# ",

Abdruck aus den Asrr, Nachr. Nr. 5085. (Band 212. - Fe bruar 1921.)

New Theory of the Aether. By T. J. :J. See. (Fourth Paper.) (With 3 Plares.)

By \i'ay of introduction, we remark at the outset that 1 and reprinted in Lloyd'« Miscellaneous Papers connected with this Fourth Paper is occupied chiefly with the foundations Physical Science, London, 1877, pp .. 19-I48.It Will be of the wave-theory of light. The subject is presented from remembered thatL/o)'dhad experimentally confirmed Hamillon'$ a new point of view; in harmony with the electrodynamic theoretical prediction of conical refraction, and therefore speaks

~'aYHheory of magnetism, to 'which I have been led by the with authority. '

researches on electrodynamic action and universal gravitation After the appearance of Poisson's memoir of 1819, the

outlined in the preceding papers. French academicians were divided into two groups; the geo-

As will be remembered by those familiar with the meters, led by Laptac«, Poisson, Lame, contending that at historical development of the wave-theory of light, j\'w·ton, I great distances from the source of disturbance the vibrations H~)'ghens and Euler had not considered the modern 'theory of the particles are in the direction of the radius, as held of vibrations confined to the plane of the wave-surface, normal I by Newton, Hu)ghens, Euler, arid Lagrange; and the physical to the direction of propagation. Indeed these great founders group, Jed by Fresnel, Araro, and Cauclty, claiming that in of the physical sciences did not d-iscriminate between the light the vibrations are transverse 'to the direction of pro-· nature of the molecular oscillations which produce sound and )' pagation, ~nd thus exactly opposite to those recognized in the

those which produce light. But about 1817 Dr. Thomas Young, , theory of sound, - '. ,

in England, and Fresnel and Arago, in France, were led to I' This celebrated philosophical controversy extended over assume that in light the molecular motions of the aether are , some twenty years, but never led to any satisfactory connormal to the direction of the ray, like the lateral vibrations I elusion. The mathematical genius of CaucJly caIne' to - the of a stretched cord. This view seemed [ike a very start- rescue of Fresnel's experiments, by showing the possibility ling hypothesis, and thus for a time it encountered great of a medium transmitting transverse waves. Yet neither Cauchy opposition. nor Fresnel showed how such transverse waves could arise;

At a somewhat earlier period both Poisson and Cauchy and after the death of Poisson, in 1839, there was a gradual had been occupied with profound researches in the mathe- acquiescence in the doctrine, without any theoretical expla-: matical theory' of wave-motion, and each of these eminent nation of the origin, of the transverse waves in light. Since geometers presented a number of brilliant memoirs to the 1840 there has been no change in the theory, though it Paris Academy of Sciences, chiefIy between the years 1810 I often has appeared far from satisfactory' to eminent investiand 18~o. When the 'first of these researcheswe:ep,resented I gators who expect u.nbroken continuity for the whole body to the Academy the venerable Lagrange," who died 10 18 13, ' of wave-phenomena m nature ..

was still nurribered among its most honored members; and I 'In his l~cid article on Light, Encyclopedia Americana, Laplace continued to take a deep interest in the wave-theory i 1904, Prof. Chas. S. Hastings, of Yale University, states the till his death in 18z7. ' - - - j crucial difficulty more recently encountered by the wave-

It thus appears that Lagrange died b-efore Young and I theory ~f light as follows: ' '

Fresnel brought forward the theory of transverse vibrations [ »Thisgreat work of Fresnel was looked upon, as inde~Q (is Ii) for explaining the interference and polarization of II it well deserves to be" as ODE; of the greatest monuments to li,ght; but Laplace lived' t6 witness this development for ten the human understanding - comparable to JI/ewton's doctrine years; and, with his pupil Poisson, always held to the historiCal! of universal gravitation - and it long remained of almost views- of wave movement handed down by l'le w ton, Huygluns unquestioned authority.' Ultimately, however, one of its funarid Euler, that at a great distance' fcomthe source the i damenta! postulates, namely, that the vibrations are always ~ibrations of the particles of the aether are largely in the i at right angles to tbe direction of the motion of the _light,

,direction of the radius drawn from the center of disturbance, I began' to give rise to difficulties. The fact also, that the

as in the theory of sound. I theory could not determine specifically whether the direction

From these circumstances, and the new physical ex- ! of vibration of plane-polarized light is in the p!ane of polariperiments of Young, Frrsne] and Arago, on polarization and ' zation or perpendicular to it was not onlj' a, manifestjnlnterference, there arose a celebrated controversy on the wave: completeness; but it was a constant stimulus to a critical tpeory of light, which' occupies a prominent place in the inspection of its premises. The more these points were studied ~;emoirs of the Paris Academy,· 1819-1839. A brief but the more insoluble the difficulties appeared.iunttl there came lucid review of these papers as they 'successively appeared .to be a tolerably widespread belief that the theory was Dot IS given by Lloyd in his contemporary Report on the, Pro- only incomplete, but that in some way it must be esse?tially, gress and Present State of Physical Optics, made to the I,' III error. «' .. " British Association for the Advancement of Science, 1834, From the deyelopmentgiven· below it _ appears. that

5085

after considerable uncertainty, extending over a full century, 1 aether already unfolded we have been able to confirm the New Theory of the Aerher now makes it possible to i work of H}I7; :- as by AiajoraJla's experiment of 19UI, - reconcile the difficulties which so perplexed the :illustrious !' and also obtain a much simpler view of magnetism, electrogeometers and physicists of the Paris Academy of Sciences. 1 dynamic action and universal gravitation, - there is plain This greatly simplifies OUf view of the wa ve-theory of light, ' indication that we should attempt to harmonize the waYewithout introducing rany arbitrary hypotheses. And as the i theory of light with this theory of the aether.

new wave-theory connects the theory 'of light. directly !\,itlH,l., ~,<~ In venturing .upon this new line of thought, in acco-, the theory of sound, according to the views of Poisson, 18,30, ! dance with Ibe views of PoisSOJ1, 1830, it is of course under~ it must 'be considered not the least fortunate solution of a ' stood that investigators should welcome suggestions for irn-

pro'bl~~ \\'hich g~eatly bewildered some of the most illustrious provernents which have not yet been made, owing to diffi-

academicians of 'France.' . cul ties in the' old paint of view,' as handed down by trac1it"io~

. L ',(5 the, Ae t h e r 'is a Gas, and thus .Lsot ro p i c from the days of YOLmi, Fresnel, Arago, and Caucll_l'. ,

j~' alt'Dir'e'cti'ons for' Ordinary Terrestrial Distailce.s, . In preparing the third paper we discovered a new method' it is a Fundamental Er.r o r in the Wa v e+T'b e o r y of for determining the absolute density of the aether, and I_ight to den·y. Radial ?llotion, 'in 'Order to ,hold to developed a 'process by which we were enabled to calculate the Doct~ine of Vibrations a lm o st wholly transverse this density at the surfaces of the sun and planets of the to t h e Direction ofa Ray.· . solar systein, This new method. was found to. beapplicabl-

" It is full), ;ealized that the, modern wave-theory of light ,0 an)' 'stellar or .sidereal system, where the force of gra"ity is so vast a subje'ct. 'til-at an}" treatment, even of the foun- is known by - observation, and thus may be extended through;

dation~' merely, necessarily is much more incomplete than out the immensity of space. . ,

. those give~' in standard treatises 1) on light. Yet el'en a partial . The me:llod has proved to be of great importance in

discussion of the foundation principles, provided it unfcil:::!,s confirming and definitely establishing the small density of the a new aspect of the theory of light, may be welcome to aether, in accordance with the views of :YCU:fOll, Rei'sr/icl,

iJ1 vestigators who seek the laws of nature. Kelvin, .and jJiaxlldl. This not 011 ir does a way with the,

, Thus r'deem it worth while to present the results at strange claim put forward by electronists that the' aether which I arrived. Under ilO other principles have 1 been able to mar have an immense density (estimated to be 2000 million

. bring jbe varied phenomena oflight into harmony with those limes that of lead ll, but also definitely establishes the 'com'

of electricity, magnetism, and gravitation, ., pressibility of the aerher when powerful forces act quickly,'

. , And since Prof. ,}lJajoralla, of Rome, in the Philoso- as in the explosions of dynamite, which was successfully phical Ivlagaline, vol, 39, :".1 a)" 1920" pp. 488 - 504, has been employed by Prof. Francis E. Nipher of St. Louis, to disturb able to confirm' experimentally the conclusion respecting the q uiescence or the medium.

gravitation to which I was led iii 1917,' (Electrod. W"I"e- . Since the aether therefore is a gas,' with properties Theory of" Phys, Forc., \'01. I, p. I5S) - that the amount which make this medium approximately isotropic for ordinary of matter within the. heavenly bodies is' much greater than distances at the surface of the earth, though aeolotropic in we heretofore hare believed, actually making the sun's i true respect to' the heavenly bodies, as distant centres of wavemass three times that accepted by astronomers,,' - we see agitation, we percei ve that the doctrine of the wa ve-theory evidence of a coming transformation of doctrine in physical of light, that the vibrations ani whoil}' transverse to the science, 'greater than any which has occurred since the age direction of the ray, rests on a fundamental error, and a

.of Kepler, Galiki, and Netoto». The new theory of the lunar correction is required to take account of the gaseous character fluctuations, motion of Mercury's perihelion, and of the pro- of the aether, and its equal compressibility in all directions. blerns of the aether treated of in A~ 50,P, 5048, seems to Thus, contrary to the assumptions of.Grem, and others, who have triumphed incontestably, get rid of the longitudinal component by' arbitrarily making

U rider the' circumstances it wil1 not do to shut our that component of the yeloeity infinite, there is a longitudinal eye.s to ~elV conceptions just because they have not been component in light, as in sound'; but it is very small, because handed down' by traditions. When so many difficulties hare it depends on the ratio of the amplitude to the wave-length arisen in the wave-theory of' light, which can not be over- A/i- = 10 -5, due to the very slight compressibility of the com'e on the old theory, it seems to be a sign of error in aether. The longitudinal component. thus becomes A = the assumed foundations of the theory itself; and the need (Al},)' (!" where (! is the spherical projection factor, about for a,',modjfi~ation of the theory is therefore urgent, not' only 1/.10, deduced from Fig.l, Plate 7; so that the 10 ngitudinal comiO; the, hope of winning' new truth, but also of attaining har- ponent probably does not exceed 1/400000 0 ~). _\ccording,

mony and simplicity. . to the Yery accurate experiments described by Prof. Haslillgs,

, H by fotlol\"illg' the principles of the new theory of the in section 5 below" Hlt_rghod construction for toe extra-

line;,: theor 'is r e ordir Thus '~able from ihe o mori:

(AN

and who waye at a of t1 as 11

Ara) tradi pain was expl:

and post' Eng:

to a that oscil eqm tion theo

tbe tion that pen' You

1) Aino~g the great standard treatises on light, tij~t by Sir 70/ill . Hrrsc!ul, Encyclopedia )lelropo!itana, 1849, is'to be especially commended for its comprehensiveness, .and because' it reflects the stare of the subject JUSt after the .epoch of Y('''lIg, Fraud and ..1"06'0, n?'ud/s Theory of Optics, translated by .J10"" and ,l/ii!iiaJfl, :Longmans, Green & Co" London and C'\e,v York, 191 if is the best recent treatise with which I am familiar. Lord Rayl"'gks article ,\,,,"c,Theory, Encyclopedia Britannica, 9'h ed. 1887., presents a masterly survey of the subject, ·based· on great personal experience, and may be unreservedly recommended. . ,

~) Compare the later calculation in the notes of Sept. 12 in section 4, and in section 8, below, which indicate that this component is about" r : (66420' 10"),

wou' not to " men of J

.n the'

9, - tetro· " plain waYe·

leCOr_ Inder·' 'f irri.'

diffi· :IidO~'

ethod

and :urai~ f the' cable 'a':;t; Jugb,~'

ce in ,f th~' schii( ) thi ethe; iliion com~' ckl;/

,.,

;fully sturb

. nary

iclll

.'. ;

.ave-

eery,

th,e ld:~ Icter

.Ofl S, ,_

who king lina!

~ ; ~

S08j

390

-'.

ordinary wave surface certainly is accurate to 1: r06, which Ii therefore lends a remarkable support to the new theory of ! tranwerse waves in light. i

Finally, it remains to point out that although in our I new theory of the aether we' usually speak of the waves as I 'r~sembling the waves on the surface of still water, - which! con "ey to the mind the image of particles revoh'ing j~ cir- II cular or elliptical paths, while tbe wave form moves on, - yet, 'as in the theory of sound, it is allowable, in many! phenomena; to conceive the oscillations of the particles to I take place in such narrow ellipses as to be practically recti- i linear, in the normal to the wave front, according to Poisson's : theory of J83o.Such approximate rectilinear motion always

is referable to simple harmonic motion, according to the ordinary theory of uniform motion in the circle of reference, Thus our theory is not restricted in any way, but is applicable to any possible elliptical oscillation of the particle, from a circle .on the one hand, to a straight-line ellipse on the other, as in the displacements referred to simple harmonic motion in the theory - of sound.

In the third paper on the new theory of the aether (AN Sc;()), near the end ofsection 8, equations (86) to (88) and beyond, we have carefully cited the reasoning of Poisson, who devoted over z S years to the mathematical theory' of waves, and in his last papers (18 I 9-1 839) maintained that at a g-reat distance from the source of disturbance the motion of th~ molecules always is sensibly normal .to tbe wave front, as in the theory of sound.

Thus Poisson never concurred in the views of Frcsne], 'Arago, and Cauchy, which rwer e gradually adopted in the traditional wave-theory of light. And it must be plainly pointed out that Fresnel's doctrine of purely transverse wa yes was an assumption pure and simple, which offered a needed explanation of the inrerference of polarized light .

It is a matter of authentic record that at 'first Fresnel and Arag a hesira ted to take such a radical departure as to postulate transverse waves (cf. Arago's. Eulogy on Fresnel, English translation, Boston, 184(), pp. 212-2 13 l,

In regard to the reluctance of the early investigators to admit a lateral vibration in light, it may be pointed out that Hltyghtns, Aewtofl, and Euler had held to the, view of oscillations chiefly in the 'line of the rays; though Euler'» equations involve no necessary restrictions as' to the direction of vibration', being of the same general form as in the theory of sound,

1857. pp. 332-333, Dr. Wheu'dl quotes the remarks of Fresnel: »:'1. Yeung, more bold in his conjectures and Jess confiding in the views of geometers, published it before me, though perhaps he thought of it after me.« And from. personal information of the progress of the theory of transverse 'wa ves,

Dr. vI/lieU/til adds: . .

"And 1\1. Arago was afterwards wont io relate, (l.take the liberty of stating this from personal .knowledge) that when he and Fresnel had obtained their joint experimental results, of the non-interference of oppositely-polarized pencils; and when Fresnel pointed out tbat transverse vibrations were the only possible translation of this fact into the undulatory theory, he himself protested that he had not the courage to publish such a conception; and accordingly, the second part of the memoir was published in Fresnri's name alone. What renders this, more remarkable is, that it occurred when M. Arago had in his possession the very letter of Young (Jan. 12, 18 I 7), in which he proposed the' same suggestion.e

From the circumstances here reported it will be seen that Fresnel and Arago did not feel vel}' secure I) in their position, under the criticisms of Laplace, Poisson a'nd' their followers. Accordingly Fresnel and Arago were more than glad to have the mathematical support of Cauchy, in favor of the possibility of transmitting transverse waves, if once they existed. But that was all that Cou(h~y's analysis proved. It did not indicate how such transverse waves would arise in nature, nor did Fresnel and his followers throw any light

on this di rticult problem. . , ,

, Accordingly it appears that the origin of the transverse

vibrations in light has never been explained on a satisfactory basis; and for that reason it is hoped that the simple theory in section 4 below may commend itself to geometers and

natural philosophers. . .' . . .'

Another difficulty of quite fundamental 'character m the wave-theory of light has been before me for many years. We commonly .: have offered to 'us for illustration of transverse waves the vibrations of a single stretched cord: this looks obvious and convincing, when we deal only with, a single

cord free, to, vibrate in empty ,space. . .'

, But in the theory of light we should have to imagine

all space, in the sphere V= 4/STr r3, r = 0, r = r , about the source of light, filled entirely -full of such cords, which wOlild thus mutually crowd each '·other on 'e'v'eryside; so that no one of them would have the assumed freedom of the single cord used 'in our class-room illustrations. The surface of the sphere has the area S = 4Trr~, and for a spherical shell of thickness dr, the volume is, 4Trr2dr,and

the integral of volume isV= 4/'TSr2dr. :'

: Now by no possibility can the sphere surface S = 4n"r2 be increased. Accordingly no ODe cord can be moved side' wise, in transverse vibration, without crowding all the other cords extending outward from the centre, unless we assume simultaneous motion of all the. cords in the same direction

u=asin[zrr/X·(Vt-x)]." _(I)

But at' length, Young began to .entertain the idea. that" the molecules of the aether might oscillate in parallel directionstransverse to the direction of the ray, though he thought that longitudinal vibration's might exist also. Fresne! independently rea ched the idea of transverse vibrations, but like Youfl,f he could not account for it dynamically,

In his History of the InductiveSciences, vol, II, j,d ed.,

. . ... ,.- ' - ,-';" .;., ~..." • -, . • ~.I· :.:-,

') In another place,l--list. ofthe Induct. Sciences, vol; II, p. 350, Dr. Wh<"wdl expl~in~;he embaTra's~'ment',;~f AT~io asfollows :M.j;~g:~ would perhaps have at once adopted the conception of transverse vibrations, when it WaS suggested by his fell<m'.!abourer,"Fnmd, if ,it had not been that he was a member of the Institute" and had to bear the brunt of the war in the frequent discussions of the undulatory theory, ,to which theory Lap/au, and other leading members, were so vehemently opposed, that' they would not even 'listen with toleration to the ·"iirgu: ment, in its favour. I do no! know how far influences of this kind might operate in producing jbe delays which took place in the publication

of Fresnei'« papers.' ' . , , "v.

•

'.,.

39~

, for the -sph~rical shell 41C r1 dr.', The chances are infinity to JI

one: against this occurrin¥. '- ,: ':",: " . ' _ -

, -ct ,-!;,r,' -These c6nsider~ti ons - alone ,sha _w t ~a~ the 01 d wave .the~ry 'I of light: is' inadmissible. The same difficulty does not arrse I iii"Pbiisi'n's'·,theory'of- 183'0,' which-makes the vibrations norma] to the 'wave -front, as in sound, and thus allows vibrati.dnanrid-ea~e'cif' space 'equal to d V- , .in,..2 or, where dris "th~~mplitude' of the oscillations." Wit~ the new theory

. £s"to-,',ihy' the waves' are mainly transverse,' mote' 'fully set II forth in' section 4, below, it is believed that the last out- , standing difficulty in the wave-theory of 11ght has been re- . moved. But before quitting this subject, we' may stale the expansive difficulty pointed out above with somewhat greater mathematical rigor; If (]) be the .velocity-pctential, we ha ve

the 'usual' differential expression '

, d~ = udx+vdf+wdz. - {z} I

Now it i~ \vell k'no\vu t'hat the" line integral' of the I tangential component velocity around any closed curve of a ' moving (incompressible) fluid remains constant throughout a;n'time; so that when d$ is a complete differential" the

circulation S d$ is zero, just as in the, obvi~u~case' when the fluid is at. rest: .

dSdW = oS(udx+vdj'+~dz) = o , (3)

_ ' . 'When the fl u'i d i sincompressi bleth i s integral round a' closed Circuit' is evanescent, and the. momentum, like the circulati-on, is zero; but for a compressible fluid. the existence of a velocity-potential f1! does not imply evanescence of the integral momentumround a closed circuit (cf. Lord Rayleigh,

Theory, of Sound, 2nd ed., 1896, vol. '2, pp. 8-9). _'

,I~ the case' of the aether, however, the fluid is so nearly absolutely incompressible that 'the above. theorems wili hold, and we may taked(]) to. be, essentially an exact differential; so that the velocity in any direction is expressed by the corresponding rate of change of (]) r_ and therefore

du/dx+dv/dy+dw/dz = '0 t(])/ox2 + 02([JjOy2 + 02 (])/Oz 2 • (4) Let us now consider any closed surface; such as that of the sphere already spoken of, S = 4 Tp·2 .: Then the rate of flow of the fluid outward, across the:element dS, becomes:

dS·d$!dn. i

" ;: ,', " --, "i

And, 'when the density is constant, the total Joss of i

fluid in: time. dt is given by the double integral:, i

- (ojOt)(4/S1rfU,3).:= SSd<D/dn.dSdt· . (S) I

where the integration is to be exten~ed over the entire sur-

face'S = 41ry2. ' I

. ~ow when the sphere surface S is full ~oth a; the I beginning and. at the end of dt, the loss of fluid vanishes, I'

so that, .

(%t)(4/s1C(Tr3) = SSd(])!dn.dSdl= o , (6)

I

jl

b)

The equation of continuity, for an incompressible fluid deduced from the spacial element dx dy ds , under this conditionof no loss of fluid across the boundary, is

, a~([J/ax2+a2(])lay2+021D!az2 = o

or briefly. y2(]) = o.

5085

392

And as Poisson's equation of wave motion is--

a~(])!OtZ =' a2 \72(]) - <..'

we see that y2$ = 0,' excludes the existence 'of this condition held rigorously: for the time CIt.. , , : Wherefore we conclude tliat - in traversing -the surfac'e

5, 'the condition in( 6}" will hold for' the- wave from the 'centre at"tbe beginning' and "also at the end of the time dt; corresponding - to the propagation of a wave through all its phases': over the wave-length)', which represents a complete oscillation of the' fti1i-d.'· . ,

But -for shorter in ter I'll I s, the equati on (6) will not hold rigorously; so th~t'iemporarilr, over an interval less than the wave frequency, T = - 2/t/V = ;.; V,' there is both slight compressibility and a flow of the flu~d across the bOem:

dary S:_-- 4;Ty2; and, for ot<T we have: '

(O/dt)(4/S1r(T1'S) '. S S d(])/dn' dSdt =±dm (9)

where dm is the total fluid )emporarily lost, an infinitesimal mass positive or negative ..

Accordingly, in the wave motion of the aether, there is slight compressibility, and a minute temporary radial motion of the fluid does take place. Hence we cannot have purely transverse motion, as, assumed in the traditional form of the wave-theory of light due to Fresnel and Cauchy.

During the last half century these problems have been

. discussed by many eminent natural philosophers - Lord Kelrin, Maxwell, Lord RaJ'leigh, Larmor.: Gla::.ebrook, etc., - but whilst they give up Grces'e views, they- do not reach sarisfactory accord in their views of the aether. A useful summary of their reasoning is given in Daniell's Principles of Physics, 3rd edition, 1895, p, 51o .. Under the circumstances we have felt that the older views must be entirely abandoned, and th~ waves in the aether treated as in Poisson's Theory of '1830, There is no experimental evidence of different velocities for compressional and distortional waves, and no such assumptions are authorized by the existing state of our knowledge.

_ 2. _M'ax/lJell's Electromagnetic Theory of Light

rests 'on Vibrations wholly transverse to the Direction of a Ray,"and thus in View of the above Considerations the Electromagnetic Theory also must be rejected as not based strictly on the Laws of ~ature.

,lie have just outlined the geometrical and physical difficult}" encountered by Fresnel's classical conception of vibrations wholly transverse to the direction in which light is propagated i and have shown how 'waves flat in the equators of the atoms, under haphazard arrangement of the atomic planes, would be equi valent tq the uniform spherical distribution of the elliptical vibration paths exhibited to the eye in Fig. I, Plate j. This new principle in the wave-theory of light gives two remarkable results:

I, From any 'spherical source of light, or luminous mass, where the number of atoms is large, it would Jead to vibrations so nearly transverse, that the longitudinal component probably would not exceed the value I /(4' ro6), and thus be insensihle+] to observation in optic a I experiments.

') A much smaller value 1/(66420"106), is reached in section 4 below, Sept, 12, 1920,

Fresno' < 'I~ieral \ cally, as

.. of }lcwtl

ttat a tt ~~cn as 'ratio A] ray, witt

iiiipropri

" T

o(the : the sma lacking o';er it between The emi analyst 'and La,

strious

arose, b the' ord:

If th&1 the Euler, ; correct, disturbs front. ( is nearl

. spherica at the! . cording: cruci al I philosoj

If treat of u>ell, w underlv:

1904, I might 1 principl 'that it rnechan follower light in late. II seems i I mean perpenc

':"';'

393

tface , the e di 11 it; p(ete

2. It makes the molecules oscillate primarily in the

direction of the normal to the wave-front, as held by Hu),ghms, Nadon, Euler, Lagrangr, Laplace and Poisson, prior to the theory of lateral vibrations of the stretched cord introduced bj' Young, 'Fresnel and Cauchy. Thus we have at once a vindication of .the profound wave-theory of Poisson, 1830, without need for recourse to the artificial and dynamically inadmissible theory of Fresnel, that the vibrations are wholly

transverse. ,

The above citations from Whl'weil show that Young, Fresnel and Arago were loth to entertain the theory of purely lateral vibrations, which they could not account for dynamically, as contrary to the views of geometers since the age of i'lewton. Apparently it never occurred to YO/tTli? and Fresnel that a theory of projection for Poisson's normal elliptical paths, such as is shown in Fig. J, Plate 7, multiplied by the small ratio A/)" would give mean vibrations almost normal to the ray, without the strained and unnajural theory of lateral motion appropriate to a stretched cord.

The theory of lateral vibrations, drawn from the example, of the stretched cord, is approximately correct, as respects I the smallness of the longitudinal component, . but it 'is wholly lacking in physical basis, as shown 'above in section I. Moreover it introduces an unfortunate and unnecessary conflict between the doctrines of experimental physics and geometry.

, The eminent experimenters, Fresnel and Arago, and the great analyst' Cauchy, were thus arrayed against Laplace, Fa isson , 'and Land; yet apparently it was not possible for these illustrious academicians to settle the controversy which thus arose, because the premises in their reasoning departed from the order of nature.

, If the theory above traced be admissible, it follows that the claims of geometers since the days of jVi?1()ton and Euler, as put forth by Laplact and Poisson, certainly were correct, that at a great distance from' the source of the disturbance the molecular oscillations are normal to the wave i front. On the other' hand, the average vibration in light i is nearly normal to the ray, owing to the effect of the I spherical projection from the variously tilted elliptical paths '! at the source of the .light, and the smallness of A,/j..- Ac- I'

. cordingly we are impressed with the necessity of, the most crucial test of the, premises underlying our reasoning in natural I philosophy. I

In order. to outline this defect clearly, we shall no,,-: i treat of the difficulty of the electromagnetic theory of A-fa:x- i well, which will also, show the unwarranted assumptions I underlying the Fresnel-Cattell)" wave-theory.. ' . i

-rr I knew,« says Lord Kelvin, (Baltimore Lectures, I 1904, p. 9) »what the electromagnetic theory of light is, ~I i m~ght be able, to think of it in relation to the fundamental i pnnciples of the wave theory of light. But it seems to me : that it is rather a backward step from an absolutely definite i mechanical motion that is put before us by Fresnel and hisi

.. f?llowers to take up the so-called electromagnetic theory of , lIght in the way it has been taken up by several writers of I late. In passing, I may say that the one thing about it that i seems intellia-ible to me, I do not think is admissible. What I 10,

mean. is, that there .should be' an ~lectric displace-me?t',

perpendIcular to the line 'of propagation and a magnetic

1).

(8)

not' less botb ,oun:

, "l'"

(9)

imal

> .... ~

here )tion rrely f the

)eeu {vin,

but aris;uln-, s 'of aces oed! ~ory ·rent .

no

our \.

';(, a hi

I;} ~'

e c-.

o~'~ be .

,<1' ..

re;, " ical

of' ght .,

ors

nic"

5085

394

disturbance perpendicular to both. It seems to me that when we have an electromagnetic theory of light, we shall see electric displacement as in the direction of propagation, aod simple, vibrations as described by Fresnel with lines' of vibration perpendicular to the line of propagation, (or the

motion actually constituting ligbt.« .

If Lord, Kelui» had such difficulty in understanding the electromagnetic theory of light, it undoubtedly is very allowable for the present writer to attempt to put the theory

of light on a simpler basis. .

The figure from llfaxlUtll's Treatise on Electricity and Magnetism, vol. H, 'p. '439, cited below, will put before our minds the electric and magnetic vibrations, conceived to be in planes at right angles to each other,and thus calling forth the above severe criticism by Lord Kdvin, who was long an associate and friend of Maxwdl. It seems to be certain that Lord Kelvin was very much bewildered by the unnatural complications of the electro-magnetic theory, and,

thus it proved of little or no value to him. .'

, In his Electricite et Optique, 190 I, p. 73, Poincare has pointed out the difficulties and contradictions he found in following Maxwell's processes. ~ II ne faut pas attribuer a cette contradiction trop d'importance. rai expose plus haut en effect les raisons qui me, font penser que Maxwell ne regardait.la theorie du deplacernent -electriq ue ou du fluide inducteur que comme provisoire, et que ce fluide inducteur auquel il conservait le nom d'electricite, n'avait 'pasa ses yeux plus de rea lite objective que les deux fluides de Coulomo, 'I-

The importance of having a pe-rfectly clear understanding of- ~Y£a:r:lOell's electromagnetic theory is so great that we quote his reasoning in full. It is not very long, and the deductions "viii justify it (pag. 438- i 9 -40).

»790. Letus now confine our attention to plane waves, the fronts of which we shall suppose normal to the axis. of Z. An the quantities, the variation of which constitutes such waves, are functions of z and t only, and are independent of x and y. Hence the equati7>ns'of magnetic induction, (A), Art. 5q I, are reduced to .

a = -dG./dz b = dFjdz

c= 0

or the magnetic disturbance' is in the plane of the wave. This agrees with what we know of that disturbance which constitutes light.« , - ,. .

) Putting pa, pft 'and Pl for n, band C 'respecti vely,

the equations of electric currents, Art. 60/, become '

47(~~u = -db/dz =' -dzF/dz2 4T(~~ = da/d:: = -d~G/dz2

[14] (I ii

47(PW. o ,

~ Hence the electric disturbance, is also in 'the .plane of the wave, and if the magnetic disturbance is confined to one direction, say that of x , the electric disturbance is confined to the perpendicular direction, or that of y.«

~ But we may calculate the' electric disturbance in another, way, for. if f,. s. ,h are: the components o( ,e!ectric displacement iq a' non-conducting ,medium,.' "", _ .

, ;:' u . dj/di' v.~· dgldi.-~w -,dh/dt.. [~SLJ12_), " .'If,P' Q, R are the components of the electromotive

intensity,

. ~. ~. ..,.

. .

'i'.";;'.~.;1!1f" => .

..., .. ' '.~ ',: .,'

"or " ",~ ......

'~~"." .F~rlp"a I:i~ ~i;1 h. e?~ ;,':~ l.I:e~, "'! it~, }h.ose. giye,r,lj in- .ecpla tipp,: [~4 J,

., '. we find' ",' d" 0 In' 0 l' '2""""'" ,., .. ,', • '.

:.. . ..... , -Fjdz'.. J(Wd'FJd'-" _" ,,_ . ,.j "

";i::d2~/~,~;~ :,,;{~,:.d~(;/d.~~;.~,;, '. fi~.' d2Hjdi2.:·~ ,~~:r,( :6}

:'::

.-', ';.-

: •• y"

:;.~; fl'2:; Itl appears- that, ;Jfaxwdi did, ilotreg~rd the elect/f or magnetic vibrationsashaviog ariy kind-of vertical rotati~'c; asl: I ~e'~il't'e f?r~f'm'o,~es" on .. "' .be~lluse·' ,he\ expressl y stat~t~ near·t,~e. close o[ 5ectjon:~9il~' that: ~thjs'correSponds to./: ray·c of. plane~polari.zed:·ligh~,«')V?ic~.,!n:he o~tJjO?ox. c1assic~; ~h~o.ry of !r~~n.d_}s· .:o~ceJ\:ed to be' direct .linear ~'lbratTonS;: a_t'nght angles:.t? !he_dlT~CtH:nrof th~.rar;i ~S":~hon'n inJl.{axp wdfsfigure:i ~'.;~I,:·:'.;' i 'Y,:J ':,: ,,:,,"'. ,;~' .. ~l;.>"'r'T .;, i.ii.(;(

.. '::.,3' After .. u_:u:h in~'~~tigati(:m, :v~.h~ve reached t~e c6~f elusion that such: SUpposItJonsare. pure hypotheses.vnot justified by arrything in. nature. For we C.a!lDot· hold the aether to be ia superfine" gas;' the: aetherons having" all· the degret'i

. ~:;;'.,." » The first and second of ,t.hese equations are the equa- of rr~edom appropriate to' Poisson's equation':' ,:., : .. : .- 1

~;~~e~~,~~O~~t:::'nfo~~.~, ~l~,ne way~:: ... a,~,~' their so;~tiO~' is ~~;"h,":a?w.!c!,2,:'.".'~2 (~2,o//:a~:'~?,2~/O/~ f),~£?/a~)'-'_':'" (.1 9)

".;~~: _:;;::~:~/ F = 11 {i"':: Vt)+h(z~ rtY·, . =: •. ' :~~Y"·,.c -. ~~~.i.~." .. t.z .. o ..... :·,a.d.m.: ... :~~: .. _:,t.~r~e :c()m._, ..• ·.p. ?n~nt·~bti6n,'~,.'.:,'.~,~p~:n~ing o~,

; r .s . ,:: -''". G~/3{z~Ft}+j;(z·+Vi)-: 'Izo) (~'il . '

..• "~ ';''', ;.,' . .-. . . ", ,,' .. 4, .' r .' ,,1';.;" +, -:(here. wa.s. a. c.eie~r.~.t~d cont~oy:ersy "on. this poini

,The solution of the third. equation is. .. between Poisson and F1'tsl1e( and their, fo.lJo~'ers" in. the In:

. ".:'., ;' If' __ 'A+Bt" ':,";, O~f lI'S) ~t.itute:.p~ Fran~e,.(ISI9-:-f.83·9},)·ut.:to--jJie end PiJ;-SSQ~ held

''''here. A and B .are functions' of x: H is therefore' either to the,. conclusion, .that)n generaL"' ibe.Yibral.jQ~s_' are not . constant or varies directly with the time. In neither case I normal .t~ the ~directiori. ':o( ,tt1e.: :ray.: F1-csn(l_hi~~elf held can it take part in the propagation of 'waves. ",:' ','; i suc.h views, .in,.)·ir·tu,e',:or,_ 'the' necessitYc'"of'~xplairilng p~larj;

:::: ~ 7 9 j':"It appears frain this that' ;, ' .. , ,. i za!!on,!nterfer_en~e,.etc. i. and Ca.ueh)"s mathematical.researches

thedirections, both of the magnetic -: __ '.' j seemed to indicate, that if ~:ibrations e'xis'tedno'rmat' to the

. ~nd ,the' eledric'disturbanc~s, lie .. .: , :"..0"< II rar. they could be propagated in:the .. aetheT:'· .- .. c.· ~ ','

ITI the plane oft~e wave. The' __'. Ji4"#." ..•... '. 5· There i~no .. doubt. t]la,t.' any kj~d. ~fvibr~~i:on~.· once

mathematical form of the' distur- <,-p;::'.'~, . established in the.aethe·i:-,.may, be propagate-d inth~tmedi~l~';'

barice therefore agrees with that'" :J':':;:-; " I but this does not show that the actual vibrations in polarized

of the. distur~an'ce which .consti-· .,',. j~" I light ar.e,of;b~:>. typ~;,; Bert': is af~ndamental ~rror in tbe

rules light; being transverse to the,' ::-.:...' ~~ -, ''4 ... 1~.''.:',..· J. " •• " II ;;.aye.th~r~ of. light, . which the ~·a\;e·th"eo~y or' m~g-netis~

direction of propagation.« .' ... ~ '''''';' ~ as ena .e ... us tP. correct .. ::.: ,._- .. j __ ' .,">. ,') :,':. '.:' .

':': >>If we suppose G = 0, the \'-.:: _::;:. , !" 6. We hold. that, light must have- a' longitudinal co'm·

disturbance will correspond to a /~. i ponent depending on the ratio of the amplitude. to the wave

plane-polarized ray of light.« /~..f:J. !Ic:ngth, which is small but finite, In the Philosophical ~Iaga'

'. »The magnetic force is: in :;:..---:::.-;. /. _' I zme. for Sept." 1896, Fitzgerald has a thoughtful .. and useful

this' case parallel to the ~is of y. :::.. ~ paper on this subject, beginning as follows: :~' " ,

and equal to I/~' dF/dz, and the. :., ~ ; .. ", I . »In most investigations' 00 the propagation of light;

electromotive intensity is' parallel I attention has been 'concentrated 00 the transverse nature of

. to.. the" axis. of.';'; and equal to I the vibration. Longitudinal. inotio?s ha:'e been relegated to

-:::-dF/dt.,· . The' magnetic force .:,' I the, case- .of pressural waves, and mvesngators have devoted.

is therefore' in'i plane' perpen- . I" themselves. to separating: the: two as rnuch as possible. In

dicular.to. rhat.. which contains·· .. .' Sir George Stokes's classical paper on Diffraction.jand in Lord

, ' . ". ., l'ig;'2 = lifa.;:-wdfsFig. 67.

the electrici';1lensity,c... .,,) Kelvin's. Bal?mor~. !--ectures, th.e ~xistence of a longitudinal

» The values of the magnetic force and of the electro. component IS mentioned: but it 1S 'mentioned only to show motive intensity at a' given instant at (jiff-erentpoints of the that it: is' very. small and that the motion is mostly transray; are represented in Fig. 6i, [cf, Fig-:'2},. for the case of verse. Now the lon~itudjDal. comp?nent i~ no doubt geneasimple harmonic disturbance in one plane. This corresponds I rally small; .except m- the immediate, neighbourhood of a to a ray of plane-polarized light,but whether the plane of I source; but ~t by no mean~ fol!ows that, as a consequence, polarization corresponds 10. the plane of the magnetic distur- the actual dIrectIOn of m?tlOn 1S transverse at all points in bance, or to the plane of the electric disturbance remains a .ware. ' In' every complicated wave there are points and

to be seen. c.. '. ' . .' " ' .: ! ,', .. ' :' . r . . often lines along which the transverse, component vanishes;

., .. and. at all' these. places' the small longitudinal component' may be; and 'often is; of great relative importance, so that tbe a.ctual motion is. largely in the direction of wave-propagation at these places . e (cf. Fitzgerald's Scientific Writings •.

190Z, p. 418.) . :

, .7. The principle of the dependence of the longitudinal component- in light on the' ratio of the amplitude to the-

Critical Analysis of''JI£axwcll's Processes.'

, . ;, Ma;iocll co~c·ei~~eci.the vibrations 10 be entirely in the wave-front, normal to the axis. of'z', and' thus'whollv dependent 00 x and)._' This is a pure assumption, in acco;· dance with the orthodox theorv but indefensible, as' .. i.S mor .. e I

fully shown hereafter. ' ' , ,

\ e.: ." •

j~'·.··{h{;: 3 :- '--{o'er s e

" Ga'seol

ihtom

. ~Pi': II

~ -~-'. ~.

dopedi •

-points ." ~jlO\\'S .' thesi~, ·'·n:afu:e.

, sumptic in orde Lord fi_ .hj'poth.

Ra)'lcig /~. ,.> :'f_

relucta: fluid; s is esse seems .~'ibrati( conclu

. ·qf a n we mt

t;~;o d: rigidit: second compr' to the -in an:

-thus s

this tl: .an elf rigidit

-displa 'to'the .In op

',' ,with

. ",of lor

·.jn~ gn

i1'aye'len~ht, A = ..1/)., ~, wlll enable us on the one hand .. -Ori page 422 Rayleig/l had. already indicated the ,limito reconcile the views of Poisson,on wave propagation,with rations or 'the elastic-solid theory iv- (;;,;. ':., ::.; '.' :.:< tho;eofFnmd'and CaucJq; .and on the other hand to »fQr these and other reasons, especially :the awkwardcorrect a fundamental defect in the wave-theory of light, ness with .which it lends itself .to ,the. explanation of dis-:

rations, which has stood for nearly a century. '." persio~< ~he .~Iastic-solid}heo.r}~(valu,~.ble 'as apie~e .6fpl:lrely

n Ma.'t_' 8, Thus it 'will be seen that lWa"~llJel!,S' figure above ! dynamical reasoning, and probablynotwithout mathematical

.; 1'.":' given has handed down the defect of lack 'of rotation of I analogy to the truth, can in optics' be 'regarded only ~s' an

the wave elements, whatever be the amplitude,' . and therefore I illustrat'io·n.«· ',' -' ;.i: .•.•• '. ..' . '.:.; ;". .:.'. ..

doe, not represent nature. No wonder that Lord Kelvin and'J ":. In: o~d~r' 't;: ~et'forlh this" difficJlty ~-omewba~ more

others have failed to understand the electromagnetic theory. ; clearly ,~e shallotiiljn~ th~' tr;3.theiniti~a! 'theory Of planewa ves As given. by Jll'!lxlvcll it is contrary to the .profound and in homogeneous elastic solids. Th.e·ne\v:theorY·o·f rr;iig~e'; conscientious researches of Pois,wt,.~\'hich were. critically tisrn, in relation 't6 light; recently' 'developed, 'requires for e'\;l'nined b)' L~place and Fourier, .a nd not .at all authorized comparison a 'definite 'outline -of the theory of plane waves b}' the researches . of. Caltch}' .. With Poincard, Ot.~er-erore, we in a homogerieous elasticjsolid. «It. is only in this way that dIsmiss jl1axl(Jdl's elecrrornagnetic theory .. as ,p'roi:isolre', 'l1Ot we .can decide whether the waves from it magnet are similar deduced Trorn the Ia IVS of nature, b'ut from certain arbitrary , to those of a solid, or are of a somewhat ' different mature, ~ssumpt:ons, and therefore fundat~entaIiY,defecti~-e,., '. I. '.' ;:!he~~1!6~i~g :y~;y'brie'f~)Ut!in'e ::i~J~un~~d ou Lord , 3. The· (:auchy-Frmul Theory of Wl1011y TranS-!' Ke!I.Jlrlss,rtlcle. ElasticIty, EDCY, 1?nt,9thed."p. 824':"5; but

.re .not verse Vibfil.iions dynamically 'Inadmissible, for.a is in accord with the.researches_.ofCai£d~y,' RaNkine, Green,

lfbel' Gaseous:NIediuhl. of High Et as ti c i t y and practically I Lord Ra),leigh, Love, and many othereminent authorities.

p6lai~ ~. In com p r e 5 sib ,r~; ~v h e t'h ~r . Isotr op i c ? r'; A (i'ol,o t r op i C. ! (i) . Definitio~s:'Let [h~'~edan'g'ul1iraxies 'OX;'OY, OZ

earch~; ';'. In. his celebrated article on tbeWa\'e~TheofY, Ene)': : be so oriented thatb~'ls perpendicularfo the. \~'atefr06t:

to the, clopedia Britannica, 9'h ed., the .Iate Lord R;nleigll often : and 0 y, OZ in theplan'eor the wa\ie front.' Then if a,

. '~;-i::: points out the . weakness . of the \~::ive~the6ry' of 'light, and ; /3,. r be the idisplacements ' of a particle' of-the solid, whose ,sllows .that although we may adopt it as"'a working hypo- undisturbed coordinates are .(x, y, z) we have for any time thesis, we are not to trust the theory as a representation of the disturbed .coordinates .x+a, y-+P,·Z+l,.Ac<;ordingly iiatu~e.Thus'on pp. 422-445-446, 'he points out Grem's'as- thedisplacement~.a~ p" are_'runcttons of x'and I,"'and this sumption that the longitudinal component has infinite' velocity, , is' the definition·.of wave nlOtion.· , ' .. ' .. ~ '; it! order to "get rid of this difficulty; but it is' evident .that i There 'is therefore a simple longitudinal strain .~ in,the Lord Ra}'leighregarded this procedure as a somewhat .violent : direction" of.OX, and 'twodiff=rential slips, ,~parallel 'to hypothesis; scarcely justified 'by 'any -known phenomenon. lOY, 'and' paraileito OZ, ;;"hichare·si0'pledi5t~Jrtiop.s,

Jr.a_deigh "says:" -. .... in the shear of planes of the material - orie 'over the' other,

~The"idea :of'traris\'erse' vibrations wasadrnitted with; The values are ., ..... '. -; ..•..

reluctance, even by Young and Frcsndthemselyes. A perfect: I 1'- , ,_ . 1'"' ---'~./' 1 (20'.)

~'= da/dx '1) = 1 2 ·d;ldx· !> = V 2 . drl dx ,

'fluid, such a~ the ethereal medium W;IS then supposed to be, : . . - " \ -" "-

is essentially incapable of transverse vibrations, .But '. there i (ii) Calculation of .thelYod" -·d;ne"to.:p~o_duce strain."

seems to benc re~son a priori. fi)f preferring' 'one ,kind' of i If W denote the.work pe,r .. unit volume required to

~i~ratio~ to another: .and ,~he:rh,et'l,?me,n~ of polarizationprove i produce this !'train,th'e 'stress q badri(:_' becomes:' conclusively ..tJiat,' if luminous vibrations, are analogous to those i IV = 1/2 (1 . .:i2--r: B.~2+)::.f~ + \1j'li ~'-+:c~j; t ~ +'~F"?,,) .. (2 I) .ora mater ial medium, 'it i~ to solids,a,nd ~ot ·.to fluids; ,th,a! i which isan'ellip:;oidal surface, A; 13;· C,' lJ,E, P being 'We m~st Jook:.~,An isotropic solid is ;'ppible 9f propagatiri? ! modui~ses'ofelasti'titv'o(the s6Hd: '0 =:: ... :./ .: '.

·two distinct kinds of waves - the first dependent upon, . --' I~' .' -; :; . .'b·· ·:'h : h ,.;.' "" ..• "f' .: 'h" C", •••.•. ,,-.

. . . ' ... ': ... ,' _ - .,' '.. I /) ,q, ret e tree components. 0 t e ·tracoon

rigidity, ,,?r,' the /o;c: by ~'?Ich: sbe;lf_ \s~esl~ted, at.;d ~h~ i 'per unit'area' of 'the ~~a~'e' front, ' ':"'e shall .'113 ve" the"jinear' 'Second a~~l?gous to waves of sound. an~_, ~~pendent;_u~,on i equations connecting the strain and ~lip,;.'''~ith':the'''modu~

compress.lblh:l" In th.e. forn:e,r the v~bra!tons are transverse i luses of el;sticity';: .' ;c,.i· i'ee , . ,! .',' .. ,.\

t.iO . the . dl~ectt?n :.ofp:ofagatlOo, tha,t 1;5; :;they :nay take ~lace '1, ; ; .. ,:' " '~.' '.';;""'1 = ·.A ~.+'F" 4-'E~. '., .. "0" ;i::.,': '.:

n any direction parallel to the wave front, and they are :~'.: ,':: "'V1/· "':'::::":'F~ B ;"'-D"",' .,,' . (.)

thussuitlibJe representatives of the vibrations of light: In' qr .2 ~ . -;,.+:' ~' .. : .' ~.:_. '. ':::::."::: ,~~~

this theory the 'luuiiniferouserher is distinctly assimilated to I rli 1/2 = EI;-+'v 'i-r;-,C~:,.· , ". '.

,a,n. elasti~sol~d,'an? the,velocity'()f light depends upo~ .the ! Now let it be further,assumed £,1),i -fulfill I[~e~r ~~la:i~n~, ". ~giditfarid ':~e'n:sity' assigned to: -the'<medium. 1':';'. :~: :;; :. ,._ I with . .tpe,~:od~Iusd,_.Qf __ e_!asticlty )11 "the three' dire~t~ons: '

.. "-:', ! .... ~?e·:'p.oss~b~lity"'.~~ 10n,g1tu~iriaJ,:waves, .. in .w~i~h ~he ,., <r:«: ,.- -""'. :'jj~·'·"·A'f+'j'1/-{E(', .. :"; <,~,_.~,>~

"," ~tsplacemen~ IS perpendicular .to·-the wave-front, IS an objection .' ." ,.. . ..__. ", . '"-.''' .. '.> ~_, ••• ,.",L:.

~o' the ~lastic-solid the.6if·of Iight," for there-iS. nothi~g ~kn'o\~n ','.'! i, I~ .;:c:";Jf>~.:. ;:.f:~:,~ ~ -+ -e.: ,,',:~ . .... (2.,~),;

·m OptICS corresponding thereto. If, however, we suppose . 0/.", ~,./!c~+./21j + ,e;, ~ r- .v'. .' .... : "" ..

, ,1I'ithGncn'.t hat the .medi urn ~s .incomp ressible, the VI: loci ty . ': : :;-" The l~~pl ~!rg.l;ie~enl~ina ntal S':. b~: Si ves, ,~pre~}~;e<.:l posi- ,

·?f longitUdinal wavesbecoinesJ.rifinite;:·and~he. objection -is ti1:'e)valu~s :fo:£' (#".;~!llcb!de?n.~ ,th_e:"yays}o.,~qicll,\1i7 ~oJi~1 .:' . . Jln great degr,ee. .obviated.s r ,~.':';i, 1;";'1 ~ . 7,;,('~"~; ,;.,'~:' ';:: may' be strained, :If we substitut~;:;~n}:. 'PB~ ~l!~h~~~~d':~J,~n: '

5

397

5085

he cnn! )t jusr]. .

• aether jegrel:'s

..

..

1:;:":

C,.;.

•. c_., _;, •. "'·-c .. :_.: .;:, •. _ .. , .,,~ -.. :,/;", .. II::. _""::"" """,=<c _::,,,,-~= .. ,<,-~= .. =. =,,:~7~7:~::'''::';"'''.''- --,._ .. ~~"'11111!111~~'"

,', ,i.,' .v..'.

, ;..,'

'.

" ;

'r~~~:3l~'~~1~'~;~~~~ri,;tl~'~,.:f~HP's;~"fj~y.~:i·K~{~n~: th_~:: ~~~)p~~"i

_l]i "';,M_:Jixjdfc::""q:.,'M.dP!dX':>"r" ,.Af·drldi"~ '-(2 4) ":-' .• i:,.- "''Th((''ihree' -tOn-;:pb'D~DtS:"o('the! '~:hol~ ·f~r.ce' dtie'to"tlie t'iac~r6n i' c.:( die' sides o~ ~n i'l)finhesi~iti pit-raIlelopiped' ox 0)' o~

"of the' solid""obvi,ously are: 'l ;-:' :,:', :" c, «', _." ,:' '_CO',,:

(,~ ~ ,"'I:" "Iir- . ~~ ; ...... 1-- f";';'''''! 'I· .... y~ '''1 v.r- .... ,~;. ·cF 11~1"jo ,~!." ~ •... , ~. ~

>' __ d)/dx', Ox O)~ os';'- dq!dx' oi oj oz"" 'd,}dx_: of 0)' ~z ~ ", (~5)

~9~,tP¥s~,c~.ipp,9nenl_l?i~~s,a~~\jri ~;,'q~i(ibii,~~" ~,i,th-' the

mass e In _ th~" salI!e element of sp!l~e;ll:nr;t__ ~e:.nce ,'we have

theresultingequatlons: ' , ", . ,L '- ,.,,' _.'

,,' :' " d?a/dtt:e&xO)!OZ ~ dp/dx·oxoy'oz: ' . j

; ..•. z ':.,. d2!J/di2• ft Ox oy oz = dq!dj;: Ox oy Oz .. :C";'" ,,(26) ,

:1; "·,d2r/dt!· (lOX d) o~ . dr/dx· ox oy 0;;, «-»: ,

"", :;~ (iii), Equ'a!ionsof 'motion for waves in an elasticsolid, :.';, .- 'Witho'ui'rega'nf!o the space ;oithe'eJ~~ent; therefore, !h'e' equations of motion are: ," ".:' -.: -' - " .' ", cii/dx' .......:. "(I' d ~a/dt2. dqld~- ' ft-· d ~jJ/dt2

"'".. ,.,,' ~ - . d~ldx' = e: d2r/d,t2 ~ '.;;, .'

Substituting ,the',values .of ~, "Ii /; from (;~): iU'(Z3)

and iDi_egr-ati~g in respect' to, x, we 'get, •

.. '_Aa+(,f'!J+Er)V2 .' ~~fa· _

. Fa+(BP+Dr)V2 = lIfPV2 Ea+(D!J+Cr)V2 = MrV2~

The thiee"r~ots' of his dererminantaf cu bic may-be called ifl, ~, _,'l1's;' and the' corresponding values of' the- ratios !JI a, r! a, determined by (2 8), may. be denoted by, b1, Cl,

b2, C2, bs' ,Cg ; :".' , .

. Accordingly- the' complete solution of (;7);' subject·t~ - (z8)" becom,es of)he form: '

.~~ ..

: a":_ ai+at+aa

. P = hI !Xl ~b2 a2:ba as

r = cIal+C2a2+(S!XS

. ;; ~al' j;[x+tV(Mlle)l+Fl(.x~tV(.Af;Jf!n . '·'a'2'A[x+tV(M2!e)]+Fz[xc."-,-tV(M2!ft)]

,- ~~ = fdx+tV(Ms/(I)J +Fs[x_:_ /V(~1-f3Ift)J . '.

.' " (i v) Th;~e different wa ve vel6~i ti es i~ferred.· -

I~ the ~bove equations /1, A, h ,F.., fl2, Fa ~r~ arbitrary fun~tion,s. Owing to the form. of these expressionsTt is therefore. inferred that there are' three different wave

'velocities, namely:. . "

'1'1 = V(Mde) V2,= V(~l(!) Vs =v{¥S!(!) bo}

and three different kinds of waves, determined by (28), and depending on the _ aeolotropic character of the solid. The waves are therefore very complex, but- are m-uch simplified

in an isotropic medium. -,

, ' Simple- case of waves in an isotropic solid.

- Letth~ solid be isotropic; and then themoduluses of

elasticity reduce to the Form: . ,

B=(;,'D"E'P=o

'., . Accordingly, the' above three different kinds of waves with three different velocities now reduce to just two: Compressional or Longitudinal.

" . tt: ~;:,;( c~~p~~~s;onal:~a\;ei' ~ike: tha't: of~o1Jnd . i~~'

O'r' other,' elastic fluid! with the motion normal 10 the fronf:> This" corresponds to the conclusion reached -by in ,.liisce1ebrated' memoir of I 830;' and holds- fofimy med i urn .,' ~ '" : ~:' , .. .- '(\-,',: .. ~ ~I~~ ).~, :( ~ ~>-(-,., ~-,; ~ :('-.: ;;/, :~:r r', .

., 2 .• ·· A .. tnins"e~se".wave, wi~h.~ tJ:!E; motion parallel to wave fr,<mt:) ,This:, wa\'e:.dep~nq_!!-oQ.,·.th~ a_ssllm.e,~ , of.an. elastic solid, which .. resists she~ringl_l!()~jgn, _as.

one, Jayer:_sJioes. oyer _a.nother,J~' , _"f;:" ': .:"" ~, ' ····(v) 'Th~ simplest case' ~f~ave-s in

solid, aeolotropic or isotropic." ,,{_, • .''\''he~· the solid isinc0l!lpressibJe,_Gru.rj has ~how~ from' equati_on (2,1) above, that the.' modulus of_elasticifY.'

A = 00; - and hence" the displacement along the' ;;--axis va'! ~ishes,"or"a::: '0'; f' 0': ,the~efo~e. (21) becomes simplf

, _. .. " . W,-''j31j2..t- q?+ 2D~, ~. '" ' .',: (32)\

And the first of (23) vanishes, leaving merely:";

· ,_ - , B"fDr.' , »«, -P1],+'Cr =. Me.: __ ,f, (331.

· _. This restriction of th~ oscillations jo the planed 1] (' gives a deterrninantal quadratic instead of" cubic, yielding tw~ wave velocities and two wave modes. The velocity along the' axis of x is thereby taken to beinfinite and a disappears;

leaving the two velocities: , -' )

1-2 . V(il1'2/e):"VS ., V(·M:~/e). (34i

And in, the case of isotropy, V2• Vg;,is 'in, (3 J), ,~nd i~ and lIfs are principal moduluses, each equal to the modulus of rigidity, - • ,' .. " , ,. " .' "

, '. A_s Lord Krlvin points out" 1Ift-, is a mixed modulus of compressibility and. rigidity. - not. a principal modulus generally, because the distortions by differential motions of planes of particles parallel to the wave front give rise to tangential stresses orthogonal to them, which, do not influenc~. the wave motion .. -

(vi) Conclusion applicable to the elastic medium of the aether gas .

. . This- outline of. the thebry of plane waves in homo-

geneous' elastic solids enables' us to form' a fair idea of the' possible types of motions o"f waves in the' .lether.' When the' motion of the- nether wave is not through ponderable bodies, it is f~ee of most restrictions,' and' follows rectilinear paths: if through ponderable masses;" the action: always follows Fermat's minimum path, defined by Haimlton's . stationary condition, oS dr = o. ,-.., .

Accordingly we learn from the above analysis that most any kind of motion may be transmitted by the waves of an' elastic solid :.- and the question to be discussed is therefore not the type of waves :which may be transmitted, but rather the type of wa ves which actually exist in 'nature, and ha ve therefore to be transmitted by the aetherial medium: .

This is mainly an observational 'qu~s'tjo~, .imd the observations should therefore be extended to the phenomena of magnetism _ and gravitation .as well as to 'those of light

and heat" .],

I. Since the aether is a gas, and therefore compressible, by extremely powerful quick-acting .forces, it follows from the kinetic theory, that even if the propagation of waves

· llioD§ ',"'physi · the £

::~its d the I the ( thesi: tbat

pf th

, than

geneaeolc as sh AN IS di direc

place type Kelv . tudir prop posit

by j impl: like wave as il of 5!

u

idea! schd subs! phys whic of s'

· the

, merr

5085

402

1 ~i~\ wa~{ ?iss~k . lastic'

~_,:i'l~~':

bv means of vibrations wholly transverse to the direction of a ray of light be a geometrical possibility, and Cau,hy' showed, and Airy and Hersc/u] confirmed by independent researches, it is physically inadmissible to assume transverse

displacements, and deny corresponding longitudinal displacements, such as was implied in the theory of Poisson, 1830, and suggested by Fitzgerald's paper on the Longitudinal

Component in Light, 1896. '

2. For such an arbitrary restriction would give the

aether gas anisotropic properties, - symmetrical' as respects the ;:q-plane, but unsymmetrical in respect to the s-axis, along which the light is propagated, - for no assignable physical. reason, except that the light is propagated along I the z·aXIS .

. 3. ;'-nd. this unsy~metrical anisotr.opy wou.ld ~hange I its direction 10 space with the change In the direction of : the ray of light, or the mere rotation of the axis about I the origin of coordinates; and hence we see_that the hypothesis is physically inadmissible. Such a physical doctrine I that the property of the aether change.s with the ~irection I of the ray can no more hold a place In natural philosophy I than can an established ieductio ad absurdum in geometry. I

4. lf we view' the aerher in free space, as homogeneous and isotropic, except as rendered heterogeneous and I aeolotropic at great distances, as of the celestial bodies, - I as shown in the first paper on the New Theory of the Aether, I AN 5044, - we cannot admit that its vibratory motion I is different in different directi ens, and changing with the r

. direct-ion in which the light is allowed to travel. i

5. Therefore if we admit a series of transverse dis- I placements of the aether particles for making waves of the I type imagined by Fresnel, Cauch)', Sir Johll Hcrsche], Air)', , Keivzn and "Jfaxwdl, we must admit also corresponding longi- I'

tudinal displacements of the aether in the direction of wave 4. Geometrical Reasons why th e Vibrations'.

propagation - thus giving rise to. rotations about mean I of Ordinary Light are mainly Transverse .

positions, Or true waves of the type imagined by Poisson.

If we contemplate the hemisphere presented to our

6. Instead of the epecial polarized waves imagined I view by a luminous spherical source of light, such as the by iWa.1.·zodi of the type described in section zabove, and ' sun, it is evident that the waves propagated towards the implying merely' a rectilinear side oscillation of the particles, I observer will cover a surface of area

like that of a stretched cord, we should therefore imagine I wa,:es of the Poisson type, referable to simple harmonic motion I as Illustrated by the modified figure of Airy for .the surface I And in orthogonal projection this area will be reduced by of still water.' one half, and become merely the area of a single great circle

The geometrical conditions are fixed by the equations: I of the sphere A' = IT r2 . \37)

u = acos(zrrt/T+p)- • I The sphere surface seen by us in projection is enor-

(u,laF+(v/b)2+(W/cF = I 1'- h d d d . h b d

v = bcos(z;, tl'c+q) (35) ; rnous y ror e-s ortene an coritracte 10 area at tear er,

( Is = V(tt~+v~+w?)., - I· while at the centre no decrease in apparent area takes place.

W=CCOS2:rt,T-f-r} I

. If therefore the atoms emit waves which are flat in the

. i· It is therefore evident that in adopting Calfchfs II planes of their equators, and a haphazard arrangement of Ideas of vibrations similar to that of a stretched cord, Her- , the atomic planes holds true, as should occur in a nonHhd was misled, and he in turn misled Airy and others - i 'magnetic sphere, it follows that the beam of light emitted

:. SUbstituting a mere geometrical abstraction, and practically a by the sun should have its vibrations so largely peripheral

physical impossibility, for the valid physical theory of Poisson, that, with A/l very small, it will present practically the which makes the vibrations of the aether similar to those appearance of transverse vibrations, - as long taught in the

of sound, but AI). very small. wave-theory of light.

, 8. The result has been a traditional false teaching in I In order to examine into this subject somewhat more

<.thew~l'e-theory of light, as hinted at by .Fi'tzgerald in the I critically we may proceed as follows. Let Fig. 3, Plate 7, .. lUemolr »On the Longitudinal Component of Light,« (The i represent an orthogonal projection of the sun's hemisphere,

Scientific- Writings of Fitzgerald, p. 4'(8), and by Professor Chas. S. Haslings, Encyclopedia Americana, I904, article Light, quoted in section I above, where it is pointed out that the conviction has grown that the wave-theory is in some way wrong,

9. It is obvious that waves of the types imagined by Cauchy and Fresnel could be transmitted by the perfectly elastic aether if they existed - as is correctly held by Herschel and Airy - but the question of fact remains: Do they in general exist?

10. This important question must be answered in the .( negative. For in magnetism we recognize, from Farada)"s rotation of a beam of polarized light, 1345, the rotations of the elements of the aether, the atoms having their equators lying in parallel planes. In common luminous bodies, on the other hand, no such parallelism in the atomic planes can be assumed: indeed this parallelism must be emphatically denied.

eities when

;sible

laWn' :icity ; va:

:np"iy'

Gyt

(33)

'1/(-

twei long ears; '~ . : ~1:.:~

{3;4"i

, " ~ ',~,

¥i

ulus

I I. And as we: cannot have luminous bodies, with the atomic planes ail parallel, as in magnetism; so also we can not imagine these atoms so tilted as to send rays to us only from their combined poles. Hence the wave-theory of light as heretofore taught is physically inadmissible.

12. We must hold that the waves of light in general are flat' in the 'planes of the equators of the atoms, and these planes tilted at ail possible angles, as explained below in Section 4. If the axis of z be in the plane of the equator of the vibrating atom, the oscillation will be of the plane wave type commonly shown. If the axis of z lies in the northern hemisphere of the atom, the approaching waves, as we look at them, will seem to rotate' left handed, in the form of a left handed helix. If the s -axis lies in -the southern

-:~-;t~ I ulus

ulu;

"0o.,c ,}~, -

.nce

---:~fo

. ~ .{~.~"

hemisphere of the atom, the waves received will seem to rotate right handed, like the coils of a right handed helix.

~ , '~.,'

with the centre at 0, and the coordinate axes OX and 0 Y as shown in the diagram, Then, if we subdivide the quadrant of the circle into 20 parts, corresponding at the centre to an '~ngular distance of 4?5 between the small circles about

- that point 0 as a pole, we may plat a curve along the radius OX which will represent a section, of the visible surface of the hemisphere, as if the area were not decreased by the orthogonal projection. The equal distances along the radius OX will represent equal values of the sine of the pol ar distance, e,or equal values of the cosine of the latitude reckoned from the base of the hemisphere here represented by the lower circle.

The' curve may be drawn from a table of natural sines or cosines by taking y proportional to this function, so that the, change wilI make a curve of the kind -indicated in the Fig. 3, Plate 7, which is repeated on both sides of 0, in orderto show to the eye the enormous condensation of surface near the circumference of the projected hemisphere. In fact the double curve on both sides of 0 is a semicircle, drawn about Y as a centre, and thus exceedingly simple.

The coordinates of the curve, to four places of decimals, and the surface integral I for the component of Poissoll's radial wave motion in line of sight, equation (38), are;

Angle e'=-~'~e; 1= ,-=:cose, I I

**00 0.0000 0.0000 0.00308
**

4·,5 0.0785 0.003! 0.00915

9.0 0.1564 0.0123 0.01502

13·5 0.2334 0.0277 0.02052

18.0 0.3090 0.0489 0.0255'

22·5 0.3826 0.0762 0.02985

27.0 0.4540 0.1090 0.03342

31.5 0.5224 o. 1474 0.03634

(: 36.0 0.5878 0.1910 0.03814

40.5 0.6494 0.2397 0.03916

45.0 0.7071 0.2929 0.03906

49·S 0.7603 0.3506 0.03824

54.0 0.8090 0.4 I 2 2 0.03625

58.5 0.8526 0.4776 0.03350

63.0 0.8910 0.5460 0.02978

67·5 0.9238 0.6174 0.02761

72.0 0.95 r I 0.6910 0.02050

76.5 0·97 2 3 0.7666 0.01501

81.0 0.9877 0.8436 0.009 I 4

85·5 0.9969 0.92 IS i= 20 0.00310

90.0 1.0000 1.0000 1: = 0.50238

i= I

I = (sin8.- SiD 8'-11 sin 1/2(8,-+-8'-1) Ion = 0.025119

= 1/40, nearly. Im = 1/ .. 0 A = A/J.·Im. (38)

From these considerations it is evident that if we imagine the atoms in the sphere to have their equatorial. planes directed radially, which will be the average position in a large mass, under haphazard atomic arrangement, the effect will be to give us an enormous preponderance of transverse vibrations near the periphery of our luminous

\.

\

5085

globe.: or in a ray of ordinary light from a globe like th~:!;: sun or a star. This reasoning applies to any luminous. body::i. or flame, such as that from a Bunsen-burner in our la; boratories, which have haphazard arrangement of the atomic" planes, all atoms vibrating so rapidly that from any single atom several hundred waves of the same type will reach the'_ eye. of the observer before the translatory motions of the; luminous atoms will produce appreciable change.

I? his Undulatory Theory of Optics, 1866, pp. ISS-I56, Ai,)' says:

»Cornmon light consists of successive series of ellip.: tical vibrations (i Deluding in .this term plane and circular vibrations), all the vibrations of each series being similar to each other, but the vibrations of one series having no rela. tion to those of another. The number of vibrations in each series must amount to at least several hundreds; but the series must be so short that several hundred series enter the eye in every second of tirne.«

This criterion of Airy obviously is fulfilled by the light from any luminous source, since even in a very small mass the' atoms are numbered by the trillion, and no change in their average orientation occurs with the lapse of time, though indi vidual atoms in their mutual interactions will slowly shift their individual equatorial pJanes to new positions, as the millions of millions of vibrations are emitted.

The centre of the yellow light of the' spectrum has a frequency of 517500000000000 vibrations per second; and thus with such an enormous flow of waves, they might be subdivided _ into ten thousand million successive series and still leave a flood of 51750 groups of waves beating upon the eye in a second. Accordingly, Airfs criterion is perfectly consistent with the motions of the individual atoms, in mutual collisions at the rate of say 10000000000 per second (cf. AN 5044, p. 66), which is about the average for terrestrial gases under laboratory conditions.

Returning now to our figure for illustrating the enormous preponderance of transverse rays in a beam of ordinary light, we easily find by calculation that 62 percent of the light comes from the lone fj = 900 to e ----:- 5 1045' 27", near the -peripher:5' of the orthogonally projected sphere surface. We may even extend this zone inward to e = 44° 25' 30: and still not approach the centre of the circle more than 0.30 of th e radi us; yet this outer zon e to 6' inel udes 7 I. 4 percent of the luminous sphere surface. Thus we see from the corresponding small circles drawn in th-e figure a bout the pole 0, why in ordinary light it may be described as practically transverse - since a great preponderance of the light from the atoms acts as if the vibratory motion were in the plane' of the wave surface.

The great hollowing out of the curve of light near the centre of the figure, from which alone indications of a'

'longitudinal component could be expected to come, and the smallness of the factor A/i., shows why there is such a feeble indication of this longitudinal component in our actual ex' periments. It is not surprising therefore that in his Undu' latory Theory of Optics, J 866, p. 9 I, Sir George Aily says:

»The reader W)lO' has possessed himself fully of this hypothesis, will see at once the connection between all the experiments given above. ~

plan pent

of I 'dieu stitu

is ~:

UDS!.

J~ fc - of il

tions

of v

to s' tion on t peril such

in tl expr-

To , emit-

and, we I the

of a:

mOr< of th only tral the »;

:nic Igle, the the

lip~' ular

r to elaach the the

ight rass

! in ugh' ;hift the

15 a and·

be and pOll per· )ms, -

per, , for

nor' ' )afY

the [lear, 'ace.

. 30f

thap P:4, 'rom bout .. :I as

5085

- »For the general explanation of these experiments, and I the very conclusion announced by Airy, in the above passage, for the accurate investigation of most of the phenomena to that there is not the smallest trace of visible fringes of interbe hereafter described, it is indifferent whether we suppose I ference due to the longitudinal component, which of COurse the vibrations constituting pol ariz ed light to take place parallel I has to come from the light near the centre of the canopy. _ 10 the plane of polarization, or perpendicular to it. There I Airy personally repeated the experiments which he described are reasons, however, connected with the most profound in- and reduced to mathematical expression: so that his con~'estigations in to the nature, of crystalline separation and into elusions have been widely accepted by natural philosophers, the nature of reflection from glass, etc., and confirming each It is by virtue of Airfs careful experimentation and othet: in a remarkable-degree, that incline us to choose the! analysis of the wave-theory of light, following the indepcnlatter: and thus:« dent and profound analysis of Sir John Herschel, in the great

"When we say that light is polarized in a particular treatise on Light, Encycl. Metropol., 1849, that we adopt plane, we mean that the 'vibration of every particle 15 _ per- Airy's presentation of the subject as authoritative. Our con-

pendicuJar to tbat plane.« elusions therefore are as follows:

=Thus, in the undulation constituting the ordinary ray 1. About 71.4 percent of the sphere surface is in-

of Iceland spar, the vibration of every particle is per pen- eluded within the elevation of 4 S° 34' 30# from the base of dicular to the principal plane of the crystal: in that con- the hemisphere. 'This part of the sphere is a zone so near stituting the extraordinary ray, the vibration of every particle the circumference as to appear to the observer to be essenis parallel to the principal plane. Wlren light falls upon , tially peripheral. Hence the origin of the belief, in view of unsilvered glass at the polarizing angle, the reflected wave I the smallness of the ratio A/)., that the vibrations are actually is formed entirely by vibrations perpendicular to the plane ! transverse, and the integral for the longitudinal component

of incidence: the transmitted wave is formed by sornevibra- f insensible to the experimenter. .

lions perpendicular to the plane of incidence, '~ith an ~xcess ,i ;L Light vibrations comirig from this periphery would

of vibrations parallel to the plane of incidence. «, appear essentially as transverse waves; and by proper optical

»The reader will perceive that it is absolutely necessary ,I appliances could be polarized into right handed, left handed, to suppose, either that there are no vibrations in, the direc- I circularly polarized or elliptically polarized light, as seen in lion of the wave's motion, or that they make no impression I that transmitted through crystals,

on the eye. For if there were such, there ought in the ex- I 3. As only 28.6 percent of the sphere surface remains perirnent of (98) to be visible fringes of interferences: of i in the larger zone, near the pole, and a 'considerable part such however there is not the smallest trace.« ! of the vibrations on that polar surface could be resolved

If we examine the figure, we find from the integral Ii likewise into circularly or elliptically polarized light, we see in the plane xJ', that the' total light emitted is given by the that in ordinary light, the average vibration is described as

expression x , made up of elliptical. vibrations (Air)', Undulatory Theory

L=Sydx (39) I of Optics, 1866, p. 156),

o _ 4. In discussing experiments leading up to Lll))'d's

To derive a corr esponding expression for the Poisson waves !. observations on conical refraction, Airy notes, in regard to

emitted radially from the sphere surface, we put i' polarization. of light, that » if common light be incident,

x=sinB Y'= I-cos8 dx= cost/dB. (40) j (which not improbably cODsists.ofsuccessive series of waves

And we integrate for (J between the limits 0 and 1!2JT, polarized in every conceivable plane) rays will be formed and, for the surface generated by revolving the axis of x, I directed t.o e:ery point of t.he [~ewton'sl~in!S, each ray bav.ing we use w between the limits 0 and 2 IT. Thus we have as I the polarization proper to Its POIDt of the nng; and a conical the surface integral of the hemisphere I sheet of light will be formed within the crystal « [Undula-

I tory Theory of Optics, p. 106). Again, summarizing the '/?'IT. l"

- L' = S S( I - cos 8) cos 8 dB dw = ZIT. (4 t) description of ordinary polarization, Airy draws three con-

I elusions: (I) »If from common light we produce, by any

o . 0 ' known contrivance,' light that is polarized in one plane,

_ To find the light in a beam we calculate the reduction there is always produced at tbe same time light more or

of area by orthogonal projection. 'less polarized "in the plane perpendicular to the former»

- If now we integrate for the light distributed over a ( ) .

( ) p- 89 .

more limited surface S =.1 8, w, we shall find the value S. On this first conclusion Air)' comments as follows:

of the integral so trifling, that till t/ = 44Q 25' 30d, x= 0.7,/' » The first leads at once to the presumption that polarization only 28.6 percent of the light will be included in the cen- , is not a modification or change of common light, but a tral canopy. Moreover the average factor for the part of I' resolution of it into two parts equally related to planes at ~je POlSSOn radia.1 wa;·.e ~otio~5 in ~h~ li~e of s~ght is only right angles to each other; and that the exhibition of a beam

4Q, and the ratio AI J. - 10 ,makIng) I of polarized light requires the action of some peculiar forces

, A = A/}.,(! = 1/(4'10')), (42) (either those employed in producing ordinary reflection and

Accordingly one would expect experimenters to reach refraction Or those which produce crystalline double refrac-

, 1

I

I

! ') The calculations made Sept. I-~, 19~o, as given in the note to section 8, below, make A/).. = I : 1660508000, which would make

, = I : (66,1-20- 1 o~) - a value hopelessly beyond the range of observation. - Note added, Sept. 12" 1920. ,

6. In view of the considerations here deduced by Airy, 10. If the considerations on the spherical distribution

we see why the spherical distribution of waves from atoms of the planes of the fiat wavelets a boved educed be valid, in every conceivable plane will give rays directed to every Airy': results could be true, and yet give us an unlimited point of the circumference of the end of a beam of light; number of component flat wavelets not originally normal just as in Airy's discussion of the polarization in Newton's- to the direction of the wave' propagation, but inclined to it rings, it is held that the waves »are polarized in every con- by the angle E, as in the electr odynarn ic formula' of Frans ceivable plane «, and »rays will be formed directed to every iVcu17laJl71J J 345:

M = II'S S 1/1" COSE ds ds' = ~ X l.J S(I/r)(dx/ds' dx'/ds'-+-d}/ds' dJ"/'ds' +dz/'d's. dz'/ds')ds ds' (43)

cos E = cos « cos a' -+- cos /J cos,8' -+- cos yeas y' 'cos (;; 1') (44)

and yielding the general formula for electrodynamic action in universal gravitation, or Ampere's theory of elementary

electric currents about ·the atoms: ...

.Q = S SS S S S I I' cost£' 1') [(x~x'F-+-()I-:j/)2-+-(~ ~:;;')2J-!I' dx d)' dz dx' dy' dz' . (45)

I S· Other Fundamental Objections to Frrsml's I Theory th~t Lig,ht Waves are purely Linear Trans' f .v e r s e Mo t i o n s,

I, (il Certain circumstances favorable to the old wavetheory of light permitted it to progress but did not establish I it on a permanent basis.

' ..

lion) which will enable the eye to perceive one of these ,parts without mixture of the other. This presumption is

strongly supported by the phenomena of partially polarized ,light; If light falls upon a plate of glass inclined to the ray, the transmitted light, as we have seen, is partially pola ri zed, If n ow· a secon d plate of g J ass' be placed in the path of the transmitted light, inclined 'at the same angle as the former plate. but with its plane of reflection at right angles to that' of the former plate, the light which emerges from' it has lost every trace of polarization; whether it be examined only with the analyzing plate B, or by the interposition' of a plate of crystal in the manner to be explained hereafter (i4 5l· This seems explicable only on the suppo- I sition that the effect of the first plate of glass was to, diminish that part of, the light which has respect to ·olle II plane (without totally removing it), and that the effect of the second plate is to di minish . in the same proportion that part of the light which has respect to the other plane; and therefore that, after emergence from the second plate, the two portions of light have, the same proportion as before. On considering this presumption in conjunction with the second and third conclusion, we easily arrive at this simple I ~·p..9thesis explaining the whole s : I

»Cornmon light consists of undulations in which the I vibrations of each panicle are in the plane perpendicular I to the direction of the wave's motion. The polarization of I light is" the resolution of the vibrations of each particle into I two, one parallel to a given plane passing through the direc-I tion of the wave's motion, and the other perpendicular to that plane; which (from causes that ;ve shall not allude to at present), become in certain cases the origin of waves that travel in different directions, When we are able to separate

, one of these 'from the other, we say that the light of each is polarized. When the resolved vibration parallel to the plane is preserved unaltered, and that perpendicular to the plane is diminished in a given ratio (or vice versa), and not separated from it, we· say that the light is partially pola-

rized.« '

Thus the difficulties of reconciling the wave-theory of light with the electrodynamic theory disappear. The resolved waves in polarized light are largely normal to the direction of propagation, but their original component flat wavelets were not, being in atomic planes inclined at all angles.

point of the ring, each ray having the polarization proper~ to its point of the rin~,«'

i. To view this reasoning "graphi cally, imagine a series; of planes drawn through the centre of the sphere and fixed " at eq ual intervals normal to a meridian of the circumference' having its pole in the observer's eye. Then imagine the, whole set of fixed planes rotated about the pole through· the observer's eye" and stopped at· successive intervals of the circumference equal to those between the fixed planes~ The equatorial portions of the hemisphere will thereby be divided into rectangular compartments with areas equal to 1'2 cos J. di. dw, were w is the angle about the pole, and j, is the latitude. To get compartments of equal areas in higher latitudes, the revolving system must stop at intervals equal to dm/cos}, = (:ZiT/lIlsecL From these considerations we percei ve that in higher latitudes the number of rectangular compartments decreases rapidly;, and if the number of fiat wa velets . of light are proportional to the rectangular areas on the sphere, the wave disturbance in light will be almost wholly peripheral, or transverse.

8. Small as is the amount of light depending on the vibrations in or near the line of vision, our sphere shows that ~he central great circles distributed in haphazard fashion, do not lie in the line of vision, but pass around it on all sides; and hence we perceive that the disturbance necessarily is rotational in character, and nearly transverse to the direction of propagation.

9. From considerations based on polarisation, ~ tending to sho-w that in the ordinary ray of Iceland spar the vibration of every particle is perpendicular to the principal plane of the crystal, while in that constituting the extraordinary ray, the vibration of every particle is parallel to the principal plane ~ the polarized iight in hoth cases being already systematically resolved by the action of the crystal - Ai7)' concludes in article ,01 of his Undulatory Theory of Optics, that there is not the smallest trace of visible fringes of interferences.

that Cell Can,

tions '·'~O\lgj . heca1 ·,ue f

is ot

norm be p expla to tJ": the: perti: what

form of p: earn 181 { exarr was, but .cond :trav€ vibrs there wave

the I left case this offer

Whi Fres was -outli ding theo

assu neve base

'.'

409

1. In his memoir of 1830 Poisson showed that in ~lastic media waves propagated from a centre are essentially like sound waves, and at great distances the molecules move mainly in the direction of the normal to the wave front. But poisson died in 1839, while Cauchy lived on till 1857; aod moreover the deceptive argument drawn from the vibrations of an' elastic cord misled, Herschel and Airy, who failed to perceive that the underlying premise implies anisotropy in the medium.

2. As Poisson never concurred in the theory of vibrations normal to the direction of the ray, Fresnel and Arago

. sought comfort -in the analytical results. of CauellY, And because such waves are theoretically possible when once they, are generated, it was inferred that light has such motion as is observed in the vibrating cord.

3. Cauchy's analysis seems to have proved that if waves normal to the direction' of propagation be started, they could be propagated by such transverse motion; yet he did not explain how they would arise, or would be started normal to the direction of propagation. Nor did his associates see the anachronism implied in a medium with anisotropic pro'perties along x, )" s, -z being in the direction of the ray, whatever that may be.

4, After a visit from Arago, 18 I 6, YoulIg began to form a theory of waves with motions normal to the direction of propagation. They were herd to be similar to undulations carried along a stretched cord, as stated in a letter April 29, 1818, (cf. Ff/hittaker's History of the Aether, p. r22), This example of the vibrating cord gave a physical analogy which was afterwards adopted by Fresnel, rIerschd, Airy and others, but it was really an anachronism; for it implied a >stringy« condition in the aether, in any direction the wave might travel, but not in other directions. The s-component of the vibration along the ray vanished, which made r = 0, and therefore s = V(;-2+'1j2) becomes confined wholly to the wa ~'e surface.

S. As we have seen above, Green took the velocity of the longitudinal component of the waves to be, infinite; which left the finite motion wholly in tbe wave surface. In the -case of a gaseous aether of course there is no authority for this procedure; and thus it simply begged the question, by ,offering an arbitrary hypothesis.

6. Hamilton's prediction of conical refraction [conf.

Whittaker's History of the Aether, p. J 3 I) only showed .Frwrds ideas of the theory to be correct in general, but was not an accurate test in all details. The theory above outlined will explain conical refraction equally well. Accordingly in the absence of definite objections, the old wavetheory triumphed by default, 'at least temporarily; yet the' assumptions made to get rid of the longitudinal component ne\'er were satisfactory, and could not be justified, because

based 00 ~n arbitrary hypothesis. '

7. The physical inadmissibility of Grem's postulate that the longitudinal' component has infinite velocity (Grem's COllected Papers,p, 246) is easily shown by the following ,consi derations:

a) In his work on Sound, Chap. V, T)'7Iaall shows that -when the bow of a violin is given. a stroke along the violin

tion did, ited rna) a it

,.,

'an:

')~: .

ries'

xed

nee

the

ugh

of

aes,

be

I to

). is

.her

lilal

We

ular

flat

reas

cost

the

ow,

ion,

all

ssa-

the ten' the ipal .tra-

I to' ~ing 'stal .ory Iges

.i, .

tury

•

5085

410

string a shrill sound arises, owing to the rapidity of the wave along the string, - glymg high pitch to the sound. Owing to its higher elasticity, waves travel say ten times more rapidly along the string than through the air. '

b ) Now it is easy to see that this is analogous to Green's unauthorized procedure, which amounts to assuming a »stringy« condition of the aether in any direction in which light is sent. And the chance that the assumed longitudinal component would not manifest itself in some way is very slight, sin'ce the aether,' with excessively small density, is naturally taken to be a gas, and the velocity of the aetheron ii . 47 IOOO kms .

8. Again, in his work on Sound, (p. 73), Tyndall shows that a sharpness of shock, or rapidity of vibration, is necessary for producing sonorous waves in air. »It is still more necessary in hydrogen, because the greater mobility of this light gas tends to prevent the formation of condensations and rarefactions.«

Therefore the aether should present enormous difficulties in the generation of waves therein, and such is observed to be a fact. By way of experiment Prof. lVipher alone has generated disturbances in the aether ; and to produce them he had to use dynamite, which gives intense forces quickly exerted. Observation thus verifies the high velocity, of the aetheron, and will not permit us to assume different velocities of the aether wave in different directions.

(ii) Purely transverse vibrations in light would imply only transverse undulations in magnetism and electrodynamic action, which is contrary to observation,

The theorv of transverse waves was first admitted somewhat reluct~ntly- by Young' and "Fresnel in the early part of the T9th century, (1802-.1829)' 'But under the celebrated

I experiments on diffraction, double refraction, polarization and 'interference conducted by Fresnel and Arago, the theory , became a new means of discovery. This apparent experi-

mental triumph of the undulatory theory aroused such interest that a long series of brilliant mathematical researches were entered upon by the eminent natural philosophers then resident at Paris - Naoier, Poisson, Cauchy and Lame.

It is true that these mathematicians were by no means agreed among themselves as to the details, yet their work was mathematically so impressive' that it created great interest in other countries, more especially in England, and was adopted by Airy, Hamilton and Hi.'rsclul, and subsequently by Green, TJIONlson (Lord Kelvin), Strikes, ,:Jfa::C1oell, and Rayleigh. In this way the undulatory theory as now taught came into wide. use; and yet it was always suspected to be somewhat defective, and we shall now point out some additional reasons why the traditional view can not be valid.

I. The theory of purely transverse waves in light is directly inconsistent with the rotations actually known in magnetism, and with the electrodynamic action of a current on a magnetic needle, in such phenomena as Oersted's experiment of 1819.

2. For 'if the Illation of the aether is linear and transverse in light, it would be logical to conclude that it must be of the same type in the waves by which electrodynamic \tion is propagated across space. Indeed, experiment proves

\ -,

','

•

'-

41 I

that both actions have the' same velocity, and take place ·in the. same .medium. And we have no grounds for assuming a difference of wave type ..

J. Yet we know by actual observation that in Oersted'e experiment of 1819 the magnetic needle Dot only is directed in a definite way, depending on the direction of the current, but - also attracted 'to the conductor by the action of electrodynamic waves propagated from the wire, as first pointed out by the present writer in 19 J 4.

4. Now the electrodynamic waves discovered in 1914 can not be wholly transverse, as held by Fresnel and his followers in the wave-theory of light; for in that case there could be no actual attraction on the needle. On the contrary, Maxwell held (Treatise on Electricity and Magnetism, 3d ed., § 793) that such transverse waves exert a slight repulsion, and on thepremise employed it is difficult to refute his conclusion.

5. In order to exert the observed attraction, the electrodynamic waves must have rotations somewhat like those observed in water waves; and the needle must so orient itself that' the elementary Ampere-currents of electricity about the atoms coincide in direction with those in the electrodynamic waves propagated from the wire.

6_ The observed attraction of the magnetic needle to the wire therefore is inconsistent with Fresnel's doctrine of purely transverse waves, as taught in the theory of light and adopted by Maxwell in his electromagnetic theory. Now magnets themselves have circulation of currents about their atoms, as first shown by Ampere's experiments with currents in 1822; and these currents about the atoms give rise to the rotations about the Faraday-lines of force, thus forming the wa ves propagated outward from magnets. It is only in this way that we can imagine how magnets presenting unlike poles attract; and, when like poles are presented, repel, by a mechanism at last disclosed to our vision.

7, Therefore the magnetic needle is attracted to a cond ucting wire - by the electrcd ynami c w a yes propagated outwardly from it; and magnets themselves also attract by sending out waves defined by the well known rotations about the Faraday-lines of force, Accordingly it follows that all such waves must necessarily involve rotations in the aether to make up the waves: and the waves incontestably are not wholly transverse, but only transverse in somewhat the same way that water. waves are transverse.

8. Tile Fresnel theory of purely transv~r5e light waves thus again is definitely disproved, and we may reconcile the varied mathematical researches of Poisson, Cal((h~v, and Lame, It should be noticed, however, that Cauchy's reasoning had DO physical basis, to control the legitimacy of the hypotheses underlying it, except the artificial analogy with the vibrating cord. Poisson and Lame on the other hand never were fully convinced that the motion in light is wholly transverse. The theory outlined ill section 4 above probably had never occurred to them.

9· Accordingly there are real weaknesses in the traditional wave-theory of light; and the difficulties noticed by the earlier investigators have never been satisfactorily overcome. The objections here pointed out appear to be new, and absolutely fatal to the theory of wholly transverse waves

5085

412

as held by Fresnel. He was essentially a specialist in light:' rather than a mathematician and all around natural philo> sopher, like Poisson, who never did believe that in nature-: the aetherial- vibrations could be as Fresnel imagined. The temporary scepticism of the illustrious Poisson is now verified_ from a new point of "jew, ,after the lapse of nearly a century •...

r o, It is remarkable that such a palpable penersion: as the theory of wholly transverse vibrations gained currencyin science through the misdirected reasoning of the followers of Cauchy. They seem to have been misled by beautifUl general formulae, valid enough as applied to wave motion in crystalline media, but utterly deceptive as a'pplied to the'

. simple case of the aether -itself, viewed in free space as a uniform isotropic medium, which furnishes the general basis for the undulatory theory of light. This outcome is the mOre remarkable and unfortunate, since Poisson was a greater and more sagacious physical philosopher than Cauchy, who was chiefly a p,ure mathematician.

(iii) Difficulties in the wave theory of light as outlined by Prof. Cilas. S_ Hastings,

In a letter to the present writer, dated Aug. 17, J 9l6,

Prof. If'astlllgs speaks as follows; I

» That light vibrations necessarily are transverse only

is proved in many ways - perhaps most obv iously by the fact that complete polarization is possible.«.,

»If light waves_fall normally on a refracting surface, any free element of volume in the first medium is sustained ~ in permanent transverse vibrations of definite period, - but if

it is attached to an element of the second medium as at the interface, the second medium having either a greater density or a greater rigidity, it will not (although necessarily retaining the same period) move so far from its position of equilibrium, Jllst at this region, therefore, so far as the first medium is concerned, we must add a system of waves of opposite phase and of an amplitude easily calculable from the ratio of light velocities in the two media - this con-. sti tutes the reflected li g h t.«

»Now consider the refracted light. The element of volume just below the interface has the .sarne period and amplitude as the attached element above; it is therefore a portion of a system of waves propagated in the same direction as the incident waves but with a velocity determined very simply by the density and rigidity_('

{/)=

free ~

'-ease infini! the f~ there '~ith

which the p wave fourth

repre:

repre: origir.

the r:

»It is when we consider oblique incidences that we get into difficulties, F1'tS7/el assumed that the same condition held in these cases also, but, as you can readily see from the diagram,

there could exist a stable

"

, state of vibrations at the

Fig,4_ Professor HaJting's diagram of the surface" only when there

path of light at the interface, is a system of compres-

sional waves also proceeding from the interface in a direction t and of an intensity easily calculable if the ratio of volumeelasticity to rigidity is known. XOIV no such system of

This In th

and \

and I

.~.j~~,

, -,'·1.'.>"1'

~'~'>~~<.

.ighf;;:'.,:

'hiler,',' lture" The'

:ified tury.

rsion

::~~i It~ful ,I

)hon

) the' as a

Dasis

nors and was

ined

916,' f

{,

:~:I

ined lit 'if .s at ~ater £trily , n 'or" first saf

rom

con-

tof and :e} irec- "

i~~((

con- , Jees' , liffi.' , mea' tlOn , IlS(i,,~ dilj

-am, lble "

1_ .. ,

.

longitudinal waves exists under any circumstances, because I measurements (the first made since Hu)'ghmst) but failed in the energy carried by the .refiected and refracted wave systems attaining 'adequate precision. Fina 1Iy I demonstrated (Amer. tAken together always equals the energy carried to the re- Jour. Sci. somewhere) that Huyghens' construction is certainly fracting surface by tbe incident wave system. (This, by the accurate to I' IO-6.C

way, is the direct answer to your principal question. I might \,» More recently Kelvin, who was especially desirous of stop here but the fixed habit of an old teacher leads me getting a defensible elastic-solid theory of light, proposed a to add: -) In order to get rid of the obvious difficulty zero _ volume-elasticity, or a collapsible ether. This gives Ereirltl assumed that the volume-elasticity of the ether, both zero velocity for compressional waves and hence no energy free and associated with matter, is infinitely great, in which is carried away from the interface. Kelvin apparently left elise the velocity of the longitudinal wave system would be his readers to imagine an outer boundary condition which infinite and it would carry no energy with it. Aside from would .prevent the ether-universe from collapsing.« the fact that absolute incompressibility is difficult to conceive there are other serious difficulties in the theory connected with the phenomena of double refraction,«

»Slokes is said to have invented an elastic-solid theory, which, however, carried with it as a necessary consequence the proof that Hltyghms' construction of tbe extraordinary wave surface in Iceland spar is slightly erroneous, say in the fourth decimal. Fitzgerald attempted to test this by accurate

As we have [Memoir of 1830, pagatioo of waves

seen, in the third paper, Poisson reduces p. 556) the sextuple integral for the proto the double integral:

6. Outline of the General Theory of the Waves from any Body, whether due to Light, Magnetism, Electrodynamic Action or Universal Gravitation.

(i) Res-ults of Poisson's analysis for wave motion.

It ~lt

CD =]I!47l)S S F(x+at cosd, )'+atsind sinw, z+at sind cosoo)tsint1 dt1 dw

o 0 7t"~"

+(r/41T) (d/dt)S Sn(x--+--atcos8, y+atsinCl sin 00, z+a/sinCl cosoo)tsind dS doo.

o 0

NOll,' the' equation

Ix+tn)'+nz=o (4i)

represents a plane through the origin. And

I x+my+n z - (al+s) = 0 (48)

represents a plane with perpendicular p = (at+s) from the origin.

If plane waves proceed from the equator of an atom, the radius of the spherical wave surface about the atom will

II increase with at; and the disturbance, in the plane of the flat wave, in the equator of the atom, will travel away with

/ the velocity at, and remain parallel to the original in all 'parallel planes. Thus Ix+: m y+n z - (at+ s) = 0 represents the disturbance in the equatorial plane of the flat waves from

any atom, propagated in every direction parallel thereto.

Out integration should include the disturbances along these planes in which the waves are flat. Accordingly, for the waves from any atom we have

1t 2it

(j) = { r/47l")S SF {l(x+at cosCl)+tn ()+at sinCl sin w)+n(z+at sinB cosoo) -(at+s)) tsin CI dt1 dw

o 0

r.: 2Tt

+(1/ 41l')(d/dt)S SII {l(x+at cosd)+m(y+alsin d sin w)+n(z+at sind cosw) - (at+s)} tsincl dt1 dill . o 0

And if we integrate this expression for the waves"lfrom all the atoms of a body; we shall have

r r.: 27C it 21t

fD. S S S(O-/4n:)J S F{l(x---t--at cosd)+m(y---t--atsin e sin w)+n(z+alsinB cos w) - (at---t--s)} r2 siD_d dr dS dill '/sinB dB dill

000 00 _ _

r t: 21t 1t 2it

+ S S S(a!4n:)(d!dt)J SII{!(x+at cosB) +m()'+at sind sinw)---t--n(z+atsin S cosw)-(at+ s)} r2sin S drdS dor- tsinO dB dw :

00 0 00

This equation may be simplified somewhat by a transformation employed by Poisson in _ his Memoir of 1819, p. 127. In this we put:

and then the second

la = p cosS' m a = p sine' sin w', n a = p sin Sf COSfJl

terms 'under the integral signs become of the form

t (cosO" cos d---t--cos (w - 00') sind' sinS} = tcos '/J

dQ = sinSdS dro

'and therefore

. r 1: 2-::": T.:" 2';t

(]J = J J S(o-/41l')SJF{!x+~)'+nz-(at---t--s)+tcost/J} r2sinddrdSdw-tsinBdddw

000 o 0

r;t 2r:; 1t 21t

-+- S J S(a/4rr) (d/dt)S Sn {l X+,.rlY+~Z-(~t-+-S)+/C05¥-,},.,2 sinO drdil dw·tsinO dd dw.

000

00

5085',

. '

(ii) Simplified expressions for all the elements of a spherical surface with motion making any' angle with the radi\l;;;.

Accordingly, when we have equations of-the type found in Poisson's expression (I\femoir of 181,9! p. i27): -.:< 1t 2"

P= S S/(gc-osu+hsinusinv+ksinucosv)sinududv 00·

we 'may put

and thus obtain:

s . f sintt' sin yl

g = pcosu'

k = p sin zz' COSy'

1t 21t

P = S S / {p lCOSl/ cosu+cos(v-y') sin zs' sinu]} sinu du dv o 0

By usmg the simplifying formulae:

cos1jJ == cosv' cosu+cos(ll-_v')sintt' sinu

dw = sinu du dv

this reduces to .

i:t 27t

P = S S /(p cOS1jJ)dw.

00

:tThus this quantity P represents the sum of all the elements of the spherical surface, multiplied each by a given function of the- cosine of the angle comprised between its radius and a radius determined in position.«

A wave flat in the equator of the atom is defined by

Ix+m):-I-7tz-(at+s) = o , (48)

The coordinates for the spherical propagation of the wave are

x+atcos8 y+atsin8 sinw z+atsin(cosw. (59)

Hence rt 21t

W = (I/4n)S SF {l{x+at cos B)+IlI{y+atsin & sin w)+n(z+al sinS cos w)-(at+s)} t sin8 dt1 dw 00

c: 2C:

+I I/4rr)(d/dt)J S I1 {l(x-l-at cosB)+m()+at sinS sinw)+~(z+at sin8 cosw)- (at+s)} t sin B M dre .

00

These solutions are general for wave motion in ligbt, I coordinates of the disturbed molecules at any time I, will be magnetism or similar natural phenomerna : and thus it remains I found in a sphere surface: (at) 2 = (x- x'F+ (y- J"F+ (z - z')2

to exa min e certai n expressi on s in Gauss' Th eory of Terrestri al x - x' = r cos B = at cos S

Magnetism, to ascertain if these phenomena are consistent with the wave-theory. But before entering upon magnetic phenomena, we summarize the hypotheses underlying Poisson's analysis as briefly as possible.

. (iii) The equations for wa-res propagated spherically in an elastic medium.

Consider a system of waves propagated spherically, from any point, whose coordinates are x, )', s, t. Then the

And Poissort's surface (at, S,

)'- y' = r sin d sin to = at sind sin 00

z-z' = rsin~cosw = atsind cosw_

Accordingly at the time t the coordinates of the disturbed molecules will be:

z+atsind cos ee .

x+at cos &

y+ at sin 8 sin w solution' vields the

(0) : .

integral over the sphere

W= (r/4iT)S Jl'{x+atcosB, )'+atsinSsinw, z+atsinBcosw}tsinBdBdw 00

. ....

~ 21t

-I- (I/4;T)(d/dtlJ Sll {x+at cos e , .J'+at sin B sin w, z+at sin S cosw} t sin B dB dw . 00

And the equation of wave motion is:

O~(/)/C,2 = a2 (&2 wjox2 + C~W/Oy2+ O~W/Gz?) . (64)

The fundamental equations

du/dt=a2ds/dz dv/dt=azds/dy dw/dt=a2ds/dz ( ~)

ds/dt = du/dx+d1J/d)'+ dte/ciz s = (I la2) d(lJ,.'dt 6 ~

lead to the components of the velocity of any molecule

u __:_ dW/dx+ [! u = dW/dy+ V It! = dW/dz+ W (66) . where U, V, TV are arbitrary functions of x, )', z, in accordance with the conditions laid down by Lagrange in the Mecaniqus Analytique.

(iv) Gauss' theorem 'that the sum total of positive and negative magnetic fluid in any magnet is zero confirms the wave-theory of magnetism.

In his Allgemeine Theorie des Erdrnagnetismus, 1838,. p. 2 r , Gauss has shown that the sum total of positive and negative fluid in the entire earth is zero, so that

S df!' = 0 • (67 )

The expression for the potential, due to the magnetic mass p, is

V= -SI/e-dfL - (68)

where the integral to be extended over the whole magnet. and e denotes the distance of the element of magnetic masS. dfL from the point acted on (x',)", z').

,4

1

(60)'

- seen j ,,-hile

, >,,:,: ·This " -may b

':~

"action

of two "dynam

-'p=1

<

':'taken the ea

reduct

, ) I (63 '

The]

, depen angle:

ClTCU!"

(77) ,

force, - b( space of at

Plan and

L

if the integration is rigorously restricted to the limits speci-

bz) lied in (74) and bs). '

Now it happens that the actions between two globes which lYI and 1Il are not restricted to their centres as seen from each other; but the globe's subtend measurable angles 2 to, 2W', and the atoms are correspondingly dispersed. When the mass is concentrated at the centre, suppose it restricted to

a minute measurable area of unit size; then' the actual expanded bodies will be larger than 'this minute area in proportions of y and y' times. If the action on unit mass in the minute area be one unit, the action of all the mass in (73) 1V will be p (f times that powerful; and that of all the mass in 111 will be y' 0". Hence the necessity of integration over every area however small and however minute the density.

(vi) The wav€' action positive as in the obs~r'ved case :iJ of gravitation.

If the concentration of the action of the distant body in the centres .iff and mbeindicated by integration with 2. In the same way the coordinates of the earth as rigidly fixed limits, 2IT in the case of ),' and n in the case seen from the centre of the sun may be regarded as fixed of u, - which restricts the mutual action to a single minute while the integration is being' extended over the sun's mass. area - we may write two integrals for the whole action: 3. The two masses as respects each other are .thus one with no spacial distribution, and the other variable

reduced to weighted points of mass jlf and m, The action throughout the solid angles 2W" 2W':

p , S S S S S S( I! ~) ii' cos (i,i') 0' dxdJ! ds 0-' dx' dy';dz'

tion

(48)

I be 'z'j2

the

'62} \ ,

here

'.

In rectangular coordinates we have:

Q _:_ V[(x-x'j2+(y-Y)2+(z-z'FJ (69)

.and. in the spherical coordinates used by Gauss,

I! = V{r2+roZ- 2rro[cosucosuo + sinusinuocos(}.-).o)]} (70) u and Ito being polar distances, r and TO being radii of the earth, )'0 a fixed longitude, and A. a variable longitude, to be used in extending the integration throughout.the mass.

T~ ~

v= - Sl!rl'd,tv. ' - SSS(a/e) dx aydz

. - SSS(O'!e) dr·,. du·rsinu dL b r)

Accordingly when we extend Gauss' theorem to the entire terrestrial globe we have the expression for the potential:

-418

of the sun on 'the earth's atoms is equivalent to the action of the earth on the sun's atorns : .

27t rc:: r

P = mS d)Jsinu duS(a/el 1'2 dr

. 0 0 0 '

2"71:.re: . r

= M S'dlSSinu duS(o/el r~ dr.' b 5)

o 0 0 .

And both of tbese expressions are zero, in accordance with (67) and b 1); for if in the case of the earth's magnetism

involving 1!187Sth part of the atoms, S dfl' , 0, ,:hich means SSSCI/I?)O'dxd)'dzii'cos(i,i') = 0; so also in the case of electrodynamic action depending on all the atoms, it follows

that Sd,u,!e = SSS(I!e}ii'cOs(i,i')adxdydz = 0 (!6)

,2-;:: ;: r

V.= - Sd1Ssintt duJ(a/e) r2dr.

000

.This will give the potential at any point (;'0 ,Ito, )'0), may be outside the' earth, as in the moon or sun.

2"":t 1t T 2'it ~ t". __ ,

- J S J S S S(I!e) ii' cos(i, i') 0',.2 drdJ. sin « d~ a' T02 dro d)·o sinuo duo = 0

o 0 0 000

2{j} 200 T ZOO· 2(~' TO •

+S S S S S J(I/e)ii'cos(i,i')ar2drd)'sin1idItO""oZdro8}osi~tJod/(o =A.

000000 .

(77)

The latter expression A is positive, because all the factors (i) Mechanical. analogies are convincing.

depending on the cosine, i i' c05(i, i') are positive - the total I. We have found the aethet to be enormously elastic,

angles o'f integration being in excess of a whole or semi- so that when any pencil of the medium is filled with a beam ,Circumference by the amounts Z(d, 2W'. This last expression. of light, ~hich consists of waves tilted at all angles and (77) explains why gravitation always appears as a positive flowing on in almost infinitely rapid succession, the pencil force, though the electrodynamic action on a point vanishes, i may be viewed as maintaining its figure by the elasticity of

, -becau,e also it emits no waves. Both bodies Iii! measurable I the medium and the rapid succession of the waves.' If the .' space, and the angular overlap .is 2 W, 2 to' when the action I pencil of light strikes a solid or liquid surface; the speed of

of all the atoms in both bodies is considered.' , the wave motion is suddenly checked, and reaction on the

::, 7. Why Reflected Light is 'Po'Jarized in it equilibrium of the 'pencil at the boundary takes place: so .Pl a n eat Rig h tAn g I e s to th e PIa n e .0 fIn c ide nee that the vibrations in certain directions are altered by the

" and Re fl e ct i o n : Confirmation of .Fresnel'S Views. contact with the solid or liquid surface.

(v) Extension or' this theo~em to the electrodynamic action between tWD spheres, as the earth and, sun.

Imagine el~ctric currents tocircuiate' around the atoms of two globes: it is required to considerthe resulting electrodynamic action. We have the sextuple integral

p = S S S S SS( I Ie) i i' ~os (i, if) a d.e d;v dz 0" dx' dy' ds' .

and the'

'The two masses may be called j}f and m , the latter being the sun.

1. The coordinates of the sun (ro, 11'0, )'0) may be taken as fixed, while the integration js being extended over the earth ..

5085

419.

c .• : 21 To judge by a tangible and familiar experiment, 'as to what may happen to a pencil of light, we may compare it to the stream of water flowing froni the nozzle of a garden hose. The cross-section of ·the stream of water is assumed

_ to be circular, and we recognize that the forces which keep it ·so, are chiefly the forward motion and surface tension, - - the -attraction 'of the water for itS'elf. In the case of the _ p-encil of light, the equilibrating forces depend on the elastic power within the aether, and thus are different; but the effect produced is very similar, for any slender cylinder filled with a flow of waves.

. 3. Now we know by daily observation, that when a round -, stream of water is thrown by a hose against a .solid wall, the . cross-section of the stream ceases to be circular, and. becomes' highly flattened, so that the new cross-section of the stream becomes an ellipse, having its longer axis normal to the plane of incidence and reflection. The flattening of the reflected stream of water- is easily seen by the most careless of observers: and thus analogy leads us to expect a similar flattening of the vibrations in a beam. of reflected light. It is true that the _ flattened stream of water is not vibrating like the nether, yet the reflected stream is flattened, and. tends to retain that figure, . with elliptical cross-section.

4. It has been proved by flash-light photography that when liquid drops are forming and falling, the detached spherules oscillate. about a mean figure, - being alternately prolate, then spherical, -and finally oblate. In the case of drops therefore the particles of the fluid oscillate about a mean position, under the influence of surface tension. The

:;; figure of the drop is drawn out of shape at the instant of detachment, and in falling the. action of surface tension restores the normal figure, and carries it beyond, so that the globules oscillate about their mean form, which is spherical.

5. Now in the same 'way, when a pencil of light is reflected from a solid or liquid surface, the act of reflection I brings into play, for an infinitesimal -tirne dt, certain forces I which tend to flatten the beam, as reflected, in a plane normal to the plane of incidence and reflection. Considering merely the relative motion of the beam ·in respect to the solid or liquid, we may regard the circular pencil as - struck by the solid or liquid in the act of reflection .. 'Owing to itseJasticity, each element of the pencil rebounds like a rubber ball - flattened in the plane normal to that of polarization, as we. see in the actual behavior of rubber balls in collision. Since each element of the pencil is elastic, there is incessant recovery from the flattening effect - so that the pencil continues to' vibrate, but by relative crowding of the vibrating aetherons it has lost 'itscircularity of cross-section, arrd become elliptical, owing to the restricted freedom imposed in the. process of reflection.

. 6. In Fig, 6, Plate 8; we may imagine equal amplitudes of vibration, in' all directions from the centre of the incident beam as shown above ; but after reflection the resistance thus encountered forces the circle into the ellipse, as shown. The mutual crowding towards the centre, owing to restricted freedom at the instant of reflection, forces tbe pencil as a whole out at the sides, and thus it takes on a very elliptical form for the cross-section. In spite of a notable flattening of the

pencil of light the aether particles describe ellipses not

straight lines, as often stated; in the theory of polarization.-)

7. It has been ShOWn by the recent researches o{ .Hcibtrg that Archhmdes· used mechanical means of provirig his theorems, at least in the first instance, and then made them rigorous by improving the geometrical demonstration. Accordingly," in dealing with" polarization, . we are justifie"d in adopting similar methods. . And th-e only question is One of devising a valid model which affords a true analogy. To this end we rely upon the evidence of experiments, in phe-_ nomena easily understood and admitting of but one interpretation.

8. The model of a reflected .stre a m of water above outlined certainly is mechanically valid._ And it may be con. firmed and extended by considering the instantaneous forced form of a series of rubber balls, in such close succession as to be united into a solid tube, like the stream of water yet not actually touching prior to reflection. At reflectio; each ball would be flattened by the resistance of the reflector so that the vi brations of the aether in the pencil take the same form, as observed in polarized light.

9. When liquid drops are formed, in the breaking of a jet, flash-light photography shows them in rapid vibration, owing to surface tension. They form, and vibrate up and down, under gravity; but the waves of the aether pencil would vibrate normal to the plane of polarization, when they are reflected. The vibrations in a plane at right angles to the plane of polarization thus' necessarily results from the reflection of waves in an elastic medium.

10. Accordingly, on the basis of actual experience, in well idefined phenomena, it is impossible to imagine any kind of vibrations in reflected light other than that at right

angles to the plane of polarization. .

If we adopted the jJ.£aaullagh-}\,TeumQ1l?1 theory of vibrations, in the plane of polarization, we should have to expect mechanically a similar effect when a circular stream of water is reflected by a smooth solid wall. No such effect is observed. And as reflection is equivalent to a blow against the round moving stream, renewed at every instant, at infinitesimal intervals dt, we see clearly that the distortion of the vibrations should take place, with the longer axis of the new elliptical motion at right angles to the plane of polarization. No other result is mechanically possible.

(ii) Analysis of light vibrations.

Let the three components of the revolving light vector be:

In \1 of t amp squa

we u

.. of tI:

- sio(z ·tos(z

sit

And princ p-q ber

This ·mav

. '

cond

It =

.u=

line, or p

u = a cos (ZIT t/r--+-a} 1) = b cos (2fr t/r--+- jJ) U) = ccos(21rt!r+y}

(uja)2+(v/W+(w/cj2 = I S = V(u2 --+-v~ --+- 1lJ2) •

lariz inter right of d

The fou-rth of these equations indicates that the path descri bed by the end of the light vector is an ellipse; and ' the fifth equation gives the displacement relatively to the equilibrium position of the aether particle at any time.

By altering the angles through 1/2;1: - & ;. we are enabled to' use sines in the place of cosines:

sin(z;T t,'r--+-p) =It/a=sin(z/Tt!r) cosp-+-cos( z n. the) SiDp

sin( 271: t/r:+q) ~v/b=sin(z;r tlr)cosq +COS(ZIT tlr) sinq - (79) sin( Z/T tlr -r- r}= «[e=: sin] 2 IT t/r-) C05 r--+- cost 2 IT t/r) sin I' •

com

b9)

but this then

'1/=

.s •

bove can)fced ssion ·'ater, :tion .ctor, : the

nottion. .s of' lving' nade tion. tified

: oue

To phe· nter-

The quantities u, u, w, represent the rectangular coordinates of the end of the revolving light vector; and the eanation for the path, q nite independently of the time, may b~ obtained by eliminating t from equation h g), by the following process. If we multiply the expanded form of these equations by sin(q-r}, sin(r-p), and sin(p-q) respectively, and add them, the right hand members will be found to

. vanish, and we get:

(u,/a) sin(q-r)+(1J!b) sin (r-p) + (w/c) sin(p-q) = o. (So)

This linear equation connects u, v, UJ, which are the rectangular coordinates of the end of the light vector; and hence we see that the path described by it lies in a plane passing through the origin. The path of the vibration there-

.fore is a plane curve.

To get the path as projected on the coordinate planes, we use two of the equations b 9)' Thus from the first two of these equations we obtain:

sin{zrr t/1:)(cospsinq - cosqsinp)= (u/a)sinq_:_ (L'/b)siop (81) cos] 1 i7: t/T){COSp sinq -cosq sinp) = - (u/a)cosq+ (v/ b)cos p.

If we square' and' add these two expressions,. we get

sin?(p-q) =. uZ/at+v?jb2,- z (u/a)(v/b) cos (p-q). (8 z) And we see that this equation is that of an ellipse whose principle axes coincide with the coordinate axes when .p - q = l/Z1f, so that only the first terms of the right member remain, and the left member is unity:

I = It?/a~+v?/b?. (83)

This represents elliptically polarized light, in which a and b "may have any ratio.

. If ~ve put w = 0, a _.:. b, r :» = ± 'l-», we have the conditions for circularly polarized light:

11 = a sin (zrrt/T) v = a COS:(ZiT ftC) (right handed) ( .)

11 = a sin (zn tiT) v = - a_ cos (m: t/r) (left handed) . 84

When the vibration ellipse reduces to a straight line, Of in practi c e approximately so, w = 0, r=: q = 0, or p - q = IT, we have by taking the square root of (8 z):

u/a±v/b . o . (8S)

In wave motion, the intensity or'the action, or the energy of the disturbance, is proportional to the square of the amplitude. Hence we add, for the geometrical sum, the squares of the component amplitudes A,B, C, and thus obtain:

7 = A2+-B~+Cz.· (86)

We shall no, .... apply this composition theorem to po~arized light. It is well known that such light is free from I I~terference, when polari zed in planes mutually inclined at I right angles, but always gives an intensity equal to the sum

. of the intensities of the separate rays. I

(iii) Analysis of the cOmpOSitl,'On of polarized light I'

compared with the evidence of observations.

-, Let us superpose upon the ray defined by equations :,_ (,9) and traveling along the s-axis, a ray of equal intensity, " but polarized at right angles to it. If the components of ,this new ray be u', .u", w', and the phase difference. be 0;

.then we shall have

~'=Bsin(:mt/r+q-+-o) v' . -Asin{mt/T+p+O)

. , .(871"

/1/ = Csin (Z1l' i/T+ r-+- 0)

.g of .tion,

and encil they ~s to' I the'

ence, any right

ibrarpect vaiii

ob-ainst

5085

422

Except for the phase" difference 0, these components become identical with those in h g), by rotating the coerdinate .system through goO, about', the s-axis. Accordingly, we have by taking the sums of the components, thus geo-

metrically compounded: .

u + u' = A'2 = A2+B2+Z,ABcos(0+q':"'p)

v+v" B'2 = A2+B2- 2ABcos(o+p-q) (88)

UJ+w' = C'2 = ZC2(I+COSO).

By simple addition we have from (86) the following geometrical composition of the components of the light vectors;

7' = A'2+B'2+C'2 (8g)

. which is equivalent to

Y' = 27+ zC2 COSO-4Af'sin a sin (q-p). (go)

But it is found by experiment that we have sensibly 'J' = 27, or the intensity of the compound ray is equal to the sum of the intensities of the separate rays, and independent of the phase differenced. Hence _it follows 'that the . secon d a rid third terms in (90) are so small as' to be insensible to observation. Therefore we conclude that witl1in

the limits of observation: '

C~o

p-q= 0."

That is, in polarized light the- radius vector is sensibly perpendicular to the direction of propagation of the ray, and the motion therefore sensibly transverse. Also from equations (S2) or (85) it follows that the particles vibrate sensibly in a straight line.

From this analysis, it follows that rays which "have suffered double refraction or reflection at the polarizing angle are plane-polarized, and thus consist of vibrations which are sensibly transverse. We use the term sensibly .transverse, instead of absolutely transverse, in order to reconcile other' facts of observation with mathematical theory.

It is shown by experimental research that when plane polarized light is reflected from metals, the" effect -is to convert it into elliptical polarization, -the degree of the ellipticity depending on the direction of the incident ray, and on its plane of polarization, as well as on the nature of the reflecting metal (cf. Ganot's Physics, J 4th ed., 1893,

§ 672, p. 656).' , .. '

When the plane-polarized light is 'reflected from silver, ·the resulting polarization is almost circular - probably because silver is so perfect a conductor of electric or aether wave motion, that the normal tendency ·to elliptical motion is largely restored. But if the plane-polarized light be reflected from galena', a lead ore of low electric conductivity, the resulting polarization cremains almost- plane.

. Now since elliptically polarized light never vanishes,

when examined in a Nicol prism, though at alternate positions it becomes fainter, 'such elliptical motion in light must be considered the general type of vibration of the aether particle. If therefore plane-polarized light, by. reflection from metallic surfaces, is rendered decidedly more elliptical . in its' motion, it would seem to follow that .in plane-polarised light the motion is never strictly rectilinear; on the contrary such

"light always has in its motion a' slight elliptical element, which permits of notable restoration, by reflection from silver .

and other high conduc!ingmetals. ".

. ...~.

' ..

..

423

50.85

_ , __ .... _., . ,

:;>:::.~: It is Jar these reasons that;· in. our discussion. of the similar reasoning will hold true for elliptical paths of an),.', .•. above equations, (82) to (qI), 'we admit.' the: plane-polarized kind, and hence the results- here shown are true for every"..

vibrations to be only sensibly transverse, not rigorously trans- kind of vibrations in 'polarized light. . ; -- - . :i'

verse, in' rectilinear paths:, _ :.. ' .... . It will be noticed that the rays consist of plane waveS.

_. This conclusion from the' combination of experimental with amplitude A and wave length )., and the ratio AI), is':

i-~5earch:with mathematical analysis fully .sustains our .view comparatively small, but here drawn on a scale large enough:: that_ there necessarily is 'aTongitudinal component in light. to enable us to see the rotation of the elements of the waveAn-f 'other view than' that here set forth is contradicted by at every_ point. The waves are imagined to be' flat in the';. well established facts' of observation, which appear to admit. plane. _ of the paper, and hence they have a 10Dgitudina~

of but one interpretation. - component depending on the amplitude A. -

8. The Undulatory Explanation of the Phe- 4. The adjacent diagram of light and dark bands shows

no rn e n o n of Interference in Polarized Light con- the interference effects, and is seen to have strips of darkness f'o r m s to Poisson's The 0 ryo f the Ell j P t i ca 1. Vi bra t i 0 D S and of light, where the motions of the split rays are such. as of the Aether Particles mainly in the Direction to destroy the rotation, or augment it by the superposition of the Normal to the Wa veBu r fa c e. of the separate effects. If, for example, the wave travel along

(i) Explanation'of j~terference when the particles de, the .r-axis, the displacement of the particle parallel to the

scribe. ellipses. y-axis is Tt, and ~ parallel to the ·x·axis:

We have shown in section I and- 4 above that the 1i = asin (2;Tx/),+a) = dy

traditional theory of the transverse vibrations in light is not ~ = acos(2rrxjJ.+a) = dx _ (92)

strictly rigorous, but Teq uires rational revision, to take account s = VClf2+ ~2) = a, 'in circular motion.

of 'the geornetri cal con di ti ons speci fied by Paisson.. and th e

related electrodynamic waves from each atom, which underlies Now a detailed treatment of the leading phenomena the theory of magnetism. Thus it is advisable to reexamine of interference is beyond the scope of this paper; yet we the bearing of these results on the theory of interference of I may sketch briefly the wave-theory of this subject, in order polarized light. - 1 to show that in spite of the ~efect above pointed out in the

. ' I. The ordinary explanation of interference handed. I fo.nn . of the o: wave-theory ~f lJ~ht handed down by YouJ/g, down from the days of YOu/lg and Fresnel is based upon an Fresnel, Ara,,:o, and. Cauch), this defect does not invalidate assumed analogy with the side vibrations of an. elastic cord. the explanation of interference,

This theory allows disturbances given the cord to travel along (ii) - Theory of the light and dark bands.

it, while the particles of the cord have only a transverse An adequate treatment of diffraction phenomena would

motion. But we have seen that this explanation begs the require a mathematical discussion of FresucTs integrals (Drude's question, in that it practically assumes a »stringy« condition Theory of Optics, pp. 188-196), which have the form:

of the aether, whereas Poisson's theory of elliptical vibrations, v v,

with their major axes in the direction of the normal to wave ~ = S cos 1/2rrl!~ db . 1j = SSiD 1/2nv2 dv.-

surface, gives an almost identical .result, without physical 0 0

premises involving the anisotropy of the medium, or geo- These functions may be thought -of as the rectangular

metrical postulates of purely lateral motion which cannot be -

coordinates of a point in the light plane ~1l. Accordingly,

admitted. )

from. (93 we have at once:

2. Accordingly, the analogy of the waves conveyed

along a twisted cord seemed so plausible to. those who did not study the problem deeply, that it came into general use, and stili holds its place today. Yet a more mischievious doctrine seldom has been introduced into science, because although plausible, it is dynamically and geometrically un-

sound in principle. . .

For wby is the aether, in the traditional form of the wave-theory, assumed to be capable of a transverse motion of appreciable dimensions, but incapable of an equally large longitudinal motion? The chief reason for this hypothesis - for it is merely a convenient hypothesis - is the problem of explaining interference, and polarization. It is known from modern research that diffraction only requires that the length of the waves shall be small compared to the dirnension of the aperture.

3· Fig. 8, Plate 9, shows how a split beam of planepolarized light may produce interference fringes when they, differ in phase by l/Z)"

For reasons of simplicity in construction the oscillations of the particles in the figure are taken to be circular, yet

And when the spacial length s is measured from the

origin, we have s = v. (96)

The functions ~,'1j 'are iIJu~trated by the following fig. 9 (Dmdc's Theory of Optics, p. J9z), which has been calculated by the method originally due to Corm; (Jour. de Phys·,.3, 1874).

Fig. 9.

Diagram of the double spiral of Fume/'s integrals, for the dif-

~ fraction of light. The curve coils about the two asymptotic points

"J Fand F'. where v= +00, and

t: _,/ u = -00. The maxima and mi-

____ =_-+-::.,:"',;;',~" .. ,,'----'--,. ---C.,c- __ ."'", _ ~ nima of intensi ty lie approxima-

tely at the intersections of the line F F' with the spiral curve. If the free intensity be I, the 'maxima are.l\ == I~34,7'l= 1.20)" 7,=J·16;theminima7,=o.,8, 7,=0,84, JB=o.S7:cf.Fnme/ Oevres Completes, I, p, 322).

q ..

/

ACI

,.

trical sur lO'evalual that the

00

;F SCO

." . 0

with con nates are

In the two

) • I

C He

and a is extent; \\ _. (b =

'P.; angl Ne

x'

C=s

-00-

x

S S

-00,

------'----

ragalion .hove into priale frir

much occ

·faire qu', later~les t de 1" lun .-t-il dit (etail u~ ~O·Lent sor

~ited tha memoirs·

than first

5085

426

'. 425

.ows ness h as tion ang

the

:; ~:

---00 -OCt

(iii) AppJication of the theory to the formation of diffraction patterns.

It is found in the application of the above functions,

that the equatious vgive the central fringe intensely white, with adjacent blackest area, near the centre of the pattern; here the double integ~1 totally vanishes, but on either side of the centre there remains some illumination. When a space has been traversed along the x-axis equal 19 a certain length, the light reappears as it should do by the above equations

depending on sine and cosine, with corresponding periodicities. The sharpness of the boundary is an essential condition of a well defined diffraction pattern; and without the fulfillment of this geometrical condition; a satisfactory exhibition of the theory can hardly be realized.

In practical experience therefore. the values of the double integrals often are somewhat approximate, - the formulae being rigorously true in the centre of the dark and light bands, when the screen effect is mathematically sharp, bu t at other places on I y pa rtl y . true, - and thus we have, interference bands, shading away gradually and attaining

maxima at intervals, 'where the contrast reaches a maximum, as shown 'in the diagram.

In general the researches of experienced physical investigators have shown that the theory of transverse waves accounts for the diffraction pattern with great accuracy. In section 5 above Prof. Hastings states that be found Huygkens' construction accurate to I: 106, which is a remarkable degree

of precision, and equally valid as applied to diffraction phenomena.

Now in our slight correction of the foundations of the (Ioz) 'wave theory of light, given in sections ,I and 4, we found that an accuracy of I: (4' :1:06) might be attained before any sensible outstanding phenomena would be likely to arise 2).

And as this is below the limits of our perception in modern (1'03) experiments, we may dismiss the question as beyond the limits of detection in the present state of physical science. But to show that a real longitudinal component should exist in light waves, though it is excessively small, we may

recall an actual measurement of the smallness. of the lougi-

( ) I tudinal component in a well dete. rmined experiment with. 104 sound. The late Lord Raj/leigh observed the musical note r: due to a pipe of an organ which could be heard 'at a

distance of 820 metres; and found by measurement that the

It is shown by Cornu's method that for the asymptotic points F and F' we have

co

;F = S cos 1/27fV2 dv o

-. :,.

co

S· 1/· "d

1jF = SIll t7fV- V .

o

aves

.l. is' Itrgh' .aYe

the ina!

, "

These integrals may be evaluated by putting x, y as the rectangular coordinates of a point P, X2+y2 = 1'2, where r is the distance from the ongm. If, therefore, we put:

co

S e-'''" d.e = jlj o

co

Se-Y'dy = M o

we get for their product the double integral: co ex:

S S e-Cx'+Y')dx dY'= ,M2.

o °

Accordingly dx dy may be looked upon as a geornetrieu! surface element in the xy plane; and the problem is to evaluate Fresnei : integrals for the diffraction. It 15 shown that the asymptotic point F has the coordinates

.ena

we .der'

the .. mg/ late

=

ir= S cos 1/27f1l2 dv,= l/~ o

=

S' 1/ 9d 1/

'1jF= Sill 2TW- v= 2

o

with corresponding integrals for the point F', whose coordi-

nates are n ega ti ve,

In the more general problem of diffraction we have the two integrals:

c= Seos [/(x,y)J da

s = Ssin [/(x,y)] do'. '

Here the function

/(x,y) = (n/I.) (I/rb + 1/(>0) [x2 cos? p+y?]

Jla( gly,

and (J is a small opening of any form in a screen of infinite extent, while the radii vectores

fJl _:_ V(X12+Y12+Z12}f}o = V(xo2+J'o2+z02)

p =angle of s-axis with (>1 cos(neo)= -COS(71!?1)=COSP.

Near a straight edge 1) these functions C and S become:

x

C=s

+oo

Sd.1:dycos ({7f/;.)(I/~h + rleo) [x~cOS\p+y2J}

94) , 9ir th{'

96)" ~. 9 ", . ted 74('

<

,rl"<, riral ' ,dif-':'

:oils oints·'?

and :. .~

x'

S=s

-r-oc

S dx dy sin {(7f/i.) (r,:'et + Ilea) [X2COS2g>+y2J} ,

-00-=

I) One of the great histo~ical difficulties in the wave-theory of light was the problem of explaining with geome~ical rigor the pro . pagatiOIl in straight lines, since on 9""ghens' principle each particle of the aether in the wave front becomes a centre of disturbance. Tbe above integrals, as worked out by mode~" g~ometers, have their limits So fixed as to include the whole region of disturbance, yielding appro- , ..

pmte fringes due to interference, bur otherwise giving rectilinear propagation. '.

'The celebrated geometer Poisson, as we lean! from the careful note appended to his posthumous memoir of 1839, pp. {S I-I 52, was much occupied with the problem of the rectilinear propagation of light during his last illness

, -Quand le mal rnoins avarice lui perrneuo ir encore de causer science avec ses arnis, il a die qu'il avnit trouve comment il pouvait se

:al:c, qu'un ebranle rne nt ne se prop"gdt dans un milieu elastique que suiva nt une seule direction; le mouvernent propage suivant les directions l.terales etant insensible aus.itot que I'angle de ces directions avec celle de Ia propagation erait appreciable. 11 a rrivait ainsi a I" propagation de .fa .Iumiere en ligne droite, Plus tard, cedant au mal, et se decidant enfin a inrerrornpre I'impression de son mernoire : c'etai; pourtant, a;~-'l. dir, la partie briginale, c'erait di'cisif pour la lum iere ; 'et che rchant avec peine le mot pour expr-imer son idiie, il a repete plusieurs fois : ' teta,! un filet de lumiere , Puissent ces paroles, religieusement conservees par les arnis de M. "Poisson, Jes dernieres ,paroles de science qui sn',ent sorties de sa bouche , mettre les savants sur la trace de pensee, et inspirer un achevement de son oeuvre digne du comrnencemenr.«

': :'. It is unfortunate that Poisson's memoirs have become very scarce, and thus little known to modern readers. It has long been recog·. n'_cd .that there is great need for the reprinting of Poisson's. Collected Works. But for my good fortune in obtaining a set of Poisson's celebrated memo,rs on waves,' formerly belonging to the library of Sir Jolm flersci,,!, the results brought out in this paper would not have been possible.

tb ') In the note dated Sept, t z, below, it is shown that the longitudinal component is A = (At}.) p = 1 ; (66420' 10"). very much smaller

an first estimated. ' C

3mplitir'~e -'of the 'oscillation in the:se- waves' could not begreater

-than 0.06'-6fa millimetre. .: - :;," ---

:' Now 10' case of r: there are 27 8.i'vibrations per second, 3.~d 'the length' of the wave; under a velocity, of 332000 mm pet"secbnd;is therefore' 120 mrn.: IL the amplitude be , '0.06 mm, as found by Lord Ra),leigh, i1: follows that the wave

, length is- 2000 times the amplitude; -As a concrete example, Of the - molecular+oscillations which produce musical sound, this result is quite remarkable.

, , In: the case of light we can determine the wave length very accurately, but we cannot measure the amplitude of the aether waves by any direct process 1). Yet if the length of typical musical waves be some 2000 times their amplitude, it will follow, from the nature of the similar cause involved, that for so elastic a 'medium as the aether the waves should also be enormously longer than their amplitudes - much grea ter than 100 times, - as, assum ed by Kelvin, kfaxwdl, and Larmer. This value of A/). = 10-2 is a relatively small' amplitude, but it gives, a lq,ngitudinal component 20 times larger than that noted in the sound wave above cited,

. Accordingly there is reason to believe that in the case of so highly elastic a medium as the aether the amplitude A is- less than 1: H)OOOO,h of the wave-length, or at least 1000 times smaller than, Kc1l,1in, _J,faxwell, and Larmar assumed __ This would make the ratio in the case of the very elastic aether fifty times smaller than was observed by Lord

Ro)'leigi for typical musical sound in our air. Such a value (iv) Integral expressions for the kinetic and potential

as I: 105 certainly is not too large, but it may be that the energy of. the medium when filled with waves.

ratio should be considerably smaller yet. Let ~, "l' , be the rectangular coordinates of a .particle

The following, figure' I r illustrates the interference a-t the time t, then the differentials d~, d1j, d~win denote phenomena observed when light passes through a glass wedge, the component velocities of a particle in the medium which with the sides mutually inclined, at a small angle. is propagating the waves. The particle is oscillating periodi-

This too represents interference, much like that of cally about a mean position, at any time t, and thus has polarized light shown in the preceding discussion, but it both a velocity of which the components are d~, d1{, dr, exhibits the phenomena more ln detail; and the phenomena and a distortion from its mean position, or displacement. exhibited are consistent with rotating elements in the waves It is well known that in wave motion half the energy is like those in the first diagram. The wedge of glass explains kinetic, half potential: tberefore the kinetic energy due towhy, the waves interfere in bands at right angles to the the component velocities of the particles becomes;

T = l/s (f.t/nlS S S[(d~/dt)2+ (d1j!dt)2+(drjdtFJ dx d)' dz .

5,085

height of the wedge" In the first diagram the direction of the height of the ",'edge, for separating the phase of the wave by 1/2)., would have to be imagined horizontal, and' the light returned along a path parallel to its ,emission .'

Fig I I. Familiar illustration of interference and reenforcement, when the light of a candle falls upon a glass wedge (lliillikan and Gale). This gi\'~s bfight and dark bands, parallel to the edge of the wedge, exactly as in th e case of Newton-rings about the centre, in the case of a lens.

428 ..

., propagatio

the· tender

;;T,.,' 2. I

. f{'~,

'''1849, sect

~·:·~I.I) ~

·p~()duclOg

It is now

~r·' .

41spersion

::;'i'5e 'to th . '1i~' ~eferre "~OItance ,

·,~,t.:: 3·

or partie:

';nethods

, bi; . FM(l1 century a. Q(mathel ~ifferentia 'action, m .methods,

waves in I

4· , in proble undergoir domains partial dil

',to be so it, should "philosoph and Pais.

9·

'under E

nuo us E Wa v e s 1

I.".

In we have. to be ii .that the . not here

is about 'lision pe normal I clity tow analogie! fully rev _caution.

Al the ligh

For the potential energy due to the distortion -of the elements of the medium we have:

-W = (1/47fK)j S j[(d(jd)'-d1!dz)2+ (d~!dz- drJdx)2+ (d'lj!dx-d~/d_l')2J dx OJ' ds , (106)

In these equations the component velocities of the wave disturbances are ciS/dt, d1'j/dt, dr/dt, and the distortions of the form .of the elements d1 = d~/dy-d1jldz, 02 = d~/dz-d~/dx, O's = dlj/dx-d~/d)1 give the displacements of the elements

of the medium along the coordinate axes: -

The total energy in the medium at any point is the sum of these two energies: T -+ W = g, or

Sl= (1!47t)S S S{l/~ftl.[(d~/dt)2::::-(d'lj/dt)2+(dr/dtF]-t-1/K' [(dr/d)'-d"lldzF+(d~/dz---'dr/dxF+(d1jlox-d;/d);n}dxd_l'dz (107) which iliustTutes the agitation of a mediumfilled with waves.

1) Since wriling the above paragraph, it bas occurred. to me that we may calculate the theoretical ratio of the amplitude to the wave- 1en¥,h of t~e aether by the following process. We have proved that the aether is z = 689321600000 times -more elasricfban air in proportion to Its density. And it is this 'elasticity which gives the a ether waves their enormous velocity; and, as compared 10 air, the 'amplitude shovld be smaller.in' proportion 10 the square root of this number. For when a wave in the aether begins to be generated it speed} away so rapidl)" under the enormous elasticity, that the amplitude is small in the fame proportion that the velocity v is great. Now from the above value we find that -Va = 830254; and as the ratio in air furnished by Lord Rayleigh's experiments is 1 '2000, we have for the a ether the re lative rado: 1(830254'1/2000 = 1/1660508000, Or 830254 times smaller than the ratio of rhe amplitude to the wave-Iength in the musical sound investigated of Lord j(a)'high. The true ratio thus appear to be 16605 limes smaller than that indicated above, and should be Ail. = 1 : (16605 - 10',; = 1:(1,66°508- 10"), which makes A = (Ajl-.)p = I :(66420' lO"). - Note added Sep. 12, J920.

,- ,

5085

430

1.

'\. ,

. \ , The wave-theory indicated by all the phenomena of papers on the New Theory of the Aether, I believe we may

nsture. safely conclude that, notwithstanding the veryextraordinary

In concluding this discussion we draw attention to the physical properties of the aether, in a certain sense it behaves indications of nature from the widest survey of physical as an elastic solid for quick acting forces; namely, that the phenomena; , ' aether will faithfully transmit any kind of vibration communicated to it, whether it involve dilatation of volume or

I. In the whole domain of mathematical physics, modern

. b d bl N mere change of form of any element dx dy dz .

investigattons lay great stress on oun nry pro ems, 0\"

DOLlndary cond i ti 0 ns natu ra II y would have great importance . Un 1 ess we gran t th is extraord inary power 0 f trans-

, . f mission of wav" e motion, we can scarcely reconcile' the

if natural forces are' due to the action 0 waves: because

at the boundary of solid orliq uid bodies tbe velocity of new theory, incl uding the extreme velocity of the ,-aetheron v = 1/2 it V .:_ 47 1239 kms 'per second, with the known propagation is changed very suddenly by the resistance, and, - extreme elasticity of the .aether, which is e = 6893 Z 1600000 the tendency to refraction and dispersion.

times greater than that of air in proportion to its density.

2, In his celebrated article on light, Encyc\. Metrop., It is evident that the aether _ is so different from air, in 1349, section 561:, Sir JO/ln Hersdul shows that the forces

respect to the high velocity of the aetheron, and the en or-

producing refraction are such as • may be termed infinite s , mous elasticity of the medium, that no movement of any

It is now recognized that these powerful actions appear in kind can occur in it without the most perfect response to dispersion and diffraction, as well as in refraction, and gi ve whatever waves arise.

rise 'to the molecular forces, which in a future paper will In this sense I regard 'the aether as' an infinite aeolobe referred to wave motions, thus confirming the great im- tropic elastic solid; but I do not assume" that all the physical partance of th~, wave-theory for all the phenomena of nature. restrictions of the ordi nary elastic solid, which we: can sub-

3· Now quite aside from the physical rconsideratiocs jeer-to experiment in+our laboratories, necessarily hold foi: of particular phenomena, weha ve general mathematical the aether. Some of these physical restrictions, which we methods for' treating partial differential equations, invented ascribe to molecular forces in solids, may be and probably by Fourier, Poisson, Ca!lChy and other geometers about a are missing in the aether, - owing to the absence of the -century ago.. Thus in our time practically all the equations complex molecular structure known in solids, and to, the {If mathematical physics turn on the treatment of partial enormously greater rapidity of the motion of the aetherons,

differential equations, as in sound, heat, light, electrodynamic OUf conclusions therefore are as follows:'

.action, magnetism, etc. And these general mathematical methods, so largely devised by Fourier and Poisson, point to waves in the aether as the underlying cause of physical forces.

4. Accordingly, the importance of boundary conditions, in problems of the transmission of energy through matter undergoing sudden transition of property, by virtue of fixed domains of discontinuity, and thus requiring the methods of parti al d i fferen ti al eq ua tio ns for t hei r exact treatm en t, seemed to be so remarkable an argument for the wave-theory that -it. should eugage the attention of geometers and natural 'philosophers who aim, at extending the researches of Fourier

and Poisson. '

ntial

tide

note hich iodi-

has dL

lent.

y"F

e to

I, Any movement whatever given to the aether will be faithfully transmitted, - owing to the extremely high velocities or the aetherons, which gives the medium both extremely great elasticity and high rigidity, --:- yet the medium is not like ordinary solids, in that it has an extraordinarily small density.

z. The aether, th eref are, has most of the wave transmitting properties of an elastic solid - will transmit any

'kind of wave; yet always with one, velocity only, .a uniform veloci ty V = 3' 1010. ems, which is so mew ha t -di fferent from what is attributed to ordinary elastic solids, with two different velocities, of the following kind, namely:

'(el A -compreSSional wave in an extended mass, say of steel" depending on both - the compressibility k and the rigidity 11;

Vc ~ V(k-+-i!3n) = 655000 ems per second ( J08)

n = 0.95'1012• ,k = 1.84.1012.

(T) A purely transverse distortional wave (withchange of volume) _ expressed by the simpler formula:

Vd = V(n/o') = 348000 ems per second ( 109)

- -~,

':\'

~r

.. <.~·r·

9, Theory 0 f the Propagation of Wave Energy, 'under Poisson's Equation O~(])/Ct2~a2v2([J, in a Conti-UUOllS Elastic Solid, with an Analysis which shows ~Yaves traveling in Different Directions .

In the New Theory 'of the Aether (AN 5044-, 5048)

we have calculated the mean molecular velocity of the aetheron ,to be v = 1/2IT V = 47 1139 kms per second, and shown -'that the aether obeys certain -laws of density .and rigidity' not heretofore suspected. The length of the mean free path is about 5 i 3000 k rns, and in free space less than one col- , in the case of an extended mass of steel, 0" = density = 7.85 '

lision per second occurs between the free aetherons, under I Thus tor steel the former value is nearly twice the normal motion. Owing to the decrease of density and rigi-r' latter, which renders the theory doubtful, in, view of the ..(iity towards large bodies like the sun and earth, all our _old non-separation of the earthquake waves of these two classes. _-ana,10gie~ with the traditional elastic solid have .to be care-·I 3. In certain respectsthe nether is more like a. gas ,~ully revised, and adapted to the new theory with extrelPe than a solid, and up to this time it is probable that ext:caution. - .' ,,'. - periment has not full)' established the two velocities theoreji: .< ' After very careful consideration of these problems, in I cally predicted for an ' elastic solid by Poisson, Cauchy and light of the data contained in the first, second, third I other mathematicians. In ,his Tides and Kindred Phenomena

the

out

en-Is: ""V,', '\=f~~ .:_:;":j;

~

.~

"

"

.~

(

, ......... J'

l'

l

.~ .

I

+-

&

~,

~

",

t

j:

~.

t:

I'·

]:.

,"

I"

."

r

1. I

. Yr.

·,·i.

431

"C of the Solar System, ' 1899, pp. 26 I~:i!, Sir George Darwin '\'re'marks' in .regard to earthquake phenomena: -

"j,'-, e The -vibrations- which are- transmitted through the' 'eaith"are of two kinds. The first sort of wave is one in

- '<whi~h th~matter through which it passes is alternately 'corn'-pressed and: dilated, it may be described as a wave of 'compression;'- In the second sort the shape' of each minute "portion of the solid is distorted, but the volume remains

,',' unchanged, and it may -be called a wave of distortion. These 'twQvibrations travel at different speeds, and the compressional wave outpaces the distortional one. Now the first sign of a. distant earthquake is that the instrumental record shows a succession' of minute tremors. These are 'supposed to be due to waves of compression, and they are succeeded by a much more strongly marked disturbance, which, however, lasts only a short time. This second phase in the instrumental record is supposed to be due to the wave of distortion.«

. »If the natures of these two disturbances are correctly 'ascribed. to their respective sources, it is certain that the matter through which the vibrations have, passed was solid. For, although a compressional wave might be transmitted without much loss of intensity, from a solid to a liquid and back again to a solid, as would have to be the case if the interior - of the earth is molten, yet this cannot be true of the distortional wave, It has been supposed that vibrations due to earthquakes pass in a straight Jine through the earth; if tben this could be proved, we should know with certainty that the earth is solid, at least far down towards its center.«

This reasoning implies that this eminent natural philosopber was in doubt as to the validity of the two-velocity theory, in practice, with actual masses like the earth.

In studying earthquake seismographic records and discussions I - find the disturbance to rise very gradually and die down equally gradually. Thus I have not been able to verify the assumption of two distinct types of waves: we .merely find that at a great distance from the source of

, disturbance the earthquake waves are spread out like a . spectrum. .This spreading out might be due to varying , resistance. to waves of one type, but of different length,

as in optics. '

, On purely physical grounds it seems difficult to imagine the distortional wave being actually separated from the cornpressional wave. That actual nature would effect this ideal separation seems very doubtful. And so far it is not supported by ,earthquake phenomena admitting of verification by observation on the propagation of waves through om globe.

: 5085

. 432

5.' Accordingly, it appears that the actual propagati6 of waves in solids deserves further study. Our premises s~ frequently are false that the actual facts, in regard to sOlid~ both homogeneous and heterogeneous, deserve more statistical inquiry, in cases where a definite decision maybe attained In his article on Light, Encycl. Brit" 9th ed., § 19, p. 446~ the. late Lor-d Rayleigh says that in such bodies .as jelly tb~ velocity of the longitudinal vibrations is a. large mUltiple bf the. velocity of the: corresponding tr-ansverse vibrations. - Nt} doubt there is some assumed evidence (or such' a statemeht besides the calculations above given, but as no authoritie~ for conclusive experiments of this type are known to me I think a result of such delicacy should be received With great caution.

6. A few cases, however, even if true, are not enough to establish general conclusions; and in view of the difficulty of conceiving how the two classes of waves can be actually separated in nature - one set of waves inevitably tendin. to run into the other - the only safe course is to appeai 'to a variety of experiments, under conditions which may I~ad to an experimentum crucis.

Notwithstanding this uncertainty as to the true order of nature - the theory being not certainly verified bv ' 0), .experiment, - it seems best to examine briefly the chi;f'';O:-

- h '1 di . . d b h . ,We may r

mat ematica con mons impose y t e propagation of waves

in an elastic solid. In an elastic solid, the equation of Poisson '~(m, n in as a sum

'(P(lJ/o/'! = a2y2(lJ (1 JO)

is satisfied by the dilatation and three components of rotation as follows: a = OC(j'dx+ 2P/o).+ ['y/8z (II <!

WI = l/dBrJO)t-Oz/CP) £02 = 1/2(Oa/Cz-Ox/Cy)

£03 = IjdCfJJ'Ox-C)'I'Ca) (I n)

In (

a, p, r being the displacements at any point p (x, )', _zl.

([)=

, where the .. -±m, ±11, for" each' Put Theory 0

In the elastic solid solutions, the components of rotation represent,

002, 008' are connected by the well known relation: But from

owd'Ox+ OW2/o),+ ows!oz __:__ 0 (1 13! with the

and only two of the three sets of solutions are independent If

tides ha Combining these with the solution for 0, we have, in all, .-

priate to

three sets of independent solutions. finest me

Take a rectangular volume of the elastic substance dium so

x = 0, x =r a , y = 0, )' = p, z = 0, z = T> Then at

any dire any time t = 0, (]) = ([)o; and by Fourier's theorem the

medium val ue of ([)o for any point within. aPr may be expressed lost in t

by the following triple summations, which include all positive from One integral values of I, m, " from 0 to 00;

velocity

\"1 expended wave froi tion, disj .with pal changed, and loca can onlj moleC'Ula OWing tc develope V, cOmpresi

1=0r0 l!l=::>O n=::::::o

([)o =.2 2 _2Alm" cos{!rrx/a)cos{1lliT_v/p}cos{nrrz/y)

1=0

?Tl-=O

n=O

1=00 t!l=C"C l~=CO

+ '2 L: 2.; Elm" sin (/n:x/a) cos(m~)/.B) COS(717TZ/r)+· .•

1=0

'~=O

(cf. Lord Ra).zeigh, Theory of Sound, 2nd ed., r896, p. 70).

.The full set of eight coefficients, for all possible arrangements of sines and cosines,

expressions: Aim" = (8/ u-P y)S S S e, cos (1;r xl ex) cos (Ill ll)'/(J) COS (:rt IT z/rl· d.'l.· d)! dz

Blm" -=(B/cx .By)S S S(/)o sin ((ill x/a) cos(mlT)'/p) cos(n;rz/y)· dx d)' ds

are given by the integral (I J 51 (Il

,: .:

2

)agatiou nises so o solids ,atistical ,ttained,

p, 446, jelly the ltiple"of 'us, ~o ltement thori\ie~

to me

< •• ,

ed witn

enough Ii fficulty actually tendipg ) appeal .iay lead

re order fied 'by be '

of waves , Pois's'On

(Il:il

rotation

(I I II

.-.r."

(Ill)

_.;f:

y, z):;

. -:.;"

rotation

ion: ;;

(I ;j)

'.

433

434

Ci-,« = (sj a fJ ylS S S(/)o cos (l tt xl a) sin (/1/ ITy/,a) cos (!lit 4y)' d."!: dy de DIm" = (8/(( fJ ylS S S(/)o sin (In xl a) sin (I/I;-r y/ fJ) COS(I/IT $/ y). dX dy dz EIIIHZ = (8Fa fJ y)S S S(/)o cos (in xl a) cos (711 IT sl fJ) sin (n IT sir) . dx d,v ds Flmn = (8/afJylS S S(/)o sin {In x/a} cos(mIT)'lfJ) sin{ll:Tz/rl' d.'\: dy ds Gl",,, = (S/afJylS S S Wo cos (/nx/a) sin (mn:y/fJ) sin {nnz/rl . d."!: dy ds lilmlL = (S,'afJ y)S SJ (/)0 sin (/iTx/a) sin{m ITy/fJ) sin (n7u/yl . dx dy ds ,

, (117) ( lIS) (I 19) (120) (IZ r) (I22)

As (JJ is a scalar quantity, we may suppose the rate of increase at any time t = 0, to be denoted by o@o/8t, whicb'may be expanded in series similar to that in (1I4), but with accented coefficients, A'lmn, B'lm" , Clm" , etc.

Knowing the initial values of (/) and 0 (/)/'Ot, we may at once write down the complete solution of (I I 0), which is

~asily seen to be: ,1== m== n==

([J =.2: 2 '2 cos (In;I:!a) cos (711 7r),/P) COS (Jl ITz/yl (Aim" cospt+A'tmn sinpt)

1=0 m=o n=o

/== m== n==

+ L: 2:. 2sin'(i:rx/a) cos (IJI iTy!fJ) cos (nITz/yl (Blmn cospl+B'lmn sinjt)+ ...

/=0 ",=0 n=O

I.,

In order to satisfy (JIO) we must have:

r'> a~Jt2(t'!1((2+1JI~/jJ!+,,2112). (1z4)

We may now combine terms which bavethe same values of l, 111, 71 in (123), and thus we find that (/) can be expressed as a sum of terms of the form:

W = 2)Kcos(pt±lITx/et±tIl7rJ/fJ±nm:/r-t) . (JZS) "where the summation is to be extended to all values of ± I" ±m, ±n, and the constants K and £ are of course different for each set of values.

Put in this form, it is clear,- as Jeans remarks [Dyn amical Theory of Gases, 2nd ed., p. 383, 19 I 6), that the solution represents sets of plane waves traveling in different directions. But from (IZ4) it follows that all the waves are propagated with the same velocity a, as in the luminiferous aeth er.

If the elastic solid has continuous character, its pa rticles have dynamically all the degrees of freedom appropriate to the aerher, 'which is an absolute continuum, the finest molecular or atomic -structure ill the universe. _ A me'dium so constituted has the capacity to transmit waves from

any direction. And in case the medium is the ultimate medium underlying the physical universe, no energy can be lost in the movement of the waves, which move incessantly from one body to another, and in free space travel with the velocity of light.

When the velocity of the waves is retarded, energy is ,expended, and pressure developed by the retardation of tbe wave front. Forces of a more intricate kind arise when refraction, dispersion, diffraction, etc., develop, as in the encounter with particles or bodies in which the velocity is suddenly i

_changed, and the wave-field redistributed, so that the density If -,:and local internal pr~ssure .of the aether is altered. But :vel .can only treat of this tOpIC when we come to deal with I ':,.rnoieC'Ular forces, wbich - heretofore have' defied explanation, ,:owing to lack of a kinetic theory of the aether and the un- "

c,_ developed state of the wave-theory. ." ."-

Usually it is assumed that in an elastic solid both ,"Compressional and distortional waves co-exist, though pro-

pagated with different speeds. The two equations of Poisson . thus become:

al = 1-"(k+ 4/3n) (J2([)I'Ot~ = a1 v~([)

for the compressional wave ; \

a~ = V(n/a) ()'!([)/Ot2 = 02,,2(/)

for the distortional wave,

(126)

With most solids the latter velocity U2 is considerably

'smaller 'than 01. the velocity of the compressional wave. In the numerical example of steel above cited, at is nearly twice as large as a2, but it still is uncertain to what extent a real separation of the two kinds of waves takes place. 'In other words, the two kinds of ",'aves are distinct and should be separated, in theory; but it is quite uncertain ~hethe'r this occurs in actual practice, owing to the limitations, of freedom of movement in such, material bodies as we find in nature. There is only one velocity of waves in the aether.

In the case of earthquake waves, there is no evidence of separation of the two kinds of waves, - all the seismographic records being explicable by the unequal velocities incident to mere wave-length, and thus having different speeds

of propagation. '

It, is true that the earth's crust is a very complex structure, and the movement incident to an earthquake involves release of strain, and thus consists of a series of adjustrnents of the quasi-solid Java beneath faulted and mutually crowding blocks of granite some zo miles thick, Perhaps we could not expect distinct separation in such a mass of tremors, partly direct and partly reflected, by the faulted blocks of the crust.

Yet if the two classes of waves actually separated in practice, we ought to perceive two distinct shocks from earth waves incident to explosions, as of powder magazines, _masses of TN. T., and other high explosives, which are powerful enough to be felt at' a great distance, but do not involve complex direct' andreflex actions' in 'the crust, as in the lava

adjustments due to earthquakes,·"' .'

So f~r as I have been able to ascertain there is n o>, , well established record of double waves from such explosions

-",

.. ;":

,a~dth~s' th~ experimental evidence would seem ~erging of the two classes of waves intoone. '."':'/' In the' case of the aether it is certain that only one ;~l;ss'of,'vaves is observed, which in fre'e space travel with .' unifoi'i~l~elocity, as in 'the case of sound in' gases: Accor}tY;.'·diiiglythe aether certainly behaves is a gas; yet its elasticity

. "is so gi'e~t that .waves of any ki~d~ay be transmitted.ias lD.an elastic solid, but apparently the velocity is uniform, 'whether'the waves involve a-rigidity, with sliding of one layer 'over another, or compression, as in gases.

, 10. Geometrical Theory of the Transmission

'of Light and other Physical Forces along Fermats Mi n i m u rn Path, aT- = dJ I/V' ds = o. .

. 0) The problem of refraction in the minimum path.

For any path in space, with radius of curvature (!, and curvature 1 : (!, we have for the length of I he curved path s and the curvature:

.... ;.

".:: . .' .. ,.~; .~: .

On

", r'

... ~ ,.... -. I ., '"

the form of the function /d~pending<' on the

of the parts' of the med_ium.,... . '" .' .....'.:

: Making use of this value ofv in (133) we. obtain~~·:.

, " 01: = S(~/v) (dx dOx+d~;·dOy.+dz doz)/ds ,f:

' .. -:- S(;/v2) ds (dvld~·O'). +'~vldx'~x+ dv/d)"oy+dvj dz·os} . (1 3's)"'

or &T- . [(I/l!) (dx/ds·ox+d_l/ds·oy+dz/ds·oz)J '. '"

. -:::- o;-S( r/v2) dv/d}.:ds;:- .ox S( I/V?) dv/dx·ds -O_l"S( ,Iv?) dV/dXds:_ oz]( I/V2) dv/dz·ds.

The last three integrals of (135), under Hamilton's stationary condition, vanish, because the fixed terminal points make ox, a)'! 6z each equal to zero, The rest of the expression depends on the 'terminal points of the path, and on the wave-length only,

These conditions 'therefore lead to four equations

oT/6x = (r/v) dx/ds O''I:/oy = (l/V) dyjds

chjoz = (l/V) dz/ds 'J·f/rJJ. = - S(I/v2)dv/d}"ds. (r37)

s = ex

N ow the tangent to the curved path ds is defined by the three differential direction cosines, fulfilling the condition':

(dx/ds}2-:t-(d_v/ds}2+(dz/ds)2 = I. (l38)

i where X is the angle between the osculating tangent planes, : and ds is the element of the curve, and e the radius of curvature, for the osculating circle passing through three

consecutive. points. And therefore if we square and ad d the first three equations

The curvature for any path is of (137) we shall obtain

. lie = V[(d~x/ds2F+(d:;/ds2r+(d?Elds2)2J. (128) 1 (OT-/ox)? + (OT-/(l'j') 2 + (OT-/ozj2 = I/V2 _

A nd the direction cosines of the radius of curvature ; (ii) Geometrical conditions fulfilled by Hamilton's

71 = ed2x/ds2 72 = (!d?y/ds2 i'$ = ed2z-jds2• (129) I characteristic function,

Now in refraction, the path. must be consistent with . In ?S~3, ,\:hen only eighteen years of age, .. Hamilton

the principle of least time, and also conform to the principle I obtalOed. insight mto his m~tho~, ,an~ gra.dually introduced of least action. The principle of least time was recognized the consld:rat\o~ of a. characte~lstlc runction A defi~ed by by the Greek geometers at Alexandria, about 300 B. C., in the following differential equation for a single particle of the constructions of Ettklid, [cf. Electrod. Wave-Theory of unit mass,

. '''''Phys, Forc., vol. I, I917, pp. 63-66), but the principle of oA = [dx,/dt'Qx+dy/dt'o)"+dz/dt'ozJ

the minimum path, in simple refraction, was discovered by I • -(dxo!dt'oxo+dYo/dt'oyo+dzo!dt'ozo)+tO'H (140)

Fermat (1601-1665), who found the actual path to conform where H is the constant of the total energy H= T+ V. to the law: T-= llVl+lsv2 . (130) If the moving particle be entirely free, the. seven where the second member is made up of the sum of" two variables in the right member of (140) are independent of terms, each a product of the length of path, I, by the velo- one another; and thus the characteristic function A fulfills

city in that path, v. '. the following remarkable differential equations:

In 'gradual refraction, such as that of light in the 8A/8x = d:"'jdl 'iJA/oxo = - dxo!dt

atmosphere, the direction of the ray changes at every point, GA/o), = d)l/dt oAjo),o = - d}'o/dt

chiefly because of the varying density. And thus if T- be' the

'time of passage, we have the integral GA/'O:;; = dz/dt OA/oso = -dzo!dl

S OA/'OH= t .

T-,= r/v·ds. (1.3J)

Therefore we have at once

And Fermat's condition of the minimum path becomes: (OA/cx}Z+(OA/O_y)2+(oA/oz)? =

o't" = CSIl1,.ds = 0, (132) = (dx/dt) 2 + {d;/dt}Z+{dz/dtF = v2 = 2(H- V) (142)

To bring out the geometrical conditions of the theory ('OA/8xo)z+(&A/o)!o)!~+(8A/ozoF =

of the minimum path; we have to develop the subject some- = (dxo/dt)2+(d)'o/dt)2+(dzo/dt)2= vo2 . 2(H- Vo) (r43) what as outlined in the author's work of 19 I 7· Now it is obvious that if physical forces be due to B! the method of the Calculus of V.'ariations, eqUation/ wave-action, these forces also will conform to the remarkable

(13 I) yields O'T = SI ,Iv' dos -51 Iv?' dsov . (I) geometrical properties of I-Ia711iltoT/'s characteristic function,

I ,. I : 3_3 ! and his analysis will be applicable alike to the propagation

If). be the wave-length, ,It )S ObVlO,US that the vel.oclty I of light, electrodynamic action and universal gravitation,

would be defined by the functional relation . Since the characteristic function A satisfies the partial

v = f(}., X, )', z) (134) differential equation:

.are giver

::-. BA/e

'where a By '. through

"~i· 02

. Si tively, g c~

o dx!dt = . And as . comporJ

.... -:. ~~

r 35),

. ~ .....

13~):

'ton's ' Jints'

, ex,' ,

don'

i byition"

138) :i005

:I,

,:,.:.

ilton'

2c_e~r I by e of

.' .. ~~<

I~· .

146'),'

-V",',

;.,.~.,

~ven It of Ilti.Il~ "

437

it follows that the partial di ffer ential coefficients with respect to t be coordinates represen t the components of the velocity in a motion possible under the forces whose poten rial is V.I And as V is the potential energy of the system, this result is very remarkable; for it assimilates the propagation of wave disturbances, such as light, and electrodynamic action, to the action of universal gravitation, which also fulfills the same condition.

By partial di fferen tiation of_( 144) with respect to the co-ordinates we have

BA/ox' ()2Ajox2+ OAlO),. 02 A/Ox cy+ cAles' oj A/cxf)z = = - a V/cx = x = d~x/dt~ = (d/dt){dx/dt)

'OA/ax- B2A/'Gx3y+oA/'Oy- (}2Aj8y2+oA/oz. 02A/cyfJz = (.45) = - B v/Oy = Y = d2)';dP = (d/dt) (ciy/dt)

vA/ox' CZ AlC:.>; oz+ 'OA/'Oy· C?A/By cz+ (JAjo;. ()2 A/{)S2 -

= -0 vloz -----,--- z - d~z/dt2 = (d/dt) (dz/dt).

Also, differentiating in respect to t, we have

dx/dt· ~?A/8x2 + dy/dt· (J?A/c:c ey+ dz/dt· 02 A/ex as = .. ' (d/dt) (fJAIO~T)

d:r:!dt· 02 A/ex oy+dy/df' 'OtA/B_v2 + ds/dt· ,(FA/ey oz - , (d/dt)(oAlcy)

d:r:/dt· (PA/O:"C Cz + dJ'/df' G~A/ay cz__,__ dz/dt - ?~A/az2 = = (d/dt) (eA/os) .

On compa~ing equations (I4S) and (146), we find that

dx/at = GA/ox d)'/dl ----'-- GA,/o), d;,/dt .'. GA/e:; (r 4 7)

satisfy si mul tan eously the two se is 0 f equa ti ons.

If now we take IX, P to be co nstan ts Vi h i ch mar combine with H to give the complete integral of (I 44), it follows that the corresponding path and the time of its description are given by the equations:

'()AjGa = al (jA/ap = P1 'iJA/fJH = t+ f (148)

where al, /31, e are three additional arbitrary: constants,

By complete differentiation of (14 8) with respect to I, through the' three coordinates x, )" Z, we have at once: f}~AI'iJx 0,,· dx/dt+fJ?A/ByCa· dyldt+ +_G~AFJz'Oa'dz/dt = 0

(}?A/exo/J. dx/dt +'02AjOy o/J. d)'/d/+

+o?AI8zo,8'dzjdt = 0

c2A!C . ..,oH· dxjdt+02A/eyoH' dy/dt+

+a~A/&zoH' dzjdt = o. differentiation in respect to a, p, H, respec-

. Similar lively, gives:

a~AICa Ox' oA/ox+fJ?A/fJa ay- 'iJAjOy+

. +otAlo,Jiz,oAjcz = 0

o?A/c/3 ex· CA/ox+ezA/O,a Oy . (}A/Oy+ +o2A/&fJGz'GA/'Oz = 0 o~Aj(jHcx' fJA/O.T+ fPA/oHOy . OA/cy+ +o2A/oHOz' GA/'az= I •

On comparing these two sets -of equations, we find ,

- . .dX/dt = OA/ex ,dy/d! = GAiay d;/d!' . ,., OA/o;; .. (15 I) "

-And as the first members of these equations represent the

;t?mponents. of the velocity of the moving particle, it follows

5085

that the second members also represent the same thing. Accordingly the proposition stated after equation (144) above is established, and obviously applies equally to light, electrodynamic action and gravitation.

(iii) The physical interpretation of Hamiltoll's analysis points to wave-action.

We have now to consider the physical interpretation of Hamilton's analysis, and we note first that the celebrated function A was invented by Hamilton for the treatment of light, Yet if all physical forces depend on waves, due to vibrations in at-oms, - with equatorial planes lying haphazard, or mutually inclined at various angles, ~ it will apply also to- magnetism, gravitation, and all kinds of electrodynamic action. Hamilton's characteristic function A is therefore above all a wave-function, equally applicable to all the forces of the universe.

. To interpret the above analysis, for the path of light, through a physical medium like the luminiferous aether, we resume the equation

(OT/O'X)2+t{h-jOyF+(OT/(h)2 - r/v2• (I52)

And we see that if we can obtain a complete integral of' this equation, containing therefore two arbitrary constants a, p, in the form

T = F(x, J', z, J., IX, /3)

"

'\

"

then the

derived- equations

2r/,oa = o_F(x, )', s, 1, tc, p)joa ~ (I' 8'rlOp = CF(x, y, z, )., (I, /3}/ofJ = /3'

will represent two series of surfaces, whose intersections give the path of the light in the medium.

As a' and 13' are also arbitr ary. constants,' the four constants [I, p, a', fj' are necessary and sufficient for the purpose of making the two intersecting surfaces each pass through any two given points Po(xo, Yo, zo) and p(x, y, z}_

These Hamiltonian considerations, on simple refraction in non-homogeneous media, show, as was originally found by Fermat, that the actual path is that of least time, as well as that of least action.

, Now in the case of light the physical cause of such action is known to be waves in the highly elastic aether, and propagated with unequal velocities, in different media, according to density, effective elasticity, and wave-length. Increase of density, due to the presence of ponderable matter, hinders the progress of the wave of given length, while increase of elasticity under thinning out of the matter acc-elerates it. Ana in general decreasing the wave-length increases the retardation in velocity.

Equiactional surfaces, orthogonal to the path oflight, are so distributed that the distances between them, for geometrical reasons, are always inversely as the velocity in the

corresponding path. _

Now [it is Clearly shown in the third paper on the New Theory of the Aether (AN 5079), that electrodyr;amic action is conveyed by waves, traveling in free aether with the velocityof light, .and therefore itbese waves will follow the same general laws as the waves of light: Such. a physical' cause necessarily takes the path of least time and .of least action, which is also. that of least resistance to the distur-

! ! j

.. - ....

~ ~.: ,.

439 '

bances'of the medium; And as the motions' of the planets

, conform: to these principles," the question may properly be asked whether any other cause than electrodynamic waveaction- could be imagined to produce, the attractions of the

heavenly bodies.' " ,

- - -" This question' has been dealt with at, some length in the second paper' (AN 5048), and from the additional discussion included in section 12 below it would seem to follow incontestably that no cause other tl1an' wave - action' could explain the> phenomena of universal gravitation.

11. The Ne w Wave-Theory or'Light accounts for a l l' Known Optical Phenomena - Refraction, Dispersion, Anomalous Dispersion, Diffraction, Interference, and the Aberration of Ligbt from the Fix'ed Stars.

(i) The problem of refraction.

It now remains to survey briefly the leading optical phenomena, to see if the new wave-theory of light will explain the observed phenomena as well or better than the oJdwave-theory, which assumes vibrations entirely normal to the direction of the ray, as in the motion of a stretched cord" but does not assume vibrations flat in the planes of

- the equators of the atoms.

, And, first, the phenomenon of simple refraction presents

no difficulty .' For the bending of the light always is due to the unequal resistance offered to the two sides of the wave front, - the one which is more resisted being held back in its advance and the other therefore propagated more rapidly, - and' thus turning the direction of the ray of light towards the denser medium. This reasoning bolds for refraction in water, a, prism of glass, or such a slightly heterogeneous medium as the earth's atmosphere, 'where the air

_ is nearly homogeneous for small distances, yet in the larger problems of the globe arranged in concentric layers, with increased density and refractive power towards the earth's surface.

On the old wave-theory of light this explanation has always been considered satisfactory; and on the new wavetheory it is equally valid, because we consider a beam of light to be made up of an infinite number of independent waves from the separate vibrating atoms. And as each wave is transmitted independently of the rest by the superfine medium of the aether, - just as on a telephone or telegraph wire large numbers of independent messages may be sent at the same instant - it follows that in transmitting the infinitely complex waves of common light, each atomic wave will be refracted exactly as if the others did not exist, and the integral effect after traversing a distance ds will be that all the waves will be refracted in the same direction, owing to the greater resistance on the same side' of their

common wave front. '

Accordingly, the explanation of refraction remains unchanged, while that for dispersion is improved, as shown

below. '

(ii') The phenomena of dispersion, including anomalous dispersion.

In ordinary .refraction, as we have seen, all the rays depending on the waves emitted by the individual atoms,

.- ...

5085

440

~t~·. are bent in the same direction; and thus it is evident th~{ if waves be of unequal Jength,. they will encounter; unequal' resistance, - the shorter waves, owing to their more rapid oscillation, being .relatively more' resisted than the long~t

ones. The result of this unequal resistance is that the waveS_ are dispersed, as in the spectrum, the _longer wa,:es being least refracted, while the shorter waves, innormal dIspersion;

suffer maximum refraction, thereby producing the spectrum effect of dispersion, as in a grating.

, Now however, many separate waves enter a refracting medium, the refractive action on each vibration Occurs as if the other vibrations did not exist: thus we have not merely refraction but also dispersion. In fact dispersion, depending on difference of wave length, seems to imply that the separate atoms, Or same atoms, are emitting not only their OWn distirtctwaves, but in most cases each atom gives quite a variety of these waves, as we see by comparing the table of wave-lengths for the different elements, as sodium, calcium,

hydrogen, iron, titanium, etc. .

The observed phenomenon of dispersion is therefore favorable to the new wave-theory; for we realize from the known phenomena of the spectral lines that each atom has its own several periods 'of vibration; and thus dispersion, or unequal refraction depending on wave Jength, ought to occur.

As for anomalous dispersion the problem is more complex, because the substances gi ving this phenomenon ex' hibit extremely variable effects. But as each atom of a given substance emits its GWn characteristic waves, there is no reason

I 'why the effect of a given refractive medium should affect

I atoms of the different substances in the same way. The proportion of energy absorbed changes with each substance, I and the resistance to each color is a function of the wave-

I· length, but not, the sam.e for all. wave-J~ngth.s, owing to the variable molecular reaction on the passmg hght waves,

. Accordingly just as refraction depends on the wave' length, for homogeneous waves of one color, so also anomalous

dispersion must depend on different resistances' for different colors Or wave-lengths, - due either to the absorptive effects of the substance, by which different' wave lengths are unequally affected, with the thinning out of particular wave' lengths, or to the increased resistance of the substance to certain waves, thus causing them to crowd over into an adjacent part of the spectrum.

In the well known case of fuchsine, 'with the abnormal deviation of the violet rays, by which this color is Jess deviated than the longer red rays, we may suppose the fuchsine to have an inherent attraction for the violet rays great enough to offset its shorter wave length as compared to the red.

Kundt's careful observations on anomalous dispersion showed that it was common in bodies having surface color - Or a different sh;de by reflected light from that given by transmitted light. "Now since in reflection we perceive the colors which are not absorbed, it follows that bodie' presenting surface color, different from that shown by light transmitted through them, must absorb the colon; which ther

- do not transmit. And therefore in transmission the spectrum is deficient', - certain waves being absorbed or taken up by the vibrating molecules, - so as to make possible the ob

abem

'!Jf th: by w. ~'hjch p~en< short, fact j aD ap .{In th confir also plane from diffic

as \H since 'of tt 'of tl: depe feren the desc:

expl: very .diffi< been of t' anal den the: the of t1 with

'of t fore eX3( in :

true -dra-

. and ,rati

44I

5085

442

, '.

onger waves'

being

[sion, ' ctrurri

.r- ..

<erved deviation of the rernammg waves from their arrange- (iv) Stokes' investigation of 1845 harmonizes with the

~ent in the normal spectrum shown by a grating, new theory of stellar aberration.

It would appear from these considerations that the I In the Phil. Mag" 1845, 27·9, Sir Gabriel Stokes at-

phenomenon of anomalous dispersion is highly favorable to , tempted to examine the theory of aberration .so as to find the wave-theory. Unless all molecules emitted and absorbed lout what distribution of velocity may be imparted to the waves appropriate to their own molecular structure, according aether about the earth, without changing the path of the to KirdtllOif's law, it does not seem possible to account for ') rays of light in space. As the new' kinetic theory of the the actual results of observation. The theory that each mole- aether (AN 5044) was not yet developed, Stokes was unwilling cule or atom vibrates in its own period, so as to absorb cer- ! to accept the view' that the earth could pass freely through rain waves in trans?1ission, but reflect othe:s from the surface I· the aether whithout setting it in motion; and he tried to find of it body so constituted, seems to harmonize all known facts the conditions which would leave the observed aberration

in a simple way, I unchanged.

(,.,) Tb bl f iff "' Ii' If c be the velocity of light in the stagnant aether, in

I,ll e prot em 0 di traction, interference, ste lar 'I

a direction whose direction-cosines relative to axes fixed in

aberra non. , I

space are I, 111, 11, and the components of the supposed

" The phenomenon of diffraction consists in the bending I I' f

ve ocity 0 the aether at any point are 11, v, ta ; then prior

of the waves through small apertures and at sbarp corners, h I

by which light is spread around and gathered into fringes i to t e development of the new kinetic theory, with v = 1 2:t' V,

I the velocity of the ray in space at the point in question

which become distinct. The wa ve-theory accounts for the I

phenomenon, under the hypothesis that the waves are very I wou d be V= c+Lu +rn u+rn to. (155)

short, which is fully verified by actual measurements. In I Fermat's minimum path and Hamilton's principle of'

fact for a given width of slit, different colored light gives stationary time, as applied by Stokes, would lead to the geoan appreciable change in the position of the fringes, depending metrical condition

,on the length of the waves in the light used: which obviously' Jz; = JSds/(c-i-lu+1tIv+nw) = o.

confirms the wave-theory, not only as heretofore taught, but To quantities of the first order in (It, v, w)/ c , this is

als'Q. as now modified to take account of waves flat in the I equivalent to

plan~ of the equators of the atoms, The theory of the waves. 0 T = 0 S dSfc~ oS(r/cZ), (tt dx+v dy+w ds] = o. (15,7)

from the individual atoms therefore does not add to the I .. - f ." . •

difficulty of the problem of diffraction in any war. I, !f the ,medlUm ~U:i111S hydrodynamically lrrotatlOnal·

I conditions. WIthout \Vhlrh.ng motion ot the parts en mass, In the matter of interference, the conclusion is similar,

I so that dlD = I( dx+v d_v+ 10 ds = 0 is a perfect differential, as we have already found in section 8 above. This is natural,

the second integral will ,depend on the values of u , u , U!

since the waves from each atom are by hypothesis independent

. . ~ , " . I at the terminal points, and thus will be independent of the

of those rrorn the oth er a toms; and w ha tever tne p OSHlOns I t' in th th b t th th Wh tl h d d '1

f .' , mo Ion 111 e ae er a OU e ear , 'v en 1115 v ro ynarmcar

o the equators, each wave IS transmitted bv the aether m-I c ditio . . t' fi d th th f tl f r I' b t

de endentlv of the waves from the ,othe; atoms. Inter- o~ I I n 15 sa 15 ~: e pa ~ 0 ?e ray 0 • 19 rt, etween wo f< P " k 1 ' b difi d h . . I points whose velocities aye glyen, IS determined wholly bv the erence tcus ta ces pace in t e mo 1 ~ • t eory Just as 111 values of tbese velocities and does not depend on the motion

the older theory, except for the detailed changes already I of th th b t u ints vi t' h f h I' h

described.· e ae er e ~een n.ese poin s in ne pat . 0 t e Ig t.

r '. • , If the terminal points be ."Co, )'0' Z.), and XI, Yl, zl,

In A:'< 5048, p, 183, we have gIven a new anc SImple I d th int " di b til' d ith 'r

I ' .., - an e In ervenmg mernurn e re WI a umrorm

exp anation of the problem of stellar aberration. It 15 so I t f th fiowi ith a 'f I' h'

" . s ream 0 ae er OWlD<Y WI a um orm ve ocrty W ose com-

,very direct and SImple as to be remarkable, In view of the e ts ,;0 th h 11 h '

.'I'ffi I fl' "d' , ' , ' , pon nts are u , v, ta , - • en we s a ave

<11 cu ty e t srnce Bradley s, iscovery ID 17 2 T. which has . 0 e ",

been increased rather, tha~ d.ecrease~ .by the in ve,stig~tions '0 T _:CjSYII7r. ds ~~SY~ ;~~. (u d,x+v dr+ w ds) = 6

of the last half century, It IS surpnsmg that this simple I I . I J . • "

analysis of the problem of stellar aberration has not been :CoY, °0 :Co Yo 20

developed before, It presents no difficulty from the old or "', y, 2, .,

the new point of view of the, wave-theory, but rests wholly on 1 = 0Sds-olr/c' {U(XI ~xo)+v()'; -}'O)+IO(:::t -zo)}] :,., (159)

the motion of the earth relatively to the independent motion I "'oY" ~"

(If the rays of the star, in the moving wave-field carried along But by hypothesis the second term of the-right rnember

with the "earth in its orbital motion about the sun. ,I of this last equation is zero, and therefore we have

I 1

i

I, ~Yo~

I Accordingly the path s obviously is a straig'hf, line, in

, the free aether, from (xo, )'0 I zo) to (Xl) Yl,' ZI ),\vhich,' are the"

II terminal .points of the path. Stokes found that the differentiidly irrotational condition would be fulfilled it the~ether

I· behaves like aperfectftuid for the slow motion' of materia] bodies through it., "

XLylsr

Jr=oJds=o,

" (I60)

Ictin~

o

irs as

lereiy nding seps. , own Jitea'

table ,ciUl~;,

s. wave: . ialous

ferririi effects. .e un·

All that we need consider is the independent motion Dr the rays 0: light relatively to the moving earth. We therefore give the parallel rays of light a common backward motion ,~xactly equal and opposite to the forward motion of the earth

-in its orbit, The diagonal of the parallelogram .gives the .~~rue motion of light relati vely to the moving earth; and by ,.:(Irawlng this diagonal of the' parallelogram we have it direct ,',?n~ perfectly satisfactory explanation of the steJlaraber.,'.T,utlOn,

, ." ..... '

!rii~!~~t~J:!1:'n'~ thio~ ~ith' "th,i (AN 5 044, 50; :;~:'~I: "explained ·00 ., old C~"PtiM:~ of W,:,:4

{"':;,j,:,, \v~'ha~~' shO-~vri: tliat the aether particles. fulfill the law' of transverse ,'fo the direction of propagation" :---~t':'"-rileari:- ,;e-Io'ciry h _ 'lMr V = 4 i1 239 kms/sec. <c-, - . ", '. 3. the extei:~alconjcal refraction mathematically Pf~~" i~:,::~:,\"i:i ;;'A'ccbrdingiy'/,'the earth's motion is only I; j- 5708th part dieted by Sir Tv. R. Hamilton about r 832; and SOOn after'

'''':';'a:the meari~el6city of the particles: And-since the velocity <wards experimentally verified by Llo)ld for aragonite wa's ;~::;~;i'/""6(tb'€'~arth:, is-'~'ery small and nearly uniform, owing, to ,th~ _ found to be definite and decisive; Yet .in examining th6' " ;" . 'circularity of the orbit, it follows' that our planet experiences cusp-ray vrefracriori Llo)'d found that the ~ boundaries were , ' . t: no- secular resistance from the nether." . - . no longer rectilinear, but swelled out in the form of an Oval

',' ":"I\10reo\";er, the earth carries it's aether wave-field with curve s - showing a very gradtlal' diffusion, due to' appre:' it, -,ali arranged in perfect kinetic equilibrium, with, law of , cia ble scattering of light [cf, LI{l_rd's Miscellaneous Papers

density" and wave amplitude I Connected with Physical Science, London, 1877, p. 14;

(J = v r A = klr (J6I) figures i and k).

'~~ten-ding away from it indefinitely. Thus a ray of light from 4. Nearly all the very exact measurements on polarized

'a fixed star enters the earth's aether wave-field as if this light by Lord Ra)'lcigh, Drude, 'Yam;'l, and others bring out medj~m were absolutely stagnant. And under the relative residuary phenomena which show a sensible departure from motion of the rays of light and the moving earth, the stellar the classic undulatory theory (cf. Glazebrook, Physical Optics aberration discovered by Bradley, I7 27, really takes place, London, 19 (4, pp. 355-387). . , just as in the emission theory' of light. 5. In the domain of electro-optics, the Kerr phenomenon

.. For the ray of light from the star pursues a straight directly points to the wave-theory, including the rotation of Hne in the earth's wave-field, and the identical component the plane of polarization by magnetism; and all this is even of the earth's motion forward; but directed backward, may more consistent with the new wave-theory than with the old: be transferred to the moving rays of light before they reach If the poles of an electro-magnet are" polished, and plane

. our globe. Thus, relatively to the moving earth, the rays of polarized light is reflected therefrom, it is found that when light really come from the direction in which the stars appear, no' current passes the plane of polarization is not rotated. and ds is a straight line. . .. If then the current flows in one direction, there is a cor.

This explanation of stellar aberration is therefore geo- responding rotation of the plane of polarization; and the metrically rigorous and perfectly satisfactory. And since in moment the current flows in the opposite direction, and thus the new wave-theory of light, "no change is made in a ray changes the pole to opposite 'polarity, the plane of polari-

- of light as respects velocity and direction, but only as regards zation is rotated in the opposite direction. This is very the internal tilting of the planes of the vibrations from the definite proof of the wave-theory, both for optics and mag' individual atoms, we perceive that the explanation' of aber- , netisrn, for the Kerr and Zeeman phenomena.

ration leaves nothing to be desired. 1 6. The production of elliptically polarized light by

Accordingly it follows that in respect to aberration letting a polarized beam fall upon a transparent' insulator, not the smallest difficulty is encountered in the confirmation such as glass, liquids or gases, under strong electric stress, of the new wave-theory of light. Such entire agre~ment, in _ the region being filled by electric waves rotating in definite such diversified optical phenomena, can have no other mea- direction, as in a magnetic field - was first discovered by ning than that the new wave-theory of light accords with Kerr, and confirmed by Becquerd, Kundt, Rdntgm, Quinckt, the order of nature. . I Lippieh, Du Bois, and others. When the medium is connected

Other phenomena examined under the new wave- with the poles of an electric machine, the waves constituting

theory of light. I the discharge make it possible to produce double refraction,

In addition to the above general phenomena there are as in a crystal, and in Zeeman's phenomena, where the specmany special phenomena which might be used to investigate tral lines are doubled. All these phenomena are found to the nature of light. With this object in view 1 have looked harmonize with the new wave-theory, quite as well as or into a variety of observed data to ascertain if any contra- better than with the classic theory of Fresne].

diction of the new wave-theory "could be established, Or even rendered probable. No such result could be brought out, though I have gone over the principal phenomena in optics and electro-optics.

'J. Polarization in crystals, which presents complex and intricate interference phenomena, and would be likely to offer a contradiction if any existed in nature.

;!.Brewstcr's law, 71 = tg rp, where n is the index of' refraction, 'and rp the angle of polarization by reflection. The partial failure of this law discovered by 'Yamin and others, when rp differs from 550 35' 3 c", seems to point to the new theory rather than the old. It appears that the outstanding residuary phenomena, not in conformity with this law, but yielding maximum polarizing effect when n = tgrp, is not

and presses

, gravitati byexpe "cf"the ' , .are "ali,

the oth'well est

I2. The Wa ve-T'h e o r y of Gravitation towards a Single Body extended to the Case of Wa v e s from Two Equal Bodies by means of the Geometrical Theory of Confocal Conics, in Conformity with the Ob s e r ve d Mo t i o ns of Planets and Comets under the

3 , towards

city of

,ticity, • air in homoge action ( with wr all atoi of the reactiOl

Newtonian Law.

(i} \Vhy the aether remains heterogeneous towards a single body like the sun.,

~ give b en the kinetic to the powerf tnediul from t the W~

'" with tl balanc

I. In our theory of the emission of light and heat waves from the sun, (AN 5044), we have shown that under the spherical expansion of the wave surface in free space, the amplitude of the waves follows the law

A = k/r

'to The mere existence of waves, as 'of light a~d heat; :;_' 'which certainly radiate from the sun with tremendous 'prii. 'energy, - thus necessarily operates t()' make the aether

"',', 'be'terOgeneQus, according to the law' (J = _,yr. _ There is D.,O

pers' ~

<,-;S, d 'oubt of this law holding for light and heat waves; and if 1.4/

?,lt~,-,' gravitational and magnetic waves exist, they too ,will follow

," 'th'is same law. It appears that ,Magnetic Storms .and ,Mag-

'izeif netic Tides', are referable 'only to waves, as shown in my

work on Physical Forces, 19 I 7; and aside from the connection of electrodynamics' with gravitation previously shown

,to exist, it is .fair to ,ask the broad question : ,

, What is the 'pr6b~bility that the fo~cef = A~ = k2/r2 _~o~ld give an appropriate wave -amplitude A =k/r, unless 'gravitational waves_~ls6 exist? Nosuch coinCid.ence could ',occur by 'mere chancel In fact the 'chances agamst such a coincidence occurring 'for all, the atoms of a body in the

p~tential ':;"'" ' , <,."

y 'SSS[(x-'XI)2+'(f-y')2+(~-$I)~1~11,.O'dxd;'d-z, (I64)

'is 'at least infin'it'( of the third 'order (=8) 'to I. "

.. Moreover, ~ince" electrodynamic action certainly is due 'towaves, and these exert-a mechanical action like magnetism 'and gravitation, what' is the chance that there is 'a sudden break in thecon'tinuity of natural forces at 'the boundary which is assumed to' divide electrical action from universal gravitation? Evide~tiY·.the probability is zer~. _ F9r w~ find by experiments on 'all the forces of nature that the doctrines ofthe correlationiof forces and the 'conservation Of energy

, 'are valid, Thus it ·isimpossibJ.eio separate gravitation from the other forces of 'nature, whose electrical character is so

well established.-,.-' " ",

, 3" The aethet 'i~- th'us thinn~d out by ~a:ve agitation,

.·to?tards all singlemasses: and as the aetheronsEave a veloeit)· of 47 12.3 9 krns per second, we perceive that the e las-

,ticity,<e , 68.932 I-!500 ooo-:times' :greater 'than rthat of our

air. in proportion to its density, 'would secure 'an-instant ~6mogeneity6f the aether everywhere 'but for the incessant action of the'recedjng"wave5~-'A.ccordingly the world is filled -with waves, constantlyreceived and constantly emitted from ',all atoms. thewa~;es are' in' Some 'way due to Othe'motions of the aetherons 'which collide with and are reflected by the

're~ctions of the' atoms.:", _. , ; .

- -. ",' 4. Thus ori th'~'one hand,the\e'~edin'g 'waves would gi~e by reaction the central, pressure of, gravitation; and On the other, the resulting heterogeneity of such an elastic' f~~,eti(; mediumalso implies the same central pressure. Owing

't;~,Jl1e enormously rapid motions :of the aetherons they tend, " ,., llyto ber:om'e,eq~aili:d~st,ribiite~t"a:nd Jhusm'aket~e: " "'" .ho_m9geneQu~,!~ut "s:s. "~~ey,aI~,ie~fl~c:teq,,x~o].e,nqy. "the atoms,":--: the collisions. and re,fl~,cti(),riskeeping up .wavss incessantly,~;:---<.,the medium~refi.1ainsheFei'ogen'eous;' radiation;' , the' energy,; oCthe cent;aI-,iprush,,_ ?_(,tll~ ~~tl-jerons j,us.!e:

'the loss. :o( 'energy by ~,the!iYa~>:es ,aw~y.;

. . '. >" j'~~~?.,~~~:~.y<: ' .. ; : ."\.'~~.:.: r." ",", . ·"_f'~.· •

.t.:

'" 4-15

force towards the centre due to the receding waves the square. of the amplitude:

f ' A2 =k2/r~

the form of the law of gravititionobserved in

'-,

..,".

-446

5· In Drude'e Theory of Optics, 19I7,PP. q-9-180,'

(English translation by Mann and .kldlikall), a very rernar- .. kable theorem is drawn from the rigorous .f9rinulation o(

HU)'ghm/principle, as' follows;. , ' ,

»When the origin lies within ,the surface S,

47rSo = S{o[s(t-r/V)/rl/&r.co~(n'r) :'.:

. -(r/r)as{t-_r/ P)/On}dS. [.35J _ (165)

)This equation may be interpreted in the following way: The light disturbance So at any point Po (whiCh has

been taken as' origin) may be looked 'upon 'as "the-superposition 'of disturbances which are propagated with a velocity, V toward Po from the surface elements ,dS of any closed surface which includes the point Po. For, since the elements <of the surface integral [.15J .areTunctions of thearguD-ient _ ' . t-r/ V, any given phase of the _elementary disturbance will ,;'j exist' at Po, -tr. seconds after it has existed at d.S.« '-"'-J

. 6.~It·thus' appears i

that disturbances emana-, . "': '1

ling from PI) towards dS '", in .aconical solid angie dw, may be ascribed to

S disturbances :from' the>

element dS '0'[ the' Same -" conical solid angl~ d.w, ".

"from ianv closed surface about Po ;,J~st ;'a5 the integr<:l of.~he o~tflowing . waves gives 4lTSo,- equa- ,

--tion (165), so also the in- , '

-regral of" the' energy - of '

-- , the in, 'flowing-distur,'bances ,

Fig. 12, Diagram of the disturbance ,so ~

, reflected from the surface i.S are equal, and oppositely about _the' point Po, and thus directed, which proves maintained in perpetuity .. ',. 'the proposition, ". ,--' " '.'"

'i. Tb·e integral oi'the vibra·tions in the s~~p·~ra'te:~soJrc·~s ,.

of the in flowing disturbances dS has to beiakenover th'e,

whole closed surface, and thus the calculation is complicated; : involving .a surface integral .. at-, the'j nterval dt , over 'the'soli d :', :" ,"" angle (0 = 4ir about the point Po .. Arid in 'o.rder tornain- ;_';.,_' "J tain the action ' the integral 'has to be 'renewed at infi'nltely ",j'

short " in tervals, dt, corresp~~di_~~' ;tod~.urface , thickness s :,,: .;': .:' .,~~ ".-'::1

, "_- e-. ,: ~ V = ,4~r2S d~:';:_:" <r-:: :,':-:;'~-'l

Bu~ - ';5' thes~renewed'_inte~ra~~};ave the sa~e;,-~~;i;~e. -- ,<) for the 'inte:~,,?l ~t, ,w~-::n.;~y,;!<;kder:'~ '. _c~?tan:t, ;;";L':<:':./:,;;;~L;>·'"

:" d ~= 47r~2Sd;' ' c' ':- ': ":':(;'67) ,

'-':0 .' _.,:, ",":" ""',,".'>1

, , .. :i:

, ~

!

I , i

!

i ..••• I.~. ~

'j:

I:

;'., H

" ;.

--- I'

-,--. "1:

, '; i

I:

,

infin ity" , 1l1~"lllIlt'U' ,'pi-obi

Accor~irlgly j( tl]e in: ""'" " .th(ethe'! f~cu< be cut at theinstantofstarhng; yet it , integrated effect be inchi

,.~~:~.it::~>{:~~i:I~t~i~i;,t;,!I~~;,~~i;,-~t~~it$K~~~~!:~', *t~~ .. "

finite. systems ,. of. hyperbolas, para,bDlas~ , ellipses, which be' d~5c;jbeci: dependin-g' 0;:;:. tn~' joiijar'~ondltion's,. as

fully _set forth below. " "

"".''',';','?', .. ', """'-' "(;;~~3~~::ipg~t;·· •

g t, , •• i(~:::r-./<;,\:~-.'~~::,::: . ~~~;'.:'~'." '._ ~_ ..... ," .v-, ... ,.',' ~~:"~:: , :" ..

§,n:/.:i;':;;'.-,';·,""':';:'·,'; -r" ,meOo"s' fluctuations show that gravitational 'waves

exist;"and are long enough to traverse the earth's mass, just'! 8:5, similar, gravitational waves, traverse the bodies,.of

'''': Jl1pit'er, Saturn,. etc.:, If also follows that the aether' is ex' ;;"}':5:ejsfy,;elr, ~ne grained, :otperwise t~es~, .refractive ph~no~.ena 2,>'<~;"v()uld:l?_ot~ _be distinctly. realized, SQ as to become senst bl e ;",>,'~')cL".o9}ei:vat~on in'the. effects they: produceion the moon's

?:;,:~.;~Mi8~· ;',:-;', .... '.. "/:,:. .". .

.. ' __ ,,);""\ 9:. The' above, mathematical theorem, relative to ,the t.;,:;,\'}Dw;r(i propagation of the disturbances from 'a closed con<··~';'ceritric· surface' S, with velocity V,equal to the velocity of ;"}>~>, th.~~";~"e~ traveling outward from the centre Po~ will be fuli:"\"filleq by tbe energy flow conveyed through the aetber by ·t_he individual aetherons from any 'spherical surface S= 4nr2. .... It; is 'not necessary that the disturbances So from the elements ~~l'!he"ericJosing .surface dS be waves; they may be stresses dueito' the 'energy of the individual aetherons produced by th'e" heterogeneity. Incident to the receding. waves, and thus

"c()nvergingto· the centre whence the waves come, ;

" . . Accordingly, the above integral (I 6 5) rigorously fulfills

. the geometrical condition for a heterogeneous aether ; it is . ,kept to the 12\\' of density .. cr = p r by the receding waves, ",andthe .aetherons always pressing inward, by 'virtue' of .. this' very heterogeneity, and the enormous elasticity E .. 689321600000 times greater than that of our' air in' pro-

., portion to its density. ., , ..

.. ~ ".'.'... (ii) Phrsi{:al ill~stration of the effects of waves from the two foci of .an ellipsoid, corresponding to a double star with equal components.

."." '.C' ~., "The ··acco~pa.riyil;g wave plate Fig. 13 (Guzileint'n, Les

Phenomenes de la Physique, 1869, p. 182) represents a faint system of confocal conics due to waves receding from two. equal' centres, such' as a double star of equal components:

(a), The confocal hyperbolas represent the: reacting pressures at the ellipsoidal boundary, if reflection' were to take place there, or the inwardly directed stresses fulfilling the above equation' for" 47fSo, under Hr~)'ghms' principle fer

, this. more complex system of -two bodies; instead of, the one cen tral mass already consi d ered,·

(b) Each wave from any centre as it reaches the hypothetical ellipsoidal boundary is met there 15y a wave from the other centre; and in reflection the reaction from the assumed bounding surface is in the direction of the hyperbolas, as shown in the figure. The reflection is perpendicular to the surface of the bounding ellipsoid;' and, whether re; ficcted or. not, the stresses' are along the hyperbolas shown;

" {t} If one offhe bodies be nearly insensible in mass,

, it is obvious that the other will emit practically ali the waves, and the. reaction or reflection IV ould be central, as in the case of a spherical body like the sun.: When there is a: single centre of waves, a comet may be made to move about it iri

.".f

3' hay,

. , focus w , for resi the aetl constan 'dation of the

Fig. 13. The upper' figure is a diagram of the waves propagated from Iwo equal foci. As, reflected from the enclosing ell ipsoidal surface, they produce . the 'confocal hyperbolas normal thereto, The en tire svstem of confocal conics is m ade more distinct 'in the lower figure.

(d) These novel considerations throw a new light OD dynamical problems, and bring the lii\\'s;'or celestial mecba' nics into harmony with the wave-theory. They are therefore of deepest interest in the theories of the motions of bodies. Every' possible' Dlotionina system of two bodies is :ie: counted for, by the effects of perfectly simple waves,' and the resulting stresses in the' aether, towards centrai masses Celestial mechanics thus acquires a hydrodynamical basis:,

no

449

",

s08S

45°

the aether being always subjected to stresses, owing to the effects for the two centres ,are combined as shown by the waves receding from the stars and other bodies of the phy- system of confocal hyperbolas. The system of confocal conics sical universe. , shown in the accompanying illustrations is thus of the highest A very' remarkable comparison may now be made be- dynamical interest.'

'tween the waves from two foci reflected from an enclosing {iii} The wave-theory rigorously extended to a system ellipsoidal surface, and that above given for waves reflected of two bodies, by means of the geometrical theory of con-

from a spherical surface enclosing a single centre. focal conics. .

I, We have seen ,that if the waves emanating from a , We have just investigated the' physical theory of waves ..

single centre be 'reflected, from the enclosing spherical sur- propagated from the two foci of an ellipsoid, and shown

face S ~ 41Tr2, we have the equation {16S}· that very remarkable phenomena may thus arise, As the

z. From this equation it follows that ,if we imagine a theory thus outlined may have great dynamical importance, wa ..... e-field established, in kinetic equilibrium, about a radi- it is necessary to examine the problem somewhat more critiating star, and suddenly enclose -that star by a perfectly cally from the point of view of geometrical rigor.

reflecting surface, S :- 41T r'. the energy near the centre will

I Perhaps it is Dot immediately obvious what' all the

flow outward, till reflected at the enclosing boundary, while '

physical phenomena would be in a wave-field 'about two equal ':that near the boundary will as steadily flow inward" to I

restore the energy lost by the central spherical shells, stars. Yet there obviously is ample assuran, ce 'that 'should

r the wave-theory triumph for a pair of equal binary stars, it

S would necessarily hold for triple and quadruple stars, and

4IT r2 dr. sidereal systems of higher order such. as we find in the;

o globular clusters, . These splendid sidereal systems 'are so

3, And as the velocity of propagation Vis constant, crowded with stars in their inner spherical shells as to attain

we have", r+dr '" R-dR aperfect blaze of starlight ,towards the centre, and thus the

41TSr2dr = 4ITSr2dr. (J68) glory of globular clusters, like M.,1$ 'in Hercules,w'Cen-

r R tauri, and 47 Toncani, is unrivaled by any other objects in

'Accordingly, the loss of wave energy from the centre and its the starry heavens. ' . .

perfect restoration goes on without ceasing, and the motion Accordingly we' recall briefly the geometry of confocal .. ~,'

of the waves thus confined is eternal. conics; in the hope 'of illuminating the wave-fields in sidereal

4. Now in the same' way, let us imagine waves ema- 'systems of high 'order, so much studied by the elder Herschel.'

Dating from two equal foci, as' in the case' of a double star .The equation' of a system of confocal conics in the

"with equal components, and suppose 'both foci suddenly xy-plane is ~2/(a27-).)+Y~/(P+),) =L ' (qo)

, enclosed by a perfectly reflecting, confocal, concentric, eHip-' And for the ~ore general system of confocal cci~i~s, intri-:

'soidal surface: dirnensionaljspace, the corresponding equation is ®

x2j(a'!+).)+y2j(b2+l)+z2j(C2+l) = I, (169) '--.x2j(a2+).)+y2/W+).i+z2/(c2,+).) -:- I, (169)

Then the waves fro~ either focus will vreturn to the From the forms of these equations, we 'perceive that, what

.: ,'other in an interval of time dt, corresponding to the distance applies to the plane of xy,will applyalso to the system of 'I:.;

,;2a, traveled before and after reflection, "in any plane section confocal conoids in xyz. Thus for the sake of simplicity we

,':,ofthe ellipsoidal surface; an'd thus the wave-field about either shall consider the system of confocal conics chiefly in the "!!',

" focus will be perpetual. 'And just 'as the wave-field reflected plane xy, as sufficiently general for .the requirements of our

" . for 'restoration is perpetual, so also the inward stress, from present prcblemjn tri-dimensional space, ". "',,' !

",.the aetber outside the surface, is equal tothe radiant'energyIf l is positive in 'the equation, the resulring icurvs

.. constantly reflected, and thus also' eternal. This is the foun- is an ellipse; but if l, is negative the curve 'becomes',an ·'dation of celestial dynamics; resulting from the new theory hyperbola. The transition from the ellipse to the hyperbola

,4"of the aether, . .' ',' ' ..,' . is explained as follows, \', , '

,'"' s. The inwardly directed system ,of' confocal hyper: -,From the form of (lJo) weperceivertbat the principal

::bolas indicate thedirection of the wave stresses sustained by axes of the curve will increase as). increases, and their ratio'

:J;the ellipsoidal. reflecting surface. And' since if we remove will tend more and more to eguaJityas). increases. Accor-'

>~\the surface, the' waves will proceed . into infinite space,' we " dingly a .circle of infinite radius, (a ,_:_ b = 00), gives the

.. recognize' that a wave-field about the two radiating foci must limiting form of the elliptical' confocals," ,

',' have its equilibrium sustained by tbe inward stress of the 'On the 'other hand,' when .t is negati"ve, the principal

. external aether, which is therefore at every point normal to axes will decrease as ). increases; and the ratio

'~nclosing' _ellipsoidalsurface.~he,exte,r,nal aetheri-tbus '.:,,', ,.. :', ,,:.:~J!., , (b~+~)/(a2~ I.)" ".,' ,<,j 17» ; Ii

its stress 'al~mg the tangents to tii e' systems of con-,' wilI' ~150 "decrease as 1 increases. ' ;.The-"elli pse thus' becomes .:., <,;

hyp~rb91.~s:, ,;"",:'.:c~;,~::,.;.:~,,'::,,;~,,', 'l"~';,:,,:: .: '. ',:, _fia~t~r .a~d~a~t~r, 'un tg l, _~~ .: :q il~l'.~~o _':: bZ, ;~.?~~,,~hf~.in9:~::>.:,,:~ ;,:"f

, 6.' This geometrical description conveys to'our minds axlsvanishes, b2+1 -:- 0; ,and the major axis IS equal to "<\"'~"T very dear 'dyminiicaI .illustrationrof the.behavior ofjthe the idistance between the foci, The curve thus narrows down :~:; '.11 ther about .as'ysiCm of jwo equal s'tars.,The inward stress to the line-ellipse joining the foci, which is a limiting form ',' r;

longer,ditect.~dto each centre separately,' but t~e", total of one of the confoeals,,' "'(,>'f< ~.

, ->: <>. ,;!;~rJl

I,.'

........ ". ~ ~ .

_ '" '_ '::,"t;IJA2.~i!>-~-i;' a~~all _ x'2/(~2+J.)jY'~/(b2+l) "'j.,'.;:,""

. l:ii,'erse 'axis~of the hyperbola is very find"·thisoi~tid:d.-·(q{'t~ii·:~b:~d~ti~~;··-''''e remem

~~<~is.ta.ncebet~'e~·~.~thb .fb.C~;· ~nd t?ecor;i': " .. ' A.2 ' ... ,' at'_:__a2e2,.,. and' put,' _b!.+J. '~" 7j_'." . 'at~a2 e' +

h~e .J?mmg the foci Is.a hmltmg. form of th~, .and·.thusobtain from (I7 3r''><;·'?'" . . ... ;,. _ "- "

;,.lI'~~; ir~:?,:;::~:!!~~~:i;~,J:~I~ t:,~;~~it;;; :f;'":=r;~:;~~r~~; ±;:~rJ ~t~)( ~Ti ;:j1~:,····~ .. ,c;·

hvnerbota spreads its branches more .. widely and. the ve~tex·T.hiS: is'!!: quadraticwith two _roots, both real, but of ~nn~."'.c

, " .. '. . .. ' , ,.' " . ' .. ,. " ... 2' ' . ,

becomes more distant from the 'foci on. the horizontal axis, slgns, and thus there.' are two conics; b, +), :______; +1J' being:.

as. shown,,in. the central part of': the figure. As J. becomes' the ellipse, and b2+J. = -1/ being the corresponding"

, 'greater and greater, the .angle between the asymptotes' of the I hyperbola. ' .

. ,.",,' J1yperpola increases, arid in the limitboth branches coincide ,(b) One conic of a confocal system and !=lDly one will,

with the' axis of y:." ' '. '. "" . . ,touch a given straight line.. . ,

. " ". ,~A,_~ccirdingly,we perceive that by' making J. approach From the equations "

:.--: b~,; we narr~w up the confocal ellipses i1_lto a st~aight line ' lx+my-'- I = 0 x2!(aLI--).) +y2J(b2+J.) = I

'.' JOlllmg the fOCI. And when the change continues still further, .

2 " J. ' . . we find for tangency:

b -t- • = -"I, a very small negative quantity, the curve passe's

frcm jbe straight line joining the foci into another straight I (a2+}.)12+{b2+}.}m2 = I

l!.n.e running ,fr?rn' either. rocus to, ~nfinity, ;vhich. give t~e I which is linear in J., and yields one value of )" correspon.' hne-hyperbola,correspondmg to the internal line-ellipse. The ding .to one confocal conic, and only one, bounding the:' point describing the line-ellipse thus ceases to move between given straight line, This might be tangent to the ellipse,; the 'foci, 'and' returns to the other focus through infinity, or to the hyperbola, but not to both at the same point, be'when the limiting elliptical confocal passes into the limiting cause the hyperbolas always are at right angles to the ellipses' hyperbolic confocal. When J. is .negative and numerically at their intersections.

'greater than a2, the curve is imaginary. '. By subtraction we have from the two equations

",' Let us now return to the above, figures, and imagine I .x2/a2+y2Jb2 =' I. x2J(a2+J.) +y2j(b2+J.) ---:- I

two equa.l.,~:a.ve c.~nt:e:; as from a.double star of equal com-l ,.x'2J[a2{a2+i.)]+y'2/[b2(b2+}.)] = o.

ponents, h~e. r." irgrms ; then obnous.!y we have two equal And as the condition of tangency is

wave-fields, one about each focus, WIth the double system I 'J 2 'JbZ _ '/( 2 .) , '/(b2")' --

, f ' f I' h b 1 '. h . . b . x x a + y y - I X X a + I. + Y Y + I. = I ( Ii 8)

0, con oca fper. 0 as, as sown 10 the a ove figure. Th_e'

entire solid angle about the centre of the confocal ellipses we. sef that (.178) shows the rectangularity of the curves at

is Sl __:_41r. their intersections ..

f;\l But W(!" may split the' system of confocal hyperbolas . (v) Application of the theory of confocal conics to

.--,,"·'inta two equal parts; on either side of the median plane, the motions of comets, as under the wave-theory of physical'

each equal to lJZ.Q = Z7T. , , ' forces. .'"- .

It will be evident on reflection" that all the hyperbolas ,Referring to the. figure given above for the waves from

about the lower focus will curve about the right star .f, just ~wo equal stars, we notice t~at the boundary there represented as in the. case of comets revolving about our sun: and all IS one of the conrocalelhpses; others of greater oblateness about the upper focus' will curve about the left star f'. And are ,shown neare: the .centre of the figurer but the approxithese infinite systems of hyperbolas will include curves of all matrons to the line-ellipses very near the centre are omitted, possible eccentricity, with a perihelion distance less than a, for reasons of clearness.

half the distance between the two foci. ' . It will be found that the spherical waves propagated

from these two centres give the confocal ellipses, and also the confocal hyperbolas, as clearly outlined in' this figure. The independent circles about the two foci are at distances a~, a~, a~, .. ,a~.

A t the boundary the' waves from the two foci are refleeted, with reaction in the" direction of the perpendicular to the surface. Hence we see that the normals at these points of reflection, give the confocal hyperbolas. Accordingly, if waves were traveling with uniform velocity from both foci, and reflected at the confocal elliptical boundary, there would thereby result stresses in the aether directed along the con' focal hyperbolas at the intersections of these two systems, This result of the intersecting system of confocals is very remarkable, since it will hold for ev<;ry point of infinite space, and thus for ellipses and hyperbolas of every possible form, mutually intersecting at right angles, as shown in the figure.

The waves propagated from two equal stars by generating .a doubly infinite system of confocal conics - the ellipses cutting the corresponding hyperbolas at richt angles

. "

- fix the paths of infinite varieties of comets about either

focus, as will be more fullydiscussed below.

"

(IV) Geometrical properties of confocal conics.

(a) Two conics 'of a confoc~l system pass through any given point - one an ellipse, the other an hyperbola. After the above outline this is almost obvious, without further discussion, for if the equation of the original conic be

x2/a2+y2JP = l' (I72)

the equation of the confocal conic is .x2/(a2+J.)+),z/(b2+).) = I.

And it is obvious that this curve will pass through the given point (x' y'), if

"453

'.

454

,

It was established by the researches of Prof. Stromgren, might be greatly extended, but we shall not enter upon the Royal Observatory, Copenhagen, about 19 I O~I I, that it here.

the cornets heretofore observed describe ellipses about In conclusion, it only 'remains to add that in the fifth sun in one focus. II had previously been supposed that and sixth papers I hope to throw some light on the obscure ihe orbits of certain comets were hyperbolic, yet greater physical cause underlying molecular and atomic forces. The refinemenJ of research proved the elliptical character of calculation of the wave-stresses at the .boundary of a liquid :n these orbits; so that they return to Our sun, and thus globule, such as a .rain drop or a drop of dew, will lead, ~e' relics of our primordial solar nebula, as, set forth in' us to the cause of surface tension, constantly acting for the :IIf Researches on the Evolution of the Stellar Systems, generation iof minimal surfaces throughout nature .

. !OJ. II, I 9 10. It is not by chance th at all Iiq uid drops take' th e

If the comets had greater than the parabolic velocity 'lSphericaI form! The geometer may discover therein a great Df movement relatively to OUr sun, v >k V( 1 +m)' V( 2/r)' the I secret of the physical universe!

~aths would be hyperbolas; such orbits, however, are not! If so, this advance will illuminate also the difficult let of record. It is obvious that we can now" interpret the . problem of capillarity, which has already engaged the atphysical significance of the system of confocal conics; in tention of so many eminent geometers. Whence we hope conformity with the observed laws of celestial mechanics" to. attack the subject of cohesion and adhesion, and even md the indicat.ions of the Wave-Theory of Physical Forces"l of explosive forces, which heretofore have appeared even

For example, if a comet with zero velocity were to more bewildering. •

cross the boundary to enter the field about the two foci, in . Mr. W. S. Trankk has laid me under lasting obligathe above wave-figure, the instantaneous stresses to the foci, I tions .by facilitating the completion of this 'fourth paper. on the line of the reflected waves, 'would cause the .comet And Mrs. See's sympathetic interest in' these researches has to pursue the indicated hyperbola, passing through the point II lent a support which often proved so invaluable as to be (x, y). Under slightly modified conditions this reasoning beyond all praise.

Starlight on Loutre, Montgomery City, Mo., 1920 Sept. 6. T, y. J. See.

,.'

-' .

-" .....

"_ : .:.,'

":'-.~ ,. " -

-"r"

: ",.i..-

. _ . . ~' • 1

,".' ",

, .' -,"

-c

"

,;

~,stronom, 1 .... ace rrcnten LU. ',!. l ".l.

-:.,: .:

T.].}. See.

,-,-,-,- ,~.:.,

, ~":' ,~~ ..... ::.: ' .. :<r: .,~:

'/~ .

+ f

i

I

1 . +

! j .. :.

Fig. I _ Illustrruio n of the effect of orthogonal prOieCrlOllt '0)" v vhich molecular motions, in PtJ£ssou's elongated ellipses normal 1;0 the wave front, at different pans of a sphere surface, become mainly transverse to the direction of the ray, at a greut distance from the source. The outer circle is In:1gnined to distinct visibil ity, so as to render the cause of the transverse vibrations in light more obvious to the imagination, as shown also by the darkened areas of the enlarged r~y at the centre of the figure.

"i

'~'.

\

'\

~ ).,

-:...... ,,, .. ~ - .. ~

Fig. 3. Graphical illustration, by means of the shaded portion, of the enormous concentration of light vibrations in the periphery of a beam, under orthogooal projection of the sphere, with Poisson's elliptical paths for the molecular oscillations along the radii from the centre, and, by means of the small factor A,'A, thus making nearly all the vibrations transverse to the direction ol the ray.

!~ 'J

'1

Inhaber"_Georg Oheim, Kiel.

!.;.

\

Fig_ 5- General view of the magnetic field about the earth, with a specimen of the waves to which the field is due, shown On the right. The magnetic needle lies in the hollow of the waves, and thus we see why it points to the north, Gauss' Theory of the Earth's .magnetism corresponds to the wave-theory for the part of the atoms which are lined up in parallel, to produce the earth's magnetic field, about 1'138oth of the whole. The rest of the atoms, 13i9il3Soth, of the whole, give the central action called grn\-itation, but the gravitational wave field is too complex to be. shown in the diagram_

Ic haber GeQri!" Oherrn. K~eL

Fig. 6.

III ustra ticn of the restriction of free vibrations when the wave motion is suddenly resisted at the boundary of a solid body. -Owing to ihe resistance to one side of to e ray th e beam of light is fla tteried like a 'reflected stream of water, and thenceforth becomes pc lar ized , vibrating with greatest freedom and largest arnplitude in the plane perpendicular to the plane of polarizction, as held by Prt{"d.

, ;\.5trOnom. N achri chten Bd. 212.

T.}.]. See. New Theory ot trie Aetner,

)'1'·

~~.'~. FIg. 7.

llh.stra\ion of light pclarized by reflection from the blue sky. The vibrations are normal to the plane of polo rization passing throug h the sun a nd the zenith, and the polarization attains a maximum at a poi n t eq uall y di sta nt from

the zenith.

11".

':;;:

11,,-

-'.,

I~ ,

"'"

I~ ~ •

"11.

o ~_~

. ~ ~. _' ..

:IN,.

fl •..

....

Fig. 8.

Undulatory explanation of the interference of polarized light, when the paths of the aetherons are circles. It will hold for ellipses, and even for straight lines, but such restrictions are not necessarv. In the u pper part of the figure the' wa ve phases differ by 1/,1.; in the lower part the phases concur, and gi"e double intensity. The light and dark bands above correspond to the present position of the wave, indicated by the heavy line, while the arrows show the advanced position of the wave when it has moved to the

right after an interval dz.

Fig. 10.

Illustration of the diffraction fringes due to a rectangular aperture, with the corresponding visibility curve above it, on slightly different scale (.)fiche/Jlm:' The central band of light is nine times more in tense than the first secondary maximum, while the higher orders of bands, all parallel to the sides . of the rectan-

gular slit, are still fainter.

Iahaber GeOrg' Oheiur, Kid.

~ .