]\gt ; .
E. of .
" " The ture of the Antagonism between Toxins and toxins By Dr. C. J. MARTIN and Dr. T. municated by Dr. , F.R.S. VI .
" " Some Differences the Behaviour of Beal Fluids that of the Mathematical Perfect Fluid By A. Conlmunicated by Lord RAYLEIGH , F.R.S. VII .
" " the Heat dissipated by a Platinum Surface at Temperatures By J. E. .
Communicated LORD , F.R.S. " " An Extension of -magnetic Theory of to include Dispersion , Reflection , and allied nomena By DlVIN EDSER , A.B.C.S. Commumcaxl by Captain DE W. ABNEY , , F.B.S. 18 , \mdash ; Read March 10 , TLe Electro-magnetic Theory of Light , as left by Maxwell , gaveo explanation of dispersion , and led to concIusions in some inconsistent with the results of experiments on metallic reflectu .
There seems to be little donbt as to the general direction in would be necessary to modify { hat theory in order to give a factory ac.count of these phenomena .
ction has considered by many to be inscparably connected wibh the : charged atoms , whilst the properties of a dielectric have been foul to admit of an explanation on similar lines ; it vould appear , where fluctuations in the electric fieldf equencies as as those of light are concerned , to formulate the reactions of the atoms or molecuIes composing medium through the turbances are propagated .
mechsnical theories oflight , when modified in a similar manner , : been found capable of more or less compl'ehensive account sion and metallic eHection ; and it would appear that assumptions necessitated case are at least as admissil as which have been made elsewhere .
No doubt a theory etension to finality must include a satisfactory account of of the luminiferous and of icity , even the ultinlate constitution of atoms ; and Mr. 's investigatio we call oven now go in this direction .
On the oth , a less comprehel]ve , depending only on the known oi my prove not without value , if , whilst ed hcnomenn , it us to a mental pictu of processcs lved .
's -macf netic of .
375 what follows , I have endeavoured to extend Maxwell 's so incIude dispersion and metallic reflection .
The initial assumptn that both conducting and dieIectric media sist of molecules , ( :hrisitl ( in the simplest case ) two oppositely charged atoms , iessentially that made by Helmholtz in his paper on the " " Electrognetic Theor of Dispersion * The methods I have employed lwever different , and my final results , bearing a eneral ance to those by Helmholtz , diffel from them in some iportant particulars , besides being simpler and more directly related results of investigations on other theories of the natu of besides the gain in simplicity and consequent physical definiteness , other reasons for not Hclmholtz 's metbod .
The Jity of employing the principle of least action in this case may bc [ estioned , since the only difficulty appears to lie in the wtions due to the charged atoms , and the nature of these ious iimplied in the energy equations assumed .
Moreover , the lations obtained by Helmholtz are themselves open to criticism , : some of the resuIts obtained do not reduce to Maxwell 's equations the terms involving the ] al'ization of the dielectric are equated zero .
2 .
This paper contains ( 1 ) a simple explanation of the fundamental enomena observed connection with a state of steady electrical in a dielectric ; ( 2 ) a consideration of the law of propagation electrical disturbances in a polarizable medium ; equations are tained which dispersion , ordinary and anomalous ; ( S ) a consideration of metallic reflection , in connection th Kundt 's iments on the velocity of light in metals .
State of Steady , lectrical Strain in lectric .
3 .
It is here ] to explain Faraday 's , that the introction of a material dielectric betweell the pl tes of a clmrged ttdense - , diminishes the potential difference belween the plates .
H. von Helmholtz , 'Wied .
Ann lS93 , vol. 48 , pp. nslated by Dr. Howard in the ' ' vol. 37 , 404\mdash ; 408 .
Helmholtz 's of as been criticised by Rieff ( ' Wied . ?
5 , vol. 55 , pp. 82\mdash ; 94 ) and Henv iside ( ' Electrician , ' vol. 37 , lition it would appear that nlcnt of energy ( other Ray leigh 's " " ipation is ible i Inciplc of least tlction ( ( Brit. ' " " Action on Light 1893 ) ) .
len the terms to atio of the nature and ( vector ) ) ensit force resp in 's equations . .
E. dser .
A tension of Let us consider a dielectric medium , other than the ether , tb composed of molecuIes each , in the simplest case , h ) oppositely charged atoms at a definite distance apart .
The voluS actually containt by the atoms is provisionally assumed to be sml in comparison with the volume of the interatomic spaces .
In what follows the ternl axis of a molecule will be applied to vector distance between a negative and its associated positive attJ .
In an isotropic medium the axes of the various molecules will , in absence of ical strain , be inclined indiiferently in all so th any element of volume will have no resultant electric If now a difference of potential bo established between any ) parallel planes in tlJe medium , the positive atoms will all move towal points of lower , and the negative atoms toward points of high potentia ) .
As a result each element of volume will now possess resultant electric moment .
AIso , if we define a molecular electn as the product of the atomic charge ( taken as positive ) in ' the vector distance between a negative and its associated positi atom , the resultant electric moment due to a strained element volume containing a number of molecnles , will coincide in directi with the direction of fall of potential .
In the unstrained medium we may represent the several momen of the molec , ules contained in an element of volume by lines radiati } uniformly from a point , and ending on the surface of a sphere .
* To find the electric moment of an element of the strained it is only necessary to determine the sum of the aIterations of component moleculal moments in the direction of the fall of potenti ' 4 .
Let be the electromotive ensity at a point in the dielectr and let ( 1 be the atomic , the distance between two atoms being .
Then a molecule whose axis makes an angle the ection of fall of potential , will be subjected to a coup tending to decrease and each atom will be subjected to force tending to increase .
Assuming that the forces called into play by ] acement of the atoms are cases proportional to the displacements , } shall indicates the ( infinitesimal ) moleculal .
rotation , and linear displacelnent of either atom in a ] with the axis of lolccule .
increase of the component molecular moment in the fall of potential due to these ains , will be Ugetis]n , ' vol. 2 , S 143 .
's Electro- ?
nagnetic of ting for and this expression becomes , if volume of the element of the dielectric , and if there molecules per unit volume , the number ofmolecules whose inclined at between and will be Hence the resultant moment of the element strained edium will be PMdv .
( 1 ) , . . . . . . . . . . . . . .
The -motive intensity , pal'allel to the direction of the fall ' potential , due to this element of the strained medinm , at a point istant from it in a direction making an angle with the direction the resnltant moment , will be -PMdv Taking the point at which we wish to determine the electronlotive atensity as origin , and the axis of to the direclion of fall potential , we shall have for the electromotive intensity due to ' lab of the strained dielectric , perpendicular and at a stance rom the origin ( if thick-ness of the slab , and hll The only slab of the dielectric which ibutes anything to -motive intensity at the origin is that lying lanes at - and .
The eJectro-motive intensity due to be Hence if be the displacement other than rization of the medium . . . . . . . . . . . . . . .
dser .
Extension of Hence is the value of the specific inductive lred i the eJectro-static system .
In the eleC r system ( 3 ) must be written is the dielectric constant of the ether .
5 .
A few remarks may be made in connection wibh the mnde the course of the above argument .
An objection might raised supposition that two oppositcly charged atoms cou emain in equilibrium at a definite distance apart .
To this it may 1 replied that the form of the argument does not necessitate assumption that electrical forces are the only forces acting ; az even if these were the only forces , it would be possible to account the atomic separation by both atoms to be revolving their common centre of gravity .
If the atoms are to be small conducting spheres , and the volume ualIy occupied by the atoms is a large fraction of tt volume occupied by the dielectric , a correction on the lines of lIosotti-Clausius dielect1ic theory would become necessary .
On otlrer hand , if electrical conductivity is inseparably colmect with the motion of atoms , it vould appea ] .
to be wrong I consider the atoms themselves to be conductors .
Tf finite motions of atoms were necessary to expIain tb properties of a dielectric , a treatment of the problem similar to used by Weber , in his molecular theory of ma , would necessitated .
It may , however , easily be shown that infinitesima atomic displacements will account for the observed phenomena .
the case of two plane and parnllel condenser plates .
Let the -motive intensity at a point between the plates in .
If now , the cbarge same , pure wate ] Jslntroduced behveeu the plates , the -motive intensity will become if the average molecular displacemcllf .
Takiug Cohn and for the specific inductive capacity of water , , 76 , we volts .
Then , since will equal the total callied by the oxygen in 1 gram of water , we shall have .
's Electro-magnetic of Hence the nitude of the displacement , necessary to account .
the dielecbric properties of a mcdinm possessing the highest cific inductive capacity known , is small , even when with nitudes .
determining the force acting on a atom within a larised dielectric : a difficulty arises somewhat to that expewhen the force on a mag.netic molecule within a net is lgh The above method , I think , is not open to vely serions ection , and as the results obtained give a fair explanation of perimental facts , it may , perhaps , be tentatively adopted .
[ Added May .
Larmor gives I as a first approximato the value of the total electric field at a point within a ized , I intensity of ion , or the ment of the polarized medium per unit volume .
As a consequence , Lorenz refraction equivalent is obtained , whilst the relation density ] lows from the .
employed in present vper .
there , a dispersion formula , differing somewhat from Kettcler , is obtained by Mr. Larmor ; it would be to low with what degree of accuracy formula is ] ) able of reprenting the dispersion of tl'anspareut substances , as determined , stance , by Paschen and Rubens .
The somewhat simple ument is submitted injustificaof the assumption made nve , that the field iu a ized edium is equal to P. Let the space veen two and parallel condensel :parated by a distance small in comparison with the area of either , filled with an isotl'opic polarizable medium , ch as that assumed .
The molecules are assumed to be in a state of equilibriulll nongst themselves when the two plates are the same ; nsequetltly the potential energy , due to the small rotations 1 , , . . . .
of the several molecules composing the nedium may presented by a quadratic function of the form A cxpression obtain the potcntial due to he separation of the atoms within the molecules .
conditions onnecting the val.ious coefficients ftrc such is esscnially ositive .
Considering for moment only the molecular rotations , the * See ] ectricity , 'vol . .
S3 , lS92 eclition , footnotc by .
fhomson .
' A Dynamical heory of and Luminifcro ) III Phil. Trans , S18-21 , ) .
) , ' Wied .
, pp. 812\mdash ; 822 ; , pp. .
Edser .
A Extension of of restitution called into play by the rotation will be Sal to , where .
Thus The first term represents the force acting on an om of the mecule whose subscript is unity , when that molecule alone is displao .
; succeeding terms represent the force on the same atom , the displacement of remaining molecules throughout the m A number of relations , similal .
to those given by lfaxwel deduced , but require no further notice here .
Let the conducting surfaces , each of area , and separated ) distance , be charged by means of a battery to a constant lifference V. Let .
be the molecular rotations produck .
If , when electrical equilibrium has been acquired , the rotations are further increased by .
, the electrical energy supplied by the battery be equal to A. V. Forther , writing , we obtain for this energy , the expressions AD .
indicates summation for the whole of the molecules out the medium .
Since , on the molecules , the above of ical energy will be to the battery ( the resistance leads being considered negligible ) , must Since the arbitrary , Hence .
Consequently the acting on a pol } molecule will be equal to that due to a field P. there , if , it is easily seen Similar reasoning will ) the inte -atomic separatio considered .
Consequently the total ical field the molecules the medium is nunlerically equal , and in a direction opposed to , the ] former obtained .
] .
' .
3 , opagation of iElect)Disturl ) Only a very slight modification of 's equations be necessary in order to the law governing the proof electrical disturbances in a polarized medium as previously considel.ed .
bet be the components of the electromotive intensiby at a being the components of the vector potential there .
electromotive intensity due to the displacement of the atoms , if the velocity of the atoms is smft coptl'o( with the elocit ( light , be delived from a . . . . . . . . . . . . . . .
Here it must be remembered that will be a lnction of We may define vector so .
In that case . . . . . . . . . . . .
( 6 ) any element volume , if tak-en as comprise a sufficient mber of molecules , will contain as many as ' atoms .
Let represent the coml)oncnfs of Theu ith quations f Hence .
from . . . . . . . . . . . . . . . .
( of These equatiollS a identical with those used by Professor J. .
Thomson in his developments of Maxwell 's They ths law that owever the elt.omotive intensity at any placs be modified by presence of the atoms , its line integl ound a closed curve will ) equal to the rate of decrease of induction the curve .
These equations differ from those obtained by Helmholtz .
7 .
If are the components of the magnetic force , and of electric current , theIl 4- , The currcnt componcnt ?
will clude the total displacem rent icction of , i.e. , not only that due to the of the duced electromove intensity , bnt that also due to { : continual alteration in position of the atoms roughout the This will be equal to .
1Ir .
has further shown tb in order to close the displaccment current by a charge with a velocity nlllf , t add at the position occupied current element such that its monlent is .
Hence in ) prese case the value of will be given by , where cates the , for unit volumc , of the of the charges into their respe tive }elocities parallel to the axis .
Hence ( S ) may be ) itten A A * in cctricity and ' p. 252 .
The latiuns oined by ] , besidcs diffcl.ing in sigIl from the c quations of lIaxwell , vould 1 to lusion that } displw 1 tho ntoms is circuita ] .
, ' .
Mag , 1889 , p. 324 ; nlso J. J. omson , Pb , 1881 , July , 1889 ; ' lIaxwell 's of Light .
a medium whose permeability is unity , we all finally ) tain , 011inating between ( 9 ) and ( 7 ) , . . . . . . . . . .
( 10 ) .
' .
Consider the propagation , parallel to the axis of , of a of waves , the electrical disturbances being parallel the axis .
reduces to .
( 11 ) .
only remaillS to express the last tel.m on the -hand side in ms of P. onsider a molecule whose axis , as previously defined , an with the axis of .
For plicity , suppose both atoms compsed in the to have masses .
It has been shown by .
Heaviside and Professor J. J. Thomson , *that a charged sphere ving with a velocity small in comparison with the velocity of , has an apparent mass greater iffi true lnass by is the , and is the radius of the sphere .
When charged spheres are moving in the neighbourhood .
a further Tection might be necessary .
Let , then , esent the appal'ent of either atom in the molecule .
The differential equation tor rotational displacement of the molecule will be of the form Here is the molecnlar viscosity , the other letters aving t -ne meaning as in S 4 .
is so small that the time of free rotational vibration is $preciably affected thereby , this equation may be Wl.itten time occupied by complete ) ration .
O. Hcaviside , ' Phil. , ISS9 ; 'Electrical ; J. J. omson , ' Recent ' p. 21 . .
Edser .
An Exte , lsion of As are here concerned only forced vibrations , and fore the pal.ticnlar lntegral of the above equation alone is reqreA may rite the solution the above equation indicates the erse of the operator , and the equation to the atomic vibration .
the the molecule will be of the form Ths ' of his equation may be written here D is inversc of the operator , and .
It has beon assumed that the coefficient of viscosity is the for both kinds of vibration .
the component of ibuted by the two atoms posing the molecule under consideration will obviously be by differentiating with regard to time , the expl.ession and by the aid of ( 12 ) and ( 13 ) , this expreSl']j ecomes employing preciseIy similar reasoning to that used in } ally determine that will be obtained* by differentiat ith to time the expression resellCC of force ions , considered charged , is not of materially affecting tb6 mula f light old follow no less from theoretioal rations t from such as that thoabsorption of ' llot appr roltl that of pnrc 11 ater .
vel.y long old be ] nagnetic Tof value , we obtain ming the operation denoted by on the ) , we . .
( 14 ) . . . . . . . . . . . . . . . . . .
since , with a similal expl.cssion , we obtain as the condition tflat ( 15 ) should be a solution .
( 14 ) Also is the velocity of of in vacuo .
ting this and we . .
Here is acve index of the medium , and ' transparent media , will be small , the ve ay itten .
Mr. .
Edser .
An xtension o Since both and are directly proportional to the molecules per unit volume , it follows that in a medium which compressed without eciably altering and we shall have density .
When is very nearly equal to unity , this may be written cc density , which is Gladstone and Dale 's well-known law .
nrther , the efracbive index for infinitely waves is } from the equation .
is the value ) btained for the specific inductive oapacity oa medium by ( 4 ) .
Further , we may -write ( 15 ) , Also ii this may be written . . . .
, . . . . .
( 1 , ' This is Ketteler 's dispersion formula , which he has shown is capal of replesenting the optical properbieS of a large number of substanc over a range of values of .
Tt is the same as that obtained I .
Glazebrook in his paper ' On the Extension of Kelvir Contractile to include spersion , \amp ; c. , \ldquo ; * and by a altelation form reduccs to Lord Kelvin 's dispersion formula As its properties have been fully discussed in the papers referred { nothing further need be said about it The molecular constitution considered above is , of course , simplest imaginable .
With a more icate imolecu l sillce eac * B. cbrook , ' Phil. December , 1888 .
Lord in , Addrcsscs , 's Electro-magnetic may have its own tion periods , a much more complicated ( tpersion formula be ated .
this connection the ft thab the absorption spectl'a of the pel'manganates of potassium , ium , lithium \amp ; c. , are appears to be very estive .
infinitely quick ations f becomes equal to unity .
If the 1iationf ; discovered by ultimatcly proved to be to tnsverse periodic disturbances of the erhe ] , of very short wave I , this would explain why ction of these tions is duced by matel.ial media .
10 .
The application of ( 16 ) to explain the optical properties of crysand metals is ious .
In a the -efficients of the ; -olving and will depend on the direction of vibration , hich resuIt in varying the direction of propagation of lrther , accordance with a well-known nical principle be taken iuto account , when - is equaI to or In case ( 16 ) may be reduced to the .
is essentially positive , whilst may bc posive or ative .
This is considered sufficient to explnin the ; bical properties of , and of the quasi-metallic aniline dyes lsll in a solid state .
Kundt has pointed out that the velocity of red lioht in a metal is oporbional to the conductivity of that metal .
A stive relation in this ection may be derived from ( 16 ) .
Let xperiments with metallic films and prisms ke show that is small for the mnjority of metals .
We part of the index is therefore to ( 16 ) may also be wrilten in the form .
, Mr. Tomlinson has shown that ' cases the of metals the order of ) , ' .
, ' .
, ' , ) 1 nagnetic Theory of nces .
Hence if the present instance mighI ) identified with the molecular as detormined by Mr. linson , a connectio1l similar to that derived experimentally by KnAt be established .
Calculating by the above approxintg the real palb of the index from Drude 's lllental data , a fair )eement , as to order of magnitude , is the same quantity culated aately .
The agreennt is not , vever , to explain the very accurate tionality between the velocity of light in a metal and the tivity of the la , tter which Kundt 's figules imply .
Since .
howew , Pfluger has shown that the temperature co-efficients of the of li ght and the conductivities are not of the same ol.der of tude , the process employed above may perhaps represent the of the physical counection betweell two quantities to degree of ation .
11 .
-ldded JIay 10.\mdash ; Another relation of some importance cane readily obtained from ( 17 ) .
It is a well-established law , made the basis of exact chemical determinations , that Ie efficient sorption of a solution of an absorbent substan in a transparent liquid is proporbional to the number of ; of the present in tlnit volume of the solntion .
A sime extension of the tsoning formerly used will , for the squaref the refiti idex of such a mixture the value .
( 1 , are the transparent medium , whi .
and refe } to properties previously of I sorbent substance .
Further we may write , i solutions , is independent of the amount jlouring mattelesent , ilst . . . . . . . . . . .
( 19 possessing one ption band , and is to a ffi ( lroximation inddent of he of the solvenb .
photographic Investigation of Spectra of Chlorophyll , .
389 hen is small , , approximately .
is the coefficient of of the solution , and is ional to , { he number of molecules of the absorbent substance per unit volume .
Since it is essentinl that ( 19 ) should a well-marked maximum value for a certain ( that light absorbed ) , additional evidence is obtained support J ; ntroduction of a viscous term into the equations to the atomic i.ations .
le equation for the squars of the refracbive index of a substance crystallized coppel .
sulphate , which possesses marked selective ( hrption without selecbive reflection to any extent , be represented by an equation of the form of ( 18 ) ; , \amp ; c. , referring to the conditions of motion of a parbicular atom , the ons of the .
atoms giving rise to torms involvin Photographic Investigation of the Absorption Spectra of Chloroph .
ill and its Derivatives in the Violet and violet Region of the \ldquo ; By C. A. SCHUNCK .
Communicated by Dr. E. SCHUNCK , F.R.S. Received March 17 , \mdash ; Read March 24 , ) , 4 , 5 .
is well kuown from the investigations of Soret , and , haemoglobin and its coloured derivatives show a characterabsorption band lying between the lines and of the solar trum .
The band has been shown to vary in posit1on between limits ; in some derivatives it is nearer the red , and in othel .
S ( the violet , end of the spectrum , and is of all the blood absorphands the most 1he very near that as been exist Schunck betwcen pbylloporpbyrin a and lJaematoporphyrin ( tive ) , and resemblance of their ptio spectra\mdash ; one lJlay almosb * ' Arch. Sciences Phy .
et Not vol. , p. 322 ; 1 ' Arch. des Scienccs Phys. et Not Dcc .
, .
Soc. Proc vol. 59 , p. 233 .

