zotero/storage/WT6HA4BI/.zotero-ft-cache

.i.
b-
. ~
.

174 VASCO RONCHI
La storia della scienza, con numerosi episodi del genere. d!mostra con tutta evidenza quanto sia dannoso per iI progresso it formarsi di compartimenti stagni, che impediscono 10 scambio delle Idee e, rendendo dlfficile la critica, procurano lunga vita agli errori.
Vasco RONCHI Firenze, /stituto Nazionale di Ottica, Arcetri

A~ c.~tvts l~{~\-",G..{' d \ \iht.

rr. ~ ~ \~J (\'e"",pj - vo \.

11S- 1'10

( \'\('S)

c··\)

Georges Sagnac and the Discovery
of the Ether(1)
\,~ J'.o~" ('h"'rr e \\

Most historians and philosophers of science today would

probably affirm their belief that their field of study is its own

justification. Without attempting to defend the extreme opposite

view that all scholarship must be tied to immediate practical needs,

I would like to suggest that any discipline is falling short of its true

potential unless it recognizes and seeks to accomplish tasks related

to the needs of the larger community of scholars, and indeed to

the needs of humanity as a whole. History and philosophy of

science is in a unique position to perform an extremely valuable

service in this regard. It can help science revitalize its theoretical

approach, by re-emphasizing the interpretation of evidence and

equations from the point of view of the natural philosopher, as

was carried out in the heroic era of early modern c science :..

Today, practicing scientists are very largely unconcerned with

the history of their own disciplines, or at least contemptuous of

opinions held in past eras; and in addition they largely ignore the

careful logical analysis of the metaphysical framework which gives

structure to their own working paradigm (2). They iail to realize

that metaphysics plays at least as great a part as measurements

and mathematical symbols, in guiding the advance of science or

natural p.hilosophy.

:

But not only do we have, at present, a group of scientists not

well informed or deeply concerned, on the whole, in history and

philosophy; also, to make matters worse, most philosophers are

not well-informed on the specifics of experimental evidence, so that

(1) An earlier version of this paper was delivered at the 8th Annual

Meeting of the Midwest Junto of the History of Science Society, Still·

water, Oklahoma, on 2 April, 1965.

)

(2) The importance of the c metaphysical framework :. is stressed

by Joseph AGASSI in his c The Nature of Scientific Problems and Their

Roots in Metaphysics :t, in Mario BUlfGBL (ed.), The Critical Approach to

Science and PhllOlOplall, 1964, p. 189-211. The c paradigm ., which I

take to mean a metaphysical framework plus various habitual, organiza­

~loDa", and instrumental appnrtenances, is discussed by Thomas KUHN

1il hiS The Structure 01 Sclenll/fc Reuolullom. 1962.

I

178 JOHN E. CHAPPELL Jr.
their philosophizing. even when intended to be about science, remains on a remote and artificial level. Add to this the very often prejudice of many historians for accumulating fact without
concern for scientific or philosophical issues per sc, and the result
is one of the most dangerous and unfortunate examples of over. fragmentation of intellectual endeavor facing twentieth-century mnn. It is in an attempt to bridge the gaps between history, philosophy, and science that I write this paper. A hit of geography is involved also. The novelty of the attempt can be read in tbe title: as tbe result of historical investigation and philosophical analysis, I am led to affirm the undoubted existence of a lumini.
ferous ether, contrary to the current beliefs of the great majorit~
of physicists.
Experimental evidence is not always what it seems, particularly to those scientists who unduly ignore rigorous logical analysis. For instance, in 1932, the report of the Kennedy-Thorndike inter­ ferometer experiment boldly. announced « Experimental Establish. ment of the Relativity of Time:. (3). In 1937, Herbert Ives unleashed a brief but devastating argument which delllonstrated that the relativity of time (if such a thing exists) has nothing to do with the r('sults of this cxpC'riment, and that the main signilicancf they have is to disprove the Lorentz-Fitzgerald contraction hypo­ thesis. Ives' argument fell short of a definitive solution to the problems raised by Kennedy and Thorndike, hut it led him to attempt another very important experiment to try to supply the evidence which Kennedy and Thorndike did not find. When he and Stilwell had completed this experiment in 1938, they still quite properly refrained from claiming that it proved the relativity or time; but adherents to relativity theory do use those results to make such a claim (4). For reasons which are beyond the scope or this paper to expound in detail, I believe that not even Ives and Stilwell made the proper interpretation of their 1938 experiment.

(3) Roy KENNEDY and Edward THORNDIKB, c Experimt'ntal Establish. 1m9e3n2t), opf. th4e00R.4e1l8a.tivity of Time >, Physical Review, v. 42 no. 3 (1 No,·,

(4) Herhert IVBS, c Graphical Exposition of the l'Ilichelson-Morley

Experiment >, Journal of the Optical Society of America, v. 27 no. 5 (MaY

1937), p. 177-180; Hf'rbert IVES nnd G. R. STH.WELL, « Experimental

Study of the Rate of a Moving Atomic Clock >, ibid., v. 28 no. 7 (July

1938), p. 215-226. The 1938 IvllS-STILWBLL results were confirmed by

__ Htili"r.

sRc,h.l

a

'M;..A,;NCD!fE;nLhl"E\i'BJo G__

al~n_d

Louis
E"lI" .... _.

_

W~I.T_~T.JBN,

c Experimental Verification of

p,..

.. ,....

.. ....... _.

___
 ..... JIo

GEOBGES SAGNAC AND THE DISCOVEBY OF THE ETHER 177
They did not because they were still handicapped by the Maxwell
theory of light, which declares that the velocity of light is inde­ pendent of the velocity of the source.
The Ritz emission theory, which adds the velocity of the source to that of the light. was introduced in 1908 but was quickly discarded. Not only did it disagree with Maxwell's theory, but it apparently was faced with an insurmountable obstacle in the form of evidence from the opposite limbs of the rotating sun and from binary stars, adduced by Tolman and by de Sitter respectively (5). But the fact that evidence is not always what it seems is again
strikingly demonstrated in this case; for now in 1965, J. G. Fox
makes clear that the Tolman and de Sitter evidence is no longer considered an adequate refutation of the Ritz theory (6). Fox introduces other evidence which he believes docs constitute such
a refutation, but admits that it makes a rather sketchy case. Once
again, full discussion of all the issues involved in this problem would be beyond the scope of this paper. I have mentioned the problem of relativity of time and the problem of the Ritz emission theory because they do bear importantly on the issues raised by Georges Sagnac anfl other investigators of the problem of ether
e drift, and because I do not think a complete understandin~ of
Sagnac's results is possible without realizing that the ~ilz approach to electromagnetic theory is essentially correct and represents the main path for the future development of physics. But for those whose present beliefs are strongly at odds with these opinions of

, 
.....

...\

I

· r~h~ji


~.·l.i

i

;

'. J"t',
I
~
! 


(5) Richard TOLMAN. « The Second Postulate of Helatlvity >, Physical ReuieLQ,v. 31 no. 1 (July 1910). p. '2~-40; Willem DB SITl'BR, « A Proof

of the Constancy of the Velocity of Light >, ProceedinUs of the Section

of Sciences of KOllillkliike Akademie van Weteruchappen te Amsterdam,

\'. 15 pi. 2 (1913). p. 1297-1298; Willem DB SITl'BR, c On the Constancy

of the Velof'ity of Light :., ibid., v. 16 pt. 1 (1913), p. 395-396. On Hitz's

Ihc'ory see Alfred O')lAHILLY, Electromagnetics, 1938.

J ((i).J. G. Fox, <!( Evidence against Emission Theories :t. American

o~lrnal of Physics, v. 33 no. 1 (Jan. 1965), p. 1-17. The recenlly-found

nJ(tenee which is supposed to refute emission theories consists of small·

H':~Il' analogues to the binary star evidence (no more valid as a refu­

'11:"t!1lh!Ininthcarneaissintghatveelvoidclteyn.eeA)s,

and Fox

the claim docs not

that meson lifetime increases appreciate, the latter claim

('tends entirely on the assumption that mass increases with velocity,

..... II(,~ tloes not occur nccordi ng to the Ritz theory,

SitU another type of argument aRainst the dcSitter evidence has heen

ot}~amdde~bmyenPlaaslcaPlriRnAcPiIpElRe,

in of

« A New Cosmology, Mechanics >, Revista

Based de la

upon the Hertzian Real Academia de

It'nClas Exaclas, Fisicas II Naturales de Madrid, v. 57 no. 4 (1963),

D. 7?'7_'7.u

iJ I!'
178 IOHN E. CHAPPELL Ir.
mine, it will be sufficient to read tbe following discussion with an open mind, to become convinced that drastic changes of one sort or another must now be made in our accepted ideas about light velocity and the ether.
•••
In the case of the Sagnac interferometer experiment, we do not have evidence which needs new interpretation, so much as we have evidence which needs to shed a harmful cloak of re-interpre­ tation which was placed over it by theoreticians who did not fully understand what the experimenter had done. We need to return to the interpretation of Sagnac, who justly claimed that he had discovered the existence of a luminiferous ether.
Ether as discovered by Sagnac need not be considered to be an imponderable elastic solid, as it was considered to be through most of the nineteenth century, after light came to be thought of as a c transverse wave J. Ether is simply a medium through which light travels, with respect to which it keeps a fixed velocity. The ether exists in, but is distinct from, any other body or medium which will transmit a light wave. The ether may be of variable density. The ether is not necessarily itself at rest, in a given coordinate system, but may move en masse with a given velocity. I believe that the ether consists of photons, i. e. of radiation particles, mostly of the lower frequencies. Ligbt passing through this ether bounces off the photons it meets ond thus travels in u zig-zag course. This view enables us to accept a constant point­ to-point velocity for all photons, at the same time that we recognize the possibility that such photons may vary in absolute speed along their respective zig-tag paths, which may be of unequal length. The variation in absolute speed results in variation in the energy received (Doppler effect). Again, I do not wish so much to explain or to confirm this view of the ether in great detail, as to suggest to the reader that a plausible conceptualization of the ether may indeed be possible; for no doubt he has read or bas heard that all attempts to posit the existence of an ether have proved to be failures.
As will be made clear below, not only Georges Sagnae but also Albert Michelson,perhaps the most skillful experimenter in the history of optics, proved by experiment that there is a luminiferous

GEORGES SAGNAC AND THE DlSCOVERY OF THE ETBER n.
ether. In view of the widespread misconception that Michelson is
responsible for proving that there Is no luminiferous ether (In the
Michelson-Morley experiment of 1887), this is a very important point to make. But let me proceed with the historical facts.
:.
The famolls « negative evidence , experiment conducted by Michelson and Morley proved merely this: that in the latitude of Europe and the United Slates, one cannot prove Iran&lational motion of the earth with respect to an ether, by means of a cross­ type interferometer. This result led to a crisis in the minds of many physicists of the day, and later became a strong argument in favor of the special theory of relativity, even though Einstein apparently used it only indirectly in his own mental operations leading to the theory (7).
Strangely enough, Michelson himself was not particularly interested in the negative result of 1887. Although he did perform another experiment in 1897, using a vertically-arranged rectan­ gular interferometer to test possible variation of ether drift with altitude, again with negative results (81, he left to E.W. Morley, D'.C. Miller, and others the work of repeating and refining the original test with the cross-type interferometer. In fact, Michelson felt that the chief value of the 1887 work was in perfecting the interfero­ meter for use in later experiments of a different nature (9). He was much more impressed with the positive results he and Morley obtained in their less famous collaboration of 1886, a repeat of the Fizeau experiment of 1851 on the behavior of light in movtng media (10). Furthermore, Michelson reacted unfavorably to the interpretation of h~s negative results which was embodied! in

(7) Albert MICHELSON and Edward MORLEY, « On the Relative Motion

of the Earth and the Luminiferous Ether -, American Journal of Science

(ser. 3), v. 34 no. 203 (NoY. 1887), p. 333-345. See also Robert SHANKLAND, .

c The Mkhelson-Morley Experiment J. Scientific Amerlcan, v. 211 no. 5

(Dec. 1964), p. 107-114.

;

(8) Albert MICHBLSON, c Relative Motion of the Ear1h and Ihe Elher _,

American Journal 01 Sclencf.' , (,er. 4), v. 3 no. 18 (Sune 1897). p. 475~79.

(9) Bernard JAFPE, Michet.on and the Speed of Light, 1960, p. 89ttO.

(10) Albert MICHELSON and Edward MORLEY, c InOuence of Motion

or the Medium on the Velocity of Light -, American J<JUrnal of Science

(ser. 3), Y. 31 no. 185 (May 1886), p. 377-386. Hippolyte FIZEAU c Sur les

hypothhes relatives A l'Ether lumineux••• ", Annates de Chlmie et' de

Physique (ser. 3), v. 57 (1859). p. 385-404.

It
i'l
f

110 IOHN E. CHAPPELL It.

GEORGES SAGNAC AND THE DISCOVERY OF THE ETHER il1

~.

t

Einstein's special theory; and. having limited mathematical skill

Georges Sagnac, born October 14. 1869. received his doctor'S

anyway (11), he made no effort whatsoever to understand and

degree in 1900 from the University of Paris. In the early yearS of

interpret the general theory when it appeared later.

his career, he worked with Pierre Curie in recording and describing

Perhaps this was one reason why Michelson did not insist on a

the properties of the newly-discovered element radium. But hi.

non-relativistic interpretation of his 1924 experiment to test ether

most important work was carried out during the years 1906 to

entrainment. Michelson had done little himself with the ether drift problem since 1897. But in the 1920's, numerous scientists urged him to make a test for rotational ether drift, to supplement his

1914, on the optics of moving media. During this time he served as professor of theoretical and celestial physics at the University of Paris. His laler years were marred by bad health, and he passed

previous work on translational ether drift. This experiment, conducted in collaboration with Henry Gale, involved sending light

away on February 26, 1926 (13). The Sagnac experiment of 1913 employed a rotating circular

around a large rectangle, over a mile in circumference. It sup­ posedly tested whether or not the ether was carried along with the

platform complete with mirrors, light source, and camera. The platform was about one meter in diameter, and the mirrors were

earth, and by obtaining first-order fringe shifts, supposedly proved

so placed that they reflected light around the perimeter of the

that it was not (12). Michelson himself said that he had merely proved what he expectcd to prove: namely, that the earth rotates

lllatform, inscribing a polygon. The source of light was divided by a mirror thinly silvered on one side, and was passed around the

on its axis. This is a strangely tame rcsponse to results which. except for the substitution of rotational motion for translational motion, serve just as well to ]lrove thc existcncc of a luminiferous ether as the 1887 results served to prove the ether did not exist! Without the development of the theory of relativity. and parti­ cularly of general relativity (applying to rotational lIlotion) in the meantime, Newtonian mechanics might never have been banished from its dominant position in the theoretical outlook of modern physicists.
Actually, by the time of the Michelson-Gale experiment of 1924, Ule existence of a luminiferous ether should not have been in doubt. For this was not in fact the first test for rotational ether drift, but merely the first test involving a rotation of the earth with

circumference of this polygon in both directions. When the platform was at rest, the two beams were exactly superimposed. The beams were then re·united and sent into an interferometer for the observation of fringe shifts. A camera took pictures of the fringes during the rotation of the apparatus. It should be empha­ sized that light source, mirrors, and interferometer with camera were all mounted on the moving platform. The only thing which might be expected not to move while the platform rotated was the path of the light from mirror to mirror. If the camera recorded fringe shifts, this indicated that the path of the light was fixed in the coordinate system of the room, and that this room therefore contained a luminiferous ether. While rotating the apparatus at about two revolutions per second, Sagnac obtained first·order fringe

respect to the sun and stars. The first rotational cther-drift expe­ rimcnt of any kind was performed in 1913 by Georges Sagnac.
(11) Michelson made clumsy mathematical crrors ill his report of bis 1881 interferllll1etel' experiment (3 eruder version of Ih(' 1887 expI" riment with Morley). and also in a 1904 paper giving the theory of the experiment he carried out with Gale in 1924. These errors were easily spotted by others and pointe!1 out to Michelson. and they do not affed the discussion in this paper materially. Sl"C Albert 'MH:m·;LsoN. « The Relativl' Motion of the Eartb and the Luminiferous Ether ~. American Journal of Scien('~ (ser. 3). v. 22 no. 128 (Aug. 1880, p. 120-129; Albert MICHEI.SON, « Relative Motion of the Earth and Aether », PMlosopllicu/
Muoozine (ser. 6), v. 8 (1904), p. 716-719.
(12) Albert MICHELSON and Henry GAI.E, « The Effect of the Earth'~ Rotation on the Vcll1cily of Light :., Astrophysical Journal, v, 61 no. 3
LA ...... 'fn.l~\ _ f"J"" iAr.

shifts, just as predicted on the hypothesis that the room contained
sllch nn ether. 
 This experiment has scarcely heen acknowledged yet in the 

literature of the history of science. The only mention of it I have 
 heen able to find, by searching through numerous volumes written 
 in several languages, is an obscure sentence in a long footnote in 
 Whittaker's Wslory of the Theories of Aelher and Electricity, 
 which reads as foIl0":5 in its entirety: II: An interesting experiment 


nE' ~(U1.3R)D

This brief biographical in Societe Francaise

sketch summarizes the remarks by Henri de PhyJique, Bulletin, no. 259 (1928),

p. 45.S; bound with Journal de Physique et Ie Radium (ser. 6), v. 9-D

"'-·-t/"

,

~ .~

1,,:i'_!1;;''"""••

112 10HN E. CHAPPELL lr.

with a rotating interferometer was performed by G. Sagnae in
J913 :t. Immediately following, references are made to the original reports (14).

Of course, physicists have taken more note of the experiment, and at least one optics text gives a careful description of it, com­ plete with schematic diagram (15). In a review of experimental evidence dealing with the optics of moving media. in his well­ known book on relativity. Wolfgang Pauli also mentions the Sagnac experiment, and evaluates it as follows: c The rotation of a reference system relative to a Galilean system can be deter. mined by means of optical experiments within the system itself. The result of this experiment is in perfect agreement with the theory of relativity ~ (16).

Given comments like PauU's, it is little wonder that cven those

few historians of science working on the twentieth century have

paid scant attention to the Sagnac experiment. Apparently it is

just another in a long and somewhat confusing string of tests

which serve as confirmation for Einstein's theories, and further­

~

more it played whicI11ed up to

ntohepianrtt ~idnuctthieonlinoef

of reasoning and those theories.

experiment

•••
But let us now attempt, by careful analysis of the basic concepts involved, to see what Sagnac really proved. Did he prove the existence of an ether, or did he prove the validity of general relativity? Or both? It is most commonly said that his results may be interpreted by either of these two theories; and yet the theories seem to be so radically different that such a compromise is difficult to accept as representative of the true condition of the physical world.
If the ether theory is accepted, mllst we think of the ether liS fixed or as mobile? One might follow Michelson's suggestion that his 1887 results could be explained by assuming that the ether is attached to the earth. But of course, Michelson, with Gale, disproved this himself in 1924, showing that the earth rotates

Blec(1tr4i)citElld.mvu. n2d,

1W96H0ITT(1A9K5E3R),,

A p.

History 43.

of

the

Theories

of

Aethu

and

(15) .8. W. DITCHBURN, Light, 1953, p. 337-339.

(16) WaUsaus PAULI, The Theory of Relativity. 1958 (1921), p. 18•• 9.

GEORGES SAGNAC AND THE DISCOVERY OF THE ETBER 183
relative to a luminiferous ether. This result does not affect the Interpretation of Sagnac's results, however, since the earth rotates much more slowly than did Sagnac's apparatus. In other words, In interpreting Sagnac's results, we can eonsider that the ether is virtually at rest relative to the laboratory, or to the earth.
At any rate it appears necessary to affirm that rotational motion in the ether has been discovered, but that translational motion bas not. This may appear to be a difficult dilemma. But when one considers what happens to the envelopes of water and air which surround the earth, the diffieulty fades away. Air and water move along with the earth in its orbit, and yet tend not to rotate with it.
Therefore they rotate with respect to the earth, in the same tempe­
rate latitudes in which all the ether drift experiments have been conducted. The ether, of course, behaves in the same way (17). But this rotational motion. also known as Coriolis force, is absent at the equator. There, air, water and ether alike do not move in a curved path bllt in a straight path, relative to the surface of the earth. This suggests, incidentally, that a repeat of the 1887 inter­ ferometer test at the equator might yield positive results (indicating rotation, not revolution). Bear in mind that the cross-type inter­ ferometer used in 1887 is equipped to detect rectilinear motion but not curvilinear motion (with respect to itself); and the rectangular or polygonal interferometer arrangements used in 1913 and 1924 are able to detect curvilinear hut not rectilinear motion.
Let us examine in more detail how the fringe shifts obtained by Sagnac indicate the presence of an ether. In terms of the wave theory, the fringe shifts seen by the camera are evidence that the two superimposed light beams arrived out of phase. This means that two waves which started together, after travelling in opposite directions, arrived at different times. This difference is visible as :t total difference in fringe positions, between two rotations In
= opposite directions, of the amount fl.p 8'ITNS/v., where N is
frt·qucncy of rotation, S is the area of the polygon enclosed hy the light path (with platform at rest, apparently). and v. is the velocity of light in the coordinate system of the room. The relationships or velocity and length to the two coordinate systems (of the Illntrorm and of the room) are especially important to keep cleahy

rE (17) Victor ANCET. in his Nouvelle Theorie de fa Relativite el de

ab,oleuct ttrhoedYelUthUenriQfrlloem,

Lyon, Sasn8

1964, p. 5, c's results.

draws

an

equivalent

conclusion