find a formula that derives [S] simply from
the velocity [of the electron] without the auxiliary hypotheses for
the [experimental] energy measurements."61 In its simplest form,
he set the product of impulse duration r and the electron's kinetic
energy equal to h. Like Wien and Stark before him, Sommerfeld
here formally adopted the nonclassical concept that the impulse
duration alone suffices to determine the energy contributed to the
pulse. Recognizing moreover that, although pulse width varies
with azimuth, the duration of the event producing the impulse
does not, he went on to claim that the inverse pulse duration
defines a sort of frequency for the impulse. In retracting his claim
of 1909 that the quantum condition does not apply to the Brems-
anteil impulse rays, he admitted, "I have only recently come to
realize that the pulse duration r can enter in the place of Planck's
i/v" 62
So, Sommerfeld's very formulation of the quantum condition
for nonperiodic systems was predicated on the novel realization
that the energy contained in impulses may be characterized by
pulse duration, or frequency, rather than pulse amplitude, the
classical measure of intensity. He elected this condition: In each
discontinuous emission, just one quantum of action h is expressed.
In the general case he required that the integral of the energy,
60
Ibid., 36.
61
Ibid., 38-9.
62
Ibid., 4311. It was commonly accepted at the time, especially in Germany, that
the pulse width varied with the quality or hardness of the x-rays. See, among
others, Pohl, Physik der Rontgenstrahlen (1912), where the idea permeates the
text. There was little experimental evidence to support this claim. The only
significant question was raised by Seitz, AP, 26 (1909), 448.
Table 6.1. Ratio of kinetic energy offi-p articles to

that ofy-rays {Ep/Ey)for three elements
Element
RaE UrX RaC
Experimental 120 44-9 2

Predicted by Sommerfeld 330 150 100
expressed over the time interval of the acceleration, is always equal

to h:63
Jo
f
Thus, Sommerfeld reformulated the quantum condition in terms
(6.15)
of action, the product of energy with time, rather than energy

alone. This change altered the need to formulate the problem in
terms of frequency. Sommerfeld succeeded in finding a general-
ized quantum condition that would apply even when no frequency
could be isolated, and this study was the beginning of work that
would lead him by 1915 to the phase integral conditions for atomic
systems.
With the quantum condition of equation 6.15, Sommerfeld
sought agreement between his theoretical formulation and experi-
mental data. Hefirstexamined the ratio of electron kinetic energy
to the resulting impulse radiant energy for both x-rays and y-rays.
Russell and Soddy, H. W. Schmidt, and William Wilson had all
given values of the y-ray ratio for various velocity /?-rays.64 Som-
merfeld's quantum formula for the ratio predicted values that
agreed only approximately with the experiment for two of three y
sources (see Table 6.1). Applying the same test to Wien's x-ray
ratio, Sommerfeld found "surprising" agreement, within 15 per-
cent.65 He was also able to compare the value of expected pulse
63
Should E be only the kinetic energy or the total energy? Should the time
dependence of E be taken into account? Sommerfeld did not know, but he
selected equation 6.15 as the most useful representation.
64
Russell and F. Soddy, PM, 21 (1911), 130-54. Schmidt, PZ, 8(1907), 361-73;
AP, 23 (1907), 671-97; PZ, 10 (1909), 6-16. W. Wilson, PRS, 85A (1910),
141-50.
65
Sommerfeld, Miinchen Sb, 41 (1911), 33.
width of y-rays from some radioactive elements. The resulting
values ranged from 3 x io~ n cm to 5 x io~13 cm, some 1.0 to 0.01
percent those of hard x-rays. Sommerfeld's quantum regulation
predicted an x-ray pulse width of 2 x io~9 cm. This agreed with
Wien's quantum value when corrected for Sommerfeld's use of an
energy ratio involving only the polarizable (impulse) part of the
x-rays. Wien, of course, had not known in 1907 that some of the
x-ray energy he measured was due to characteristic x-rays.
But Sommerfeld's chief concern was to show that the difference
between his purely electromagnetic calculation and the quantum
treatment was more illusory than real. "The two theories do not
exclude, but augment each other," he said.66 Wien had been
unable to show a connection between his electrodynamic treat-
ment of x-rays in 1905 and his quantum analysis in 1907. Stark had
given up the electrodynamic solution entirely. Sommerfeld hoped
that his y-ray theory showed that a reconciliation was not only
possible but reasonable. His electromagnetic calculation for the
emission of the y-pulse was successful as far as it went, but it left the
acceleration, or the impulse duration, unspecified. That was where
the quantum hypothesis, reformulated in terms of action, offered
assistance. "In optics," Sommerfeld said, "consequences of the
wave theory must be retained even though they remain unex-
plained by Planck's discovery . . . In the same way consequences
of the impulse theory of x-rays can be integrated with the quantum
of action."67
SECOND-GENERATION TESTS OF ANISOTROPY
Encouraged by the success of Sommerfeld's y-ray theory, Edgar

Meyer continued with the fluctuation analysis in 1911. Campbell
had been quick to point out that Meyer, like Schweidler, had
neglected a possible variation in co, the number of electrons pro-
duced by each impulse.68 Precisely this variation had stood in the
way of Campbell's tests. The issue had taken on new significance
by 1911 because Bragg had begun to claim more forcibly that the
value of a; is 1; each neutral-pair y-particle gives rise to one and
only one secondary /^-electron. That ^-particle is then responsible
66
Ibid, 41.
67
Ibid., 41-2.
68
Campbell, PZ, 11 (1910), 826-33.
Figure 6.8. Meyer's apparatus for a second attempt to determine an-

isotropy of ionization by y-rays. [AP, 37 (1912), 705.]
for all subsequent ionization in the gas, according to Bragg.69 If

Bragg's hypothesis was true, it seemed that the fluctuation analysis
could not distinguish dependent from independent events at all.
Meyer undertook a new series of tests that were simultaneously
designed to answer Campbell's criticism and test Bragg's hy-
pothesis.
Meyer shot y-rays across a double capacitor in a direction
parallel to that of the field gradient, as shown in Figure 6.8. He
compared the ionization produced in the isolated gas volumes on
both sides, reasoning that if the ionization events were fully inde-
pendent, the fluctuations in saturation current in chambers i and i
would remain uncorrelated.70 The mean square fluctuation in the
differential current measured at the grounded plate C should not
change appreciably when the polarity of thefieldin one chamber is
changed. If plates A and B are both positively charged to io3 V, a
case Meyer denoted by the subscript ++, the fluctuation e++ will
reflect no more than the random production of electrons. If the
polarity of plate B is reversed, the positive-ion products change
69
W. H. Bragg and Porter, PRS, 85A (1911), 349-65.
70
E.Meyer, PZ, 13 (1912), 73-81; AP, 57(1912), 700-20.
direction in chamber 2, but the absolute number of electrical
impacts at plate C, and hence the fluctuation in the current, will be
unchanged. Therefore, for uncorrelated events, ++/+_ ~ i. But if
a given y-ray produces more than one electron along its path, the
correlated reception of electrical charges at plate C will construc-
tively reinforce the + + fluctuation and cancel the H; the ratio
++/+_ will then approach some maximum value much greater
than i.
Meyer reasoned loosely that the ratio ++/+_ represents the
"coupling" between events in the two chambers; it gives the
number of electrons produced, on the average, by each y-pulse. He
was apparently ignorant of the extensive British literature on
differential ion mobilities; these showed that he could not count on
the simultaneous arrival at C of the electron produced in chamber
i with the ion produced in chamber 2 even if they were produced
by the same y-ray. But a typical test produced the ratio e+Je+_ =
2.94; to Meyer, this high value indicated a "strong interdepen-
dence between ionization processes in [regions] 1 and 2."71
Then he turned the capacitor on its side and shot the y-rays in
parallel to plate C He called this the transverse case and contrasted
it to the earlier longitudinal test. The longitudinal case, he claimed,
had measured the correlation between ion products produced by
the same y-ray pulse or-particle along the line of motion. The
transverse case promised a measure of events produced along the
hypothetical pulse front. Again he found a correlation; the trans-
verse ratio ++/+_ was also greater than 1. Meyer concluded that
neither Bragg's nor Einstein's hypotheses could explain the degree
of correlation he had demonstrated. Sommerfeld's electrodynamic
theory of y-rays was the "simplest explanation," even if it was not
fully able to explain the high secondary /?-velocities.72 A student of
Sommerfeld's, Eberhard Buchwald, analyzed Meyer's results.73
He felt that Meyer's data allowed him to determine roughly the
size of the angular region over which correlations could still be
found in y-ray-induced ionization. The half-angle he obtained
was 170, which was in reasonably good agreement with the angular
width Sommerfeld assigned to y-rays from radium C caused by
electron emission at 0.95c.
71
E. Meyer, AP, 57(1912), 708-9.
72
Ibid., 719.
73
Buchwald, AP, 39 (1912), 41-52.
74
Campbell applied his own tests to Meyer's data. He allowed for
six possible outcomes, three each for the two possibilities of (A)
spatially localized and (B) spreading waves. The three were (i)
Planck's suggestion that atoms are predisposed by chance to ioni-
zation, (2) Bragg's suggestion that ionization is due to ^-particle
intermediaries, and (3) the long-honored triggering hypothesis.
Campbell concluded that the evidence favored option A3, al-
though he had known since 1907 that any triggering hypothesis was
difficult to reconcile with the number of electrons in the atom.
"This combination of theories," he concluded in despair, "appears
to unite the disadvantages of all theories and to possess the advan-
tages of none."75
Although the tests of radiation by fluctuation analysis were not
over, they had run aground as of 1912. Uncertainty over the
variation in the parameter a) effectively masked the significance of
the results. All those who tried the experiments concluded that
y-rays are anisotropic. The question was, how far from spherical
symmetry is the pattern of radiated energy? To the extent that any
quantitative result was offered, the empirical data matched Som-
merfeld's ring-pulse y-rays the best.76
BRAGG'S APPEAL TO DUALISM
Bragg had held fast to the neutral-pair hypothesis. It was partly the
completeness of Barkla's analysis for the softer x-rays that gradu-
ally pushed the domain of their discussion closer to that of the hard
x-rays, where Bragg held the advantage. When Rutherford and
Geiger showed in 1908 that the oparticle carries twice the charge
of the electron,77 Bragg cast aside the a-particle as the positive
constituent of a pair in favor of an undefined "positive electron."
In 1909 Bragg responded directly to Thomson's proposal that each
electron radiates into a single narrow cone. By then, overwhelming
evidence had accumulated for the forward scattering of secondary
products of x-rays and y-rays. Bragg claimed that even Thomson's
"energy bundles of very small volume" must interact with individ-
74
Campbell, PZ, 13(1912), 81-3.
75
Ibid, 83.
76
Stark, PZ, 13 (1912), 161-2. E. Meyer, PZ, 13 (1912), 253-4. Laby and
Burbidge, PRS, 86A (1912), 333-48. Burbidge, PRS, 89A (1913), 45-57.
77
Rutherford and Geiger, PRS, 81A (1908), 162-73. Bragg and Madsen, PM, 16
(1908X938.
ual atoms to explain this behavior. "It seems hard," he sighed in
this, his last paper before returning from Australia to England, "to
understand the distinction between such bundles and entities
generally classed as material."78
But in 19io, following his removal from Adelaide to Leeds, a
variety of influences caused Bragg to recognize that his neutral-
pair hypothesis could not be the full answer to the problems of
radiation. First, he came into closer contact with the continental
literature. The debate in Germany between Stark and Sommerfeld
mirrored many of the issues still unresolved between himself and
Barkla. A review article on his position, prepared for both British
and German readers in 1910, illustrates the more reflective and less
dogmatic expression of his hypothesis, which he tailored for a
European audience. When a neutral pair is disrupted, the constitu-
ent electron proceeds on course; "the form of the entity may
change, y into /?, X into cathode ray, and so on," he said, "but there
is so little change in anything but form that practically we may
assume a continuity of [particle] existence."79
So too did the narrowing gap between x-rays and visible light
influence Bragg's position. In 1910 two independent tests, one by
Bragg's compatriot R. D. Kleeman, showed that forward scatter-
ing was also found in the photoelectric effect using ultraviolet
light.80 Bragg had never presumed to extend his hypothesis to
periodic light waves;findingthat they too demonstrated one of the
chief arguing points for neutral pairs caused him to step back for
reflection. Perhaps there was some truth in the growing conviction
of many physicists by 1911, British as well as German, that
Barkla's characteristic x-rays bridged the gap between periodic
light waves and impulse x-rays.
Bragg was also impressed by how far toward a particlelike
concentration of y-ray energy Sommerfeld had managed to come.
But Bragg felt that Sommerfeld had not gone far enough. Soon
after reading the y-pulse study, he wrote to Sommerfeld to com-
plain that the rings still dissipated their energy too rapidly.81 How
78
W. H. Bragg and Glasson, PM, 17 (1909), 855-64. Bragg gave a final talk on the
scattering of x-rays at the January 1909 meeting of the Australasian Associa-
tion for the Advancement of Science but no text remains, Report, 72 (1910),
113.
79
W. H. Bragg, PM, 20 (1910), 396.
80
Kleeman, Nature, 83 (1910), 339; PRS, 84A (1910), 92-9. Stuhlmann, Nature,
55(1910), 3 i i ; / W , 20(1910), 331-9. See Chapter 9.
81
W. H. Bragg to Sommerfeld, 17 May 1911. Steuwer, BJHS, 5 (1971), 271-2.
could the ring-pulse pass on virtually its entire energy to an
electron after traversing macroscopic distances? But while he ar-
gued with Sommerfeld the dangers of continued attempts to make
the pulse hypothesis conform with the ionization data, Bragg
simultaneously hedged his bet. His basic point, he concluded, was
not the physical nature of x-rays but the claim that in ionization,
particlelike accumulations of energy are transferred to single elec-
trons. One to one. He could avoid the quaking ground of radiation
theory by separating the x-rays per se from the ionization event.
With his new colleague at Leeds, H. L. Porter, Bragg proposed at
the Royal Society that x-rays and y-rays do not ionize atoms
directly. One need only assume that a neutral pair disintegrates
before ionization can occur. It is the resulting high-speed electron
that strikes and ionizes the atom, they claimed.82
Just two weeks before, C. T. R. Wilson had reported to the Royal
Society on his most recent experiments to make visible the tracks
of ionizing particles by means of vapor condensation.83 Here was
the clear evidence that Bragg had long sought to demonstrate
unambiguously that x-rays ionize gases in precisely the same
manner as do particulate /?-rays. Wilson's photograph, reproduced
in Figure 6.9, shows only multiple short tracks; there is no sign of
the diffuse cloud that ionization by spreading impulses would
seem to require. Bragg made certain that the message got to
Sommerfeld. If the energy of the rays were not concentrated in
unchanging particlelike units, he claimed, exceedingly long pe-
riods would be necessary for sufficient energy to build up on a
single electron from a succession of impulses.84
More perplexed than ever by the ambiguous action of x-rays and
y-rays, Bragg returned to the dual nature he had first ascribed to
x-rays in 1907.85 He began to call for a theory of x-rays, and to a
lesser extent of y-rays, that would embrace both particle and wave
characteristics. The existence of two sorts of x-rays, the polarizable
but inhomogeneous rays from the x-ray tube and the unpolarized
characteristic rays from irradiated matter, reflected for him a
82
W. H. Bragg and Porter, PRS, 85A (1911), 349-65. W. H. Bragg, PM, 22
(1911), 222-3; 23 (1912), 647-50.
83
C. T. R. Wilson, PRS, 85A (1911), 285-8.
84
W. H. Bragg to Sommerfeld, 7 July 1911. Stuewer, BJHS, 5 (1971), 273.
85
W. H. Bragg, TPRRSSA, 31 (1907), 94-8. He had suggested that along with the
neutral pairs, impulses coexist in an x-ray beam. This was to answer Marx's
velocity measurements.
Figure 6.9. C. T. R. Wilson's cloud-chamber photograph of ionization

by x-rays. "The whole of the region traversed by the primary [x-ray]
beam is seen to be filled with minute streaks and patches of
cloud . . . mainly small thread-like objects not more than a few milli-
metres in length, and many of them considerably less than 1/10 mm in
breadth." [PRS, 55^4 (1911), facing 286.] The x-ray beam enters from
the left.
fundamental duality. Even before the riveting news that x-rays

could be made to interfere, Bragg issued a clear call for a dualistic
interpretation of x-rays. At the British Association meeting in
1912, he sought guidance in the inherent ambiguity of speculations
on light by Newton and Huygens.86 Earlier he had recommended
forgetting "that idea of keeping touch with electromagnetic theory
as we fancy it must be, which is hampering every movement."87
Elegant and interesting as it was, Sommerfeld's pulse analysis
for y-rays raised more problems than it solved. Grappling with
some of those problems soon led Sommerfeld to significant ad-
vances in quantum theory. But in rejecting the triggering hy-
pothesis, Sommerfeld opened himself to Bragg's criticism: No
spreading-wave explanation could explain how x-rays and y-rays
retain their localized energy across macroscopic distances. Som-
86
W. H. Bragg, BAAS Report (1912), 750-3. "I am very far from being averse to a
reconcilement of a corpuscular and a wave theory: I think that someday it may
come." Bragg to Sommerfeld, 17 May 1911. Stuewer, BJHS, 5 (1971), 271-2.
For earlier strong hints of the direction of Bragg's thoughts, see BAAS Report
(1911), 340-1. For the influence of Bragg's dualistic claims, see Jeans, BAAS
Report (1913), 380; Report on radiation (1914), 63ff.
87
Bragg, Studies in radioactivity (1912), 191.
merfeld's hypothesis could not answer why, nor could Thomson's.
Either the energy in the rays must be localized like actual particles,
or some mechanism must be found whereby many impulses can
conspire to act on a single electron before most of them have
arrived at that electron.
This latter problem had first been debated, not for x-rays or
y-rays but for visible light. And in the period 1909- 11, it had
become clearer and clearer that whatever problems complicated
the understanding of x-rays and y-rays affected periodic light waves
as well. For the latter, Sommerfeld's techniques provided no ref-
uge, nor could one even claim the advantage of temporally discon-
tinuous impulses. The full weight of accumulated evidence that
light is a periodic, spherically propagating wave stood squarely in
the way.
7
Problems with visible light
The speaker does not wish to deny the heuristic value of

the [lightquantum] hypothesis, only to defend the
[classical] theory as long as possible.1
Between 1909 and 1912, many influential physicists first realized

that the problems preventing a consistent understanding of x-rays
applied as well to ordinary light. But unlike the relatively new
impulse theory of x-rays, the wave theory of light was exceedingly
well founded. It rested on a century's accumulation of experimen-
tal evidence. Spatial concentration in the energy of light could not
be attributed to a temporal discontinuity, as it could for x-ray
impulses. Light was known to be a repeating periodic wave. It was
not fully realized at first that hypotheses about the nature of
ordinary light required modification. For a decade after 1910, the
significance of the growing empirical evidence favoring spatial
localization of luminous energy went largely unrecognized. This
occurred because the data ran counter to the orthodox view of
visible light, the spectral region wherein the periodic properties of
radiation were most easily demonstrated and most firmly estab-
lished.
H. A. Lorentz attacked the problem in much the same spirit, but
with quite the opposite intent, as had Einstein. He showed in 1909
that the lightquantum hypothesis is incompatible with the quan-
tum transformation relation itself, let alone with classical ideas
about radiation. In so doing, he laid the foundation for a restate-
ment of a major difficulty with any classical explanation of the
photoelectric effect: It takes an extraordinarily long time for the
observed quantity of energy to build up from periodic waves
incident on an electron. Any nontriggering explanation of the
1
Lorentz, PZ, 11 (1910), 1250.
Problems with visible light 169
photoeffect had to clarify how this delay time, first measured in
1913, could be two orders of magnitude shorter than that predicted
classically. As if to emphasize the difficulty, the triggering hypoth-
esis itself was shown to conflict with new experiments. By 1911 the
photoefFect had become a recognized problem for physicists.
This new concern over the nature of light was set against a
growing awareness that x-rays and light are manifestations of the
same phenomenon. A few, such as Stark, had believed this was so;
between 1908 and 1912, experiments repeatedly confirmed the
suspicion. The homogeneous characteristic x-rays seemed in all
respects to constitute the x-ray analog to optical emission spectra.
Homogeneity in the penetrating power of the secondary x-rays was
increasingly attributed to uniformity of impulse width; by analogy
to light, this implied that the characteristic x-rays are, in a sense,
monochromatic. The existence of a well-defined threshold for the
emission of characteristic x-rays suggested a mechanism similar to
Stokes' law of optical fluorescence. Finally, in 1912, it was found
that x-rays can interfere. These results made it imperative that a
resolution to the paradoxes be found not just for x-rays but for light
as well.
The sharpness of the issues was assured by the double-edged
sword of analogy. In the early years of the century, impulse x-rays
were thought to be sufficiently different from periodic light so that
notions of discontinuity, for example in the operational definition
of x-ray intensity, could be applied without requiring that similar
considerations hold for light. As the data confirming the equiva-
lence of x-rays and light increased, the need to combine under-
standing of both raised anew for light all of the paradoxes of x-rays.
The analogy of x-rays to light, at first divergent and now conver-
gent, began to raise serious doubts about the classical understand-
ing of light as a periodic-wave phenomenon.
H. A. LORENTZ AND ENERGY ACCUMULATION

The highly respected Dutch theoretical physicist Hendrik Antoon
Lorentz was uneasy about the lightquantum. He was universally
regarded as a master mathematical physicist, and his views were
taken very seriously by scientists of all nations. Although he
himself had done no work on x-rays, Lorentz accepted the impulse
hypothesis and thought that there was a qualitative difference
between x-rays and periodic light waves.2 Hefirstbegan to take the
quantum transformation relation seriously in 1908, but he raised
strong arguments against the spatially localized Hghtquantum.3 In
a talk on the subject in April 1909 to the Dutch Congress of
Scientists and Physicians, he pointed to the "striking" successes of
the Hghtquantum: It explained Stokes' law; it seemed to be consis-
tent with the findings of recent experiments on the photoeffect;
and it appeared to explain Stark's Doppler shift in the emission
spectrum from canal rays.4 But, Lorentz concluded, "on closer
examination serious difficulties arise for the hypothesis."5
It was primarily phenomena of interference that convinced
Lorentz that light could not consist of localized lightquanta. Split
beams still interfere after a separation of two million wavelengths.
To achieve this, each Hghtquantum would have to extend almost a
meter in the direction of propagation. This argument weighed
heavily against the Hghtquantum for Planck too.6 Lorentz pointed
out that if each Hghtquantum spread its effect over so long a
distance, it conflicted with the principle of the quantum hypoth-
esis. To receive its full quantum of energy, a molecule would have
to wait for the entire train of the lightquantum's "vibrations" to
pass over it. But how can this ever occur if the minimum quantum
of energy that the molecule can take up exceeds the amount
contributed by each impulse? How does the molecule sense that
enough vibrations are yet to come so that it can start to absorb
them? How can the process of absorption even begin?
Another version of the interference argument showed that light-
2
This is borne out more by what he did not say than by what he did say. In
addressing the Dutch Congress of Scientists and Physicians in 1907 and 1909,
talks on "Light and the structure of matter" and on the "Lightquantum
hypothesis," he never mentioned x-rays. Lorentz, NNGC Handelingen, 11
(1907), 6-21; 12 (1909), 129-39. In his famous Wolfskehl lectures in 1910 he
discussed a variety of quantum issues, including the photoelectric effect. But he
did not discuss x-rays. PZ, 77(1910), 1234-57. Nor did he include x-rays in his
1911 article on the "Nature of light" for the Encyclopaedia Britannica, 1 ith ed.,
16 (1911), 617-23.
3
For the background and story of Lorentz' conversion to the quantum transfor-
mation hypothesis, an event of major significance in the widespread adoption
of the concept, see Kuhn, Black-body theory (1978), chap. 8.
4
Lorentz, NNGC Handelingen, 12 (1909), 129-39.
5
Ibid, 136.
6
Planck to Lorentz, 7 October 1908, 16 June 1909, 7 January 1910, and draft of
Lorentz to Planck, 30 July 1909. Lorentz papers. See also Planck's remarks
following Einstein's lecture, PZ, 10 (1909), 826.
quanta must also spread their effect in directions perpendicular to
propagation. Imagine a telescope adjusted so that a star image is as
sharply defined in the focal plane as the theory of diffraction will
allow. If the objective lens or mirror is now partially blocked, so
that only light from a small area of the original objective is seen,
the image of the star is degraded. A smaller solid angle of light
produces a larger image in the focal plane. The light must therefore
spread its influence over the full objective aperture of the telescope.
This is as true for the largest telescope (at the time, Hale's 150-cm
telescope on Mt. Wilson) as for the pupil of the eye. But any single
lightquantum that extends laterally across the Hale telescope can
deliver only io~4 of its energy to the eye; how then can the
quantum transformation rule apply to the retinal stimulation that
constitutes the "sight" of the star? 'The conflict shows," Lorentz
concluded, "that one cannot speak of propagating lightquanta
concentrated in small regions of space that at the same time
remain undivided."7
Lorentz was invited by the Wolfskehl commission to deliver six
lectures at Gottingen in October 1910. The general theme of the
lecture series - entitled "Old and new questions in physics"- was
the aether: the reasons for the concept and modifications in it that
had recently been suggested.8 In thefifthlecture, Lorentz discussed
the lightquantum hypothesis, with particular reference to the pho-
toelectric effect. The triggering hypothesis he found wanting. A
simple heated cathode, with no light shining on it, emits electrons
with kinetic energy on the same order of magnitude as that of the
cathode molecules. According to equipartition of energy, this is
the expected behavior of the free electrons within the metal but not
of those bound to atoms. If the light triggers the release of only
unbound electrons, their velocities must also be comparable to the
kinetic velocity. Experiment showed that they arefivefoldgreater.
Moreover, it had been shown that heating the photocathode does
not increase the measured velocity as it certainly should have if
free electrons were being triggered.9
Lorentz' argument did not, in fact, address the true triggering
7
Lorentz, NNGC Handelingen, 12 (1909), 139.
8
Lorentz, PZy 11 (1910), 1234-57.
9
Lorentz cited only Ladenburg's studies, but there had been others beginning in
1906. Lorentz' claim in 1910 that the linear law had been confirmed was
premature. Wheaton, Photoelectric effect (1971), chap. 13.
hypothesis because he did not consider the bound electrons. The
velocity of an electron released from its atom is a function of the
binding energy; depending on the details of the then unclear
atomic structure, it could very easily be five times the thermal
energy of a cathode molecule. The triggering hypothesis was never
intended to apply to unbound electrons; the photoelectron must
have a frequency, by virtue of its motion within the atom, in order
to be triggered by the light.
In regard to the opposing view, that the photoeffect is a transfor-
mation of light energy to electron kinetic energy, Lorentz showed
that Einstein's lightquantum explanation seemed to accord well
with observations. But Lorentz judged it "impossible" if light-
quanta were thought to be fully independent of one another.10
"The speaker does not wish to deny the heuristic value of the
hypothesis," he said, "only to defend the old theory as long as
possible."11
Johannes Stark addressed the same issues at about the same
time, asking rhetorically if it was really a "fantastic obstinacy," as
some had suggested, that drove him to explain events "by observ-
ing their course" rather than from a "mathematical theory [such as
Sommerfeld's] based on abstract concepts."12 He felt that the very
existence of the concept of an aether stood most in the way of a
clear view of the problem. Like Campbell, Stark wished to do away
with the aether entirely. But his hypothesis that interference effects
are due to "coherent" aggregations of lightquanta that possess a
periodic structure hardly convinced Lorentz. Light not only had to
come in indivisible quanta, Lorentz objected, but to listen to Stark,
it had to come in fixed groupings of quanta as well.
CLASSICAL ABSORPTION OF LIGHT
Lorentz analyzed classical waves to see how they might deliver

quantum units of energy to bound electrons. A monochromatic
light beam acts on an electron bound to an atom by a quasi-elastic
force. Depending on the match between the frequency of the
incident light and the oscillation frequency of the bound electron,
10
Lorentz, PZ, 77 (1910), 1249.
11
Ibid., 1250.
12
Stark, JRE, 7(1910), 387.
energy is transmitted from the light to the electron. At the same
time, the accelerating charged electron radiates energy - the so-
called radiative damping. Lorentz used the case of maximum
energy transfer, which occurs when the incident frequency is equal
to the natural frequency of the electron. When the radiative loss
equals the rate energy is incident from the light, the electron
reaches equilibrium. The total energy available to the electron is
the total contained in the periodic spherical wave that passes over
the electron during the time the electron takes to reach equilib-
rium. The equilibrium energy of the electron is thus determined by
the intensity of the incident light, although not all of the radiant
energy is necessarily transferred to the electron. Lorentz showed
that for visible wavelengths the minimum light intensity that could
deliver one quantum of energy to an electron at equilibrium is
almost io5 ergs/cm2-sec. Yet the photoeffect was known to occur
for visible light with an intensity at least as low as 2 ergs/cm2-sec.
In this way, Lorentz raised a corollary to the paradox of quality:
the dilemma of energy accumulation. Campbell had noted the
problem first, but when Lorentz explained the issue, it could no
longer be ignored.13 Removing the damping term so that the
electron could absorb an indefinite amount of energy, Lorentz
calculated the number of wavelengths of light that must combine
their effect before the required quantum can be absorbed. He
found that for visible light the number is
/ 2hc2 2X1010
Jl is the radius of the electron, X is the wavelength of the incident

light, and / is its intensity in ergs per square centimeter per second.
Bright sunlight, even if its full intensity were concentrated at a
single wavelength, would have to supply several million wave-
lengths. The low-intensity yellow light known to induce photo-
emission had to supplyfivebillion. This would take an "extraordi-
narily long time," he said - several millionths of a second.14
Lorentz did not give the derivation of equation 7.1, but he did
sketch enough of the assumptions involved so that one can recon-
13
Campbell, Theory (1907), 226ff.
14
Lorentz, PZ, 11 (1910), 1251.
15
struct it. Despite all the simplifying assumptions - he neglected
radiative losses and the intra-atomic interaction of electrons, as
well as the fact that one cannot expect to have monochromatic
wave trains of sufficient length - it is a revealing calculation. This
is not because the result requires so many oscillations but because
15
The equation of motion for a bound electron acted upon by a sinusoidal field
and not subject to radiative losses is
mx + fx = eAE cos cot (7.1 a)
At resonance, co2 = f/m; hence:
eA
x = - cos cot - co2x (7.1b)
m
Lorentz found that the maximum amplitude of oscillation grows in proportion
to t, so we may set x = Ct sin cot + B cos cot. Then
x = Ceo sin cot + Ccot cos cot Bco sin cot (7.1c)
and
x = coC cos cot + Ceo cos cot (tCco2 sin cot + Bco2 cos cot)
or
Jc = 2coC cos cot - co2x (7. Id)
We wish to calculate the total energy of the electron E = T + U. To accom-
plish this, note that dT/dt = d(mx2/2)/dt = mxx. From (7.1b):
AT1
= XeAE cos cot mco2xx = XeAE cos cot fxx (7. le)
But ~fxx = d(-fx2/i)/dt which is dU/dt, so

dE = dT+ dU = xeAE cos cot dt (7.1 f)
We now set the total energy equal to hv and solve for the time r required to
reach that threshold.
hv = eAE cos cot [(C Bco) sin cot + Ccot cos cot] dt (7. lg)
Jo
The first two terms drop out when integrated over a whole number of oscilla-
tions, and because cos2 a = (1 + cos 2a)/2,
r . /I +cos2a)A ,
hv= eAEcot( \dt
Again the last term drops to zero, leaving only

AEeCco T2 ._ ,, x
hv = -^- (7. In)
Substituting the value C = AEe/2com derived from (7.1b) and (7. Id):
T2 = 8hvrn/e2AE2 (7.1i)
it requires so few. If one compares the length of a wave train
needed to supply the same energy over just the cross section of the
electron, or indeed over the entire volume of an atom, the result at
visible wavelengths and intensity is some eight orders of magni-
tude greater than Lorentz' estimate!16
Lorentz gave an analytic solution to the problem of absorption
of electromagnetic energy; he did not interpret the result mechani-
cally, as others, notably Stark and Bragg, would do. Lorentz
treated the electron as a virtual point; its interaction with the field
occurs only through its charge and is therefore independent of its
"size." The electron radius ft enters equation 7.1 only as a conve-
nient combination of quantities, not as the controller of the frac-
tion of incident radiation absorbed. The effective area of the wave
front that contributes energy to the electron, according to Lorentz'
calculation, is inversely proportional to the square root of the light
intensity. Only for unrealistically great light intensity (on the order
of io18 ergs/cm2-sec at visible frequencies) does this area become
comparable to that defined by the maximum amplitude of the
electron just prior to its release from the atom.17 For lower intensi-
ties, which are known to be able to induce photoelectrons, the
To reconcile this result with Lorentz' formula (7.1), we need to express the
field strength AE in terms of Lorentz' /, which is an energy flux, / = AE2l^nc;
then substitute the classical radius of the electron Si = ie2l^m [M. Abraham,
AP, 10 (1902), 151]; and recognize that r = n/v = nk/c:
4hc2
which, except for a factor of V2, is equation 7.1.

16
The total energy through cross-sectional area Q per second is IQ, so the time
needed to attain energy hv is T' = hv/IQ. The ratio of the time calculated from
energy flux to that found by Lorentz is therefore T'/T = ^T^nhWfiU c/Q. For
the cross section of an atom, Q = io~16 cm2, and for a reasonable visible light
intensity of / = 500 ergs/cm2-sec, T'/T io8.
17
The maximum amplitude of the harmonically bound electron is reached one-
half cycle before its release when the potential enegy fx2/2 is equal to hv.
Thus, x2 - 2hc/\mw2 = h\/2Tr2mc. For visible light, x2 turns out to be 6 x
i o 1 6 cm2, the same order of magnitude as the cross-sectional area of the
atom. As Lenard had noted as early as 1902, a linearly oscillating electron
will have to obtain its entire T during that last half-cycle, for it will lose its
potential energy as it leaves the atom unless some special mechanism is called
into play to sever the binding force. The same problem afflicts an electron
rotating in a central potential formed by an effective positive charge 4- e.
For a circular orbit, U = 2T, so the total energy is on the order of mv2 =
m(2urv)2. When this equals hv, r2 for a visible frequency is also about 10 16
cm2.
electromagnetic cross section is considerably greater than any
mechanical interpretation would seem to allow.
Thus, when Lorentz concluded that the results predicted by
equation 7.1 were "not at all favorable for the old theory," he had
hardly the same discrepancy in mind that others would soon
mention. But Lorentz was primarily concerned with showing that
accumulation of energy over any finite time interval is incompati-
ble with the quantum transformation hypothesis for absorption of
light. How can an electron accept any of the energy from a classical
light wave before the full quantum is presented? Since most physi-
cists supported the classical model of light, this conflict cast doubt
on the new quantum theory just as its influence was beginning to
grow.18 Lorentz concluded, despite his objections to lightquanta,
that "without the assumption of lightquanta, these phenomena
raise the greatest difficulties."19
Lorentz did not extend the argument to the more intractable
case of x-ray ionization. Light possesses a repeating field oscilla-
tion, so that the accumulation of energy proceeds relatively swiftly.
X-rays each contribute only a single impulse. For the case of
x-rays, equation 7.1 gives a result equivalent to the total number of
impulses required. Moreover, for x-rays, the advantageous reso-
nance condition can no longer apply.
Stark liked Lorentz' argument. He devoted the second volume
of his Principles ofatomic dynamics entirely to "elementary radia-
tion." There he marshalled all arguments at his disposal against the
spreading-wave model of radiation. Thefirstobjection was that of
energy accumulation; he merely reprinted Lorentz' words verba-
tim, judging that his opponent had made the case for lightquanta
as well as he could.20
THE DECLINE OF THE TRIGGERING HYPOTHESIS

Philipp Lenard, the originator of the triggering hypothesis, had
attained substantial recognition by 1911. He had won the Nobel
18
In 1907 Einstein proposed a theory of specific heats based on the quantum
postulate that soon attracted the approbation of leading physicists and physical
chemists. This was the first subject, aside from the arcane problems of cavity
radiation, that drew favorable reactions to the quantum. Klein, Science, 148
(1965), 173-80. Kuhn, XIV Congres international d'histoire des sciences,
Proceedings, 1 (1974), 183-94.
19
Lorentz, PZ, 11 (1910), 1251.
20
Stark, Elementare Strahlung (1911), 267 - 70.
Prize in 1906 for his early investigation of cathode rays. The next
year, he was appointed Ordinarius and director of the new Radio-
logical Institute at Heidelberg. He developed a qualitative atomic
model during the first decade of the century based, in part, on
evidence for bound electron energies derived from experiments on
the emission of photoelectrons.21 In 1909 the triggering hypothesis
was described by Rudolf Landenburg as being among "the gener-
ally accepted truths" of physics.22 But at the same time that
Lorentz raised questions about the lightquantum, Lenard discov-
ered evidence that cast serious doubt on the triggering hypothesis.
In 1910 Lenard began an extensive series of experiments with his
skilled colleague, Carl Ramsauer, to study the action of ultraviolet
light on gases. They published five monographs, each directed to a
different aspect of the effect.23 Lenard was at first encouraged by
the "surprising" discovery that his source of ultraviolet light re-
leased a great number of electrons from the gas, although the
wavelengths he used were only minimally absorbed by the gas.24
This seemed to be even more evidence in favor of the triggering
mechanism: The incident light is not expected to lose much energy
in breaking the bond that holds the electron to the atom. In solid
bodies energy is lost through thermal effects, but in a gas these
losses should be negligible. In 1911 Lenard and Ramsauer studied
the phenomenon more closely because it "appeared to be of
particular interest from the energy point-of-view."25
The result of their careful analysis cast strong doubt on the
triggering hypothesis. They showed that most of the electrons
released in the photoeffect came from a small number of highly
absorbing atoms - impurities - mixed with their gas sample. The
conclusion was unavoidable. "For gases, as for solids and fluids,
the photoelectric effect is correlated not just with absorption [of
light] but with very strong (metallic) absorption . . . Our obser-
vations leave no support for the assumption that the energy of
[photoelectrons] originates in the inner atom; they indicate much
more that the incident light provides the energy."26 "A complete
theory of the energy transformation in the photoelectric effect is
21
Wheaton, HSPS, 9 (1978), 299-323.
22
R. Ladenburg, JRE, 6 (1909), 427.
23
All appeared in the Heidelberg Sb, I (1910), Abh. 28, 31, 32; 2 (1911), Abh.
16, 24.
24
Lenard and Ramsauer, Heidelberg Sb, 7 (1910), Abh. 31, pp. 2 iff.
25
Lenard and Ramsauer, ibid., Abh. 24, p. 1.
26
Ibid., 5-6.
not yet possible," they concluded.27 At the eighty-third Naturfor-
scherversammlung held in Karlsruhe in September 1911, Carl
Ramsauer announced publicly that the photoelectric effect could
no longer be thought of as exclusively a triggering action. "The
energy of the ejected electron," he said, "does not come from the
atom as originally assumed by Herr Lenard, but from the absorbed
light."28 If any sizable fraction of the energy of a photoelectron is
transferred from the light to the electron, either the paradoxes of
quantity and quality must be resolved or the mechanism by which
accumulation occurs must be found. In 1913 Lenard declared the
photoelectric effect a "difficulty" with no apparent solution.29
MAX PLANCK'S "SECOND THEORY"
Max Planck was strongly influenced by Lorentz' Wolfskehl argu-

ments, having heard them in correspondence with Lorentz before
the lectures in 1910.30 Early in 1911, Planck offered a new approach
to the problems of cavity radiation with which he bad grappled
since his derivation of the empirical black-body law in 1900. This,
his so-called second quantum theory, appears, in light of recent
research, to have been the first in which he accepted the discontin-
uity in energy transfer implied by the quantum transformation
hypothesis.31
At first Planck had attempted to quantize the absorption of
radiation by atoms, but the difficulties that Lorentz raised forced
him to reconsider that approach.32 The quantum hypothesis de-
mands that a minimum quantity of energy be supplied before
absorption can take place. Classical waves can deliver that energy
only a bit at a time. How can the first bit be absorbed at all? And if
it is absorbed, what happens if the incident radiation is cut off
immediately thereafter? Either that first subquantum of energy is
absorbed, violating the quantum hypothesis, or it is not. If it is not,
how can light ever be absorbed?
To avoid this dilemma, Planck proposed that the absorption of
radiant energy occurs continuously; only its subsequent emission
27
Ibid., 7.
28
Ramsauer, PZ, 12 (1911), 998.
29
Lenard, AP, 41 (1913), 82.
30
See note 6 in chapter 7.
31
Kuhn, Black-body theory (1978), chap. 10.
32
Planck, AP, 31 (1910), 758-68.
33
is quantized. At any moment, the electrons in irradiated atoms
possess among themselves a continuous distribution of energy. A
small fraction have energy just less than an integral multiple n of
the quantum hv (where v is the frequency of the oscillating elec-
tron). Given the small increment of energy needed to attain this
threshold by the incident light, emission of the entire energy nhv
can occur. The likelihood that the emission will or will not occur is
determined by the character of the oscillator; for any individual
oscillating electron, only a statistical probability for emission can
be determined. In this last suggestion, Planck may have been
influenced by the analogous recourse to a statistical theory of
radioactive decay. Meyer's statistical analysis of fluctuations in the
ionization produced by y-rays, as well as the seemingly random
emission of these rays, may have encouraged Planck to apply
similar concepts to the emission of light.
Planck mentioned the photoelectric effect as one of the cases to
which his new hypotheses might apply. It is not difficult to see how
this would occur, although Planck did not provide the details. It
was not until 1913 that he extended his theory of quantum emis-
sion to include the emission of electrons.34 His new hypothesis
circumvented the accumulation problem altogether. Electrons
absorb energy continuously, but some of them might be ready to
fly out of the atom with the stored quantity nhv in kinetic energy
almost instantaneously after the illumination begins. Further-
more, the velocities of emitted electrons should be independent of
both the intensity of the light and the temperature of the metal.
Light energy is not transformed into electron kinetic energy; the
latter already exists within the atom. Although heating the plate
will increase the energy of individual electrons, photoemission
occurs only at a fixed threshold, so that the resulting electron
energies remain unaffected.
Planck's theory was the quantum analog of the triggering hy-
pothesis. That was its problem. The theory suffered most of the
drawbacks of the triggering hypothesis. Only a few months after
Planck's announcement of it, Lenard and Ramsauer showed that
photoemission occurs only where strong absorption of light is also
found. According to Planck's interpretation, this should not be the
case. He thought that photoemission requires only a very small
33
Planck, VDpG, 75 (1911), 138-48; AP, 57(1912), 642-56.
34
Planck, Berlin Sb (1913), 350-63.
energy input from the incident light, just enough to raise the most
energetic electrons to the quantum threshold. But Lenard and
Ramsauer showed that strong absorption of light is correlated with
strong photoemission.
There was yet another difficulty that the triggering hypothesis,
whether the classical or the quantum version, had to confront. The
specific form of the photorelation between incident frequency and
electron velocity was still uncertain in 1911. But it was clear
enough that a continuous distribution of incident light frequency
sparked a continuous distribution of electron velocity. It could no
longer be claimed that the data supported resonance effects at
specific wavelengths. Planck's theory seemed to require that so-
dium atoms, for example, contain an electronic structure suffi-
ciently complex so that a full continuum of emission threshold
energies exists, with each value of the energy keyed to a particular
exciting frequency.
Atfirst,Planck denied the resonant nature of the interaction. He
said, "one cannot directly determine the frequency of the excited
oscillator from the wavelength of the incident ray."35 But at the
Solvay Congress later that year he retracted the claim, leaving
obscure just how the immense number of required electron fre-
quencies was to be reconciled with the limited number of electrons
in the atom.36 The difficulty would not soon go away. In 1911 it was
still possible to think that the unknown structure of the atom
might encompass enough electron frequencies to validate the
quantum triggering hypothesis. But in 1913 Niels Bohr offered a
new quantum theory of atomic structure that allowed only a small
number of electron frequencies. As the acceptance of the Bohr
theory grew, the unresolved problems of the photoeffect simulta-
neously became more intractable and less visible.
SOMMERFELD'S THEORY OF THE

PHOTOELECTRIC EFFECT
Arnold Sommerfeld was even more profoundly influenced by

Lorentz' discussion of the photoeffect than was Planck. He knew
that he could derive a certain degree of spatial localization for the
energy of y-rays, and even for that of x-rays. But for visible light his
weapons were of no avail. A different sort of explanation had to be
35
Planck, VDpG, 75 (1911), 147.
36
Planck, Solvay I, pp. 93-114, in particular 112.
found, and Lorentz had severely restricted the options. Applying
the quantum of action was hindered by the fact that both absorp-
tion and emission of energy are singular, nonperiodic events. The
quantum relation demands a frequency. To apply quantum theory
to discontinuous events, which have no frequency, Sommerfeld
had to generalize the quantum condition. When dealing with the
emission of y-rays earlier in the year, he had faced the same
problem and had conceded ignorance. He had simply used Wien's
experimental value for the ratio of x-ray to cathode-ray energy in
order to estimate the y-pulse width. Now, however, fresh from the
realization for y-rays that impulse duration can "take the place of
Planck's i/v," Sommerfeld went considerably further. Putting the
quantum transformation relation in its action form, E/v = h, and
applying his newly found insight from radiation theory, he arrived
at the general condition Ex = h, where r is the temporal duration
of the discontinuous event.
Sommerfeld, like Planck, formulated the quantum condition in
a way that sidestepped the problem of energy accumulation; he
made accumulation the very basis of the new principle.37 His
general expression depended only on the time interval over which
an event occurs. It could therefore be applied directly to nonperi-
odic events. As in Planck's theory, an atom was assumed to absorb
radiant energy continuously until a specific condition was at-
tained. Unlike Planck, Sommerfeld specified the threshold condi-
tion for emission in terms of a quantum of action, the product of
energy and time, rather than of energy itself. When the time
integral of the energy-determining Lagrangian taken over the
duration of the process attains the value h/m, an electron is
released with the full kinetic energy T of the system. In other
words, if U is the potential energy of the bound electron, the
condition for emission is fixed by the relation
r
Jo
(T- U)dt = h/2n
With his talented former assistant, Peter Debye, Sommerfeld

(7.2)
formulated a theory of the photoelectric effect based on this quan-

tum principle. In itsfinalbut still incomplete form, it was the most
successful treatment of the photoeffect that did not require the
37
Sommerfeld, PZ, 12 (1911), 1057-68. See also Nisio, XIVidme Congres interna-
tionale d'histoire des sciences, Proceedings, 2 (1975), 302-4.
assumption of the lightquantum. It was first presented at the
eighty-third Naturforscherversammlung in 19u.38 An electron is
to be thought of as bound by an elastic force fx, and loses no
energy by radiation. It is acted upon by an oscillating electric field
of maximum amplitude AE and angular frequency co. The equa-
tion of motion of the electron is therefore
mx +fx = eAE cos cot (7.3)
2
The instantaneous kinetic energy of the electron is T = mX /2\ its
potential energy Uisfx2/2. Sommerfeld inserted these relations in
the quantum condition (equation 7.2) and integrated by parts.
When the equation of motion, equation 7.3, is substituted, the
quantum condition becomes
mxx - x xAE cos cot dt = h/ln (7.4)
0 2 Jo
The left-hand side of equation 7.4 describes the variation in the
action of the electron as energy is pumped in by the incident wave.
It is a rapidly varying sinusoid with an envelope that grows steadily
but slowly from peak to peak. The quantum condition is achieved
when the amplitude of the oscillation finally reaches the fixed
magnitude h/m. This can occur only very close to one of the many
peaks of the growing sinusoid; otherwise the threshold would have
been reached on the preceding upswing. Consequently, at the
threshold the slope of the left-hand side of equation 7.4 must be
close to zero. But that expression is simply J(T U)dt, so that the
slope of the function is just the Lagrangian, T U. The quantum
condition is therefore reached when T U. In other words,
mX2/2 = fx2/i. But y/f/m is coOJ the resonant frequency of the
undamped electron. Thus, when the quantum condition is ful-
filled at time T, X = COOX, and the first term in equation 7.4 is just
T/coo. The second term vanishes entirely at resonance: As long as
the incident frequency equals the frequency of the electron, the
integral is taken over a whole number of wavelengths of a sinusoid
and consequently equals zero.
Thus, for the idealized case of monochromatic light in reso-
nance with an undamped electron, Sommerfeld's quantum condi-
tion (equation 7.2) reduces to
38
Ibid., 1064-5 a n d associated notes.
If we express the kinetic energy of the electron in terms of the

electric potential P that just suffices to arrest its motion, we obtain
Pe = hv0 (7.6)
This is Einstein's photoelectric effect law, equation 5.4, before the
work function p is taken into consideration.
The accumulation time is an essential part of Sommerfeld's
theory. The integral in equation 7.2 is always to be taken over the
accumulation interval T. Unlike Planck's theory, Sommerfeld's
required the energy of the liberated electron to come from the
incident light. He could not ignore the accumulation time. This
was especially true if the unavoidable radiation damping was
included and the incident light was more realistically granted some
range of frequency. Sommerfeld had clearly known of these prob-
lems before his talk in Karlsruhe. Bragg had written in response to
Sommerfeld's y-ray theory and cited Campbell's reservations
about the accumulation time for x-ray ionization. "The x-ray tube
must be at work for several days," he said, "before an electron in
an atom can acquire as much energy as a [released] electron is said
to possess."39 The problem is naturally more severe for aperiodic
x-ray pulses than it is for the rapid and regular succession of
impulses that constitute periodic light. But Lorentz had made it
quite clear that a difficulty lurked even for light. His relation
expressed in equation 7.1 can easily be modified to predict a delay
time for visible light equal to
For yellow light with an intensity of 2 ergs/cm2-sec, which is

known to stimulate photoelectrons, the accumulation time, ac-
cording to equation 7.7, is io~5 sec.40
39
W. H. Bragg to Sommerfeld, 7 July 1911. Stuewer, BJHS, 5 (1971), 273.
40
The simpleminded (and misleading) textbook calculation of the delay time, by
assuming that all the energy incident on the cross section of the atom accumu-
lates on an electron, produces the result r ~ 8 X io4 sec! See Chapter 7, notes
15 and 16. The analytical classical solution given by Lorentz is nowhere nearly
so far out of line with observations. Cf. Weidner and Sells, Elementary modern
physics (i960), 94, 123.
184 Seeking an electwdynamic solution, 1907-1912
In presenting his theory of the photoeffect, Sommerfeld did not
mention that the accumulation time constituted a potential prob-
lem. But Stark immediately objected. He asked if Sommerfeld had
calculated the magnitude of the accumulation time, claiming that
Lorentz' calculation showed that it "cannot be brought into agree-
ment with the observed conditions of [light] intensity and the
maximum kinetic energy of emitted electrons."41 Perhaps Som-
merfeld did not think that the problem was so great, at least not for
ordinary light. After all, one knew very little about the photoelec-
tric effect delay time experimentally; perhaps the io~5 sec predic-
tion was not so far from the truth. Sommerfeld responded that the
"difficulty of the time interval" might be alleviated when the
theory was extended to nonmonochromatic light.42 Stark per-
sisted, saying that many of the most accurate photoeffect measure-
ments had been done with sharply defined spectral lines. Two days
later, the accumulation time took on even greater significance
when Carl Ramsauer delivered the news that the triggering hy-
pothesis was no longer viable.
It was true that not much was known about the photoelectric
delay time in 1911. Some knew that Alexandr Stoletov had found
in 1889 that the delay was not longer than io~3 sec.43 Wien had
given an estimate of 0.01 sec in 1905 for the accumulation of x-ray
pulses, and that absurdly long result had induced him to accept the
triggering hypothesis.44 Einstein, recognizing that a critical point
had been raised, judged that an experimental investigation of
photoeffect delay times would be "of the greatest interest."45 Erich
Marx took up Einstein's challenge, but it was almost two years
before his results were known. In the meantime, Sommerfeld
continued making attempts to solve the photoeffect conundrum.
THE SOLVAY CONGRESS
In October 1911 the First Solvay Congress was held in Brussels. It

was here that the quantum principle was recognized officially to be
a significant and unavoidable aspect of physical theory. Although
41
Stark's comment in the discussion following Sommerfeld's presentation, PZ,
12 (1911), 1068-9.
42
Ibid.
43
Stoletov, Physikalische Revue, 1 (1892), 723-80. See Wheaton, Photoelectric
effect (1971), chap. 5.
44
Wien,^P, 7^(1905), 1005.
45
Einstein's contribution to the discussion, PZ, 12 (1911), 1068-9.
the conference was titled "The Theory of Radiation and Quanta,"
there was little serious discussion of the lightquantum hypoth-
esis.46 Virtually all those assembled felt that Einstein's proposal
was an unwarranted rejection of the highly verified wave theory of
light. Stark was not invited to participate. Einstein was, but was
asked to speak on his quantum theory of specific heats.
Sommerfeld and Planck each presented his new version of the
quantum regulation. After Planck spoke, Wien asked if the long
accumulation time really constituted an insurmountable problem
for what he called Planck's "first theory."47 Planck answered that it
did, not so much because it disagreed with experiment but because
the very process of accumulation of energy seemed to be incom-
patible with quantum absorption.48 Sommerfeld repeated his ar-
guments on the localized energy of x-rays and y-rays. The de-
manding problems of absorption were submerged in his action
integral representation of the quantum principle.
At the Solvay Congress, Sommerfeld expanded the treatment of
the photoeffect that he had given at the earlier Naturforscherver-
sammlung. The spatial concentration of light energy implied by
the photoeffect he called one of the "greatest difficulties" then
facing electrodynamics.49 He asserted that "radiation theory and
electromagnetic theory complement, rather than exclude, one
another,"50 but despite his best efforts, the photoeffect remained a
difficulty. He was able to treat off-resonance with the help of Peter
Debye; the result was "an analog to Stokes' law" of fluorescence.51
They showed that only waves with a frequency greater than the
resonant frequency of the electron are effective in passing on
energy to the electron. To solve equation 7.4 when co = coo, they
defined a figure of merit e = (co &>O)T, where T is the accumula-
tion time for the resonance case. They then showed that the
quantum condition, equation 7.4, reformulated in terms of e,
became
2maJ h
( 1 + e) - (cos 6 + sin c) = [(co2 - co02)/co0eAE]2 (7.8)
2n
46
La theorie du rayonnement et les quanta, Langevin and M. de Broglie, eds.
(Paris, 1912), henceforth Solvay I.
47
Wien's comments followed Planck's presentation, Solvay /, p. 126.
48
Planck, Solvay I, pp. 126-7.
49
Sommerfeld, Solvay I, p. 344.
50
Ibid., 333.
51
Sommerfeld, PZ, 72(1911), 1065.
/
s
#*=
S
>
i y /
i/
/
J
// 1 / +-^
-*/y-
\
'

\
\
\

/ air
2
\\
\
/
/
r x
Figure 7.1. Graph used by Debye and Sommerfeld to derive Stokes' law
for the photoelectric effect. [AP, 41 (1913), 890.]
The left-hand side of equation 7.8 is the superposition of a sinusoi-

dal oscillation, cos e + sin e = 2 sin (e + 71/4), on the linear
relation 1 + e. Plotted as a function of e, it is the rising sinusoid
traced by the solid line in Figure 7.1. The right-hand expression of
equation 7.8 is a constant for fixed OJ, CDO, andAE. The solutions of
equation 7.8 therefore lie at the intersection of the rising sinusoid
with the horizontal line that represents the value of the right half
of the equation. If there is more than one point of intersection,
the one for the smallest value of e has the shortest accumulation
time r.
The right-hand side of equation 7.8 is always positive. Thus, the
only conceivable solutions for negative values of 6 are restricted to
the angular range e= n/2 to 0. A negative value for e means that
the incident frequency is less than the resonant frequency. The
anti-Stokes case can occur only within an extremely small fraction,
some icr11, of the resonant frequency. The Stokes solution for
positive values of e can occur any number of times, for any co
greater than a>0. As in any resonant interaction, the efficiency of
the energy transfer drops rapidly as e increases. But beyond this
neat trick, Sommerfeld and Debye could not go. They used what
they called their "ersatz" model to derive "a few of the most
general consequences of photoelectric behavior without coming
into conflict with electrodynamics."52
Even in their final study of the photoeffect, published in 1913,
Sommerfeld and Debye could not solve half of the problems that
remained.53 They attempted to extend the analysis from mono-
chromatic to "natural" radiation, and this required an appeal to
statistical arguments. But despite their effort (one "somewhat
tiresome calculation" of an integral extended to nine printed
pages) they made no significant progress toward a resolution of the
paradoxes. They did, however, calculate an expression for the
expected delay time for monochromatic light. The value was
"unmeasurably short" but was little different from those derived
from Lorentz' equation 7.1. For the green line at 5,460 A in the
mercury spectrum, the expected delay was 8.7 X io~5 sec, an in-
terval too short to ensure that the light frequency remains well
defined at all.54
One month after their prediction, thefirstauthoritative determi-
nation of a photoeffect delay time was announced.55 Erich Marx,
who had earlier measured the velocity of x-rays, accepted the
challenge proposed by Einstein in Karlsruhe. The experiments
were done in Leipzig with the professor at the Technische Staats-
lehranstalt, Karl Lichtenecker.56 They used a rotating mirror to
limit the duration of the exposure of a sensitive potassium photo-
cell to light from various sources. Photoemission could be mea-
sured for exceedingly low intensities, down to 6 X io~8 erg/cm2-
sec. For the 5,900-A mercury line, even at the subvisible intensity
of 0.5 erg/cm2-sec, they could find no delay in the onset of photo-
emission right down to the limit of their equipment's temporal
resolution, 1.5 X io~7 sec. This was less than 1 percent of the value
predicted by Sommerfeld and Lorentz as the minimum time
needed for accumulation from classical waves. Marx pointed out
that the small value could not be explained even if light energy was
restricted to a diverging cone.57 The divergence too quickly re-
duced the energy density of the rays. Marx referred explicitly only
52
Debye and Sommerfeld, AP, 4V (1913), 927, 930.
53
Ibid.
54
The data came from Buisson and Fabry, CR, 152 (1911), 1838-41.
55
There had been several attempts. The most significant were those by Elster and
Geitel, PZ, 13 (1912), 468-76; Ioffe, Munchen Sb, 43 (1913), 19-37; and E.
Meyer and Gerlach, ASPN Geneve, 35 (1913), 398-400.
56
Marx and Lichtenecker, AP, 41 (1913), 124-60.
57
Marx, AP, 41 (1913), 161-90.
to Thomson's unit hypothesis, but the implication was clear that
even Sommerfeld's x-ray and y-ray models were thus rendered
impotent.
Sommerfeld's theory failed in other respects too. It was neces-
sary to leave out radiation damping entirely from the calculation.
If it was included, the photoeffect could apparently not occur for
any but abnormally high light intensities. Moreover, Sommerfeld
and Debye had emphasized in 1911 that theirs was a true reso-
nance model of the photoeffect, in contrast to Einstein's treat-
ment.58 And Sommerfeld persisted in the belief that experiments
would confirm that the photoresponse shows peaks at wavelengths
characteristic of the irradiated atoms. In 1913 he called "con-
trived" the idea that a "great number of independent resonators of
all possible [frequencies] are present in the atom."59
A certain ambiguity surrounded the question of whether reso-
nant peaks appear in the photoresponse from various materials.
Robert Pohl and Peter Pringsheim had found in 1910 that two
kinds of photoresponse are superimposed in most experiments.60
The first is the normal effect, the second a "selective" effect
controlled largely by the state of polarization of the light. Before
this discovery, it was not realized that carelessness about the
incident light - whether it was reflected and if so at what angle -
could produce peaks in the measured photoresponse. Afterward,
the evidence from the normal effect, although still uncertain,
began to converge. Sommerfeld's hopes notwithstanding, it ap-
peared that the maximum electron velocity increases monotoni-
cally in response to the frequency.61 The peaks seemed to be an
artifact of the selective photoeffect.
In the meantime, the chief impediment to accurate studies over
a large wavelength range in the normal photoeffect - oxidation
impurities on the photocathode - had been overcome. Karl
Compton and Owen Richardson did experiments in 1912 on
cathodes scraped clean and maintained in a vacuum.62 Although
the specific relation of stopping potential to light frequency was
58
Sommerfeld, Solvay I, pp. 355-6.
59
Debye and Sommerfeld, AP, 41 (1913), 930.
60
Pohl and Pringsheim, VDpG, 12 (1910), 215-28, 249-60, 682-96, 697-710;
PM, 27(1910), 155-61.
61
Wheaton, Photoelectric effect (1971), chap. 13.
62
O. Richardson and K. Compton, PM, 24 (1912), 575-94. Hughes, PTRS,
212A (1913), 209-26.
still uncertain in 1913, the new evidence made it even clearer that
the maximum photoelectron velocity grows monotonically with
frequency.63 If this was a resonance effect, as Sommerfeld claimed,
one had to explain how an essentially infinite number of resonant
frequencies are represented in the atom. Sommerfeld had no
answer. He held to the faint hope that the evidence was "afflicted
with great uncertainty."64 Thereafter Sommerfeld viewed the pho-
toeffect as a "real difficulty," but he did not publish on the subject
again.65
THE EQUIVALENCE OF X-RAYS AND LIGHT
Two formerly separate strains of thought became firmly united in

the period 1911 -13: the theory of light and the theory of x-rays.
The net result was to highlight the paradoxes of radiation on a
general level. Before light and x-rays were recognized as common
forms of radiation, localized properties could be relegated primar-
ily to the "new" radiations: x-rays and y-rays. About the nature of
these rays there was still some uncertainty. But a constellation of
events, beginning in 1911, made it clear that light is implicated too.
First came Sommerfeld's y-ray study, written in response to the
criticisms of Stark and Meyer, and an attempt at a quantum theory
of nonperiodic events. Then Planck offered his quantum theory of
light emission, his first to employ the quantum transformation
hypothesis. Soon Sommerfeld selected the photoeffect for applica-
tion of his own generalized quantum principle; it provided the
frequency condition that the quantum formalism seemed to re-
quire. The reaction to Sommerfeld's theory at the Naturforscher-
versammlung in September led to the first detailed experimental
test of photoelectric effect delay times. And at that same meeting,
Carl Ramsauer announced that the triggering hypothesis had been
virtually abandoned by its creator, Philipp Lenard.
Sommerfeld had already recognized that the characteristic
x-rays form a direct extension of the optical frequencies. He
concluded that the Eigenschwingungsdauern, or oscillation pe-
riods of the characteristic x-rays, could be substituted in the quan-
tum transformation relation "to determine [x-ray] energy, from
63
Pohl and Pringsheim, VDpG, 15 (1913), 637-44.
64
Debye and Sommerfeld, AP, ^7 (1913), 928.
65
Ibid., 873.
the shortest right down to the visible and infra-red."66 Sommerfeld
provided a single quantum condition that applied equally to the
periodic and aperiodic x-rays.
In England, even Barkla had begun to refer to the characteristic
x-rays asfluorescentrays, implying that they were related to visible
light.67 C. T. R. Wilson presented his cloud-chamber photograph,
Figure 6.9, showing dramatically that x-rays produce only a small
number of electrons in their flight through a gas. And Bragg,
frustrated with the inconsistencies that he had puzzled over for five
years, and now nonplussed that ordinary light had to be included
among the forms of radiation that act so peculiarly, began to call
publicly for a dualistic conception of all forms of radiation.
On the heels of these events of 1911 came the startling news from
Germany that x-rays, like light, can interfere. Our next chapter is
devoted to an analysis of the impact of this influential discovery.
But thefindinghad a discouraging effect on studies of the nature of
radiation. Most physicists took the result as proof that x-rays are
periodic waves like light. This was the straightforward conclusion
to be drawn from seemingly uncontestable evidence. An alterna-
tive was available only to the relatively small number of physicists
conversant with the evidence for paradox in the behavior of x-rays,
and it was not appealing to contemplate extending those para-
doxes to the venerable wave theory of light. There was very little
difference in the response between British and German investiga-
tors or between mathematical and empirical physicists to the new
x-ray evidence. X-ray interference did not create a crisis in radia-
tion theory; what crisis there was existed already. For most physi-
cists, the discovery reduced much of the uncertainty about the
nature of x-radiation, and so discouraged continued research on
the energy structure of free radiation.
The discovery of x-ray interference had a temporary stimulating
effect on empirical studies of absorption of all forms of radiation,
as shown by Figure 7.2. But that interest soon peaked and declined
steadily in the middle of the decade.68 Of course, much of the
decline was due to the disruptive effect of World War I on research,
especially in Britain. But in addition, the very fecundity of crystal
66
Sommerfeld, Munchen Sb, 41 (1911), 43.
67
Barkla, PM, 22 (1911), 396-412. He also spoke of x-ray "line spectra."
68
Figures are taken from the Fortschritte der Physik, abstracting journal of the
Deutsche physikalische Gesellschaft.
40 T
Experimental studies on absorption of:

30 --
light
1909 10 11 12 13 14 15 17 18
Figure 7.2. Number of experimental studies on absorption of radiation,

1909-1918.
diffraction attracted some of the research that might have been

directed to the nature of x-rays. W. H. Bragg, to cite the most
significant example, devoted virtually all of his research after 1914
to the analysis of crystals based on an assumption of wavelike
x-rays. At the same time, the new methods of x-ray spectroscopy
introduced by crystal diffraction encouraged increasing study of
the x-ray emission spectra for the analysis of atomic structure.
Since neither of these otherwise important research fields contrib-
uted directly to an understanding of the nature of radiation or of its
quantum absorption, they do not appear in Figure 7.2.
DECLINE OF THEORETICAL INTEREST IN

FREE RADIATION.
The attempts of mathematically adept physicists to grapple with
the problems of radiation absorption were largely unsuccessful,
and all were short-lived. Figure 7.3 traces the number of mathe-
matical studies on the critical problems during the decade from
1909 to 1918.69 The absolute number of studies is quite small, but
69
The graph includes all articles concerned with the "theory" or the "nature" of
one of the three forms of radiation. It also includes all articles known to treat
mathematically the absorption or transformation of radiant energy. The
counts come from four sections of the Fortschritte: Electricity and Magnetism
10B Becquerelstrahlen, and 10C Rontgenstrahlen, also 14 Lichtelektrizitdt and
Optics 1 Allgemeines.
Mathematical studies on absorption of:
light
1909 10 11 13 14 15 16 17 18
Figure 7.3. Number of mathematical studies on absorption of radiation,

1909-1918.
the marked decrease in interest, which parallels the decline in

empirical studies, is clear. Note, in particular, the precipitate drop
in the sum of the curves after 1913. There were a few later attempts
to solve the riddle of the photoeffect, but the problems seemed to
be growing and the momentum established in 1911 -12 had been
lost.
Even at the peak of interest in absorption phenomena, many of
the studies were directed more at resolving questions of atomic
structure than at resolving the paradoxes of radiation. Karl Her-
mann at the Technische Hochschule in Berlin had argued in 1912
that no resonance mechanism could explain the release of photo-
electrons bound in elliptical orbits. The frequency of revolution
changes as energy is added, and, according to Hermann, this
would prevent any considerable accumulation of energy.70 J. J.
Thomson considered the photoelectric effect in 1913 when he
proposed a contrived atom model that, like the one he had pro-
posed in 1910, offered a continuum of electron frequencies and
energies to a passing light wave.71
70
K. Hermann, VDpG, 14 (1912), 936-45.
71
J. J. Thomson, PM, 26 (1913), 792-9.
Interest in questions of atomic structure increased sharply after
1913 when Niels Bohr published his famous theory of atoms.72
This was the major reason internal to physics for the concomitant
decline in research on radiation. The disruption of all research in
Europe during World War I played a role, but even so, considera-
tion of radiation theory dropped far behind the treatment of
atomic theory. Bohr's theory relied on the anticlassical assumption
of discontinuity in both emission and absorption of radiant en-
ergy; it bypassed the problem of energy accumulation by postulat-
ing instantaneous absorption. It was also as surprisingly successful
in its prediction of spectral frequencies and chemical properties as
it was unorthodox. The theory was quickly taken very seriously,
and it set rigid constraints on the number of possible electron
frequencies in the atom as well as on the relationship between
electron frequency and orbital velocity. The triggering hypothesis,
indeed any resonance model of the photoelectric effect, was effec-
tively rendered impossible.
For almost ten years after the adoption of the Bohr theory, most
studies of the photoelectric effect were intended to clarify not the
nature of light but the structure of atoms. In 1916, for example,
Paul Epstein used the photoeffect to test his hypothesis that hyper-
bolic orbits followed by ejected photoelectrons are compatible
with the Bohr-Sommerfeld model of the hydrogen atom.73 He
found that such orbits were possible but that the atom would have
to respond selectively to particular incident frequencies, and by
1916 it was quite clear that no resonant peaks occur. Epstein did
not comment on the discrepancy and did not consider the problem
of how the incident light can concentrate its entire quantum effect
on a single electron. Owen Richardson developed a theory of the
photoelectric effect in 1914 that treated only macroscopic quanti-
ties and was entirely independent of assumptions about light. The
effect was considered to be analogous to the evaporation of mole-
cules from a liquid; the quantum condition became a sort of latent
heat of vaporization.74
A few younger physicists outside the German and British insti-
tutes-Abram Ioffe, Jun Ishiwara, and Mieczesfaw Wolfke- per-
sisted with Einstein in promoting the advantages of the lightquan-
72
Bohr, PM, 26 (1913), 1-25. See Heilbron and Kuhn, HSPS, 1 (1969), 211 -90.
73
Epstein, AP, 50 (1916), 815-40; PZ, 77(1916), 313-16.
74
O. Richardson, PM, 27(1914), 476-88.
turn. But this fell on virtually deaf ears.75 Influential mathematical
physicists besides Einstein avoided the issues almost to a man.
Sommerfeld turned to the elucidation of his phase-integral quan-
tum conditions for the Bohr atom, and to detailed examination of
optical and x-ray emission spectra. Lorentz undertook no study of
the nature of free radiation, except for one study of x-ray impulses
to which we shall return in the next chapter. Debye moved steadily
toward integration of the new quantum theory of the atom with
problems on the larger scale of molecular forces. Planck continued
in his attempts to reconcile his thermodynamic interpretation of
radiation with quantum theory. Thomson concentrated on new
experimental studies of positively charged ions. He was heavily
involved in war work after 1914, gave up his directorship of the
Cavendish Laboratory in 1918, and effectively stopped publishing
research altogether until the early 1920s.
Two who expressed the most continuing interest before 1920 in
the structure of radiation were Wilhelm Wien and Max Laue. A
year after receiving the Nobel Prize for his work on cavity radia-
tion, Wien gave a series of lectures at Columbia University on
"New problems of theoretical physics." He pointed again to the
great difficulties that faced the quantum theory of light absorption
and showed that Einstein's analysis of fluctuations did not neces-
sarily demand recourse to the lightquantum.76 But he soon turned
almost exclusively to theoretical analysis of emission spectra from
positive ions. Laue, as we shall see, was deeply involved with
theoretical analysis of x-ray interference. In the second decade of
the twentieth century he worked on a thermodynamic treatment
of emission and absorption of light and, although unsympathetic
to Einstein's lightquantum well into the 1920s, he was unusual in
his belief that a statistical solution could resolve even interference
phenomena.77
75
"That [Einstein] may occasionally have missed the mark in his speculations, as
for example with his hypothesis of lightquanta, ought not be held too much
against him, for it is impossible to introduce new ideas, even in the exact
sciences, without taking risk." These were the words of Planck, Rubens,
Nernst and Warburg to the Kultusminister in their attempt to have the
non-German Einstein appointed to the Kaiser Wilhelm Gesellschaft and the
Prussian Academy of Sciences in 1913. Quoted in Kahan, Archives Interna-
tionales d'histoire des sciences, 15 (1962), 337-42.
76
Wien, Neuere Prohleme (1913); WAS Journal, 5(1913), 273 -84; AP, 46 (1915),
749-52.
77
Laue, VdPG, 79(1917), 19-21.
Einstein himself was the most conspicuous exception among
mathematical physicists; in 1916 he wrote a brilliant restatement of
the problems of free radiation and offered a refined version of his
statistical solution.78 But for most physicists, the problems were
too demanding and the temptation of the successes of the Bohr
theory too great. No resolution to the paradoxes of radiation was
found by physicists who sought a consistent mathematical repre-
sentation of nature. As we shall see in the next part of this study,
the important steps toward solutions in the decade from 1911 to
1921 were taken not by mathematical theorists but by empiricists
who plumbed the manifold consequences of x-ray diffraction by
crystals.
78
Einstein, Zurich Mitteilungen, 18 (1916), 47-62; PZ, 7^(1917), 121-8.
PART IV
Interference of x-rays and the corroboration
of paradox 1912 -1922
8
Origins of x-ray spectroscopy
[X-rays] are a kind of wave with properties no wave has

any business to have.1
In the spring of 1912 two research assistants in Munich directed an

x-ray beam through a crystal and found that the beam was re-
formed into a well-defined interference pattern. The property most
characteristic of periodic waves - their ability to interfere - is
shared by the x-rays. Max Laue, the man chiefly responsible for the
discovery, thought he had found proof that characteristic second-
ary x-rays from the crystal are periodic waves. But H. A. Lorentz
quickly pointed out that impulses should interfere too. He showed,
in a tour de force argument, that the accepted square pulse is an
impossible representation of x-rays. William Henry Bragg and his
son concluded that the interference maxima could, indeed, be due
to irregular x-ray pulses. But, as such, x-ray impulses were not
different from ordinary white light. They supported this claim with
a new technique of crystal analysis fully analogous to ordinary
optical spectroscopy.
Crystal diffraction provided a new tool for the analysis of x-rays.
Pushed furthest by Henry Moseley and Charles Darwin, the tech-
nique soon showed that some x-rays comprise periodic wave trains
of great length. The extremely sharp angular resolution of ob-
served x-ray interference maxima indicated beyond doubt that
x-rays are no different, except in frequency, from ordinary light.
Rutherford soon extended the technique to the y-rays. Not only
could one isolate characteristic y-rays, he believed, one could
show, with some effort, that they interfere too.
The successful integration of the new spectroscopy with the
Bohr atom came, as had x-ray diffraction, from Sommerfeld's
1
Moseley to his mother, 21 January 1913. Heilbron, Moseley, (1974), 199-200.
200 X-ray interference, 1912-1922
Munich. Walter Kossel suggested in 1915 that x-rays are the
emission product of the redistribution of electrons in the close-
lying orbits of the atom. This gave theoretical justification to what
empiricists had known since 1913: x-rays are no more or less than
high-frequency light. Rutherford analyzed the y-ray results in the
context of his recently formulated nuclear model of the atom.
Certain groupings of electrons in the superstructure of the atom
are responsible, he suggested incorrectly, for the characteristic
high-frequency vibrations that produce monochromatic y-rays.
Although the final justification of y-rays as the product of transi-
tions within the nucleus did not come until 1922, the extension of
the electromagnetic spectrum to include x-rays and y-rays was
essentially complete by the start of the Great War.
X-RAYS CAN INTERFERE

Max Laue, who replaced Peter Debye as Sommerfeld's theoretical
assistant in 1912, thought that the periodic secondary x-rays char-
acteristic of their scattering element might show interference ef-
fects if a suitable diffracting device could be found. It was his
inspiration to suggest that the regularly spaced atoms presumed to
constitute crystals would yield their characteristic x-rays from a
regular array of points; this might allow observation of interfer-
ence. The experimental test was performed by Walther Friedrich
and Paul Knipping, both working at Sommerfeld's Institute for
Theoretical Physics in Munich. The work was apparently done
over Sommerfeld's objections. The resulting photographs, one of
which is reproduced as Figure 8.1, vindicated Laue, and the suc-
cess was announced to the world in June 1912.2
Sommerfeld's skepticism is easily understood. As Forman has
shown, Sommerfeld was not likely concerned that thermal agita-
tion of the lattice points would disrupt possible interference.3 He
had found long before that x-ray impulses seem not to interfere; no
fringes appear at an x-ray shadow edge. Sommerfeld had other
serious objections. Only a month before Laue's discovery, he had
published a detailed study of x-ray diffraction.4 Sommerfeld re-
2
Friedrich, Knipping, and Laue, Miinchen Sb, 42 (1912), 303-22. Reprinted
with additional notes in AP, 41 (1913), 971-88.
3
Forman, AHES, 6 (1969), 38-71.
4
Sommerfeld, AP, 38 (1912), 473-506.
Origins ofx-ray spectroscopy 201
Fig. S
+ + '
CO " +

+ & oo
Figure 8.1. One of the original x-ray interference photographs taken in

1912 by Friedrich and Knipping at Laue's suggestion. The crystal is zinc
sulfide; the x-ray beam enters parallel to the optic axis. At the bottom
are Laue's predicted positions for interference spots from five homo-
geneous x-rays. His best estimate of the crystal spacing a was
3.38 x 10~8 cm; the X range was 1.3-5.2 x 10~9 cm. [Friedrich, Knip-
ping, and Laue, Munchen Sb, 42 (1912), Tafel II.]
sponded shortly after the Solvay Congress to Robert Pohl and

Bernhard Walter, who had been unable to find evidence of diffrac-
tion in an x-ray photograph taken under conditions similar to
those originally employed by Wind and Haga.5 Peter Koch devel-
oped an optical photometer to analyze the photograph quantita-
tively for Sommerfeld.6 The resulting intensity contours allowed
5
Walter and Pohl, AP, 25 (1908), 715-24; AP, 29 (1909), 331-54.
6
Koch,^P, 38 (1912), 507-22.
Sommerfeld to reach the gratifying conclusion that the new photo-
graph, like the old, gave evidence of diffraction of aperiodic im-
pulses with a width of about 4 X io~9 cm. The analysis showed no
sign of fringes, and Sommerfeld concluded that the "more peri-
odic" characteristic x-rays, by then known to accompany the
impulses, played no part in diffraction from a slit.7 If the character-
istic x-rays from the anticathode are unable to pass through the
2-jum slit, how can they pass through the spaces between crystal
atoms, some io4 times narrower? Sommerfeld did admit that the
data favoring diffraction were hard won; clearer evidence would,
he said, "form a sort of capstone to the [impulse] theory and
definitely exclude any corpuscular theory of x-rays."8
It was thus well before, not after, the discovery of x-ray interfer-
ence that Sommerfeld distinguished impulse x-rays from the more
periodic secondary x-rays in his writings. In his 1909 response to
Stark and his 1911 study of y-rays, he had separated the two forms
of x-rays; at first, he thought that the quantum theory could not
apply to the impulse part. But it might yet have been that the
homogeneous x-rays were no more than strong Fourier compo-
nents of the impulses. Each element might scatter x-ray impulses
and alter their shape in a characteristic way. In light of his 1912
diffraction analysis, however, the fluorescent x-rays seemed to be
more than especially strong Fourier components because they are
apparently excluded from regions accessible to impulses. To make
the emancipation explicit, Sommerfeld assigned a new name to
what had, until then, been only the Bremsanteil in x-rays. The
inhomogeneous and unpolarized Bremsstrahlung thus assumed a
status separate from the homogeneous, partially polarized, and
temporally extended fluorescent x-rays.
It was only the homogeneous x-rays that concerned Laue. Fol-
lowing Sommerfeld, he thought that Bremsstrahlung impulses of
varying widths, diffracted from a confused set of crystal planes
oriented in various directions, could produce only a diffuse black-
ening of the photo plate. Because of this, Laue expected to see only
the superimposed effect of all secondary rays. The copper sulfate
7
Sommerfeld, AP, 38 (1912), 483.
8
Ibid., 474. Pohl, Physik der Rontgenstrahlen (1912), appeared in the summer of
1912. Before the final proofs were in, Pohl added a short section at the end on
the new evidence for x-ray interference. The rest of the book is based on the
impulse hypothesis.
Origins of x-ray spectroscopy 203
crystal used in the first experiments was chosen carefully to maxi-
mize the homogeneity of characteristic x-rays, but it was of suffi-
ciently low density so as not to absorb x-rays inordinately.9 For a
full year, Laue persisted in his belief that only fluorescent x-rays
exhibit interference. What we now take as the distinguishing fea-
ture of x-ray diffraction - that the crystal imposes its own perio-
dicity on the incident radiation - was, at first, unrecognized by
Laue.
Problems quickly arose for his interpretation. Laue could calcu-
late, with some ambiguity, how many specific "wavelengths" were
responsible for the interference spots. He listed them beneath the
photograph in Figure 8.1. When he encountered difficulty in
attributing these solely to the fluorescent rays from the crystal, he
raised the possibility that some of them came from the anticathode
in the x-ray tube. If Sommerfeld had been wrong about the exis-
tence of interference, perhaps he was also wrong about the absence
of fluorescent rays from the anticathode. Laue concluded: "that
the radiation emitted by the crystal has a wave character is proven
by the sharpness of the intensity maxima." But it was still unclear
"whether the periodic radiation arises first in the crystal by fluore-
scence or whether it already is present alongside the impulses in the
primary radiation and is only separated by the crystal."10
In his analysis of the interference photographs the following
month, Laue concluded that the five responsible wavelengths
ranged from 1.27 to 4.83 X io~9 cm.11 But the method used to
obtain these values did not determine them uniquely. Some spots
that should have appeared were conspicuously absent. Laue also
had difficulty explaining why an intense interference pattern was
cast by diamond, when carbon, its only elemental constituent, had
been shown by Barkla to be a weak emitter of fluorescent x-rays.
Even more perplexing was the evidence presented by Friedrich and
Knipping that the absorption coefficient of some of the interfering
rays from zinc was markedly lower than that found for the charac-
teristic x-rays from zinc.12 A year later, in a Zusatz added to the
9
Friedrich, Knipping, and Laue, Miinchen Sb, 42 (1912), 314. Forman, AHES,
6(1969X63.
10
Friedrich, Knipping, and Laue, Miinchen Sb, 42 (1912), 310-11.
11
Laue, Miinchen Sb, 42 (1912), 363-73.
12
Friedrich, Knipping, and Laue, Miinchen Sb, 42 (1912), 331-2. They obtained
3.84 cm"1 in aluminum. Barkla and Sadler, PM, 14 (1907), 413, had measured
// for zinc x-rays passed through aluminum as 96.0 cm"1.
reprinting of thefirstpaper, Laue still concluded that the role of the
crystal is only to select certain well-defined wavelengths from the
primary incident beam.13
IMPULSES CAN INTERFERE TOO

The discovery of x-ray interference was not without influence on
the theory of the Bremsstrahlung impulses. A few months after
x-ray interference had been announced, H. A. Lorentz, who had
never written on x-rays before, recognized that, in principle, im-
pulses should show interference too. In December 1912 he offered
an elegant argument that cast serious doubt on the accuracy of the
square pulse as an approximation for the Bremsstrahlung The
intensity of scattered impulses at point B due to a primary beam
striking matter at point A consists of more than simply the second-
ary beam directed along line AB. According to Huygens' principle,
the effect at B must take into account the elementary spherical
impulses arising from every point that the scattered pulse passes.
In the case of a rapidly oscillating vibration such as light, these
elementary wavelets largely cancel out. But Lorentz realized that
Sommerfeld's singly directed pulse shape did not allow this can-
cellation. A singly directed pulse is one in which the displacement
of the electric field is in one direction only, transverse to the
direction of propagation of the pulse, as shown in Figure 8.2.
Lorentz calculated what the net effect of the scattered square
impulse would be at point B.
He invoked the well-known rule that the sum of the distances
from each focus to a point on an ellipse is constant. One must
imagine space broken up into a series of nested ellipses of revolu-
tion around A and B as foci. See Figure 8.3, where one such shell is
indicated in cross section. The secondary impulse from A will pass
through different parts of a given shell at different instants, but the
Huygens wavelets from each differential element on a given ellipse
all converge at the same instant on point B. Therefore, if the
impulses are singly directed, a progressively weaker constructive
maximum at B is contributed by each of an essentially infinite
number of shells. Huygens' principle applies even without
13
Laue, AP, 41 (1913), 989-1002.
14
Lorentz, Amsterdam Verslag, 21 (1913), 911-23.
singly-directed
doubly-directed
Figure 8.2. Comparison of singly and doubly directed impulses.
primary beam
Figure 8.3. Lorentz' ellipsoid of revolution used to show that singly

directed x-ray impulses cannot be valid.
material scattering centers in the surrounding space, but the actual

spherical impulses that are induced at each surrounding gas mole-
cule show the same interference effects at B. A conservative esti-
mate was that the intensity due to the constructive interference of
secondary impulses radiated from ambient air would exceed that
amount radiated directly through B by a factor of 95! Barkla's
repeated measurements since 1904 on the intensity of scattered
x-rays showed no hint of this massive effect.
This was not the only problem. Lorentz showed that in a gas the
scattered intensity at B should vary with the square of the number
of scattering molecules. This directly conflicted with empirical
evidence that the intensity varies as the first power. Furthermore,
the temporal distinctness of the primary impulse would disappear
entirely in the secondary because secondary impulses arise in each
of a tremendous number of electrons spread over macroscopic
distances. The net effect at B of a single primary impulse would last
for an appreciable time. For a scattering body i cm across, the
impulse seen 900 from the primary beam would persist 3 X io4
times as long as the primary impulse.15 Add to this the rapid
succession of primary impulses in any discernible x-ray beam and
the net disturbance at B "would last continuously."16
Lorentz offered a way out of the dilemma. If the impulses are
made doubly directed, so that the net excursion in amplitude is
balanced on both sides of equilibrium, the interfering effect can be
reduced to zero. As long as the net integral over the impulse
vanishes, all is saved. But this solution brought into question the
long-accepted mechanism of x-ray stimulation. How can a speed-
ing electron, brought to rest against the anticathode, create a
doubly directed impulse? "One can scarcely refrain from think-
ing," Lorentz said, that the "electric force in the excited im-
pulses . . . must chiefly lie in the same direction as would be the
case if the stoppage of the particle occurred in a straight line."17
Lorentz had touched a sensitive nerve. His was not the first
published consideration of the precise shape of the x-ray impulse,
but it was the first to clarify the issues. Stokes had assumed a
doubly directed impulse in 1897, having explicitly rejected a singly
directed one. But he published a sketch that may have misled
others. Thomson adopted the square pulse later used by Sommer-
feld. Stokes had merely mentioned in passing that "the dynamical
theory [of diffraction] shows that [the square pulse] is not possi-
ble."18 He very likely foresaw what Lorentz later made explicit. To
15
Lorentz found that the duration of the effect at B of a single pulse is given by 2/
sin (6/2)/c, where / is the linear dimension of the scattering body and 6 is the
angle line AS makes with the extension of the incident beam. At 6 ^ 900, and
/taken to be 1 cm, the duration is about 5 x io"11 sec. Wien had estimated that
a typical beam carries 7 x 1014 pulses per second, each of them being some
1.4 x io~15 sec at most. Each secondary is thus 3 x io4 longer than its primary.
16
Lorentz, Amsterdam Verslag, 21 (1913), 920.
17
Ibid., 923.
18
Stokes, Papers, 5 (1905), 266.
Figure 8.4. A singly directed impulse modulated by a doubly directed

one.
my knowledge, the only person aware of the potential ambiguity

before 1912 was Paul Ehrenfest, who wondered privately in 1905
whether the pulse might not be a step function.19 When Wien
derived a value for pulse width from energy considerations in that
same year, it was only 1 percent of the value that Sommerfeld
found from Wind and Haga's photograph. J. D. van der Waals, Jr.,
raised the possibility that Wien had measured only part of an
irregular impulse produced by many zigzag motions of the elec-
tron as it stopped.20 Wien replied accurately that his analysis was
independent of a zigzag assumption.21 In doing so, he missed the
importance of the argument. For the zigzag is precisely what
Lorentz used to extricate the impulse hypothesis of x-rays from the
tight corner in which he found it.
Lorentz's solution depended on the ways in which impulses are
scattered by matter. The singly directed impulse is preferentially
scattered at long wavelengths; this is a consequence of the fact,
already remarked by Sommerfeld, that its Fourier spectrum peaks
at zero frequency. Physically, this implies that in passing through
matter, singly directed impulses lose energy to atoms much more
readily than do doubly directed impulses. A zigzag motion of the
electron in slowing down will produce rapid fluctuations in the
amplitude of the pulse, yet the impulse itself will remain singly
directed because the net electron deceleration is in one direction
only. A singly directed impulse results, modulated by in the
simplest case, a doubly directed impulse, as shown in Figure 8.4.
As soon as the net disturbance passes through air or the glass wall
of the x-ray tube, the singly directed impulse is absorbed, leaving
only the doubly directed part accessible to measurement.
19
Ehrenfest notebook 1-06, second entry number 508, dated 31 December 1905.
Ehrenfest archive.
20
van der Waals, Jr., AP, 22 (1907), 603-5.
21
Wien, AP, 22 (1907), 793~7-
THE BRITISH INTERPRETATION OF
X-RAY INTERFERENCE
As we noted in Chapter 6, W. H. Bragg had considerably softened
his opposition to periodic x-rays by late 1912. He began to develop
a dualistic concept of both x-rays and y-rays in which a wavelike
and a particlelike component exist side by side in the beam. The
idea owed something to remarks he had made as early as 1907
when confronted by Marx's velocity measurement for x-rays. By
1912, Bragg was convinced that visible light as well was implicated
in this duality. He expressed his dualistic speculations at the 1912
meeting of the British Association even before he assimilated
Laue's interference results. He argued for a theory that would
allow "a corpuscular and a wave theory of light at the same
time."22 The conceptual difficulty in this theory he thought was
"one of our own making" for having in the past so rigidly demar-
cated the concepts of particle and wave.
Thus, even when confronted with the evidence for interference,
Bragg's position could still be defended. It is quite true that he first
thought of a corpuscular explanation of the x-ray interference
photographs; the directions toward the spots might coincide with
the directions down the "channels" between atom planes in the
crystal.23 But to emphasize this claim is to miss the dualistic stance
of Bragg's revised position of 1911. In the larger context of his
thought, it was hardly more than an attempt to see if the expected
corpuscular interpretation was forthcoming for what clearly al-
ready had a wave interpretation. It was less particles versus waves
as particles and waves: Bragg actively sought "one theory which
possesses the capacities of both."24 He was far from opposed to the
wave view, but he did have doubts about attributing the entire set
of interference spots to a few well-defined wavelengths in the
primary x-ray beam. Even Barkla expressed doubts regarding
Laue's methods.25
Bragg's son, William Lawrence Bragg, under the influence of his
22
W. H. Bragg, Nature, 90 (1913), 529-32, 557-60. He sought a "scheme of
greater comprehensiveness, under which the light wave and corpuscular x-rays
may appear as the extreme presentments of some general effect." Bragg,
Studies in radioactivity (1912), 192-3.
23
W. H. Bragg, Nature, 90 (1912), 219.
24
Ibid, 361.
25
Barkla to Rutherford, 29 October 1912. Rutherford papers. Heilbron, Moseley
(1974X62-3.
al
T~
OpTical pa^
A=
SB
ab - at sin
* 4 * cos
%*& \ - zosZB c ^ sin* a
Tkvs
A-
Figure 8.5. Derivation of the Bragg condition for x-ray interference in

crystals.
father and of the Cambridge pulse theorists, invoked the British

theory of the diffraction grating to gain an intermediate solution.
Gouy, Rayleigh, and Schuster had concluded that a diffraction
grating imposes its own regularity on the irregular impulses
thought to constitute white light.26 The replicated impulses then
interfere with one another, reinforcing the Fourier components
that make up the impulse. Each component frequency appears at
its own characteristic angle relative to the incident beam. The
younger Bragg recognized that Laue had found the x-ray analog to
this optical process.
However irregular and aperiodic the incident impulses may be,
the regularity of the crystal forces the scattered impulses to follow
one another at regular intervals. The effective separation is id sin 6
where dis the separation of the crystal planes and 9 is the glancing
angle of the x-ray beam against the face of the crystal (see Figure
8.5). One does not need to have Laue's few monochromatic wave-
26
See Chapter 2.
lengths, each with an independent existence in the x-ray beam.
The crystal manufactures them by the natural process analogous
to Fourier analysis of the pulses.
Before the beginning of December, the younger Bragg had
obtained clear interference photographs from x-rays reflected off
mica at glancing angles.27 By the time that this first demonstration
was given that x-rays can reflect, it had taken on even greater value
as evidence that x-rays must be periodic waves. Whenever the
incident angle 6 is such that the reflected impulses follow one
another spaced by a whole multiple of a component wavelength,
that frequency is constructively reinforced. An ideal crystal with
only one effective spacing (any real crystal, of course, contains
many effective spacings) should produce an accurate Fourier spec-
trum of the impulses thrown at it. The new approach was more
than reminiscent of optical spectroscopy. The earliest attempts to
chart the variation in x-ray intensity as a function of angle were
done using modified optical spectroscopes: An x-ray tube and a
collimating slit replaced the light source, and a crystal replaced the
grating. For the most sensitive of these early surveys, an ionization
chamber replaced the photographic plate.
The first question that occurred to the elder Bragg was whether
the paradoxes of quality and quantity apply to the interfering
x-rays. Could these monochromatic x-ray waves pass on abnor-
mally high velocities to electrons? Or was that property limited to
the nonperiodic impulse x-rays? He joined his son in the investiga-
tion, and in January 1913 wrote to Ernest Rutherford to report that
indeed "the ray travels from point to point like a corpuscle [yet]
the disposition of the lines of travel is governed by a wave theory.
Seems pretty hard to explain, but that surely is how it stands at
present."28
In April the Braggs were rewarded by finding distinct interfer-
ence peaks at sharply defined angles.29 It was most remarkable that
27
W. L. Bragg, PCPS, 17 (1912), 43-57. The younger Bragg credited C. T. R.
Wilson's lectures on the analysis of white light as having influenced his
reformulation of crystal diffraction analysis. Start of x-ray analysis (1967), 3.
28
W. H. Bragg to Rutherford, 9 and 18 January 1913. Rutherford papers.
Heilbron, Moseley (1974), 74.
29
W. H. Bragg to Rutherford, 10 March 1913. Heilbron, Moseley (1974), 75-6.
W. H. and W. L. Bragg, PRS, 88A (1913), 428-38. Three days before the
Braggs' announcement, Julius Herweg at Greifswald reported a similar suc-
cess, PZ, 14 (1913), 417-20.
Degrees.
Figure 8.6. Bragg's evidence for multiple orders of x-ray spectra. [Bragg
and Bragg, PRS, 88A (1913), 413.]
the form and spacing of the peaks did not change appreciably when
different crystals were used to diffract the x-rays. The same pattern,
shown in Figure 8.6, was found at just slightly different angles for
the x-rays reflected from rock salt, iron pyrite, zincblende, potas-
sium ferrocyanide, potassium bichromate, calcite, quartz, and
sodium ammonium tartrate. At first, Bragg tried to discern a
harmonic progression in the angle of the peaks. The distribution of
the peaks was remarkably like that produced by visible light on one
side of an optical grating. In each case three peaks, A, B, and C,
were accompanied by three more at proportionately greater
angles. The distribution looked like multiple orders generated by
optical emission spectra. The new spectroscopy had been suitably
rewarded with x-ray emission lines.
Bragg measured the absorption coefficient for each of the three
line sets. He found that each set (e.g., C,, C2, C3) has a distinct
penetrability independent of the crystal. He had found "three sets
of homogeneous rays," and he concluded, "since the reflection
angle of each set of rays is so sharply defined, the waves must occur
in trains of great length. A succession of irregularly spaced pulses
could not give the observed effect."30 Once they had estimated the
spacing of atom planes in the crystals, the Braggs gave the first
unambiguous estimate of^the wavelength of a homogeneous x-ray.
30
W. H. and W. L. Bragg, PRS, 88A (1913), 436.
The value, 1.78 X io~8 cm, was considerably longer than Laue's
hypothetical wavelengths, but in line with Sommerfeld's pulse
width estimates.
By 1913, x-ray emission lines were not unexpected; Barkla had
spoken in 1911 offluorescentx-rays as a "line spectrum."31 But the
sharpness of the angular definition was unexpected. Even so, the
i/ioth degree angular resolution of the Bragg apparatus prevented
the full realization of homogeneity. The conflict between x-rays of
definite wavelength and the paradoxes of x-ray absorption was
demonstrated clearly enough to Bragg. "The problem remains,"
he said, "to discover how two hypotheses so different in appear-
ance can be so closely linked together."32 Bragg's call for a dualistic
theory of light grew louder.33 But no resolution appeared possible.
The Braggs saw in the new spectroscopy the means to a detailed
understanding of crystal structure. From mid-1913 on, their efforts
were directed primarily to this end. They did not continue their
investigation of the nature of x-rays, and for that reason their
otherwise commendable work ceases to be of immediate concern
to us.
HENRY MOSELEY AND THE ATOM

Corroboration of the Braggs' result was immediately forthcoming
from Henry Moseley and Charles G. Darwin in Manchester.
Moseley was, at the time, physics demonstrator for Rutherford, to
whom he had come from Oxford to learn research methods in
radioactivity.34 His first study there was designed to corroborate
Rutherford's contention that only one ^-particle is emitted from
each radioactive decaying atom.35 But when Moseley heard of
Laue's spots, he shifted interests. "Some Germans," he wrote his
mother in October, "have recently got wonderful results by passing
31
Barkla and Nicol, PPSL, 14 (1911), 9-17.
32
W. H. Bragg and W. L. Bragg, PRS, 88A (1913), 436.
33
W. H. Bragg to Richardson, 20 May 1913. Richardson papers. The Braggs
thought Laue's methods were "very complicated" and "unwieldy to handle."
The reflection technique "made it possible to advance." W. H. and W. L.
Bragg, X rays and crystal structure (1915), vii, 10.
34
Heilbron, Moseley (1974).
35
Moseley and Makower, PM, 23 (1912), 203-10. Moseley, PRS, 87A (1912),
230-55.
X-rays through crystals and then photographing them, and I want
to see if the same results are to be found with y-rays."36
In preparation for his "thorough frontal attack," he first con-
vinced Rutherford that x-rays were a worthwhile experimental
subject. No previous work on x-rays had been done at Manchester.
Next, he tried to clear up the more confusing aspects of Laue's
discussion. Following a course parallel to that of William L. Bragg,
including recourse to the same theory of diffraction of white light,
Moseley arrived, step by step, at almost the same results as the
Braggs. Like the Braggs, he and Darwin tested first whether the
ionizing properties of the interfering x-rays were similar to those of
unscattered x-rays.37 Next, the pair retrieved a large optical spec-
trograph from the stores at the Manchester laboratory for a sys-
tematic survey of the intensity of scattered x-rays as a function of
angle. Moseley and Darwin had already plotted 1(6) before the
discovery of lines by the Braggs. Their equipment, and some
inexperience, caused them to miss the emission lines entirely. The
spectrograph was accurate to i min of arc, six times finer than that
of the Braggs. In breaking up the continuum of angles for the
survey, they had passed right over the extremely sharp lines.
But once informed of the peaks, they put the high resolution to
good use, showing quickly that two of the three lines are doublets
and that all are superimposed on a continuous background.38
Then, realizing that the absorption coefficients were very close to
those long known from Barkla's work, Moseley identified the
peaks as the Aând L series of secondary x-rays from the platinum
anticathode of the x-ray tube. Soon Bragg was corroborating their
results for tungsten and nickel.39 As Moseley wrote to his mother,
"there is here a whole new branch of spectroscopy, which is sure to
tell one much about the nature of an atom."40
From the angular width of an interference peak, Moseley and
Darwin tried to estimate the length of the wave train that produced
it. They were unable to determine the width precisely because they
36
Moseley to his mother, io October [1912] and 14 October [1912]. Heilbron,
Moseley (1974), 193-4.
37
Moseley and Darwin, Nature, 90 (1913), 594.
38
Moseley and Darwin, PM, 26 (1913), 210-32.
39
W. H. Bragg, PRS, 89A (1913), 246-8.
40
Moseley to his mother, [18 May 1913]. Heilbron, Moseley (1974), 204-5.
did not know the effective size of the focus spot on the anticathode
from which the x-rays came, but they could show that the width of
a line did not exceed 5 min of arc. This upper limit ensured that the
interfering rays were 'Very nearly monochromatic."41 They ar-
gued from this that the wave trains must be at leastfiftyoscillations
long, but were willing to go much further based on a purely
theoretical argument. If the radiation comes from a classically
oscillating electron, damped by the loss of radiated energy, its
amplitude could drop to half its original value in not less than
1,000 wavelengths!42
In no sense could it be maintained any longer that x-rays
differed from ordinary light. The characteristic secondary x-rays
were fully as monochromatic as visible colors. The Bremsstrah-
lung impulses constituted an exact analogy to polychromatic white
light; indeed, Moseley and Darwin referred to the continuous
x-ray background as white x-rays. They offered a qualitative analy-
sis of the background x-ray spectrum, interpreting it as the super-
position of multiple orders of Fourier spectra of the Bremsstrah-
lung impulses.
The conflict between Laue's and Moseley's views on the nature
of interference was one that had arisen before. Like Lord Rayleigh
before him, Moseley thought that impulses are separated into
Fourier component waves by the grating; like Stokes, Laue did not.
But however one chose to visualize the action of the grating, there
could be no doubt that x-rays were neither more nor less than
simple high-frequency light. Whatever the ambiguity that clouded
the relationship of the periodic to the aperiodic x-rays, it was an
ambiguity intimately shared by visible light. Whatever the resolu-
tion sought for the paradoxes of the new radiations, it was now
clear that it had to apply to visible light as well.
As a test of spatial localization of the constructively interfering
x-rays, Moseley and Darwin estimated how many crystal
41
Moseley and Darwin, PM, 26 (1913), 228.
42
They solved the equation of motion for a heavily damped electron and found
that the envelope of the oscillating displacement decreases exponentially with
a decay constant inversely proportional to the square of the wavelength. At an
intensity maximum of the x-ray spectrum, 1.397 A, half-amplitude is reached
after 2,680 ^k oscillations, where k is a constant dependent on the crystal type
and ranges from 0.125 to 1.0. Thus, at least 1,000 oscillations are needed to
reach half-amplitude.
"grooves" were involved in producing the beams of great angular
resolution they measured. "In each reflecting plane," they re-
ported, "an area containing at least a million atoms must have
been concerned in producing so perfect a reflexion."43 This corol-
lary of the paradox of quantity thus added to the problem that
Moseley called "this most mysterious property of energy."44
As is well known, Moseley went on in the next year to a
systematic survey of the K and L radiations induced in a large
variety of materials, and found a linear relation between the square
of the x-ray frequency and the atomic number of the element. As
Heilbron points out, Moseley could scarcely have avoided the
interpretation of his results in light of atomic theory. While he was
in Manchester, Niels Bohr was also there, bringing to fruition the
first statement of his atomic theory.45 Moseley was thereby di-
verted from his original interest in the interference of y-rays, and
from attempts to resolve the paradoxes of radiation. He and
Darwin had promised a closer analysis of the angular resolution of
x-ray lines, "as the matter is of great theoretical interest."46 But
they did not pursue the subject. For the remainder of his short life,
Moseley joined the increasing number of physicists who aban-
doned the unrewarding landscape of radiation theory for the
greener pastures of the Bohr atom.
CONTINENTAL REACTION TO
CRYSTAL INTERFERENCE
Because of the uncertainties of Laue's initial interpretation, the

first reaction on the Continent to the discovery of x-ray interfer-
ence was mixed. Nonetheless, by the fall of 1913, virtually all
physicists believed that crystal diffraction had indisputably dem-
onstrated that x-rays are equivalent to ordinary light. To most, this
meant that x-rays, like visible light, consist of periodic electromag-
netic oscillations. Only a few physicists, accustomed already to the
discontinuous properties of x-rays, began to ask if light should not
43
44
Moseley to his sister, [2 February 1913]. Heilbron, Moseley (1974), 200- 1
45
Bohr, PM, 26 (1913), 1-25.
46
share those unwavelike characteristics. Interest in x-ray research
grew by a factor of three from 1911 to 1913.47
Whatever the reservations Sommerfeld held before the discov-
ery of x-ray interference, they vanished afterward. He repeatedly
spoke on x-rays before general audiences in the following years,
using the Laue result as the central theme of his talk. In one of the
first of these talks, delivered in April 1913 to the German Society
for the Advancement of Instruction in Mathematics and Natural
Science, he explained that "impulses and waves are both deter-
mined by Maxwell's equations." He saw no conflict in the exis-
tence, side by side, of both impulses and periodic waves in x-rays.
It was nothing like the conflict "between the wave theory and
corpuscular theory of light."48 He also described Lorentz' concerns
about the form of the Bremsstrahlung pulses, a ratherfinepoint for
his audience. He was intensely interested in that analysis, and had
already begun an extension of Lorentz' work that he published two
years later. Although a full explanation of the origin of the periodic
fluorescent x-rays was still lacking, "one thing can be discerned,"
he said; "the theory is on the right track."49
Stark's response to crystal diffraction of x-rays was more
guarded. He had been carrying out experiments on the mechanical
effects he thought should be produced in crystals by impacts of a-
and canal-ray particles.50 He thought he had found evidence that
the particles are preferentially deflected along the "avenues" of the
crystal. In a paper published shortly after Laue's announcement,
Stark claimed to have come already to the conclusion that the
same behavior would be shown by the x-ray lightquanta. "It is
remarkable," he added gratuitously, "that I had come to these
views on the basis of the light-cell hypothesis even before I knew of
the experiments . . . taken as evidence for 'interference' of x-
rays."51
47
The Beibldtter zu den Annalen der Physik lists approximately twenty-five
papers on x-rays in each year from 1907 to 1910. The year 1911 has only twelve.
Thereafter, the numbers rise rapidly: 33, 51, 64, 61, and 63 in 1916.
48
Sommerfeld, Naturwissenschaften, 1 (1913), 708. Sommerfeld spoke fre-
quently on the new experiments on x-ray interference in the following years.
Munchener medizinische Wochenschrift, 62 (1915), 1424-30; 63 (1916),
458-60.
49
Sommerfeld, Naturwissenschaften, 1 (1913), 706.
50
Stark and Wendt, AP, 38 (1912), 941-57.
51
Stark, PZ, 75(1912X975.
Stark's inflated claims did not lead him to attempt a reconcilia-
tion of lightquanta with interference once it became clear that
more was involved than channeling of particles. By mid-1913 he
accepted the Braggs' explanation of crystal diffraction.52 Subse-
quently, he renounced the lightquantum. Interference, he said,
demands that lightquanta extend along the direction of propaga-
tion, and thus are not localized. If, on the other hand, lightquanta
are localized from front to back, their spectral lines should be
conspicuously broadened in an electricfield.53Stark would win the
Nobel Prize in physics in 1919 for his experimental researches on
line broadening in electric fields. Yet in 1912 he could find no
spectral evidence that lightquanta are localized. Soon thereafter,
Stark, the first supporter of the lightquantum hypothesis after
Einstein, altogether dropped the nature of radiation as a topic of
research.
Laue answered Stark's early claims, and in the process showed
that the Bragg reflection law, nX = id sin #, is a special case of his
own method. It applies, he said, when the interfering rays are "so
spectrally inhomogeneous as to be incapable of interfering after a
difference in [optical path] of only a few wavelengths."54 He still
thought that the spots were due to a few homogeneous component
x-rays. The Braggs, he felt, were pursuing only a special case of his
more general analysis.
Laue concluded with the frank admission that the effect on the
interference maxima of the thermal agitation of lattice points was
"less than any expectation," but he was unable to explain why.
The lattice point vibration was treated in detail in a pair of papers
by Debye. Sommerfeld's former assistant had just finished the
joint paper on the photoelectric effect, discussed in Chapter 7.
Debye gave an influential talk on the quantum theory of specific
heats and heat conduction at the 1913 Wolfskehl lecture series in
Gottingen.55 The theory, introduced by Einstein in 1907, became
the first subject after cavity radiation to excite widespread interest
in the quantum theory.56 Debye focused on Planck's remarkable
prediction of a "zero-point" energy of atoms at absolute zero
52
Stark, PZ, 74(1913), 319-21.
53
Stark, VDpG, 76(1914), 304-6.
54
Laue, PZ, 74(1913X422.
55
Debye, Kinetische Theorie der Materie (1914), 17-60.
56
Klein, Science, 148 (1965), 173-80. Kuhn, Black-body theory (1978).
temperature. He turned to the thermal influence on crystal diffrac-
tion as a possible source of corroborating experimental data.
In July 1913 Debye concluded that thermal motion would have
little or no observable effect on the definition of the Laue spots.57
He calculated the expected decrease in intensity of the interference
spots located at increasing angles from the beam direction. This
relation depends on the effective wavelength of the x-rays and
allowed Debye to derive values for x-ray wavelengths indepen-
dently of the crystal parameters required by both Laue and
Bragg.58 The end result was a relation giving the decrease in spot
intensity for increasing order of the spectrum.59 The more com-
pressible the crystal, that is, the weaker the interatomic binding
forces, and the higher the temperature, the faster the intensity
decreases. In this way, Debye provided the means to reconcile
Laue's and the Braggs' interpretations of interference. If the higher
orders of spectra decrease quickly enough in intensity, the contin-
uous x-ray spectrum may be largely free of the superposed high
orders. Such a spectrum would therefore be a more accurate
reflection of the Fourier spectrum of the impulses that produce it
than would a spectrum including the higher orders.
In the fall of 1913, Laue renounced his appeal to monochro-
matic primary x-rays. At the Naturforscherversammlung in
Vienna, Walther Friedrich reported that "it is, for the first time,
experimentally demonstrated [that] the absorption coefficient for
x-rays is a measure of the wavelength."60 He specifically denied the
assertion that only a few wavelengths in the primary beam are
responsible for the interference spots. In the discussion that fol-
lowed, Laue protested that Friedrich had credited him with a
viewpoint that he had "never held," that the primary x-ray beam is
itself monochromatic. If he had not been entirely clear before, he
claimed, it was due to his "strong reservations" about the "only
possible explanation that he had seen" at the time.61 So, Laue
brought his own interpretation into line with that of the Braggs
while still retaining its generality. Indeed, he now thought that the
57
Debye, VDpG, 15 (1913), 678-89. For the state of this issue before Debye's
treatment, see Forman, AHES, 6 (1969), 38-71
58
Debye, VDpG, 75(1913), 738-52.
59
Ibid., 857-75.
60
Friedrich, PZ, 74(1913), 1081.
61
Ibid, 1085.
primary beam consists of a "wide continuous spectrum." It was
particularly fortunate, he said, that the Braggs, Moseley, and
Darwin had found "isolated monochromatic x-rays" so that cor-
roboration of the basic laws of interference could be found.62
By 1915 Sommerfeld could speak "with complete confidence of
'x-ray light' distinguished from normal light by precise determina-
tion of its wavelength."63 As early as 1900, he had suggested the
division between periodic and aperiodic waves, two ends of a
continuum stretching from monochromatic light to the x-rays.
This idea was now widely applied to distinguish the two types of
x-rays. The characteristic x-ray lines are like monochromatic light;
Sommerfeld compared them to "distinctly colored spectral
lines."64 Like white light, the Bremsstrahlung impulses may be
thought of either as irregular pulses or as a superposition of Fourier
component waves. No longer as concerned as he had been in 1900
with keeping the two concepts separate, he concluded, "it is more a
question of ease and purpose which way of speaking we employ."65
Sommerfeld developed this idea of interchangeability between
impulse and continuous spectrum representations in a careful
study of the spectrum of Bremsstrahlung pulses in 1915. He ex-
tended Lorentz' argument against singly directed pulses using
Fourier spectrum analysis to do so, and showed inter alia that a
doubly directed square pulse has a spectrum that peaks at a
frequency approximately equal to nc/d, rather than at zero fre-
quency like the singly directed impulse.66 Moreover, it was clear
from the analysis that inherent in the very nature of the problem
was an unavoidable tradeoff between well-defined frequency and
well-defined pulse width. The more temporally restricted is the
impulse, the more component frequencies have to be added to
produce it, and the notion of a single frequency for the impulse
becomes less well defined. Conversely, the very sharp x-ray emis-
sion lines, which define virtually pure frequencies, have to be
thought of as periodic disturbances extending over macroscopic
distances in space. Just where in space the effect of the x-ray is
located then becomes unclear.
62
Laue, PZ, 74(1913), 1077.
63
64
Ibid., 5.
65
66
Sommerfeld, AP, 46(1915), 721-48.
It was not so much that physicists prior to this period were
unable to visualize x-ray impulses as a superposition of periodic
waves, but that there had been little advantage in doing so. The
temporal discontinuity of impulses had attracted attention to
them originally. Most physicists in thefirstdecade of the twentieth
century thought of x-rays as short temporal events rather than as
the net effect of many extended processes in the aether. After 1912
that view became dated, and the new physics of x-rays exploited
more and more the Fourier tradeoff between temporal discontin-
uity and frequency. Evidence for the interference of x-rays did not
convince physicists that x-rays are periodic waves; rather, it con-
vinced them that, whatever the true nature of radiation, x-rays act
in all respects just like ordinary light.
CHARACTERISTIC y-RAYS
Rutherford's hypothesis that y-rays are the electromagnetic im-
pulses produced by the expulsion of ^-electrons from the atom
remained purely qualitative until Sommerfeld's 1911 study. In this
view, one would expect to find a correlation between ^-particle
velocity and y-ray penetrating power. Although the latter property
could be measured, not until 1910 were methods developed to
determine with accuracy the velocities of released ^-electrons.
Consequently, there was little more than fragmentary evidence
before 1912 that y-rays possess properties characteristic of their
parent element. McClelland had found in 1904 that the y-rays from
radium are not homogeneous in penetrating power; their absorp-
tion does not follow the exponential law.67 A student of Ruther-
ford's in Montreal, Tadeusz Godlewski, noted the following year
that y-rays from different sources exhibit differing penetrability.68
Richard Kleeman, Bragg's former co-author, found at least two
distinct absorption coefficients combined in the y-rays from ra-
dium.69 In his experiments in 1904 Paschen had found two peaks in
the velocity distribution of the ^-particles emitted by radium,70
and Kleeman suggested in 1907 that these are intimately related to
67
McClelland, PM, 8 (1904), 67-77.
68
Godlewski, PM, 70(1905), 375-9.
69
Kleeman, PRS, 79A (1907), 220-33.
70
Paschen, AP, 14 (1904), 389-405; PZ, 5 (1904), 5O2"4-
71
the two strong y-ray components he had found. Otto von Baeyer,
Otto Hahn, and Lise Meitner found more compelling evidence of
/?-particle velocity peaks in 19 io.72
There was little evidence, however, that the perceived groupings
of y-rays are characteristic of their radioactive source, even if the
/^-particle velocity peaks seem to be. At first, Rutherford was
reluctant to believe that the patterns in y-ray penetrating power
could be, as he later put it, "ascribed to a characteristic radiation
set up in the radiator by the primary y-rays."73 When D. C. H.
Florance measured the intensity distribution of the y-rays scattered
from lead in 1910, Rutherford interpreted the results as proof that
y-rays cannot be stimulated by external influences on matter.74 But
the following year, J. A. Gray demonstrated that y-rays - or what
appeared to be y-rays - could, like x-rays, be stimulated in matter
by incident /^-particles.75 This set to rest several vague doubts:
Some radioactive substances are known to emit a significantly
lower intensity of y-rays than would be expected were the y-rays a
product of /^-particle emission. Actinium is a case in point. Ruth-
erford thought that this might be due to "some peculiarity in the
mode of emission of thefirays," and his doubts about characteris-
tic y-rays began to fade.76
First, Rutherford set Moseley to work to verify that only one
^-particle is emitted by any single atom. Moseley showed that the
elements called radium B and radium C emit only one ^-electron
when an atom decays. In the meantime, Gray found what at the
time seemed convincing evidence that y-rays can stimulate sec-
ondary y-rays in matter,77 although it is clear in retrospect that
these were high-frequency characteristic x-rays. Initially, Ruther-
ford doubted Gray's claim that these secondary "y-rays" are char-
71
Eve, PM, 11 (1906), 586-95. Kleeman, PM, 14 (1907), 618-44; 75 (1908),
638-63.
72
von Baeyer and Hahn, PZ, 11 (1910), 488-93. von Baeyer, Hahn, and
Meitner, PZ, 72(1911), 273-9.
73
Rutherford, Radioactive substances (1913), 282. Although it had very little
influence on his research, Rutherford came very close to adopting Thomson's
unit hypothesis for x-rays in 1913, p. 84.
74
Florance, PM, 20 (1910), 921-38.
75
Gray, PRS, 85A (1911), 131-9. Important preparatory work was done by
Starke, Le Radium, 5 (1908), 35-41; and by Davisson, PR, 28 (1909), 469-70.
76
Rutherford, Radioactive substances (1913), 271.
77
Gray,/W, 26(1913), 611-23; T^S, 56^ (1912), 513-29;^7^(1912),489-501.
acteristic of the element from which they arise. But in 1912 he
began to take seriously evidence that the groups of y-rays of
homogeneous penetrating power correspond to the peaks in /?-
electron velocity. At about the same time that the news of crystal
diffraction of x-rays came from Munich, Rutherford concluded
that "the transformation of energy from the /? ray form to the y ray
form or vice versa takes place in definite units which are character-
istic for a given ring of electrons."78
Rutherford now demanded the evidence for characteristic y-rays
that he had denied before. He set about to find it himself, and he
soon felt he had succeeded. Frederick Soddy, with his wife and
Alexander Russell, had shown in a series of experiments between
1909 and 1911 that, aside from endpoint behavior, y-rays are
absorbed exponentially along most of their path through lead.79
This is the classical symptom of homogeneity. With a student,
Rutherford measured absorption coefficients for y-rays spontane-
ously emitted from several elements, finding that radium B gives
three "distinct types" and radium C a single type.80 The coeffi-
cients differed markedly; in units of inverse centimeters of alumi-
num, they are shown in the penultimate column of Figure 8.7, a
table from a slightly later publication.81 In a burst of overconfi-
dence, Rutherford declared that the y-rays of coefficient 45 from
radium D were, with "little doubt," the "characteristic radiation of
the L type to be expected from an element of atomic weight 2 io."82
By one of those misapprehensions that so enrich and color the
history of experimental science, Rutherford was here quite misled.
The electromagnetic spectrum of a radioactive element consists of
more than just the y-rays characteristic of it. The exiting y-rays
and other energetic emissions stimulate the surrounding electronic
structure of the atom, releasing, in addition, the characteristic
x-ray spectrum of the element. It was these high-frequency charac-
teristic x-rays that Rutherford studied, calling them "y-rays" all the
while. We commit a historical anachronism to explain this point
here, and certainly no discredit is to reflect on Rutherford. Indeed,
78
Rutherford, PM, 24 (1912), 453-62.
79
F. and W. Soddy and Russell, PM, 19 (1910), 725-57. Russell, PRS, 86A
(1911X240-53.
80
Rutherford and H. Richardson, PM, 25 (1913), 722-34.
81
Rutherford and H. Richardson, PM, 26 (1913), 937-44.
82
Ibid., 331. Richardson soon used the y-ray absorption coefficients to fix the
atomic weight of uranium at 230. PM, 27 (1914), 252-6.
y Rai/$ of the Thorium and Actinium Products.
Absorption Mais, absorption

Atomic coefficient ft ooeffioient p/d
Element. weight. in aluminium. in aluminium.
Radium B 214 230 1 (cm.)"1 851 (cm.)"1

40 f 14-7 L
0-51J 0188 J
Rndium C 214 0-115 0-0424
BadiumD 210 461 16-5 1
0-99/ 0-36]
BadiumE 210 similar types to D
but very feeble.
Mwothorinra 2 ... 228 261 (cin.)- 1 9-6 1
0-116/ 0031 J ,.
Thorium B 212 1601 59 1
32 I 11-8 I
036 J 013 J
Thorium D 206 0-096 0035
Radioaetiniuni 251 9*2 1
0190 f 0*070 /
Actinium B 1201 44 1 ,
31 I .. 114 I
045 J 0165 J ,,
Actinium D 0198 0073
Figure 8.7. Absorption coefficients of mixed x-rays and y-rays from

thorium and actinium products. [Rutherford, PM, 26 (1913), 943.]
as we shall see, Rutherford's continued belief that he was studying

characteristic y-rays and thereby consolidating his hypothesis that
they originate in electron groupings in the atom (where, in fact, the
rays he studied did arise), led to much important physics and
several essential discoveries. This again illustrates Francis Bacon's
wise dictum that truth emerges more readily from error than from
confusion. To reduce confusion for the reader, I will henceforth
put crucial references to Rutherford's y-rays cum x-rays in quotes
thus: "y-rays." But for the moment, we must leave Rutherford in
his state of fruitful error.
By the summer of 1913, when Bohr published his theory of
atomic structure, Rutherford felt he had bridged the gap in fre-
quency that had separated x-rays from y-rays. Never before had it
been suggested that in radioactive decay one might find rays
identical to those produced by electric discharges. "The radiations
from the various radioactive substances can be conveniently di-
vided into three distinct classes," Rutherford claimed. Aside from
the most easily absorbed component of// = 230, which was well
within the recognized domain of x-rays, the three groups are:"(i) a
soft radiation, varying in different elements from // = 24 to // = 45,
probably corresponding to characteristic radiations of the 'L' type
excited in the radio atoms; (2) a very penetrating radiation with a
value of// in aluminum of about .1, probably corresponding to the
'K' characteristic radiation of these heavy atoms; (3) radiations of
penetrating power intermediate between (1) and (2) corresponding
to one or more types of characteristic radiations not so far ob-
served with x-rays."83 He spoke freely of x-ray "wavelength" and
took the absorption data from the "y-rays" as if they predicted new
wavelengths to be found in x-ray spectra.
INTERFERENCE OF y-RAYS
After the discovery that x-rays can interfere, many physicists
thought that y-rays ought to show interference effects too. Moseley
started his x-ray investigations with this idea in mind. Walther
Friedrich gave two talks to the Wednesday Munich Colloquium
early in 1913 on interference of x-rays and then turned directly to a
review of analogous y-ray research.84 It was Rutherford who actu-
ally put x-ray diffraction apparatus to work and showed that
"y-rays" can show interference effects. His results, obtained with
Edward Andrade, were announced in the May 1914 issue of the
Philosophical magazine*5
Exceedingly long exposure times were required to capture the
penetrating "y-rays" on photographic plates. To make the diffrac-
tion angles as large as possible, the rays from radiums B and C were
reflected at glancing angles from rocksalt crystals. This decreased
the effective grating spacing. The y-ray source was placed in a
strong magnetic field to deflect all ^-particles and secondary elec-
trons. As a final precaution, Rutherford used Maurice de Broglie's
technique of rotating the analyzing crystal to overcome its inher-
ent irregularities; he worried about "contorted and undulating
surfaces." A very slow scan by the crystal accentuates the lines of
interest by smearing out ghost lines produced by periodic irregu-
larities in the atom planes.86
83
Rutherford and Andrade, PM, 27 (1914), 855.
84
The Register of the Miinchener physikalisches Mittwochscolloquium, 1 July
1912, 19 February 1913, 7 May 1913. AHQP, 20.
85
Rutherford and Andrade, PM, 27 (1914), 854-68. A brief first announcement
was made in Nature, 92 (1913), 267.
86
M. de Broglie, CR, 157 (1913), 924-6. See Heilbron, Isis, 58 (1967), 482-4.
KADIUM B. PLATINUM.
Soft y-ray spectrum. X-ray spectrum.
Angle of reflexion Wave-length Intensity. Angle of reBexion

from roc'ksalt. (in cm.)- 1122
8 0' 793 X10" 8 in.

8 10' 809 m.
8 34' 838 m. 8 27'
8 43' 8&J m. 8 43'
9 23' 917 f.
9 4.V 953 m.
10 3' 982 B. 10 2'
10 18' 1-OOfi in. 10 13'
10 32' 1029 in.
10 48' 1-065 f.
11 0' 1-074 f
11 17' 1100 f.
11 42' 1141 m
12 3' 1175 . 12 3'
12 16' 1196 m
12U 31' 1-219 f.
13 01 1-266 f.
13 14' 1286 f.
13 31'' 1-316 f.
13 62* 1349 m
14 2' 1-366 m.
Figure 8.8. Rutherford's data for what he took to be characteristic y-ray

lines. [PM, 27(1914), 861.]
As early as October 1913, Rutherford and Andrade could report

success. Faint lines appeared at angles of io and 120. With the best
current estimate of atom spacing in rocksalt (2.814 X io~8 cm),
they calculated the wavelengths of some twenty-one lines. The
results are shown in Figure 8.8. The two strongest lines have
wavelengths of 0.982 and 1.175 X icr 8 cm. Again, I remind the
reader that what Rutherford here believed were y-rays were actu-
ally high-frequency secondary x-rays from the source element.
Rutherford immediately tried to find room for the " y-ray " lines
in Moseley's scheme of x-rays. The wavelength of the strong line
from radium B (ju = 40), if inserted in the L series of x-rays, gave
that element an atomic number of 82. This was "unexpected
corroboration" of Kasimir Fajans' and Soddy's prediction that
radium B is an isotope of lead.87 This justification of the hypothesis
of radioactive decay series encouraged Rutherford's belief that a
smooth transition would be found between x-rays and y-rays.
In August 1914 there followed studies on what Rutherford
thought were the harder y-rays from radiums B and C. This time
Rutherford found it impossible to obtain the normal crystal pic-
87
According to Fajans and Soddy, radium B, actinium B, and thorium B are all
isotopes of lead (82).
tures. Instead, he shot the primary y-rays directly through thin
sheets of crystal and recorded interference lines on a photographic
plate at a distance.88 The results contributed two more wave-
lengths: 0.99 and 1.6 X io~9 cm. These forced Rutherford to make
the first of many revisions in his attempt to link the "y-ray" data
with Moseley's x-rays. Rutherford and Andrade found a line from
radium B at i4o'. This hard line (ju = 0.115) and that from
thorium B (// = 0.36) had therefore to be taken as evidence of a
"higher mean frequency" series than the K series. 'This may for
convenience," Rutherford went on at a dizzying pace of specula-
tion, "be named the 'H' series, for no doubt evidence of a similar
radiation will be found in other elements when bombarded by
high-speed cathode rays."89 Rutherford was still convinced that the
source of the y-rays was intimately connected to that of the charac-
teristic x-rays. He was therefore very much interested when, at this
juncture, the source of the characteristic x-rays was identified in
Germany.
A THEORY OF X-RAY EMISSION

Ernst Wagner was one of many investigators of x-rays at Rontgen's
Institute for Experimental Physics in Munich; he directed his
attention to the absorption of x-rays by matter. With the new
spectroscopic techniques, it was possible to determine the precise
wavelength at which continuous x-rays are preferentially absorbed
by atoms.90 Wagner found that the absorption is not restricted to a
specific x-ray wavelength, although the low-frequency threshold
for absorption is sharply defined. A typical absorption "edge" is
shown schematically in Figure 8.9. Wagner began a systematic
catalog of these absorption edges in metals.91 He expected to find
resonant absorption of x-rays at or near the characteristic x-ray
emission frequencies of the atom; instead, he found slight differ-
ences between the onset of absorption and the slightly lower
frequencies of emission. What happened to the difference?
Wagner, following the Bohr model of the atom, suggested qualita-
88
Rutherford and Andrade, PM, 28 (1914), 263-73. See also Rutherford to
Boltwood, 20 June 1914, quoted in Badash, Rutherford and Boltwood (1969),
293-
89
90
Heilbron, Isis, 58 (1969), 451-85.
91
Wagner, AP, 46 (1915), 868-92. Kossel drew his data from Wagner, Miinchen
Sb, 44(1914), 329-38.
low v high v
Figure 8.9. Schematic representation of the difference between x-ray

emission lines and absorption edges.
tively that the discrepancies would show up as differences in

electron energy levels.
Enter Walter Kossel. Educated under Philipp Lenard at Heidel-
berg, Kossel had come to Munich in 1913 to study with Sommer-
feld. Although he never held an official position at the Institute, he
kept abreast of the research work done there. He also had the
benefit of close discussions with Bohr when the latter visited
Munich.92 Kossel tried to bring order to Wagner's data by arrang-
ing a table of x-ray absorption features as a function of wave-
length. He soon noted that, except for the more intense of the two
lines that together comprise the Remission feature, a strong x-ray
absorption edge is always found at a frequency slightly greater than
that of the associated emission line. As Germany was mobilizating
for World War I, in October 1914 Kossel hit upon a solution to this
regularity from the point of view of the Bohr atom.93 The absorp-
tion of x-rays might be responsible for the ejection of an electron
from a filled, low-lying atomic ring. Emission follows when the
atom returns to equilibrium. The idea bore much similarity to
Lenard's treatment of spectral radiation following photoioniza-
tion. Kossel had studied the emission of electrons from metals by
92
Bohr to Kossel, 19 July 1921. Bohr archive; AHQP BSC 4, 2.
93
His inspiration occurred between September and October 1914. VDpG, 16
(1914), 898-909, 953-63. Heilbron, his, 58 (1969), 451-85.
incident electron beams for his dissertation under Lenard at Hei-
delberg.94
With Kossel's suggestion that characteristic x-ray lines originate
in electron transitions to the innermost rings of the Bohr atom,
x-rays gained theoretical credentials equivalent to those of visible
light. Both were, in the context of the Bohr theory, the product of
downward electron transitions: visible light to high-lying electron
rings, x-rays to low-lying ones. Sommerfeld's subsequent elabora-
tion of the Kossel scheme further established the validity of the
hypothesis. Theoretical and empirical evidence had converged,
justifying further the equivalence of x-rays and light.
In the first months of 1915, evidence mounted that Kossel was
right. Ionization of the atom must precede the emission of charac-
teristic x-rays. Transitions to refill the lower rings produce the
characteristic lines. As Heilbron has indicated, this mechanism
would not likely have suggested itself to Bohr, who had already
rejected ionization as a prerequisite for optical emission. In Kos-
sel's hypothesis, the absorption edges directly measure the energy
levels in the atom. Differences in absorption edges allow calcula-
tion of the fine splitting of the main quantum levels, and this soon
became the chief empirical means of refining the Bohr theory.95 It
also enabled Sommerfeld and others to make predictions about
the number of electrons that shield the nucleus at each level of the
atom. The main significance of the new understanding of x-rays
was that it provided a means to probe more deeply into the
structure of the atom, not to stimulate interest in the associated
x-ray photoelectric effect. It was not until 1921 that the latter
problem was seriously examined, as we shall see in a later chapter.
In the meantime, the success of the Bohr theory attracted away the
talents ofjust those physicists who were most capable of producing
a theoretical solution to the problems of free radiation.
THE SOURCE OF THE y-RAYS

The source of the x-rays was known by 1915, but the origin of the
y-rays remained obscure. Rutherford could only grope toward a
qualitative y-ray emission mechanism based on electron configu-
94
Kossel, AP, 57(1912), 393-424.
95
The central document of this program is Sommerfeld, Atombau und Spektral-
linien (1919), and successive editions. See also Heilbron, Atomic structure
(1964), and his contribution in History of twentieth century physics (1977),
40-108.
rations in the atom. Few physicists disagreed at the time that y-rays
originate from a source in the atom equivalent to that of x-rays, but
the attempt to find room in the electronic superstructure of the
Bohr atom failed repeatedly. In part, the very strength of the
y-ray- x-ray analogy slowed the recognition that y-rays might arise
from a level of atomic structure that was entirely different from
electron orbits.
Rutherford was convinced that a smooth transition would be
found between the characteristic x-ray and y-ray spectra. But it was
very difficult to measure "y-ray" wavelengths directly using x-ray
methods. So, immediately after presenting what he believed, in-
correctly, were the first few y-ray wavelength measurements,
Rutherford used them to construct a general relation between
wavelength and absorption coefficient. This allowed him to derive
approximate y-ray wavelengths from simple absorption data. On
the basis of this extrapolation, Rutherford submitted a hypothesis
to relate the y-rays to atomic structure.96 The characteristic "y-
rays" were thought to be excited in special configurations of
electrons within the atom. Radioactive disintegration ejects an
electron "from or near the nucleus." If this ^-particle happens to
pass through one or more of the special regions in the electronic
structure, a monochromatic y-ray is excited. These, in turn, can
reconvert their energy, either singly or in combination with other
y-rays, by stimulating a secondary /^-particle. This produces the
sharp peaks observed in the ^-particle velocity spectrum.
There was little or no evidence for a continuous y-radiation
analogous to the x-ray Bremsstrahlung. But James Chadwick had
found that most of the energy radiated by the ^-particle lies in the
continous distribution of velocity, not in the sharply defined
/?-particle velocity peaks.97 Rutherford's mechanism of y-ray emis-
sion explained why: Only y-rays of well-defined frequency are
emitted by the special electron groups. The /?-rays, on the other
hand, originate in the uncharted depths of the nucleus, where
mechanisms may very well produce a wide distribution of ener-
gies. Moreover, /?-rays that leave the nucleus without stimulating
characteristic y-rays are complemented by others that are ejected,
like photoelectrons are, by combinations of y-rays. It is the latter,
relatively rare, ^-particles that form the characteristic /?-ray veloc-
ity peaks.
96
Rutherford, PM, 27(1914), 488-98.
97
Chadwick, VDpG, 16(1914), 383-91.
Rutherford could also explain why some elements, such as
radium E, appear to have no /?-ray lines at all. If few y-rays are
produced in the electron rings, there is no way that corresponding
/?-ray lines can be produced. Do atoms of radium E simply lack the
necessary electron configurations? No, answered Rutherford. If
the ^-particles are ejected in a fixed direction relative to the
electron shells, they might not pass through the special regions at
all. This he thought was "not improbable" given the "remarkably
definite way" that the atom disintegrates. This argument also
explains how two isotopes, the atomic structure of which is the
same according to the x-ray data, can differ markedly in their y-
and /?-ray spectra.
Rutherford had departed from his original inspiration about the
y-rays. No longer did they accompany the emission of each /?-par-
ticle; now, only certain /?-rays in certain atoms were priviledged to
stimulate y-rays. He was caught, as it were, between the success of
the pulse theory of y-rays and the demands of his crystal diffraction
data. He slipped into a form of expression that remarkably mir-
rored his own ambiguous viewpoint. Reflecting the contemporary
interest in quantum energy relations, he proposed that y-ray en-
ergy is emitted in integral multiples of hv, where v is the frequency
of the y-ray. Clearly, this could not apply to single y-ray impulses,
each of which carries a single hv in total energy. Rutherford was
still thinking of the y-rays more as impulses than as waves. The
interference that produces the crystal pictures was due to repeating
series of impulses. The emission of energy nhv in the form of y-rays
occurs, he said, in linked "trains" of wavepulses. And the entire
train of n pulses can act in concert to stimulate a single secondary
/^-particle; this particle obtains the entire quantity nhv of energy.
At a slightly later point, Rutherford applied the same imagery to
x-rays: he spoke ambiguously of "a train of one or more waves
each of energy /zv."98 He accordingly tried to demonstrate that the
frequencies of the known "y-ray" lines could be expressed as
integral multiples of a unit of energy.
The war intervened and disrupted all investigations at Ruther-
ford's Manchester laboratory. The y-ray studies "had to be broken
off."99 Chadwick was interned in Germany as an enemy alien. In
98
Rutherford, PM, 2^(1914), 319.
99
Rutherford, Chadwick, and Ellis, Radiations from radioactive substances
fact, the demands of true y-ray crystallography were themselves
too formidable for continued studies of y-ray wavelengths. The few
high-frequency x-rays found by Rutherford and Andrade and
thought to be y-rays in 1914 were the last measured directly for
over a decade. Extremely sensitive detectors were required; glanc-
ing angles as small as 10 min of arc proved necessary, as were
elaborate measures to protect the photoplate from high-frequency
secondary x-rays.100 The imperfect correlation of "y-ray" absorp-
tion with wavelength provided most of the estimates of true y-ray
wavelengths in this period. Some information about y-ray fre-
quencies could be extracted after 1921 by extrapolating the mea-
sured x-ray photoeffect relations to y-rays. But as we shall see in the
next chapter, the most significant of these transformation experi-
ments on the energy balance between radiation and matter were
based on principles that emphasized the particulate aspects of the
radiation.
The conceptual shift needed for further understanding of the
y-ray source seems to have arisen largely by elimination of other
alternatives. By 1922 it was clear that electron transitions, even to
the lowest layers of the atom, are not energetic enough to produce
y-rays. There simply was no room left in the atom. When Charles
Ellis showed that y-ray frequencies, like visible and x-ray frequen-
cies, could be reduced to differences among a set of energy values,
he sought the locus of the rays in the nucleus because it was all that
remained.101 His hypothesis was soon bolstered by data on y-rays
from a number of radioactive elements; and Rutherford's few
remaining doubts vanished by 1925 that true y-rays are character-
istic of the nucleus, not the electronic structure of the atom.102
After this reorganization of thought, it was possible to separate
entirely y-rays from x-rays: y-rays are x-rays of nuclear origin.103
The relatively low-frequency characteristic "y-ray" lines that
100
On the difficulty of y-ray measurements, see Rutherford to Bohr, 20 May
1914. Bohr archive. Quoted in Bohr, Collected works, 2 (1981), 593-4. Ko-
varik, PR, 19 (1922), 433. Thibaud, Theses (1925). Frilley, Theses (1928).
101
Ellis, PRS, 1 0 1 A (1922), 1-17.
102
Ellis and Skinner, PRS, 705^(1924), 185-98. Ellis, PCPS, 22(1924), 369-78.
Black, PRS, 106A (1924), 632-40; 109A (1925), 166-76.
103
Rutherford and Wooster, PCPS, 22 (1925), 834-7. Rutherford here used the
best available data on his "y-ray" frequencies, together with Moseley's rela-
tion, to show that the atomic number of radium B is 83, not 82. This result
suggested to him that the "y-ray" is emitted as a part of, or after, the ^-particle
emission that transforms element 82 to 83.
Rutherford had fitted so readily into Moseley's x-ray scheme are,
in fact, characteristic x-rays stimulated by a- and ^-particles as
they disrupt the electronic structure of an atom. To be sure, y-rays
and x-rays are not very different from one another, except in
energy, after each is released from the atom. But the result of
believing them to be equivalent to one another was ultimately to
locate the source of the y-rays in the nucleus, a markedly different
level of atomic structure than that whence come both x-rays and
visible light. The y-ray- x-ray analogy had come to an end, largely
as a result of its long-lived success.
Quantum transformation experiments
Where does the ejected electron get its kinetic energy
when its separation from the light source becomes so
great that the light intensity almost completely vanishes?1
The same x-ray spectroscopic techniques that led to the identifica-

tion of the atomic origin of x-ray spectra also offered means to test
the localization of energy in the radiation. In the second decade of
this century, x-ray spectrometers used as monochromatizers pro-
vided more purely defined x-rays than had been available from
natural characteristic radiation. This, coupled with techniques to
determine the kinetic energy of secondary electrons, provided
opportunities for precise testing of the quantum transformation
relation, E = hv. These experiments form the subject of this
chapter.
We are primarily concerned here with the absorption of radia-
tion, not with its emission. To be sure, the quantum regulation of
the emission of radiant energy has no classical explanation; yet
there is no electromechanical inconsistency implied in the cre-
ation of a spherical wave containing a definite amount of energy. It
is the inverse case that causes real difficulty. How can that quan-
tum of spherically radiating energy concentrate its full power on a
single electron? For this reason, verification of the quantum rela-
tionship for emitted radiation lies outside our direct concerns. The
Franck-Hertz experiments beginning in 1912, for example, dem-
onstrated the quantum nature of energy transfer, but they did little
to encourage acceptance or even consideration of the lightquan-
tum.2 On the other hand, the experiments detailed here that
verified the quantum nature of the absorption of light and x-rays
1
Planck, Das Wesen des Lichts (1920), 15.
2
Franck and G. Hertz, VDpG, 14 (1912), 457-67, 512-17.
gave substance to the lightquantum hypothesis because they veri-
fied the particlelike transfer of radiant energy to matter.
In none of the experiments on the energy of radiation, going
back to those by Rutherford and Wien, had it yet been possible to
make comparisons between incident radiant energy of specified
quality and the resulting velocity of released electrons. X-ray
spectroscopy made available sources of precisely defined x-ray
energy; lacking atfirstwere means to chart with equal precision the
energies contributed to electrons by the radiation. Robert Millikan
provided unambiguous proof in 1914 that Einstein's law of the
photoelectric effect is valid over a large spectral range, but resist-
ance to the lightquantum was strongest for visible and ultraviolet
light. Millikan himself remained unconvinced that the quantum
interpretation should replace the wave theory of light. Rutherford
and his students developed an improved /?-ray velocity spectro-
meter on the eve of the Great War and began to make progress
toward a full understanding of the quantum absorption of x-rays.
But the exigencies of the war stopped this research, and it was not
revived systematically until the 1920s.
Thus, although these experiments set the stage for the eventual
full-scale verification of Einstein's linear photoelectric relation for
x-rays and y-rays, and actually provided it for light, they did not
sway opinion to the lightquantum hypothesis. For that essential
step in the abandonment of deterministic physics more was re-
quired, including a fuller and clearer understanding of the energy
needed to remove a deeply bound electron from an atom. Efforts
prior to 1920 to cope with absorption of radiation effectively
mirror the dualistic interpretation that was to come. Spectroscopy,
a technique founded in principle on the wave nature of radiation,
contributed the data that most clearly demonstrated the particu-
late nature of light.
THE EMPIRICAL STATUS OF THE

PHOTOELECTRIC EFFECT
Experimental studies on the photoeffect after 1910 were made with
the growing realization that light and x-rays are common forms of
radiation. Even before Philipp Lenard's rejection of the triggering
hypothesis in 1911, evidence had pointed to analogies between
x-ray ionization and the photoeffect. It had been shown in 1907 by
Quantum transformation experiments 235
Innes, Cooksey, and Laub that the velocity of the secondary
electrons released by x-rays is independent of the intensity but
increases with the quality of the x-rays.3 Wilhelm Seitz had carried
the analogy even further in 1910 with studies on the behavior of the
x-ray photocurrent in the vicinity of zero external potential. The
result, he said, "reminds one qualitatively of the Lenard curve for
the photoelectric effect."4 Also, in 1910 Otto Stuhlman, working
with Owen Richardson at Princeton University, investigated the
photoemission from a io~7-cm-thick platinum film deposited on
the back side of a quartz plate that was transparent to ultraviolet
light. The number of electrons released in the same direction as the
transmitted light beam was some 15 to 17 percent greater than the
emission in the backward direction, measured when the same film
was laid on the front side of the quartz plate.5 Within a week, R. D.
Kleeman at Cambridge provided independent corroboration.6
Electrons stimulated by light show the same asymmetrical pattern
as do those released by x-rays.
The assumptions used to interpret the similarities between light
and x-rays changed markedly after the discovery of x-ray interfer-
ence. Neither Stuhlman nor Kleeman had felt that the asymmetry
in photoemission contradicted the wave theory of light "in its
ordinary electromagnetic form."7 Charlton Cooksey, the Yale
physicist who had first demonstrated the forward scattering of
x-rays, summarized the data on the equivalent action in the pho-
toeffect in 1912. He stated that periodic ultraviolet light should be
evaluated on the same scale of penetrating power as x-rays.8 This
conclusion was reached even before the announcement of x-ray
diffraction. After x-rays were known to interfere, the analogy
between x-rays and light began to exercise its influence to alter
ideas about light rather than to define x-rays in comparison to
3
Innes, PRS, 79A (1907), 442-62. Cooksey, AJS, 24 (1907), 285-304. Laub, AP,
26 (1908), 712-26.
4
Seitz, PZ, 77(1910), 707.
5
Stuhlman, Nature, 83 (1910), 311; PM, 20 (1910), 331-9; 22 (1911), 854-64.
Rubens and Ladenburg found in 1907-before the Bragg-Barkla debate clari-
fied its importance-that a difference of 1:100 exists between the photocurrent
stimulated in the backward and forward directions from a gold foil i o 5 cm
thick. VDpG, 9(1907), 749-52.
6
Kleeman, Nature, 83 (1910), 339; PRS, 84A (1910), 92-9.
7
Stuhlman, PM, 20 (1910), 339.
8
Cooksey, PM, 24 (1912), 37-45.
light. And the number of experimental studies on the photoefFect
peaked in 1912, as can be seen in Figure 7.2.
As discussed briefly in Chapter 7, the empirical understanding of
the precise form of the photoelectric relation was lacking even as
late as 1913. First, it was difficult to ensure that photocathode
materials used in tests were free of impurities. The mose sensitive
cathods are the alkali metals with loosely bound outer electrons.
These substances are subject to rapid oxidation, and emission
from the oxidized products clouded the results of many experi-
ments. Second, one had to be careful to separate the effects of the
normal photoefFect from the selective photoefFect discovered in
1910.
Two other influences were responsible for the strong experimen-
tal interest in the photoefFect around 1912. Of greatest significance
was Lenard's renunciation of the triggering hypothesis the year
before, an event that removed the basic presupposition that un-
derlay most previous serious studies of the phenomenon. Second
was the discovery of x-ray crystal diffraction, a result that fully
identified x-rays with ordinary light. It consequently raised for
visible light all the paradoxes of x-rays and y-rays.
One of the most significant experiments on the photoefFect in
this peak year 1912 was done by Owen Richardson and Karl
Compton at Princeton. Richardson's interest in the photoefFect
had been partly stimulated by his close ties to J. J. Thomson. He
had been a student in Trinity College, had taken a first in the
natural sciences tripos, and had done research under Thomson.9 In
1906 he had gone to Princeton University as a faculty member, and
it was there that he began research on the photoefFect with his
former student, Karl Compton. Dissatisfied with previous failures
to remove impurities, they directed single wavelengths of ultravio-
let light onto photocathodes that had been carefully scraped clean
and then kept in a vacuum.10 Their results appeared to substantiate
Einstein's relation - that light frequency is proportional to the
square of electron velocity - a result that Richardson later admit-
ted he had not expected to find.11 Criticism by Robert Pohl and
9
Swenson, DSB, 11 (1975), 419-23.
10
O. Richardson and K. Compton, PM, 24 (1912), 575-94. Hughes, PTRS,
212A (1913), 209-26.
11
O. Richardson, PRS, 94A (1918), 269.
Peter Pringsheim in 1913 brought the explicit form of the relation-
ship into question again, but it was thereafter clear that no reso-
nance peaks were to be found.12 Richardson explicitly pointed out
the similarities between x-rays and light, reflected on the paradoxes
that stood in the way of a consistent theory of x-rays, and rehearsed
the data for asymmetrical photoemission. The last, he remarked,
might lead one "to the conclusion that [like x-rays, light]
. . . consists of a shower of material particles."13
To avoid this unpalatable conclusion, Richardson offered a
treatment of the photoeffect "from a rather wider standpoint"
than either lightquanta or waves, "without making any definite
hypothesis about the structure of the radiation."14 Expressed in its
fullest form in 1914, Richardson's derivation of the photoelectric
relation rested on no assumptions about the nature of light or
about the structure of the atom. It was a purely thermodynamic
argument that bypassed completely the issues that led to the
double paradox of quality and quantity. Photoemission of elec-
trons was considered to be analogous to the evaporation of a gas.15
Richardson assumed the Planck law for the energy distribution in
the radiation, and also assumed that the number of electrons
emitted by the light is proportional to its intensity. The result
predicted a sharply defined wavelength threshold for emission,
analogous to a latent heat of evaporation from the metal of the
cathode. Moreover, he showed that the maximum energy of an
electron has to be proportional to hv C, where C, like the
photoelectric work function, is a constant characteristic of the
metal.
But Richardson's interest in the photoelectric effect was largely
motivated by concerns other than the nature of light. An under-
standing of the photoeffect was important to his study of the
12
Pohl and Pringsheim, VDpG, 15 (1913), 637-44.
13
O. Richardson, PM, 25 (1913), 145.
14
Ibid.
15
O. Richardson, PM, 27 (1914), 476-88. Richardson's interest in the photoef-
fect grew out of his research on the "glow discharge," the emission of electrons
from hot metals. The earliest statement of the method was in PM, 23 (1912),
594-627, in particular 615ff. See also PM, 24 (1912), 570-4; a note appended
in proof to the last paper appeared separately in Science, 36 (1912), 57-8; PR,
34 (1912), 146-9. Richardson's theory deserves closer scrutiny than it has
received. For a beginning, see Stuewer, Minnesota studies in philosophy of
science, 5, (1970), 246-63.
electron discharge from a heated cathode.16 The "glow discharge"
soon became his central research subject, and in 1916 he sought an
explanation of it in terms of a photoeffect stimulated by the light
emitted from the hot cathode.17 Although Richardson's photoelec-
tric research was restricted to purely phenomenological tests that
contributed precise data on a series of metals not analyzed before,
he continued to watch the developing evidence about the photoef-
fect carefully.18 Having formulated his own non-lightquantum
hypothesis, he was confident that a resolution would be found for
the problems of the absorption of radiation that would avoid the
"restricted and doubtful hypothesis used by Einstein."19
MILLIKAN'S DEFINITIVE EXPERIMENTS
Among those who had taken an early interest in the photoelectric

effect was the talented American experimentalist Robert Millikan.
Beginning his study in 1906, Millikan was unwillingly drawn ever
closer to a quantum interpretation of the phenomenon. First, he
found no increase in the maximum electron velocity with increas-
ing temperature of the cathode.20 This was one of the first recog-
nized difficulties with the triggering hypothesis, but in the face of
the then general confusion about photoeffect results, it was not
fatal. Soon Millikan showed that photoelectric fatigue - the de-
crease in sensitivity known to affect metals after long exposure to
light - is an effect of impurities created by oxidation; the fatigue
disappears entirely when the metals are kept in a vacuum.21 With
this new technique, Millikan examined the response of a number
of metals to ultraviolet light produced by sparks from zinc elec-
trodes fired from a bank of twelve Leyden jars.22 He measured
greater electron velocities than had been seen before. There was no
16
Richardson won the 1928 Nobel Prize in physics for his work on thermionic
emission of electrons.
17
O. Richardson, PM, 31 (1916), 149-55.
18
K. Compton and O. Richardson, PM, 26 (1913), 549-67. Richardson and
Rogers, PM, 29 (1915), 618-23.
19
20
Millikan and Winchester, PR, 24 (1907), 116-18.
21
Millikan and Winchester, PR, 29 (1909), 85. Millikan, PR, 30 (1910), 287-8.
See also von Baeyer and Gehrtz, VDpG, 12 (1910), 870-9; Hughes, PCPS, 16
(1911), 167-74.
22
Millikan and Wright, PR, 34 (1912), 68-70. See also Millikan, PR, 35 (1912),
74-6; VDpG, 10(1912), 712-26.
evidence of the critical potential that Einstein had predicted
should stop all emitted electrons. Moreover Millikan's student,
J. R. Wright, in ignorance of the selective photoeffect, thought he
had found evidence of a clear resonance maximum in aluminum
at 2,166 A.23
Between 1911 and 1914, Millikan's evidence against the quan-
tum relation crumbled; again the photoelectric effect confused its
students. Pohl and Pringsheim showed in 1912 that the new high-
velocity electrons Millikan had found were due to electrodynamic
effects of the sparks used as the light source; their experiments with
arc light failed to duplicate Millikan's extreme values.24 Wright's
resonance peak could be attributed to the selective effect, the lack
of a threshold to systematic errors in technique. The experiments
were not easy. Carl Ramsauer, who would later distinguish himself
with the most precise of all measurements of photoelectron veloci-
ties, was tricked in 1914 too. Applying magnetic deflection to sort
secondary electrons by velocity, Ramsauer hinted at the conclu-
sion, soon utterly disproven, that there is no threshold for photo-
emission, and that the velocity distribution is entirely independent
of light frequency.25
Millikan tried again. This time he concentrated on the fre-
quency-velocity relation, and provided unassailable proof that
Einstein had been right all along. The experiments were done at
the Ryerson Laboratory in Chicago using an ingenious and com-
plicated apparatus - "a machine-shop in vacuo" - that allowed
photocathodes to be scraped clean of impurities and tested entirely
under a vacuum.26 First, the contact potential between the photo-
cathode and the test apparatus had to be determined; then the
actual photoelectron stopping potential could be measured at
various wavelengths. Millikan tested sodium and lithium and
found that the relation between light frequency and the potential
that just turned back the fastest electrons is a perfectly straight line,
as shown by the solid line in Figure 9.1, which is based on his
graphic presentation of the results. The slope of the line is equal to
the best contemporary value for h/e within 0.5 percent. Millikan
then corrected this best-fit line for the contact potential between
23
Wright, PR, 33(1911), 43-52.
24
Pohl and Pringsheim, VDpG, 74(1912), 974-82. Millikan, P.K, 7(1913X73-5.
25
Ramsauer, Heidelberg Sb, 5A (1914), Abh. 19, 20.
26
Millikan, PR, 4 (1914), 73-5; PR, 6 (1915), 55.
1.6
1.2 Conta< :t
Potem ial
Correc tion
.8
.4 y*
intensit
Thresh Did Frejuency = 57.0 x10 1 3 Hz Di ect Ob servatic
ofTh resholc Freque ncy
.4
photoc jrrent p >er unit

8
1.2
1
55x10 13 60 65 70 75 80 85 90 95 100 55 59 63 ;10 1 3
Frequency (Hz)
Figure 9.1. Millikan's data for lithium corroborating Einstein's photoelectric law. [Courtesy of the Exploratorium, San
Francisco. From Millikan, PR, 7(1916), 377.]
cathode and detector, shifting it upward to the dashed line. The
intercept of the corrected relation accurately marked the threshold
frequency for emission, measured separately and shown in the
inset on the right side of Figure 9.I.27 Closer corroboration of
Einstein's linear law, equation 5.1, could not be desired. To my
knowledge, Millikan's results were never questioned, and were
quickly recognized by leading European physicists to be definitive.
But acceptance of the Einstein law of the photoelectric effect did
not carry with it acceptance of the hypothetical lightquantum.
MR. WHIDDINGTON'S REMARKABLE RULE

Four years before the quantitative verification of the quantum
relation for the photoeffect, significant regularities had already
appeared in the transformation between x-ray and electron energy.
Before 1912, however, it was conceptually difficult to assign x-ray
impulses a definite frequency. The corroboration of the quantum
relations, for example in Wien's pioneering work in 1907, was
therefore obscured. In these early years, one had to rely on classifi-
cation of x-ray quality according to different penetrating powers,
and in these experiments British physicists excelled. Throughout
the first decade of the twentieth-century, British experimental
work on x-rays outshone, and for a while outnumbered, that of any
other country.28
The early work on the x-ray transformation relations was pre-
dominantly British. G. W. C. Kaye showed in 1908 that x-rays
stimulated from different anticathodes by the same cathode ray
beam show a rough proportionality between their penetrating
27
Millikan, PR, 7 (1916), 355-88. The work is described further in Wheaton,
Photoelectric effect (1971), chap. 13, and Millikan, Electron (1917), 224.
28
Numbers of research papers on x-rays listed in the Fortschritte der Physik,
1900-10:
British French
German Italian
1900 01 02 03 04 05 06 07 08 09
Despite the equivalent number of British and German studies, the experimen-
tal work that was cited in this period was almost entirely British.
power and the atomic weight of the anticathode.29 The early
experiments were done without the benefit of Barkla's sharply
defined homogeneous x-rays. In 1910 Charles Sadler, who had
collaborated with Barkla in the discovery of characteristic second-
ary x-rays, found that the beam of secondary electrons emitted
from metals irradiated by a single fluorescent x-ray "line" is itself
always absorbed exponentially.30 This implied that the released
electrons were all of similar velocity. Sadler found that the velocity
was independent of both the intensity of the incident x-ray and the
nature of the tertiary radiating substance. He discovered, however,
that the electron velocity increases linearly with the atomic weight
of the secondary radiating substance that produced the homogen-
eous x-rays. Beatty found at roughly the same time that the
absorption of secondary electrons varies linearly with the absorp-
tion of the various homogeneous x-rays he used to stimulate
release of the electrons.31 But the most important relationship that
tended to confirm the close connection between the characteristic
secondary x-rays and the characteristic optical emission spectra of
chemical elements was discovered by Richard Whiddington in
1911.
Whiddington studied the secondary x-rays from chromium and
other elements at the Cavendish Laboratory in Cambridge. The
rays obediently appeared, with their characteristic coefficient of
absorption of 136 cm"1 in aluminum, only when the potential of
the discharge tube exceeded a certain threshold. When Whidding-
ton substituted various materials as anticathodes using Kaye's
techniques, he found that there was a definite order in the pene-
trating power of the emitted secondary x-rays. Thisfindingwas not
unexpected. Barkla had shown enough to suggest that such rela-
tionships should hold. But Whiddington thought that he had
found a possible violation of Stokes' law for the fluorescent nature
of secondary x-rays. "There is reason to believe," he hinted, "that
this secondary radiation can be extracted by a primary radiation
somewhat less penetrating than itself."32 Thus, it was extremely
important for him to refine his apparatus so that the electron
energy required to stimulate a given x-ray line could be measured
accurately.
29
Kaye, PTRS, 209A (1908), 123-51.
30
Sadler, PM, 19 (1910), 337-56.
31
Beatty, PCPS, 15 (1910), 416-22; PM, 20 (1910), 320-30.
32
Whiddington, PCPS, 15 (1910), 575; PRS, 85A (1911), 99-118.
T O ELECTROSCOPE
Figure 9.2. Whiddington's apparatus to show the proportionality of the

atomic weight of an element to the minimum velocity of electrons
capable of producing x-rays that will, in turn, stimulate characteristic
x-rays in that element. [PRS, 85A (1911), 324.]
Six months later, Whiddington offered the results of his new

apparatus.33 It incorporated characteristics of an early cathode-ray
velocity spectrometer and is shown as Figure 9.2. Bulb D produces
the cathode-ray beam; the brass cylinder Q wound with wire
generates a magnetic field directed out of the plane of the paper.
Thefieldselects electrons of only a single velocity; these then strike
the anticathode at T. Whiddington substituted materials of var-
ious atomic weight A for the scattering plate at R; for each mate-
rial, he determined the minimum electron velocity that would
produce primary x-rays capable of stimulating characteristic rays
from R. His results for seven elements are given in Table 9.1. They
illustrate a relation approximately expressed by
^critical ^ A X 10 8 cm/sec (9.1)
an expression soon named Whiddington s rule in England.
Whiddington expected that a similar relation could be found for
the x-ray photoeffect, the inverse case of that given above. He
33
Whiddington, PCPS, 16 (1911), 150-4; PRS, 85A (1911), 323-32. First an-
nouncement in Nature, 88 (1911), 143.
Table 9.1. Whiddington 's data for seven elements
Velocity, v
Element (108 cm/sec) A A/v
Al 20.6 27 1.3
Cr 50.9 52.5 1.03
Fe 58.3 56 0.96
Mn 61.7 58.6 0.95
Cu 62.6 63.2 1.01
Zn 63.6 65.1 1.02
Sc 73.8 78.9 1.07
promised new experiments, but problems soon arose. The velocity

induced in electrons by x-rays was too great, and the electrons
could not conveniently be turned back by a decelerating electric
field of a practical magnitude. Moreover, the electrons released
from R are not concentrated in a beam like those in tube D, and
Whiddington found that he could not apply magnetic deflection to
determine their maximum velocity. He could only provide a
plausibility argument in 1912, based on prior data taken by Beatty
and Sadler to suggest, but not to confirm, that a comparable
proportionality holds for the velocities of photoelectrons created
by homogeneous x-ray beams.34
Whiddington's attempt was premature. He was restricted to the
small number of homogeneous x-ray emission lines from a few
heavy atoms, and he had no quantitative means to measure spe-
cific electron velocities. He could only hope at best to determine
the maximum velocity of released electrons. Even after 1912, the
new specificity in x-ray quality offered by crystal diffraction had
yet to be matched by techniques to measure electron energy peaks
throughout the distribution of speeds. Only when these two new
techniques were combined could the x-ray photoeffect be ade-
quately investigated. We have already discussed the origins of
x-ray spectroscopy. The early means developed to study the veloc-
ity of radioactively emitted /^-particles now demand our attention.
/^-PARTICLE VELOCITY SPECTROSCOPY

Friedrich Paschen presented a remarkable new tool to radioactive
research through his defense of the claim in 1904 that y-rays are
34
Whiddington, PRS, 86A (1912), 360-70, 370-8.
very fast electrons. He greatly advanced the magnetic deflection
experiments done theretofore on cathode rays by Lenard, Kauf-
mann, and others. Paschen so strengthened the magneticfieldthat
unwanted /?-ray electrons from his y-ray source were bent into
closed circles and never left the apparatus.35 The technique to
deflect electrons in a magnetic field so that those of different
velocity are separated on a photographic plate was not new; Bec-
querel had made such plates since 1899 before /?-rays were known
to be electrons.36 But Paschen demonstrated that precise control of
the electrons was possible and that correlations might be drawn
between the strength of the field and the velocity of electrons bent
to a predetermined radius by the magnet.
Ernest Rutherford was the first to exploit magnetic deflection
systematically. In 1905 he placed a photographic plate above a
wirelike source of a-particles. To arrive at the plate, the particles
had to pass through a collimating slit parallel to the source and
then traverse a magnetic field; this action separated electrons by
velocity.37 In this way, he showed that after passing through in-
creasing thicknesses of absorbing matter, a-particles slow down
appreciably until they are no longer able to expose the photoplate.
William Wilson applied a similar technique to /?-rays from radium
in 1909. His apparatus was fairly sophisticated and could bend
^-particle trajectories through a full 1800. The intent was to show
that homogeneous /?-rays are not absorbed exponentially but
rather linearly in matter.38 In these studies of a- and/?-particles the
y-rays were a definite hindrance, and Wilson did not apply his
methods to secondary electrons excited by the y-rays.
The attempt to derive quantitative results for/?-ray electrons was
first made by Otto von Baeyer and Otto Hahn, physical chemists in
Berlin. They based their work on Rutherford's magnetic deflection
of the a-rays and concluded that "the line spectrum" of electron
velocities produced on the photoplate offered "the possibility to
determine the velocity of single /?-rays."39 They placed a sensitive
photoplate above a sheet with a slit to allow through only a narrow
solid angle of rays, as shown in Figure 9.3. The undeflected y-rays
35
See Chapter 3. A general discussion of the development is found in Jenkin,
Leckey, and Liesegang, Journal of electron spectroscopy and related phenom-
ena, 72(1977), 1-35.
36
Becquerel, CR, 130(1900), 206-11; PRI, 77(1906), 85-94.
37
Rutherford, PM, 70(1905), 163-76.
38
W. Wilson, PRS, 82A (1909), 612-28.
39
von Baeyer and Hahn, PZ, 77 (1910), 493.
Figure 9.3. von Baeyer and Hahn's arrangement to measure ^-particle

velocities photographically by passing them through a magnetic field.
[PZ, 77(1910)TafelXIL]
mark the straight line through the slit from the radioactive wire
used as the source. The curved paths followed by /^-electrons in the
uniform magnetic field depend on the velocity of the electrons.
Since different-velocity electrons would pass through the slit on
orbits of different radii of curvature, the bands in the photographic
record indicated that there were distinct peaks in the velocity
distribution.
Once the technique was known, modifications were instituted
that led to the recognition of a remarkable focusing effect. In 1911
von Baeyer, Hahn, and Lise Meitner turned the photoplate so that
it lay in the plane of /^-particle motion and the entire course of the
electrons was traced, in much the same manner as Villard had
done in the experiments in which he discovered the y-rays.40 Jean
Danysz, working in the laboratory of Marie Curie, placed the
plate so as to take advantage of a full 1800 deflection of the elec-
trons.41 This arrangement was soon recognized to have a distinct
40
von Baeyer, Hahn, and Meitner, PZ, 12 (1911), 273-9, 378-9; 13 (1912),
264-6.
41
Danysz, CR, 753 (1911), 339-41, 1066-8; Le Radium, 9(1912), 1-5.
Figure 9.4. The self-focusing effect of electrons projected into a trans-

verse magnetic field.
advantage. As Rutherford put the matter in 1913, "The/?-rays of a

definite velocity comprised in a comparatively wide cone of rays
can be concentrated in a line of very narrow width on the photo-
graphic plate."42
Rutherford realized that, in a homogeneous magnetic field,
electrons of a given velocity will follow circular paths with a
constant radius of curvature. This is true regardless of the initial
angle of projection of the electron with respect to the plane of the
photographic plate. Electrons that begin their motion at any small
angle from the normal to the plate will follow different orbits, but
will all strike the photoplate at about the same distance from the
origin. How this is done may be seen in Rutherford's Figure 9.5,
but the reason that it works is more clearly rendered in Figure 9.4.
The circle describes the orbit, or partial orbit, of all electrons with
the same velocity. The different chords represent the relative
position of the photoplate for electrons projected at various small
angles from the normal. As is clear in thefigure,the distance from
the origin O to the darkened points P along the photoplate will be
essentially constant for all electrons of the same velocity.43 The
42
Rutherford and Robinson, PM, 26 (1913), 719.
43
Point P is sharply defined by electrons emitted precisely normal to the plate.
Electrons emitted on either side of normal will fall progressively closer to the
origin. The edges are therefore smeared toward the origin but are sharply
defined on the other side.
JUL
Figure 9.5. Rutherford and Robinson's ^-particle velocity spectrometer.

[PM, 26(1913), 720.]
electrons projected in different directions are focused at point P in

much the same way that a lens focuses on one point in the focal
plane all the light that diverges from a point on the object.
COMBINING X-RAY AND j8-RAY SPECTROSCOPY

Rutherford and his student, Harold Robinson, successfully fol-
lowed up on Danysz's modification of the /?-ray velocity analysis.
Danysz had found some twenty-five distinct velocity peaks in the
/?-rays from radiums B and C.44 In July 1913, Rutherford and
Robinson used the refined spectrometer shown in Figure 9.5 to
analyze the system of lines and derived a value of the electron
velocity for each.45 The radioactive source is placed at S, the
magnetic field extends out of the plane of the page in the region
above the photographic plate PP. About a year later, Rutherford
realized that the method could be applied to measure the velocities
of secondary electrons released from metals by y-rays. Thus, in the
same month that the first presumed y-ray wavelengths were mea-
44
Danysz, Le Radium, 70 (1913), 4-6.
45
Rutherford and Robinson, PM, 26 (1913), 717-29.
sured in Manchester, the /?-ray velocity technique was extended to
the measurement of electrons produced by y-rays.46
In 1914, Rutherford's students used the technique to investigate
x-ray photoelectrons in a test of "Planck's theory."47 A metal plate
replaced the radioactive source at S, and it was irradiated with the
x-rays produced by a nickel anticathode especially chosen for its
relatively simple characteristic spectrum. Because, they said, elec-
trons are released from various depths in the metal, the velocity
edges are smeared into 2-mm bands. For an iron radiator they
found three such bands, for lead two. Using the outer edge of the
band for measurement, they found that the velocity of one band
agreed within 2 percent with the value predicted by the quantum
rule for electrons stimulated by the K x-rays from nickel.
Thus, by the beginning of World War I, the experimental tech-
niques needed for the detailed verification of the x-ray photoeffect
relations were available in Britain. The obstacles that had stopped
Whiddington in 1911 were obstacles no longer. But it was not until
after the war that /?-ray spectroscopy was applied in a systematic
way to this problem. When it did come, it came from America and
France, not from Britain or Germany. Rutherford had pointed the
way, and had the war not intervened, he might have followed up
on the methods. But Rutherford never intended to specialize in
x-ray research; his elected research domain had become the nu-
cleus of the atom.
THE COOLIDGE X-RAY TUBE AND THE

DUANE-HUNT LAW
One of the great problems that faced any attempt to quantify the
characteristics of x-rays before 1913 was the inconstancy of x-ray
tube output. In 1913 a great step toward the solution of that
problem was made. William Coolidge at General Electric per-
fected a new x-ray tube that had a number of advantages over the
traditional one.48 Its cathode was made of a spiral tungsten wire;
Coolidge had developed the means of forming normally brittle
tungsten into a shape that could tolerate the heating needed to
46
Rutherford, Robinson, and Rawlinson, PM, 28 (1914), 281-6.
47
Robinson and Rawlinson, PM, 28 (1914), 277-81. Moseley to Rutherford, 27
May 1914. Rutherford papers. Quoted in Heilbron, Moseley (1974), 235.
48
Coolidge, PR, 2 (1913), 409-30.
maximize the outflow of electrons. The heated cathode allowed
the tube to produce x-rays in a higher vacuum than was previously
possible; harder x-rays resulted. The tube could also operate for
longer periods with remarkably little variation in either the "in-
tensity" (quantity) or "penetrating power" (quality) of the x-rays.
Moreover, each parameter could be varied independently. The
quality is determined by the steady potential across the electrodes,
the quantity by the temperature of the tungsten filament. When
combined with water cooling of the anticathode, the Coolidge tube
produced x-rays of dependable properties in steady operation over
extraordinarily long periods.49 In conjunction with the versatile
techniques of x-ray spectroscopy, it made possible experiments on
the quantum transformation relation that were no longer limited
to isolated characteristic x-ray wavelengths.
Among the first to recognize the importance to physics of the
American-made Coolidge tube was William Duane, who offered
strong evidence for the quantum law in x-ray emission. Duane was
a relative newcomer to x-ray studies; his previous work had been
almost exclusively in radioactivity.50 When he assumed a position
left vacant at Harvard after John Trowbridge left, Duane extended
his research to include x-rays as being of possible therapeutic value
in medicine. In 1914 he tried to improve the correlation of x-ray
wavelengths with absorption, and soon turned to x-ray spectros-
copy as the best means to test the quantum predictions of x-ray
emission.51 In December he showed what was no surprise to
European physicists but still a novelty to Americans: that the
proportionality constant relating the frequency of the x-ray to the
energy of the parent cathode-ray electron is on the order of magni-
tude of Planck's h.52
Duane was fortunate in having access to two special pieces of
apparatus at Harvard. Thefirstwas an accurate x-ray spectrometer
that had been built by David Webster. Duane also profited from
the careful work that Webster had done on the threshold voltage
required for entire series of characteristic x-rays to appear.53
49
See, for example, Dauvillier, Technique des rayons x (1924), 5off.
50
Bridgman, NAS Biographical memoirs, 18 (1937), 23-41. Forman, DSB, 4
(1971), 194-7.
51
Duane, PR, 4(1914X544.
52
Duane, PR, 6 (1915), 166-76.
53
Webster, PR, 7(1916), 599-613.
Equally important was the large storage battery consisting of
20,000 Plante cells constructed and left by Trowbridge. It was
capable of supplying extremely steady dc up to 45,000 V. In April
1915 Duane and Franklin Hunt used this equipment to test the
hypothesis that a givenfixedpotential can excite continuous x-rays
with a frequency only below a specific threshold.54 They tried
many dc potentials and found an abrupt drop in Bremsstrahlung
x-ray intensity for spectroscopic angles less than a well-defined
critical angle. That limiting angle increased as the x-ray tube
potential dropped. The implication was that electrons of a given
energy are able to excite x-rays of almost any frequency up to, but
never greater than, a certain threshold. The fixed dc voltages were
known, and the corresponding wavelengths could be calculated
from the spectrometer angles. The relation followed the approxi-
mate rule used by Wien eight years before and partially verified by
Whiddington in the meantime. To a few tenths of a percent:
F0 = - f (9.2)
e ^
where Vo is the critical potential and Xo the corresponding x-ray
wavelength. Equation 9.2 soon became known in America as the
Duane-Hunt law, and was recognized for many years to be the
most accurate of all means to measure h.55
The Coolidge tube was also important to the research group at
Manchester that carried on the x-ray studies begun by Henry
Moseley. Charles Darwin, Moseley's collaborator, had hoped that
monoenergetic electron beams might be isolated using magnetic
deflection. One could then use them to produce x-rays and see if
the wavelengths, measured by absorption and diffraction, corre-
spond to the quantum law. His early experiments were prevented
by the variation in x-ray output due to the unsteady vacuum in the
experimental tube. The work was soon dropped altogether when
he, like Moseley, left Manchester for the war.
In 1915 Rutherford resumed the work for a short while using a
54
Duane and Hunt, PR, 6 (1915), first 166-71. Webster argued that finding a
threshold would cast doubt on any impulse representation of Bremsstrahlung
because the Fourier spectrum of an impulse has no high-frequency limit. This
is true in principle but, as Figure 2.7 illustrates, the amplitude of the high-
frequency components in the pulse drop rapidly toward zero. For any practical
experiment there is an effective threshold. Webster, PR, 6 (1915), 56.
55
Birge, PR Supplement, 1 (1929), 57.
tottage in thousands.
Figure 9.6. Rutherford's data showing apparent deviation from the

quantum law for x-rays produced at extremely high potentials. [Ruther-
ford, Barnes, and Robinson, PM, 30 (1915), 353.]
new Coolidge x-ray tube. With the few students still in the labora-
tory, he tried to "see how far [the quantum] relation holds for the
excitation of x-rays by electrons."56 Up to 115,000 V the tube was
supplied with electricity from a motor-driven Wimshurst genera-
tor; from there up to 175,000 V they used an induction coil capable
of producing a 20-in. spark. Because the tube produced x-rays of
unprecedented hardness, the effective wavelengths were very small
and the angles required by crystal diffraction were correspondingly
minute. To estimate the wavelengths, they were forced to use the
imprecise relation of absorption coefficient to wavelength.
The results of the survey are shown as Figure 9.6. The calculated
frequency of what Rutherford termed the end radiation is essen-
tially the highest-frequency x-ray produced by the indicated volt-
age. They found systematic deviation from the quantum relation
at high potentials; rather than increasing in proportion to the
applied potential, the end frequency reached a maximum value at
about 142 kV. Rather than following the Planck relation hy = E,
the data conformed to hv = E kE2, where k is a constant.
Rutherford realized that these data were only approximate. By this
56
Rutherford, J. Barnes, and Robinson, PM, 30(1915), 352.
time, he was convinced that the Bohr-Kossel model of x-ray
production could explain the systematic deviation as energy lost to
the electron as it penetrates to the more tightly bound orbits within
the atom. But in 1915, knowledge of x-ray absorption edge energies
was insufficient to allow evaluation of further tests, and Ruther-
ford, pressed by wartime work, did not pursue any.
Rutherford's primary reason for undertaking this study had
little to do with x-rays; he still hoped to find evidence in support of
his hypothesis that characteristic y-rays originate in the electronic
structure of the atom. He did find some x-ray frequencies at high
potentials that were shorter than those estimated from absorption
coefficients of what he still thought were selected y-ray lines. But
the saturation in frequency made it clear that a definite limit
prevented the production of x-rays as penetrating as true hard
y-rays such as those from radium C. He wrote only one more paper
on x-rays, and then after the war moved back into experiments
entirely on radioactivity. The most influential voice in Britain on
x-rays continued to be Charles Barkla's.
SELECTIVE X-RAY ABSORPTION
The chief means of measuring x-ray quality in Britain remained

the rate at which the rays are absorbed by matter. Even after the
introduction of x-ray spectroscopy, absorption coefficients contin-
ued to be the primary test. In 1912 E. A. Owen introduced a
semiempirical relation to express the absorption coefficient ji for
homogeneous x-rays defined for the standard absorbing material,
aluminum.57 If the rays originate in an anticathode of atomic
weight A, then
pt oc A ~ 5 (9.3)
In 1914, using Moseley's data, Rutherford could reformulate

Owen's relation in terms of x-ray frequency:58
ju oc v~5'2 (9.4)
This was extended by Darwin and independently by Bragg and

Peirce to
// oc Z4 W2 (9.5)
57
Owen, PRS, 86A (1912), 426-39.
58
where Z is the atomic number of the x-ray source.59 Equation 9.5
became known in Britain as the Bragg-Peirce law, and it offered
some convertibility between spectroscopic and absorption results
for x-rays. Rutherford had used it to convert his measures of highly
penetrating x-rays to frequency in Figure 9.6. As Bragg said, "the
older method of defining the quality of x-rays must naturally give
place to the more quantitative and direct definition in terms of
wavelength."60 By this he did not mean that quality determination
by absorption had outlived its usefulness, only that the results
should be expressed in terms of x-ray wavelength. There were still
many in Britain who questioned whether spectroscopy really of-
fered a "more quantitative and direct" definition of x-ray quality.
There were good reasons for skepticism. The empirical relation
between x-ray wavelength and the coefficient of absorption, equa-
tion 9.5, represents only the general course of the function. It was
strongly suspected that resonant absorption of x-rays should
occur. The analogy to visible light suggested it; the fluorescent
stimulation of characteristic x-rays harmonized with it; and the
Kossel model of x-ray origin fairly demanded it. But before 1914
only very imprecise data existed on resonant x-ray absorption.
Barkla and Sadler had found in 1909 that the absorption coeffi-
cient for characteristic K x-rays from a given material is greatest
when the absorbing substance is the same as that producing the
rays.61 Manne Siegbahn, who by 1914 was developing a first-rate
laboratory for x-ray research in Sweden, pointed explicitly to the
close parallel to preferential optical absorption in 1914.62
As discussed in Chapter 8, the first detailed study of x-ray
absorption using spectroscopic methods was done by Ernst
Wagner in 1914. He found that absorption increased dramatically
at certain wavelengths that were characteristic of the absorbing
material. But unlike the optical case, in which absorption sharply
peaks at the resonant frequency, the absorption coefficient for
x-rays begins to rise some distance on the high-frequency side of
the actual resonant peak. Superimposed on the general equation
9.5 is a series of "edges" shown schematically in Figure 9.7. The
59
W. H. Bragg and Peirce, PM, 28 (1914), 626-30.
60
Ibid., 626. Albert Hull and Marion Rice showed in 1916 that the exponent 5/2
is closer to 3; PR, 8 (1916), 326-8.
61
Barkla and Sadler, PM, 18 (1909), 739-60.
62
Siegbahn, PZ, 15 (1914), 753-56.
Kedge
low v high v
Figure 9.7. Schematic representation of the superposition of x-ray

absorption edges on the law relating absorption coefficient to wave-
length.
edge of each tooth always exceeds by a small fraction the frequency

of the associated x-ray emission line. In their most primitive form,
these were the data that Walter Kossel had used to formulate the
idea that characteristic x-rays arise from electron transitions to the
innermost shells of the atom.
The existence of the strong x-ray absorption features raised
serious obstacles for tests of the x-ray photoeffect relation. If an
incident x-ray loses energy before it stimulates an electron to leave
the inner orbits, the direct quantum balance no longer holds. The
"work function" in the Einstein photoeffect law is no longer
negligible, nor is it constant for a given cathode. Electrons ejected
from different Bohr orbits will lack differing amounts of the inci-
dent x-ray energy. This was the reason for the deviations Ruther-
ford had found from the quantum transformation relation. To
keep an accurate balance and test the quantum relations, one had
to know the energy depth of the low-lying orbits. This information
could be obtained only through careful study of the x-ray absorp-
tion edges, a research field that was in its infancy at the outbreak of
World War I.
BARKLA'S NEGATIVE INFLUENCE
There was another way to study x-ray absorption, by the tradi-

tional means of x-ray penetrating power. Charles Barkla was the
acknowledged master of the technique in Britain. Already con-
vinced that characteristic x-rays form a line spectrum, in 1913 he
briefly tried to apply crystal diffraction methods.63 The trial rein-
forced his willingness to speak of true wavelengths for what had
formerly been, for him, x-ray impulses, but he did not long con-
tinue the new methods.64 After his appointment at Edinburgh in
1913, he began studies using absorption methods on the secondary
electrons released by x-rays.65 With a student, he measured the
range of the electrons produced by homogeneous x-rays incident
on metals, varying the pressure of the gas through which the
electrons passed until the ionization current across afixeddistance
reached saturation.66 The density of the gas then gave a rough
measure of the maximum energy of the electrons, and it could be
reduced to actual values of electron velocity by applying a different
"Whiddington's law" stating that the range of an electron is pro-
portional to the fourth power of its velocity.67
The analysis convinced Barkla that, unlike the case of the
optical photoeffect, the velocity of the electron does not depend at
all on the material from which it is ejected; it depends, he claimed,
only on the wavelength of the incident x-ray. He tested calcium,
zinc, and tin using K x-rays from both silver and tin, finding that
saturation occurred at the same gas pressure within 5 to 10 percent
for each of the three test materials. Thus, Barkla convinced himself
that no work function was subtracted from the energy donated by
the x-ray. Subsequent studies, undertaken in preparation for his
Bakerian Lecture to the Royal Society in 1916, led him to propose
that x-ray energy is not always absorbed in whole quanta.68 His
convoluted argument depended on comparison of the energy of
secondary K x-rays to the energies of secondary electrons excited at
63
Barkla and Martyn, PPSL, 25 (1913), 206-15.
64
For a general account, see Stephenson, AJP, 35 (1967), 140-52.
65
Barkla and Philpot, PM, 25 (1913), 832-56.
66
Barkla and Shearer, PM, 30 (1915), 745-53.
67
Whiddington, PRS, 85A (1911), 328. See also Beatty, PRS, 89A (1913),
314-27.
68
Barkla, Nature, 94 (1915), 671-2; 95 (1915), 7.
the same time by homogeneous primary x-rays. But his experi-
ments were able only to detect maximum energies, and rested on
his assumption that no energy is lost to a work function in the
ionization process.69
Barkla's researches were accorded more respect than they de-
served. Richardson tried in 1917 to disprove his contentions, first
by demonstrating that the conclusions did not follow from the data
and second by an appeal to the fact that the Kossel model of x-ray
emission forbids the appearance of the full x-ray quantum in the
ejected electron.70 But Barkla's reputation had been built on his
solid early work and was not easily dislodged. His claims against
the quantum theory and for the existence of a new high-frequency
characteristic x radiation, the J series, muddied the waters of x-ray
research between 1916 and 1923. Part of the difficulty was certainly
that measurement of secondary electron energy was limited to
gross methods sensitive only to the maximum electron velocity in
what was actually a full distribution of velocities.
Barkla's view was remarkably influential into the 1920s.71 Even
as Owen Richardson tried to disprove Barkla's antiquantum con-
tention, it affected research being done in Duane's laboratory at
Harvard. In 1918, with his student, Kang Fuh Hu, Duane com-
pared x-ray emission frequencies with their associated absorption
edges. Despite the Kossel-Sommerfeld model, there was no clear
understanding at Harvard of the differences to be expected be-
tween absorption and emission frequencies. They found that the
wavelength of the emission peak was uniformly 0.3 percent greater
than that for absorption. "It appears," they concluded, "that either
the critical absorption wavelength is about 1/3 percent shorter
than that of the [emitted] line, or else the wave lengths do not
correspond to the centers of the peak in the emission spectrum."72
Duane's student, Hu, tried to see whether /?-ray spectroscopy
would contribute to an understanding of the differences. Applying
69
Barkla, PTRS, 217A (1917), 315-60.
70
O. Richardson, PRS, 94A (1918), 269-80.
71
Barkla's contention that x-rays are not absorbed in entire quanta was the main
point discussed by Richardson, M. de Broglie, Rutherford, Ehrenfest, Lange-
vin, Brillouin, Lorentz, Bragg, Millikan, and Barkla at the Third Solvay
Congress in 1921. The long discussion followed a presentation by Maurice de
Broglie on the x-ray photoeffect. Solvay HI, pp. 101-30.
72
Duane and Hu, PR, 11 (1918), 490-1.
a constant potential to his rhodium anticathode x-ray tube, Hu
found that the same maximum electron velocity was elicited from
both silver and lead radiators.73 Two peaks appeared in electron
velocities from silver, each within 5 percent of that predicted by the
quantum relation 1/2 mv2 = hv for the Ka and Kfi x-ray emission
lines from silver. Moreover, the x-rays from silver, incident on tin,
released electrons of the velocity predicted by the quantum rela-
tion for Ka and Kfi x-rays from both silver and tin. But the
agreement was no closer than 5 or 6 percent; Hu reported that the
electron velocity was always smaller than that expected from the
quantum relation and the tube potential. "Experiments are still in
progress," he said, "to see if there is a real difference from the
simple equation 1/2 mv2 = hv, such as would be suggested by the
term '/?' in Einstein's form of the equation."74 Apparently the
experiments, if completed at all, were inconclusive, but Hu pub-
lished no more on the x-ray photorelations. Nor were the studies
carried further by others at Duane's laboratory, where /?-ray spec-
troscopy was not a standard tool for research.
ATTITUDES TOWARD THE LIGHTQUANTUM
Before 1921 the clearest evidence for spatial localization in radia-

tion came from Millikan's results for the photoelectric effect. But
this case of quantum absorption concerned visible light, in which
the localized properties were not strongly pronounced and the
resistance of orthodox wave theory was stiffest. Futhermore, Milli-
kan's data concerned only one point in the spectrum of photoelec-
tron velocities: the maximum velocity. In this period, Millikan
himself opposed the lightquantum hypothesis, calling it "reck-
less," "untenable," and even "erroneous."75 Despite his uncon-
tested proof of Einstein's linear photoeffect relation, Millikan
never doubted, until 1922, that an explanation of the result would
be found without the suspect lightquantum. His particularly
strong opposition may be attributed both to his respect for Euro-
pean mathematical physics - he had studied under Planck and
Nernst in Berlin - and to the tradition that his teacher, Albert
73
Hu, PR, 77(1918X505-7.
74
Ibid., 506.1 have corrected an obvious misprint in the equation.
75
Millikan, PR, 7(1916), 355; Electron (1917), 230.
Michelson, had developed in American precision measurement in
wave optics.76
Nor did William Duane adopt the lightquantum before 1922.
His research had concentrated on the production, not the absorp-
tion, of x-rays. It was in the latter that the paradoxes of radiation
were most strongly expressed. Duane's work contributed to the
wider acceptance of the quantum relation, but it sidestepped the
issues that forced consideration of the lightquantum. Duane was
able to separate his empirical x-ray frequency relationship entirely
from questions regarding the energy transferred in interactions; in
1916 he derived the Planck black-body law without, he claimed,
any assumption that energy and frequency are related.77 Because
the entire kinetic energy of an individual electron is not necessarily
radiated as an x-ray quantum, he felt that the energy in the rays
could be measured only through appeal to empirical relations such
as Whiddington's fourth-power law.
"We are not therefore compelled to explain the emission of
radiation in quanta /JV," Duane said; one had only to explain how
an electron with a certain kinetic energy "can produce radiation of
frequency up to, but not greater than, that given by [the quantum
condition]."78 It was a curious evasion. Following Rutherford,
Duane offered a qualitative mechanism: High-frequency x-ray
vibrations are associated only with the innermost electrons of an
atom, to which only the highest-velocity electrons can penetrate.
Thus, in 1918, Duane himself was blind to the message that
emerged from his student Hu's /?-ray velocity studies. The work
was not continued at Harvard, and Duane nowhere, to my knowl-
edge, made any attempt to reconcile Hu's preliminary results with
his own proposed dissociation of frequency and energy.79
Two facts may be discerned about the experiments carried out
76
Millikan had studied under Planck and Nernst in Berlin and had heard
lectures by Poincare in Paris. He sometimes published major results in Ger-
man journals, including six in the period 1912-17. For the influence of
Michelson, see Millikan, Autobiography (1950), 66.
77
Duane, PR, 7(1916), 143-7.
78
Duane, Science, 46(1917), 348.
79
While Duane had worked in Paris with the Curies, he had collaborated with
Jean Danysz, one of the early developers of ^-electron spectroscopy. Duane
may have just ended that collaboration when Danysz began the electron
analysis that Duane never utilized.
on the quantum transformation relations before 1921. The first is
that the most important work was done outside Britain and Ger-
many. Although important background work had been accom-
plished in Rutherford's Manchester laboratory, the intervention of
the war effectively stopped this research. Moreover, in Britain,
Barkla refused to believe that absorption of x-ray energy occurs in
quanta, and his views were influential. Richardson strongly ques-
tioned Barkla's position but was nontheless convinced that some
interpretation of the photoeffect, perhaps his own, would solve the
problems without the "restricted and doubtful hypothesis used by
Einstein."80 J. J. Thomson was never able to accept the quantum
theory. Partly in response to his dissatisfaction with the new
direction of physics, he relinquished his Cavendish Professorship
to Rutherford in 1918. There were few serious empirical studies on
x-ray or y-ray absorption in Germany at all before 1921. The
significant empirical work came from America and, as we shall see
in the next chapter, from France.
The second thing to note is that experimentalists were no more
receptive to spatially localized lightquanta than were their more
mathematically inclined colleagues. It might be supposed that the
evidence for quantum absorption would convince empirically
minded physicists that some radical revision of the wave nature of
light would have to be found. But as we have seen, this was not the
case. Empiricists were no less demanding than other physicists that
the well-verified evidence for interference and diffraction of radia-
tion be granted priority. In any event, the evidence for quantum
absorption of x-rays and y-rays, where the localized properties of
radiant energy are most pronounced, was sparse before 1921. But
in the first months of that year, this circumstance changed
abruptly, an event to which we turn in the next and thefinalpart of
this study.
80
PART V
The conceptual origins of wave-particle
dualism
1921-1925
10
Synthesis of matter and light
A group of electrons that traverses a sufficiently small

aperture will exhibit diffraction effects.1
It is not entirely surprising that the earliest reaffirmation of the

dualism inherent in radiation came neither from Britain nor
Germany. There the issues had been recognized, discussed, and
dismissed over the preceding twenty years. And American re-
search traditions were largely derivative of German and British
models; Duane and Millikan had both studied under Planck, and
Richardson did research with Thomson in Cambridge. It took an
outsider to these traditions to raise the issues in a way that began to
convince others.
That process began in France. Maurice and Louis de Broglie
were, by any standards, unusual physicists. Amateurs in an age of
professionals, possessing a confidence born of noble status, work-
ing on experiments that had only the meagerest precedent in their
homeland, they were not as firmly bound by the predispositions
that prevented physicists of other nations from directly assessing
the spatially localized lightquantum.
Their conversion was real. Maurice had once been convinced
that x-rays and y-rays fit neatly into the electromagnetic spec-
trum.2 In his discussion, La nature des rayons de Rontgen in 1915,
there was no thought that x-rays might be anything other than
periodic waves. As late as May 1920, he spoke freely of x-ray
"wavelengths" and discussed the optical fluorescence that results
from degradation of x-ray "frequency."3 But late in 1920 he re-
viewed the new evidence accumulated at high x-ray frequencies.
There he hinted at unspecified difficulties in the orthodox wave
1
L. de Broglie, CR, 777(1923), 549.
2
M. de Broglie, Revue scientifique, 53 (1915), 581.
3
M. de Broglie, SFP PVRC, 1920 (1921), 33.
264 Conceptual origins of wave-particle dualism, 1921 -1925
interpretation of x-rays. In order to establish the wave nature of
x-rays on as "solid" a basis as that of light, he said that one now had
to ask whether "old ideas about periodic radiation propagating as
waves" might not be "subject to serious revision."4 What he had in
mind became clearer just a few months later, early in 1921, when
he forcefully revived the issue of the paradoxical properties of
radiant energy.
CORPUSCULAR SPECTRA
Maurice de Broglie was the eldest son in a family traditionally
prominent in French politics. In 1906 he assumed the hereditary
title "Due de Broglie," and at about the same time he outfitted a
private laboratory in his Parisian residence.5 Although educated in
physics at the College de France, he was a scientific amateur, a
species common in the nineteenth century but increasingly rare in
the twentieth. In 1908, when he defended his theses following work
with Paul Langevin, his research concerned ionization of gases and
Brownian motion.6 He soon took an interest, then uncommon for
a Frenchman, in the new quantum theory. He was secretary to the
First Solvay Congress in 1911 and, with Langevin, edited its pro-
ceedings.7 With the discovery of x-ray interference in 1912, he
reorganized his research plans. The following year he made a
notable discovery: The slow rotation of a crystal greatly reduces
the spurious x-ray interference maxima caused by periodic imper-
fections in the crystal lattice.8 The rotation smooths out what
Moseley had called the "horrid diffraction fringes" caused by
natural ruling errors in the crystal, which greatly confused early
crystallographic analysis.9 The new technique was an important
step forward and was, for example, indispensable to Rutherford in
his 1914 measurement of what he thought were y-ray wavelengths.
De Broglie did some of the most significant experimental re-
4
M. de Broglie, Scientia, 27 (1920), 110-11.
5
Weill-Brunschvicg and Heilbron, DSB, 2 (1973), 487-9. La Varende, Les
Broglie (1950), 265-320. The laboratory was first established in de Broglie's
apartments in rue Chateaubriand; later, another was set up in rue Lord Byron a
block away in the 8th arrondissement.
6
M. de Broglie, Theses (1908).
7
Solvay I.
8
M. de Broglie, CR, 157(1913), 924-6. See also Heilbron, Ms, 5^(1967), 482-4.
9
Moseley to his sister, 13 August [1913]. Heilbron, Moseley (1974), 207-8.
Synthesis of matter and light 265
search on x-ray absorption edges. In 1914, on the suggestion of
Bragg and Siegbahn, he successfully interpreted anomalous ab-
sorption bands recorded on a photoplate for x-rays from silver as if
they were due to preferential absorption of the x-rays by the silver
halide emulsion.10 There soon developed a lively race among
Siegbahn, Wagner, and de Broglie to measure the precise frequen-
cies of the absorption edges. Wagner found that there were two L
edges.11 In confirming this result in 1916, de Broglie found a
third.12 At war's end he dedicated his research to x-ray absorption;
the frequencies of the edges gave the most direct measurement of
the energy levels of the Bohr atom. The discovery of three L edges,
five M edges, and so on directly shaped ideas about the electronic
structure of the atom as interpreted by Bohr and Sommerfeld. De
Broglie had entered a front-rank research field and possessed the
skill and facilities to make significant contributions.
By January 1921, de Broglie realized the potential value of /?-ray
spectroscopy to his research. He had noted in 1920 that x-rays
stimulate secondary ^-electrons,13 but now recognized that this
held the key to a study of energy transformation in the absorption
of x-rays. His inspiration likely had much to do with Alexandre
Dauvillier, who joined de Broglie's research group late in 1920.
Dauvillier's research had already moved from studies of x-ray
absorption to an examination of the kinetic energies of the second-
ary electrons released by x-rays.14 He began to investigate the
chemical effects of electrons on photoemulsions and carefully
studied the work by Barkla in which the electron range was used to
fix electron energy. Early in 1921, Dauvillier completed the most
thorough review in the French language on the photoelectric
effect, with the explicit intention to apply that knowledge to the
case of x-rays. "The fastest electron emitted by a body exposed to
[x-rays] possesses precisely the same velocity as the fastest electron
in the cathode beam that produces the x-rays," he claimed.15
Dauvillier was also perplexed that the energy granted an electron
by x-rays is independent of the distance over which the rays have
10
M. de Broglie, CR, 158 (1914), 1493-5-
11
Wagner, AP, 46 (1915), 868-92.
12
M. de Broglie, CR, 163 (1916), 352-5.
13
M. de Broglie, SFP PVRC, 1920 (1921), 33-
14
Dauvillier, Theses (1920).
15
Ledoux-Lebard and Dauvillier, Physique des rayons x (1921), 208.
traveled.16 "The electromagnetic and corpuscular theories seem to
be very difficult to reconcile," he said. "The problem of the nature
of radiation remains unresolved."17
In January 1921 a South African physicist, encouraged by Owen
Richardson, published a study that further directed the Parisian
physicists' attention to these problems.18 Lewis Simons, like Rich-
ardson, was convinced that Barkla's antiquantum claims were
simply wrong: A single homogeneous x-ray beam gives rise to an
entire spectrum of electron velocities, and different amounts of
energy are lost to each electron depending on its original position
in the atom. Simons worked hard, but unsuccessfully, to extract
convincing evidence for this claim from ionization and electron-
range data. About ^-ray spectroscopy, the suggestive early work by
Robinson and Rawlinson, and the incomplete experiments by Hu,
Simons had apparently not heard.
Maurice de Broglie had the great advantage of knowing all of
these prior experiments. But he was skeptical of the claim that
x-ray energy is always absorbed by electrons in integral multiples
hv. When hefirsttook up the detailed evaluation of electron energy
in February 1921, he thought he had found evidence that electrons
leave the atom with quanta of energy hv and therefore must have
absorbed more than that quantized amount from the x-rays.19 His
studies of x-ray absorption edges and his understanding of the
Bohr atom made it clear that some energy must be spent to remove
each electron from the atom. This seemed a significant issue to
clarify for presentation at the Third Solvay Congress, to which he
had been invited and which would meet in two months' time.
Besides, a recent disappointment in his attempt to enter the Paris
Academy of Sciences made a certain broadening of his experimen-
tal inquiry attractive.20
And so, Maurice de Broglie, intrigued by the paradox that
16
Ibid., 399.
17
Ibid., 407-8.
18
Simons, PM, 41 (1921), 120-40. De Broglie knew of Simons' paper, see Solvay
III, p. 85n.
19
M. de Broglie, SFP PVRC, 1921 (1922), 10.
20
On 22 November 1920, Maurice de Broglie was nominated to replace Adolphe
Carnot in the Academie des Sciences. On the 29th, after four votes, Jules Louis
Breton was elected, de Broglie having received only one of about sixty votes
cast in each round. CR, 171 (1921), 1031, 1045. One of his biographers reports
that Maurice de Broglie was later elected on the basis of his studies on the
photoeffect. Sudre, Revue des deux mondes (July-Aug. i960), 577-82.
Figure 10.1. Maurice de Broglie's x-ray photoeffect velocity spectro-

meter. [Rayons x (1922), 142.]
Dauvillier had proposed, and confident he could overcome

Simons' methodological difficulties, recognized that progress in
quantum theory might follow from use of the electron velocity
apparatus built by Rutherford, Robinson, and Rawlinson. He
quickly adapted and modified the Manchester /?-ray velocity spec-
trometer and employed it to analyze x-ray absorption using what
he called spectres corpusculaires, or particle velocity spectra.21 In
the apparatus shown in Figure IO.I x-rays strike the sample ele-
ment placed at C. Secondary x-rays are unaffected by the magnetic
field that is imposed in a direction normal to the plane of the page,
but secondary electrons are deflected into circular paths, and thus
sorted by velocity, forming characteristic patterns of bands - the
corpuscular spectra - on the photoplate PPr.
In collaboration with his younger brother, Louis, Maurice de
Broglie immediately formulated the goals of his research in terms
of the Bohr model of the atom.22 By April 1921, they agreed with
Simons and Richardson that Barkla was wrong. Each electron in
21
M. de Broglie, CR, 172 (1921), 274-5, 527-9.
22
M. and L. de Broglie, CR, 172 (1921), 746-8. L. de Broglie and Dauvillier, CR,
772(1921), 1650-3; CR, 173(i92i)9 137-9.
an atom, granted energy hv by the incident monochromatic x-ray
beam, must spend a certain amount of that energy to escape the
atom. The amount was always just the "depth" of the orbit from
which the electron came, and those depths had been carefully
charted in de Broglie's private laboratory in rue Chateaubriand. A
monochromatic beam of frequency v will therefore produce elec-
trons of energy hv hvK, hv hvL, and so forth, where vK and vL
are the measured frequencies of the K and L absorption edges for
the cathode material. Measurement of the kinetic energy of the
released electrons therefore yields directly the energy level of the
vacated Bohr orbit. In this sense, the x-ray photoeffect law could
not be independently verified, but the systematic deviations from
the quantum relation could be explained by an appeal to the
growing understanding of the electronic structure of the atom.
De Broglie began these studies in order to shed light on the atom
and to verify his previous x-ray absorption results. 'There is no
doubt," de Broglie wrote to his friend, Frederick Lindemann,
"that radiation of frequency v expels from an element R groups of
electrons with energies corresponding to
hv energy level K
hv energy level L
[and so on]. One can therefore find the energy levels of the
secondary radiator using an [analyzing] crystal as intermediary."23
He was impressed by the clear evidence that each feature in the
x-ray spectrum is mirrored by a set of well-defined electron veloc-
ity bands in the corpuscular spectrum. "One recovers all the
properties of the x-ray spectrum," he pointed out, "including lines
and bands."24 In March 1921 he had corroborated this result in ten
elements, ranging from copper (29) to ytterbium (70).25 For exam-
ple, x-rays from a rhodium anticathode incident on a copper plate
give rise to seven distinct electron velocity bands. Three corre-
spond to the Ka and Kfi x-rays from copper, each diminished by the
absorption energies hvL and hvM for copper. What would be four
lines only appears as three because the frequency difference
Kp Ka is equal to hvL hvM, as shown in Figure 10.2. The
23
M. de Broglie to Lindemann, 12 April 1921. Cherwell papers.
24
M. de Broglie, CR, 172 (1921), 275.
25
Ibid., 806-7.
Figure 10.2. Schematic representation of the inner orbits of the copper

atom to explain Maurice de Broglie's x-ray photoelectric measure-
ments.
electron kinetic energy that remains after a Kp x-ray ionizes an

L-level electron is Kfi hvL. This is identical to that from the
process Ka hvM. So, only three lines appear. The other four
bands correspond to the Ka and Kp x-rays from rhodium, each
diminished by hvL and hvM for copper.26
By April 1921, when he reported these results to the Third
Solvay Congress, de Broglie was also aware of their significance for
the photoelectric relations.27 Millikan had been able to compare
optical frequencies only to the maximum electron velocity; de
Broglie could offer corroboration at points within the electron
distribution as well. Thus, after introducing corpuscular spectra as
a means to analyze the atom, he proceeded to a general discussion
of the photoelectric effect. Assuming the validity of the Bohr-
Kossel model for x-ray emission, it was clear that the full energy
quantum of the x-ray is transferred to individual electrons. Cou-
pled with the evidence that each x-ray frequency gives rise to a
particular set of electron velocities, it became virtually impossible
for de Broglie to avoid the issue of the lightquantum. On several
occasions in his presentation, he alluded to the "remarkable"
behavior of the absorption process when the energy of secondary
electrons is found to be the same as that of the entire spherical
26
M. de Broglie, JP, 2 (1921), 265-87; SFP PVRC, 1921 (1922), 49.
27
M. de Broglie, "La relation hv = E dans les phenomenes photoelectriques;
production de la lumiere dans le choc des atomes par les electrons et produc-
tions des rayons de Rontgen," p. 12. Typescript draft of de Broglie's contribu-
tion to the Solvay proceedings. Solvay archives.
x-ray wave.28 The radiation "must be corpuscular," he said, "or, if
it is undulatory, its energy must be concentrated in points on the
surface of the wave."29
C. D. ELLIS AND NUCLEAR y-RAYS
Transformation experiments on y-rays are limited to absorption

phenomena. There was no way to stimulate the emission of y-rays;
they are a by-product of radioactive decay. The same difficulties
that prevented analysis of the x-ray photorelation until the early
1920s also hampered research on the absorption of y-rays. In
addition, studies were complicated by the problems of directly
measuring y-ray wavelengths. Rutherford thought he had suc-
ceeded in 1914 in measuring wavelengths for a few homogeneous
y-ray groups, but, as we have explained, these were characteristic
x-rays. The experimental difficulties of true y-ray wavelength de-
termination were immense. The next direct measures were at-
tempted only in 1919 and were not successful until 1924.30
The work done on y-rays in the meantime relied almost entirely
on absorption coefficient measurements. A series of studies by
Alexander Russell and Frederick Soddy beginning in 1909 was
based on prior work by H. W. Schmidt on the absorption coeffi-
cients of y-rays from uranium X, radium C, thorium, and actin-
ium.31 As discussed in Chapter 8, these experiments led to Ruther-
ford's first interest in homogeneous groupings of y-rays and
encouraged his hypothesis that the characteristic y-rays originate
in the vibration of special groups of electrons near the center of the
atom. His students, Robinson and Rawlinson, applied /?-ray spec-
troscopy in 1914, seeking data on the electrons released by homo-
geneous y-rays from matter.32 For the same incident radiation,
electrons released from gold (79) were found to be consistently 1 or
2 percent faster than those released from lead (82). To an extent
not then found for the optical or x-ray case, photoelectron veloci-
ties due to y-rays appeared to depend markedly on the material
from which the electrons are emitted.
28
Ibid., 1, 12; SolvayM, pp. 80, 85, 89.
29
Ibid., 13; Solvaylll, p. 89.
30
Thibaud, Theses (1925).
31
F. and W. Soddy and Russell, PM, 19 (1910), 725-57. Russell, PRS, 86A
(1911X240-53.
32
Rutherford, Robinson, and Rawlinson, PM, 28 (1914), 281-6.
During the war, research on these topics virtually stopped in
Britain, the only country where it had been pursued at all. In 1917
Rutherford reassessed the few "y-ray" wavelengths he had mea-
sured before the war33 in light of new studies by Rice and Hull in
America on the compass of the presumed relation between ab-
sorption coefficient and x-ray frequency.34 Formerly, equation 9.5
or its equivalent had been shown to hold for X > 6 X io~9 cm; now
it appeared to be valid down to icr9 cm. The overlap that Ruther-
ford had found between x-rays produced at high potentials and the
secondary x-rays he believed to be y-rays was here independently
corroborated. In some cases, x-ray wavelengths derived from ab-
sorption behavior were smaller than the shortest wavelength
(0.072 A) found for "y-rays" by Rutherford.
If the smooth transition between x-rays and y-rays was to be
believed, and Rutherford did not yet doubt that it was, the major-
ity of y-rays correspond to x-rays produced by potentials greater
than a half million volts. "It would thus appear probable," Ruth-
erford concluded, "that the absorbed groups of y-rays are due to
the conversion of the energy, E = hv, of a wave of frequency v into
electronic form and that consequently the energy of the /?-ray
groups may be utilized by the quantum relation to determine the
wavelengths of the penetrating gamma rays."35 Thus, the photo-
electric relation, proven rigorously for the optical case and inferred
on the basis of the Bohr atom for x-rays, was simply assumed to
hold for y-rays. By this means, wavelength measurements could be
extended into the highest-frequency region despite ignorance of
absorption behavior.
The man who put this program into practice, within two months
of de Broglie's first papers on corpuscular spectra, was Ruther-
ford's student at Cambridge, Charles Ellis. Even before the 1921
Solvay meeting, Rutherford had sent Ellis' unpublished results to
de Broglie.36 At the congress, still before Ellis' study had appeared
in print, Rutherford announced that Ellis' results for y-rays were in
full agreement with those of de Broglie.37 But rather than empha-
33
Rutherford, PM, 34 (1917), 153-62.
34
Hull and Rice, PR, 8 (1916), 326-8.
35
Rutherford, PM, 34 (1917), 161.
36
M. de Broglie, JP, 2 (1921), 26911.
37
Rutherford, Solvay III, pp. 107-9. Ellis' first study was submitted to the Royal
Society of London on 22 April 1921, sixteen days after the end of the Solvay
Congress.
size, as de Broglie had, the significance of the result for energy
localization in the radiation, Ellis used his result to inquire about
atomic structure. By comparing the frequency of the incident y-ray
with the resulting electron velocity, "it should be possible to
deduce," Ellis said, "from what part of the atom the electron
originated."38 One can here sense the growing uncertainty at Cam-
bridge about Rutherford's model for characteristic y-rays.
Ellis used as the source a small tube of radium emanation, which
emitted the known y-ray spectrum from radium B and radium C.
The material to be tested was wrapped as a thin foil tightly around
the tube, and the resulting electrons dispersed by velocity differ-
ences in a strong magnetic field. He submitted uranium (92), lead
(82), platinum (78), tungsten (74), and barium (56) to the rays, and
in each case found three well-defined velocity peaks in the electron
velocity spectrum. He assumed that the three were due to three
frequencies of y-ray, which he denoted by energies Wl9 W2, and
W3. In each case, the observed electron energy E should be one of
the unknown y-ray energies minus an energy Wn characteristic of
the atom under test.
Because the Wn values varied widely for the same presumed
y-ray energy, Ellis concluded that the released electrons had to
come from the innermost rings of the atom. For example, the
slowest electrons of the three groups from tungsten possess
8,000 eV more kinetic energy than the equivalent electrons stimu-
lated in platinum. Using the most recent values of the K absorp-
tion edges for the test elements, Ellis could calculate the actual
values of the y-ray energies W1, W2, and W3. The results were
remarkably consistent for the five elements for which data was
available, as shown in Table 10.1. The agreement across horizontal
rows elicited Ellis' conclusion that the data were "in complete
agreement with Rutherford's theory of the connection between the
/?-ray and y-ray emission of radioactive atoms."39
Rutherford's theory was at the moment in a state of flux. Until
the early 1920s, he had proposed special groupings of electrons in
the superstructure of the atom. These were supposed to resonate to
the harmonies of a relativistically speeding ^-electron and then
emit a y-ray frequency characteristic of the atom. With the re-
sumption of research in Cambridge after the war, it was clear that
38
Ellis, PRS, 99A (1921), 261.
39
Ibid, 265.
Table 10.1. Ellis' y-ray photoelectric derivation of
three y-ray energies, Wn, incident on five elements
(X10 5 eV)
Element
Ba W Pt Pb U
wx 2.35 2.36 2.38 2.40

w2 2.9 2.89 2.91 2.91 2.92
w3 3.46 3.46 3.46 3.48
the Bohr theory allowed no special groupings of the sort earlier

imagined by Rutherford. The book that most stimulated Ellis's
analysis was Sommerfeld's influential Atombau und Spektral lin-
ien, the first edition of which had appeared in 1919. Using the
results Sommerfeld had assembled about x-ray absorption, Ellis
saw a way to justify his assumption that radium B, an isotope of
lead (82), emits electrons with velocities determined photoelectric-
ally by six well-defined frequencies of y-ray. He subtracted, in turn,
the energy of the A: shell and of Sommerfeld's hypothetical A' shell
from each of the assumed y-ray energies, and twelve of the faster
electron velocity lines emerged, accurate to less than 1 percent.
"All the lines are given without any gaps and no extra ones are
predicted," he remarked with evident pride.40
Thereafter, /?-ray spectroscopic results alone could be taken to
provide data on the energy of y-rays, an extrapolation de Broglie
had already made for x-rays.41 Using the few known y-ray wave-
lengths, Ellis calibrated a scale that went far below the 0.07 A limit
then set by diffraction techniques. His full report was received by
the Royal Society in February 1922; in it he charted the full
spectrum of true y-ray wavelengths from radium B, radium C,
thorium C, and thorium D - some twenty lines in all. "The main
y-rays from radium B correspond to energies 2.4 [to] 4.0 x io5
[electron] volts," he said. If these rays "owe their origin to a
transition between stationary electron states in the atom," then the
final electron orbit had to be "very deep in the atom and certainly
40
Ibid., 267.
41
M. de Broglie, CR, 172 (1921), 806-7; 173 O921), 1157-60. Meitner, ZP, 9
(1922), 131-44, 145-52. M. de Broglie and Cabrera, CR, 176 (1923), 295-6.
inside the K ring."42 James Chadwick had just repeated Ruther-
ford's classic a-particle scattering experiments.43 He placed partic-
ular emphasis on the region between the nucleus and the K shell
and found no evidence that any electrons were there. Ellis there-
fore concluded early in 1922 that "y-rays must come from the
nucleus."44
LES FRERES DE BROGLIE

Louis de Broglie developed intellectually under the influence of his
elder brother, Maurice. When their father died in 1906, Maurice
was thirty-one and Louis fourteen, and the elder took responsibil-
ity for the younger's education.45 "At every stage of my life and
career," Louis later told his brother, "I found you near me as guide
and support."46 After completing a licence in history in 1910, Louis
prepared for another in physics under Leopold Brizard, the same
physicist who had first taught Maurice.47 And Maurice saw to it
that Louis had an advance opportunity to study the proceedings of
the First Solvay Congress in 1911, of which Maurice was a secre-
tary.48 Louis later recalled that exposure to these discussions en-
couraged him "to determine the true nature of the mysterious
quanta that Max Planck had introduced into theoretical phys-
ics."49
After service as radio telegraph operator on the Eiffel Tower
during World War I, Louis began to work in physics. He was
drawn to problems that arose directly out of his brother's re-
search, at first an interpretation of Maurice's x-ray absorption
results according to the Bohr atom.50 With Maurice and, later,
others at his laboratory - Alexandre Dauvillier, Jean Thibaud,
Jean Trillat - Louis published a series of papers showing the
congruence between x-ray absorption and optical absorption spec-
42
Ellis, PRS, 101A (1922), 13.
43
Chadwick, PM, 40(1920), 734-46.
44
Ellis, PRS, 101A (1922), 14. Whiddington, PM, 43 (1922), 1116-26.
45
M. de Broglie in Louis de Broglie (1953), 423-9.
46
La Varende, Les Broglie (1950), 306.
47
M. de Broglie, Grandes souvenirs, belles actualites: Le recueil du jeunes, 2
(July 1948), 4-6. See also L. de Broglie Recherches d'un demi-siecle (1976),
7-16.
48
Solvay I.
49
L. de Broglie in Louis de Broglie (1953), 458.
50
L. de Broglie, SFP PVRC, 1921 (1922) 15-16.
51
tra. These studies, begun in 1920, directed Louis' attention to the
structure of the atom, not to the structure of radiant energy. Like
most physicists of the time, he thought x-rays to be a form of
high-frequency light wave.
The central problem in the application of x-ray spectra to
questions of atomic structure was then to discover the number of
electrons that occupy each of the allowed Bohr orbits.52 Arnold
Sommerfeld had gone furthest with a relativistic explanation for
the appearance in x-ray spectra of doublets, closely spaced pairs of
x-ray emission lines. The separation in frequency of the pair varies
in a regular manner with the atomic number of the parent atom.
Sommerfeld had explained this as due to a relativistic effect of the
rapidly rotating electrons closest to the atomic core.53 Louis de
Broglie therefore confronted the successful application within his
own research domain of Einstein's theory of relativity, a theory
that, he said, had fascinated him since his early student days.54
Although the younger de Broglie's interests in 1920 centered on
atomic theory, within a year they had changed significantly.
Maurice's experiments showed with increasing clarity in 1921 that
an x-ray transfers its entire quantum of energy to individual
electrons, and Louis began trying to find an explanation. Years
later, he recalled the influence that his brother's research had had
on him. It "directed my attention to the quality of waves and of
x-rays and y-rays," he said.55 "We debated the most pressing and
baffling questions of the time, in particular the interpretation of
the results of your experiments on the x-ray photoeffect," he told
his brother. 'The insistence with which you directed my attention
to the importance and the undeniable accuracy of the dual particu-
late and wave properties of radiation little by little redirected my
thought."56 In 1921, following his brother's report to the Solvay
51
L. and M. de Broglie, CR, 172 (1921), 746-8; 173 (1921), 527-9. L. de Broglie,
CR, 173 (1921), 1160-2,1456-8. L. de Broglie and Dauvillier, CR, 775 (1922),
685-8, 755-6. L. de Broglie, JP, 3 (1922), 33-45.
52
His second thesis for the doctorate in 1924 was on the closely related topic "les
proprietes chimiques des elements interpreters par leur structure electroni-
que."
53
Sommerfeld and Wentzel, ZP, 7(1922), 86-92. For de Broglie's reaction, see
his contribution in Louis de Broglie (1953), 459-60.
54
L. de Broglie, Nouvelles perspectives en microphysique (1956), 232.
55
La Varende, Les Broglie (1950), 306.
56
L. de Broglie, Savants et decouverts (1951), 302. Trillat, Societe Francais de
mineralogie et decristallographie Bulletin, 83 (i960), 239-41.
Congress, Louis turned his attention to the counterintuitive be-
havior of x-rays that, in the words of his brother, concentrates
energy "in points on the surface of the wave."57
Louis reviewed all of Einstein's publications on the nature of
light. He was now convinced, like his brother, that a synthetic
theory combining wave and particulate representations was re-
quired,58 but he confronted the formidable difficulty that Einstein
had yet been unable to resolve: a physical interpretation combin-
ing contradictory wave and particle properties. In November 1921
he presented a sketch of how he expected this synthesis might be
attained: A complete theory of radiation should be built in a
manner parallel to that of hydrodynamics, wherein wavelike laws
govern the macroscopic behavior of large numbers of water mole-
cules.59
In late 1921 he applied rudimentary relativity theory to "atoms
of light" in order to analyze their statistical properties "without
any intervention of electromagnetism."60 Einstein had already
treated an association of lightquanta as if subject to the laws of
kinetic theory. But de Broglie took lightquanta to be literally
"atoms of light," and unlike Einstein he defined a mass for them
using the quantum relations and relativistic dynamics. Light-
quanta, "supposed [to be] of equal very small mass," move at
velocities dependent on their frequency, but in all cases very near
to c.61 The effective mass m is the sum of that due to the kinetic
energy and a "proper" mass rao. Because mo for an atom of light
must be "infinitely small," the approximation m = hvjc2 will
suffice, but this is only an approximation. Using it, de Broglie
derived the Wien radiation law and showed, inter alia, that a gas of
lightquanta should exert only half the radiation pressure expected
using the classical corpuscular theory of light.
Several matters served in 1922 and 1923 to reinforce Louis de
Broglie's interest in the lightquantum and its eventual reconcilia-
tion with the wave theory of light. Maurice completed his written
contribution to the Solvay Congress proceedings; the revision,
finished early in 1922, pointed up in even greater detail the di-
57
M. de Broglie, draft of Solvay III presentation, p. 13. Solvay archives. Solvay
III, p. 89.
58
L. de Broglie in Louis de Broglie (1953), 4 6 -
59
L. de Broglie, JP, 3 (1922), 33-45; CR, 775(1922X811-13.
60
L. de Broglie, JP, 3 (1922), 422.
61
Ibid, 422n, 428.
lemma of waves and particles. At the time of Maurice's verbal
presentation in April 1921, it was clear only that the incident x-ray
frequencies are replicated in the electron spectra.62 Taking up the
subject again in November 1921, de Broglie established "several
more recent results," among them the finding that electron veloc-
ity lines appear even for stimulation of the radiator by its own
characteristic x-rays.63 That is, not only will the incident x-rays
invoke a series of scattered electron velocity lines, each decreased
in energy by the internal energy losses in the atom of the scattering
substance, but they will also give rise to a series of velocity lines
from any of the substance's ownfluorescentx-ray lines, stimulated
by the incident rays, again diminished by the energy levels of the
atom.64 And de Broglie restated the evidence for the particulate
nature of x-rays. "Today," he said, "our attention is again directed
to a set of phenomena that suggest the new idea of [quantum]
emission." These results "present facts in such a way that their
quantities are sometimes described in terms of the wave theory,
sometimes in terms of the emission theory [of light]."65
The evidence for this dualistic interpretation was in fact little
different from that recognized in 1907 by Bragg, and the elder de
Broglie pointed to it in some remarques at the end of his 1922 book
on the physics of x-rays. When a spherical x-ray wave passes
through a gas, two unaccountable things happen: (1) only a very
small fraction of gas molecules is ionized, releasing a fast elec-
tron; (2) the kinetic energy of an electron so freed is many orders of
magnitude greater than the total radiant energy in the small solid
angle of the wave that passes over the electron.66 Maurice de
Broglie thus forcefully revived interest in the paradoxes of quality
and quantity; now they were considerably strengthened by the
compelling evidence for quantum absorption of both x-rays and
y-rays.
WAXING INTEREST IN THE LIGHTQUANTUM

Maurice de Broglie's may have been the loudest voice in Louis de
Broglie's ear, but his was not the only voice. In the year following
62
M. de Broglie, draft of Solvay III presentation. Solvay archives.
63
M. de Broglie to Lindemann, 5 November 1921. Cherwell papers.
64
M. de Broglie, Solvay III, 80-100.
65
M. de Broglie, Les rayons x (1922), 17, 19.
66
Ibid., 156-9. M. and L. de Broglie, CR, 175 (1922), 1139-41-
Maurice's presentation of evidence for the spatial concentration of
radiant energy, several leading physicists in Germany and England
first gave serious consideration to Einstein's claims. Just five years
before, in 1916, Einstein had restated his position. In the second
part of a study now remembered only for its statistical treatment of
emission and absorption of light, Einstein claimed that each light-
quantum must possess a linear momentum of magnitude hv/c.67
But he did not then stimulate the interest he thought appropriate
for the hypothesis.
Planck delivered a talk on the "Nature of light" to the Kaiser
Wilhelm Gesellschaft in 1919 in which he surveyed the difficulties
that were then facing radiation theory.68 Planck was still far from
adopting the lightquantum and did not even discuss the hypothesis
directly. Instead, he suggested that the triggering hypothesis might
yet explain the photoeffect: According to relativity theory, each
gram of matter constitutes 2 x 1010 kilocalories of energy. The
mass of even the lightest atom is enough to account for the ejection
of several electrons with the observed velocities. Planck said,
however, that the quantum nature of light was the "foremost and
most difficult dilemma in quantum theory,"69 and once again
relativity theory was invoked as a factor in its resolution.
By late 1921, opposition to the lightquantum began to wane.
W. H. Bragg pointed to the "considerable new information" that
suggested localized x-ray quanta.70 He did not say so explicitly, but
the information he had in mind was certainly the results de Broglie
and Ellis had announced only a few months before. Einstein
proposed an experimental test that he thought might distinguish
between a wave and a lightquantum interpretation of Doppler
shifts in the frequency of light radiated from moving ions.71 At
first, the experimental results of Hans Geiger and Walther Bothe
seemed to support the lightquantum explanation.72 But Paul
Ehrenfest soon pointed out an error in Einstein's analysis that
rendered the result inconclusive.73
67
Einstein, Zurich Mitteilungen, 18 (1916), 47-62; PZ, 18(1917), 121-8.
68
Planck, Wesen des Lichts (1920), given 28 October 1919.
69
Ibid., 22.
70
W . H . Bragg, Nature, 107 (1921), 374.
71
Einstein, Berlin Sb (1921), 882-3.
72
Geiger and Bothe's results were not published. See Einstein, Berlin Sb (1922),
22.
73
See Klein, HSPS, 2 (1970), 1-39.
Far more remarkable than the proposed experiment was the
reaction to it. Unlike everything else Einstein had written on
lightquanta - a hypothesis taken seven years before as an example
of how an otherwise great mind can err - his work in 1921 sparked
lengthy discussion by a number of leading mathematical physi-
cists.74 Max Laue, who had recently begun to worry about proper-
ties of radiation "not yet explained by wave theory," engaged
Einstein in a spirited debate before the Berlin Academy.75 Ehren-
fest initiated a long correspondence on the significance of the
proposed experiment.76 Gregory Breit published a detailed paper
on the subject.77 C. Wilhelm Oseen, a close friend of Bohr, and
several others tried to see how close special solutions of Maxwell's
equations would come to the properties of what they now called
Einstein's "needle radiation."78 Erwin Schrodinger, influenced by
Einstein's statistical arguments, successfully derived a form of the
Doppler principle for quantum emission.79 Large numbers of
physicists did not yet take the lightquantum to heart, but com-
pared to the lack of interest just a few years before, the increase in
research studies of lightquanta by mid-1922 was significant. Fi-
nally, the paradoxes of radiation made themselves felt.
Moreover, in the period of 1914-21, there had been a remark-
able change in Einstein's public image. In 1914, when he was
offered a position in Berlin and the directorship of the Kaiser
Wilhelm Institut fur Physik, Einstein had been a demonstrably
superior physicist. But by 1921, his revolutionary suggestion that
light is bent by gravity had been experimentally verified amid great
and growing public acclaim. It was no longer easy to dismiss an
idea, however seemingly absurd, of so eminent a man. Physicists
74
See Chapter 7, note 75.
75
Laue, Berlin Sb, (1921), 480.
76
Einstein to Ehrenfest, ca. 20 January 1922. AHQP 1. See Klein, HSPS, 2
(1970), 1-39.
77
Breit, PM, 44 (1922), 1149-52.
78
Oseen, AP, 69(1922), 202-4. The lightquantum was frequently interpreted as
"needle radiation," that is, possessing longitudinal extension. This was an
intermediate stage of interpretation designed to allow for the explanation of
interference effects. See also Emden, PZ, 22 (1921), 512-17; L. Brillouin, JP, 2
(1921), 151-2; Bateman, PM, 46 (1923), 977-91. Duane, NAS Proceedings, 9
(1923), 158-64, should also be included, although it was published after
Compton's conversion to the lightquantum. Duane adamantly opposed
Compton's explanation but had adopted much of the essence of directed
lightquanta.
79
Schrodinger, PZ, 23 (1922), 301-3.
began to take cognizance of Einstein's most revolutionary hypoth-
esis of all, that of lightquanta.
In November 1922, while he was on a trip to the Far East, Albert
Einstein was awarded the 1921 Nobel Prize in physics. The honor
was not given to him for the theory of relativity, or for the
introduction of statistical techniques into physical analysis, or for
his quantum theory of specific heats. Any one of these contribu-
tions would have been worthy.80 The award came instead on
Oseen's nomination for Einstein's "achievements in mathemati-
cal physics, especially for his discovery of the photoelectric law."81
In his letter of nomination, Oseen quoted with emphasis Manne
Siegbahn's interpretation of Einstein's law, that energy transfor-
mation from radiation to matter or vice versa occurs only in
multiples of Planck's constant times frequency.82 Siegbahn's opin-
ion was of particular significance because of his knowledge of and
research on x-ray absorption features, essential preparation for
interpretation of the x-ray photoeffect. Much of Einstein's theoret-
ical work did not satisfy the criteria established in the Nobel will
for an accomplishment of practical significance. But the theoreti-
cal underpinning for the photocell, already in industrial use in the
postwar period, was eligible.
It was no coincidence that Einstein's award was the first Nobel
Prize given in physics after de Broglie's and Ellis's detailed verifica-
tion of the photoeffect law for radiations other than visible light.
The prize normally to be awarded in late 1921 was reserved and
then given to Einstein in late 1922, along with the 1922 Nobel Prize
to Niels Bohr.83 In his precis of Einstein's work for the 1921 prize,
Oseen lauded the "original and perceptive analysis" that had led to
Einstein's claim than an atom can absorb only multiples of single
quanta. This hypothesis, he said, was given the "most beautiful
confirmation" by Millikan's experiments on the energies of pho-
80
See, as one of many examples, the letter of nomination from Sommerfeld, 11
January 1922, in which all of these accomplishments figure, including the
quantum theory of the photoeffect. Nobel archives, 1922.
81
Nobel en 1921-1922 (1923), 6.
82
Oseen to Nobel committee, 2 January 1922. Nobel archives, 1922.
83
Friedman, Nature, 292 (1981), 793-8. There were several nominations for the
photoeffect work, among them ones from Sommerfeld (11 January), Emden
(29 January), and Oseen (2 January). "Forslag till utdelning av Nobelpris i
fysik ar 1922," pp. 17-18, 74, and 14-15, respectively. Nobel archives. For
other documents relevant to the Einstein award see Pais, American scientist,
70(1982), 358-65.
84
toelectrons, but there was no explicit appeal to de Broglie's x-ray
work in the justification for the Einstein award. That was reserved
for the place where it would have maximum impact: the brief for
Bohr's award written by Oseen at the same time.
Oseen saw the two awards as inextricably linked; in supporting
Einstein's, he was furthering Bohr's. De Broglie's x-ray evidence
formed the crucial empirical bridge between the two.85 The Ein-
stein-Bohr hv law - what we have called the quantum transfor-
mation relation - linked Bohr's atomic theory to the photoeffect.
If the law of the photoeffect was valid, de Broglie's results verified
Bohr's atomic structure down to the innermost shells of the atom.
Thus, Oseen first used Millikan's confirmation of the photoeffect
law to justify Einstein's award and then appealed to de Broglie's
results in justifying the award to Bohr. The Royal Swedish Acad-
emy agreed, although only after holding off the Einstein award
until both could be granted in 1922. The award of the 1923 Nobel
Prize to Millikan for his precise experimental work, including his
verification of the photoeffect law, was entirely consistent with the
position taken by the Academy. The awards of Einstein and Bohr
gave implicit sanction to the lightquantum hypothesis of radiation
absorption that underlay the photoeffect law.
One measure of the increasing attention paid to the lightquan-
tum hypothesis is the consolidation of opposition to the idea,
opposition that need not have been made explicit a few years
before. Bohr, for example, had never seriously concerned himself
with the paradoxical problems of radiation theory. But just at the
time that both he and Einstein received their Nobel Prizes, Bohr
proposed a way to avoid adopting the lightquantum.86 At the time,
Bohr's atomic theory faced formidable difficulties. Two of the
most pressing were the anomalous Zeeman effect - the multipli-
cation of spectral lines when the radiating sample is placed in a
magnetic field - and the dispersion of light. The first concerns the
emission of light, but the second involves absorption where, as we
have seen, the problems of localization of radiant energy are least
avoidable.
84
Oseen, "Einsteins lag for den fotoelektriska effekten." 12 pp. Bilage 2 to 1922
prize. Nobel archives.
85
Oseen, "Den Borska atomteorien." 34 pp. Bilage 3 to 1922 prize. Nobel
archives.
86
Bohr, ZP, 13 (1923), 117-65. See Klein, HSPS, 2 (1970), 1-39.
When light strikes matter, some of the energy is reradiated at a
different frequency. The relationship between the incident and
dispersed light was then experimentally well known; classical con-
sideration of the atom as a set of sinusoidal electron oscillators
gave a good first approximation.87 But this was accomplished by
assuming that the oscillating electron is coupled to the resulting set
of radiated frequencies. There can be no such coupling between
the frequencies absorbed by the Bohr atom and those subsequently
reradiated because radiated frequencies are differences of atomic
frequencies, not electron frequencies themselves. Although some
success attended attempts to treat optical dispersion as if it were
due to perturbations of the electrons in their periodic Bohr orbits, a
fundamental obstacle remained: The absorbed and radiated fre-
quencies are proportional to the energy differences between states,
not to the energies of the states themselves.88
A key question for Bohr was whether the full quantum of
radiant energy is concentrated on a single atom or electron within
an atom. It was, therefore, not a new dilemma, and Bohr had
shown some concern over it slightly before Maurice de Broglie's
presentation in 1921.89 By late 1922, Bohr had concluded that the
conservation laws governing energy transfer between light and
matter may not hold on the microscopic level. He argued that the
laws of energy and momentum conservation may only be approx-
imations valid for large numbers of atoms; the laws may hold only
statistically, and not apply to each microinteraction between radi-
ation and matter. Only when the individual transfers are inte-
grated over the large number of atoms in a testable sample does the
observer find that conservation laws apply. Einstein had tried a
similar approach a decade before and had eventually rejected it.90
Bohr's proposal illustrates the seriousness with which he now
viewed the paradoxes of radiation; to arrive at a solution, he was
willing to discard the microscopic validity of energy conservation.
But just as Bohr voiced this desperate hypothesis, another con-
vincing experimental demonstration was offered that further con-
87
References to the literature on the classical treatment of dispersion may be
found in Drude, AP, 1 (1900), 437-40. The empirical data appeared in a series
of papers beginning about 1900, mostly by R. W. Wood and published in the
/ W a n d PZ. See "Wood" in Poggendorff, Handworterbuch, vols. 4 and 5.
88
R. Ladenburg, ZP, 4 (1921), 451-68.
89
See, for example, Bohr's contribution to Solvay III, especially pp. 24iff.
90
Einstein to J. Laub, November 1910. Einstein archive.
firmed the lightquantum interpretation; one that explicitly denied
the escape from the lightquantum that Bohr had outlined. To
understand better the environment within which Louis de Broglie
worked in this period, we now turn to a brief discussion of this
effect named after Arthur Compton.
THE COMPTON EFFECT

It had been known for several years that if monochromatic x-rays
are directed against matter, something similar to optical dispersion
occurs. The x-rays scattered off the target have effective frequen-
cies less than that of the incident beam, and the decrease varies
with the angle between the incident beam and the direction of the
observer. Several attempts were made to find a theoretical inter-
pretation of the phenomenon, but the most notable were those by
the American experimentalist Arthur Holly Compton and, inde-
pendently, by Peter Debye in Zurich. A recent book covers the
technical aspects of these researches, so we need not give a detailed
discussion here.91 But we do need to place the results and influ-
ences of Compton's work in the context of the then growing
recognition of the lightquantum.
Compton had attempted for several years to explain the de-
crease in scattered x-ray frequency by analogy to optical scattering
from large molecules. When the wavelength of light is of the same
order as the diameter of the scattering particles, interference of
waves from opposite sides of the particle affects the angular distri-
bution of scattered light, producing a pattern similar to that seen
with x-rays. But in order to apply this argument to x-rays, the
electrons that reradiate the scattered x-rays would have to be about
io~10 cm in diameter, a thousand times larger than generally be-
lieved.92 In 1919-20 Compton discarded this uncomfortable no-
tion when he found evidence to support a second hypothesis that
scattered x-rays are a special kind of fluorescent x-ray.93 They
could not be ordinary fluorescent x-rays because they are not
characteristic of the scattering substance, their frequency varies
91
Stuewer, Compton effect (1975). Bartlett, AJP, 32 (164), 120-7.
92
A. Compton, WAS Journal, 8 (1918), 2 - n ; PR, 14 (1919), 20-43. See also
Stuewer, Compton effect (1975), chap. 3.
93
A. Compton, PR, 18 (1921), 96. He was led to this interpretation by previous
work on y-rays: PM, 41 (1921), 749-67; PR, 77(1921), 38-41.
with that of the incident rays, and they are polarized. Clearly,
many puzzles remained.
Late in 1921, influenced but not yet convinced by the evidence
that had already stimulated a new interest in the lightquantum by
others, Compton tentatively applied the quantum theory to the
problem for the first time.94 He did not appeal directly to the
radical lightquantum, but rather to the quantum transformation
relation, using it to calculate the recoil velocity of an electron
struck by the incident x-rays. He then explained the wavelength-
shifted secondary x-rays as Doppler-shifted x-rays reradiated by
the recoil electron. A year later, in December 1922, Compton
abruptly joined those who were finally taking the lightquantum
seriously. He discarded the two-step process and asserted that the
incident x-rays are scattered in localized lightquantum units.95
Each interaction of a lightquantum with an electron may then
be analyzed according to the kinematics of two-particle collisions,
the crucial assumption being that the x-ray is a particle of energy hv
and momentum hv/c. Conservation of energy and linear momen-
tum predict that a shift in wavelength follows the scattering angle 9
according to the relation
AX = h/mc(\ -cos 6) (10.1)
The experimental data that Compton had amassed corroborated
the result with great accuracy.
Compton's decision to reinterpret his data with the lightquan-
tum hypothesis was part of the reawakening of interest in the
lightquantum we have seen in others; the issues were current, at
least among European physicists. Maurice de Broglie's experimen-
tal evidence for lightquantum x-rays predated Compton's by eigh-
teen months; Ellis' argument for quantum y-rays had been in print
for almost a year. Compton cited both in his report on x-rays to the
National Research Council before his own conversion to the
lightquantum hypothesis.96 The spate of papers and discussions in
the German literature that marked the growth of interest in the
lightquantum formed the environment of (but was not initiated
by) Compton's success. Einstein's Nobel Prize was awarded before
word was heard of Compton's work. In short, Compton was
94
A. Compton, PR, 19 (1922), 267-8.
95
A. Compton, PR, 21 (1923), 483-502.
96
A. Compton, NRC Bulletin, 4 (1922), 54-5.
Figure 10.3. Debye's figure relating x-ray energy and direction to scat-
tered electron energy and direction. [PZ, 24 (1923), 164.]
reacting to a shift in opinion toward the lightquantum that had

already influenced many physicists before his results contributed
to that shift.
It is a measure of the increasing tolerance of the lightquantum
before Compton's work that a second, virtually identical explana-
tion of the anomalous frequency shift in scattered x-rays was given
by Debye before Compton's paper was published. Debye, then
Ordinarius at Zurich, started explicitly with Einstein's needle
radiation and treated the collision of an x-ray quantum with an
electron as if it were a classical interaction of two particles of
known momentum and energy. He described his work as "an
attempt to draw the most detailed possible conclusions from the
two assumptions 'quanta' and 'needle radiation' by use of the most
general laws 'energy conservation' and 'momentum conserva-
tion'." His hope was to "gain deeper insight of the quantum laws,
in particular their connection with wave optics."97 Debye's analy-
sis was based on the experimental data that Compton himself had
amassed in his report to the National Research Council just before
Compton's reconciliation to the lightquantum. Debye's results
were equivalent to, although somewhat more generally stated
than, Compton's. Figure 10.3 summarizes his conclusions about
97
Debye, PZ, 2^(1923), 166.
the special case in which the incident x-ray frequency is the
figure-of-merit value mc2/h. The numbered vectors above the
horizontal line of the incoming x-ray quantum express the fraction
of incident radiant energy transferred to the x-ray quantum scat-
tered in various directions; those below the horizontal express the
fraction of incident energy donated to the electron in each case.
The Compton effect elicited a strong response from European
physicists already primed for discussion of the unorthodox light-
quantum. Sommerfeld said that it "sounded the death-knell" of
the wave theory.98 Compton's discovery had the effect of a crystal
dropped into a supersaturated solution; the issues had been cur-
rent for two years, and the rapid crystallization of thought was due
to that advance preparation. By the time the word spread in the
first months of 1923, no physicist could ignore the lightquantum
any longer.
LOUIS DE BROGLIE'S QUANTUM-RELATIVISTIC

PARADOX
The increasing acceptance and wider adoption of the lightquan-
tum encouraged Louis de Broglie to intensify his search for a
physical interpretation whereby the seeming incompatibility of
wave and particle representations of radiation might be resolved.
The theory of relativity had thus far played no essential role in the
deliberations that had led to recognition of the paradoxes of
radiation." It had been implicated in explanation of x-ray doublets
and as the source of extra energy for photoemission. In 1923 de
Broglie tried to bring harmony between the two brainchildren of
Einstein: the theory of relativity and the lightquantum hypothesis.
He confronted in this synthesis a second dilemma, the solution to
which, he quickly recognized, was closely tied to a solution of the
paradoxes of radiation.
According to relativistic dynamics, precisely the sort that de
Broglie had been applying to "atoms" of light, the internal energy
of a particle with a rest mass of m0 is E = moc2. According to the
quantum theory, it possesses an intrinsic frequency v0, where
98
Sommerfeld to Compton, about spring, 1923. Paraphrased by Compton in
JFI, 198 (1924), 70. Millikan in Nobel en 1923 (1924).
99
The greatest use of relativity theory had been made by Emden, PZ, 22 (1921),
hvo = moc\ But to an observer past whom the particle travels at
velocity vp9 its internal energy is moc2/ V i vp2/c2, so that the
observed frequency of the quantum wave is
(10.2)
/2
On the other hand, the most directly comprehensible consequence
of the theory of special relativity is that a moving clock appears to a
stationary observer to be running more slowly than it does to an
observer moving along with it. According to relativity theory, the
observed frequency of the particle oscillator is
(10.3)
The relativistic dilation of time intervals predicts that the station-

ary observer measures a lower frequency than does the moving
observer.
Put in its baldest terms, the theory of relativity produces an
asymmetry when combined with quantum theory. Amass point's
energy , and hence its intrinsic quantum frequency , must
increase with its velocity relative to the observer; time dilation
predicts that the frequency must decrease for the same observer.
Louis de Broglie's greatest contribution to physics was to recognize
that a solution to this problem would shed light on the paradox of
wave and particle.
Casting about for a resolution of this new paradox for the
lightquantum, de Broglie reinterpreted a study published in 1919
by Marcel Brillouin, with whom he had studied physics at the
College de France. Brillouin had wished to explain how an electron
describing a continuous periodic orbit might nonetheless be de-
scribed by discontinuous quantum laws.100 His method was based
on a hypothesis originally due to Vito Volterra that, under certain
circumstances, a "hereditary" field is established by a moving
particle.101 Brillouin assumed that a particle - he called it a point
mobile - moves through a medium with a velocity greater than
that of waves in the medium. The motion is assumed to be
quasi-periodic and to be restricted to a circular region with a
100
M. Brillouin, CR, 168(1919), 1318-20.
101
Volterra, Lemons sur les equations (1913). For the fullest expression see
Journal de mathematiques pures et appliquees, 7 (1928), 249-98.
diameter less than the product of the wave velocity and the period
of the particle. Brillouin had in mind the quasi-periodic motion of
an electron constrained by the Coulomb force of the atomic
nucleus to remain within a certain region of space. A spherical
wave is to be imagined arising from the electron at the start of its
periodic motion. At the end of that period, the electron must still
lie within the spherical volume of space to which the wave has
expanded. This condition is necessary to ensure that the particle,
which moves faster than the wave, does not simply leave the wave
behind. Under these special circumstances, the particle catches up
at regular intervals to the waves it produces in moving through the
medium.
An electron circling a nucleus creates, therefore, a series of
imaginary points trailing in its wake. From each of these points a
spherical wave, emitted when the electron was there, would catch
up with the electron at the same instant as all the rest. The
hereditary field thus established embodies, as it were, one possible
past history of the particle. Brillouin showed that if the energy of
the point mobile at an anterior position is inversely proportional to
the distance to its present position, then the physical quantity that
characterizes the difference between one point in the chain and the
next has the dimensions of action (energy x time). Like Planck's
constant of action, it changes by integral multiples from point to
point.
When Louis de Broglie encountered Brillouin's suggestive
study, he was deeply involved in his search for resolutions both of
the wave-particle representation of light and of the relativistic-
quantum discrepancy. He already had come to identify atoms of
light with particles of infinitesimal, but perhaps not zero, mass.
Thus, de Broglie saw that Brillouin's analysis might apply to
lightquanta as well as to electrons. Viewed from this perspective,
the hereditary field offered the possibility of a solution to one or
both of the problems. An atom of light, like Brillouin's point
mobile, might possess an associated wave that interacts according
to some special relationship with the particle.
"Suddenly, at the end of the summer of 1923," de Broglie
recalled, "all of these ideas seemed to crystallize in my mind."102
The synthesis of particle and wave that he was developing for
102
L. de Broglie in Louis de Broglie (1953), 461.
application to atoms of light applied in an equivalent form to
atoms of matter as well. De Broglie's conviction that matter and
light should be treated equivalently lay in the symmetry inherent
in his goal: to form a synthesis between particle and wave. Much of
his motivation had come from the elder de Broglie, who, in his
book about x-rays in 1922, pointed out that "one finds certain
kinetic aspects in undulatory radiation, and certain periodic
aspects in the directedness of electrons."103 He meant that elec-
trons of definite velocity act in collisions just like x-rays of a
precisely defined frequency.104 His younger brother now trans-
formed that heuristic equivalence into a physical symmetry by
emphasizing the formal and physical similarity between matter
and light. Louis de Broglie's initial concern about the nature of
light had led to a revolutionary advance in the theory of matter.
PHASE WAVES
When de Brogliefirstpresented his synthetic fusion of particle and
wave, he did it not for the lightquantum but for a particle of
matter. Consider a particle moving with velocity vp past two
stationary observers. One observer, w, watches for the quantum-
wave effects of the moving point; the other, p, notices only the
relativistic effects of the point oscillator. According to equations
10.2 and 10.3, the two disagree about the frequency because v w >
vp. De Broglie suggested that observer w is seeing not the wave
effects of the particle itself but rather those of an associated plane
wave that travels in the same direction as the particle. This as-
sumption allows a special congruence between the particle and its
associated wave that seems counterintuitive at first: Even if two
observers disagree about the frequencies and velocities of the
internal oscillation of the particle and the wave, they nonetheless
can agree that at each instant there is a particular point in space
where the two oscillations remain in phase with one another. De
Broglie established the criterion that this point always coincides
with the changing position of the particle. That is to say, despite the
fact that the quantum and relativistic frequencies of the particle
differ, the two oscillations keep their phase - the fraction of a full
103
M. de Broglie, Les rayons x (1922), 17.
104
Ibid., 19. He displays a table giving equivalent frequencies and the derived
velocity of an electron of equivalent effect for radiations from "electric waves"
through y-rays.
wavelength that has passed - always equal to one another at the
point in space occupied by the particle. Beyond establishing the
connection of particle to wave, this relation defined the physical
location of the particle as that where the phase wave and the
internal relativistic oscillation always constructively interfere.
Imagine that the particle passes an arbitrary reference point with
velocity vp. Let us assume that it is then at the beginning of its
cycle; it has phase zero at that instant. At the same instant, the
phase of the plane wave is assumed to be zero at the same point. As
the wave propagates, it can be imagined to leave behind in its wake
an oscillation of frequency vw. To satisfy de Broglie's criterion that
the phases of particle and wave match at the particle at every
instant, the wave must travel faster than the particle. Unlike
Brillouin's point mobile, de Broglie's - which he tellingly called le
mobile - must move more slowly than its associated wave. Indeed,
one can calculate just how much faster the wave has to travel to
fulfill the criterion of phase synchronization.
The amplitude of the particle's internal oscillation at time t after
passing the reference point is proportional to sin invp t. In this
time, the particle travels a distance xp = vp t, so that its amplitude
at xp is
<S>p = sin(2nvpxp/vp) (10.4)
In the same time interval t, the zero-phase point of the plane wave
travels a distance xw > xp, so that at the same instant the amplitude
of the wave at point xp, at the particle, is
Ow = sin 2nvwAt (10.5)
where At is the time interval (less than /) that the zero-phase point
of the wave took to travel the distance from xp to xw. Equation 10.5
gives, in other words, the amplitude of the associated wave at a
point xw xp behind the zero-phase point. Clearly, At = (xw xp)/
vw. A straightforward manipulation shows that equations 10.4 and
10.5 are identically equal if and only if105
vw = c2/vp (10.6)
The velocity of the wave must exceed c by the same proportion
that the velocity of the particle falls short of c.
105
Equality of equations 10.4 and 10.5 requires vpxp /vp = vw (xw /vw xp /vw)m,
using the relation vp = v^ i vp2/c2) derived from equations 10.2 and 10.3 this
may be rewritten as
(1 - vp2/c2)xp /vp = xw /vw - xp /vw
The associated wave can therefore transfer no energy. It is not a
material wave; de Broglie called it an ondefictive when presenting
this argument to the Parisian Academy of Sciences in September
1923.106 It differs in fundamental respects from Marcel Brillouin's
hypothetical wave: de Broglie's is not physically real; de Broglie's
travels faster than the particle with which it is associated; finally,
Brillouin's wave is spherical, and it soon became clear that de
Broglie's is planar. De Broglie treated the same case, an electron
traveling in a circular orbit, that Brillouin had. According to de
Broglie, the fictional plane wave must precede the electron around
the circular orbit before the electron itself has traced that path. The
wave does not propagate spherically away from the coincident
point in the way Brillouin imagined his wave to do.
Only in this way can one understand the means de Broglie used
to calculate the time interval r between electron-wave phase
coincidences. In order to meet again at time r after the wave front
and the electron initially coincide, the wave must complete some-
what more than a single revolution in the orbit. In so doing, the
wave covers a total linear distance
XW = (T+T)VP (10.7)
where T is the period of the electron's revolution in its orbit.
Rearranging equation 10.7 with the additional constraint imposed
by equation 10.6 yields the value T = T vp2/c2 (1 vp2/c2).
With this value for r, de Broglie applied the condition of phase
synchronization. The phase of the wave is, by definition, zero
when it reencounters the electron. So the equality of phase de-
mands that the phase of the relativistic oscillation of the electron
also be zero at that point. Thus, in covering the distance VPT, the
electron must complete an integral number of internal oscilla-
tions. In other words:
Using the equality xw /vw = xp /vp, this becomes

1 - vp2/c2 =l-xp /xw
And in light of the equality xp /xw = vp /vw, this is
vp2/c2 = vp /vw
which leads directly to equation 10.6. This result, which follows from de
Broglie's own analysis, is derived on different considerations by Kubli, AHES,
7(1970), 26-68.
106
L. de Broglie, CR, 177 (1923), 507-10, 548-50, 630-2; PM, 47 (1924),
446-58.
h Vl-(tyfc) 2
where n is an integer. This may be rearranged to the form
T
=nh (10.9)
VI ~{vplcf
Equation 10.9 is precisely the action-integral expression for al-
lowed electron orbits in the Bohr atom, a mathematical condition
that had lacked, until now, any physical interpretation.107
Many years later, de Broglie described himself as always having
had "a very 'realist' conception of the nature of the physical world
and little given to purely abstract considerations."108 He had found
in 1923 what had eluded Bohr, a realistic conceptual explanation
for the discrete orbits in the Bohr atom, and it obviously impressed
him mightily. "It is almost necessary to suppose," de Broglie said
at the time, "that the electron orbit is only stable if the fictional
wave reencounters the electron in phase."109 And the following
year he repeated, "This beautiful result . . . is the best justifica-
tion that we can give for our way of addressing the problem of
quanta."110
EXTENDING DE BROGLIE'S HYPOTHESIS

Louis de Broglie began his studies in physics by trying to discover
from x-ray analysis the principles of atomic structure. He had
switched interests in 1921 to analyze the nature of radiation, and
this concern had now led him directly to a successful interpreta-
107
The generalized Sommerfeld condition is $1 p dq = nh. In de Broglie's system
this becomes
' mo(vx + vy + v2)
1K
y
V1 - (v/c)2 dx dy dz = nh
Because of the general relation f pdq = f p(dq/dt)dt, this may be rewritten as
mn
(vx2 + v2 + vz2)dt = nh
V1 - (v/c)2
or simply
m0v2T
, = nh
V1 - (v/c)2
108
L. de Broglie in Wave mechanics (1973), 12.
109
L. de Broglie, CR, 777(1923), 509.
110
L. de Broglie, Theses (1924, 1963), p. 53 of reprint.
tion of the Bohr atomic theory. This mutual influence of atomic
and radiation theory in his work was a product of his search for a
synthetic view that would place light and matter on an equivalent
footing. The revolutionary conceptual contribution of the phase
wave hypothesis was that it did not distinguish between an atom of
matter and an atom of light. Both consist of a mobile with an
associated wave. In one the wave properties predominate, and in
the other the localized particle properties predominate; but neither
is solely particle or wave.
De Broglie still had to provide an explanation of interference
and diffraction. The wave aspects of the synthetic theory of radia-
tion allowed for both, but one had to explain why the atoms of
light that carry the radiant energy behave as they do. And here he
encountered formidable problems. To solve them and explain how
a mobile can change direction, it was necessary, de Broglie said, to
alter the laws of inertia. The velocity of the energy-carrying particle
is equivalent to the group velocity of the phase wave, he claimed,
while the wave itself, carrying no energy, travels with the phase
velocity. Material points, whether lightquanta or electrons, are
analogous to singularities in a group of waves; the individual waves
in the group pass by, but the singularity may remain. The particle,
whether lightquantum or electron, follows ray trajectories con-
trolled by its phase wave; the trajectory is fixed by Fermat's
principle of least action. In an isotropic medium, the path of the
particle is always normal to the advancing wave front.
In his early studies and in his thesis of 1924, de Broglie neither
used nor discussed wavepackets; he was thinking of extended
monochromatic plane waves that, like Brillouin's spherical heredi-
tary waves, determine the possible trajectories of the associated
particle. Only after his success with the Bohr atom did de Broglie
make a tentative identification of the velocity of his onde fictive
with optical phase-wave velocity and that of the particle with the
related group velocity. The concept of representing the particle as a
superposition of phase waves that forms a localized discontinuity
in the field - the now-familiar wave-packet representation - is a
later invention. De Broglie's early suggestion of a parallel should
not be taken as more than a suggestion of how to begin to explain
interference and diffraction in his new concept of radiation.111 The
111
MacKinnon, AJP, 44 (1976), 1047-55, argues that de Broglie's identification
is inconsistent with modern understanding. But the claim is ahistorical and
physically misleading. See Schlegel, AJP, 45 (1977), 871-2.
real significance of de Broglie's inspiration lies in the influence it
had on others, notably Erwin Schrodinger, whose attempts to
rationalize its successes led to the full statement of wave mechanics
in 1925.
De Broglie's identification of the ondefictive as a phase wave led
to another prediction: the possibility of diffracting electron beams.
Near an aperture comparable in size to the phase wavelength, the
distortion produced in the wave bends the trajectory of a light-
quantum or an electron. "The new dynamics of a free material
point is to classical dynamics," de Broglie claimed, "like wave
optics is to geometrical optics."112 The paths of lightquanta that
pass through a pair of slits are controlled by the interference
pattern created in their phase waves. The energy transferred by the
light, as detected for example on a photographic plate, is distrib-
uted in fringes corresponding to the well-known interference
bands. This is true regardless of the intensity of the incident light.
The intensity measures the number of lightquanta; it does not
affect their trajectories. In an identical way, the distortion in the
phase wave structure of an electron near a sufficiently small aper-
ture could be expected to diffract the electron beam. The mecha-
nism for the complex propagation of the phase wave that in turn
directs the trajectory of an energy-carrying particle remained a
desideratum of a "generalized theory of electromagnetism."
De Broglie gave the fullest expression of his hypothesis and its
remarkable success for the Bohr atom as one of his theses for the
doctorate under Jean Perrin and Paul Langevin in November
1924.113 There, for the first time, he explicitly calculated the now-
familiar wavelength of the phase wave that "guides the displace-
ment of the energy" of a material particle:
X = h/mQv (10.10)
But with the exception of a refined treatment of the statistical
properties of a gas of lightquantum particles, he added little in the
thesis that went beyond the early papers. De Broglie had done
more than simply turn Einstein's lightquantum hypothesis on its
head; if there is a particlelike aspect to light, there must also be a
wavelike aspect to matter. He stressed the fundamental synthesis
112
L. de Broglie, CR, 777 (1923), 549.
113
L. de Broglie, Theses (1924), Annales de physique, 3 (1925), 22-128. The
reprint includes a useful historical introduction by de Broglie, pp. 22-30.
that his hypothesis implied: Light and matter are equivalent. Both
share in differing degrees the characteristics of localized particle
and spatially extended wave.
Langevin was very much impressed by the work. He called it
"an attempt at a solution of the most important problem of
present-day physics."114 Its most profound aspect was the general-
114
From the draft manuscript critique by Langevin of de Broglie's thesis for -
merly in the possession of Mme. Luce Langevin. I am indebted to her for
permission to quote it in full here:
Rapport sur la these de M. L. de Broglie 1924
Le travail de M. Louis de Broglie represente un important effort vers la
solution du probleme le plus important de la physique actuelle, la synthese
des deux theories optiques des ondulations et des quanta, jusqu'ici contradic-
toires au moins en apparence, dont chacune s'appuie sur tout un ensemble de
faits et de confirmations experimentales remarquables et dont Fopposition
vient renouveler de maniere imprevue l'ancien conflit de remission et des
ondulations.
Les enonces d'aspect paradoxal qu'a du introduire la theorie des quanta
sont eux-memes de deux ordres qui correspondent a ce qu'on peut appeler
Taspect dynamique d'un cote et celui des "quanta de lumiere" de l'autre. Pour
rendre compte de la composition du rayonnement noir et de la structure des
spectres d'emission et d'absorption il a fallu, avec Planck, Bohr, Sommerfeld,
admettre que parmi tous les mouvements des centres electrises compatibles
avec les lois de la dynamique classique, seuls sont possibles ceux pour lesquels
les orbites electroniques satisfont a certaines relations numeriques simples.
D'autre part M. Einstein (et M. J. J. Thomson)* a montre que des effets dits
photoelectriques ou l'absorption du rayonnement s'accompagne de remis-
sion de particules cathodiques, ainsi que lesfluctuationsdans le rayonnement
noir semblent bien imposer l'idee que l'energie du rayonnement lui-meme
n'est pas distribute de maniere continue dans l'espace mais se compose de
grains d'energie proportionnelle a la frequence.
L'idee fondamentale de M. de Broglie consiste a unifier les deux ordres
d'enonces de quanta en associant a chaque particule electrisee considered
comme grain independant d'energie un phenomene periodique de frequence
proportionnelle a l'energie du grain exactement comme l'hypothese des
quanta de lumiere associe aux ondes lumineuses periodiques des grains non
electrises ou leur energie se trouve concentree.
La liaison generale ainsi etablie entre le mouvement d'une particule electri-
see ou non et la propagation d'un phenomene periodique est rattachee par
l'auteur aux principes generaux de maniere tres profonde et par une serie de
remarques telles qu'une hypothese paradoxale au premier abord apparait de
plus en plus conforme a la nature des choses. II montre en particulier que les
relations introduites entre le mouvement d'un grain d'energie et la propaga-
tion du phenomene periodique qui Taccompagne sont exactement de meme
nature que celles introduites par Lord Rayleigh et Gouy entre le mouvement
d'un groupe d'ondes et la propagation de la phase d'une onde exactement
periodique.
D'autre part, et ceci est peut-etre 1'aspect le plus interessant des idees de
M. de Broglie, cette liaison semble correspondre a l'analogie, purement
*Name inserted later.
ity claimed for the relationship between particle mechanics and
wave properties, between Hamiltonian dynamics of particles and
Fermat's principle of least action for wave propagation. "Perhaps
here is the beginning of a wave interpretation of the principles of
dynamics," Langevin said, and pointed to de Broglie's method for
interpreting the stationary orbits of the Bohr atom. Although
formelle jusqu'ici, mais peut-etre tres profonde, entre les enonces d'integrales
stationnaires qui sont a la base de la dynamique d'une part avec les principes
d'Hamilton et de Maupertuis et de la theorie des ondulations, d'autre part
avec le principe de Fermat tel que Fintroduisent et le justifient les conceptions
d'Huyghens et de Fresnel.
De meme que les principes de la dynamique du point materiel prennent
leur expression la plus simple grace a Introduction d'un vecteur a quatre
dimensions, l'impulsion d'Univers dont la composante d'espace est la quan-
tite de mouvement et la composante de temps Fenergie du mobile, Fauteur
montre que le principe de Fermat et toute la cinematique des ondes condui-
sent a introduire un "vecteur d'onde" d'Univers dont la composante d'espace
est dirigee suivant le rayon et dont la composante de temps est la frequence.
L'introduction de ces deux vecteurs conduit tout d'abord a generaliser la
relation du quantum sous une forme intrinseque valable pour tous les sys-
temes galileens de reference, et a preciser la relation entre le mouvement
d'une particule et le phenomene periodique associe, en affirmant Fidentite des
deux vecteurs [:] impulsion d'Univers de la particule et vecteur d'onde pour le
phenomene periodique (ou plutot, en raison du choix des unites, la Constance
de leur rapport egal a la constante h de Planck). On etend ainsi dans un sens
tout a fait conforme a Fesprit de la theorie de relativite aux quatre compo-
santes de chaque vecteur, la relation du quantum envisagee jusqu'ici pour les
seules composantes de temps, energie et frequence.
Dans ces conditions, les principes d'Hamilton et de Fermat se confondent
et representent deux aspects d'un meme principe, Fun concernant le mouve-
ment de la particule et Fautre la propagation de Fonde associee. Peut etre y
a-t-il ici Forigine d'une interpretation ondulatoire des principes de la dynami-
que. En tout cas une consequence immediate de cette unification des deux
vecteurs qui donnerait Fenonce le plus general et le plus profond de Fhypo-
these des quanta, est que les trajectoires dynamiquement possibles de la
particule dans un champ de force se confondent avec les rayons de Fonde
associee.
Un resultat particulierement remarquable est que les conditions dyna-
miques de[s] quanta se deduisent immediatement de cette conception si Fon
admet qu'une trajectoire fermee ne peut etre stable si Fonde associee au
mobile et qui se propage le long de cette meme trajectoire avec une vitesse
superieure a celle du mobile doit, en rejoignant celui-ci se retrouver en
concordance de phase avec elle-meme. Cette condition de resonance des
trajectoires stables pour Fonde associee fournit les conditions de quanta sous
la forme generate de Sommerfeld et en donne une representation] concrete
singulierement suggestive malgre les obscurites qui subsiste[nt] encore quant
a la nature physique du phenomene periodique associe au grain d'electricite.
En appliquant les memes conceptions au cas d'un gaz contenu dans un
recipient a parois supposees reflechissantes pour les ondes associees, Fauteur
others in Paris had difficulty understanding the thesis - Perrin said
publicly only that it was "tres intelligent"115 - Langevin discussed
the matter with Einstein and then asked de Broglie to send Einstein
an advance copy of the thesis,116 Einstein immediately recognized
its value. "He has lifted a corner of the great veil," Einstein wrote
to Langevin and Langevin related to de Broglie.117 And the Pari-
sian physicists granted the younger de Broglie his doctorate.
INTERSECTION OF RADIATION THEORY AND

ATOMIC THEORY
At the time that the new synthetic understanding of matter and
light evolved in France, much effort in theoretical physics lay in
the attempt, largely carried on by German-speaking physicists, to
extend the Bohr theory of the atom to multielectron systems. By
1923 it was widely recognized that the program was failing.118
There were a number of difficulties, including the problem of
dispersion of light, mentioned earlier, the anomalous Zeeman
effect and its associated "riddles" for spectroscopy,119 the necessity
for introducing half-integral quantum numbers,120 and problems
that led eventually to the exclusion principle,121 and to electron
spin.122 The increasing realization that semimechanical solutions
to those problems were not forthcoming contributed to the envi-
ronment in Germany within which Louis de Broglie's heterodoxi-
cal hypothesis was received.
The chief architect of atomic theory was Niels Bohr, who in 1923
understood that the lightquantum could be avoided no longer if he
wished to bring harmony to a new nonmechanistic atomic
retrouve sans nouvelle hypothese Fexpression donnee par Planck pour la
constante de Fentropie ou constante chimique du gaz.
115
La Varende, Les Broglie (1950), 333. See also the comments of another
member of de BroghVs thesis committee, Charles Mauguin, in Louis de
Broglie (1953), 434-
116
Langevin to L. de Broglie, 27 July 1924. "Autographes," L. de Broglie dossier.
117
Langevin to L. de Broglie, 13 January [1925]. Ibid. (The original is quoted in
German.)
118
Lande, PZ, 24 (1923), 441-4.
119
Forman, his, 59 (1968), 156-74; HSPS, 2 (1970), 153-261.
120
Cassidy, HSPS, 10 (1979), 187-224.
121
Serwer, HSPS, 8(1911), 189-256.
122
Kronig in Theoretical physics (i960), 5-39.
123
theory. Bohr found the Compton effect a particularly intransi-
gent obstacle. It seemed to prove that individual interactions of
x-rays with electrons conserve energy and momentum. Even
Bohr's daring statistical interpretation of energy conservation now
seemed insufficient to circumvent the lightquantum. The story
need not be repeated here how the American theorist, John Slater,
approached Bohr at this time with the idea that lightquanta are real
and that their emission and absorption by atoms are dictated by a
"virtual radiation field."124 The resulting development of the
Bohr-Kramers-Slater theory of atomic "virtual oscillators" reaf-
firmed in a transitional form Bohr's statistical interpretation of
radiant interaction.125 Although experimental tests of major pre-
dictions of the BKS theory by Bothe and Geiger failed to justify the
idea, by the time the experimental results were known, the issues
had become superfluous.126
The basic ideas inherent in the statistical approach of the BKS
theory were used by Bohr's assistant, Hendrik Kramers, later
joined by Werner Heisenberg, to treat the thorny problem of the
dispersion of light by the Bohr atom.127 This successful work
strongly influenced the young Heisenberg in 1925 when he found a
mathematical formalism that solved most of the pressing problems
that the Bohr theory faced.128 Heisenberg's matrix mechanics,
although directed to the problems of atomic theory, was stimu-
lated in essential ways by the renewed appreciation of Einstein's
lightquantum hypothesis.
In the meantime, a separate line of development based on de
Broglie's hypothesis of phase waves led to another means of ana-
lyzing the difficulties of atomic theory. This followed the route
from de Broglie through Einstein to Erwin Schrodinger. The
resurgence of interest since 1921 in Einstein's lightquantum had
led him to try to restate its inherent flaws. He had known since
1909 that lightquanta alone were impotent to resolve the difficul-
ties that faced radiation theory - hence his plea to Ehrenfest that
123
Bohr, AP, 71 (1923), 228-88.
124
Klein, HSPS, 2 (1970), 1-39. Slater in Wave mechanics (1973), 19-25.
125
Bohr, Kramers, and Slater, PM, 47(1924), 785-802.
126
Bothe and Geiger, ZP, 26 (1924), 44; ZP, 32 (1925), 639-63. See also
A. Compton and Simon, PR, 25 (1925), 306-13; 26 (1925), 289-99.
127
Kramers, Nature, 773 (1924), 673-4. Heisenberg and Kramers, ZP, 31 (1925),
681-708.
128
Heisenberg, ZP, 33 (1925), 879-93.
129
the problem was sufficient to drive him to the madhouse. In-
stead he developed a concept of a Gespensterfeld, or "ghost field,"
that might itself carry little or no energy, yet guide the trajectories
of lightquanta in accord with laws of wave optics.130 Little came of
the concept, but it bore sufficient similarity to de Broglie's ideas so
that Einstein was quick to recognize what was significant in the
heterodox suggestion from France.
When Einstein encountered de Broglie's thesis, he was just
developing a new statistical approach to the kinetic theory of an
ideal gas using a method suggested by the Indian physicist S. N.
Bose. Bose had derived the black-body law using a statistical model
of interacting particles based on the lightquantum, entirely free of
assumptions from classical electromagnetic theory.131 In 1925 Ein-
stein pointed out that de Broglie's hypothesis allowed one to treat
equivalently the fluctuations in black-body radiation and the sta-
tistical fluctuations of a gas according to Bose's statistics. He then
went on to propose a possible experimental test for diffraction in
atomic beams.132 The particular statistical application of de Brog-
lie's hypothesis that Einstein made in 1925 "stuck the matter right
under [Schrodinger's] nose."133
Among the physicists now interested in the lightquantum,
Erwin Schrodinger was in a unique position to appreciate the
merits of de Broglie's phase wave hypothesis. In 1922 he had called
attention to a "remarkable property of the quantum orbits of an
individual electron."134 A mathematical restriction that acts on an
electron in a periodic orbit suggested that with each revolution the
electron "adjusts" itself to its new point in four-dimensional
space-time. Thus, Schrodinger had already recognized a periodic,
spatially extended effect of each electron in space. It has been
suggested that Louis de Broglie's reputation among most German-
129
Einstein to Ehrenfest, 15 March [1921]. AHQP 1, 77.
130
Einstein, Berlin Sb (1923), 359-64, was as close as Einstein came to publishing
his ideas. See Lorentz, Problems (1927), 156-7, who lectured on the idea at
Caltech in 1922; Lorentz to Einstein, 13 November 1921, Einstein archive.
131
Bose, ZP, 26 (1924), 178-81.
132
Einstein, Berlin Sb (i924)> 261-7; (1925), 3-14.
133
Schrodinger to Einstein, 23 April 1926. Einstein archive. Quoted in Przibram,
Briefe zur Wellenmechanik (1963), 24. Schrodinger's first use of de Broglie's
hypothesis was in PZ, 27(1926), 95-101. For the influence of Einstein's work
in kinetic theory on Schrodinger's derivation of wave mechanics, see Klein,
Natural philosopher, 3 (1964), 3-49.
134
Schrodinger, ZP, 72(1922), 13-23.
speaking physicists, notably those such as Bohr and Sommerfeld
who were working on atomic spectroscopy, was not high.135 But
Schrodinger, like Einstein and de Broglie, was committed to rela-
tivistic and statistical formulation of theory; he was not directly
concerned with problems of atomic structure.
Late in 1925, Schrodinger began a four-paper series in which he
extended the formal analogy between Hamiltonian mechanics of
particles and wave optics.136 Using de Broglie's hypothesis, Schro-
dinger reinterpreted quantum transitions as changes in the vibra-
tional mode of the superposition of "phase waves" that character-
ize the entire atomic system. He was relieved to have found an
alternative interpretation of quantum theory that avoided what to
him was the uncomfortable notion of mechanically indescribable
electron jumps from one orbit to another. The transition of an
electron could now be interpreted as a change in the vibrational
mode of the entire atom. The quantum wave mechanics that
resulted from Schrodinger's analysis was a direct descendant of
physicists' accommodation to the lightquantum hypothesis.
In the spring of 1926, Schrodinger showed that Heisenberg's
matrix mechanics, which Schrodinger said had "frightened" and
"repulsed" him by its purely formal character, was mathemati-
cally equivalent to his own wave mechanics.137 Schrodinger wrote
to de Broglie in March 1926 to describe his work. "It is possible to
calculate all the Born-Heisenberg matrix elements by simple
differentiations and quadratures if one has the complete solution
of the wave equation," he said.138 He was especially gratified to
135
Raman and Forman, HSPS, 1 (1969); 291-314. Hanle, Isis, 68 (1977), 606-9.
136
Schrodinger, AP, 79 (1926), 361-76. See also Wessels, SHPS, 10 (1979),
311-40.
137
Schrodinger, AP, 79 (1926), 734-56. Eckart, PR, 28 (1926), 711-26.
138
Schrodinger to L. de Broglie, 26 March 1926. (L. de Broglie dossier.) Extracts
from the original letter written in Schrodinger's French, follow [reprinted
with permission of Ruth Braunizer and l'Academie des Sciences de Paris].
"Pour des phenomenes petites ou fines il faut aussibien dans la mechanique
que dans Foptique remplacer le traitement geometrique par un traitement
ondulatoire-on pourrait aussi dire par un traitement physique,-en tenant
compte de ce que la longueur d'onde estfinieet non infiniment petite comme
le suppose l'optique geometrique (et de meme l'ancienne mechanique)." "II
est possible de calculer toutes les elements de matrices de Born-Heisenberg
par des simples differentiations et quadratures, quand on possede la solution
complete de l'equation d'onde mentionnee plus haut, c'est a dire un systeme
complet de 'Eigenfunktionen'."
report that the wave mechanical formulation based on de Broglie's
hypothesis replaced the purely "geometrical formalism" by a wave
treatment, "one could also say by a physical treatment."
Two formerly independent research fields - the spectroscopic
analysis of atoms and the statistical analysis of radiation - were
brought into contact in 1925 by the long-delayed discussion of the
physical reality of lightquanta. By 1926 each had developed into
equivalent formulations of the new quantum mechanics. That
there were two independently formulated versions of the new
quantum mechanics was in large part an effect of the virtual
separation of the two parent fields - atomic theory and radiation
theory - for the preceding decade. And this, in a larger sense,
constituted the fullest corroboration of Louis de Broglie's remark-
ably fruitful but decidedly unorthodox proposal that matter and
light are fundamentally one and the same thing.
Epilogue
"We'll wait awhile," the master said, "that thus

our senses may grow used to this strong smell,
and after that, it will not trouble us."
Dante, Inferno, XI10-12
A new and fundamentally indeterminate approach to physical

theory had been established, and the duality of waves and particles
became an inescapable aspect of the successful new physics. After
Max Born proposed a probabilistic interpretation of Schrodinger's
wave function in 1926, younger physicists learned that it was
improper to ask what the consistent outcome would be from given
initial experimental conditions.1 One could calculate the relative
likelihood of all possible results and assign a probability to each.
Seeking to express the "conceptual content" of the new probabilis-
tic mechanics - which by its very nature denied conceptualization
in the ordinary sense of the word - Werner Heisenberg analyzed
an issue in classical physics in which an equivalent limitation acts.
The example he selected touches our study to the core, for he
reverted to concepts that his teacher, Arnold Sommerfeld, had
long before recognized to be true for impulse x-rays.2
Heisenberg's symbolic formulation of the noncommuting char-
acter of measurements to determine the position x and the mo-
mentum p = h/X of a particle is his famous principle of indeter-
minacy:
AxApx~h (E.I)
This states that it is impossible to determine simultaneously, with
arbitrary accuracy, both the position and the momentum of a
1
Born, ZP, 37 (1926), 863-7; 38 (1926), 803-27. An earlier version appeared in
Gottingen Nachrichten (1926), 146-60.
2
Heisenberg, ZP, 43 (1927), 172-98. Indeterminacy principle is preferable to
uncertainty principle: The former denotes the claim oiignorabimus essential to
the Copenhagen interpretation, the latter merely ignoramus.
The tiger and the shark 303
particle along a line. If the position x is known accurately, then the
momentum along that axis is correspondingly indeterminate. The
product of the indeterminacy in both measurements must exceed
a certain small number set by Planck's constant. Momentum,
according to de Broglie's relation, may be expressed in terms of
wavelength. So, the wave-mechanical interpretation of the inde-
terminacy relation is equivalent to a prohibition of simultaneously
determining the position of a particle and the wavelength of its
wave function. Conceptually, it is a direct analog of the tradeoff
that Sommerfeld had utilized to distinguish between localized
x-ray impulses and periodic waves.
One may not speak of a piece of a monochromatic wave; its
wavelength can be precisely defined only if the wave extends to
infinity in space. If one interprets the momentum of the particle in
terms of its de Broglie wavelength, Heisenberg's relation E. 1
makes the same point about the electron. Either the velocity or the
position of the particle may be defined by experiment with arbi-
trary accuracy, but the two cannot be determined simultaneously.
If the electron is known to occupy a particular location in space,
the Fourier integral of its wave function spreads over many fre-
quencies and the momentum becomes accordingly less definite. If,
on the other hand, the momentum is specified, the wave function
must be monochromatic; the wave must extend to great distances
in space. But then the range of possible positions that the particle
may occupy with equal likelihood becomes correspondingly great.
In a later study of the continuous x-ray spectrum as an example
of continuity in physics before and after quantum mechanics,
Sommerfeld referred explicitly to the resuscitation of his action-
integral representation of the quantum conditions in Heisenberg's
principle of indeterminacy.3 He might have gone further. In the
form relating position and momentum, Heisenberg's indetermi-
nacy relation is a conceptual extension of the Fourier interpreta-
tion of impulse and periodic waves that Sommerfeld had first
raised in 1901.
ELECTRON DIFFRACTION
The wave properties of light and the localized properties of matter
are familiar characteristics. The accumulated experimental evi-
3
Sommerfeld, Scientia, 51 (1932), 41-50.
304 Epilogue
Figure E.I. Electron diffraction pattern from gold obtained in 1927 by

G. P. Thomson. Compare this to the x-ray picture in Figure 8.1,
imagining that photograph spinning rapidly. [PRS, 117A (1928), 604.]
dence that we have discussed convinced physicists in the early

1920s that radiation also demonstrates particlelike properties. Ex-
perimental evidence in corroboration of the remaining permuta-
tion of wave and particle with light and matter was soon offered.
A beam of electrons directed toward a slit of molecular dimen-
sion should show diffraction effects. "In this way," de Broglie had
predicted, "one might seek experimental confirmation of the [pilot
wave] hypothesis."4 Louis de Broglie was, by his own admission,
not able to undertake the tests. And he had no success in convinc-
ing the experimentalists in rue Lord Byron to assist him.5 In
mid-1925 Walter Elsasser suggested that previously unexplained
results of electron-beam experiments might be interpreted as due
to the wavelike behavior of electrons.6 Anomalous peaks in the
intensity of electrons scattered from matter noted in experiments
two years before by Clinton Davisson and Charles Kunsman
could, Elsasser argued, be evidence for diffraction of electrons.7
4
L. de Broglie, CR, 177(1923), 549.
5
See the interview with L. de Broglie by T. S. Kuhn in AHQP.
6
Elsasser, Naturwissenschaften, 73(1925), 711.
7
Davisson and Kunsman, PR, 22 (1923), 242-58. See also Gehrenbeck, Electron
diffraction (1973).
Within two years, Davisson and Lester Germer in America and
George Thomson in Britain confirmed Elsasser's prediction.8 (See
Figure E.i.)
Elsasser also suggested that an anomaly in the absorption of slow
electrons by noble gases might also be due to a wave effect. The
absorption generally increases as the electron velocity decreases.
But the Swedish physicist Nils Akesson noted in 1916 that elec-
trons accelerated across only a few volts seem to be more penetrat-
ing than faster ones.9 The effective cross section for the transfer of
electron kinetic energy to atoms of the gas drops abruptly for
electrons with energy under a few electron volts; the electron mean
free paths become extraordinarily long. Starting in 1921, extensive
research on the phenomenon was undertaken by Herbert Mayer
and Carl Ramsauer in Heidelberg.10 Neither classical nor semi-
quantum treatments succeeded in explaining the "Ramsauer ef-
fect."11 But when a full explanation using the new quantum
mechanics was provided in 1929, it was, as Elsasser had suggested,
of an effect similar to colloidal scattering of light. The phenome-
non sets in only when the de Broglie wavelength of the incident
electrons exceeds a threshold value.12
In 1929 Louis de Broglie received the Nobel Prize in physics for
his theoretical hypothesis of matter waves.13 The award followed
results by Davisson and G. P. Thomson that verified the hypoth-
esis experimentally.14 It had taken seventeen years from the time
8
Davisson and Germer, Nature, 119(\î% 558-60. G. P. Thomson, PRS, 117A
(1928), 600-9. See also Russo, HSPS, 12 (1981), 117-60.
9
Akesson, Lunds Universitets Arsskrift, 12 (1916), part 11. See also his earlier
studies under Lenard, Heidelberg Sb, 5A (1914), Abh. 21.
10
Ramsauer, PZ, 21 (1920), 576-8. Mayer, AP, 64 (1921), 451-80. Ramsauer,
PZ, 22(1921), 613-15; AP, 6^(1920,513-40.
11
The classical treatment was by Hund, ZP, 13 (1923), 241-63. The semiquan-
tum was presented by Zwicky, PZ, 24 (1923), 171-83.
12
Faxen and Holtsmark, ZP, 45 (1927), 307-24. Wentzel, PZ, 29 (1928),
321-37. Mott, PCPS, 25 (1929), 304-9.
13
The award was part of a remarkably continuous series of Nobel Prizes given for
studies on the interaction of radiation with matter. The prizes in 1921-3 to
Einstein, Bohr, and Millikan have already been discussed. In 1924 Manne
Siegbahn's work on x-ray spectroscopy was honored, and in 1925 Franck and
Hertz's work on the quantum excitation of atoms. In 1927 Compton was
awarded half of the prize for his work on x-ray scattering, and the 1928 prize
went to Owen Richardson for work on thermionic emission (closely related to
the photoeffect). Louis de Broglie's prize in 1929 recognized the importance of
the wave interpretation of matter; it was matched in 1933 by t n e half-prize
awarded to Erwin Schrodinger.
14
Nobel en 7929 (1930), 16-27.
306 Epilogue
Einstein first suggested the lightquantum until he received the
Nobel Prize for related work, and in virtually all of those years his
revolutionary hypothesis was rejected by most physicists. It was
onlyfiveyears after de Broglie's equally revolutionary idea that he
was awarded the prize. Both Louis and Maurice de Broglie had
been nominated for the Nobel Prize as early as 1925.15 At that time,
the evaluation of their work described in detail the importance of
Maurice's research on the x-ray photoeffect, but Louis's theoretical
research was judged to have thus far "little significance for
science."16 Although at first even de Broglie's closest associates in
France had great difficulty believing in matter waves, in Schro-
dinger's formulation the inherent idea was widely accepted. There
was still great opposition to the conceptual demands of the new
synthetic view of matter and light. But the difficulties it relieved
and the symmetry inherent in the new dualistic interpretation
served to recommend its attractions despite its drawbacks. Regard-
less of their intuitive predispositions, by 1927 most influential
European physicists had learned to accept the duality of waves and
particles.
Scientists raised on mechanical representations of nature found
the new indeterministic physics hard to swallow. J. J. Thomson
described the new dualistic basis of physics as a struggle "between a
tiger and a shark, each is supreme in his own element but helpless
in that of the other."17 He persevered in attempts to find a me-
chanically consistent model for lightquanta. In 1925 he pictured
them as disconnected closed loops or rings of electric force, their
frequency determined by the time light takes to loop around the
circumference.18 William Henry Bragg joked in resignation that
physicists trot out a wave description on Mondays, Wednesdays,
and Fridays and a particle description on Tuesdays, Thursdays,
and Saturdays.19
15
Khvol'son to Nobel Committee, 8 January 1925. In "Forslag till utdelning av
Nobelpris i fysik ar 1925," p. 61. Nobel archives.
16
Oseen, "Utredning betraffande M. Siegbahn samt M. och L. de Broglie," 16
pp. Bilage 1 to 1925 prize. Nobel archives.
17
J. J. Thomson, Structure of light (1925), 15.
!8
Ibid., 2 iff. Thomson's proposal bears a resemblance to a casual suggestion by
Philipp Lenard in 1910, Heidelberg Sb, 1 (1910), Abh. 16, pp. 17-18.
19
W. H. Bragg, BAAS Report, 98 (1928), 20.
20
Planck, JFI, 204 (1927), 13-18. Lorentz, Nature, 114(1924), 608-11. For the
classical portrayal of German willingness to accept dualistic theory, see For-
man, HSPS, 3(1971), 1-115. That this sort of issue represented a crucial break
with tradition is sensitively treated in McCormmach, Night thoughts (1981).
But many physicists in Germany, already used to a less mecha-
nistic ontology than their British counterparts, and many younger
physicists in other nations were able to adopt the new approach to
physical theory with less difficulty.20 Some, such as Schrodinger
and Einstein, saw the indeterminacy relations as a statement of
man's limited ability to penetrate the details of natural processes,
not as a complete description of how nature acts. Others, such as
Heisenberg and Bohr, felt that in the indeterminacy relations and
in Bohr's associated principle of complementary physical descrip-
tions, humans had discovered the deep truth that nature itself
follows only statistical rules. In its fully developed form - the
Copenhagen interpretation - this latter view represents the public
philosophical view of most physicists to this day. But these issues
are largely unrelated to doing physics. Whether nature is only
statistically consistent or whether humans are simply unable to
penetrate to a consistent core soon became matters of only mar-
ginal professional concern to physicists.
Physics had undergone a remarkable transformation. The in-
consistencies that had plagued atom theory and confounded radia-
tion theory were resolved in the new quantum mechanical for-
malism. But successful form was gained at a high conceptual
price - the abandonment of deterministic theory. Problems in
atomic theory were instrumental in bringing about that change in
world view. But it was the attempt to formulate a consistent theory
of radiation after 1921 and the influence of this program on atomic
theory that sparked the most dramatic shift in understanding. And
the specific problems that led to the adoption of the new dualistic
interpretation of light extend back, as we have seen, to the earliest
years of the century.
BIBLIOGRAPHY
All works mentioned in the notes and text are listed here by author
or editor. A few liberties have been taken in the interests of
consistency. Lords Cherwell, Kelvin, and Rayleigh appear here
under their family names - Frederick Lindemann, William
Thomson, John W. Strutt and Robert J. Strutt - even though
many of their works are published under their baronial title.
Similarly, English transliteration is used for Russian physicists
who published papers in a German or French transliteration; in
particular, Chwolson appears here as Khvol'son, Joffe as IofFe.
Names with prepositions (de Broglie, Des Coudres, van der Waals)
are listed under the final part.
Abraham, H., and P. Langevin, eds. Ions, electrons, corpuscles. 2 vols.

Paris, 1905.
Abraham, M. "Prinzipien der Dynamik des Electrons." AP, 10 (1903),
105-79.
- Theorie der Elektrizitdt, vol. 2, Elektromagnetische Theorie der Strah-
lung. Leipzig, 1905.
Akesson, N. "Uber die Geschwindigkeitsverluste bei langsamen Katho
denstrahlen." Heidelberg Sb, 5A (1914), Abh. 21.
- "Uber die Geschwindigkeitsverluste bei den langsamen Kathoden-
strahlen und uber deren selektive Absorption." Lunds Universitets
Arsskrift, 12 (1916), part 11.
Allen, S. J. "The velocity and ratio e/m for the primary and secondary
rays of radium." PR, 23 (1906), 65-94.
Andrade, E. N. da C. "William Henry Bragg." RSL Obituary Notices of
Fellows, 4 (1943), 277-300.
Angstrom, K. "Beitrage zur Kenntnis der Warmeabgabe des Radiums."
PZ, 6(1905), 685-8.
310 Bibliography
Ashworth, J. R. "y rays from radium." Nature, 69 (1904), 295.
Badash, L., ed. Rutherford and Boltwood: Letters on radioactivity. New
Haven, 1969.
Baeyer, O. von, and A. Gehrtz. "Die Anfangsgeschwindigkeiten licht-
elektrisch ausgeloster Elektronen." VDpG, 72(1910), 870-9.
Baeyer, O. von, and O. Hahn. "Magnetische Linienspektren von /?
Strahlen." PZ, 11 (1910), 488-93.
Baeyer, O. von, O. Hahn, and L. Meitner. "Uber die /? Strahlen des
activen Niederschlags des Thorium." PZ, 12 (1911), 273-9.
- "Nachweis von /? Strahlen bei Radium D." PZ, 12 (1911), 378-9.
- "Das magnetische Spektrum der /? Strahlen des Thoriums." PZ, 13
(1912), 264-6.
Barkla, C. G. "Secondary radiation from gases subject to x-rays." PM, 5
(1903X685-98.
- "Polarization in Rontgen rays." Nature, 69 (1904), 463.
- "Polarized Rontgen radiation." PTRS, 204A (1905), 467-79.
- "Secondary Rontgen radiation." PM, 11 (1906), 812-28.
- "Polarization in secondary Rontgen radiation." PRS, 77A (1906),
247-55-
- "The atomic weight of nickel." Nature, 75 (1907), 368.
- "The nature of x rays." Nature, 76 (1907), 661-62.
- "Note on x rays and scattered x rays." PM, 15 (1908), 288-96.
- "Der Stand der Forschung iiber die sekundare Rontgenstrahlung."
JRE, 5(1908), 246-324.
- "Phenomena of x ray transmission." PCPS, 15 (1909), 257-69. Ger-
man translation in JRE, 7(1910), 1-15.
- "The spectra of the fluorescent Rontgen radiations." PM, 22 (1911),
396-412. German translation in JRE, 8 (1911), 471-88.
- "Problems of radiation." Nature, 94 (1915), 671-2.
- "X-ray fluorescence and the quantum theory." Nature, 95 (1915), 7.
- "Bakerian Lecture - on x-rays and the theory of radiation." PTRS,
217A (191JI 315-60.
Barkla, C. G., and G. H. Martyn. "Interference of Rontgen radiation
(preliminary account)." PPSL, 25 (1913), 206-15.
Barkla, C. G., and J. Nicol, "Homogeneous fluorescent x radiations of a
second series." PPSL, 14 (1911), 9-17.
- "X ray spectra." Nature, 84 (1910), 139.
Barkla, C. G., and A. J. Philpot. "Ionization in gases and gaseous
mixtures by Rontgen and corpuscular (electronic) radiations." PM,
^(1913X832-56.
Barkla, C. G., and C. A. Sadler. "Secondary x rays and the atomic weight
of nickel." PM, 14 (1907), 408-22.
- "Homogeneous secondary radiations." PM, 16 (1908), 550-84.
Bibliography 311
- "The absorption of Rontgen rays." PM, 77 (1909), 85-94-
- 'The absorption of Rontgen rays." PM, 18 (1909), 739"6o.
Barkla, C. G., and G. Shearer. "Note on the velocity of electrons expelled
by x-rays." PM, 30 (1915), 745-53-
Bartlett, A. "Compton effect: Historical background." AJP, 32 (1964)
120-7.
Bassler, E. "Polarisation der X-Strahlen, nachgewiesen mittels Sekun-
darstrahlung." AP, 28 (1909), 808-84.
Bateman, H. "On the theory of light quanta." PM, 46 (1923), 977-91.
Beatty, R. T. "The production of cathode particles by homogeneous
Rontgen radiations." PCPS, 15 (1910), 416-22.
- "The production of cathode particles by homogeneous Rontgen radia-
tions, and their absorption by hydrogen and air." PM, 20 (1910),
320-30.
- "The energy of Rontgen rays." PRS, 89A (1913), 314-27.
Becquerel, H. "Sur les radiations invisibles emises par les corps phospho-
rescents." CR, 122 (1896), 501-3.
- "Sur quelques proprietes nouvelles des radiations invisibles emises par
divers corps phosphorescents." CR, 122 (1896), 559-64
- "Sur les radiations invisibles emises par les sels d'uranium." CR, 122
(1896X689-94.
- "Sur les proprietes differentes des radiations invisibles emises par les sel
d'uranium, et du rayonnement de la paroi anticathode d'un tube de
Crookes." CR, 122 (1896), 762-7.
- "Influence d'un champ magnetique sur le rayonnement des corps
radio-actifs." CR, 129 (1899), 996-1001.
- "Sur le rayonnement de l'uranium et sur diverses proprietes physiques
du rayonnement des corps radio-actifs." In Rapports presentes an
Congres International de Physique, C. Guillaume and L. Poincare,
eds., vol. 3, pp. 47-78, Paris, 1900.
- "Contribution a l'etude du rayonnement du radium." CR, 130 (1900),
206-11.
- "Sur la dispersion du rayonnement du radium dans un champ magne-
tique." CR, 130(i9Oo\ 372-76.
- "Deviation du rayonnement du radium dans un champ electrique."
CR, 130(i9oo), 809-15.
- "Sur la transparence de l'aluminum pour le rayonnement du radium."
CR, 130(i9oo), 1154-57.
- "Surlaradio-activitesecondairedesmetaux." CR, 132 (1901), 371-73.
- "Sur l'analyse magnetique des rayons du radium et du rayonnement
secondaire provoque par ces rayons." CR, 132 (1901), 1286-9.
- "Sur quelques effets chimiques produits par la rayonnement du ra-
dium." CR, 133 (1901), 709-12.
312 Bibliography
- "Sur quelques proprietes du rayonnement des corps radioactifs." CR,
134 (1902), 208-11.
- "Sur le rayonnement du polonium et du radium." CR, 136 (1903),
431-4-
- "Sur unepropriete des rayons a du radium." CR, 756(1903), 1517-22.
- "Uber die von der Strahlung radioactiver Korper hervorgerufene se-
kundare Strahlung." PZ, 5 (1904), 561-3.
- "Sur la radio-activite de la matiere." PRI, 77 (1906), 85-94.
Benoist, L., and D. Hurmuzescu. "Nouvelles proprietes des rayons x."
CR, 722(1896X235-6.
Birge, R. T. "Probable values of the general physical constants." PR
Supplement, 1 (1929), 1-73.
Black D. H. "The P ray spectrum of mesothorium 2." PRS, 106A (1924),
632-40.
- "/? ray spectra of thorium disintegration products." PRS, 109A (1925),
166-76.
Blondlot, R. "Sur la vitesse de propagation des rayons x." CR, 135 (1902),
666-70, and related studies, ibid., 721-4, 763-6, 1293-5.
Bohr, N. "On the constitution of atoms and molecules [Part I]." PM, 26
(1913), 1-25.
- "L'application de la theorie des quanta aux problemes atomiques." In
Solvay III (1923), 228-47.
- "Linienspektren und Atombau." AP, 71 (1923), 228-88.
- "Uber die Anwendung der Quantentheorie auf den Atombau. I. Die
Grundpostulate der Quantentheorie." ZP, 13 (1923), 117-65.
Translated as "On the application of the quantum theory to atomic
structure. Part I, The fundamental postulates of the quantum
theory." PCPS Supplement (1924).
- Collected works, vol. 2. Ulrich Hoyer, ed. Amsterdam, 1981.
Bohr, N., H. A. Kramers, and J. C. Slater. "The quantum theory of
radiation." PM, 47 (1924), 785-802.
Boltzmann, L. "Rontgen's neue Strahlen." Elektro-Techniker, 14 (1896),
385-9. Reported in Electrician, 36 (1896), 449. Reprinted in Boltz-
mann, Populdre Schriften, pp. 188-97. Leipzig, 1896.
- "Uber ein Medium, dessen mechanische Eigenschaften auf die von
Maxwell fur den Elektromagnetismus aufgestellen Gleichungen
fiihren." AP, 48 (1893), 78-99.
Born, M. "Zur Wellenmechanik der Stossvorgange." Gottingen Nach-
richten (1926), 146-60.
- "Zur Quantentheorie der Stossvorgange." ZP, 37 (1926), 863-7; 38
(1926), 803-27.
Bose, S. N. "Planck's Gesetz und Lichtquantenhypothese." ZP, 26
(1924), 178-81.
Bibliography 313
Bothe, W., and H. Geiger. "Ein Weg zur experimentellen Nachpriifung
der Theorie von Bohr, Kramers, und Slater." ZP, 26 (1924), 44.
Bragg, W. H. "The 'elastic medium' method of treating electrostatic
theorems." AusAAS Report, 5 (1891), 57-71.
- "Presidential address to Section A." AusAAS Report, 4 (1892), 31-47.
- "The energy of the electromagnetic field." AusAAS Report, 6 (1895),
223-31.
- "On some recent advances in the theory of the ionization of gases."
AusAAS Report, 10 (1904), 47-77.
- "On the ionization of various gases by the a particles of radium." PM,
/7(i9o6), 617-32.
- "A comparison of some forms of electric radiation." TPRRSSA, 31
(1907), 79-93-
- "The nature of Rontgen rays." TPRRSSA, 31 (1907), 94-8.
- "On the properties and natures of various electric radiations." PM, 14
(1907X429-49
- "Thenatureofyandxrays."Nature, 77(1908), 270-1,560;Nature, 78
(1908X271.
- "The consequences of the corpuscular hypothesis of the y and x rays,
and the range of /? rays." PM, 20 (1910), 385-416. German transla-
tion in JRE, 7(1910), 348-86.
- "The mode of ionization by x rays." PM, 22 (1911), 222-3.
- "Corpuscular radiation." BAAS Report, (1911), 340-41.
- "On the direct or indirect nature of the ionization by x rays." PM, 23
(1912X647-50.
- "Radiations old and new." BAAS Report, (1912), 750-3; Nature, 90
- Studies in radioactivity. London, 1912.

- "X rays and crystals." Nature, 90 (1912), 219, 360-1.
- "The reflection of x rays by crystals." PRS, 89A (1913), 246-8.
- "Aether waves and electrons." Nature, 107 (1921), 374.
- "Craftsmanship and science." BAAS Report, 98 (1928), 1-20.
Bragg, W. H., and W. L. Bragg. "The reflection of x rays by crystals."
PRS, S&4 (1913X428-38.
- X rays and crystal structure. London, 1915.
Bragg, W. H., and J. L. Glasson. "On a want of symmetry shown by
secondary x rays." TPRRSSA, 32 (1908), 301-10; PM, 17 (1909),
855-64.
Bragg, W. H., and J. P. V. Madsen. "An experimental investigation of the
nature of the y rays - No. 1." TPRRSSA, 32 (1908), 1 -10; PM, 15
(1908X663-75.
- "An experimental investigation of the nature of the y rays - No. 2."
TPRRSSA, 32 (1908), 35-54; PM, 16 (1908), 918-939.
314 Bibliography
Bragg, W. H., and S. E. Peirce. "The absorption coefficients of x rays."
PM, 2^(1914), 626-30.
Bragg, W. H., and H. L. Porter. "Energy transformations of x rays." PRS,
55.4(1911), 349-65.
Bragg, W. L. "The diffraction of short electromagnetic waves by a
crystal." PCPS, 77(1912), 43-57.
- The start of x ray analysis. Chemistry background book. London, 1967.
Breit, G. "The propagation of a fan-shaped group of waves in a dispersing
medium." PM, 44(1922), 1149-52.
Bridgman, P. W. "Biographical memoir of William Duane, 1872-1935."
NAS Biographical memoirs, 18 (1937), 23-41.
Brillouin, L. "L'agitation moleculaire et les lois du rayonnement ther-
mique." JP, 2(1921), 142-55.
Brillouin, M. "Actions mechaniques a heredite discontinue par propaga-
tion: Essai de theorie dynamique de 1' atom a quanta." CR, 168
(1919), 1318-20.
de Broglie, L. "Sur la degradation du quantum dans les transformations
successives des radiations de haute frequence." CR, 173 (1921),
1160-2.
- "Sur la theorie de l'absorption des rayons x par la matiere et le principe
de correspondance." CR, 173 (1921), 1456-8.
- "Considerations theoriques sur l'absorption des rayons x par la ma-
tiere." SFP PVRC, 1921 (1922), 15-16.
- "Rayons x et equilibre thermodynamique." JP, 3 (1922), 33-45.
- "Rayonnement noir et quanta de lumiere." JP, 3 (1922), 422-8.
- "Sur les interferences et la theorie des quanta de lumiere." CR, 175
(1922), 8 1 1 - 1 3 .
- "Ondes et quanta." CR, 777 (1923), 507-10.
- "Quanta de lumiere, diffraction et interferences." CR, 177 (1923),
548-50.
- "Les quanta, le theorie cinetique des gaz et le principe de Fermat." CR,
777(1923X630-2.
- "A tentative theory of light quanta." PM, 47 (1924), 446-58.
- Theses. Recherches sur la theorie des quanta. Paris, 1924. Republished
in Annales de physique, 3 (1925), 22-128, with a useful historical
introduction by de Broglie, pp. 22-30. German translation, Unter-
suchungen zur Quantentheorie. Leipzig, 1927. The French text was
republished by Masson. Paris, 1963.
- "Allocution prononcee au jubile de M. Maurice de Broglie le 13 Juin
1946." Savants et decouverts, pp. 29J8-305. Paris, 1951.
- "Vue d'ensemble sur mes travaux scientifiques." In Louis de Broglie:
Physicien et penseur, A. Georges, ed., pp. 457-86. Paris, 1953.
- Le dualisme des ondes et les corpuscules dans I'oeuvre de Albert
Einstein. Paris, 1955.
Bibliography 315
- "Une rencontre avec Einstein au conseil Solvay en 1927." In Nouvelles
perspectives en microphysique, pp. 232-7. Paris, 1956.
- "The beginnings of wave mechanics." In Wave mechanics: The first
fifty years, W. C. Price, S. S. Chissick, and T. Ravensdale, eds., pp.
12-18. New York, 1973.
- "Avant propos." Recherches d'un demi-siecle, pp. 7-16. Paris, 1976.
de Broglie, L., and A Dauvillier. "Sur le systeme spectral des rayons
Rontgen." CR, 175 (1922), 685-8.
- "Sur les analogies de structure entre les series optiques et les series de
Rontgen." CR, 175 (1922), 755-7.
de Broglie, M. Theses. Recherches sur les centres electrises de faible
mobilite dans les gaz. Paris, 1908.
- "Sur un nouveau procede permettant d'obtenir la photographie des
spectres de raies des rayons de Rontgen." CR, 157 (1913), 924-6.
- "Sur la spectroscopie des rayons secondaries emis hors des tube a
rayons de Rontgen, et les spectres d'absorption," CR, 158 (1914),
1493-5-
- "La nature des rayons de "Rontgen . . ." Revue scientifique, 53
- "Sur une systeme de bandes d'absorption correspondant aux rayons

L des spectres de rayons x des elements . . ." CR, 163 (1916),
352-5.
- "La portee des nouvelles decouvertes dans la region des rayons de tres
haute frequence." Scientia, 27 (1920), 102-11.
- "Les ecrans renforgateurs et le spectre des rayons x." SFP PVRC, 1920
(1920,32-3.
- "Sur les spectres corpusculaires des elements." CR, 772(1921), 274-5.
- "Remarques sur l'equation photo-electrique d'Einstein." SFP PVRC,
7927(1922), 10, 13-14.
- "Sur les spectres corpusculaires des elements." CR, 772(1921), 527-9.
- "Sur les spectres corpusculaires. Lois de remission photo-electrique
pour les hautes frequences." CR, 172 (1921), 806-7.
- "Les phenomenes photo-electriques correspondant aux rayons x et les
spectres corpusculaires des elements." SFP PVRC, 1921 (1922),
48-9.
- "Les phenomenes photo-electriques pour les rayons x et les spectres
corpusculaires des elements." JP, 2 (1921), 265-87.
- "Sur les spectres corpusculaires et leur utilization pour Petude des
spectres de rayons x." CR, 173 (1921), 1157-60.
- Les rayons x. Paris, 1922.
- "La relation hv = e dans les phenomenes photoelectriques." In Solvay
HI, pp. 80-100.
- "Le prince Louis de Broglie enfant." Grandes souvenirs, belles actual-
ites: Le recueil dujeunes, 2 (July 1948), 4-6.
316 Bibliography
- "La jeunesse et les orientations intellectuelles de Louis de Broglie." In
Louis de Broglie: Physicien et penseur, A. Georges, ed., pp. 423-9.
Paris, 1953.
de Broglie, ML, and L. de Broglie. "Sur le modele d'atom de Bohr et les
spectres corpusculaires." CR, 172 (1921), 746-8.
- "Sur les spectres corusculaires des elements." CR, 173 (1921), 527-9.
- "Remarques sur les spectres corpuscularies et Feffet photo-electrique."
CR, 775(1922), 1139-41.
de Broglie, M., and J. Cabrera. "Sur les rayons gamma de la famille du
radium et du thorium etudies par leur effet photo-electrique." CR,
776(1923), 295-6.
Brunhes, B. "Une methode de mesure de la vitesse des rayons Rontgen."
CR, 750(1900), 127-30.
Buchwald, E. "Zur Berechnung der y Strahlschwankungen." AP, 39
(1912X41-52.
Buisson, H., and C. Fabry. "Sur la mesure des intensites des diverses
radiations d'un rayonnement complexe." CR, 152 (1911), 1838-41.
Bumstead, H. A. "The heating effects produced by Rontgen rays in
different metals, and their relation to the question of change in the
atom." PM, 11 (1906), 292-317.
Burbidge, P. W. "The fluctuation in the ionization due to y rays." PRS,
59^(1913), 45-57.
Campbell, N. Modern electrical theory. Cambridge, 1907.
- "The study of discontinuous phenomena." PCPS, 15 (1909), 117-36.
- "Discontinuities in light emission." PCPS, 15 (1909), 310-28; 15
(1910), 513-25.
- "The aether." PM, 19 (1910), 181-91. German translation in JRE, 7
(1910), 15-28.
- "Uber Schweidlersche Schwankungen." PZ, 11 (1910), 826-33.
- "Nachschrift." PZ, 13 (1912), 81-3.
Carvallo, E. "Sur la nature de la lumiere blanche." CR, 130 (1900),
79-82.
- "Sur la nature de la lumiere blanche et des rayons x." CR, 130 (1900),
130-2.
Cassidy, D. "Heisenberg's first core model of the atom: The formation of
a professional style." HSPS, 70(1979), 187-224.
Chadwick, J. "Intensitatsverteilung im magnetischen Spektrum der /?
Strahlen von Radium B + C." VDpG, 16 (1914), 383-91.
- "The charge on the atomic nucleus and the law of force." PM, 40
(1920), 734-46.
Chwolson. See KhvoFson.
Compton, A. H. "The size and shape of the electron." WAS Journal, 8
(1918), 2-11, PR, 7^(1919), 20-43.
- "The degradation of gamma ray energy." PM, 41 (1921), 749-67.
Bibliography 317
- "The absorption ofy rays by magnetized iron." PR, 77(1921X38-41.
- "Secondary high frequency radiation." PR, 18 (1921), 96.
- "The spectrum of secondary x rays." PR, 19 (1922), 267-8.
- "Secondary radiations produced by x-rays, and some of their applica-
tions to physical problems." NRC Bulletin, 4 (1922), 1-56.
- "A quantum theory of the scattering of x rays by light elements." PR,
27(1923), 483-502.
- "The scattering of x rays." JFI, 198 (1924), 57-72.
Compton, A. H., and A. W. Simon. "Measurements of /?-rays associated
with scattered x-rays." PR, 25 (1925), 306-13.
- "Directed quanta of scattered x-rays." PR, 26 (1925), 289-99.
Compton, K. T., and O. Richardson. "The photoelectric effect II." PM,
26 (1913X 549-67.
Cooksey, C. D. "On the corpuscular rays produced in different metals by
Rontgen rays." AJS, 24 (1907), 285-304.
- "The nature of y and x rays." Nature, 77 (1908), 509-10.
- "On the asymmetry in the distribution of secondary cathode rays
produced by the x rays; and its dependence on the penetrating power
of the exciting rays." PM, 24 (1912), 37-45.
Coolidge, W. "A powerful Rontgen ray tube with a pure electron dis-
charge." PR, 2 (1913), 409-30.
Des Coudres, T. "Uber Kathodenstrahlen unter dem Einflusse magne-
tischer Schwingungen." VDNA, 7^(1895), 86-8.
- "Elektrodynamisches uber Kathodenstrahlen." VDNA, 16 (1897),
157-62.
Crookes, W. "Bakerian lecture on the illumination of lines of molecular
pressure and the trajectories of molecules." PTRS, 170:1 (1879),
135-64.
- On radiant matter. London, 1879.
- "Radio-activity of uranium." PRS, 66 (1900), 409-23.
Curie, M. Theses. Recherches sur les substances radioactives. Paris, 1903.
Reprinted in the Mallinkrodt classics of radiology. St. Louis, 1966.
Curie, P. "Action du champ magnetique sur les rayons de Becquerel.
Rayons devies et rayons non devies." CR, 130 (1900), 73-6.
Curie, P., and M. Curie. "Sur une substance nouvelle radio-active,
contenue dans la pechblende. CR, 727(1898), 175-8.
- "Sur la radioactivite provoquee par les rayons de Becquerel." CR, 129
(1898), 714-16.
- "Les nouvelles substances radioactives et les rayons qu'elles emittent."
In Congres international de physique, C. Guillaume and L. Poincare,
eds., vol. 3, pp. 79-114. Paris, 1900.
Curie, P., and G. Sagnac. "Electrisation negative des rayons secondaires
produits au moyen des rayons Rontgen." CR, 130 (1900),
1013-15.
318 Bibliography
Danysz, J. "Sur les rayons ft de la famille du radium." CR, 153 (1911),
339-41, 1066-8; Le Radium, 9(1912), 1-5.
- "Sur les rayons /? des radiums B, C, D." Le Radium, 10 (1913), 4-6.
Danysz, J., and W. Duane. "Sur les charges electroniques transportees
par les rayons a et/?." CR, 155 (1912), 500-3.
Dauvillier, A. Theses. Recherches spectrometriques sur les rayons x.
Paris, 1920.
- La technique des rayons x. Paris, 1924.
Dauvillier, A., and L. de Broglie. "Sur la structure electronique des
atomes lourds." CR, 172 (1921), 1650-3.
- "Sur la distribution des electrons dans les atomes lourdes." CR, 173
(1921), 137-9.
Davisson, C. "Note on radiation due to impact of)? particles upon solid
matter." PR, 28 (1909), 469-70.
Davisson, C , and L. H. Germer. "The scattering of electrons by a single
crystal of nickel." Nature, 119 (1927), 558-60.
Davisson, C , and C. H. Kunsman. "The scattering of low speed electrons
by platinum and magnesium." PR, 22 (1923), 242-58.
Debye, P. "Zustandsgleichung und Quantenhypothese mit einem An-
hang iiber Warmeleitung." In Vortrdge uber die kinetische Theorie
derMaterie undderElektrizitdt. Mathematische Vorlesungen an der
Universitdt Gottingen, D. Hilbert, ed., vol. 6, pp. 17-60. Leipzig,
1914. An abstract appeared in PZ, 14 (1913), 259-260.
- "Uber den Einfluss der Warmebewegung auf die Interferenzerschei-
nungen bei Rontgenstrahlen." KD/?G, 75 (1913), 678-89.
- "Uber die Intensitatsverteilung in den mit Rontgenstrahlen erzeugten
Interferenzbildern." VDpG, 75(1913X738-52.
- "Spektrale Zerlegung der Rontgenstrahlen mittels Reflexion und War-
mebewegung." VDpG, 15 (1913), 857-75.
- "Zerstreuung von Rontgenstrahlen und Quantentheorie." PZ, 24
(1923), 161-6.
Debye, P., and A. Sommerfeld. "Theorie des lichtelektrischen Effektes
vom Standpunkt des Wirkungquantums." AP, 41 (1913), 873-
930.
Donath, B. Die Einrichtung zur Erzeugung der Roentgenstrahlen und ihr
Gebrauch. Berlin, 1899.
Dorn, E. "Uber die erwarmende Wirkung der Rontgenstrahlen." AP, 63
(1897), 160-76.
- "Versuche uber Sekundarstrahlen." Archives Neerlandaises (2), 5
(1900), 595-608. Issued separately as Recueil de travaux offerts par
les auteurs aH. A. Lorentz. The Hague, 1900.
- "Versuche iiber Sekundarstrahlen und Radiumstralilen." Halle Ab-
handlungen, 22(1901), 39-43.
Bibliography 319
- "Elektrostatische Ablenkung der Radiumstrahlen." Halle Abhandlun-
gen, 22(1901), 47-50.
- "Eine merkwtirdige Beobachtung mit Radium." PZ, 4 (1903), 507-8.
Drude, P. "Zur Geschichte der elektromagnetischen Dispersionsgleich-
ungen." AP, 1 (1900), 437-40.
Duane, W. "Relation between the wavelength and absorption of x rays."
Title only in PR, 4 (1914), 544.
- "On [the] relation between the wavelengths of x rays and the voltages
required to produce them." PR, 6 (1915), 166-76.
- "Planck's radiation formula deduced from hypotheses suggested by x
ray phenomena." PR, 7(1916), 143-7.
- "Radiation and matter." Science, 46 (1917), 347-9.
- "The transfer in quanta of radiation momentum to matter." NAS
Proceedings, 9(1923), 158-64.
Duane, W., and K. F. Hu. "On the critical absorption and characteristic
emission x-ray frequencies." PR, 11 (1918), 489-91.
Duane, W., and F. L. Hunt. "On x-ray wave-lengths." PR, 6(1915), page
numbers repeat in this volume; first 166-71.
Eckart, C. "Operator calculus and the solutions of the equations of
quantum dynamics." PR, 28 (1926), 711-26.
Edison, T. "Are Rontgen ray phenomena due to sound waves?" Electri-
cal engineer, 21 (1896), 353-4.
Einstein, A. "Uber einen Erzeugung und Verwandlung des Lichtes be-
treffenden heuristischen Gesichtspunkt." AP, 77 (1905), 132-48.
- "Zur Theorie der Lichterzeugung und Lichtabsorption." AP, 20
(1906), 199-206.
- "Zum gegenwartigen Stand des Strahlungsproblems." PZ, 10 (1909),
185-93. See also the discussions between Einstein and Ritz, ibid.,
224-5,323-4.
- "Uber die Entwicklung unserer Anschauungen uber das Wesen und
die Konstitution der Strahlung." PZ, 10 (1909), 817-25.
- Contribution to the discussion following Sommerfeld's presentation at
the 1912 Naturforscherversammlung. PZ, 72 (1911), 1068-9.
- "Zur Quantentheorie der Strahlung." Zurich Mitteilungen, 7S (1916),
47-62. Reprinted in PZ, 75 (1917), 121-8.
- "Uber ein den Elementarprozess der Lichtemission betreffendes Ex-
periment." Berlin Sb (1921), 882-3.
- "Zur Theorie der Lichtfortpflanzung in dispergierenden Medien."
Berlin Sb (1922), 18-22.
- "Bietet die Feldtheorie Moglichkeiten fur die Losung des Quanten-
problems?" Berlin Sb (1923), 359-64.
- "Quantentheorie des einatomigen idealen Gases." Berlin Sb (1924),
261-7; (1925), 3-14.
320 Bibliography
Einstein, A., and J. Laub. "Uber die Elektromagnetische Grundglei-
chungen fur bewegte Korper." AP, 26 (1908), 532-40.
- "Uber die im elektromagnetischen Felde auf ruhende Korper aus-
geiibten ponderomotorischen Krafte." AP, 26 (1908), 541-50.
Eisberg, R. M. Fundamentals of modern physics. New York, 1961.
Ellis, C. D. "The magnetic spectrum of the ft rays excited by y rays." PRS,
99A (1921), 261-71.
- "/? ray spectra and their meaning." PRS, 101A (1922), 1 -17.
- "Uber die Deutung der /?-Strahlspektren radioaktiver Substanzen."
ZP, 70(1922), 303-7.
- "The high energy groups in the magnetic spectrum of the radium C /?
rays." PCPS, 22 (1924), 369-78.
Ellis, C. D., and H. W. B. Skinner. "The interpretation of)? ray spectra."
PRS, 105A (1924), 185-98.
Elsasser, W. "Bemerkungen zur Quantenmechanik freier Elektronen."
Die Naturwissenschaften, 75(1925), 711.
Elster, J., and H. Geitel. "Der photoelektrische Effekt am Kalium bei sehr
geringen Lichtstarken." PZ, 13 (1912), 468-76.
Emden, R. "Uber Lichtquanten." PZ, 22 (1921), 513-17.
Epstein, P. S. "Versuch einer Anwendung der Quantenlehre auf die
Theorie des lichtelektrischen Effektes und der ft Strahlung radioak-
tiver Substanzen." PZ, 77(1916), 313-16; AP, 50(1916), 815-40.
Eve, A. S. "Rontgen rays and the y rays from radium." Nature, 69 (1904),
436.
- "A comparison of the ionization produced in gases by penetrating
Rontgen and radium rays." PM, 8 (1904), 610-18.
- "On the secondary radiation caused by the P and y rays of radium."
PM, 5(1904), 669-85.
- "On the secondary radiation due to the y rays of radium." Nature, 70
(1904X454.
- "The absorption of the y rays of radioactive substances." PM, 11
(1906), 586-95.
Faxen, H., and J. P. Holtsmark. "Beitrag zur Theorie des Durchganges
langsamer Elektronen durch Gase." ZP, 45 (1927), 307-24.
Florance, D. C. H. "Primary and secondary y rays." PM, 20 (1910),
921-38.
Fomm, L. "Die Wellenlange der Rontgen-Strahlen." Munchen Sb, 26
(1896), 283-6.
Forman, P. "The doublet riddle and atomic physics circa 1924." his, 59
(1968), 156-74.
- "The discovery of the diffraction of x-rays by crystals: A critique of the
myths." AHES, 6 (1969), 38-71.
Bibliography 321
- "Alfred Lande and the anomalous Zeeman effect." HSPS, 2 (1970),
153-261.
- "Weimar culture, causality, and quantum theory, 1918 -1927: Adapta-
tion by German physicists and mathematicians to a hostile intellec-
tual environment." HSPS, 3 (1971), 1 -115.
- "Bragg, William Henry." DSB, 2 (1973), 397-400.
- "Duane, William." DSB, 4 (1971), 194-197.
Forman P., J. L. Heilbron, and S. Weart. "Physics circa 1900." HSPS, 5
(1975), 1-185.
Franck, J., and G. Hertz. "Uber Zusammenstosse zwischen Elektronen
und den Molekiilen des Quecksilberdampfes und die Ionisierungs-
spannung derselben." VDpG, 14 (1912), 457-67.
- "Uber die Erregung der Quecksilberresonanzlinie 253,6///* durch Elek-
tronenstosse." VDpG, 14 (igi2), 512-17.
Franck, J., and R. Pohl. "Zur Frage nach der Geschwindigkeit der
Rontgenstrahlen." VDpG, 10 (1908), 117-36,489-94.
- "Bemerkung zu den Versuchen des Hrn. Marx uber die Geschwindig-
keit der Rontgenstrahlen." AP, 34 (1911), 936-40.
Frank, P. Einstein, his life and times. New York, 1953.
Friedman, R. "Nobel physics prize in perspective." Nature, 292 (1981),
793-8.
Friedrich, W. "Rontgenstrahlinterferenzen." PZ, 14 (1913), 1079-84.
Friedrich, W., P. Knipping, and M. Laue. "Interferenz-Erscheinungen
bei Rontgenstrahlen." Miinchen Sb, 42 (1912), 303-22. Reprinted
with additional notes in AP, 41 (1913), 971-88.
Frilley, M. Theses. Spectrographie par diffraction cristalline des rayons y
et lafamille de radium. Paris, 1928.
Fiichtbauer, C. "Uber die Geschwindigkeit der von Kanalstrahlen und
von Kathodenstrahlen beim Auftreffen auf Metalle erzeugten nega-
tiven Strahlen." PZ, 7(1906), 748-50.
Gehrenbeck, R. K. C. J. Davisson, L. H. Germer, and the discovery of
electron-diffraction. Unpublished Ph.D dissertation, University of
Minnesota, 1973.
Geiger, H. "The irregularities in the radiation from radioactive bodies."
PM, 75(i 9 o8),539-47. ..
Geiger, H., and W. Bothe. "Uber das Wesen des Comptoneffektes, ein
experimenteller Beitrag zur Theorie der Strahlung." ZP, 32 (1925),
639-63.
Georges, Andre, ed. Louis de Broglie: Physician et penseur. Paris, 1953.
Giesel, F. O. "Uber die Ablenkbarkeit der Becquerelstrahlen im magne-
tischen Felde," AP, 69 (1899), 834-7.
Gifford, J. W. "Are Rontgen rays polarized?" Nature, 54 (1896), 172.
322 Bibliography
Gocht, H. Die Rontgen-Literatur, 5 vols. Stuttgart, 1911 -25.
Godlewski, T. "On the absorption of the /? and y rays of actinium." PM,
70(1905), 375-9.
Goldstein, E. "Uber die Entladung der Elektrizitat in verdiinnten
Gasen." Berlin Mb (1888), 82-124.
Gouy, L. G. "Sur le mouvement lumineux." JP, 5 (1886), 354-62.
- "Sur le refraction et la diffraction des rayons x." CR, 123 (1896),
43-4-
- "Sur la constitution de la lumiere blanche." CR, 130 (1900), 241-4.
- "Sur le mouvement lumineux et les formules de Fourier." CR, 130
(1900), 560-2.
Gray, J. A. "Secondary y rays produced by /? rays." PRS, 85A (1911),
131-9.
- "The nature of the y rays excited by /? rays." PRS, 86A (1912), 513 - 29.
- "The similarity in nature of x and primary y rays." PRS, 87A (1912),
489-501.
- "The scattering and absorption of the y rays of radium. "PM, 26(1913),
611-23.
Green, G. "On the laws of the reflection and refraction of light at the
common surface of two non-crystallized media." TCPS, 7(1839),
2-24.
- "On the propagation of light in crystallized media." TCPS, 7 (1839),
121-40.
Guillaume, C. E. Les rayons x. Paris, 1896.
Guillaume, C. E., and L. Poincare, eds. Rapports presentes an congres
international de physique. 3 vols. Paris, 1900.
Haga, H., and C. H. Wind. "De buiging van x-stralen." Amsterdam
Verslag, 7(1899), 387-8.
- "De buiging der Rontgen-stralen." Amsterdam Verslag, 7 (1899),
500-7. An English translation appeared in Amsterdam Proceedings,
1 (1899), 321, 420-6. A German version appeared as "Die Beugung
der Rontgenstrahlen." AP, 68 (1899), 884-95.
- "Die Beugung der Rontgenstrahlen." AP, 10 (1903), 305-12.
Hanle, P. "Erwin Schrodinger's reaction to Louis de Broglie's thesis on
the quantum theory." Isis, 68 (1977), 606-9.
Heaviside, O. "On the electromagnetic effects due to the motion of
electrification through a dielectric." PM, 27 (1889), 324-39.
Heilbron, J. L. A history of the problem of atomic structure from the
discovery of the electron to the beginning of quantum mechanics.
Unpublished Ph.D. dissertation, University of California, Berkeley,
1964.
- "The Kossel-Sommerfeld theory and the ring atom." Isis, 58 (1967),
451-85.
Bibliography 323
- "The scattering of a and /? particles and Rutherford's atom." AHES, 4
(1968), 247-307.
- H. G. J. Moseley: The life and letters of an English physicist,
1887-1915. Berkeley, 1974.
- "Lectures on the history of atomic physics." In History of twentieth
century physics, C. Weiner, ed., pp. 40-108. New York, 1977.
- "Physics at McGill in Rutherford's time." In Rutherford and physics at
the turn ofthe century, M. Bunge and W. Shea, eds., pp. 42 - 73. New
York, 1979.
Heilbron, J. L., and T. S. Kuhn. "The genesis of the Bohr atom." HSPS, 1
(1969), 211-90.
Heilbron, J. L., and B. R. Wheaton. Literature on the history of physics in
the 20th century. Berkeley papers in history of science, vol 5. Berke-
ley, 1981.
Heisenberg, W. "Uber quantentheoretische Umdeutung kinematischer
und mechanischer Beziehungen." ZP, 33 (1925), 879-93.
- "Uber den anschaulichen Inhalt der quantentheoretischen Kinematik
und Mechanik." ZP, 43 (1927), 172-98.
Heisenberg, W., and H. A. Kramers. "Uber die Streuung von Strahlen
durch Atome." ZP, 31 (1925), 681-708.
Helmholtz, H. von. Vorlesungen uber die elektromagnetische Theoriedes
Lichts. Hamburg, 1897.
Hendry, J. "The development of attitudes toward the wave-particle
duality of light and quantum theory, 1900-1920." Annals ofScience,
37 (1980), 59-79.
Hermann, A. "Albert Einstein und Johannes Stark: Briefwechsel und
Verhaltnis der beiden Nobelpreistrager." Sudhoffs Archiv, 50
(1966), 267-85.
- "Die friihe Diskussion zwischen Stark und Sommerfeld uber die
Quantenhypothese." Centaurus, 12 (1968), 38-59.
- Frilhgeschichte der Quantentheorie 1889-1913. Mosbach, 1969.
Hermann, K. "Zur Theorie des lichtelektrischen Effektes." VDpG, 14
(1912X936-45.
Hertz, H. "Versuche uber die Glimmentladung." AP, 19 (1883),
782-816.
- "Uber die Ausbreitungsgeschwindigkeit der elektrodynamischen Wir-
kung." AP, 34 (1888), 551-69. A first announcement appeared in
Berlin Sb (1888), 197-210.
- "Uber Strahlen elektrischer Kraft," AP, 36 (1889), 769-83.
- "Uber die Beziehungen zwischen Licht und Elektrizitat." In GDNA
Tagehlatt, pp. 144-9. Heidelberg, 1890.
- Die Principien der Mechanik, in neuem Zusammenhange. Leipzig,
1894.
324 Bibliography
- Gesammelte Werke, vol. 3, Schriften vermischten Inhalts, P. Lenard,
ed. Leipzig, 1895.
Herweg, J. "Uber die Beugungserscheinungen der Rontgenstrahlen am
Gips." PZ, 14 (1913), 417-20.
Hirosige, T. "Origins of Lorentz' theory of electrons and the concept of
the electromagnetic field." HSPS, 1 (1969), 151-209.
Home, R. W. "W. H. Bragg and J. P. V. Madsen: Collaboration and
correspondence, 1905-1911." Historical records of Australian
science, 5:2 (1981), 1-29.
Hu, K. F. "Some preliminary results in a determination of the maximum
emission velocity of the photoelectrons from metals at x-ray fre-
quencies." PR, 11 (1918), 505-7.
Hughes, A. L. "On the velocities of the electrons produced by ultra-violet
light." PCPS, 76(i9ii), 167-74.
- "On the emission velocities of photo-electrons." PTRS, 212A (1913),
209-26.
Hull, A., and M. Rice. "The law of absorption of x rays at high frequen-
cies." PR, 8 (1916), 326-8.
Hund, F., "Theoretische Betrachtungen iiber die Ablenkung von freien
langsamen Elektronen in Atomen." ZP, 13 (1923), 241-63.
Ikle, M. "Literatur der Radioaktivitat vor dem Jahre 1904." JRE, 1
(1904), 413-42.
Innes, P. D. "On the velocity of the cathode particles emitted by var-
ious metals under the influence of Rontgen rays, and its bear-
ing on the theory of atomic disintegration." PRS, 79A (1907),
442-62.
Ioffe, A. F. [Joffe]. "Eine Bemerkung zu der Arbeit von E. Ladenburg:
'Uber Anfangsgeschwindigkeit und Menge der photoelektrischen
Elektronen usw.V AP, 24 (1907), 939-40.
- "Beobachtungen iiber den photoelektrischen Elementareffekt." Miin-
chen Sb, 43 (1913), 19-37.
Jammer, M. The conceptual development of quantum mechanics. New
York, 1966.
Jaumann, G. "Longitudinales Licht." Wien Sb, 104:11a (1895), 747-92;
revised in AP, 57(1895), 147-84.
- "Uber die Interferenz und die elektrostatische Ablenkung der Katho-
denstrahlen." Wien Sb, 106:11a (1897), 533-50.
Jeans, J. H. "Discussion of radiation." BAAS Report, (1913), 380.
- Report on radiation and the quantum theory. London, 1914.
Jenkin, J. G., R. C. G. Leckey, and J. Liesegang. "The development of
x-ray photoelectron spectroscopy: 1900-1960." Journal of electron
spectroscopy and related phenomena, 72(1977), I ~35-
Bibliography 325
Joffe. See Ioffe.
Kahan, T. "Un document historique de Facademie des sciences de Berlin
sur l'activite scientifique d' Albert Einstein." Archives internation-
ales d'historie des sciences, 15 (1962), 337-42.
Kargon, R. H. "Model and analogy in Victorian science: Maxwell's
critique of the French physicists." JHI, 30 (1969), 423-36.
Karnozhitzkiy, A. N., and B. Golitsyn. "Researches concerning the
properties of x-rays." Nature, 53 (1896), 528; described by Stokes,
"On the nature of the Rontgen rays." Manchester LP, 41 (1897), no.
15, P- 2.
Kaufmann, W. "Die magnetische und elektrische Ablenkbarkeit der
Becquerelstrahlen und die scheinbare Masse der Elektronen." Got-
tingen Nachrichten, (1901), 143-55.
Kaufmann, W., and E. Aschkinass. "Uber die Deflexion der Kathoden-
strahlen." AP, 62 (1897), 588-95.
Kaye, G. W. C. "The emission and transmission of Rontgen rays."
PTRS, 209A (1908), 123-51.
- "The emission of Rontgen rays from thin metallic sheets." PCPS, 15
(1909), 269-72.
Kelvin. See Thomson, William.
KhvoFson, O. D. [Chwolson]. Lehrbuch derPhysik, vol. 2, H. Pflaum, tr.
Braunschweig, 1904.
Kirchhoff, G. "Zur Theorie der Lichtstrahlen." Berlin Mb (1882),
641-69.
- Vorlesungen uber mathematische Physik, vol. 2, Mathematische
Optik, K. Hensel, ed. Leipzig, 1891.
Kleeman, R. D. "On the recombination of ions made by a, /?, and x
rays." PM, 12 (1906), 273-97.
- "On the ionisation of various gases by a-, /?-, and y-rays." PRS, 79A
(1907X220-33.
- "On the secondary cathode rays emitted by substances when exposed
to the y rays." PM, 14 (1907), 618-44.
- "On the different kinds of y rays of radium, and the secondary y rays
which they produce." PM, 15 (1908), 638-63.
- "A difference in the photoelectric effect caused by incident and diver-
gent light." Nature, 83 (1910), 339.
- "On the direction of motion of an electron ejected from an atom by
ultra-violet light." PRS, 84A (1910), 92-9.
Klein, M. J. "Max Planck and the beginnings of the quantum theory."
AHES, 1 (1962), 459-79.
- "Einstein's first paper on quanta." The natural philosopher, 2 (1963),
59-86.
326 Bibliography
- "Einstein and the wave particle duality;" The natural philosopher, 3
(1964), 3-49.
- "Einstein, specific heats and the early quantum theory." Science, 148
(1965), 173-80.
- "Thermodynamics in Einstein's thought. Science, 757(1967), 509-16.
- "The first phase of the Bohr-Einstein dialogue." HSPS, 2 (1970),
1-39.
- "Mechanical explanations at the end of the nineteenth century."
Centaurus, 77(1972), 58-82.
Kleinert, A., and C. Schonbeck. "Lenard und Einstein. Ihr Briefwechsel
und ihr Verhaltnis vor der Nauheimer Diskussion von 1920." Ges-
nerus, 35 (1978), 318-33.
Klickstein, H. Wilhelm Conrad Rontgen. On a new kind of rays. Mal-
linkrodt classics of radiology, vol. 1. St. Louis, 1966.
Koch, P. P. "Uber die Messung der Schwarzung photographischer Plat-
ten in sehr schmalen Bereichen. Mit Anwendung auf die Messung
der Schwarzungsverteilung in einigen mit Rontgenstrahlen aufge-
nommenen Spaltphotogrammen von Walter und Pohl." AP, 38
(1912X507-22.
Kohlrausch, K. W. F. "Uber Schwankungen der radioaktiven Umwand-
lung." Wien Sb, 115:11a (1906), 673-82.
Kossel, W. "Uber die sekundare Kathodenstrahlung in Gasen in der
Nahe des Optimums der Primargeschwindigkeit." AP, 37 (1912),
393-424.
- "Bemerkung zur Absorption homogener Rontgenstrahlen." VDpG, 16
(1914), 898-909, 953-63.
Kovarik, A. "On the high frequency rays in the y ray spectrum of
radium." PR, 19 (1922), 433.
Kramers, H. A. "The law of dispersion and Bohr's theory of spectra."
Nature, 113 (1924), 673-4.
Kronig, R. "The turning point." In Theoretical physics in the twentieth
century: A memorial volume to Wolfgang Pauli, M. Fierz and
V. Weisskopf, eds., pp. 5-39. New York, i960.
Kubli, F. "Louis de Broglie und die Entdeckung der Materiewellen."
AHES, 7(1970X26-68.
Kuhn, T. S. "The quantum theory of specific heats: A problem in
professional recognition." XIV congres internationale d'histoire des
sciences, Tokyo, Proceedings, 1 (1974), 183-94.
- Black-body theory and the quantum discontinuity, 1894-1912. New
York, 1978.
Kuhn, T. S., J. L. Heilbron, P. Forman, and L. Allen. Sources for history
of quantum physics: An inventory and report. Philadelphia, 1967.
Bibliography 327
Kunz, J. Theoretische Physik auf mechanischer Grundlage. Stuttgart,
1907.
Laby, T. H. "A string electrometer." PCPS, 15 (1909), 106-13.
Laby, T. H., and P. Burbidge. "The nature of y rays." Nature, #7(191 1),
144.
- 'The observation by means of a string electrometer of fluctuations in
the ionization produced by y rays." PRS, 86A (1912), 333-48.
Ladenburg, E. "Uber Anfangsgeschwindigkeit und Menge der photo-
elektrischen Elektronen in ihrem Zusammenhange mit der Wellen-
lange des auslosenden Lichtes." VDpG, 9 (1907), 504-14.
Ladenburg, R. "Die neuren Forschungen uber die durch Licht- und
Rontgenstrahlen hervorgerufene Emission negativer Elektronen."
JRE, 6 (1909), 425-84.
- "Die quantentheoretische Deutung der Zahl der Dispersionselek-
trons." ZP, 4 (1921), 451-68.
Lande, A. "Schwierigkeiten in der Quantentheorie des Atombaues, be-
sonders magnetischer Art." PZ, 24 (1923), 441-4.
Langevin, P., and M. de Broglie, eds. La theorie du rayonnement et les
quanta (Solvay I). Paris, 1912.
Larmor, J. "On the theory of the magnetic influence on spectra; and on
the radiation from moving ions." PM, 44 (1897), 503-12.
- Aether and matter. Cambridge, 1900.
- "On the constitution of natural radiation." PM, 10 (1905), 574-84.
Laub, J. J. "Uber sekundare Kathodenstrahlen." AP, 23 (1907), 285-
300.
- "Zur Optik der bewegten Korper." AP, 23 (1907), 738-44.
- "Zur Optik der bewegten Korper." AP, 25 (1908), 175-84.
- "Uber die durch Rontgenstrahlen erzeugten sekundaren Kathoden-
strahlen." AP, 26 (1908), 712-26.
Laue, M. "Eine quantitative Prufung der Theorie fur die Interferenz-
Erscheinungen bei Rontgenstrahlen." Miinchen Sb, 42 (1912),
363-73.
- "Kritische Bemerkungen zu den Deutungen der Photogramme von
Friedrich und Knipping." PZ, 14 (1913), 421-3.
- "Rontgenstrahlinterferenzen," PZ, 7^(1913), 1075-9.
- "Zusatze (Marz 1913)." AP, 41 (1913), 989-1002.
- "Ein Versagen der klassischen Optik." VDpG, 19 (1917), 19-21.
- "Antrittsrede." Berlin Sb (1921), 479-81.
Ledoux-Lebard, R., and A. Dauvillier. La physique des rayons x. Paris,
1921.
Lehmann, O. Die elektrischen Lichterscheinungen oder Entladungen.
Halle, 1898.
328 Bibliography
Lenard, P. "Uber Kathodenstrahlen in Gasen von atmospharischem
Druck und in aussersten Vakuum." AP, 51 (1894), 225-68.
- "Uber die elektrostatischen Eigenschaften der Kathodenstrahlen." AP,
64(1898X279-89.
- "Erzeugung von Kathodenstrahlen durch ultraviolettes Licht." AP, 2
(1900), 359-75-
- "Uber die lichtelektrische Wirkung." AP, 5(1902), 149-98.
- "Uber Ather und Materie." Heidelburg Sb, 1 (1910), Abh. 16, sepa-
rately paginated.
- "Uber Elektrizitatsleitung durch freie Elektronen und Trager, II." AP,
41 (1913), 53-98.
Lenard, P., and C. Ramsauer. "Uber die Wirkungen sehr kurzwelligen
ultravioletten Lichtes auf Gase und uber eine sehr reiche Quelle
dieses Lichtes." With various sub-titles in the Heidelberg Sb, 1
(1910), Abh. 28, 31, 32; 2 (1911), Abh. 16, 24, all separately pagi-
nated.
- Ibid., part 2 (Abh. 31), subtitled "Wenig absorbierbares und doch auf
Luft wirkendes Ultraviolett."
- Ibid., part 5 (Abh. 24), subtitled "Wirkung des stark absorbierbaren
Ultraviolett und Zusammenfassung."
Lodge, O. "On the rays of Lenard and Rontgen." Electrician, 36 (1896),
438-40.
- "On the present hypotheses concerning the nature of Rontgen's rays."
Electrician, 36 (1896), 471-3.
- "The surviving hypothesis concerning the x-rays." Electrician, 37
(1896), 370-3.
Lorentz, H. A. "Vereenvoudigde theorie der electrische en optische
verschijnselen in lichamen die zich bewegen." Amsterdam Verslag,
7 (1899), 507-22. Reprinted in Collected papers, P. Zeeman and A.
Fokker, eds., vol. 5, pp. 139-55, as "Theorie simplifee des pheno-
menes electriques et optiques dans des corps en mouvement." The
Hague, 1937.
- Untitled Nobel lecture in Lesprix Nobel en 1902, separately paginated.
Stockholm, 1905.
- Lehrbuch der Physik zum Gebrauche bei akademischen Vorlesungen,
2 vols. G. Siebert, tr. Leipzig, 1907.
- "[Het licht en de bouw der materie]." NNGC Handelingen, 11 (1907),
6-21. Collected papers, P. Zeeman and A. Fokker, eds., vol. 9, pp.
167-81. The Hague, 1939. German translation in PZ, 8 (1907),
542-9.
- "De hypothese der lichtquanta." NNGC Handelingen, 12 (1909),
129-39; Collected papers, P. Zeeman and A. Fokker, eds., vol. 7, pp.
Bibliography 329
374-84. The Hague, 1934. German translation in PZ, 11 (1910),
349-54-
- "AlteundNeueFragenderPhysik."PZ, 77(1910), 1234-57; Collected
papers, P. Zeeman and A. Fokker, eds., vol. 7, pp. 205 -57., The
Hague, 1934.
- "Nature of light." In Encyclopaedia Britannica, nth ed., vol. 16, pp.
617-23. London, 1911.
- "Over den aard der Rontgenstralen." Amsterdam Verslag, 27 (1913),
911 - 23; Collected papers, P. Zeeman and A. Fokker, eds., vol. 3, pp.
281-94. The Hague, 1936.
- "The radiation of light." Nature, 114 (1924), 608-11.
- Problems in modern physics. New York, 1927.
McClelland, J. A. "The penetrating radium rays." RDS Sci trans, 8
(1904), 99-108; PM, 8 (1904), 67-77.
- "On secondary radiation." RDS Sci trans, 8 (1904), 169-82; PM, 9
(1905), 230-43.
- "On secondary radiation (Part II); and atomic structure." RDS Sci
trans, 9 (1905), 1 - 8; and three succeeding works, ibid., 9 - 26,27 - 36,
and 37-50.
McClung, R. K. "A preliminary account of an investigation of the effect
of temperature on the ionization produced in gases by the action of
Rontgen rays." PCPS, 12 (1903), 191-98.
McCormmach, R. "J. J. Thomson and the structure of light." BJHS, 3
(1967), 362-87.
- "H. A. Lorentz and the electromagnetic view of nature." Isis, 61
(1970), 459-97.
- "Einstein, Lorentz, and the electron theory." HSPS, 2 (1970), 41 -87;
and "Editor's foreword," ibid., ix-xx.
- "Hertz, Heinrich Rudolf." DSB, 6 (1972), 340-50.
- Night thoughts of a classical physicist. Cambridge, Mass., 1981.
Mach, E. "Durchsicht-Stereoskopbilder mit X-Strahlen." Zeitschrift fur
Elektrotechnik, 7^(1896), 259-61.
MacKenzie, A. S. "Secondary radiations from a plate exposed to rays
from radium." PM, 14 (1907), 176-87.
MacKinnon, E. "De Broglie's thesis: A critical retrospective." AJP, 44
(1976), 1047-55.
Madsen, J. P. V. "Secondary y radiation." TPRRSSA, ^2(1908), 163-92;
PM, 77(1909X423-48.
Malley, M. From hyperphosphorescence to nuclear decay: A history ofthe
early years of radioactivity 1896-1914. Unpublished Ph.D. disser-
tation, University of California, Berkeley, 1976.
Marx, E. "Die Geschwindigkeit der Rontgenstrahlen." VDpG, 7(1905),
330 Bibliography
302-21; PZ, 6 (1905), 768-77. An expanded report is in AP, 20
(1906X677-722.
- Grenzen in der Natur und in der Wahrnehmung vom Standpunkte der
Elektronentheorie und des elektromagnetischen Weltbildes. Leipzig,
1908.
- "Theorie der Vorgange im Nullapparat zur Geschwindigkeitsmessung
der Rontgenstrahlen." VDpG, 70(1908), 137-56.
- "Zur Frage der Geschwindigkeit der Rontgenstrahlen, Erwiderung auf
die gleichlautende Arbeit der Herren Franck und Pohl." VDpG, 10
(1908), 157-201.
- "Experimentelles Verhalten und Theorie des Apparates zur Ge-
schwindigkeitsmessung der Rontgenstrahlen." AP, 28 (1909),
37-56.
- "Uber den Einfluss der Rontgenstrahlen auf das EinsetzenderGlim-
mentladung." AP, 28(1909), 153-74.
- "Zweite Durchfiihrung der Geschwindigkeitsmessung der Rontgen-
strahlen." AP, 33(1910), 1305-91.
- "Sind meine Versuche uber die Geschwindigkeit der Rontgenstrahlen
durch Interferenz elektrischer Luftwellen erklarbar?" AP, 35 (1911),
397-400.
- "Die Theorie der Akkumulation der Energie bei intermittierender
Belichtung und die Grundlage des Gesetzes der schwarzen Strah-
lung " AP, 41 (1913), 161-90.
Marx, E., and K. Lichtenecker. "Experimentelle Untersuchung des
Einflusses der Unterteilung der Belichtungszeit auf die Elektronen-
abgabe in Elster und Geitelschen Kaliumhydriirzellen bei sehr
schwacher Lichtenergie." AP, 41 (1913), 124-60.
Mauguin, C. "La these de doctorat de Louis de Broglie." In Louis de
Broglie: Physicien et penseur, A. Georges, ed., pp. 430-6. Paris,
1953.
Maxwell, J. C. "On physical lines offeree." PM, 21 (1861), 161-75,
338-48; 23 (1862), 12-24, 85-95. Reprinted in Maxwell, Scientific
papers, W. D. Niven, ed. vol. 1, pp. 451-513. Cambridge, 1890.
Mayer, H. F. "Uber das Verhalten von Molekiilen gegenuber freien
langsamen Elektronen." AP, 64 (1921), 451-80.
Meitner, L. "Uber die Entstehung der /?-Strahl-Spektren radioaktiver
Substanzen." ZP, 9(1922), 131-44.
- "Uber den Zusammenhang zwischen /?- und y-Strahlen." ZP, 9 (1922),
M5-52.
Mendelssohn, K. The world of Walt her Nernst. London, 1973.
Meyer, E. "Bericht iiber die Untersuchungen der zeitlichen Schwankun-
gen der radioaktiven Strahlung." JRE, 5 (1908), 423-50; 6 (1909),
242-5.
Bibliography 3 31
- "Uber Stromschwankungen bei Stossionisation." VDpG, 12 (1910),
253-74.
- "Uberdie Struktur der y Strahlen." Berlin Sb, 32 (1910), 624-62; JRE,
7(1910), 279-95.
- "Uber Schweidlersche Schwankungen." PZ, 13 (1912), 73-81.
- "Uber die Struktur der y Strahlen. II." AP, 37 (1912), 700-20.
- "Zur Diskussion uber die Struktur der y Strahlen." PZ, 13 (1912),
253-4.
Meyer, E., and W. Gerlach, "Sur remission photo-electrique d'electrons
par des particules metalliques ultra-microscopiques," ASPN Gen-
eve, 35(1913), 398-400.
- "Sur l'effet photoelectrique des particules ultra-microscopiques aux
basses pressions." ASPN Geneve, 37(1914), 253-45.
Meyer, E., and E. Regener. "Uber Schwankungen der radioactiven
Strahlung und eine Methode zur Bestimmung des elektrischen Ele-
mentarquantums." VDpG, 70(1908), i-iy,AP, 25 (1908), 757-74.
Meyer, S., and E. von Schweidler. "Ober das Verhalten von Radium und
Polonium im magnetischen Felde." PZ, 1 (1899), 9 0 - 1 , 113-14.
Michelson, A. A. "A theory of the 'x-rays'." AJS, 1 (1896), 312-14;
Nature, 5^(1896X66-7.
Miller, A. I. Albert Einstein's special theory of relativity: Emergence
(1905) and early interpretations {1905-1911). Reading, Mass., 1981.
Millikan, R. A. "Some new values of the positive potentials assumed by
metals in a high vacuum under the influence of ultra-violet light."
PR, 30(1910), 287-8.
- "The effect of the character of the source upon the velocities of
emission of electrons liberated by ultra-violet light." Pi?, 35 (1912),
74-6.
- "Uber hohe Anfangsgeschwindigkeiten durch ultraviolettes Licht
ausgeloster Elektronen." VDpG, 70(1912), 712-26.
- "On the cause of the apparent differences between spark and arc
sources in the imparting of initial speeds to photo-electrons." Pi?, 1
(I9I3X73-5.
- "A direct determination of 'A V Pi?, 4 (1914), 73-5.
- "New tests of Einstein's photo-electric equation." PR, 6 (1915), 55.
- "A direct photoelectric determination of Planck's 7z'." PR, 7 (1916),
355-88.
- The electron. Chicago, 1917.
- "The electron and the light-quant from the experimental point of
view." Separately paginated in Lesprix Nobel en 1923. Stockholm,
1924.
- The autobiography of Robert A. Millikan. New York, 1950.
Millikan, R. A., and G. Winchester. "Upon the discharge of electrons
332 Bibliography
from ordinary metals under the influence of ultra-violet light." PR,
24 (1907), 116-18.
- "The absence of photoelectric fatigue in a very high vacuum." PR, 29
(1909), 85.
Millikan, R. A., and J. Wright. "The effect of prolonged illumination on
photo-electric discharge in a high vacuum." PR, 54 (1912), 68-70.
Moseley, H. "The number of /? particles emitted in the transformation of
radium." PRS, 87A (1912), 230-55.
Moseley, H., and C. G. Darwin. "The reflection of the x rays." Nature, 90
(1913X594.
- "The reflection of the x rays." PM, 26 (1913), 210-32.
Moseley, H., and W. Makower. "y radiation from radium B." PM, 23
(1912), 203-10.
Mott, N. F. "The quantum theory of electronic scattering by helium."
PCPS, 25 (1929), 304-9.
Nernst, W. Theoretische Chemie vom Standpunkte der Avogadroschen
Regel undder Thermodynamik, 5th ed. Stuttgart, 1907.
Nisio, S. "Sommerfeld's theory of the photoelectric effect." XIV congres
internationale d'histoire des sciences, Tokyo, Proceedings, 2 (1975),
302-4.
Les prix Nobel en 1902. Stockholm, 1905.
Les prix Nobel en 1921-1922. Stockholm, 1923.
Nye, M. "N-rays: An episode in the history and psychology of science."
HSPS, 7/(1980), 125-56.
Oseen, C. W. "Die Einsteinsche Nadelstichstrahlung und die Max-
wellsche Gleichungen." AP, 69 (1922), 202-4.
Owen, E. A. "The passage of homogeneous Rontgen rays through gases."
PRS, 86A (1912), 426-39.
Pais, A. "How Einstein got the Nobel Prize." American scientist, 70
(1982), 358-65.
Parkinson, E. M. "Stokes, George Gabriel." DSB, 13 (1976), 74-9.
Paschen, L. C. H. F. "Uber die durchdringenden Strahlen des Radiums."
AP, 7^(1904), 164-71.
- "Uber die Kathodenstrahlen des Radiums." AP, 14 (1904), 389-405.
- "Uber eine von den Kathodenstrahlen des Radiums in Metallen er-
zeugte Sekundarstrahlung." PZ, 5 (1904), 502-4.
- "Uber die y Strahlen des Radiums." PZ, 5 (1904), 563-8.
- "Uber die Warmeentwickelung des Radiums in einer Bleihulle." PZ, 6
(1905), 97-
Phillips, C. E. S., ed. Bibliography of x-ray literature and research.
London, 1897.
Bibliography 333
Planck, M. "Uber die Natur des weissen Lichtes." AP, 7 (1902),
390-400.
- Remarks following Einstein's talk to the GDNA. PZ, 10 (1909), 826.
- "Zur Theorie der Warmestrahlung." AP, 31 (1910), 758-68.
- "Eine neue Strahlungshypothese." VDpG, 13 (1911), 138-48.
- "La loi du rayonnement noir et l'hypothese des quantites elementaires
d'action." In Solvay I(1912), 93-114.
- "Uber die Begriindung des Gesetzes der schwarzen Strahlung." AP, 37
(1912), 642-56.
- "Uber das Gleichgewicht zwischen Oszillatoren,freien Elektronen und
strahlender Warme." Berlin Sb (1913), 350-63.
- Das Wesen des Lichts. Berlin, 1920. Originally in Die Naturwissen-
schaften, 7(1919), 903-9.
- "The physical reality of light quanta." JFI, 204 (1927), 13-18.
Pliicker, J. "Uber die Einwirkung des Magneten auf die elektrischen
Entladungen in verdunnten Gasen." AP, 103 (1858), 88-106.
J. C. Poggendorff's biographisch-literarisches Handworterbuch zur Ge-
schichte der exacten Naturwissenschaften. Vol. 4, Leipzig 1904. Vol.
5, Leipzig, 1926. Vol. 6, Leipzig, 1936-40.
Pohl, R. Die Physik der Rontgenstrahlen. Braunschweig, 1912.
Pohl, R., and P. Pringsheim. "Die lichtelektrische Empfindlichkeit der
Alkalimetalle als Funktion der Wellenlange." VDpG, 12 (1910),
215-28, 249-60.
- "The normal and the selective photoelectric effect." PM, 21 (1910),
155-61.
- "Weitere Versuche uber den selektiven lichtelektrischen Effekt."
VDpG, 72(1910), 682-96.
- "Der selektive lichtelektrische Effekt an K-Hg-Legierungen." VDpG,
72(1910), 697-710.
- "Zur Frage hoher Geschwindigkeiten lichtelektrischer Elektronen."
VDpG, 74(1912), 974-82.
- "Uber die langwellige Grenze des normalen Photoeffektes." VDpG, 15
(1913X637-44.
Poincare, H. Theorie mathematique de la lumiere, vol. 1,. Lemons sur la
theorie mathematique de la lumiere. Paris, 1889.
- "Les rayons cathodiques et les rayons Rontgen." RGS, 7 (1896),
52-9.
Poynting, J. H. "On the transfer of energy in the electromagnetic field."
PTRS, 175 (1884), 343-6i.
Precht, J. "Untersuchungen uber Kathoden- und Rontgenstrahlen." AP,
61 (1897), 350-62.
- "Die Warmeabgabe des Radiums." VDpG, 6 (1904), 101-3.
Przibram, K., ed. Briefe zur Wellenmechanik. Vienna, 1963.
334 Bibliography
Pyenson, L. "Einstein's early scientific collaborators." HSPS, 7 (1976),
83-123.
Quincke, G. Response to Hertz. GDNA Tageblatt, p. 149. Heidelberg,
1890.
Raman, V. V., and P. Forman. "Why was it Schrodinger who developed
de Broglie's ideas?" HSPS, 1 (1969), 291-314.
Ramsauer, C. "Uber die Wirkung sehr kurzwelligen ultravioletten
Lichtes auf Gase." PZ, 12 (1911), 997-8.
- "Uber eine direkte magnetische Methode zur Bestimmung der licht-
elektrischen Geschwindigkeitsverteilung." Heidelberg Sb, 5A
(1914), Abh. 19, separately paginated; AP, 45 (1914), 961 -1002.
- "Uber die lichtelektrische Geschwindigkeitsverteilung und ihre Ab-
hangigkeit von der Wellenlange." Heidelberg Sb, 5A (1914), Abh. 20,
separately paginated; AP, 45(1914), 1121-59.
- "Uber den Wirkungsquerschnitt der Gasmolekule gegenuber langsa-
men Elektronen." AP, 64 (1921), 513-40.
- "Uber den Wirkungsquerschnitt der Gasmolekule gegenuber langsa-
men Elektronen." PZ, 21 (1920), 576-8.
- "Uber den Wirkungsquerschnitt der Edelgase gegenuber langsamen
Elektronen." PZ, 22 (1921), 613-15.
Rayleigh. See Strutt.
Richardson, H. "Analysis of the y rays from uranium products." PM, 27
(1914), 252-6.
Richardson, O. W. "Some applications of the electron theory of matter."
PM, 23 (1912), 594-627.
- "The theory of photoelectric action." PM, 24 (1912), 570-4.
- "The laws of photoelectric action and the unitary theory of light
(Lichtquanten Theorie)." Science, 36 (1912), 57-8.
- "The application of statistical principles to photoelectric effects and
some allied phenomena." PR, 34 (1912), 146-9.
- "The asymmetric emission of secondary rays." PM, 25 (1913), 144-50.
- "The theory of photoelectric and photochemical action." PM, 27
(1914), 476-88.
- "The complete photoelectric emission." PM, 31 (1916), 149-55.
- "The photoelectric action of x rays." PRS, 94A (1918), 269-80.
Richardson, O. W., and K. T. Compton. "The photoelectric effect." PM,
24(1912X575-94.
Richardson, O. W., and F. J. Rogers, "The photoelectric effect III." PM,
29(1915X618-23.
Righi, A. "Sur les changements de longeur d'onde obtenus par la rotation
d'un polariseur, et sur le phenomene des battements produits avec
les vibrations lumineuses." JP, 2 (1883), 437-46.
- "Sulla propagazione dell' elettricita nei gas attraversati dai raggi de
Rontgen." Bologna Memorie, 6 (1896), 231-301.
Bibliography 335
- "SulF elettrizzazione prodotta dai raggi del radio." Lincei Atti, 14
(1905), 556-9.
Robinson, H., and W. F. Rawlinson. "The magnetic spectrum of the /?
rays excited in metals by soft x rays." PM, 28 (1914), 277-81.
Rontgen, W. C. "Uber eine neue Art von Strahlen (vorlaufige Mitthei-
lung)."Wurzburg5^(1895), iî-ÂP, 6^(1898), 1-11. (Rontgen
- "Uber eine neue Art von Strahlen II. Mittheilung." Wiirzburg Sb

(1896), 11-16; AP, 6^(1898), 12-17. (Rontgen II).
- "Weitere Beobachtungen uber die Eigenschaften der X-Strahlen."
Berlin Sb (1897), 576-92; AP, 64 (1898), 18-37. (Rontgen III).
Rubens, H., and E. Ladenburg. "Uber die lichtelektrische Erscheinung
an dlinnen Goldblattchen." VDpG, 9 (1907), 749-52.
Runge, C, and J. Precht. "Die Warmeabgabe des Radiums." Berlin Sb
(1903), 783-6.
Russell, A. S. "The effect of temperature upon radioactive disintegra-
tion." PRS, 86A (1911), 240-53.
Russell, A. S., and F. Soddy. "The y rays of thorium and actinium." PM,
2/(1911), 130-54.
Russo, A. "Fundamental research at Bell Laboratories: The discovery of
electron diffraction." HSPS, 72(1981), 117-60.
Rutherford, E. "Uranium radiation and the electrical conduction pro-
duced by it." PM, 47 (1899), 109-63.
- "Penetrating rays from radio-active substances." Nature, 66 (1902),
318-19. Translated, with omissions, in PZ, 3 (1902), 517-20.
- "The magnetic and electric deviation of the easily absorbed rays from
radium." PM, 5(1903), 177-87. Translated in PZ, 4 (1903), 235-40.
- "Radio-active processes." Nature, 68 (1903), 163.
- "Does the radio-activity of radium depend on its concentration?"
Nature, 69 (1904), 222.
- "The heating effect of the radium emanation." AusAAS Report (1904),
86-91.
- "Nature of the y rays from radium." Nature, 69 (1904), 436-7.
- Radioactivity. Cambridge, 1904. (2nd ed., Cambridge, 1905.)
- "Some properties of the a rays from radium," PM, 70(1905), 163-76.
- "The origin of f$ and y rays from radioactive substances." PM, 24
(1912), 453-62.
- Radioactive substances and their radiations. Cambridge, 1913.
- "The connexion between the /? and y ray spectra." PM, 28 (1914),
305-19.
- "The structure of the atom." PM, 27 (1914), 488-98.
- "Penetrating power of the x radiation from a Coolidge tube." PM, 34
(1917), 153-62.
- Comment on M. de Broglie's report. In Solvay 777 (1923), 107-9.
336 Bibliography
Rutherford, E., and E. N. Andrade. "The wavelength of the soft y rays
from radium B." PM, 27(1914), 854-68. A brief first announcement
appeared in Nature, 92 (1913), 267.
- "The spectrum of the penetrating y rays from radium B and radium
C"PM,28 (1914X263-73.
Rutherford, E., and H. T. Barnes. "Heating effect of the y rays from
radium." PM, 9 (1905), 621-28.
- "The charge and nature of the a particle." PRS, 81A (1908), 162-73.
Rutherford, E., and H. Geiger. "An electrical method of counting the
number of a particles from radio-active substances." PRS, 81A
(1908), 141-61.
Rutherford, E., and R. K. McClung. "Energy of Rontgen and Becquerel
rays, and the energy required to produce an ion in gases," PTRS,
196A (1901), 25-59; German translation in PZ, 2 (1900), 53-5.
Rutherford, E., and H. Richardson. "The analysis of the gamma rays
from radium B and radium C." PM, 25 (1913), 722-34.
- "Analysis of the y rays from radium D and radium E." PM, 26 (1913),
324-32.
- "Analysis of the y rays of the thorium and actinium products." PM, 26
(1913X 937-44-
Rutherford, E., and H. Robinson. "The analysis of the /? rays from
radium B and radium C." PM, 26 (1913), 717-29.
Rutherford, E., and W. A. Wooster. "The natural x ray spectrum of
radium B. PCPS, 22 (1925), 834-837.
Rutherford, E., J. Barnes, and H. Robinson. "Maximum frequency of the
x rays from a Coolidge tube for different voltages." PM, JO (1915),
339-60.
Rutherford, E., J. Chadwick, and C. D. Ellis. Radiations from radioactive
substances. Cambridge, 1930.
Rutherford, E., H. Robinson, and W. F. Rawlinson. "Spectrum of the ft
rays excited by y rays." PM, 28 (1914), 281-86.
Sadler, C. A. "Transformations of x rays." PM, 18 (1909), 106-32.
- "Homogeneous corpuscular radiation." PM, 19 (1910), 337-56.
Sagnac, G. "Sur la transformation des rayons x par les metaux." CR, 125
(1897X942-4.
- "Transformation des rayons x par transmission." CR, 126 (1898),
467-70.
- "Emission de rayons secondaires par Fair sous Finfluence des rayons
x." CR, 726(1898), 521-3.
- Theses. De Voptique des rayons de Rontgen et des rayons secondaires
qui en derivent. Paris, 1900.
Schlegel, R. "Louis de Broglie's thesis." AJP, 45 (1977), 871-2.
Schmidt, H. W. "Uber Reflexion und Absorption von p Strahlen." AP,
23 (1907), 671-97.
Bibliography 337
- "Einige Versuche mit p Strahlen von Radium E." PZ, 8 (1907),
361-73.
- "Uber die Strahlung des Uranium x." PZ, 10 (1909), 6-16.
Schrodinger, E. "Dopplerprinzip und Bohrsche Frequenzbedingung."
PZ, 25(1922), 301-3.
- "Uber eine bemerkenswerte Eigenschaft der Quantenbahnen eines
einzelnen Elektrons." ZP, 12 (1922), 13-23.
- "Zur Einsteinschen Gastheorie." PZ, 27 (1926), 95-101.
- "Quantisierung als Eigenwertproblem (erste Mitteilung)." AP, 79
(1926), 361-76.
- "Uber das Verhaltnis der Heisenberg-Born-Jordanschen Quanten-
mechanik zu der meinen." AP, 79 (1926), 734-56.
Schuster, A. "The discharge of electricity through gases." PRS, 5 7 (1890),
526-59.
- "On interference phenomena." PM, 57(1894), 509-45.
- "On the new kind of radiation." British Medical Journal (18 Jan.
1896), 172-3.
- "On Rontgen's rays." Nature, 53 (1896), 268.
Schweidler, E. von "Uber Schwankungen der radioaktiven Umwand-
lung." Premier congres international pour Vetude de la radiologie et
de I' ionization, Liege, paginated section no. 3. Brussels, 1906.
- "Zur experimentellen Entscheidung der Frage nach der Natur der y
Strahlen." PZ, 11 (1910), 225-7, 614-19.
Seelig, C. Albert Einstein, eine dokumentarische Biographie. Zurich,
1954.
- Albert Einstein, Leben und Werk eines Genies unserer Zeit. Zurich,
i960.
Seitz, W. "Bemerkung zu der B. Walterschen Mitteilung uber das Ront-
gensche Absorptionsgesetz." AP, 26 (1909), 448.
- "Geschwindigkeit von Elektronen, welche durch weiche Rontgen-
strahlen erzeugt werden." PZ, 11 (1910), 705-8.
Serwer, D. The rise of radiation protection: Science, medicine, and
technology in society, 1896-1935. Unpublished Ph.D. dissertation,
Princeton University, 1976.
- tfUnmechanischer Zwang: Pauli, Heisenberg, and the rejection of the
mechanical atom." HSPS, 8 (1977), 189-256.
Siegbahn, M. "Uber den Zusammenhang zwischen Absorption und
Wellenlange bei Rontgenstrahlen." PZ, 15 (1914), 753-6.
Simons, L. "The beta ray emission from thin films of the elements
exposed to Rontgen rays." PM, 41 (1921), 120-40.
Slater, J. C. "The development of quantum mechanics in the period
1924-1926." In Wave mechanics: Thefirst fifty years, W. C. Price,
S. S. Chissick, and T. Ravensdale, eds., pp. 19-25. New York,
1973.
338 Bibliography
Soddy, F., W. M. Soddy, and A. S. Russell. "The question of the
homogeneity of y rays." PM, 19 (1910), 725-57.
Solvay I. La theorie du rayonnement et les quanta, P. Langevin and
M. de Broglie, eds. Paris, 1912.
Solvay III. Atomes et electrons. Paris, 1923.
Sommerfeld, A. "Zur mathematischen Theorie der Beugungserschein-
ungen." Gottingen Nachrichten (1894), 338-42.
- "Diffractionsprobleme in exacter Behandlung." Deutsche Mathema-
tiker-Vereinigung Jahresbericht, 4 (1895), r 7 2 "4-
- "Mathematische Theorie der Diffraction." Mathematische Annalen,
47 (1896), 317-74.
- "Theoretisches uber die Beugung der Rontgenstrahlen." PZ, 1 (1899),
105-11; PZ, 2 (1900), 55-60
- "Die Beugung der Rontgenstrahlen unter der Annahme von Aether-
stossen." VDNA (1900), part 2, first half, separately paginated.
- "Theoretisches uber die Beugung der Rontgenstrahlen." Zeitschriftfur
Mathematik undPhysik, 46 (1901), 11-97.
- "Uber die Verteilung der Intensitat bei der Emission von Rontgen-
strahlen." PZ, 10 (1909), 969-76.
- "Uber die Verteilung der Intensitat bei der Emission von Rontgen-
strahlen." PZ, 11 (1910), 99-101.
- "Das Plancksche Wirkungsquantum und seine allgemeine Bedeutung
fur die Molekularphysik." PZ, 12 (1911), 1057-68.
- "Uber die Struktur der y Strahlen." Munchen Sb, 41 (1911), 1 -60.
- "Application de la theorie de l'element d'action aux phenomenes
moleculaires non periodiques." In Solvay I (1912), 313-72.
- "Uber die Beugung der Rontgenstrahlen." AP, 38 (1912), 473-506.
- "Unsere gegenwartige Anschauungen uber Rontgenstrahlung." Die
Naturwissenschaften, 1 (1913), 705-13.
- "Uber das Spektrum der Rontgenstrahlen." AP, 46 (1915), 721-48.
- "Die neueren Fortschritte in der Physik der Rontgenstrahlen." Mun-
chener medizinische Wochenschrift, 62(1915), 1424-30.
- "Die neueren Fortschritte in der Physik der Rontgenstrahlen." Die
Naturwissenschaften, ^(1916), 1-8, 13-18.
- "Die Wellenlange der Rontgenstrahlen als Hartemass." Munchener
medizinische Wochenschrift, 63 (1916), 458-60.
- Atombau und Spektrallinien. Braunschweig, 1919.
- "Das Spektrum der Rontgenstrahlung als Beispiel fur die Methodik
der alten und neuen Mechanik." Scientia, 51 (1932), 41-50.
- "Autobiographische Skizze." In Geist und Gestalt: Biographische Bei-
trage zur Geschichte der Bayerischen Akademie der Wissenschaften,
vol. 2, pp. 100-109. Munich, 1959.
Sommerfeld, A., and G. Wentzel. "Uber regulare und irregulare Dub-
letts."ZP, 7(1922X86-92.
Bibliography 339
Sommerfeld, A., and E. Wiechert. "Harmonischer Analysator."
Deutsche Mathematiker-Vereinigung Katalog mathematischer
Modelle, Apparate und Instrumente, pp. 214-221. Munich,
1892.
Stark, J. "Elektricitat in Gasen." In Handbuch der Physik, 2nd ed.,
A. Winkelmann, ed., vol. 4, pp. 454-653. Leipzig, 1905.
- "Uber Absorption und Fluoreszenz im Bandenspektrum und iiber
ultraviolette Fluoreszenz des Benzols." PZ, 8 (1907), 81-5.
- "Beziehung des Doppler-Effektes bei Kanalstrahlen zur Planckschen
Strahlungstheorie." PZ, 8 (1907), 913-19.
- "Elementarquantum der Energie, Modell der negativen und der posi-
tiven Elektrizitat." PZ, 8 (1907), 881-4.
- "ZurEnergetikundChemiederBandenspektra."PZ, 9(1908)585-94,
356-8.
- "Neue Beobachtungen an Kanalstrahlen in Beziehung zur Lichtquan-
tenhypothese." PZ, 9 (1908), 767-73.
- "Die Valenzlehre auf atomistisch elektrischer Basis." JRE, 5 (1908),
124-53.
- "The wave-length of Rontgen rays." Nature, 77 (1908), 320.
- "Uber die Ionisierung von Gasen durch Licht." PZ, 10 (1909),
614-23.
- "Uber Rontgenstrahlen und die atomistische Konstitution der Strah-
lung,"7>Z, 70 (1909), 579-86.
- "Zur experimentellen Entscheidung zwischen Atherwellen- und
Lichtquantenhypothese. I. Rontgenstrahlen." PZ, 10 (1909),
902-13.
- Die Prinzipien der Atomdynamik, vol. 1, Die elektrischen Quanten.
Leipzig, 1910.
- "Zur experimentelle Entscheidung zwischen der Lichtquantenhy-
pothese und der Atherimpulstheorie der Rontgenstrahlen." PZ, 11
(1910X24-31.
- "Folgerungen uber die Natur des Lichtes aus Beobachtungen iiber die
Interferenz." JRE, 7(1910), 386-404.
- "Weitere Beobachtungen iiber die dissymmetrische Emission von
Rontgenstrahlen." PZ, 11 (1910), 107-12.
- "Zur experimentellen Entscheidung zwischen Lichtquantenhypothese
und Atherwellentheorie. II. Sichtbares und ultraviolettes spektrum."
PZ, 77(i9io), 179-87.
- Comment following Sommerfeld's presentation to the GDNA. PZ, 12
(1911), 1068-9.
- Die Prinzipien der Atomdynamik, vol. 2, Die elementare Strahlung.
Leipzig, 1911.
- "Zur Diskussion uber die Struktur der y Strahlen." PZ, 13 (1912),
161-2.
340 Bibliography
- "Bemerkung uber Zerstreuung und Absorption von /? Strahlen und
Rontgenstrahlen." PZ, 13 (1912), 973-6.
- "Zur Diskussion des Verhaltens der Rontgenstrahlung in Kristallen."
PZ, 7^(1913), 319-21.
- "Schwierigkeiten fur die Lichtquantenhypothese im Falle der Emis-
sion von Serienlinien." VDpG, 16 (1914), 304-6.
Stark, J., and W. Steubing. "Fluoreszenz und lichtelektrische Empfind-
lichkeit organischer Substanzen." PZ, 9 (1908), 481-95.
- "Weitere Beobachtungen iiber die Fluoreszenz organischer Substan-
zen." PZ, 9 (1908), 661-9.
Stark, J., and G. Wendt. "Pflanzt sich der Stoss von Kanalstrahlen in
einem festen Korper fort?" AP, 38 (1912), 941-57.
Starke, H. "Die magnetische und elektrische Ablenkbarkeit reflektierter
und von diinnen Metallblattchen hindurchgelassener Kathoden-
strahlen." VDpG, 5 (1903), 14-22.
- "Recherches sur les rayons secondaires du radium." Le Radium, 5
(1908), 35-4L
Stephenson, R. J. "The scientific career of Charles Glover Barkla," AJP,
55(1967), 140-52.
Stewart, O. "Becquerel rays, a resume." PR, 11 (1900), 155-75.
Stokes, G. G. "On the dynamical theory of diffraction." TCPS, 9 (1849),
1-62. Reprinted in Mathematical and physical papers, J. Larmor,
ed., vol. 2, pp. 243-328. Cambridge, 1883.
- On the nature of light. London, 1884.
- On light. London, 1887.
- "On the nature of Rontgen rays." PCPS, 9 (1896), 215-16.
- "On the Rontgen rays." Nature, 54 (1896), 427-30.
- "An annual address (chiefly on the subject of the Rontgen rays)." VI
Journal of Transactions, 30(1898), 13-25.
- "On the nature of the Rontgen rays." Manchester MP, 41 (1897),
no. 15, 29 pp. Reprinted in Mathematical and physical papers, J.
Larmor, ed., vol. 5, pp. 256-77. Cambridge, 1905.
- Memoir and scientific correspondence of the late Sir George Gabriel
Stokes, J. Larmor, ed., vol. 2. Cambridge, 1907.
Stoletov, A. G. "Aktinoelektrische Untersuchungen." Physikalische
Revue, 1 (1892), 723-80.
Stoney, G. J. "Evidence that Rontgen rays are ordinary light." PM, 45
(1898), 532-6; PM, 46 (1898), 253-4.
Strutt, J. W. [third Lord Rayleigh]. "Wave theory." Encyclopaedia Bri~
tannica, 9th ed., vol. 24, pp. 421-59. London, 1888.
- "On the character of the complete radiation at a given temperature."
PM, 27(1889), 460-9.
- "Rontgen rays and ordinary light." Nature, 57 (1898), 607.
Bibliography 341
Strutt, R. J. [fourth Lord Rayleigh]. "On the behavior of the Becquerel
and Rontgen rays in a magnetic field." PRS, 66 (1900), 75-9.
- "On the conductivity of gases under the Becquerel rays." PTRS, 196A
(1900,507-27.
- "On the intensely penetrating rays of radium." PRS, 72 (1903),
208-10.
- The Becquerel rays and the properties of radium. London, 1904.
- The life of Sir J. J. Thomson. Cambridge, 1942.
Stuewer, R. "Non-Einsteinian interpretations of the photoelectric ef-
fect." Minnesota studies in philosophy of science, R. Stuewer, ed.,
vol. 5, pp. 246-63. Minneapolis, 1970.
- "William H. Bragg's corpuscular theory of x rays and y rays." BJHS, 5
(1970,258-81.
- The Compton effect: Turning point in physics. New York, 1975.
Stuhlmann, O., Jr. "A difference in the photoelectric effect caused by
incident and divergent light." Nature, 83 (1910), 311.
- "A difference in the photoelectric effect caused by incident and emer-
gent light." PM, 20 (1910), 331-9; 22 (1910, 854-64.
Sudre, R. "L'oeuvre scientifique de Maurice de Broglie." Revue des deux
mondes( July -August i960), 577-82.
Swenson, L. "Richardson, Owen Willans," DSB, 11 (1975), 419-23.
Taylor, G. I. "Interference fringes with feeble light." PCPS, 15 (1909),
114-15.
Thibaud, J. Theses. La spectroscopie des rayons y. SpectresP secondaires
et diffraction cristalline. Paris, 1925.
Thomson, G. P. "Experiments on the diffraction of cathode rays." PRS,
117A (192$), 600-9.
Thomson, J. J. Notes on recent researches in electricity and magnetism.
Oxford, 1893.
- "Longitudinal electric waves and Rontgen's x-rays." PCPS, 9 (1896),
49-61.
- "On the discharge of electricity produced by the Rontgen rays, and the
effects produced by these rays on dielectrics through which they
pass." PRS, 59 (1896), 274-6.
- "Presidential address to Section A," BAAS Report (1896), 699-706.
- "The Rontgen rays." Nature, 53 (1896), 391-2.
- "The Rontgen rays." Nature, 54 (1896), 302-6. The Rede Lecture
delivered June 10, 1896.
- "Cathode rays." PM, 44 (1897), 293-316.
- "On the charge of electricity carried by the ions produced by Rontgen
rays." PM, 46 (1898), 528-45.
- "On the diffuse reflection of Rontgen rays." PCPS, 9 (1898), 393-7.
- "On the connection between the chemical composition of a gas and the
342 Bibliography
ionization produced in it by Rontgen rays." PCPS, 70(1898), 10-14.
- The discharge of electricity through gases. New York, 1898.
- "A theory of the connection between cathode and Rontgen rays." PM,
45.(1898), 172-83.
- Conduction of electricity through gases. Cambridge, 1903.
- "The magnetic properties of systems of corpuscles describing circular
orbits." PM, 6 (1903), 673-93.
- "Radium." Nature, 67(1903), 601-2.
- Electricity and matter. New Haven, 1904.
- "On the positive electrification of a rays, and the emission of slowly
moving cathode rays by radio-active substances." PCPS, 13 (1904),
49-54-
- "On the emission of negative corpuscles by the alkali metals." PMy 10
(1905X584-90.
- "Do the y rays carry a charge of negative electricity?" PCPS, 13 (1905),
121-3.
- Conduction of electricity through gases, 2nd ed. Cambridge, 1906.
- "On the number of corpuscles in an atom." PM, 11 (1906), 769-81.
- "On secondary Rontgen radiation." PCPS, 13 (1906), 322-4; 14
(1906), 109-14.
- Corpuscular theory of matter. London, 1907.
- "On the ionization of gases by ultra-violet light and on the evidence as
to the structure of light afforded by its electrical effects." PCPS, 14
(1907), 417-24.
- "On the light thrown by recent investigations on electricity on the
relation between matter and ether." Victoria University of Man-
chester Lectures, 8 (1908), separately paginated.
- "The nature of the y rays." PCPS, 14 (1908), 540.
- "On the velocity of secondary cathode rays from gases." PCPS, 14
(1908), 541-5.
- "On the theory of radiation." PM, 20 (1910), 238-47.
- "On a theory of the structure of the electric field and its application to
Rontgen radiation and to light." PM, 19 (1910), 301-13. Hints of
Thomson's radical proposal were given in BAAS Report (1909),
3-29; and in Nature, 81 (1909), 248-57.
- "Rontgen rays," Encyclopaedia Britannica, 1 ith ed., vol. 23. London,
1911.
- "On the structure of the atom." PM, 26 (1913), 792-9.
- The structure of light. Cambridge, 1925.
- Recollections and reflections. London, 1936.
Thomson, J. J., and E. Rutherford. "On the passage of electricity through
gases exposed to Rontgen rays." PM, 42 (1896), 392-407.
Bibliography 343
Thomson, W. [Lord Kelvin]. Baltimore lectures on molecular dynamics
and the wave theory of light. London, 1904. Notes by Hathaway
distributed by Johns Hopkins University, 1884.
- "On the generation of longitudinal waves in ether." PRS, 59 (1896),
270-3.
- "Note on Lord Blythswood's paper." PRS, 59 (1896), 332-3.
- "Continuity in undulatory theory of condensational-rarefactional
waves in gases, liquids, and solids, of distortional waves in solids, of
electric waves in all substances capable of transmitting them, and of
radiant heat, visible light, and ultra-violet light." PM, 46 (1898),
494-500; BAAS Report (1898), 783-7.
- "On the reflection and refraction of solitary plane waves at a plane
interface between two isotropic elastic mediums - fluid, solid, or
ether." PM, 47 (1899), 179-91.
- "On the application of force within a limited space, required to
produce spherical solitary waves, or trains of periodic waves, of both
species, equivoluminal and irrotational in an elastic solid." PM, 47
(1899), 480-93; 48 (1899), 227-36, 388-93.
- "Contribution by Lord Kelvin to the discussion on the nature of the
emanations from radium . . ." BAAS Report (1903), 535-7;/W, 7
(1904), 220-2.
Topper, D. "Commitment to mechanism: J. J. Thomson, the early
years." AHES, 7 (1971) 393-410.
Trenn, T. "Rutherford and the alpha-beta-gamma classification of
radioactive rays." Isis, 67 (1976), 61-75.
Trillat, J. J. "Maurice de Broglie." Societe Frangaise de mineralogie et
de cristallographie Bulletin, 83 (i960), 239-41.
La Varende. Les Broglie. Paris, 1950.
Villard, P. Les rayons cathodiques. Paris, 1900. Originally presented in
(Rapports presentes au congres international de physique, C. Guil-
laume and L. Poincare, eds., vol. 3, pp. 115-37. Paris, 1900.
- "Sur le rayonnement du radium." CR, 130 (1900), 1178-9.
- "Sur la reflexion et la refraction des rayons cathodiques et des rayons
deviables du radium." CR, 130 (1900), 1010-12.
- "La formation des rayons cathodiques." In Ions, electrons, corpuscles,
H. Abraham and P. Langevin, eds., pp. 1013-28. Paris, 1905.
Volterra, V. LeQons sur les equations integrates et les equations integro-
dijferentielles. Paris, 1913.
- "Sur la theorie mathematique des phenomenes hereditaires." Journal
de mathematiques pures et appliquees, 7 (1928), 249-98.
van der Waals, J. D., Jr. "Zur Frage der Wellenlange der Rontgenstrah-
len." AP, 22 (1907), 603-5; 23 (1907), 395-6.
344 Bibliography
Wagner, E. "Spektraluntersuchungen an Rontgenstrahlen nach Versu-
chen gemeinschaft mit Joh. Brentano." Miinchen Sb, 44 (1914),
329-38.
- "Spectraluntersuchungen an Rontgenstrahlen I." AP, 46 (1915),
868-92
Walter, B. "Die Nature der Rontgenstrahlen." AP, 66 (1898), 74-81.
- "Uber die Haga und Wind'schen Beugungsversuche mit Rontgen-
strahlen." VDNA (1901), part 2, p. 43; PZ, 3 (1902), 137-40.
- "Uber die Geschwindigkeit der Rontgenstrahlen." Fortschritte auf
dem Gebiete der Rontgenstrahlen, 9 (1906), 223-5.
Walter, B., and R. Pohl. "Zur Frage der Beugung der Rontgenstrahlen."
AP, 25(1908), 715-24.
- "Weitere Versuche iiber die Beugung der Rontgenstrahlen." AP, 29
(1909X331-54.
Webster, D. L. "The x-ray spectrum of tungsten at a constant potential."
PR, 6(1915), 56.
- "Experiments on the emission quanta of characteristic x rays." PR, 7
(1916), 599-613.
Weidner, R., and R. Sells. Elementary modern physics. Boston, i960.
Weill-Brunschvicg, A. R., and J. L. Heilbron. "Broglie, Louis-Cesar-Vic-
tor-Maurice de." DSB, 2 (1973), 487-9.
Wentzel, G. "Die unperiodischen Vorgange in der Wellenmechanik."
PZ, 29(1928), 321-37.
Wessels, L. "Schrodinger's route to wave mechanics." SHPS, 10 (1979),
311-40.
Wheaton, B. R. The photoelectric effect and the origin of the quantum
theory of free radiation. Unpublished M.A. thesis, University of
California, Berkeley, 1971.
- On the nature ofx- and gamma-rays: Attitudes toward localization of
energy in the "new radiations," 1896-1922. Unpublished Ph.D.
dissertation, Princeton University, 1978.
- "Philipp Lenard and the photoelectric effect, 1889-1911." HSPS, 9
(1978), 299-323.
- "Impulse x-rays and radiant intensity: The double edge of analogy."
HSPS, 77(1981X367-90.
Wheaton, B. R., and J. L. Heilbron. An inventory ofpublished letters to
and from physicists, 1900-1950. Berkeley papers in history of
science, vol. 6. Berkeley, 1982.
Whiddington, R. "Preliminary note on the properties of easily absorbed
Rontgen radiation." PCPS, 15 (1910), 574-5.
- "The production and properties of soft Rontgen radiation." PRS, 85A
(1911), 99-118.
Bibliography 345
- "The production of characteristic Rontgen radiations." PCPS, 16
(1911), 150-4; PRS, 85A (1911), 323-32. A first announcement
appeared in Nature, 88 (1911), 143.
- "The transmission of cathode rays through matter." PRS, 86A (1912),
360-70.
- "The velocity of the secondary cathode particles ejected by the charac-
teristic Rontgen rays." PRS, 86A (1912), 370-8.
- "X-ray electrons." PM, 43 (1922), 1116-26.
Wiechert, E. "Bedeutung des Weltathers." Konigsberg Schriften, 35
(1894), 4-11.
- "Uber die Grundlagen der Electrodynamik." AP, 59 (1896), 283-323.
- "Die Theorie der Elektrodynamik und die Rontgen'sche Entdeck-
ung." Konigsberg Schriften, 37(1896), 1-48.
- "Experimentelle Untersuchungen uber die Geschwindigkeit und die
magnetische Ablenkbarkeit der Kathodenstrahlen." AP, 69 (1899),
739-66.
Wien, W. "Temperatur und Entropie der Strahlung." AP, 52 (1894),
.132-65.
- "Uber die Energievertheilung im Emissionsspektrum eines schwarzen
Korpers." AP, 58 (1896), 662-9.
- "Uber die Moglichkeit einer elektromagnetischen Begriindung der
Mechanik." Archives Neerlandaises (2), 5 (1900), 96-107.
- "Untersuchungen uber die elektrische Entladung in verdiinnten
Gasen." AP, 5 (1901), 421-35; AP, 8 (1902), 244-66.
- "Uber die Natur der positiven Elektronen." AP, 9 (1902), 660-4.
- "Uber die Selbstelektrisierung des Radiums und die Intensitat der von
ihm ausgesandten Strahlen." PZ, 4 (1903), 624-6.
- "Uber die Energie der Rontgenstrahlen." PZ, 5 (1904), 128-30.
- "Uber die Theorie der Rontgenstrahlen." JRE, 1 (1904), 215-20.
- "Uber die Energie der Kathodenstrahlen im Verhaltnis zur Energie der
Rontgen- und Sekundarstrahlen." Festschrift Adolph Wullner ge-
widmet zum siebzigsten Geburtstag," pp. 1 -14. Leipzig, 1905.
- "Uber die Energie der Kathodenstrahlen im Verhaltnis zur Energie der
Rontgen- und Sekundarstrahlen." AP, 18 (1905), 991 -1007.
- "Uber eine Berechnung der Wellenlange der Rontgenstrahlen aus dem
Planckschen Energie Element." Gottingen Nachrichten (1907),
598-601.
- "Uber die Berechnung der Impulsbreite der Rontgenstrahlen aus ihrer
Energie." AP, 22 (1907), 793-7-
- "Uber die absolute von positiven Ionen ausgestrahlte Energie und die
Entropie der Spektrallinien." AP, 23 (1907), 415-38.
- "Gesetze und Theorien der Strahlung." Wurzburg Sb (1907), 103-12.
346 Bibliography
- Vorlesungen uber neuere Probleme der theoretischen Physik. Leipzig,
1913; WAS Journal, 3 (1913), 273-84.
- "Zur Theorie der Strahlung." AP, 46 (1915), 749-52.
Wigger, O. "Zur Characteristik der a- und y-Strahlen." JRE, 2 (1905),
391-433.
Wilson, C. T. R. "On a method of making visible the paths of ionizing
particles through a gas." PRS, 85A (1911), 285-8.
Wilson, W. "On the absorption of homogeneous/? rays by matter, and on
the variation of the absorption of the rays with velocity." PRS, 82A
(1909), 612-28.
- "The decrease in velocity of the ^-particles on passing through matter."
PRS, 84A (1910), 141-50.
Wind, C. H. "Uber die Deutung der Beugungserscheinungen bei Ront-
genstrahlen." AP, 68 (1899), 896-901.
- "Over helderheidsmaxima en -minima als gevolg van een gezichtsbe-
drog." Amsterdam Verslag, 7(1899), 12-19.
- "Zur Anwendung der Fourierschen Reihenentwickelung in der
Optik." PZ, 2 (1900), 189-96.
- "Zur Beugung der Rontgenstrahlen." PZ, 2 (1900), 297-8.
Winkelmann, A., ed. Handbuch der Physik, 2nd. ed. Leipzig, 1905.
Wright, J. R. "The positive potential of aluminum as a function of the
wavelength of the incident light." PR, 33 (1911), 43-52.
Wiillner, A. Lehrbuch der Experimentalphysik, vol. 3, Die Lehre vom
Magnetismus und von der Elektricitdt, 5th ed. Leipzig, 1897.
Zehnder, L. W. C. Rontgen, Briefe an L. Zehnder. Zurich, 1935.
Zwicky, F. "Das Verhalten von langsamen Elektronen in Edelgasen."
PZ, 24(1923), 171-83.
INDEX
All names mentioned in the text and substantive footnotes, as well

as selected subjects, are noted here. Important contributions are
listed by topic; significant discussions are indicated by boldface
numbers. References to authors cited in the bibliography are in
italic type. Discussions that extend well beyond the page cited are
marked here by "ff", but no attempt has been made to index
subsequent page numbers for topics that extend simply to the next
page.
Abraham, Henri (1868-1943), 309 Barkla, Charles Glover (1877-1944),

Abraham, Max (1875-1922), 64, 135, 139, 190,203,205,212,
105,111, 153,509 213, 253-54, 257, 260, 265, 311;
absorption coefficient, 79, 253 and Bragg, 94ff; education, 44;
aether: Campbell's, 142; properties, 9; on homogeneous x-rays: charac-
Thomson's, 140 teristic, 100, discovered, 93, two
Akesson, Nils, 305, 309 components, 101; neutral-pair
Allen, Lini, 326 hypothesis, 95; and Stark, 119; on
Allen, Samuel James (1877-1966), x-rays: absorption, 80, 256ff,
309 characteristic, 100, elements' re-
alpha (a) rays, 16,52,55,83 sponse, 91, polarization, 45,
Andrade, Edward Nevelle da Costa rejects quantum theory, 256, sec-
(1887 -1971),224,231,309, 336 ondary, 45
Angstrom, Knut (1857-1910), 65, Barnes, Howard Turner (1873-1950),
309 336
anomalous dispersion, 30 Barnes, James (1878-1962), 336
Aschkinass, E., 325 Bartlett, A., 311
Ashworth, J. R., 58,310 Bassler, Eugen, 126,311
Bateman, Harry (1882-1946), 311
Bacon, Francis (1561 -1626), 223 Beatty, R. T. (1882-1941), 242, 244,
Badash, Lawrence, 310 311
Baeyer, Otto von (1877-1946), 221, Becquerel, Henri (1852-1908), 245,
245, 310 311; y rays' effect, 54; and
348 Index
Becquerel, Henri (cont.) Bragg, William Lawrence (1890-
Paschen, 62; on radioactivity, 49; 1971),213,575, 314; on x-ray
and Villard, 54; on x-rays, 20, 21 interference, 208
Becquerel rays, 49ff Bragg-Peirce law, 253
Benoist, L., 312 Breit, Gregory (1899-1981), 279, 314
beta (P) rays, 16; e/m of, 52; and y Bremsanteil x-rays, 127
rays, 152; spectroscopy, 62, Bremsstrahlung, 202, 216, 219
244ff, 265ff, 273 Breton, Jules Louis, 266
Birge, Raymond Thayer (1877-1980), Bridgman, Percy Williams (1882-
312 1961), 314
Black, D. H., 312 Brillouin, Leon (1889-1969), 314
black light, 16 Brillouin, Marcel (1854-1948), 287,
Blondlot, Rene (1849-1930), 46, 312 291,293,574
Bohr, Niels (1885-1962), xii, 4, 180, Brizard, Leopold, 274
193,215,227,297,300,307, Broglie, Louis de( 1892 -1987), xii,
312; Nobel prize, 280ff; opposes xv, xxiii, 4, 263, 267, 299, 314,
lightquantum, 281 316; education, 274; on electron
Bohr theory of atom, 215, 223, 269, diffraction, 294, 304; and
297; as attraction for mathemati- lightquantum, 286ff; Nobel prize,
cal physicists, xvii, 4, 193; and 305; and phase waves, 289ff
L. de Broglie, 292; and dispersion Broglie, Maurice de (1875-1960), xi,
of light, 281; and Rutherford's 224,263,271,280,306,575,
view of y rays, 273; and triggering 327; and L. de Broglie, 274ff; and
hypothesis, 193; and x-ray lightquantum, 269, 277, 284;
photoeffect, 267 Solvay congress, 264, 276; and
Boltwood, Bertram B. (1870-1927), x-ray photoeffect, 266ff
226 Bronson resistance, 145
Boltzmann, Ludwig (1844-1906), 18, Brunhes, Bernard, 576
107, 312 Buchwald, Eberhard (1886-1975),
Born, Max (1882-1970), 302, 312 162,576
Bose, Satyendra (1894-1974), 299, Buisson, Henri (1873-1944), 576
312 Bumstead, Henry Andrews (1870-
Bothe, Walther (1891 -1957), 278, 1920), 576
298, 313, 321 Burbidge, P. W., 576, 327
Bragg, William Henry (1862-1942),
xi,2, 61,81ff,89,94ff, 135, 137,
139, 148, 190-1, 199,208,213, Cabrera, J., 576
265, 278, 306, 313; on a-particle Campbell, Norman Robert (1880-
range, 83; and Barkla, 94ff; 1949), 183, 576; on fluctuations,
delimits waves and particles, 82; 148, 160, 163; on photoelectric
on dualistic theory, 180; educa- effect, 142ff
tion, 81; on y rays: and secondary canal rays, 16, 56
electrons, 97ff, paniculate, 82ff, Carnot, Adolphe, 266
stimulate one electron, 160; Carvallo, Emmanuel (1856-1945),
neutral-pair hypothesis, 87; on 42, 576
paradox of quality, 85; on Cassidy, David, 576
paradox of quantity, 84; criticizes cathode rays, 5ff
Sommerfeld, 150, 164, 183; on Chadwick, James (1891 -1974), 229,
x-rays: absorption law, 253, 231,274,576,556
paniculate, 85, finds secondary characteristic: y rays cum x-rays, 220;
peaks, 210 x-rays, lOOff
Index 349
chemical rays, 9 Edison, Thomas Alva (1847-1931),
Cherwell, see Lindemann 579
Chwolson, see Khvol'son Ehrenfest, Paul (1881 -1933), 109,
Compton, Arthur Holly (1892-1962), 207, 279
ix, 4, 283, 305, 316 Einstein, Albert (1879-1955), ix, xii,
Compton effect, 283ff xvi, 105ff, 176, 195, 276, 278ff,
Compton, Karl Taylor (1887-1954), 297, 307, 579; Bragg's ignorance
188,236,577,554 of, 90; and lightquantum
Cooksey, Charlton Downs, 99, 235, hypothesis, 106ff, 278; Nobel
317 prize, 280, 305; on photoelectric
Coolidge, William David (1873- delay time, 184; and Stark, 118ff,
1975), 249, 577 121, 130
Copenhagen interpretation of Eisberg, Robert, 320
quantum mechanics, 307 electric discharge, 5, 76
Des Coudres, Theodore (1862-1926), electromagnetic world view, 6ff, 104ff
577 Ellis, Charles Drummond (1895-
Crookes, William (1832 -1919), 6,57 7 1980), 231, 284, 320, 336; on
Curie, Marie Sktodowska (1867 - y-ray photoeffect, 270ff
1934), 50ff, 56,577 Elsasser, Walter M. (1904 - ),
Curie, Pierre (1859-1906): deflection 304, 320
ofa-and^-rays, 5Iff, 577 Elster, Julius (1854-1920), 320
Emden, Robert (1862-1940), 320
Danysz, Jean (1860-1928), 246, 259, energy accumulation, 173ff, 183
318 Epstein, Paul Sophus (1883-1966),
Darwin, Charles Galton (1887-1962), 193, 320
199, 251, 332; on x-rays, 212ff equipartition theorem, 171
Dauvillier, Alexandre Henri (1892- Eve, Arthur Stewart (1862-1948),
1979), 265, 575, 318, 327 320; on y-rays, 67; on x-rays, 59;
Davisson, Clinton Joseph (1881 - and Paschen, 62
1958), 304, 318
Debye, Peter Joseph (1884-1966), ix,
200, 283, 318; on Compton Fabry, Charles (1867-1945), 576
effect, 285; on lattice point vibra- Fajans, Kasimir (1887-1975), 225
tion, 217; on photoelectric effect, Faraday's experiment, 30
181ff, 185 Faxen, Hilding (1892-1968), 320
delta (S) rays, 88 Fermat's principle, 293ff
diffraction of impulses, 22, 3 Iff FitzGerald, George Francis (1851-
discharge rays, 16 1901), 20
dispersion of light, 282, 298 Fizeau's experiment, 30
Donath, B., 318 Florance, D. C. H., 221, 320
Dora, Ernst (1848-1916), 318; fluctuations: in y-ray ionization,
energy of x-rays, 72; velocity of 147ff, 160ff; in photoelectric ef-
x-ray photoelectrons, 73, 85, 112 fect, 145; in radioactive decay,
doubly directed impulse, 23, 205 143
Drude, Paul (1863-1906), 579 fluorescent x-rays, 103, 190, 202; see
Duane, William (1872-1935), 250, also characteristic rays
257ff, 318, 579 Fomm, L., 320
Duane-Hunt law, 251 Forman, Paul, 320, 326, 333
Fourier series: of singly directed
impulse, 34ff; and indeterminacy
Eckart, Carl Henry (1902-1973), 579 principle, 303
350 Index
Franck, James (1882-1964), 151, Heilbron, J. L., xv, xix, 228, 527,
305, 321 322, 326, 344
Franck-Hertz experiment, 233 Heisenberg, Werner (1901 -1978),
Frank, Philipp (1884-1966), 321 xii, 298, 300, 302ff, 323
Friedman, Robert, 321 Helmholtz, Hermann von (1821 -
Friedrich, Walther (1883-1968), 200, 1894), 10, 18,525
203, 218, 224, 321 Hendry, Jon, 323
Frilley, Marcel, 321 hereditary field, 288
Fiichtbauer, Christian (1877- ) Hermann, Armin, 323
321 Hermann, Carl (1898-1961), 192,
323
gamma (y) rays: and x-rays, 16, 61, Hertz, Gustav L. (1887-1975), 305,
65, 222; asymmetry of, 155; 527
characteristic, 22Iff, 229, 270; Hertz, Heinrich (1857-1894), 323;
discovered, 53ff; divorced from on cathode rays, 7, 18; on
x-rays, 231; named, 56; photo- electromagnetic world view, 8;
electric effect, 271 on Maxwell's theory, 10
Gehrenbeck, Richard K., 321 Herweg, August Julius (1879-1936),
Gehrtz, A., 310 324
Geiger, Hans (1882-1945), 144, 163, Hirosige, Tetu, 324
278, 298, 313, 321, 336 Holtsmark, Johan Peter (1894-1975),
Geitel, Hans (1855-1923), 320 320
General Electric, 249 Home, Rod, 324
George, Andre, 321 homogeneous x-rays, 80; see also
Gerlach, Walther (1889-1980), 331 characteristic rays
Germer, Lester (1896 -1971), 304,318 Hu, Kang Fuh, 257ff, 266, 579, 324
Gespensterfeld, 299 Hughes, Arthur Llewelyn (1883-
Giesel, Friedrich Oskar(1852-1927), 1978), 324
51,321 Hull, Albert Wallace (1880-1966),
Gifford, J. W., 321 254,271,524
Glasson, J. L., 313 Hund, Friedrich (1896 - ), 324
glowworms, 16 Hunt, Franklin L., 251, 579
Gocht, Hermann, 322 Hurmuzescu, Dragomir (1865 -1954),
Godlewski, Tadeusz (1878 -1921), 572
220, 322 Huygens, Christiaan (1629-1695),
Goldstein, Eugen (1850-1930), 322 22, 205
Golitsyn, Boris Borisovich (1862-
1916), 19,525
Gouy, Louis Georges (1854-1926), Ikle, Max (1872-1918), 51, 324
21,40,209,522 indeterminacy principle, 302
Gray, Joseph Alexander (1884-1966), infrared, 9
221,522 Innes, P. D.: on x-ray photoeffect,
Green, George (1793-1841), 21, 322 137, 235; 75,115,524
Guillaume, Charles Edouard (1861 - impulse x-rays: acceptance, 2Iff, 29,
1938), 322 43; diffraction 22, 33ff; electric
force 25; energy, 27; frequency,
Haga, Hermann (1852-1936), 31, 114ff, 158, 181; interference,
201,522 204ff; properties, 15ff, 43;
Hahn, Otto (1879-1968), 221, 245, proposed, 21; variation of width
310 with azimuth, 111, 129; Wien's
Hanle, Paul, 322 interpretation, 1 lOff
Heaviside, Oliver (1850-1925), 322 intensity of x-rays, 79ff
Index 351
interference: of presumed y-rays, 224; Kovarik, Alois (1880-1965), 326
of impulse x-rays, 15, 204; of Kramers, Hendrik A. (1894-1952),
light, 16, 140; as obstacle to 298, 572, 323, 326
lightquantum, 109, 170; of phase Kronig, Ralph de Laer (1904 - ),
waves, 293; of x-rays, 194ff 326
Ioffe, Abram Fedorovich (1880- Kubli, Fritz, 326
1960), 147, 193,524 Kuhn, Thomas S., xvi, xviii, 323, 326
ions, 73 Kunsman, Charles H.( 1890- ), 304,318
Ishiwara, Jun (1881 -1947), 193 Kunz, Jacob (1874-1938), 327
ISHTCP, xviii, xxiii
J series, 257 Laby, Thomas Howell (1880-1946),

Jammer, Max, 324 144, 327
Jaumann, Gustav (1863-1924), 18, Ladenburg, Erich R. (1878-1908),
324 136,235,527,555
Jeans, James Hopwood (1877-1946), Ladenburg, Rudolf (1882-1952),
324 177,527
Jenkin, John, 324 Lande, Alfred (1888-1957), 327
Joffe, see Ioffe Langevin, Paul (1872-1946), xii, 264,
294, 309, 327; on L. de Broglie's
Kahan, Theo, 325 thesis, 295ff
Kaiser Wilhelm Institute, 279 Larmor, Joseph (1857-1942), 327;
Kargon, Robert H., 325 radiation from accelerated
Karnozhitskiy, Aleksandr charge, 110, 113; on x-rays, 43
Nikolayevich (1867-1906), 19, Laub, Jakob Johann (1882-1962),
325 114,235,520,527
Kaufmann, Walther (1871-1947), Laue, Max (1897-1960), 194, 199,
30, 64, 245, 325 214,279, 527, 327; on Bragg
Kaye, G. W. C. (1880-1941), 92, law, 217ff; on interference of x-
241,325 rays, 200ff
Kelvin, see Thomson, William Leckey, R. C. G., 324
KhvoFson, Orest Danilovich (1852- Ledoux-Lebard, R., 527
1934), 325 Lehmann, Otto (1855-1922), 327
Kirchhoff, Gustav (1824-1887), 10, Lenard, Philipp (1862-1947), 29, 176,
18,33,525 179, 189,227,234,236,245,
Kleeman, Richard Daniel (1878- 328; on cathode rays, 7; and
1932), 83, 98, 220, 325; forward Einstein, 108; e/m of electron,
scattering in photoeffect, 164, 30; on photoelectric effect, 74; on
235; monotonic response of triggering hypothesis, 74, 177ff;
elements to y-rays, 92; secondary velocity of photoelectrons, 73
electrons from y- and /?-rays, 86ff Lichtenecker, Karl (1882- ),
Klein, Martin J., xviii, 325 187,550
Kleinert, Andreas, 326 Liesegang, J., 324
Klickstein, Herbert, 326 lightning and thunder, 44, 90
Knipping, Paul (1883-1935), 200, lightquantum hypothesis, 3; Bohr
203, 527 resists, 281; Bragg ignores, 89;
Koch, Peter Paul (1879-1945), 201, M. de Broglie resuscitates 269;
326 Compton uses, 284; Einstein
Kohlrausch, Karl Wilhelm Fritz proposes, 105ff, 278; Lorentz
(1884-1953), 144,526 rejects, 169ff; most oppose, 109;
Kossel, Walter (1888-1956), 227, Nobel prize legitimizes, 281;
255, 269, 326 Stark adopts, 121,217
352 Index
Lindemann, Frederick Alexander Michelson, Albert Abraham (1852-
(Lord Cherwell) (1886-1957), 1931), 258, 331
268 Michelson -Morley experiment, 30
Lodge, Oliver (1851-1940), 19, 328 Miller, Arthur I., 331
Lorentz, Hendrik Antoon (1853- Millikan, Robert Andrews (1868-
1928), 152, 168, 183, 194, 199, 1953), ix, 234,258,263,269,
328; and electron theory, 8, 30, 281, 331; on photoelectric effect,
110; and interference of impulses, 238ff
204fF; opposes lightquantum, monochromatic light, 15, 41, 303
169ff; and absorption of light, Moseley, Henry Gwyn Jeffries
172; Wolfskehl lectures, 171; and (1887-1915), 199, 212ff, 221,
x-rays, 43, 48, 67 224,251,264,552
Moser rays, 16
Mott, Neville, F. (1905 - ), 332
McClelland, James Alexander
(1870-1920), 65, 92, 329; on N-rays, 16
y-rays: absorption, 220, charge, needle radiation, 279
59, periodic response of elements, Nernst, Walther (1864-1941), 67,
91, secondaries, 67; and Paschen, 194, 258, 332
62; and triggering hypothesis, 75, Nernst lamp, 145
85 neutral-pair hypothesis, 87ff, 94
McClung, Robert Kenning (1874- Nicol, J., 310
1960), 72, 78, 329, 336 Nisio, Sigeko, 332
McCormmach, Russell, 329 Nye, Mary Jo, 332
Mach, Ernst (1838-1916), 30, 329
MacKenzie, Arthur Stanley (1865- Ockham's razor, 20
1938), 329 Oseen, C. Wilhelm (1879-1944),
MacKinnon, Edward, 293, 329 279ff, 332
Madsen, John Percival Vissing Owen, Edwin Augustine (1887-1973),
(1879-1969), 97, 313, 329 253, 332
Makower, Walter D. (1879-1945),
332 Pais, Abraham, 332
Malley, Marjorie, 329 paradox of quality: denned, 85ff;
Martyn, G. H., 310 restated, 277
Marx, Erich (1874-1956), 151, 208, paradox of quantity: recognized, 77;
329; and Bragg, 89; photoeffect restated, 277
delay time, 184, 187; velocity of Paris Academy of Sciences, 266
x-rays, 46ff Parkinson, E. M., 332
matrix mechanics, xii, 298, 300 Paschen, Louis Carl Heinrich
Mauguin, Charles, 330 Friedrich (1865-1947), 220, 332;
Maxwell, James Clerk (1831-1879), on /?-ray spectroscopy, 244; on
5, 10, 330 y-rays,61ff, 68
Mayer, Herbert F. (1900-1968), 305, Peirce, S. E., 253, 313
330 Perrin, Jean (1870-1942), 294, 297
mechanics: British view, 5; German phase waves, 289ff
view, 6, 104 Phillips, Charles E. S., 332
Meitner, Lise (1878-1968), 246, 330 Philpot, A. J., 310
Mendelssohn, Kurt, 330 photoelectric effect: W. H. Bragg, 85,
Meyer, Edgar (1879- ), 144, E. Ladenburg, 136, Lenard, 74ff,
151, 179, 189, 330; and y-rays, Millikan, 238ff, others, 234ff;
148, 160ff delay time, 172, 184, 187;
Meyer, Stefan (1872-1949), 331 forward scattering, 235; normal
Index 353
and selective, 188; seen as triggering hypothesis, 278
problem, 185, 189; theory of by: Rice, Marion, 254, 271, 324
Debye, 18Iff, Einstein, 108, Richardson, H., 334, 336
Lorentz, 17Iff, Planck, 178ff,O. Richardson, Owen Willans (1879-
Richardson, 193, 237, 1959), 188, 194,235,260,263,
Sommerfeld 18 Iff; of x-rays, xv, 266, 305, 317, 334; and Barkla,
73, 228, 255, 266ff, 276 257; and photoeffect, 193, 237
Planck, Max (1858-1947), 148, 170, Righi, Augusto (1850-1920), 334
185, 189, 194,217,258,263, Ritz, Walther (1878-1909), 122
274, 333; on oscillators, 108; Robinson, Harold (1889-1955), 248,
radiation theory, 278; second 261,210,335,336
quantum theory, 178; white light, Rontgen, Wilhelm Conrad (1845-
42 1923), 44, 72, 76, 79, 87, 335;
Plucker, Julius (1801 -1868), 333 discovers x-rays, 16ff
Pohl, Robert W. (1884-1976), 151, Rogers, F. J., 334
188, 201, 236, 239, 321, 333, 344 Rubens, Heinrich (1865-1922), 148,
Poincare, Henri (1854-1912), 10, 18, 194,235,555
333 Runge, Carl (1856-1927), 335
Poincare, Lucien, 322 Russell, Alexander Smith
polarization of x-rays, 19, 43ff (1888-1972), 159,222,270,
polonium, 51 335, 337
Porter, H. L., 165, 314 Russo, Arturo, 335
Poynting, John Henry (1852-1914), Rutherford, Ernest (1871-1937),
80, 333 199, 210, 220, 228, 234, 259, 264,
Precht, Carl Julius (1871 -1942), 333, 267, 270, 335, 342; on
335 a-particles, 50, 55, 163; on
Pringsheim, Peter (1881-1963), 188, ^-particles, 50, 55; and fi
236, 239, 333 spectroscopy, 245ff; on y-rays:
Przibram, K., 333 characteristic, 221, 229, 272,
Pyenson, Lewis, 334 believes has found inter-
ference, 222, ionizing proper-
ties, 54, quantum theory, 230;
quantum efficiency, 146
number of atomic electrons, 137;
Quincke, Georg H. (1834-1924), 334
and Paschen, 64; on x-rays, 72,
79,213
radiant intensity, 80, 103, 294
radioactinic rays, 16 Sadler, Charles A., 93, 242, 244, 254,
radioactivity, 50ff 310, 336; and characteristic
radium, 51 x-rays, 102
Raman, V. V., 333 Sagnac, Georges M. (1869-1928), 44,
Ramsauer, Carl (1879-1955), 177, 317,336
179, 184, 189, 239, 328,334 saturation of ionizing current, 76, 84
Ramsauer effect, 305 Schlegel, Richard, 336
Rawlinson, W. F., 267, 270, 335, 336 Schmidt, H. Willy (1876-1914), 159,
Rayleigh, Lord III, see Strutt, John 270, 336
William Schonbeck, Charlotte, 326
Rayleigh, Lord IV, see Strutt, Robert Schrodinger, Erwin (1887-1961), xii,
John 4, 279, 294, 302, 305, 337; wave
Regener, Erich (1881 -1955), 144,331 mechanics, 298ff
relativity theory: and L. de Broglie, Schuster, Arthur (1851 -1934), 209,
275, 286ff; and y-rays, 152; and 337; on cathode rays, 6; on
quantum frequency, 286; and grating, 42; on x-rays, 19
354 Index
Schwarzschild, Karl (1878-1916), 38 Starke, H., 62, 340
Schweidler, Egon von (1873-1948), Stephenson, Reginald J. (1903-
331, 337; fluctuation analysis, 1973), 340
143ff Steubing, Walther (1885-1965), 340
Seelig, Carl, 337 Stewart, Oscar (1869-1945), 55, 340
Seitz, Wilhelm Georg (1872-1945), Stokes, George Gabriel (1819-1903),
235, 337 10, 42, 206, 214, 340; impulse
Sells, Robert, 334 hypothesis of x-rays, 2 Iff; on
Serwer, Daniel, 337 refraction, 22
Shearer, G., 311 Stokes' law of fluorescence, 103, 107,
Siegbahn, Manne (1886-1979), 254, 185
265, 305, 337 Stoletov, Alexandr Grigor'yevich
Simon, A. W., 317 (1839-1896), 184,540
Simons, Lewis, 266, 337 Stoney, George Johnstone (1826
singly directed impulses, 28, 205 1911), 42, 340
Skinner, Herbert Wakefield Banks Strutt, John William (third Lord
(1900-1960), 320 Rayleigh) (1842-1919), 41, 209,
Slater, John C. (1900-1976), 298, 214,540
312, 337 Strutt, Robert John (fourth Lord
Soddy, Frederick (1877-1956), 159, Rayleigh) (1875-1947), 341; on
222, 225, 270, 335, 338 y-rays, 56, 66; on x-rays, 52
Soddy, W. M , 222,338 Stuewer, Roger, xviii, 341
Solvay Congresses: of 1911,184ff, 338; Stuhlman, Otto (1884-1965), 235,
of 1921, 257, 266, 269, 276, 338 341
Sommerfeld, Arnold (1868-1951), 2, Sudre, R., 341
105, 135, 148, 185, 189,206, Swenson, Lloyd, 341
216, 273, 275, 280, 286, 300, 302,
318, 338; on y-rays, 150ff, 157; Taylor, Geoffrey Ingram (1886-
on photoeffect, 180ff, 187; 1975), 141,547
quantum theory of atom, 159; tertiary x-rays, 45
on x-rays and y-rays, 152; on Thibaud, Jean, 274, 341
radiant intensity, 129, 157; and Thompson, Sylvanus P. (1851 -1916),
Stark, 122, 126ff, 130; on x-rays: 21
azimuthal variation, 128, 157, Thomson, George Paget (1892- 1975),
Bremsstrahlung, 127,219, 304, 341
continuum of waveforms, 38, Thomson, Joseph John (1856-1940),
303, diffraction, 33ff, 36, energy, 2, 103, 135, 152, 188, 192, 194,
127, interference, 200 206, 260, 306, 341; on cathode
spathofluoric rays, 16 rays, 6, 46; and electromagnetic
spectres corpusculaires, 267ff field, 78, 138ff; paradox of
Stark, Johannes (1874-1957), xi, 2, quantity, 76ff; and Paschen, 65;
105, 135, 160, 176, 189, 339; on and photoeffect, 136ff; and
aether, 172; and Barkla, 103, quantum theory, 142; on
119; on canal rays, 118; conflates radiation damping, 113; on
wave and impulse, 116; on radioactivity, 58; and O.
lightquantum, 12Iff; and Richardson, 236, 263; and
Sommerfeld, 130ff, 184; and secondary impulses, 44; and
Wien, 116ff; on x-rays: triggering hypothesis, 75, 136;
aetherwave hypothesis, 125, 130, unit theory of radiation, 138ff; on
azimuthal variation, 123ff, x-rays: impulse hypothesis, 24ff,
quantum theory, 117, ionizing effect, 76, polarization,
interference, 216 19, scattering, 92, 96
Index 355
Thomson, William (Lord Kelvin) 160, 184, 194, 207, 234, 345; on
(1824-1907), 18, 19,21,29,57, blackbody law, 106; on
343 electromagnetic world view, 8,
"tiger and shark," 306 110;andLorentz, 110; on
Topper, David, 343 paradox of quality, 112; and
Townsend, John Sealy (1868-1957), Stark, 116; on triggering
85 hypothesis, 75, 113; on x-rays:
Trenn, Thaddeus, 343 azimuthal variation, 111,
triggering hypothesis, 74ff, 85, 136, frequency, 114, and light, 113,
171, 179 quantum theory, 114
Trillat, Jean (1899 - ), 274, 343 Wigger, Otto, 66,546
Trowbridge, John (1843-1923), 250 Wilberforce, Lionel Robert, 45
tubes of electric force, 78, 140 Wilson, Charles Thomas Rees (1869-
1959), 165, 190,546
ultraviolet, 9 Wilson, Harold Albert (1874-1964),
unit hypothesis, 138ff, 188 51
Wilson, William (1887-1948), 159,
Varend, La, 343 245, 346
Victoria Institute, 21 Winchester, George, 557
Villard, Paul Ulrich (1860-1934), Wind, Cornelis H. (1867-1911), 38,
53ff, 246, 343 201, 322, 346; diffraction of
Volterra, Vito (1860-1940), 287, 343 x-rays, 3Iff
Winkelman, Adolph (1848-1910),
Waals, J. D. van der, Jr. (1873-1971), 346
207, 343 Wolfke, Mieczysfaw (1883-1947),
Wagner, Ernst (1876-1928), 226, 193
254, 265, 343 Wood, Robert Williams (1868-1955),
Walter, Bernhard Ludwig (1861- 282
1950), 48, 151,201,544 Wooster, W. A., 336
Warburg, Emil (1846 -1931), 194 Wright, J. R., 239, 557, 346
wave packets, 293 Wullner, Friedrich Adolph (1835-
wave theory of light, 9ff, 189ff 1908), 43, 110,546
Weart, Spencer R., 321
Webster, David Locke (1888-1976), x-rays: absorption coefficient, 17, 79;
250,251,544 absorption edges, 226, 254;
Weidner, Richard, 344 Bremsstrahlung, 127, 202, 216,
Weill-Brunschvicg, A. R., 344 219; confused with y-rays, 220ff;
Wendt, Gerald (1891 -1973), 340 discovered, 11; energy, 72, 110;
Wentzel, Gregor (1898-1978), 338, impulse hypothesis, 15ff, 20ff;
344 intensity 79, 294; interference,
Wessels, Linda A., 344 200ff; interpretations, 16ff;
Wheaton, Bruce R., 323, 344 photoeffect, 266ff, 276; quality,
Whiddington, Richard D. (1885- 78; radiative nature, 17;
1970), 242ff, 256, 344 reflection, 44, 73, 210; scattered,
Whiddington's law, 256 95, 283ff; tertiary, 45; velocity,
Whiddington's rule, 243 46ff; white, 214
white light, 40ff
white x-rays, 214 Zeeman effect, 30, 297
Wiechert, Emil (1861 -1928), 8, 30, Zehnder, Ludwig A. (1854-1949),
33, 338,345 17,546
Wien, Wilhelm (1864-1928), 2, 105, Zwicky, Fritz (1898-1974), 346

The Tiger and The Shark Empirical Roots of Wave-Particle Dualism

Uploaded by

Copyright:

Available Formats

The Tiger and The Shark Empirical Roots of Wave-Particle Dualism

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

The Tiger and The Shark Empirical Roots of Wave-Particle Dualism

Uploaded by

Copyright:

Available Formats

The tiger and the shark

Empirical roots of wave-particle dualism

Cambridge University Press

This publication is in copyright. Subject to statutory exception

First published 1983

Library of Congress Cataloguing in Publication data

ISBN 978-0-521-25098-6 hardback

Transferred to digital printing 2008

Foreword by Thomas S. Kuhn page ix

The first three decades of the twentieth century embraced two

Leopards break into the temple and drink dry the

The first years of this century witnessed the final rejection of

For all practical purposes the wave theory of light is a

This is the story of a radical change in humans'concept of light. In

STRUCTURE OF THE ARGUMENT

BRITISH MECHANISTIC PHYSICS

THE ELECTROMAGNETIC WORLD VIEW

THE WAVE THEORY OF LIGHT

The Rontgen emanation consists of a vast succession of

Between 1898 and 1912, a majority of physicists thought that

THE NEW RAYS

THE BRITISH VIEW OF ELECTROMAGNETIC

Figure 2.1. G. G. Stokes' doubly directed electromagnetic impulse.

appreciable loss of energy, and consequently suffer little refraction

Figure 2.2. Diffraction effect of Stokes' impulse.

where 6 is the thickness of the pulse and v is the original velocity of

Figure 2.3. Electricfieldaround an electron t units of time after it stops

Figure 2.4. J. J. Thomson's derivation of radial to transverse electric

lines.45 The total magnetic force is given by ^ncAE) or

Temporal width = 5/c

Figure 2.5. Singly directed square impulse as a function of time.

Note, in particular, that the energy is inversely proportional to the

CONTINENTAL RESPONSE TO X-RAY

SPECTRAL DECOMPOSITION OF AN IMPULSE

where/(0 is the known functional dependence of the waveform on

Figure 2.7. Fourier spectrum of square impulse.

assigned to that frequency in order, by superposition of all waves,

SOMMERFELD S DIFFRACTION ANALYSIS

Figure 2.8. Sommerfeld's diagram of x-ray diffraction by an edge.

screen, and by Huygens' principle a second cylindrical impulse

Figure 2.9. Distribution of diffracted impulse x-ray intensity. [Sommer-

Figure 2.10. Change in diffracted intensity with impulse width. [Based

Figure 2.11. Schematic distribution of x-ray impulse intensity after

photo as evidence of periodic-wave x-rays. He responded to Som-

THE NATURE OF WHITE LIGHT

C. G. BARKLA AND X-RAY POLARIZATION

Beam 3 is plane- As seen at A

Figure 2.12. Schematic diagram of Barkla's demonstration of x-ray

the penetrability of secondary x-rays produced in five gases was,

ERICH MARX MEASURES THE VELOCITY

Figure 2.13. Marx' apparatus to measure x-ray velocity. [Franck and

His apparatus is shown in Figure 2.13. A high negative potential on

The y-rays are very penetrating Rontgen rays [produced

The discovery of x-rays reduced much natural reluctance on the

Figure 3.1. Number of studies of radioactivity, 1896- 1903.

By 1900 it was clear that the /?-rays were high-speed electrons.

HIGHLY PENETRATING RAYS

Figure 3.2. Marie Curie's schematic diagram of differential magnetic

cles. Strutt believed that electromagnetic impulses would arise