.[ 287 X. An Expel'imental Determination of the Values of the Velocities of Normal Propagation of Plane Waves in different directions in a Biaxal Crystal , and a Comparison ef the Results with Theory .
By R , .
T. GLAZEBROOK , B.A. , Fellow ef Trinity College , Cambridge .
Communicated by J. CLERK MAXWELL , M.A. , F.R.S. Received June 19 , -Read June 20 , 1878 .
P.ART I. Prelimiuary.-Prefessor STOKEs 's Report to the British Association , 1862 , with Outline of the Method .
Section !
.-Review of Previous Experiments and Criticism of FRESNEL 's .
IN bis report to the British Association in 1862 , Professor STOKES says : " The exactness of the spheroidal form , assigned by H UfGHENS to the sheet of the wave surface within Iceland spar , does not seem to have been tested to the same degree of rigour as the ordinary refraction of the ordinary ray ; for the methods applied by W OLLASTON ( Phil. Trans. , 1802 , p. 381 ) and MALUS ( Mem. de l'Institut Sav .
Etran .
, tom .
ii .
, p. 303 , 1811 ) for observing the extraordinary refraction can hardly bear comparison for exactness with the method of prismatic refraction adopted for the ordinary ray ; and observations on the absolute velocities of propagation in different directions within biaxal crystals are almost wholly wanting .
" This bas long been recognized as a desideratum , and it has been suggested to employ for the purpose the displacement of fringes of interference .
" It seems to me that a slight modification of the ordinary method of prismatic refraction would be more convenient and exact .
" Let the crystal to be examined be cut , unless natural faces or planes of cleavage answer the purpose , so as to have two planes inclined at an angle suitable for the measure of refractions ; there being at least two natural faces or cleavage planes left undestroyed , so as to permit of an exact measure of the directions of any artificial faces .
The prism thus formed having been mounted as usual and placed in any azimuth , let the angle of incidence or emergence ( according as the pr.ism remains fixed or turns with the telescope ) be measured by observing the light reflected from the surface , and likewise the deviation for several standard-fixed lines in the spectrum ; ea.eh observation furnishes us with an angle of incidence and the corresponding angle 288 MR. R. 'l ' .
GLAZEBROOK ON PLANE W .AVES IN A BI .
AX.AL CRYSTAL , of emergence , the angle of the prism being known .
For if \lt ; p be the angle of incidence , D the deviation , i angle of prism , t/ J angle of emergence , D=\lt ; f , +"1-i .
\#183 ; .t/ J=D+i-\lt ; f , " But \#183 ; without making any supposition as to the law of double refraction , or assuming anything beyond the truth ef HuYGHENs 's principle , which , following directly from the superposition of small motions , lies at the base of the whole theory uf undulations , we may at\#183 ; once deduce from the directions of incidence and emergence the direction and velocity of propagation of the wave within the crystal .
For if a plane wave be incident on any plane refracting surface , it follows directly from HuYGHENs 's principle that the refracted wave or waves will be plane , and if \lt ; p be the angle of incidence , \lt ; f , ' the inclination of the refracted wave to the surface , V the velocity of propagation in air , v the wave velocity in the crystal ~in cp'_ sin cp ' V V " And if \lt ; p1 t/ J1 be the inclination of either refracted wave to the faces of our crystal prism we have v sin \lt ; p= V sin \lt ; f , ' v sin t/ J= V sin t/ J ' \lt ; p ' + " 1 ' = i adding and subtracting ( 1 ) and ( 2 ) and remembering ( 3 ) , we get respectively or or v sin \lt ; p_++ cos \lt ; p_-_"_V= V sin i cos \lt ; p_ ' - _"f_r ' \lt ; p + \#183 ; '~ v cos __'_f ' sin .
, J.,.,._ -\#183 ; _''~f ' i , J..f -\#183 ; '~ ' = V cos - sin _'l'_ 'f ' " By division tan \lt ; /J++ cot \lt ; p-"fr = tan i cot cp ' -+ ' tan\lt ; !\gt ; ' ; + ' tan~ tan \lt ; p 2 + cot \lt ; /J~'t .
" Equations ( 3 ) and ( 6 ) determine \lt ; f , ' and t/ J ' , and vis known fron1 svin \lt ; p = sin cp ' - v sin " fr sin " fr ' ~ = -v- '- .
( 1 ) ( 2 ) ( ~1 ) ( 4 ) ( 5 ) ( 6 ) MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAX .
A.L CRYSTAL .
289 " Hence we can find the velocity of propagation of the wave , the normal to which lies in a plaue perpendicular to the faces of the prism , and makes known angles with those faces , and hence with the crystallographic axes .
" In accordance with these suggestions , I undertook a series of observations at the Cavendish Laboratory , Cambridge , which I propose in the present paper to describe ; I also wish to discuss the results arrived at , and to compare them with those deduced from FRESNEL 's and some of the rival theories of double refraction .
But previous to this it seems natural to devote some space to the consideration of the experiments that have already been made with the same object .
These we may divide into two classes : ( 1 .
) Those which have reference to Iceland spar and other uniaxal crystals ; ( 2 .
) Those jn which biaxal crystals were used .
BREWSTER ( Thirteenth Report of the British Association ) proved conclusively that one wave in Iceland spar obeys strjctly the ordinary law of refraction .
SWAN ( Edinburgh Trans. , vol. xvi .
, p. 375 ) obtained by direct measurements with prisms placed in the position of minimum deviation values of the ordinary refractive index , which differ at most by \#183 ; 00002 .
To verify his construction for the extraordinary ray , HuYGHENS himself made but few experjments , and it was not till 1802 that WoLLASTON ( Phil. Trans. , 1802 , p. 381 ) attempted to test the theory with anJ degree of exactness .
In 1810 MALUS ( " Theory de la Double Refraction , " Paris Mem. des Savants :Etrangers , tom .
ii .
, p. 303 ) undertook a series of experiments with the same object in view .
Rather later , BroT undertook the same task , while more recently RuDBERG ( PoGG .
Ann. , vol. xiv .
, p. 45 ) and MASOART have measured the values of the principal indices by means of prisms cut parallel to the axis .
Since 1862 Professor STOKES has applied the method indicated above to prisms of Iceland spar , and finds the results of experjment agree with HuYGHENS 's construction with a possible difference of \#183 ; OOOl in the values of the refractive index .
The details of the experiments are as yet unpublished .
( 2 .
) We must consider next the case of biaxal crystals .
Various experimenters have determined the valnes of the principal refractive indices for different crystals .
FRESNEL alone has endeavoured to verify his theory by experiment ( ' CEuvi\#183 ; es Complete es de FRESNEL , ' tom .
ii .
, p. 415 ; second supplement to th0e " Memoir sir la Double R efraction " ) .
The method adopted was to observe the displacement of the fringes of interference , formed from two parallel slits , produced by introducing two plates of topaz of the same thickness cut in different directions from the same crystal into the paths of the pencils proceeding from the two slits respectively , and to compare this with the displacement calculated on FRESNEL 's theory .
The first set of observations corn bined with the angle between the optic axes serves to determine the principal velocities a , b , c. a being assumed to be unity , let b2=a2-/ 3=l-/ 3 o2=a2-y=l-y Then f3 =0\#183 ; 00338 y =0"01222 y-/ 3=0\#183 ; 00884 In the second set of observations the plates are inclined to the incident light , so that the light when in the crystal passes along an optic axis .
To calculate the displacement of the fringes on FRESNEL 's theory , it becomes necessary to know the velocity of the light in the crystal ; this is given by v/ = l-y sin2 X X being the angle between the direction of vibration and the normal to that circular section of the wave surface whose radius is a or unity .
Hence neglecting y2 sin4 X , \amp ; c. Theory gives for the displacement measured in wave lengths N=l2\#183 ; 73 Experiment gives N=ll\#183 ; 87 For the other two \#183 ; waves theory gives N=46\#183 ; 03 Experiment gives N=45\#183 ; 49 In the third experiment the light was incident at about 60 ' .
In this case , \lt ; p being the angle between the direction of vibration and the normal to the circular section radius b v ?
= 1-/ 3 cos2 \lt ; p-y sin2 \lt ; p and t0 the same approximation Whence theory gives Experiment gives N=20\#183 ; G3 Thus in each case the agreement between theory and experiment is fairly close ; but we must remember that the quantities to be measured are very small , and that in obtaining the theoretical results , small quantities have been neglected , without showing MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
291 the effects they would produce , and which , though small , may be sufficient to constitute the difference between FRESNEL 's and some rival theory .
Moreover , the values of the principal refractive indices in topaz are so nearly equal that any conceivable wavesurface will differ but little from two spheres .
To estimate then the weight which we may attach to these results , let ns see how tar the experiments are consistent with one of the rival theories .
Now the equation to the surface of wave slowness in the theories of GREEN and C~.\..UCHY involves more constants than the three principal refractive indices of the crystal .
To determine these we must make the principal sections agree with experimental results in more points than the extremities of the axes ; and it is probable , therefore , that they wm differ but little throughout .
In fact , the existence of these unnecessary constants is a radical defect in both these theories .
Lord RAYLEIGH , however , has proposed a theory in which the only constants are the principal velocities ( Phil. Mag. , vol. 41 , series 4 , 1871 ) .
The shape of the surface of wave slowness is also considerably different from that given by FRESNEL 's theory .
Its principal sections consist of a circle and the inverse of an ellipse instead of a circle and an ellipse .
He supposes " the density of the ether in a crystal to be a function of the direction of displacement , " while the forces called into play are the same as in a homogeneous medium .
From this he deduces as the equation to determine the velocity of a wave front , of which the direction cosines of the normal are l , m , n , the equation a , b , c being the principal velocities .
Let us find the velocity for a wave perpendicular to the plane zx , making an angle a with Ox , l=- sin a , m=0 , n= cos a. or since a= l , c2= 1-y neglecting y2 , \amp ; c. V2 { ( 1-y ) cos2 a+ sin2 a}= 1-y V2= ( 1-y)(l-y cos2 a)-1 V = ( 1-}y)(l +\#189 ; Y cos2 a ) V=l-\#189 ; Y(l- cos2 a ) = 1 - il y si .
n~ " a But replacing a by x , this is exactly the form found above for v1 in FRESNEL 's second experiment .
Now let us take a wave perpendicular to the plane yz inclined at an angle \lt ; p to Oy .
Then l=O , m= sin \lt ; p , n= cos \lt ; p , and the equation for V2 becomes b2 sin2 cp c2 eos2 cp y2_b2 + y2_c2 0 V 2(b2 sin2 \lt ; /l+c2 cos2 \lt ; /l)=b2c2 or V2(1-/ 3 sin2 \lt ; p-y cos2 \lt ; p ) = ( 1-/ 3)(1-y ) = 1-(.B+y ) to the same approximation V = { 1-\#189 ; ( .B+y ) } { 1 +\#189 ; ( / 3 sin2 \lt ; /l+y cos2 cp ) } = l-\#189 ; ( / 3 cos2 \lt ; /l+y sin2 cp ) and this agrees with the formula used in FRESNEL 's third experiment .
Thus FRESNEL 's experiments afford as much a verification of Lord RAYLEIGH'S theory as of his own , and are , therefore , an insufficient test of the truth of either .
They are , however , the only attempts made to verify the theory .
Section II.-Description of the Apparcitits used and Method of making the Experiments .
My own work was undertaken in the hope of obtaining results sufficiently accurate to decide between some of the various theories which have been propounded .
I proceed to describe the apparatus used .
The crystal itself was a piece of arragonite , very clear and of a light colour , obtained from Germany by the Demonstrator at the Cavendish Laboratory , Cambridge , through HILGER , of Tottenham Court Road , London , W.C. Originally it was in the form of an hexagonal prism .
Four of the faces were those marked m in Professor MILLER'S mineralogy , the other two were the faces marked a. The natural ends had been cut off , and the artificial ends were approximately perpendicular to the m and a faces .
The crystal consisted of several twins , the twin planes being parallel to two of the m faces .
At this stage Professor STOKES kindly undertook to examine the crystal , to determine in which direction to cut it with the greatest advantage for the purposes of the experiment .
He found that fair reflexions were obtained from three of the m faces , and decided that they should be left untouched to determine the position of any other cut faces .
Again , the mean axis of elasticity bisects the angle between the contiguous m faces ; MR. R. T. GLAZEBROOK ON PLANE WAVES IN A .BIAXAL CRYSTAL .
293 the optic axes lie in the plane perpendicular to this axis .
A prism formed with its refracting edge nearly parallel to the axis of b , would permit of observations being made in a plane passing nearly through the optic axes ; and it was fonnd that , by inclining the faces so that the axis of ci nearly bisected the angle of the prism , and making that angle about 42 ' , light could be passed through the crystal in the direction of either optic axis .
It was therefore thought right that the prism should be so cut .
Moreover , by polishing one of the a faces and cutting a face nearly coincident with the fourth m face , a second prism was formed with its refracting edge nearly parallel to c. It was important to get observations in different zones from the same piece of arragonite , for its chemical constitution is more or less variable , and the constants-i.e .
, the principal refractive indices--may vary in different specimens .
The work of cutting was successfully performed by HILGER , and the faces of the prisms were very fairly plane .
The crystal being thu8 cut , it remained to determine accurately the position of the cut faces with reference to the crystallographic axes .
For this purpose , and throughout the observations , I used a goniometer , by GRUBB , kindly lent me by Professor STOKES .
The collimator is fixed to a graduated circle of about 9 inches diameter .
The graduations , which are in silver , are on the flat face , not the limb of the circle , and the circle is divided to arcs of 10 ' .
Attached to the reading telescope there are two verniers ; the graduations on the verniers being hi the same plune as those of the circle ; and by means of these the circle can be read to angles of 1 O ' ' .
In the centre of this circle there is a table to hold the prism ; the table being attached to a second graduated circle with verniers which can be read to half minutes .
It was requisite throughout the experiments to have the edge of the prism parallel to the axis of revolution of the telescope .
To attain this the prism was attached to a stand , adjustable with set screws , which rested in slots on the table of the goniometer .
The reading telescope was fitted with a needle point instead of the usual cross wires .
In order to level the prism , I covered half the slit , so thnt when viewed directly the line of junction of the covered and uncovered portions coincided with the needle point .
I then moved the telescope until the images of the slit , formed by reflexion at the faces of the prism , were brought in turn into the field ; .
and altered the levelliug screws till the same coincidence was obtained .
Homogeneous light was secured by using a BUNSEN'S burner ~ith a platinum wick , one end of which was immersed in a saturated solution of salt-and-water .
By the advice of Professor STOKES , the observations were made at angles of incidence increasing in arithmetic progression .
The direct reading was taken several times during each day 's work to avoid error .
The method of observing was as follows :--I set the telescope to a given scale reading , and turned the prjsm till the image of the slit seen by reflexion at one of its faces coincided with the I).eedle point .
The difference between the direct reading and the known one to which the telescope was set gave the supplement of twice the angle of incidence , whence the angle of incidence is easily obtained .
I then turned the telescope till one image of the slit seen by refraction through the prism coincided with the needle point and took the readjng .
The difference between this and the direct reading gave the deviation for that wave .
The observation was then repeated for the second wave .
I then reset the telescope , so that its axis made an angle of 4 ' with its former positjon , and took a similar set of observations .
This gave a series of deviations corresponding to angles of incidence increasing by 2 ' .
Just in the neighbourhood of the op~ic axes , observations were made at every 1 ' of incidence .
For the first prism , each observation in the set finally chosen was repeated three tinies , and the mean of the results ta ken .
The difference between this mean and any observation was only in a few cases as great as 1 o'/ .
In the case of the second prism , the observed values of the deviation on different days for the same angle of incidence agreed so closely , rarely differing by more than 20 " , that it seemed unnecessary to take a third set .
I found also that the refracted images were much more distinct for angles of incidence greater than that for minimum deviation , than for angles less .
On passing through the position of minimum deviation , I therefore reversed the prism , so that the face of incidence became that of emergence .
Fig. 1 .
Thus , if A B C D be the ray undergoing minimum deviation , for a ray such as PB QR , I made P B the incident ray , QR the emergent ; whjle for a ray like SB TU , I reversed the prism , so that U T became the incjdent , B S the emergent ra.y .
The cause of the increased distinctness lies in the fact that as the angle of incidence increases from that giving minimum deviatjon , the breadth of the ima.ge of the slit decreases , and i , ice versa .
MR. R. T. GLA .
ZEBROOK ON PLANE W.A.VES IN A BIAXA .
L CRYSTAL .
2 !
)5 Section III.-Deterniination of the Position of the Principal Plane of the First Prism with reference to the Crystallographic Axes.-Values of the Quantities observed in the Experiments , and Calculation of the Reciprocal of the TiVave Velocity for the First Prism .
We proceed now to the measurements made to determine the position of the principal plane of the first prism , of which the faces are P P , .
Fig. 2 .
Let A B C be the points in which the axes of the crystal cut a sphere , with its centre at the origin . .
Let M M ' be the poles of the m faces of the crystal , M M ' lie in the great circle A B , and the angle M M/ is known ; also B M=B M ' ; let P P , be the poles of the cut faces of the prism .
Let PM=0 , PM'=0 ' P , M= 0 , , P)\tl ' = 0 , ' Let a , / 3 , y be the angles which O P makes with the axe : : ; .
Let MA=fL\#183 ; From triangle P A M , cos PM= cos AP cos AM+ sin AP sin AM cos PAM or From PAM ' , cos 0= cos a cos fL+sin a sin fL cos A cos fi ' = - cos a cos fL + sin e.t sin JJ .
cos A .\#183 ; .
cos 0- cos a cos fL= cos 0 ' + cos a cos fL cos 0- cos 0 ' ... CJ S a=- 2 cos\#181 ; - .
0+0 ' .
0'-0 SIU - 2 - Slll 2 - cos\#181 ; ( 1 ) Again , from triangle P M B , cos PM= cos BP cos BM+ sin BP sin BM cos PBM or and from P B ' M , cos 0= cos f3 sin 11-+ sin / 3 cos / L cos B cos 0 ' = cos f3 sin / L - sin / 3 cos / L cos B .
, 13_ cos 0+ cos 0 ' . .
cos - 2 .
sm\#181 ; 0+0 ' 0'-0 cos--cos-- sm\#181 ; Now the angles 0 0 ' were determined by observation .
Each angle was observed six times on two different days .
The mean of the results was , 0 = 77 ' 18 ' 44 " 0'=100 ' 43 ' 3011 Some difficulty was experienced about the angle / L\#183 ; ( 2 ) Professor MILLER gives it as 58 ' 5 ' ; but this , when substituted in the aboYe formulae , gives a. value for P P , the supplement of the angle of the crystal , which differs considerably from the measured angle .
I therefore measured the angle / L and found it to be 5 8 ' 30 ' nearly .
Mr. GARNETT ( the Demonstrator at the Cavendish Laboratory ) kindly communicated this result to the Rev. H. P. GURNEY , of Clare College , who was then lecturing for Professor MILLER , and he measured the angle with very nearly the same result as I had already arrived at .
Taking , then , the value he obtained , and assuming as before that O B bisects the angle MOM ' , we have 11-= 58 ' 28 ' . .
, : whence from the formulm ( 1 ) ( 2 ) a=67 ' 10 ' 35 " / 3= 88 ' 52 ' 4011 For the position of P , we have similarly , 0 = 80 ' 39 ' I 0'=101 ' 16 ' 23 " I Whence , a == 69 ' 59 ' 20 " I / 3 , =91 ' 6 1 30 " ( 3 ) ( 4 ) ( 5 ) ( 6 ) * This value of the angle fl is given in DANA'S 'Mineralogy , ' for specimens of aLTagonite coming from Silesia .
The piece used was obtained from that ueigltbonrhood .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
297 To determine the angles BA P , BA P , , the triangles BA P and B ' A P , give ( since AB= i ) whence similarly from B ' A P , From triangle P A P , cos f3 cosPAB=-.s1n a BAP=88 ' 46 ' 4011 B'AP =88 ' 49 ' 10 " ' Fig. 3 .
cos PP , = cos AP , cos AP+ sin AP , sin AP cos PAP , Whence pp =137 ' 9 ' 5511 I The mean of numerous experiments gave pp= 137 ' 9 ' 3011 ' The smallness of the difference affords a strong presumptjon m favour of the measurements .
Let P P , cut A C ' in L. Then .
APP-sinPAP , .
AP SJn - .
pp Sln ' Slll I AP , P=0 ' 3 ' 23 " .
Whence Again cot LP , sin AP , = cos AP , cos AP , L+ sin AP , L cot LAP , .
\#183 ; .
LP = 67 ' 27 ' 25 " .
' .
ALP sin AP , sin LAP , Slll = -- .
LP Sln , ALP=l0 12 ' 4 " ( 7 ) ( 8 ) ( 9 ) .
AL sin AP , sinAP , L Slll -sin ALP , AL=2 ' 31 ' 58 " .
( 10 ) The solution of the triangle A L P leads to the same results .
We now come to the experimental determination of the deviation and the calculations of the refractive indices .
Throughout the work for this prism , \lt ; p represents the angle between the normal to the wave incident on or emerging from the face P , when in air , \lt ; /J ' the corresponding angle in the crystal , while l/ ; y ; ' are the angles for the face P. I propose then to give a table of the values of \lt ; p or l/ ; according as the wave was incident on P , or P. ( 2 .
) Of D+i , i being the angle of the prism , D+i is given because the expression cf.\gt ; :+ , which is equal to D ; \ occurs in the formula , and it saved trouble to calculate D+i at once from the readings .
( 3 .
) Of the values of \lt ; p ' deduced from the formulae tan cf.\gt ; '---t ' = tan i_ tan p-t_ cot cf. , + , fr C 2 2 2 2 proved in the first section and \lt ; p1 + y ; ' = i. ( 4 .
) Of the values of \#181 ; , the refractive index as the mean of the results , given by the two formulae sin \lt ; p = \#181 ; , sin \lt ; /\gt ; ' sin y ; = \#181 ; , sin y ; ' Any difference in the two values of \#181 ; , found from these two formulae was due only to the errors necessarily arising from the use of proportional parts , and rarely exceeded \#183 ; 00001 .
I also introduce a list of errors in the calculated quantities which would occur from errors of 1 O " in each of the observed quantities , taken so as to produce the maximum effect in the result .
This list was obtained by setting down from the tables the errors to which the supposed errors in D and \lt ; /\gt ; would give rise .
The numbers in the table of errors for\#181 ; , are the digits in the fifth place of decimals .
The next page gives the complete work for one observation as a sample .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BI .
AX.AL CRYSTAL .
299 Workfor ; \#183 ; educing one Observation .
Direct reading 263 ' 39 ' Re-fl exion reading 180 ' 5 ' Deviation reading 297 ' 34 ' 12 " i=angle of prism 42 ' 52 ' 30 " \lt ; /J + 'fr Lcot~ 10\#183 ; 1012497 \lt ; /J ' - , y , ' -2- 4 ' 54 ' 30 " L sin \lt ; /J 9\#183 ; 8725466 L sin , y , 9\#183 ; 6792905 fl from \lt ; /J 1\#183 ; 68123 4\gt ; =\#189 ; f 180-(Direct-Reflexion ) } =-l(180 ' 5'-83 ' 39 ' ) =48 ' 13 ' .
Error-5 " .
D+i=D-220 ' 48 ' 30 " =76 ' 45 ' 42 " .
Error+ 10 " .
tf ; =D+i-4\gt ; =28 ' 32 ' 42 " .
Error+ 15 " .
cf , +"fr =D+i=38 ' 22 ' 51 " .
Error+5 ' .
cp 2 'o/ = cp~"fr -tf ; = 9 ' 50 ' 9 " .
Error-10 " .
1 , L tan 2=9\#183 ; 5936354 .
------ Error .
L tan \lt ; /J- 'Y ' Error .
L tan cj/ - , y , ' 21G 9\#183 ; 2389842 -1252 8\#183 ; 9338693 \lt ; /J ' , y , ' -4 ' ' 26 ' 19 ' 45 " -4 " 16 ' 30 ' 45 " L sin \lt ; /J ' log f. -94 9\#183 ; 6469205 -170 \#183 ; 2256261 - - L sin 'fr ' log 1 .
580 9\#183 ; 4536616 284 \#183 ; 2256289 fl from , y , mean -3 1\#183 ; 68124 +15 ] \#183 ; 681235 Error .
-1036 +4 " -76 TABLE giving results of experiments on first prism in the first positioD .
Outer sheet .
Inner sheet .
Table VI .
Table V. Table IV .
Table I. Table II .
0\#181 ; 2\#183 ; ftz- oq/ .
q/ .
D+i .
q\gt ; .
D+i .
( / \gt ; ' .
ocJ\gt ; ' .
II 0 I II 0 I " 0 I 0 I / I 0 I / / , , 9 1\#183 ; 68500 6 16 17 45 77 4 20 28 13 76 50 16 16 20 2 6 1\#183 ; 68540 17 22 26 76 39 45 30 13 76 24 fi3 17 25 8 1\#183 ; 68568 18 26 14 76 21 43 32 13 76 6 30 18 29 16 1\#183 ; 68578 19 29 9 76 9 10 34 13 75 54 30 19 32 21 1\#183 ; 68589 20 30 55 76 2 30 36 13 75 47 30 20 34 27 5 1\#183 ; 68568 5 21 31 50 75 59 46 38 13 75 45 51 21 35 22 5 1\#183 ; 68558 22 31 23 76 2 26 40 13 75 48 58 22 35 3 1\#183 ; 68517 23 29 57 76 8 41 42 13 75 56 28 23 ; 33 29 1\#183 ; 68467 24 27 12 76 19 8 44 13 76 8 43 24 36 22 1\#183 ; 68421 25 22 58 76 34 3 45 13 76 24 55 25 25 56 3 I 1\#183 ; 68376 4 26 17 11 76 53 13 48 13 76 45 42 26 19 45 4 !
1\#183 ; 68302 27 10 5 77 15 40 50 13 77 10 5 27 12 4 I 1\#183 ; 68226 28 118 77 42 10 52 13 77 38 48 28 2 33 I 1\#183 ; 68190 28 26 13 77 57 3 53 13 77 54 33 28 27 10 I 1\#183 ; 68153 28 50 41 78 12 53 54 13 78 11 33 28 51 13 -- 3 1\#183 ; 68131 4 29 14 32 78 30 12 55 13 78 29 15 29 14 54 4 1\#183 ; 68123 29 37 41 78 49 2 5ti 13 78 47 35 29 38 16 1\#183 ; 68122 30 0 16 79 9 3 57 13 79 6 25 30 1 20 1\#183 ; 68122 30 22 19 79 30 8 58 13 79 26 34 30 23 47 1\#183 ; 68114 31 4 55 80 15 15 60 13 80 9 33 31 7 20 2 1\#183 ; 68125 3 31 45 53 81 5 10 62 13 80 56 36 31 48 39 3 1\#183 ; 68117 32 23 6 81 58 55 64 13 81 48 26 32 27 48 1\#183 ; 68123 32 58 34 82 57 30 66 13 82 44 26 33 4 33 1\#183 ; 68122 33 31 38 84 0 23 68 13 83 44 56 33 38 51 1\#183 ; 68134 34 1 56 85 8 18 70 13 84 50 20 34 10 28 2 1\#183 ; 68130 3 34 29 48 86 20 25 72 13 86 0 32 34 39 24 3 1\#183 ; 68121 34 55 0 87 37 10 74 13 87 15 4'3 35 5 31 1\#183 ; 68127 35 17 10 88 59 20 76 13 88 35 51 35 28 48 1\#183 ; 68121 35 36 39 90 26 0 78 13 !
)0 1 2 35 49 8 - \lt ; / ) is the same for both sheets .
Table III .
I\#189 ; - of\#189 ; 1\#183 ; 68ll9 9 1\#183 ; 68119 1\#183 ; 68125 1\#183 ; 68134 1\#183 ; 68130 1\#183 ; 68129 5 1\#183 ; 68128 1\#183 ; 68119 1\#183 ; 68123 1\#183 ; 68116 1\#183 ; 68123 3 1\#183 ; 68113 1\#183 ; 68114 1\#183 ; 68104 1\#183 ; 68106 1\#183 ; 68103 3 1\#183 ; 68073 1\#183 ; 68032 1\#183 ; 67996 1\#183 ; 67918 1\#183 ; 67841 2 1\#183 ; 67756 1\#183 ; 67674 1\#183 ; 67591 1\#183 ; 67519 1\#183 ; 67450 2 1\#183 ; 67388 1\#183 ; 67328 1\#183 ; 67274 c. , : , 0 0 ~ ; ;o ~ :3 Q t-i P\gt ; N trj t:d ~ 0 0 Pi 0 z 1-ij t-i ~ trj ~ P\gt ; \lt ; 1 trj w. H z P\gt ; t:d ~ P\gt ; ~ P\gt ; t-i 0 ~ \gt ; -1 w. ~ TABLE giving results of experiments on first prism in the second position .
Outer sheet .
Inner sheet .
Table XII .
Table XI .
Table X. Table VII .
Table VIII .
Table IX .
0112\#183 ; I\#189 ; \#183 ; oq/ .
cp ' .
D+i .
yr .
D+i .
cjJ ' .
oc/ J ' .
111\#183 ; 0111II 0 I II 0 I , I 0 I II 0 I II 0 I II II 8 1\#183 ; 68502 6 23 24 20 76 8 53 83 47 5 75 16 5 23 31 25 6 1\#183 ; 68117 8 1 1-68539 22 32 28 76 1 57 35 47 5 75 48 24 22 29 20 1\#183 ; 68122 2 1\#183 ; 68570 21 31 43 75 59 53 37 47 5 75 45 32 21 28 7 1\#183 ; 68120 3 1\#183 ; 68574 20 31 58 76 1 53 39 47 5 75 47 40 20 28 9 1\#183 ; 68119 4 1\#183 ; 68569 19 33 28 76 8 17 41 47 5 75 54 34 19 2\:l 32 1\#183 ; 68123 5 5 1\#183 ; 68560 5 18 36 17 76 19 0 43 47 5 76 5 45 18 32 16 5 1\#183 ; 68125 5 6 1\#183 ; 68541 17 40 30 76 33 43 45 47 5 76 21 18 17 36 30 Hi8125 7 l \#183 ; 68517 16 46 12 7G 52 38 47 47 5 76 41 0 16 42 17 1\#183 ; 68126 8 1\#183 ; 68480 15 53 18 77 15 2 49 47 5 77 4 48 15 49 42 1\#183 ; 68126 9 1\#183 ; 68439 15 2 14 77 41 52 51 47 5 77 32 47 14 58 49 1\#183 ; 68122 10 3 1\#183 ; 68386 4 14 12 45 78 12 16 53 47 5 78 4 34 14 9 48 4 1\#183 ; 68122 3 11 l \#183 ; 68333 13 25 9 78 46 47 55 47 5 78 40 55 13 21 49 1\#183 ; 68131 12 1'68277 12 39 29 79 25 24 57 47 5 79 21 4 12 37 43 1\#183 ; 68127 13 1\#183 ; 68223 11 55 55 80 7 57 59 47 5 80 5 40 11 54 50 1\#183 ; 68130 14 1\#183 ; 68204 11 35 2 80 31 30 60 47 5 80 29 40 113415 1\#183 ; 68140 15 80 55 0 61 47 5 80 55 0 1113 55 1\#183 ; 68122 16 yr is the same for both sheets .
Thus Tables III .
, VI .
, IX .
, and XII .
give the values of the refractive indices , or reciprocals of the wave velocity , for the different directions in the crystal given by II .
, V. , VIII .
, and XI .
\Ve leave here the experimental work for the first prism , and consider the second ; reserving for the future the comparison of our results with theory .
rs:\#183 ; ~ ~ t"-i ~ ~ t ; J td ~ 0 0 Pi 0 z 1- 'd t , ~ t ; J ~ t , :j [ / ) .
H z ~ td H ~ ~ ~ t , 0 ~ [ / ) .
1-3 ~ ( : , :\gt ; 0 I- ' 302 MR. R. 'f .
GLAZEBROOK ON PLANJ !
J WAVES IN A BIAXAL CRYSTAL .
Section IV.-Description ef the Second Prism : Position of it8 Axes.-Observed Valiics of the Deviations and Angles ef Incidence , and Co.lculation ef the Reciprocals of the Wave Velocity .
It will be remembered that , in describing the cutting of the crystal , I stated that a second prism was formed with its edge nearly parallel to C , and its faces not differing much from the m and a faces of the crystal .
Fig. 4 .
Let the figure , fig. 4 , represent a section of the crystal by a plane perpendicular to the rrn faces .
The traces of the faces of the second prism are A Band CD .
The greater part of the face A B lies in the twin crystal , so that the results for the second prism apply to that crystal .
On examination , the m face B C was seen to consist of two parts , which are not in the same plane .
Observation showed that the upper haif ( the prism being placed as in the figure ) is in the same zone as the faces D E , E F , while the lower portion is considerably removed from this zone .
I found also that when the faces D E , E F , were levelled , the line of junction of the dark and bright portions of the slit , when seen by reflexion from A B , coincided with the needle point ; so that A B is in the same zone as the rrn faces ; while the reflexion from CD , though in the field , was too high .
Thus if Q Q , , fig. 5 , be the points where the normal to A B , fig. 4 , and that to C D , fig. 4 , produced backwards meet the sphere of reference , A , B , C , fig. 5 , being the poles of the principal planes ; Q lies on the great circle A M B ; Q , is slightly below it .
Let M , M , ' , fig. 5 , be the points where the normal to E F , and that to E D produced backwards cut the sphere .
Then AM=AM , '=58 ' 28 ' .
The mean of several observations gave for MQ= 103 ' 58 ' .
Hence AQ=45 ' 30 ' .
( 1 ) The mc:1n of the observations for M Q , and M , ' Q , , gave MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BI..A.XAL CRYSTAL .
303 MQ , =57 ' 51 ' 30"=0 say ; M , 'Q , = 59 ' 9 ' = 0 ' say ; and the triangles M A Q , M , ' A Q , give , where 11-=AM=M , ' A. AQ cos 0+-cos 0 ' cos = ' 2 cos\#181 ; Whence substituting for 0 and 0 ' AQ , = 2 ' 43 ' 20 ' ' Fig. 5 .
( 2 ) In the triangle MA Q , the three sides are known ; we can thence find the angle MA Q , and obtain In the triangle Q A Q , the sides A Q , A Q1 and the angle Q A Q , are known AQ = 45 ' 30 ' AQ , = 2 ' 43 ' 20 " QAQ , = 103 ' 35 ' 15 " Whence from the formula we get cos QQ , = cos AQ cos AQ , + sin AQ sin AQ , cos QAQ , QQ , = 46 ' 12 ' ( or rather less ) .
The mean of several experiments had already given for the value of QQ , 4-6 ' 1 o ' ( 3 ) These results , differing as they do by less than 2 ' , afford a verification of the accuracy of the measurements .
The position of the plane Q Q , is completely determined by the arc A Q and the angle A Q Q , ; we have already found , AQ=45 ' 30 ' For the angle A Q Q , the triangle Q A Q , gives .
AQQ sin AQ , .
QAQ Sln = .
QQ Sln 1 1 Slll , whence AQQ , = 4 ' 37 ' 20'/ .
( 4 ) Thns we see that the principal section of the second prism is inclined to the plane c at an angle of 4 ' 37 ' 20 " .
I mentioned above that the second prism lay in the twin crystal .
This was proved by passing light through the crystal , in a plane nearly parallel to the principal plane B C , when four images of the slit appeared .
On cutting off the end containing the second prism , two images vanished ; on cutting off the other end , only the second pair vanished , and the first pair reappeared .
The work for the first prism was done with the second pair , the first pair being rarely sufficiently distinct for measurement .
Throughout the work , \lt ; p \lt ; /\gt ; ' denote the angles which the normals to the incident or emergent wave , and the wave in the crystal make with the normal to the face Q. The same remarks as to the reversal of the crystal hold as in the case of the first prism .
: : : ,0 ... , ; . .
; rt-\lt ; \gt ; \lt ; :-\gt ; \lt ; ~ ~ , ' " , ' I Table XV .
cpl .
/ '1 \#183 ; i(j ) ' .
-- II 5 1\#183 ; 53026 5 1 \#183 ; 53036 1\#183 ; 53031 1\#183 ; 53028 1-5 ; 1020 1\#183 ; 53025 1\#183 ; 53033 1\#183 ; 53036 1\#183 ; 53037 -- 2 1\#183 ; 53040 3 1'53041 1\#183 ; 53043 1'53045 l \#183 ; 53049 TABLE giving the results of experiments on the second prism .
Inner sheet .
Outer sheet .
Table XIV .
Table XIII .
Table XVI .
Table XVII .
--- \lt ; // .
D+i .
( / \gt ; .
D+i .
\lt ; J/ .
o\lt ; // .
0 I ii 0 I II 0 I , , 0 I , , 0 I II , , 27 9 1 74 12 20 44 17 30 82 43 10 24 30 Hl I 32 2 41 I 76 rn 1.5 54 17 30 84 ~5 15 28 49 40 34 34 40 78 11:l 5 60 17 30 86 21 40 31 2 40 rp is the same for both sheets .
Table XVIII .
l'2\#183 ; 0\#181 ; 2-- 1\#183 ; 68312 5 1\#183 ; 68330 I 1\#183 ; 68336 1\#183 ; 68347 1\#183 ; 683fi7 I 1\#183 ; 68363 3 1\#183 ; 68378 1\#183 ; 68381 I l\#183 ; G8391 l \#183 ; G8400 I 1\#183 ; 68401 2 1\#183 ; 68413 1\#183 ; 68421 ~ ~ ~ ~ Q t " p.N t ; J to ~ C 0 r1 0 z 1- , j t " ~ t ; J ~ t ; J w z p.- 1:d H p.p\lt ; j p.t " 0 ~ w ~ Ci : ) 0 \lt ; :)l EXPERIJYIENTAL results for second prism , second position .
I Inner sheet .
Outer sheet .
Table XXI .
Table XX .
Table XIX .
Table XXII .
Table XXIII .
\#183 ; - c\#181 ; 1\#183 ; / 11\#183 ; orf/ .
rj/ .
D+i .
1fr .
D+i .
\lt ; j ) ' .
o\lt ; :/ J ' .
-- ~I II 0 I II 0 I II 0 I II 0 I II 0 I II II 8 1\#183 ; 53027 6 26 4 53 73 59 25 31 42 30 83 53 55 27 58 54 6 1\#183 ; 53019 24 54 17 73 49 40 33 42 30 83 28 20 26 55 30 1\#183 ; 53032 23 44 57 73 45 20 35 42 30 83 1 10 25 53 10 1\#183 ; 53027 22 36 37 73 44 50 37 42 30 82 46 5 24 51 53 I 1\#183 ; 53017 21 29 28 73 48 15 39 42 30 82 38 0 23 51 53 5 1\#183 ; 53018 5 20 23 47 73 56 10 41 42 30 82 36 10 22 53 14 5 1\#183 ; 53010 19 19 24 74 7 35 43 42 30 82 39 40 21 55 57 1\#183 ; 53023 18 16 48 74 23 30 45 42 30 82 48 40 21 0 7 1\#183 ; 53028 17 15 44 74 42 55 47 42 30 83 1 50 20 5 38 1\#183 ; 53013 16 16 9 75 5 30 49 42 30 83 20 45 19 13 4 3 1\#183 ; 53017 4 15 18 40 75 32 25 51 42 30 83 43 40 18 21 59 4 1\#183 ; 53010 14 22 57 76 2 45 53 42 30 84 11 40 17 32 54 1\#183 ; 53013 13 29 24 76 37 15 55 42 30 84 44 10 16 45 43 1\#183 ; 53010 12 37 57 77 15 30 57 42 30 85 21 30 16 0 36 1\#183 ; 53003 11 48 45 77 57 40 59 42 30 86 3 10 15 17 33 3 1\#183 ; 53010 4 11 2 10 78 44 25 61 42 30 86 48 50 14 36 2H 4 1\#183 ; 53010 10 18 0 79 35 10 63 42 30 87 39 40 13 57 54 1\#183 ; 53017 9 36 35 80 30 30 65 42 30 88 35 45 13 21 57 1\#183 ; 53003 8 57 33 81 29 35 67 42 30 89 35 25 12 48 1 1\#183 ; 53014 8 21 52 82 34 10 69 42 30 90 40 30 12 16 53 1fr is the same for both sheets .
Table XXIV .
fl2\#183 ; 0fl2\#183 ; 1\#183 ; 68388 8 1\#183 ; 68376 1\#183 ; 68368 1\#183 ; 68344 1\#183 ; 68333 1\#183 ; 68330 5 1\#183 ; 68320 1\#183 ; 68317 1\#183 ; 68295 1\#183 ; 68289 1\#183 ; 68269 3 1\#183 ; 68264 1\#183 ; 68256 1\#183 ; 68256 1\#183 ; 68~54 1\#183 ; 68231 3 1\#183 ; 68237 1\#183 ; 68240 1\#183 ; 68220 1\#183 ; 68218 !
) l9 co 0 ~ ~ ~ ~ 1-3 Q t , I\gt ; t--l l:tj td ~ 0 0 ~ 0 z r-o t , I\gt ; z l:tj ~ ~ l:tj U1 H z I\gt ; td H I\gt ; ~ I\gt ; t , 0 ~ U1 1-3 I\gt ; ~ MR. R. T. GLAZEBROOK ON PL .
AN W .AVES IN .A BI .
AX.AL CRYSTAL .
307 Section V. --Determination of the Values of the Principal Indices .
-Angle between the Optic Axes .
Our next step is to determine the values of the refractive indices on FRESNEL 's theory .
It ha8 been shown ( Section III .
) that the principal plane of the first prism cuts the plane CA C ' in a line inclined at an angle of 2 ' 31 ' 58 " to O A , and makes with that plane an angle of 1 ' 12 ' 4 " .
Let us consider the intersection of this plane and the plane B 0 , B ' C ' .
Fig. 6 .
It will be a line , M O M ' , say , inclined at 1 ' 12 ' , or rather less to C O C ' .
Take O A , 0 B , 0 C , as axes of :c , y , z , and let J-1 , a , J-1 , b , \#181 ; , 0 be the principal refractive indices , and therefore the semi-axes of the surface of wave slowness .
The sections of the surface of wave slowness on FRESNEL 's theory by the plane B O C are a circle of radius J-1 , a and an ellipse axes J-1 , b , \#181 ; , 0 in the directions of O C , 0 B respectively . .
Hence the values of the radius vector in this direction given by experiment-i.e .
, the values of the refractive indices for light traversing the crystal in this directionare J-1 , a , and the radius vector of an ellipse axes J-1 , b , \#181 ; , 0 , inclined at about 1 ' 12 ' to J-1 , b\#183 ; To determine which of the experiment.al values of \#181 ; , is in this direction we require to find L M , L being the point of intersection of MM ' and A C ' .
From the triangle C ' L M , since C ' is a right angle , cos C'LM= tan C'L cot LM Whence LM= 87 ' 28 ' 4 " And if O P , be as before normal to the face P , of the first prism , LP , = 67 ' 27 ' 25 " [ Section III .
( 8 ) ] .
Whence MP , =20 ' o ' 30 " ( 1 ) ( 2 ) and MP , measures the angle between the wave normal required in the crystal and the normal to the face P , .
This is the angle we have denoted by\lt ; /\gt ; ' .
From Tables II .
, III .
, Vlll .
, and IX .
we have corresponding values of Other observations gave 1\gt ; ' 19 ' 32 ' 21 " 20 ' 34 ' 27 " 20 ' 28 ' 9 " 19 ' 29 ' 32 " \lt ; /\gt ; ' 19 ' 39 ' 22 " 20 ' 41 ' 22 " f1 , 1\#183 ; 68134 1\#183 ; 68130 1 \#183 ; 68119 1\#183 ; 68123 f1 , 1\#183 ; 68126 1 \#183 ; 68129 Again , none of t , he experimental values of p , in Table III .
, lines 1 to 11 , and Table IX .
, differ greatly from 1 \#183 ; 68125 , and this is not far from the mean of the six observations recorded above .
We take , then , this as the radius vector of an ellipse axes / J-b , \#181 ; ,0 , inclined at 1 ' 12 ' to / J-b\#183 ; To determine P , c we must have recourse to the second prism .
Did the plane of the second prism coincide with that of x y , the values of \#181 ; , in Table XXI .
would all be equal to \#181 ; , 0 .
We shall consider later the effect of the inclination of the plane of the prism to that of x y , hut for the present may remark that the values in Table XXI .
differ little from 1 \#183 ; 53013 , the value of / J-c found by R u DBERG .
Let us take therefore P , c= 1 \#183 ; 53013 We have then in the plane B O C -1 - -cos2 0 + -sin2 0 \#181 ; ,2 - / J , b2 \#181 ; , ?
0 being the angle between p , and / J-h\#183 ; Now Whence \#181 ; , = 1 \#183 ; 68125 P , c= ] \#183 ; 53013 0=1 ' 12 ' The value found by RuDBERG is 1 \#183 ; 68157 .
The d ifference is considerable , being \#183 ; 00025 .
( 3 ) ( 4 ) :M:R .
R. T. GLAZEBROOK ON PLANE W .A .
VES IN .A .
BI .
A.XAL CRYSTAL .
309 We may note , however , that his values of / J , b deduced from two different pieces of crystal differed by \#183 ; 0004 .
Now fLa is the value of the refractive index for the first prism for the wave E when \lt ; /\gt ; ' = 20 ' o ' 3 9 " The values of\#181 ; , at about 30 ' on either side of this are The value l \#183 ; 68578 1\#183 ; 68574 Table VI .
, 1 \#183 ; 68589 Table XII .
, 1\#183 ; 68569 fLa= 1 \#183 ; 68580 .
( 5 ) seems to represent these fairly .
Let us take it at JJresent , and see how far it gives consistent results .
RUDBERG gives fJ , a= l \#183 ; 68589 To verify these results let us find the angle between the optic axes .
If 28 ' be this angle in the crystal Whence ( 1 1 1 I I tan 8'= '\/ ~ \#181 ; }- ; } ~ I 1 1 I -- - ~ I L \#181 ; c~q \#181 ; bQ~ J 8'=9 ' 4 ' 58 " .
( 6 ) and if 28 be the angle in air as seen through a face normal to the least axis of the crystal sin 8 = / J , b sin 8 ' whence 8= 15 ' 23 ' 30 ' ' KIRCHOFF found 15 ' 27 ' as the result of experiments .
The value given by RuDBERG 's values of the indices is 8=15 ' 7 ' ( 7 ) which differs from KrRCHOFF 's result by 20 ' , whereas the value given above as the result of my experiments differs only by 3 ' 30 " .
Section VI.-Calculation of the Reciprocal of the Velocity of Wave Propagation on FRESNEL 's Theory for the First Prism and Comparison w\#163 ; th Experiment.-The same for Lord RAYLEIGH'S Theory .
We are now in a position to calculate the reciprocals of the velocity of wave propagation for the first prism on the theory of FRESNEL .
Fig. 7 .
Let O O ' , fig. 7 , be the points in which the optic axes cut a unit sphere .
L M the trace of the principal plane of the prism cutting AC ' A ' in L , B C ' B ' in M. Let N be any point in the plane , in which a wave normal meets the unit sphere .
Let NO=0 NO'=0 ' .
C'LM = x = 1 ' 12 ' [ Section III .
( 9 ) ] .
cos 0= cos LN cos OL+ sin LN sin OL cos x= cos OL cos ( LN-\gt ; . .
) cos i\ if tan A.=tan OL cos X\#183 ; Similarly if tan A'= tan O'L cos X\#183 ; Now whence cos 0 , = cos Oi\''L cos ( LM - / \\ ' ) cos AO=90'-9 ' 4 ' 58 " [ Section V. ( 6 ) ] .
= 80 ' 55 ' 2 " AL=2 ' 31 ' 58 " [ Section III .
( 10 ) ] .
OL= 78 ' 23 ' 4 " O'L=96 ' 33 ' ( 1 ) ( 2 ) MR. R. T. GLAZJ !
JBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
311 .\=78 ' 22 ' 55 " ( 3 ) , \'=96 ' 33 ' 5 " ( 4 ) Also P , being the point in which the normal to the face P , meets the sphere LN=LP , +P , N =LP , +\lt ; /\gt ; ' LP , = 67 ' 27 ' 25 " [ Section III .
( 8 ) ] .
LN -.\=LP , +\lt ; f\gt ; ' -.\ =\lt ; /\gt ; ' -10 ' 55 ' 30 " LN -.\'=LP , +\lt ; /\gt ; ' -.\ ' =\lt ; /\gt ; ' -29 ' 5 ' 40 " ( 5 ) ( 6 ) Again , if v1 v2 are the velocities of propagation in the direction 0 0 ' , we know that v/ +vz2=a2+ci1\#183 ; -(a2-c2 ) cos 0 cos 0 ' v12-vz2= ( a2-c2 ) sin 0 ' sin 0 ' a c having here their usual meanings of the greatest and least principal velocities .
Hence 2v12=a2 +c2-(a2-c2 ) cos ( 0+0 ' ) 2vz2=a2+c2-(a2-c2 ) cos ( 0-0 ' ) From these formulae we can find v1 v 2 , and thence the values of \#181 ; 1 \#181 ; ,2 for Taking f1 , a= l \#183 ; 68580 f1 , c= l \#183 ; 53013 c2=-"=\#183 ; 35187G \#181 ; a " I I 1 I 2 I 17 I 18 I 22 TABLE XXV.-Comparison of FRESNEL 's theory with experiment for the first prism in its fust position .
Outer sheet .
Inner sheet .
Excess of \#181 ; z \#181 ; 2 1'1 Excess of Experiment Experiment .
Theory .
0-0 ' .
0+6 ' .
Theory .
\#181 ; I Experiment over Theory .
Experiment .
over Theory .
0 ' II 0 I -\#183 ; 00006 1\#183 ; 68500 1\#183 ; 68506 7 21 20 18 21 2"0 1\#183 ; 68123 1\#183 ; 68119 - \#183 ; 00004 -\#183 ; 00002 1\#183 ; 68540 1\#183 ; 68542 5 13 20 18 20 0 1\#183 ; 68124 1\#183 ; 68119 -\#183 ; 00005 \#183 ; 00001 1\#183 ; 68568 1\#183 ; 68567 3 7 10 18 19 30 1\#183 ; 68124 1 \#183 ; 68125 \#183 ; 00001 \#183 ; 00000 1\#183 ; 68578 1\#183 ; 68578 1 2 20 18 19 20 1\#183 ; 68125 1\#183 ; 68134 \#183 ; 00009 \#183 ; 00011 1\#183 ; 68589 1\#183 ; 68578 1 0 10 18 19 10 1\#183 ; 68125 1\#183 ; 68130 \#183 ; 00005 \#183 ; 00001 1\#183 ; 68568 1\#183 ; 68567 3 0 50 18 19 30 1\#183 ; 68124 1\#183 ; 68129 \#183 ; 00005 \#183 ; 00012 1\#183 ; 68558 1\#183 ; 68546 4 59 0 18 20 0 1\#183 ; 68124 1\#183 ; 68128 \#183 ; 00004 \#183 ; 00003 1\#183 ; 68517 1\#183 ; 68514 6 54 40 18 20 40 1\#183 ; 68123 168119 -\#183 ; 00004 -\#183 ; 00007 1\#183 ; 68467 1\#183 ; 68474 8 47 20 18 22 0 1\#183 ; 68122 1\#183 ; 68123 \#183 ; 00001 -\#183 ; 00004 1\#183 ; 68421 1\#183 ; 68425 10 36 40 18 24 0 I 1\#183 ; 68121 1\#183 ; 68116 -\#183 ; 00005 \#183 ; -- \#183 ; 00005 1\#183 ; 68376 1\#183 ; 68371 12 21 40 18 27 0 1\#183 ; 68118 1\#183 ; 68123 \#183 ; 00005 -\#183 ; 00010 1\#183 ; 68302 1\#183 ; 68312 14 1 30 18 32 30 1\#183 ; 68114 1\#183 ; 68113 -\#183 ; 00001 -\#183 ; 00025 1\#183 ; 68226 1\#183 ; 68251 15 32 40 18 43 40 1\#183 ; 68104 1\#183 ; 68114 \#183 ; 00010 -\#183 ; 00033 1\#183 ; 68190 1\#183 ; 68223 16 12 10 18 53 50 1\#183 ; 68096 l \#183 ; 68104 \#183 ; 00008 -\#183 ; 00046 1\#183 ; 68153 1\#183 ; 68199 16 45 20 19 9 40 1\#183 ; 68083 1\#183 ; 68106 \#183 ; 00023 -\#183 ; 00049 1\#183 ; 68131 1\#183 ; 68180 17 10 0 19 32 50 1\#183 ; 68063 1\#183 ; 68103 \#183 ; 00040 -\#183 ; 00044 1\#183 ; 68123 1\#183 ; 68167 17 26 50 20 2 50 1\#183 ; 68036 1\#183 ; 68073 \#183 ; 00037 -\#183 ; 00036 1\#183 ; 68122 1\#183 ; 68158 17 37 30 20 37 40 1\#183 ; 68005 1\#183 ; 68032 \#183 ; 00027 -\#183 ; 00030 1\#183 ; 68122 1\#183 ; 681.52 17 44 50 21 15 40 1\#183 ; 67970 l \#183 ; 67996 \#183 ; 00026 -\#183 ; 00032 1\#183 ; 68114 1\#183 ; 68146 17 52 50 22 34 20 1\#183 ; 67893 1\#183 ; 67918 \#183 ; 00025 -\#183 ; 00018 1\#183 ; 68125 1\#183 ; 68143 17 56 50 23 52 40 1\#183 ; 67814 l \#183 ; 67841 \#183 ; 00027 -\#183 ; 00023 1\#183 ; 68117 l \#183 ; 68140 17 59 50 25 8 20 1\#183 ; 67731 1\#183 ; 677.56 \#183 ; 00025 -\#183 ; 00016 1\#183 ; 68123 1\#183 ; 68139 18 1 30 26 19 30 1\#183 ; 67652 1\#183 ; 67674 \#183 ; 00022 -\#183 ; 00016 1\#183 ; 68122 1\#183 ; 68138 18 2 30 27 27 10 1\#183 ; 67574 l \#183 ; 67591 \#183 ; 00017 -\#183 ; 00003 1\#183 ; 68134 1\#183 ; 68137 18 3 30 28 29 30 1\#183 ; 67499 l \#183 ; 67519 \#183 ; 00020 -\#183 ; 00007 1\#183 ; 68130 1\#183 ; 68137 18 4 10 29 26 40 1\#183 ; 67428 1\#183 ; 67450 \#183 ; 00022 -\#183 ; 00016 1\#183 ; 68121 1\#183 ; 68137 18 4 40 30 18 20 1\#183 ; 67361 1\#183 ; 67388 \#183 ; 00027 -\#183 ; 00009 1\#183 ; 68127 1\#183 ; 68136 18 5 10 31 4 30 1\#183 ; 67301 1\#183 ; 67328 \#183 ; 00027 -\#183 ; 00015 1\#183 ; 68121 1\#183 ; 68136 18 5 20 31 45 40 1\#183 ; 67247 1\#183 ; 67274 \#183 ; 00027 - I co '--- t~ ~ ~ ~ t , il\gt ; ' N trj ttl ~ 0 0 r1 0 z " O t , ~ trj ~ ~ trj U1 H z ~ ttl H ~ 0 ~ U1 ~ t'1 ~ 0 0 0 0 t"i ; , \lt ; ; , \lt ; H ?
\lt ; t- : ) w TABLE XXVI.-Comparison of the results of FRESNEL 's theory with experiment for the first prism in its second position .
Outer sheet .
Inner sheet .
Excess of \#181 ; 2 Excess of Experiment orer \#181 ; ~ fl ' .,.9 , 0+0 ' .
\#181 ; l \#181 ; l Experiment over Theory .
Experiment .
Theory .
Theory .
Experiment .
Theor_v .
- -0 I II 0 I II 1 -\#183 ; 00008 1\#183 ; 68502 1\#183 ; 68510 7 21 20 18 21 20 1\#183 ; 68123 1\#183 ; 68117 -\#183 ; 00006 2 -\#183 ; 00008 1\#183 ; 68539 1\#183 ; 68547 5 13 20 18 20 0 1\#183 ; 68124 1\#183 ; 68122 -\#183 ; 00002 3 \#183 ; 00003 1\#183 ; 68570 1\#183 ; 68567 3 7 10 18 19 30 1\#183 ; 68124 1\#183 ; 68120 -\#183 ; 00004 4 -\#183 ; 00004 1\#183 ; 68574 1\#183 ; 68578 1 2 20 18 19 20 1\#183 ; 68125 1\#183 ; 68119 -\#183 ; 00006 5 -\#183 ; 00009 1\#183 ; 68569 1\#183 ; 68578 1 0 10 18 19 10 1\#183 ; 68125 1\#183 ; 68123 -\#183 ; 00002 6 -\#183 ; 00009 1\#183 ; 68560 1\#183 ; 68569 3 0 50 18 19 30 1\#183 ; 68124 1\#183 ; 68125 \#183 ; 0G00l 7 -\#183 ; 00008 1\#183 ; 68541 1\#183 ; 68549 4 59 0 18 20 0 1\#183 ; 68124 1\#183 ; 68125 \#183 ; 00001 8 -\#183 ; 00005 168517 1\#183 ; 68522 6 54 40 18 20 40 1\#183 ; 68123 1\#183 ; 68126 \#183 ; 00003 !
) -\#183 ; 00009 1\#183 ; 68480 1\#183 ; 68489 8 9 22 18 21 32 .
1\#183 ; 68123 1\#183 ; 68126 \#183 ; 00003 10 -\#183 ; 00008 1\#183 ; 68489 1\#183 ; 68447 9 49 40 18 23 0 1\#183 ; 68121 1\#183 ; 68122 \#183 ; 00001 11 -\#183 ; 00015 1\#183 ; 68386 1\#183 ; 68401 11 26 3 18 25 13 I 1\#183 ; 68120 1\#183 ; 68122 \#183 ; 00002 12 -\#183 ; 00018 1\#183 ; 68333 1\#183 ; 68351 12 57 35 18 28 35 1\#183 ; 68116 1\#183 ; 68131 \#183 ; 00015 13 -\#183 ; 00021 1\#183 ; 68277 1\#183 ; 68298 14 22 48 18 34 28 1\#183 ; 68112 1\#183 ; 68127 \#183 ; 00015 14 -\#183 ; 00024 1\#183 ; 68223 1\#183 ; 68247 15 38 55 18 45 15 1\#183 ; 68103 1\#183 ; 68130 \#183 ; 00027 15 -\#183 ; 00019 1\#183 ; 68204 1\#183 ; 68223 16 11 55 18 53 55 1\#183 ; 68096 1\#183 ; 68140 \#183 ; 00044 ~ ~ ~ ~ Q t " ~ N trj ttl ::0 0 0 rl 0 z 1-i:J t " ~ trj ~ ~ UJ .
H z pttl H p~ 0 ~ U1 ~ ~ C ; :i ~ c.:\gt ; The Tables XXV .
and XXVI .
give the results of theory for the first prism .
Let us take the outer sheet first .
In lines 1-12 , Table XXV .
, and lines 1-10 , Table XXVI .
, the agreement between theory and experiment is fairly close , but these series of experiments overlap and are in the immediate neighbourhood of an axis of the section .
From this point onwards , we see that the experimental curve lies entirely inside that given by FRESNEL ' : : !
theory , and the difference between the two increases as we approach the neighbourhood of the optic axes , line 16 , Table XXV .
, 14 and 15 Table XXVI .
, while for the rest of Table XXV .
the differences diminish .
The tables extend from about 8 ' on one side of the axis C , to nearly 16 ' on the other .
This is obvious , for between the optic axes 0-0 ' , beyond them 0+0 ' is very approximately equal to twice N C. Now let us consider the inner sheet , as before in lines 1-14 , Table XXV .
, lines 1-11 , Table XXVI .
, the agreement is fairly close ; but these observations extend from about 6 ' on one side , to 7 ' on the other of the maximum radius vector of the section .
As in the case of the outer sheet , the differences here increase as we approach the neighbourhood of the optic axes , reaching a maximum of \#183 ; 00044 , line 15 , Table XXVI .
After passing the optic axes , the differences decrease , and for the rest of Table XXV .
average about \#183 ; 00024 , and in this case the curve given by experiment lies outside that given by theory .
On examining the differences given in Table XXV .
for the outer sheet , I noticed that theory and experiment agree closely up to the immediate neighbourhood of the optic axis , line 16 ; while in the neighbourhood of the optic axis , Table XXVI .
, line 15 , the increase in the differences is much more gradual .
In fact , if we compare the experimental values of \#181 ; 1 in Table XXV .
, lines 11-14 , and Table XXVI .
, 12-15 , for about the same values of 0+0 ' we find considerable differences .
This led me to calculate the results of other sets of experiments covering the same ground , which had been already made .
The following table gives the results of one such set , taken about two months previously ; during the jnterval the goniometer had been taken partially to pieces , and afterwards refocussed and set again .
It .
seems worth while to give the whole set of observations which cover nearly the \#183 ; Rame ground as Table II .
, and form therefore a test of the degree of accuracy of the results .
The measurements were made at angles of incidence , nearly , but not quite , coincident with those of Table II .
In order to compare results , so far as the wave E ( inner sheet ) is concerned , a correction by interpolation is necessary therefore .
The values of \lt ; /\gt ; ' , it will be observed , differ by about 5 ' from those in Table II .
In the tab1e , the first column gives the values of \lt ; /\gt ; ' ; the second the values of \#181 ; ; l\fR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
315 the third those of \#181 ; , ( corrected by interpolation for the wave E ) ; and the fourth the values of \#181 ; , from Table III .
; while the last column gives the differences between the two sets .
Comparison of values of\#181 ; , deduced from two independent series of observations at an interval of about two months .
cp ' .
0 ' II 34 rn 53 - TABLE XXVII.-Wave 0 .
- \lt ; /\gt ; ' .
0 16 27 ' 1 ' rn 30 2 6 ' 0 28 56 4~ 2 fL 1\#183 ; 681:n 1\#183 ; 68127 1\#183 ; 68128 l \#183 ; 681 : !
G 1\#183 ; 68129 1\#183 ; 68120 1\#183 ; 68128 1\#183 ; 68125 1\#183 ; 68116 1 \#183 ; 68127 1\#183 ; 68116 1\#183 ; 68130 1\#183 ; 68116 1\#183 ; 68127 1\#183 ; 68106 TABLE XXVII .
continued.-W ave E. Values of \#181 ; , Values of \#181 ; , from reduced to the values Table III , corresponding f'\#183 ; of , p ' in Table II .
by to the same values of , P ' interpolation .
as in previous eolun:rn .
- 1\#183 ; 68108 1\#183 ; 68115 1\#183 ; 68066 1\#183 ; 68075 1\#183 ; 68073 1\#183 ; 68029 1\#183 ; 68038 1\#183 ; 68032 1\#183 ; 67991 1\#183 ; 68000 1\#183 ; 67996 1\#183 ; 67905 1\#183 ; 67914 1\#183 ; 67918 1\#183 ; 67828 ' 1\#183 ; 67837 1\#183 ; 67841 1\#183 ; 67743 1\#183 ; 67752 1 \#183 ; 67756 1\#183 ; 67668 1\#183 ; 67678 1\#183 ; 67674 1\#183 ; 67590 1\#183 ; 67600 1\#183 ; 67591 1\#183 ; 67508 1\#183 ; 67518 1\#183 ; 67519 1\#183 ; 67450 1\#183 ; 67461 1\#183 ; 67450 1\#183 ; 67368 1\#183 ; 67380 1\#183 ; 67388 - Excess of the fir , !
series over the second .
-- -\#183 ; 00002 \#183 ; 00006 \#183 ; 00004 -\#183 ; 00004 -\#183 ; 00004 -\#183 ; 00004 \#183 ; 00004 \#183 ; 00009 -\#183 ; 00001 \#183 ; OOOll -\#183 ; 00008 Taking the observations for the wave O first , we observe that they do not differ greatly from those given for the same values of \lt ; /\gt ; ' in Table II .
, except in the case of those which correspond to lines 12 , 13 , 14 , of Table III .
, at which points the new values 2 s 2 of / Li are greater than those given in Table III .
, and agree better with the observations in the neighbourhood of the other optic axis .
Another set of measurements in the same neighbourhood gave as values of 1-'-i 1\#183 ; 68124 1 \#183 ; 68123 1 \#183 ; G8131 l \#183 ; G8125 Hence , on the whole , it seems likely that the results of the observations given in Table III .
, lines 12 , 13 , 14 , are too small , and that an error has been made in the measurement at that point , owing to the indistinctness of the images which nearly overlapped , so that we may fairly replace Table II .
, lines 12 , 13 , 14 , by Table XXVII .
, lines 12 , 13 , 14 , and instead of the results of Table XXV .
referring to these we shall have \#181 ; 1 Theory .
\#181 ; 1 Experiment .
I Difference .
-- 1\#183 ; 68114 1\#183 ; 68130 \#183 ; 00016 1\#183 ; 68104 1\#183 ; 68116 \#183 ; 00012 1\#183 ; 68096 1\#183 ; 68127 \#183 ; 00031 - This modification renders the results similar in the neighbourhood of the two optic axes .
Let us now consider the results given in the latter part of the table for the wave E. The two sets of observations there recorded were made under entirely different circumstances .
The telescope and collimator of the goniometer vrnre removed in the interval between them , and both reset and refocussed .
; and yet in but one case is the difference between the results of the two observations as great as \#183 ; 0001 , while the average difference , irrespective of sign , is about \#183 ; 00005 .
This comparison forms a strong test of the accuracy of the experiments , and .
the close agreement of the results seems to show that the error in them is at any rate not greater than \#183 ; 00005 .
It would appear that morn trustworthy results might have been obtained by combining the series of results registered in this last table with those discussed throughout .
But experiments covering the whole grouud embraced in the series we have chosen had not been made at the same time pi-eviously .
Moreover , we mu.st recollect that each value of \lt ; p and D+ i , given in Tables I. , IV .
, \amp ; c. , is the mean of three for the firnt prism and of two for the second .
And , again , this comparison assures us that the difference between the results of experiment and theory amounting , as it does in some cases , to \#183 ; 0005 in the value of \#181 ; , is far too great to hiwe ari.sen from experimental errors or errors of observation .
It seems , then , that assuming the position of the plane of the prism , and the values of the principal axes to be accurately determined , we may assert that in a central i- : ; ection of the wave surface inclined at a sn1all angle to the plane of the optic :MR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BI .
A.XAL CRYSTAL .
317 axes , there is considerable difference between FRESNEL 's theory and experiment : that the differences between the two are most mark \lt ; l in the neighbourhood of the optic axes , and amount there to \#183 ; 0005 about .
For the outer sheet of the surface ( except just iu the neighbourhood of the principal axes ) , the theoretical values of the radius vector are uniformly greater than the experimental , while for the inner sheet the reverse is the case .
In fact , the curves of section of the surface , as given by experiment , apJJroach more nearly to one another than those of the surface of vrnve slowness on FRES .
NEL 's theory .
Fig. 8 .
In the above figure , fig. 8 , the dotted line gives the result of experiments , while the strong line gives the form of the section on FRESNEL 's theory .
To compare the results with those given by Lord RAYLEIGH'S theory , we have the equation to the surface , ci , b , c being the principal velocities , Z2 m2 n 2 y z +~ + ~ =O - -1 - -1 - -1 a 2 b2 cZ Put V = !
, r is a radius vector of the surface of wave slowness r Let us suppose as a first approximation that m= 0 , i.e. , that the plane of the pnsm is coincident with a principal plane of the crystal , a 2Z2(1-c2r 2 ) +c2n2(1-a2r '2 ) = 0 r 2ci~c2=ci2l:.l+ 2cn2 Also Z=\#183 ; cos 0 , n= sin 0 0 being the angle between r and the maximum radius vector r ~ " = -1 - ( '-1 - -1 ) si.n ~ " 0 cz cZ az ] =2\#183 ; 84192 c ; J From these and the known values of 0 , \#181 ; , or r can be calculated .
The table gives the result .
Comparison between Lord RAYLEIGH'S Theory and Experiment .
0 .
I \#181 ; Theory .
\#181 ; , Experiment .
Excess of Theory over Experiment .
0 ' , , 1 31 10 1\#183 ; 68570 1\#183 ; 68568 \#183 ; 00002 8 25 33 1\#183 ; 68261 1\#183 ; 68190 \#183 ; 00071 11 6 40 1\#183 ; 68028 1\#183 ; 67918 \#183 ; 00110 14 9 48 1\#183 ; 67688 1\#183 ; 67519 \#183 ; 00169 15 48 28 I 1\#183 ; 67476 1\#183 ; 67274 \#183 ; 00202 _l The difference8 between theory and experiment are so marked that it seems unnecessary to calculate the values of \#181 ; , for smaller intervals in the values of 0 .
And though some of the apparent difference may be due to the error made in assuming the principal plane of the prism to coincide with one of the crystal , that cannot account for the whole ; for we have seen that in FRESNEL 's surface the error made by the same assumption appears only in the fourth place of decimals , in the value of the refractive index , while the differences between Lord RATLEIGH 's 'theory and experiment show themselves in the third place , and tend to increase with 0 .
Thus it seems clear that Lord RA YLEIGH 's theory will not account for the phenomena of double refraction in arragonite .
This result agrees with that arrived at by Professor STOKES for Iceland spar , Section VlI.-Ejfect of Varying Constants of Theory and Position of Plane of Prism 'With reference to A xes of Crystal .
It remains now to discuss the effects of variations in the values of the axes a , b , c. The values of c and b are determined directly from observation , and in the neighbourhood of these axes the agreement is most close .
Any alteration in them would affect eRpecially that part of the section which lies between the optic axes , and for which the differences between theory and experim~mt are least ; let us consider , therefore , a variation in the value of a. :MR .
R. T. GLAZEBROOK ON PLANE W .AVES IN .A BT .
AX.AL CRYSTAL .
319 Such a variation would prodnce an alteration in the angle between the optic axes , and therefore in the values of 0 , 0 ' for any wave normal .
If , however , we take a wave normal at some distance from the optic axes , the values of 0 , 0 ' will be altered by nearly equal amounts in opposite directions , and 0+0 ' will be nearly constant .
Let us then find the value for a in order that the experimental value in line 29 of Table XXV .
may agree with theory .
We have From these we get 2v/ =a2+c2-(a2-c2 ) cos ( 0+0 ' ) =a,2{1- cos ( 0+0 ' ) } +c2 { 1 + cos ( 0+ 0 ' ) 1 0+0 v 2-c2 cos2-.\#183 ; .
a,2=------ .
0+0 ' 8Jn2 -211-1=1\#183 ; 67274 0+0'=31 ' 451 40 ' ' c2= \#183 ; 351876 as against \#183 ; 427117 , and the value of / Le is instead of fL c= 1\#183 ; 49478 1 \#183 ; 53013 A reference to the tables for the second prism shows at once the impossibility of such a change .
It is true that any decrease in fLc would produce a decrease in the angle between the optic axes .
This would increase 0 ' and decrease 0 by nearly eqnal amounts , and on the whole produce a small decrease in 0+ 0 ' .
This change would decrease v1 and therefore increase 11-1 , and hence to produce the required change in 11-1-i.e .
, to make fLl equal to the experimental value in line 29 we must take a value for / Le larger , but only slightly larger , than that given above .
There remains now for discussion the effect of a variation in the position of the plane of the prism with reference to the crystallographic axes .
We shall consider this by discussing separately the effect of a rotation about each of three lines through the centre approximutely at right angles .
These lines are ( 1 .
) The normal to the principal plane of the prism ; ( 2 .
) The intersection of this plane with the plane B C B ' ; ( 3 .
) The intersection of the principal plane with the plane AC A ' .
( 1 .
) Rotation about a normal to the principal plane of the prism .
This cannot produce the required effect , for though we can thns decrease the differences on one side of the maximum radius vector of the section , we increase those on the other side by about the same amounts .
( 2 .
) Rotation about the intersection with B C B ' .
This is of no use , for the effects produced on opposite sides of the maximum radius vector are again of an opposite kind .
( 3 .
) There remains , therefore , only the rotation about the line of intersection with the plane C A C ' , or a variation in the angle X\#183 ; This change is to be such that the sections of the outer and inner sheets may be brought into closer proximity ; X must therefore be decreased .
Now on referring to the tables of the experimental results , it will be seen that none of those in Table III .
, line8 l-11 , Table VJ .
, lines 16-29 , and Table IX .
, lines 1-16 , differ greatly from 1 \#183 ; 68125 .
If it were then permissible to neglect the angle X , this would be the value of / J-h , and the other section would be on FRESNEL 's theory an ellipse of / 1-a , / 1-c\#183 ; In the abstract published in the 'Proceedings of the Royal Society , ' No. 188 , 1878 , this bas been done , and a limit assigned to the error thuR introduced .
Professor STOKES has since pointed out that near the optic axes the limit is considerably too small , and thus Jed me to the accurate calculation of the theoretical values of 11-1 , 11-z given above .
The effect of a decrease in the angle X may , therefore , be estimated by referring to the calculations for the case in which we put x= 0 .
The accompanying table gives the results .
0 being the angle between the wave normal and the major axis of the ellipse , which is given by where 00 is the angle between the normal to the face of the prism and major axiP , and is = 20 ' O ' 40 ' ' [ Section V. ( 2 ) ] .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN .A BI .
AXAL CRYS'r.AL .
821 TABLE XXVIII .
Tables from which experi- r or\#181 ; from E xcess of menh1 , l r esult is 0 .
r or\#181 ; from Theory .
E xp eriment .
E xp eriment taken .
over T heory .
0 I XL 8 25 3"8 1\#183 ; 68193 1\#183 ; 68204 " 8 4 45 1\#183 ; 68224 1\#183 ; 68223 - 1 " 7 21 11 1\#183 ; 68284 1\#183 ; 68277 - 7 " 6 35 31 1\#183 ; 68342 1\#183 ; 68333 - 9 , , 5 47 55 1\#183 ; 68396 1\#183 ; 68386 -10 " 4 58 26 1\#183 ; 68444 1\#183 ; 68439 - 5 " 4 7 22 1\#183 ; 68486 1\#183 ; 68480 - 6 V. 3 42 55 1\#183 ; 68504 1\#183 ; 68500 - 4 XL 3 14 28 1\#183 ; 68522 1\#183 ; 68517 - 5 V. 2 38 14 1 \#183 ; 68542 1 \#183 ; 6854,0 - 2 .
XL 2 20 10 1 \#183 ; 68550 1\#183 ; 68541 - 9 V .
1 34 26 1\#183 ; 68566 1\#183 ; 68570 + 4 XL 1 23 23 1 \#183 ; 68570 1\#183 ; 68560 -10 v. 0 31 31 1\#183 ; 68578 1\#183 ; 68578 - 0 XL 0 27 12 1\#183 ; 68578 1\#183 ; 68569 - 9 V. 0 30 15 1\#183 ; 68578 1\#183 ; 68589 + 11 XL 0 31 ] 8 1\#183 ; 68577 1\#183 ; 68574 - 3 XL 1 31 3 1\#183 ; 68567 1\#183 ; 68570 + 3 V. 1 31 10 1\#183 ; 68567 1\#183 ; 68568 + 1 , , 2 30 43 1\#183 ; 68544 1 \#183 ; 68560 . .
XI .
2 31 48 1 \#183 ; 68544 1\#183 ; 68539 - 5 V. 3 29 17 1\#183 ; 68515 1\#183 ; 68517 + 2 XI .
3 33 40 1\#183 ; 68510 1\#183 ; 68502 - 8 V. 4 26 32 1 \#183 ; 68472 1\#183 ; 68467 - 5 , , 5 22 18 1 \#183 ; 68422 1\#183 ; 68421 - 1 , , 6 16 31 1\#183 ; 68367 1\#183 ; 68376 + 9 " 7 9 25 1\#183 ; 68300 1\#183 ; 68302 + 2 " 8 0 38 1\#183 ; 68230 1\#183 ; 68226 - 4 " 8 25 33 1 \#183 ; 68Hl4 1\#183 ; 68190 - 4 " 8 50 1 1 \#183 ; 68156 1 \#183 ; 68153 - 3 II .
9 14 14 1\#183 ; 681]8 1\#183 ; 68131 . .
" 9 37 36 1\#183 ; 68077 l \#183 ; 68073 - 4 " 10 0 40 1\#183 ; 68037 1 \#183 ; 68032 - 5 " 10 23 7 1\#183 ; 67997 1\#183 ; 67996 - 1 " 11 6 40 1\#183 ; 67 !
)14 1\#183 ; 67918 + 4 " 11 47 59 1\#183 ; 67831 1 \#183 ; 67841 +10 " 12 27 8 1 \#183 ; 67748 1\#183 ; 67756 + 8 " 13 3 53 1\#183 ; 67666 1\#183 ; 67674 + 8 " 13 38 11 1\#183 ; 67587 1\#183 ; 67591 + 4 " 14 9 48 1\#183 ; 67511 1\#183 ; 67519 + 8 " 14 38 44 1\#183 ; 67439 1\#183 ; 67450 + 11 " 15 4 51 1\#183 ; 67373 1\#183 ; 67388 + 15 " 15 28 8 1\#183 ; 67310 1\#183 ; 67328 +18 " 15 48 28 1 \#183 ; 67250 1\#183 ; 67274 +24 For about 10 ' on either side of the major axis ( the passage across which is indicated by the dark line in the table ) the results of theory and experiment agree fairly well .
The average difference irrespective of sign is ( if we neglect the values for 0=2 ' 30 ' 43 ' ' and for 0=9 ' 14 ' 14 " , which gives a point on the circle of radius 1\#183 ; 68125 ) \#183 ; 000047 , and only in two cases does the difference amount to \#183 ; 00-01 .
But when we come to values for 0 greater than 11 ' we see that the experimental value of fL is , as before , always greater than the theoretical , nnd that the difference increases with 0 , and at last becomes nearly the same as in Table XXV .
Thus this change does not produce agreement between the two .
We may , however , consider shortly what alteration in / Le would in this case bring theory and experiment more closely into agreement .
We have .
c ; :- _\#181 ; , .
20 , ; : . .
o\#181 ; --3 Sln Of .
Le fl , c Taking the last values on Table XXVIII .
8\#181 ; = \#183 ; 00024 0=15'48 ' \#181 ; =1\#183 ; 67274 fLe= J. '53013 Whence 8\#181 ; 0= '00248 so that the minor axis of the ellipse , which having the same major axis as the experimental curve would pass through the extreme experimental point , is instead of \#181 ; 0=1\#183 ; 53261 1 \#183 ; 53013 And on referring to Table XXI .
, which gives the values of / Le very approximately , we see that this value is quite out of the questjon .
Thus a decrease in the angle X will not produce the required effect .
Neither will an increase .
For the first part of the inner sheet in Table III .
, and throughout Table IX .
, we must have the theoretical value of \#181 ; 1 and therefore of 0+0 ' nearly constant ; and this is clearly impossible along a plane section incli11ed at a finite angle to the plane of the optic axes .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
323 Thus a change in X will not render theory and experiment consistent .
Thus no change in the position of the plane of the prism will bring FRESNEL 's theory into agreement with experiment .
Section VIII.-The The01\#183 ; etical Investigations for the Second P1-ism on FRESNEL 's Theory with Comparison with Experiment .
Our next step is the theoretical investigation for the second prism .
Now experiment has shown that this prism lies in the twin crystal ; we must therefore \#163 ; nd the axes of this crystal .
Fig. 9 .
Let the \#163 ; gnre represent a section of the crystal by a plane perpendicular to the axis on The twinning takes place about an axis perpendicular to m ' , and consists of a rotation through 180 ' about that axis .
Its effect , therefore , is to bring O B , \#163 ; g. 9 , into the position O B ' , where angle mOB=mOB ' that is , to turn the axes in the direct.ion B to A , through an angle , =2(mOB)=2(i-AOm ) =1r-2\#181 ; , =180'-116 ' 56 ' =63 ' 4 ' and if Q is the point where the perpendicular on the face Q of the second prism meets the sphere , we know that AQ=45 ' 30 ' [ Section IV .
( l ) ] .
Hence since AA'= 63 ' 4 ' and Q lies in plane A O B , A'OQ=l7 ' .34 ' .
( 1 ) Now let A ' B ' C ' , :fig .
10 , be the points in which the new axes cut the sphere , and let P Q M be the plane of the prism ; P being any point in the plane ; Q the normal to the face , and M the point in which the plane cuts B ' C ' , 0 , O ' the optic axes .
Then if P is the pole of any wave in the crystal , PQ = \lt ; f\gt ; ' .
Fig. 10 .
Let PO=0 , PO'=0 ' , we shall require the values of 0 , 0 ' .
Now in triangle A ' M Q angle Q= 4 ' 37 ' 20 " [ Section IV .
( 4 ) ] .
A'Q=l7 ' 34 ' cot MQ= cos Q cot A'Q whence MQ= 17 ' 37 ' 14 " .
Let A ' M Q=xThen from triangle A ' M Q , whence Now , put cos M= cos QA ' sin Q. x= s5 ' 35 ' 40 ' ' cos 0 ' = cos PM cos MO'+ sin PM sin MO ' cos x = cos MO'(cos PM+ sin PM tan MO ' cos x ) tan MO ' cos x= tan " A. ' 0 , cos M01 cos ( PM-i\ ' ) cos = - , .
, - .
cos/ \ , ( 2 ) ( 3 ) ( 4 } MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
325 Similarly 0 cos MO cos ( PM+ :X .
) cos = cos/ " \ ' , .
where tan ; \.=tan MO cos x Now A'O=A'O ' =90'-9 ' 4 ' 58 ' ' [ Section V. ( 6 ) ] .
From triangle A ' M Q Whence Hence = 80 ' 55 ' 2 ' ' sin A'M= sin MQ sin Q A'M= 1 ' 23 ' 52 " . .
\#183 ; .
M0=79 ' 31 ' 10 " .
MO'= 82 ' 18 ' 54 " .
A.=22 ' 33 ' 2 " .
A.1 = 29 ' 39 ' 5 " PM+A=\lt ; p ' +QM+A = \lt ; f\gt ; ' + 40 ' 1 o ' 16 " .
PM- ; \ , '=\lt ; /\gt ; ' +QM- ; \ , ' =\lt ; /\gt ; ' -12 ' l ' 51 ' ' .
( 5 ) ( 6 ) ( 7 ) ( 8 ) ( 9 ) ( 10 ) ( 11 ) ( 12 ) Substituting these values of A , MO , PM+A , \amp ; c. ( the values of \lt ; f\gt ; ' being taken from Tables XIV .
, XVII .
, and XX .
) , in equations ( 4 ) , ( 5 ) , we can get a series of values for 0 0 ' .
And as before [ Section VI .
] , remembering that 0 ' here is 7r-0 ' in Section VI .
, 2v/ =a2+c2+(a2-c2 ) cos ( 0-0 ' ) 2v22=a2+c2+(a2-c2 ) cos ( 0+0 ' ) Thus we obtain a series of values for v1 , v 2 , and hence for \#181 ; ,1 , \#181 ; ,2 , given m Table XXIX .
To obtain the distance of P from the minimum radius vector , we have to add to \lt ; f\gt ; ' the angle M Q=l7 ' 37 ' 14 " LSection VJIT .
( 2 ) ] .
I q/ from Tables XVII .
and XXIII .
0 I II TABLE XXIX.-Comparison of FRESNEL 's theory and experiment for second prism .
Outer sheet .
Inner sheet .
Excess of \#181 ; 2 \#181 ; 2 \#181 ; l Excess of Experiment over 0+0 ' .
0-0 ' .
J,11 Experiment over Theory .
Experiment .
Theory .
Theory .
Experiment .
Theory .
0 , II 0 , II -\#183 ; 00024 1\#183 ; 68218 1\#183 ; 68242 164 15 15 1 17 0 1\#183 ; 53014 1\#183 ; 53014 . .
-\#183 ; 00026 1\#183 ; 68220 1\#183 ; 68246 164 20 17 -\#183 ; 00010 1\#183 ; 68240 1\#183 ; 68250 164 25 43 -\#183 ; 00027 1\#183 ; 68227 1\#183 ; 68254 164 31 ' 42 -\#183 ; 00028 1\#183 ; 68231 1\#183 ; 68259 164 38 7 1\#183 ; 53010 -\#183 ; 00009 1\#183 ; 68254 1\#183 ; 68263 164 45 9 1 57 45 1\#183 ; 53016 1\#183 ; 53003 -\#183 ; 00013 -\#183 ; 00013 1\#183 ; 68256 1\#183 ; 68269 164 52 32 1\#183 ; 53010 -\#183 ; 00016 1\#183 ; 68256 1\#183 ; 68272 165 0 34 -\#183 ; 00015 1\#183 ; 68264 1\#183 ; 68279 165 8 56 -\#183 ; 00015 1\#183 ; 68269 1\#183 ; 68286 165 18 10 .
-\#183 ; 00003 1\#183 ; 68289 1\#183 ; 68292 165 27 46 2 35 0 1\#183 ; 53019 1\#183 ; 53013 -\#183 ; 00006 -\#183 ; 00003 1\#183 ; 68295 1\#183 ; 68298 165 37 52 +\#183 ; 00010 1\#183 ; 68312 1\#183 ; 68302 165 43 32 +\#183 ; 00020 1\#183 ; 68330 1\#183 ; 68310 165 55 42 +\#183 ; 00019 1\#183 ; 68336 1\#183 ; 68317 166 7 38 +\#183 ; 00022 1\#183 ; 68347 1\#183 ; 68325 166 19 42 4 2 10 1\#183 ; 53028 1\#183 ; 53028 . .
+\#183 ; 00025 1\#183 ; 68357 1\#183 ; 68332 166 31 32 +\#183 ; 00024 1\#183 ; 68363 1\#183 ; 68339 160 43 48 +\#183 ; 00032 1\#183 ; 68378 1\#183 ; 68346 166 55 6 +\#183 ; 00028 1\#183 ; 68381 1\#183 ; 68353 167 6 38 +\#183 ; 00031 1\#183 ; 68391 1\#183 ; 68360 167 17 56 4 43 34 1\#183 ; 53035 1\#183 ; 53036 +\#183 ; 00001 +\#183 ; 00036 1\#183 ; 68400 1\#183 ; 68366 167 28 54 +\#183 ; 00029 1\#183 ; 68401 1\#183 ; 68372 167 39 32 +\#183 ; 00035 1\#183 ; 68413 1\#183 ; 68378 167 50 0 +\#183 ; 00038 1\#183 ; 68421 1\#183 ; 68383 167 59 56 +\#183 ; 00048 1\#183 ; 68437 1\#183 ; 68389 167 9 29 5 24 56 1\#183 ; 53043 1\#183 ; 53045 +\#183 ; 00002 \lt ; /J ' from Tables XIV.and XX .
0 , II ---- 00 tv 0 ~ ~ ~ f-3 ~ \gt ; N t=.J o::l ~ 0 0 p::j 0 z 1-ij t"i ~ t=.J ~ t=.J [ / ) .
H z \gt ; - 1 : ; :j H k \gt ; t " i 0 ~ [ / ) .
8 \gt ; != " ' MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
327 Considering first the inner sheet , we see that theory and experiment agree closely .
The greatest difference is \#183 ; 00013 ; and in this case , by referring to the adjacent experimental values , it is clear that the value 1 \#183 ; 53003 is too small .
We must remember , however , that the investigations are concerned with that sheet of the surface of wave slowness of which the section by the plane , x y , is a circle , and that the section considered is inclined at an angle less than 5 ' to this plane ; so that the form of the curve of section differs but little from a circle , and we may therefore reasonably suppose that any theory in which one of the principal sections of the surface of wave slowness is a circle , would also agree fairly closely with our results .
We therefore proceed to discuss the results for the outer sheet given in the table just preceding .
And here we see at once that there is considerable difference between theory and experiment .
Moreover , from \lt ; /\gt ; ' = 12 ' 1 G ' 53 " to \lt ; /\gt ; ' =20 ' 5 ' 38 " , the theoretical values of \#181 ; , are uniformly greater than the experimental ; while from \lt ; /\gt ; '=20 ' 35 ' 20 " to \lt ; /\gt ; '=31 ' 42 ' 34 " the reverse is the case .
The difference between the two increases on the whole uniformly , and in the last case is as great as \#183 ; 0005 .
For the values\lt ; /\gt ; '= 19 ' 13 ' 4 " and \lt ; /\gt ; '=20 ' 5 ' 38 " , theory and experiment agree closely .
To find the angle between any radius vector and the minimum one of the curve , we have to add to the value of\lt ; /\gt ; ' the angle M Q , which is 17 ' 37 ' 14 " .
Let us call the angle Q O P , 0 .
Then we see that for the angles 0=0 , 0=37 ' ( about ) , and 0=90 ' , the curves given by theory and experiment agree .
From 0=29 ' 54 ' 7 " to 0=37 ' ( aboul ) , the theoretical curve lies outside the experimental .
From 0=37 ' ( about ) to 0=49u 19 ' 48 " the reverse is the case .
In the neighbourhood of the axes the two must agree more closely .
These conditions would all be satisfied by a curve which agrees with FRESNEL 's section at the extremity of the minimum radius vector ; lies inside it for about 40 ' ; then cuts FRESNEL 's curve again , and lies outside for the rest of the quadrant , agreeing again at the extremity of the maximum radius vector .
At first sight there seems some discrepancy between this result and that arrived at already in Table XXV .
for the first prism .
But closer examination shows that the two results are corroborative .
For , in the first place , the observations there tabulated apply to the neighbourhood of the major axis of the ellip6c section .
For about 6 ' on either side of this principal axis , the differences between theory and experiment are sometimes positive , sometimes negative ; that is , between these limits FRESNEL 's section satisfies the results of experiment ; but in the last eighteen observations there recorded , the expe1\#183 ; imental value of \#181 ; ,1 is always greater than the theoretical .
Thus the experimental results are represented by a curve , which near the major axis agrees with FRESNEL 's section , but which , as we go from 6 ' to 16 ' away from that axis , differs from that section , and lies outside the section .
This result , viz. : that on going away from the maximum radius vector the curve given by experiment should lie outside that given by theory , is exactly that to which we have been led by the work for the second prism .
The cutting of the crystal would not allow of observations being taken near the minor axis of the ellipse , in the case of the first prism .
Of course , it has here been assumed that the sections of the surface of wave slowness , by the principal planes , are curves of the same kind ; or rather the results of experiment lead to that as a fact .
Lord RAYLEIGH'S theory needs no further discussion here , for his curve being the inverse of an ellipse , always lies outside the ellipse with the same axes , and therefore cannot agree with experiment .
Section IX.-Possibility of an Error in the Position of the C1'ystallographic Axes discussed.-No Section of the Surface of Wave Slowness in FRESNEL 's Theory can agree with the experimental Results .
But the possibility remains that there is an error in the determination of the position of the faces of the prisms , especially in the second case , with respect to the crystallographic axes .
The probability of any error is small , as the observations were repeated on several different occasions with nearly coincident results .
To make certain , however , let us solve the inverse problem .
Having given the two values , / Li / Lz , of the radii vectores of the surface of wave slowness drawn in the same direction l rn n in the crystal to find l , m , n. We have as the equation to find v on FRESNEL 's theory Whence v4-vz{ z2(b2+c2 ) +mz(cz+ci2 ) +n2(a2+b2 ) } +zzb2c2+m2c2a2+n2a2b2=0 .
\#183 ; .
if v1 v2 are the roots of this equation , that is , the velocities for the same wave front v12vz2=l2b2c2+m2c2a2+n2ci2b2 v/ +vz2=l2(b2+c2 ) +1n2(c2+a2 ) +n2(a2+b2 ) but n2= 1-m}-l2 Whence we get b2z2(c2-a2 ) + a2m2(c2- b2 ) =v12vz2-a2b2 F(c2-a2)+rn'\#183 ; \c2 - b2)=v/ +'uz2-a2-b2 :M:R .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
329 Whence F(ci2-c2)(a2-b2)= ( a2-vi2)(a2 -vz2 ) 1nP(b2 - c2 ) ( b2 - a2 ) = ( b2-v/ ) ( b 2-v~,2 ) n2(c2-a,2)(c2-b2 ) = ( c2-vi2 ) ( c2-vz2 ) From the tables for the second prism , we see that for the inner sheet for cp'=28 ' 11 ' 21"_ !
_=1\#183 ; 53021 Vl For the outer sheet , Since the value of v1 changes very slowly , we may reasonably treat these two values for cp ' as coincident , and substitute in the above formulae .
We find cos-1 n 1 = 49 ' 35 ' cos-1 1n1 = 40 ' 27 ' Again , fur inner sheet for For outer sheet Treating these two as coincident , we find cos-1 n2= 55 ' 44 ' cos-1 rii2= 34 ' 23 ' For inner sheet for For outer sheet Whence cp ' = 2G0 5 ' 11 " 2-_ = 1 \#183 ; 53028 Vl cp'=26 ' 19 ' 10 " !
=1\#183 ; 68378 V2 cos-1 n3= 47 ' 5 6 ' cos-1 m3= 42 ' 8 ' We may justify the treating of the two values of cp ' as the same , by the consideration that Jl-i varies at the rate of \#183 ; 00001 ( about ) for 1 ' , and therefore the difference for 20 ' is inappreciable .
Now the angle between the radii vectores in the first and second case cannot be less than i.e. , notlessthan55 ' 44'-49 ' 35 ' or 6 ' 9 ' But the angle between these radii is the difference between the two corresponding values of \lt ; f\gt ; ' , or =31 ' 42 ' 34 " -28 ' l ' 21 " = 3 ' 41 ' 1311 Again , taking the second and third , the angle deduced from theory is not less than cos-1 n 2-cos-1 n3 but from experiment it is or 55 ' 44 ' -47 ' 56 ' or 7 ' 48 ' 31 ' 42 ' 34"-26 ' 19 ' 1011 = 5 ' 23 ' 24 " In both these cases the angle , as given by theory , must be very much larger than that given by experiment .
Thus , in order that the three pairs of values of / Li JJ-2 may correspond to three radii of the surface of wave slowness on FRESNEL 's theory , the angles between them must be half as large again as they are found to be .
Hence no section of the surface of wave slowness on FRESNEL 's theory can agree with the results of the experiments .
We may remark , in addition , that the differences between consecutive values of fL for the outer sheet given by experiment are about double those given by theory .
After this investigation it seems needless to discuss possible alterations in the position of the plane of the second prism , with a view to bringing the results of experiment more nearly into accordance with theory .
While as regards the constants a , b , c , it has already been seen that they have received their most probable values .
Thus our results , so far as they go , point to the fact that FRESNEL 's construction does not represent Nature , and that some other theory must be sought to explain the phenomena of double refraction .
And as regards the rival theorioo already proposed , Lord RA YLEIGH 's , we have seen , is even more than FRESNEL 's at variance with observed appearances , while I can hardly think that sufficient data have yet been collected to repay the labour of comparison with the theories of GREEN and CAUCHY , for on account of the undetermined constants it would be necessary to force theory and experiment to agree in so many points that they mm ; t , perforce , almost agree exactly along the small arcs investigated .
:MR .
R. T. GLAZEBROOK ON PL .
A. NO W .A .
YES IN .A .
BI .
A.X.A.L CRYST .
A.L. 331 PA .
RT II .
Section r. \#183 ; 1"-Description of a Second Crystal and Resitlts of Experiments .
Before asking permission to lay the results obtained in the previous part of the paper before the Socjety , I thought it would be better , if possible , to repeat the investigation with a second piece of the crystal .
With some trouble I obtained a suitable piece last February from A. HILGER , of Tottenham Court Road .
Like the former piece , it was an hexagonal pnsm m shape , but of considerably smaller cross-section .
Fig. 1 .
The axis of the prism was parallel to the axis of c , the base being perpendicular to the axis , while the top was broken off obliquely .
The planes m , m ' gave the best reflexions , and were therefore chosen to determine the position of artificial faces .
My chief aim in having the prism cut was to get as long a continuous arc on the wave surface as possible .
To attain this I had a face R , fig. 2 , polished as nearly coincident with c as possible .
Fig. 2 .
The oblique end was cut ( P , fig. 2 ) so as to be inclined at about 37 ' to R , the edge of the prism thus formed between P and R being nearly parallel to the interseution of m and c. * N.B. The two Parts being quite distinct , the numbering of the Sections and Figures is commenced afresh with Part II .
2 u 2 The face ni , was cut so as to be inclined at about 35 ' 20 ' to P , the line of inter section of this face ( Q , fig. 2 ) and P being also nearly parallel to that of P and R. So that the principal planes of the prisms formed by P and R , P and Q respectively , were nearly coincident , and as my observations extended almost from perpendicular incidence on Q to perpendicular incidence on R : they embraced an arc of over 70 ' .
The instrument used and the methods adopted for levelling , reading , \amp ; c. , have been fully described in the previous part of the paper .
The results of the experiments are contained in Tables I. , II .
, III .
, and IV .
Tables I. and II .
refer to the p~ism R P. cp is the angle made by the wave normal in air with the normal to P. cf\gt ; ' the corresponding angle in the crystal .
D is the observed deviation .
i the angle of prism .
D+i is given , for it occurred in the calculation , and was just as easily found as D. The value taken for i is i= 3 7 ' 2 ' 56 " This value is the mean of twelve observations , none of which differed from the mean by 10 " .
cp ' is found from the formulae cf\gt ; ' +"1 ' =i and tan cp ' --t ' = tan i._ cot cp_+-t tan cf , --t C 2 2 2 2 which has already been proved ( Part I. , Section I. ) .
In Table I. cp is the same for both outer and mner sheet , being the angle of incidence .
In Table II .
cp has different values for the two sheets , being the angle of emergence , for as already explained , on passing through the position of minimum deviation , the prism was reversed so that the face of incidence became that of emergence .
Throughout the work \#181 ; 1 refers to the inner , 11-z to the outer sheet .
The values of Sep ' , 8\#181 ; are calculated as before on the assumption of errors of 10 " in 2cp and D+i , the observed quantities taken ; these errors being combined so as to produce the greatest possible effect in the result .
Tables III .
and IV .
refer to P Q , for which i=35 ' 18 ' 50 " .
The symbols have the same meaning , but while in I. and II .
the wave normal falls on the same side of the normal to P as the axis O 0 , in III .
and IV .
it is on the other side of O P. In III .
cp is the angle of emergence , and therefore different for the inner and outer sheets ; in IV .
it is the angle of incidence , and the same for both .
Each observation was repeated twice on different occasions ; the results of two measurements rarely differed by 20 ' .
TABLE I. Outer sheet .
I \lt ; j ) .
o112 , \#181 ; 2 .
o\lt ; j ) ' .
\lt ; / ) ' .
D+i .
D+i .
" 0 I II 0 I II 0 I " 0 I 35 59 41 83 53 56 82 7 10 83 42 2"6 1\#183 ; 68559 35 45 .55 82 17 1 80 7 10 82 5 36 \#183 ; 00009 1\#183 ; 68558 9 35 29 24 80 44 41 78 7 10 80 33 21 1\#183 ; 68553 35 10 1 79 17 31 I 76 7 10 79 6 26 1\#183 ; 68546 34 47 54 77 54 51 74 7 10 77 44 6 1'68539 34 22 47 76 37 16 72 7 10 76 26 36 \#183 ; 00007 1\#183 ; 68532 33 55 3 75 2t 6 70 7 10 75 13 26 1-68526 8 33 24 38 74 15 31 68 7 10 74 4 51 1\#183 ; 68519 32 51 40 73 11 16 66 7 10 73 0 36 1\#183 ; 68505 32 16 18 72 11 11 64 7 10 72 0 31 1\#183 ; 68514 313814 71 15 51 62 7 10 71 4 51 1\#183 ; 68497 30 58 13 70 23 51 60 7 10 70 12 26 \#183 ; 00006 1\#183 ; 68489 8 30 15 49 69 36 6 58 7 10 69 24 31 1\#183 ; 68485 29 31 16 68 52 16 56 7 10 68 39 31 1\#183 ; 68465 28 44 50 68 11 51 54 7 10 67 58 26 - 1\#183 ; 68473 27 56 13 67 35 36 52 7 10 67 20 41 1\#183 ; 68463 27 5 54 67 2 36 50 7 10 66 46 36 \#183 ; 00004 1\#183 ; 68454 7 26 13 50 66 33 6 48 7 10 66 15 16 1\#183 ; 68457 25 19 58 66 7 16 46 7 10 65 47 16 1\#183 ; 68457 24 24 34 65 44 56 44 7 10 65 22 21 1\#183 ; 68459 23 27 40 65 26 1 42 7 10 65 0 31 1\#183 ; 68451 22 29 27 65 10 11 40 7 10 64 41 51 \#183 ; 00002 1'68448 6 21 29 52 64 57 56 38 7 10 64 25 56 1\#183 ; 68441 20 29 5 64 49 6 36 7 10 64 ]3 16 1\#183 ; 68446 19 27 3 64 44 6 34 7 10 64 3 36 - I 1\#183 ; 68461 18 23 54 64 42 56 32 7 10 63 57 26 1\#183 ; 68446 17 19 54 64 45 1 30 7 10 63 54 1 1\#183 ; 68448 6 16 14 53 64 51 36 28 7 10 63 54 6 Inner sheet .
\lt ; / ) ' .
I o\lt ; j ) ' .
fl10 I II II 36 6 16 1\#183 ; 68102 35 35 59 9 1\#183 ; 68106 35 16 20 1\#183 ; 68115 34 53 57 1\#183 ; 68114 34 28 43 1\#183 ; 68115 34 0 54 1\#183 ; 68106 33 30 24 8 1\#183 ; 68099 32 57 21 1\#183 ; 68089 32 21 52 1\#183 ; 68074 31 43 51 1\#183 ; 68069 31 3 56 1\#183 ; 68030 30 21 28 8 1\#183 ; 68016 29 37 20 1\#183 ; 67962 28 51 1 1\#183 ; 67817 27 12 48 1\#183 ; 67806 26 21 14 7 1\#183 ; 67721 25 27 58 1\#183 ; 67634 24 33 11 1\#183 ; 67532 23 36 55 1\#183 ; 67421 22 39 13 1\#183 ; 67303 214018 6 1\#183 ; 67161 20 40 3 1\#183 ; 67015 19 38 38 1\#183 ; 66854 18 36 0 I 1\#183 ; 66694 17 32 26 1\#183 ; 66502 16 27 50 6 I 1\#183 ; 66300 0fl1\#183 ; \#183 ; 00009 3 \#183 ; 00007 8 \#183 ; 00006 13 \#183 ; 00004 18 \#183 ; 00002 23 ~ ~ ~ ~ Q t " :\gt ; N t=.J to ~ 0 0 ~ 0 z hj t , ~ r::'.i ~ :\gt ; - ~ U1 ... .
; z :\gt ; td H :\gt ; ~ C " 0 ~ U1 1--3 :\gt ; r i : , :i i : , :i i : , :i TABLE II .
Outer sheet .
\lt ; p. c\#181 ; 2\#183 ; l'-2\#183 ; o\lt ; /J ' .
\lt ; /J ' .
I D+i'. .
\lt ; p. -- !
0 I II II 0 I , , 0 I II 0 I II 32 49 16 1\#183 ; 68433 18 46 18 64 42 26 32 3 51 30 50 31 1\#183 ; 68458 17 43 17 64 43 41 30 0 46 28 55 16 1\#183 ; 68446 16 41 4 64 48 26 28 1 21 27 3 31 1\#183 ; 68445 15 40 5 64 56 41 26 5 1 25 15 21 1\#183 ; 68448 I 23 30 46 3 1\#183 ; 68458 4 13 41 58 65 23 56 22 22 21 214911 1\#183 ; 68453 12 44 51 65 42 21 20 36 11 20 11 11 1\#183 ; 68448 11 49 16 66 4 21 18 53 6 18 36 46 1\#183 ; 68455 10 55 19 66 29 56 17 13 41 17 5 26 1\#183 ; 68446 10 2 51 66 .:i8 36 15 37 16 15 37 50 4 1\#183 ; 68452 5 9 12 17 67 31 6 14 4 21 14 13 41 1\#183 ; 68446 8 23 25 68 6 51 12 35 11 12 53 26 1\#183 ; 68451 7 36 37 68 46 36 11 9 41 11 36 41 1\#183 ; 68450 6 51 45 69 29 51 9 47 16 10 23 56 1\#183 ; 68454 6 9 3 70 17 6 8 29 1 9 14 51 4 1\#183 ; 68450 5 5 28 28 71 8 1 7 15 26 8 9 51 1\#183 ; 68445 4 50 10 72 3 1 6 5 26 7 9 6 1\#183 ; 68445 4 14 19 73 2 16 4 59 41 6 12 51 1\#183 ; 68453 3 41 3 74 6 1 3 58 41 5 20 21 ----- 1\#183 ; 68437 3 10 1 7513 31 3 2 11 4 33 1 5 1\#183 ; 68447 6 2 41 58 76 26 11 2 10 6 3 49 36 1\#183 ; 68428 2 16 15 77 42 46 1 23 31 3 12 1 1\#183 ; 68451 1 53 58 79 5 11 0 41 51 Inner sheet .
D+\#183 ; i. \lt ; j ) ' .
o\lt ; /J ' .
0 I II 0 I II II l'-1\#183 ; oh -- 1\#183 ; 66678 1\#183 ; 66490 1\#183 ; 66302 1\#183 ; 66095 1\#183 ; 65883 1\#183 ; 65667 3 1\#183 ; 65459 1\#183 ; 65240 1\#183 ; 65037 1\#183 ; 64818 1\#183 ; 64603 4 1\#183 ; 64400 1\#183 ; 64203 1\#183 ; 63984 1\#183 ; 63782 1\#183 ; 63615 4 1\#183 ; 63433 1\#183 ; 63258 1\#183 ; 62956 1\#183 ; 62807 5 1\#183 ; 62691 1\#183 ; 62583 c , J i : , , : , ; i,. .
~ ~ pj f-3 Q t " \gt ; N ~ td ~ 0 0 ~ 0 z l ; j t " z\gt ; ~ ~ ~ rn )-1 z \gt ; td H ~ \gt ; t " 0 ~ rn 1-3 \gt ; ~ TABLE III .
Outer sheet .
\lt ; j ; .
0\#181 ; 2- !
12\#183 ; o\lt ; /J ' .
\lt ; j ; ' .
D+i .
\lt ; j ; .
0 , " " 0 , II 0 , , , 0 , , , 1 32 40 1\#183 ; 68464 0 55 0 73 39 50 0 49 20 2 19 15 1\#183 ; 68450 1 22 39 72 26 25 0 4 30 3 10 50 1\#183 ; 68454 1 47 28 71 18 0 0 44 25 4 6 30 1\#183 ; 68450 2 26 15 I 70 13 40 1 37 25 5 6 45 1\#183 ; 68458 3 1 56 69 13 55 2 34 45 -- 6 10 50 9 1\#183 ; 68456 9 3 30 52 68 18 0 3 35 50 7 19 10 1\#183 ; 68464 4 20 14 67 26 20 4 40 40 8 31 10 1\#183 ; 68462 5 2 41 I 66 38 20 5 49 5 9 46 40 1\#183 ; 68449 5 47 9 65 53 50 7 0 50 11 6 20 1\#183 ; 68462 6 33 56 65 13 30 8 16 20 13 55 45 1\#183 ; 68462 8 12 55 64 2 55 10 56 30 15 25 25 1\#183 ; 68457 9 5 2 63 32 35 12 21 35 16 58 30 1\#183 ; 68454 9 58 50 63 5 40 13 49 10 18 35 5 1\#183 ; 68462 10 54 19 62 42 15 15 19 40 I 20 14 50 --- I 23 43 35 1-68465 13 49 6 61 50 45 20 8 30 I 29 21 o 7 1\#183 ; 68457 6 16 54 56 612810 25 22 10 31 20 30 1\#183 ; 68464 17 59 3 61 27 40 27 12 0 33 23 30 1\#183 ; 68465 19 4 4 61 30 40 29 4.50 37 40 55 1\#183 ; 68465 2116 33 61 48 5 32 59 5 39 56 20 1\#183 ; 68473 22 23 57 62 3 10 I 35 o 50 Inner sheet .
D+i .
\lt ; j ; ' .
o(J/ .
0 , , , 0 , " , , -- I\#189 ; - Of\#189 ; -- 1\#183 ; 62619 1\#183 ; 62500 1\#183 ; 62360 1\#183 ; 62208 1\#183 ; 62058 1\#183 ; 61904 9 1\#183 ; 61715 1\#183 ; 61525 1\#183 ; 61320 1\#183 ; 61120 1\#183 ; 60897 8 1\#183 ; 60671 1\#183 ; 60450 1\#183 ; 60~01 1\#183 ; 59943 1-59710 8 1\#183 ; 59427 1\#183 ; 59173 1\#183 ; 58911 1\#183 ; 58651 1\#183 ; 58365 7 1\#183 ; 58086 1\#183 ; 57815 1\#183 ; 57536 1\#183 ; 57269 1\#183 ; 56994 I ~ ~ ~ ~ Q t'1 \gt ; t ' .
; J t_tj t ; :J ~ 0 0 ~ 0 z 1-'(j t'1 ~ t_tj ~ ~ t_tj w z \gt ; -t. .
; .
: .
J. .
\gt ; ~ ~ 0 ~ kj w \gt ; --3 ~ co co CJt TABLE IV .
Outer sheet .
cp .
cp2- P2\#183 ; ocp ' .
\lt ; j/ .
D+i .
D+i .
-- II 0 I II 0 I II 0 I II 0 I II 7 1\#183 ; 68444 6 I 1\#183 ; 68456 19 19 40 61 31 40 33 53 10 57 6 35 1\#183 ; 68464 20 21 44 61 38 55 35 53 10 57 9 40 1\#183 ; 68454 21 22 47 61 49 5 37 53 10 57 15 45 1\#183 ; 68454 22 22 32 62 2 40 39 53 10 57 25 0 6 1\#183 ; 68448 7 23 21 2 62 19 20 415310 57 37 15 1\#183 ; 68458 24 18 0 62 39 35 43 53 10 57 52 25 1\#183 ; 68452 25 13 38 63 2 45 45 53 10 58 10 40 1\#183 ; 68440 26 7 44 63 29 0 47 53 10 58 32 0 1\#183 ; 68452 27 0 1 63 59 0 49 53 10 58 57 5 8 1\#183 ; 68452 8 27 50 39 64 32 10 515310 59 24 50 1\#183 ; 68451 28 39 27 65 8 45 53 53 10 59 56 20 1\#183 ; 68454 29 26 17 65 49 0 55 53 10 60 31 20 1\#183 ; 68450 30 11 9 66 32 45 57 53 10 6110 0 1\#183 ; 68459 30 53 49 67 20 25 59 53 10 61 52 0 - 8 1\#183 ; 68448 9 31 34 29 G8 11 35 615310 62 38 45 1\#183 ; 68454 32 12 39 69 7 5 63 53 10 63 29 5 1\#183 ; 68464 32 48 23 70 6 50 65 53 10 64 23 50 1\#183 ; 68461 33 21 47 71 10 25 67 53 10 65 22 55 1\#183 ; 68452 33 52 43 72 18 15 69 53 10 66 26 35 9 1\#183 ; 68465 10 34 20 43 73 31 5 715310 67 35 10 1\#183 ; 68442 34 46 29 74 47 40 73 53 10 68 48 15 1\#183 ; 68447 35 9 5 76 9 35 75 53 10 70 6 45 1\#183 ; 68452 35 28 49 77 36 20 77 53 JO 71 30 5 1\#183 ; 68446 35 45 48 79 7 45 79 53 10 72 59 10 ---- 9 1\#183 ; 68444 11 35 59 45 80 44 15 81 53 10 74 33 10 Inner sheet , cp ' .
ocp ' .
P1\#183 ; ------ 0 J II II 19 36 20 6 .
1\#183 ; 57426 20 46 52 1\#183 ; 57144 21 56 27 1\#183 ; 56879 23 5 3 1\#183 ; 56616 24 12 31 1\#183 ; 56383 25 18 47 7 1\#183 ; 56153 26 23 47 1\#183 ; 55929 27 27 22 1\#183 ; 55716 28 29 26 1\#183 ; 55512 29 29 33 1\#183 ; 55342 30 28 13 8 1\#183 ; 55157 31 24 47 1\#183 ; 54996 32 19 19 1\#183 ; 54846 33 11 39 1\#183 ; 54708 34 1 58 1\#183 ; 54561 34 49 19 9 1\#183 ; 54460 35 34 26 1\#183 ; 54348 36 16 45 1\#183 ; 54251 36 56 17 1\#183 ; 54163 37 32 52 1\#183 ; 54082 38 6 15 10 1\#183 ; 54019 38 36 45 1\#183 ; 53948 39 3 42 1\#183 ; 53900 39 27 32 1\#183 ; 53846 39 47 36 1\#183 ; 53821 40 4 31 11 1\#183 ; 53774 o\#181 ; l. -- ~ c.:\gt ; O'l ~ ~ ~ t ' ~ N t:_tj td ~ 0 0 f1 0 z " " CJ t ' ~ z t:_tj ~ ~ t:_tj m H z \gt ; b : I H ~ ~ 0 ~ m 8 t MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
:-337 Section II.-Determination of the Position of the Planes of the Prism P R ' with reference to the Axes of Elasticity .
In order to compare these results with theory we require to know accurately the position of the principal planes of the two prisms .
We will consider the prism , faces PR , first .
Let O A , 0 B , 0 C be the directions of the principal axes of the wave surface , 0 B bisects the angle between the faces m m ' , 0 C is parallel to the intersection of these faces .
Let the normals to the faces mm ' cut a unit sphere , centre 0 , in MM ' , the normal to the face Pin P , and the normal to R in R. Let the direction angles of P and R with reference to A , B , C be a 1 / 31 -y1 , a 2 / 32 y2 respectively .
[ P lies in qnadrant A B C , R in A ' B C'.J Fig. 3 .
Let PM= 01 , PM ' =.0/ RM= 02 , RM'= 02 ' Observations on the angles between P and rn , P and rn ' give 01 , 0/ respectively .
While observations on the angles between R and m , Rand m ' give 02 , 0 , / respectively .
Let MM'= 2\#181 ; , = 2MB .
Then , as in the fil'st part , from triangles M ' B P , M P B we get 01 ' + 01 0/ -01 cos - 2- cos 2 cos / 31 = -------cos\#181 ; s.i n 0-/ - + 0\#183 ; 1 s.i n -0/ --01 2 2 cos ( )(.1 = -------cos\#181 ; with similar formulae for / 32 , a 2 .
MD0CCLXX1X .
2 X ( 1 ) ( 2 ) All the angles 01 , 0/ , 02 , 02 ' , and \#181 ; , were observed several times throughout the work .
The mean of the observations give 01 =54 ' 17 ' ti " .
0/ =75 ' 16 ' 20 " 0 ; 2 =88 ' 40 ' 35 " 0/ = 88 ' 20 ' 20 " f1 , =31 ' 53 ' 3011 ( 3 ) ( 4 ) ( 5 ) ( G ) ( 7 ) In no case was the difference between the mean and any one of the observations so great as 20 " .
Substituting these values we \#163 ; .nd Also whence a 1=71 ' 49 ' 3011 } f:31 = 60 ' 25 ' 49 " a:)=90 ' 19 ' 1011 } f:3z=88 ' 14 ' 32 " sin2 y 1= cos2 a 1+ cos2 / 31 si .
n2 y 2 + 2 Q 2 = cos az cos / J2 y1=35 ' 42 ' 57 " 1r-y2=l0 47 ' 11 " ( 10 ) ( 11 ) ( 12 ) ( 13 ) Let R O meet the sphere again in R ' , and let the figure represent the sphere as seen by an eye on O C produced .
Fig. 4 . . .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
339 We have now sufficient data to calculate the angle R ' P. 0P=y1 AP=al BP=,81 CR'=7r-y:2 AR'=7r-a:2=89 ' 40 ' 5011 B'R'=,82 .
From triangle B ' A R ' whence cos B'R ' = cos B ' AR ' sin AR ' cos B'R ' .\#183 ; .
cos B'AR'= .
ARE Slll ' cos / 32 - .Slll a 2 B'AR'= 88 ' 14 ' 32 " .\#183 ; .
R'AC= 1 ' 45 ' 28 " From triangle CAP cos OP= cos CAP sin AP cos OP .
\#183 ; .
cos CAP= ----:---A-- Ps111 .
\#183 ; .
CAP= 31 ' 17 ' 20 " .
\#183 ; .
R'AP=33 ' 2 ' 48 " From triangle R ' A P cos R'P = cos R ' A cos AP + sin R ' A sin AP cos R ' AP = cos ( 7r - a 2 ) cos a 1 + sin ( 7r - a 2 ) sin a 1 cos R ' AP whence R'P= 37 ' 2 ' 53 " But R'P can be measured directly , the mean of twelve observations , none of which differed from the mean by 1 O " , gave R'P=37 ' 2 ' 56 " The close agreement between these results confirms the accuracy of the values of a1 , .B1 , \amp ; c. Let P R ' eut C A in L , we proceed to determine C L , R ' L , and the angle A L P. From triangle A R ' L cot AL sin AR'= cot AR'P sin R ' AC+ cos AR ' cos R ' AC .\#183 ; .
cot AL sin a 2= cot AR'P sin R'AC+ cos a 2 cos R'AO we have seen already that R ' AC= 1 ' 45 ' 28 " For A R ' P we have s.i n AR'P = s.in RAP'P si .
n R'AP Sln .\#183 ; .AR'P=59 ' 18 ' 37 " Un substituting we get AL=88 ' 38 ' 18 " .
\#183 ; .CL= 1 ' 21 ' 42 " For R ' L the triangle R ' AL gives .
R'I _ sin R ' AL sin AL sm -- ' - sin AR'L whence R'L= 2 ' 2 ' 37 " For ALP the triangle ALP gives si .
n ALP = si.n ALPP si .
n LAP Sln LP=R'P-R'L=35 ' O ' 19 " Whence ALP=59 ' 20 ' 11 " Thus we have determined the position of the plane R ' P completely .
Section III.-Deterrrnination of the Position of the Plane P Q with reference to the Principal Planes .
( 15 ) ( 16 ) ( 17 ) Our next step will be to determine the position of P Q ' , Q ' being the point in which Q O produced backwards cuts the sphere .
Fig. 5 .
Let the direction angles of O Q ' be a3 / 33 y3 MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
341 Let then MQ'=03 M'Q'=03 ' cos 0/ + 03 03 ' -0 . .
~-~ cos -- .
, cos Q 2 2 / J3= -------cos\#181 ; .
03 ' + 03 .
03 ' -0 .
, Slil ~-~ Slll _._ _ , , cos 2 2 a 3= -------cos\#181 ; But 03 , 03 ' are the angles between the faces Q and m , Q and m ' respectively .
And observation shows that 03=19 ' O ' 15 " } 03 ' 64 ' 37 ' o " Whence / 33=35 ' 58 ' 41 " cos Q'BA = c_os aa sm/ 33 Q'BA=33 ' 37 ' 51 ' ' cos CQ ' = sin BQ ' sin Q ' BA .\#183 ; .y3=OQ'=71 ' O ' 44 " To corroborate these values let us find P Q ' and compare with experiment cos Q'AB = C-.O-S/ 33 Slll a 3 .\#183 ; .Q'AB=21 ' 54 ' 12 " But we know from the previous work ( Section II .
) that PAB=90'-PAC=58 ' 42 ' 42 " .\#183 ; .PAQ'=36 ' 48 ' 30 " cos PQ ' = cos AP cos AQ ' + sin AP sin AQ ' cos P AQ Whence PQ'=35 ' 18 ' 35 " The mean of a large series of experiments gave PQ'=35 ' 18 ' 50 " and again we have strong evidence in favour of the correctness of the work .
( 3 ) ( 4 ) ( 5 ) ( 6 ) Let the figure , as before , represent the sphe1e as seen by an eye on O C produced .
.l\lt ; 'ig .
ti .
Let Q ' P produced cut C A in L ' we proceed to determine C L ' , P L ' , and P L ' A. From the triangle L ' A P cot AL ' sin AP= cot APL ' sin L'AP+ cos AP ' cos L'AP From triangle P Q ' A si .
n APL ' -- s.i n PAQQ ' ' .
PAQ ' Slfi Sln " 1r-APL'=64 ' 41 ' 28 ' ' Also L'AP=CAP=31 ' 17 ' 20 " [ Section II .
from previous work .
J Substituting these values we get AL'=88 ' 44 ' 20 " CL'= 1 ' 15 ' 40 " The triangle A L ' P gives si .
n AL'P = s.i n AALP ' si .
n APL ' Sln whence AL'P=59 ' 13 ' 2 " .
Also .
L'P sin AL ' .
L'AP sin = sin APL ' sin whence L'P=35 ' 3 ' 14 " .
( 7 ) ( 8 ) ( 9 ) To corroborate these results draw L K perpendicular to P L ' , then L K L ' is approximately a right-angled triangle , and angle LL'K= 60 ' nearly .
Hence L'K=\#189 ; LL ' approximately Now L'K=PL ' -PL=3 ' approximately LL'=CL-CL'=6 ' approximately .\#183 ; .L'K is equal to \#189 ; LL ' MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
343 Section IV.-Genacil Theory .
To apply these results to the theoretical calculation .
Taking the same figure Fig. 7 .
Let the optic axes cut the sphere in O O ' ; consider a plane cutting the plane A O C between C and O ' jn L say ; let N be any point on it ; let NLA=x , NO=0 , NO'=0 ' .
Then we know that if v1 , v2 are the velocities along the wave normal thro1.1gh :N , a , b , c the prjncipal velocities , a being the greatest v12-vz2= ( a2-c2 ) sin 0 sin 0 ' v/ +vz2=(a2+c'2)-(a2-c2 ) cos 0 cos 0 ' 2v12= ( a2+c2)-(a2-c2 ) cos ( 0+0 ' ) 2vz2=(a2+c2)-(a2-c2 ) cos ( 0-0 ' ) .
We require , therefore , the values of 0 and 0 ' .
Now jf Similarly where cos 0= cos OL cos LN - sin OL sin LN cos x = cos OL(cos LN - tan OL sin LN cos x ) = cos OL cos ( LN +x .
) cos ; .\ , tan X.= tan OL cos x 0 , cos O'L ( LN , ' ) COS = -- , COR -A CORA ' tnn X.'= tan O'L cos X\#183 ; ( 1 ) ( 2 ) ( 3 ) ( 4 ) Thus to use the formulae for 1 , i , v2 we regnire to know ac~mrately the position of the optic axes .
These can only be determined by finding the values of a , b , c. The results of the experiments enable us to do this with considerable exactness .
Let / La , / Lb , / Le be the principal refractive indices , / La being the greatest , / Le the least / La=~ Section V.- Values of the P1\#183 ; incipal Refractive Indices .
We are now in a position to determine from our experimental results the values of a , b , c the principal velocities .
The plane P R ' cuts the plane A Oat a short distance from 0 .
The section of the surface of wave slowness by the plane A O O consists of a circle ~\gt ; f radius / Lb and an oval curve of axes / La and / Le in the directions O 0 , 0 A respectively , which for a small distance on either side of a principal axis may be treated as an ellipse ; as in FRESNEL 's theory .
Also , if L be the point in which P R ' cuts the plane A O 0 LP=35 ' O ' 19 " [ vide Section II .
( 16 ) ] .
The value of p , corresponding to a value sheet will thus be the value of / Ll\#183 ; But we have as corresponding values we may therefore take \lt ; f , ' 34 ' 28 ' 43 " 34 ' 53 ' 57 ' ' 35 ' 16 ' 20 ' ' of \lt ; f , '=35 ' O ' 191 ' in Table I. for the inner fL 1'68115 1"68114 1\#183 ; 68115 The values of the radius vector to the outer sheet in the same neighbourhood are , Table I. , lines 4 and 5 , For \lt ; f , ' 34 ' 47 ' 54 ' ' 35 ' 10 ' 1 " \#181 ; .
1 \#183 ; 68546 l '68553 .
\lt ; t , '=35 ' o ' 19 " MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
345 we may take therefore p. , =l \#183 ; 68550 This then must be the length of the radius vector of an ellipse , axes / La and / Le , which passes through the point L--i.e .
, which is inclined at an angle .CL= 1 ' 21 ' 42 " to the axis / La [ vide Section II .
( 15 ) ] .
Now if 0 be the angle which a radius vector 1\#183 ; makes with the axis / La in the case considered 1 cos2 0 sin2 0 -=-+- 1\#183 ; 2 \#181 ; / ' \#181 ; l 0=1 ' 21 ' 42 " sin2 0= \#183 ; 000565 and the error made in giving to / Le an approximate value will be inappreciable .
Let us take , therefore , for / Le the value found in the previous part which agrees with RuDBERG 's value Now whence P. , e= 1 \#183 ; 53013 r= 1 \#183 ; 68550 - =\#183 ; 351919 rZ 1 1 ( 1 sin2 0 ) \#181 ; , ,2= cos2 0 r 2 - c2 P. , a= 1 \#183 ; 68560 It now remains to determine / Le accurately .
The natural faces a , m , ' ( fig. 1 ) formed a prism whose edge was parallel to the axis O C. By passing light through these faces and observing the incidence and deviation as before , I was enabled to calculate / L e\#183 ; To pass light through these faces it was necessary , however , to polish them , and this operation altered their position appreciably .
The values obtained by direct measurement required a small correction , and to calculate this it was necessary to determine the position of these new faces .
Let the faces approximately coincident with a , and rrn , ' be called X Y respectively .
On levelling the faces of reference mm ' , I found that Y remained very nearly jn the same zone as before , viz. : that of m m ' m , m/ ; while X fell rather below this zone .
Observations on the angles between X and Y and the faces mm ' respectively gnve XM ( =04 say)=l23 ' 32 ' XM'(=0'4 say)= 59 ' 52 ' arc YA'M ( = 05 say)= 11 7 ' 17 ' arc YA'M'(=05 say)=l81 ' 3 ' Fig. 8 .
Since Y , M , M ' , are in one zone with A and B , and B bisects arc MM ' 2BA'Y =M ' A'Y + MA'Y .
\#183 ; .
arc BA'Y=149 ' 10 ' A'Y=l49 ' 10'-90 ' =59 ' 10 ' Let a 4 / 34 y 4 be the direction angles of X. Then , as before , [ Section II .
( 1 ) J whence .
0'4+04 .
014-04 Sln .
- 2- Sln 2 cos a4=-------- \amp ; c. cos\#181 ; a4= 3 ' 28 ' 5011 / 31.=91 ' 42 ' To test these values let us calculate the angle X Y ; the triangle X A B gives XAB- cosXB cos - .
AX Slll n. whence 1r-XAB=60 ' 43 ' 5011 In the triangle X A Y we now know A X , A Y and the angle X A Y. MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAX .
A.L CRYSTAL .
347 Solving , we get Experiment gave XY=119 ' 5 ' 119 ' 4 ' Here again we have considerable confirmation .
( 1 ) Let the plane Y X L be the plane of the prism cutting C A C ' in L ; it will be necessary to find A L , X L , and the angle A L Y. The right angled triangle L A Y gives sin A Y = tan AL cot A YL .
The triangle X A Y gives .
LYA- sin AX sin YAX sin - .
XY Sll1 whence LYA=3 ' 28 ' 30 " Substituting this value , we get AL=2 ' 59 ' 10 " .
From the same triangle we get cos YL = cosA Y cos AL whence YL=120 ' 47 ' 10 " but XY=ll9 ' 4 ' .\#183 ; .
XL=1 ' 43 ' 10 " From the same triangle again cos ALY= tan AL cot LY whence r-ALY=91 ' 47 " ( 2 ) ( 3 ) Now we know that if 0 , O ' are the points where the optic axes cut the unit sphere K , the point where any wave normal meets the sphere if KO=0 KO'=0 ' 1 1 ( 1 1 ) , 2v2=-+-- --- cos ( 0-0 ) f1 , a2 \#181 ; ?
f1 , a2 \#181 ; ? .
\#183 ; .
--\#190 ; ( 1 + cos ( 0-0 ' ) ) fl , c =~-- ; ( 1- cos ( 0-0 ' ) ) \#181 ; 2 fl , a " - cos2 - =- ; ;- 0 su12 -- 1 ( 0-0 ' ) 1 1 .
( 0-0 ' ) f1 , c2 2 W fl , a " 2 ( 4 ) If then we can determine 0-0 ' since \#181 ; , is given by experiment , this formula determines the value of / Le\#183 ; Now along the plane X Y , 0-0 ' is very small , and a considerable change in the positions of O O ' will produce but small variations in the value of 0-0 ' and .
\#183 ; .
of / Le\#183 ; Let us , therefore , take the value of the angle between the optic axes given by the values already determined for / La and / Lb , viz. : / La= 1 \#183 ; 68560 / Lb= l '68115 and the approximate value of / Le fLc= l \#183 ; 53013 If 2\lt ; /\gt ; ' is the angle between the optic axes whence Hence \lt ; f , ' = 9 ' 4 ' 5 / / AO=AO'=80 ' 55 ' 55 " AL= 2 ' 59 ' 10 ' ' OL= 83 ' 55 ' 5 " O'L=77 ' 5G ' 45 , .
, x=ALK=88 ' 13 ' XL=1 ' 43 ' 10 " MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
349 Let K be on the side of L remote from X. Experiment shows that this is the actual position of the wave normal in the case considered .
if Let KX=l/ J ' .\#183 ; .
LK=KX-XL=l/ J'-1 ' 43 ' 10 " cos 0= cos OL cos LK+ sin OL sin LK cos x= cos OL cos ( LK-X .
) cos A , tan A= tan OL cos X Substituting the values of O L and x Similarly X.= 16 ' 17 ' 1 O " cos OL log - --1 \#183 ; 0427912 cos A , LK-X.=l/ J'-18 ' O ' 20 " cos 0 , = c-o-s O\#183 ; '-L cos ( LK + X. ' ) cos A , where tan X. ' = tan O'L cos x .\#183 ; .
X.'=8 ' 17 ' 30 " 1 cos O'L og cos :x , .
1 " 324:3204 LK+X.'="1'+6 ' 34 ' 20 " A table on the next page gives the values of l/ J , the angle made by the incident wave normal with the normal to X ; D+i the deviation + the angle X Y ; tfJ ' the angle made by the wave normal in the crystal with the normal to X ; and \#181 ; ,2 ' \#181 ; , being the reciprocal of the wave velocity .
TABLE to determine the Value of / .Le\#183 ; , t , .
D+i .
, J , ' .
1 \#181 ; 2 .
0 I II 0 I " 0 I II 51 16 7 101 47 15 30 39 0 \#183 ; 427076 49 41 7 101 48 45 29 53 13 \#183 ; 427061 47 49 7 101 57 45 28 57 43 \#183 ; 427049 46 5 22 102 13 0 28 6 3 \#183 ; 427039 44 26 7 102 34 45 27 14 5 \#183 ; 427029 42 55 27 103 3 5 26 25.31 \#183 ; 427018 Substituting the values of \#183 ; if/ from this table in the formulae for cos 0 and cos 0 ' we can :find 0 and 0 ' .
The values of 0-0 ' are 0 , II Substituting these values m the formula for / .Le , the mean of the results agrees closely with Thus we may take f-1-e= 1 \#183 ; 53013 J.La= 1 \#183 ; 68560 / .Lb=1\#183 ; 68115 J.Le= 1 \#183 ; 53013 And the angle between O C and the optic axes is 9 ' 4 ' 5 " .
If \lt ; p is the angle as seen in air \lt ; /\gt ; =15 ' 21 ' 50 " Th~ value KIRCHHOFF found by experiment was 15 ' 27 ' ( 5 ) The close agreement is noteworthy .
As in the first part , the agreement is much closer than it would be taking RuDBERG 's Yalnes of / J-a , / J-b , f-1-\lt ; \#183 ; MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
351 Section VI.-Application of the Theory to Experiment .
-Tables giving Theoretical Values of p,.-Discussion of Results .
We can now continue the work of calculating the theoretical values of the reciprocal of the velocity .
Consider first the plane P R ' .
With the same notation we have [ Section IV .
( 3 and 4 ) ] .
and Whence Fig. 10 .
cos OL cos 0= -cos- cos ( NL+L ) .
i\ , tan A= tan OL cos ALN cos O'L ' cos 0'= 1 cos cos(NL-\gt ; . .
' ) i\ , tan \gt ; . .
' = tan O'L cos ALN 00=0'0=9 ' 4 ' 5 " [ Section V. ( 5 ) ] LC=1 ' 21 ' 42 ' ' [ Section II .
( 15 ) ] .
\#183 ; .
OL= lOU 25 ' 47 ' ' .
O'L= 7 ' 42 ' 23 " ALN=x=59 ' 20 ' 11 " [ Section II .
( 17 ) ] .
A.= 5 ' 21 ' 47 " .
lo cos OL = 1 \#183 ; 9946698 g cos i\ , \gt ; . .
' =3 ' 56 ' 52 " .
loo- cos O'L 1\#183 ; 9970915 b cos A.1 ( 1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) ( 6 ) ( 7 ) ( 8 ) LN=LP-PN=LP-\lt ; /J'=35 ' 0 ' 19"-\lt ; /J ' .
LN +"-= 40 ' 22 ' 6 " -\lt ; /J ' .
LN-\gt ; ..'=31 ' 3 ' 27"-\lt ; /J ' .
( 9 ) ( 10 ) ( 11 ) Substituting these values we can calculate 0 and 0 ' , and hence find the values of v1 and v2 or their reciprocals / Li and f-Lz\#183 ; Tables V. and VI .
give the results of this substitution : on the right hand for the inner sheet , on the left hand for the outer .
The centre columns give the values of L N +\gt ; . .
, L N -A ' .
Then on either side , 0+ 0 ' for the inner sheet , and 0-0 ' for the outer .
Then the theoretical values of / Li / Lz , then the experimental values taken from Tables I. and II .
, and finally , in the outside column , the excess of theory above experiment .
The next step is to discuss the theory for the plane P Q. With the same figure and notation as before , if this plane cut the plane A 0 in L ' we have [ Section IV .
( 3 and 4 ) J A , A.1 1 being values of A A ' in this case .
where whence Fig. 11 .
cos OL ' I cos 0= -- cos ( L N+\gt ; . .
) COS A , / / tan A , = tan 0L ' cos AL'N 0 , _ cos O'L ' ( L'N ,. .
' ) COS - 1 COS -A , COS A , / / tan \ ' = tan 0'L ' cos AL'N 00=00'=9 ' 4 ' 5 ' ' [ Section V. ( 5 ) ] CL'= 1 ' 15 ' 40 " [ Section III .
( 7 ) J .
\#183 ; .
OL'=10 ' 19 ' 45 " 0'L ' = 7 ' 48 ' 25 ' ' .
AL'N =x ' =5g0 13 ' 2 " [ Section III .
( 8 ) ] ( 12 ) ( 13 ) ( 14 ) ( 15 ) :MR .
R. T. GLAZEBROOK ON PLANE WAVES IN .A BI .
AX.AL CRYSTAL .
:-353 X. , =5 ' 19 ' 45 " ( 16 ) x. , '=4 ' o ' 50 " .
( 17 ) L'N=L'P+PN ( N being now on the side of P remote from L ' ) =35 ' 3 ' 14"+\lt ; /\gt ; ' .
L'N +X .
, =40 ' 22 ' 5911 +c/ \gt ; ' L'N-X .
, '=31 ' 2 ' 24''+1\gt ; ' cos OL ' log--=l \#183 ; 9947864 .
cos A- , cos O'L ' - log -x , =1\#183 ; 9970226 .
COS , I ( 18 ) ( 19 ) ( 20 ) ( 21 ) and as before , we can calculate from these data the values of 0 , 0 ' and hence of / J , i JJ,2 .
Tables VII .
and VIII give the results of this .
The fifth column gives the values of L N +x .
for the inner sheet the next L N -X .
' ' ' the next 0+ 0 ' , the eighth the value of the reciprocal of the velocity calculated on FRESNEL 's theory [ by means of Section IV .
( 1 and 2 ) ] , 9 .
; 2=a2+c2--(a2-c2 ) COS ( 0+ 0 ' ) 2=a2+c2-(a2-c2 ) cos ( 0-0 ' ) fi , z the ninth , the same quantity from experiment , and the last , the excess of the theoretical over the experimental value .
The values of 0-0 ' given in the fourth column are calcnlated for the inner sheet .
Now as the values of \lt ; /\gt ; ' corresponding to the same incident wave are different for the two sheets , the values of 0 and 0 ' will be different ; but on referring to the values it will be seen that after the first twenty observations , 0-0 ' is very nearly constant throughout the arc considered ; so that JJ,2 varies very slowly , and except just at first , we may , to the degree of approximation required , use the value of 0 and 0 ' for the inner sheet in calculating the theoretical value of the reciprocal of the velocity for the outer sheet .
In the case of the first twenty observations for which 0-0 ' varies appreciably with \lt ; /\gt ; ' , the experimental values of JJ , z are obtained _by interpolation from those valnes in Table I. , to which they correspond .
Table V. contains the theory for I. , , VI .
, , , , II .
" " VII .
VIII .
" " " " III .
lV .
2 z TABLE V. Outer sheet .
I LN+ ; \gt ; ... LN- ; \ ' .
Difference .
, u2 Experiment .
I \#181 ; ~ Theory .
0-0 ' .
0 I " 0 I " 0 I 4 15 50 5 2 4"9 - 4 1\#183 ; 68558 1\#183 ; 68554 2 6 50 4 46 7 4 32 32 - 2 1\#183 ; 68553 1\#183 ; 68551 5 5 46 4 12 53 + 1 1\#183 ; 68546 1\#183 ; 68547 5 28 9 3 50 30 + 5 1\#183 ; 68539 1\#183 ; 68544 5 53 23 3 25 16 + 9 1\#183 ; 68532 I 1\#183 ; 68541 6 2112 2 57 27 +10 1\#183 ; 68527 1\#183 ; 68537 4 12 20 6 51 42 2 26 57 + 9 1\#183 ; 68520 1\#183 ; 68529 7 24 45 1 53 54 +20 1\#183 ; 68502 1\#183 ; 68522 8 0 14 118 25 + 0 1\#183 ; 68515 1\#183 ; 68515 8 38 15 0 40 24 + 9 1\#183 ; 68498 1\#183 ; 68507 9 18 10 0 0 29 + 8 1\#183 ; 68490 1\#183 ; 68498 6 44 30 10 0 38 0 41 59 + 3 1\#183 ; 68486 1\#183 ; 68489 10 44 46 1 26 7 +16 1\#183 ; 68465 1\#183 ; 68481 11 31 5 2 12 26 + 1 1\#183 ; 68473 1\#183 ; 68474 12 20 13 3 1 34 + 4 1\#183 ; 68463 1\#183 ; 68467 13 9 18 3 50 39 - 8 1\#183 ; 68454 1\#183 ; 68462 8 28 10 14 0 52 4 42 13 + 1 1\#183 ; 684,57 1\#183 ; 68458 14 54 8 5 35 29 - 3 1\#183 ; 68457 1\#183 ; 68454 15 48 55 6 30 16 - 8 1\#183 ; 68459 1\#183 ; 68451 16 45 11 7 26 32 - 3 1\#183 ; 68451 1\#183 ; 68448 17 42 53 8 24 14 - 3 1\#183 ; 68448 1\#183 ; 68445 9 10 40 18 41 48 9 23 9 1\#183 ; 68441 19 42 3 10 23 24 1\#183 ; 68440 20 43 28 11 24 49 1\#183 ; 68461 21 46 6 12 27 27 1\#183 ; 68446 22 50 40 13 31 1 - !
) 1\#183 ; 68448 1\#183 ; 68439 9 26 20 23 54 16 14 35 37 - -Inner sheet .
0+ 01 .
\#181 ; 1 Theory .
/ Li Experiment .
n I , , 18 14 0 1\#183 ; 68111 1\#183 ; 68102 18 11 50 1\#183 ; 68113 1\#183 ; 68112 18 9 50 1\#183 ; 68114 1\#183 ; 68106 18 8 30 1\#183 ; 68115 1\#183 ; 68115 18 8 10 1\#183 ; 68115 1\#183 ; 68114 18 9 40 1\#183 ; 68114 1\#183 ; 68115 18 13 10 1\#183 ; 68111 1\#183 ; 68106 18 19 40 1\#183 ; 68106 1\#183 ; 68099 18 29 50 1\#183 ; 68097 1\#183 ; 68089 18 44 30 1\#183 ; 68085 1\#183 ; 68074 19 4 30 1\#183 ; 68068 1\#183 ; 68069 \#183 ; 19 30 50 1\#183 ; 68046 1\#183 ; 68030 20 3 50 1\#183 ; 68017 1\#183 ; 68016 20 44 10 1\#183 ; 67980 1\#183 ; 67962 21 32 50 1\#183 ; 67936 1\#183 ; 67917 22 29 0 1\#183 ; 67880 1\#183 ; 67859 23 31 30 1\#183 ; 67817 1\#183 ; 67806 24 42 10 1\#183 ; 67743 1\#183 ; 67721 26 0 30 1\#183 ; 67656 1\#183 ; 67634 27 23 50 1\#183 ; 67560 l \#183 ; 67532 28 53 10 1\#183 ; 67463 1\#183 ; 67421 30 28 20 1\#183 ; 67331 1\#183 ; 67303 32 7 0 1\#183 ; 67200 1\#183 ; 67161 33 52 0 1\#183 ; 67054 1\#183 ; 67015 35 40 30 1\#183 ; 66896 1\#183 ; 66854 37 33 20 1\#183 ; 66726 1\#183 ; 66694 39 26 50 1\#183 ; 66547 1\#183 ; 66502 41 26 0 1\#183 ; 66352 1\#183 ; 66300 Difference .
09 01 08 00 01 -01 05 07 08 09 -01 i:.o C , ~ ~ ~ ~ t"\lt ; pN trj td i ; d 0 0 rl 0 z " O t"\lt ; ~ t , j s t , j [ I)_ z ~ td H -~ \gt ; -1 ; .,. .
t"\lt ; 0 ~ [ I)_ 8 p ... ~ ~ N ~ Outer sheet .
Difference .
f'~ Experiment .
\#181 ; 2 Theory .
1\#183 ; 68433 1\#183 ; 68458 1\#183 ; 68416 1\#183 ; 68445 1\#183 ; 68448 -21 1\#183 ; 68458 1\#183 ; 68437 1\#183 ; 68453 1\#183 ; 68448 1\#183 ; 68445 1\#183 ; 68446 -15 1\#183 ; 68452 1\#183 ; 68437 1\#183 ; 68446 1\#183 ; 68451 1\#183 ; 68450 1\#183 ; 68454 -13 1\#183 ; 68450 I 1\#183 ; 68437 1\#183 ; 68445 I 1\#183 ; 68445 1\#183 ; 68453 1\#183 ; 68437 I --10 I 1\#183 ; 68447 1\#183 ; 68437 1\#183 ; 68428 1\#183 ; 68451 I LN+X .
0-0 ' .
0 I II 0 I II TABLE VI .
LN-X .
' .
0+01 .
0 I II 0 I II - Inner sheet .
\#181 ; 1 Theory .
p1 Experiment .
Difference .
1\#183 ; 66702 1\#183 ; 66678 24 1\#183 ; 66534 1\#183 ; 66490 44 1\#183 ; 66341 1\#183 ; 66302 39 1\#183 ; 66143 1\#183 ; 66095 48 1\#183 ; 65939 1\#183 ; 65883 56 1\#183 ; 65730 1\#183 ; 65667 63 1\#183 ; 65520 1\#183 ; 65459 61 1\#183 ; 65306 1\#183 ; 65240 66 1\#183 ; 65098 1\#183 ; 65037 61 1\#183 ; 64884 1\#183 ; 64818 66 1\#183 ; 64674 1\#183 ; 64603 71 1\#183 ; 64466 1\#183 ; 64400 66 1\#183 ; 64264 1\#183 ; 64203 61 1\#183 ; 64055 1\#183 ; 63984 69 1\#183 ; 63872 1\#183 ; 63782 90 1\#183 ; 63678 1\#183 ; 63615 I 63 1\#183 ; 63509 1\#183 ; 63433 76 1\#183 ; 63340 1\#183 ; 63258 82 1\#183 ; 63182 1\#183 ; 63100 82 1\#183 ; 63030 1\#183 ; 62956 74 1\#183 ; 62897 162807 90 1\#183 ; 62776 1\#183 ; 62691 85 1\#183 ; 62662 1\#183 ; 62583 79 ~ ~ ~ " " 3 Q t " ~ N trj td ~ 0 0 Vi 0 z " 'd t- , ~ t , :_ ; ~ ~ t , : .
: : U1 H z p , .
t:d H p , ~ ~ t " 0 ~ U1 ,. .
: : ~ ~ ~ C)1 C)1 TABLE VII .
- I Outer sheet .
LN+X .
LN-X ' .
Difference .
\#181 ; ~ Experiment .
\#181 ; ~ Theory .
0-0 ' .
---\#183 ; - --0 I II I 0 I II 0 I II -27 1\#183 ; 68464 1\#183 ; 68437 9 29 50 I 39 52 39 30 32 4 1\#183 ; 68450 40 20 13 30 59 38 1\#183 ; 68454 40 50 20 31 29 45 1\#183 ; 68450 41 23 2 32 2 27 1-68458 41 58 27 32 37 52 -HJ 1\#183 ; 68456 1\#183 ; 68437 9 29 30 42 36 25 33 15 50 1\#183 ; 68464 43 16 25 33 55 50 Hi8462 43 58 52 34 38 17 1\#183 ; 68469 44 43 27 35 22 52 1\#183 ; 68462 45 30 22 36 9 47 I -15 1\#183 ; 68453 1\#183 ; 68438 9 28 30 46 19 11 36 58 36 1\#183 ; 68457 48 2 58 38 42 23 I 1\#183 ; 68454 48 57 29 39 36 54 1\#183 ; 68462 49 53 45 40 33 10 -28 I 1\#183 ; 68466 1\#183 ; 68438 9 27 20 50 51 59 41 31 24 1\#183 ; 68549 51 51 24 42 30 49 1\#183 ; 68465 52 52 38 43 32 3 1\#183 ; 68475 54 59 31 45 38 56 I -18 1\#183 ; 68457 1\#183 ; 68439 9 26 20 56 4 46 46 44 11 I 1\#183 ; 68464 57 11 23 47 50 48 1\#183 ; 68465 58 19 14 48 58 39 1\#183 ; 68458 59 28 9 50 7 34 1\#183 ; 68465 60 38 12 5117 37 -36 1\#183 ; 68475 1\#183 ; 68439 9 24 30 61 49 11 I 52 28 36 Inner sheet .
0+ 01 .
f'i Theory .
\#181 ; 1 Experiment .
0 I II 71 52 50 1\#183 ; 62682 1\#183 ; 62619 72 46 30 1\#183 ; 62563 1\#183 ; 62500 73 45 10 1\#183 ; 62430 1\#183 ; 62360 74 48 50 1\#183 ; 62268 1\#183 ; 62208 75 58 0 1\#183 ; 62131 1\#183 ; 62058 77 12 10 1\#183 ; 61965 1\#183 ; 61904 78 30 20 1\#183 ; 61786 1\#183 ; 61715 79 53 30 1\#183 ; 61600 1\#183 ; 61525 81 20 20 1\#183 ; 61400 1\#183 ; 61320 82 52 0 1\#183 ; 6119'2 1\#183 ; 61120 84 27 50 1\#183 ; 60974 1\#183 ; 60897 86 7 40 1\#183 ; 60748 1\#183 ; 60671 87 51 10 1\#183 ; 60514 1\#183 ; 60450 89 38 10 1\#183 ; 60285 1\#183 ; 60201 90 ' + 1 28 40 1\#183 ; 60025 1\#183 ; 59943 - 90 '+ 3 23 0 1\#183 ; 59769 1\#183 ; 59710 " 5 19 50 1\#183 ; 59510 1\#183 ; 59427 " 7 20 20 1\#183 ; 59246 1\#183 ; 59173 " 9 23 40 1\#183 ; 58977 1\#183 ; 58911 " 11 30 0 1\#183 ; 58705 1\#183 ; 58651 90 ' +13 38 30 1\#183 ; 584:32 1\#183 ; 58365 " 15 49 30 1\#183 ; 58155 1\#183 ; 58086 " 18 3 30 1\#183 ; 57880 1\#183 ; 57815 " 20 19 30 1\#183 ; 57605 1\#183 ; 57536 " 22 37 40 1\#183 ; 57329 1\#183 ; 57269 90 ' +24 57 50 1\#183 ; 57058 1\#183 ; 56994 - --- , .
D ifforenoe .
74 ut : : co c. ; 1 C ) ~ ?
d ?
d t\lt ; p , -N t , :j t:d ~ 0 0 Pi 0 z 1-"d t\lt ; ~ t , :j ~ l:rj w. H z p , -- t:d H p , -~ p , -- t\lt ; 0 ~ w. -'3 p , -~ I Outer sheet .
- .
LN + ; \gt ; ... Difference .
\#181 ; 2 Experiment .
\#181 ; 2 Theory .
/ J-0 ' .
0 I / / 0 I / / - 5 1\#183 ; 68444 1\#183 ; 68439 9 25 10 59 59 19 I 1\#183 ; 68456 61 9 51 I 1\#183 ; 68464 62 19 26 I 1\#183 ; 68454 63 28 2 1\#183 ; 68454 64 35 30 - 8 1\#183 ; 68448 1\#183 ; 68440 9 22 30 65 41 46 1\#183 ; 68458 66 46 46 H18452 67 50 21 1\#183 ; 68440 68 52 25 1\#183 ; 68452 69 52 32 -12 1\#183 ; 68452 1\#183 ; 68440 9 21 50 70 5112 1\#183 ; 68451 71 47 46 I 1\#183 ; 68454 72 42 18 I 1\#183 ; 68450 73 34 38 1\#183 ; 68459 74 24 57 - 7 1\#183 ; 68448 1\#183 ; 68441 9 20 0 75 1118 1\#183 ; 68454 75 57 25 1\#183 ; 68464 76 39 44 1\#183 ; 68461 77 19 16 1\#183 ; 68452 77 55 51 -23 1\#183 ; 68465 1\#183 ; 68442 9 19 50 78 29 14 1\#183 ; 68442 78 59 44 l \#183 ; 68447 79 26 41 1\#183 ; 68452 79 50 31 1\#183 ; 68446 80 10 35 - 2 1\#183 ; 68444 1\#183 ; 68442 9 19 20 so\#183 ; 27 39 TABLE VIII .
Inner sheet .
LN- ; \gt ; . .
' .
0+U ' .
\#181 ; 1 Theory .
0 I / / 0 0 I / / 50 38 44 90+21 20 50 1\#183 ; 57482 514916 " 23 40 0 1\#183 ; 57209 52 58 51 " 25 57 30 1\#183 ; 56944 54 7 27 " 28 13 0 1\#183 ; 56690 55 14 55 " 30 26 20 1\#183 ; 56446 56 2111 90+32 36 10 1\#183 ; 56215 57 26 11 , , 34 45 50 1\#183 ; 55991 58 29 46 , , 36 5110 1\#183 ; 55779 59 31 50 , , 38 54 20 1\#183 ; 55580 60 31 57 , , 40 52 50 1\#183 ; 55393 61 30 47 90+42 48 50 1\#183 ; 55216 62 27 11 : , 44 40 50 1\#183 ; 55051 63 21 43 46 28 40 1\#183 ; 54899 64 14 3 " , , 48 12 0 1\#183 ; 54756 65 4 22 " 49 51 40 1\#183 ; 54623 65 51 43 90+51 24 20 1\#183 ; 54515 66 36 50 52 54 40 1\#183 ; 54393 67 19 9 " 54 18 20 1\#183 ; 54293 67 58 41 " 55 36 40 1\#183 ; 54203 68 35 16 " " 56 49 0 1\#183 ; 54123 69 8 39 90+57 55 10 1\#183 ; 54051 69 39 9 " 58 55 30 1\#183 ; 53988 70 6 6 59 49 0 1\#183 ; 53934 70 29 56 " 60 36 10 1\#183 ; 53887 70 50 0 " " 6116 0 1\#183 ; 53848 I 71 6 55 90+61 49 20 1\#183 ; 53817 P1 Experiment .
Difference .
1\#183 ; 57426 56 1\#183 ; 57144 65 1\#183 ; 56879 65 1\#183 ; 56616 74 1\#183 ; 56383 63 1\#183 ; 56153 62 1\#183 ; 55929 62 1\#183 ; 55716 63 1\#183 ; ,55512 68 1\#183 ; 55342 51 1\#183 ; 55157 59 1\#183 ; 54996 55 1\#183 ; 54846 53 1\#183 ; 54708 48 1\#183 ; 54561 62 1\#183 ; 54460 55 1\#183 ; 54348 45 1\#183 ; 54251 42 1\#183 ; 54163 40 1\#183 ; 54082 41 1\#183 ; 54019 32 1\#183 ; 53948 40 1\#183 ; 53900 34 1\#183 ; 53846 41 1\#183 ; 53821 27 1\#183 ; 53774 33 ~ ?
J ?
J :3 Q t " P\gt ; N t.zj t:d ~ 0 0 ~ 0 z '"d t " ~ t.zj ~ U1 H z ~ t:d r-i ~ P\gt ; t " 0 ~ r-1 U1 ~ ~ ~ CJ ] '-l 358 J\IR .
R. 'l ' .
GLAZl~BHOOK ON PLANE WAVES IN A BIAXAL .CRYSTAL .
Let us consider first the inner sheet .
In Table V. we see the differences are small to begin with , and increase fairly regularly throughout the whole of the table , the experimental value being almost always less than the theoretical .
_ Lines 4 , 5 , and 6 were those chosen to determine the value of / Lb which explains the coincide11ce of the two curves in that neighbourhood .
The table covers an arc of the surface of nearly 20 ' , as is seen by referring to the values of \lt ; { \gt ; ' in Table I. Throughout this arc the theoretical section lies outside the experimental , the difference between the two increasing as we proceed from the axis / Lb\#183 ; This continues throughout Table VI .
The first three lines of Table VI .
just overlap the last three of Table V. The difference reaches a maxim nm value towards the end of Table VI .
at a point for which cf\gt ; ' = 1 ' 19 ' 54 " ( vide Tables II .
, 21 , and VI .
, 21 ) , that is at about 34 ' away from the axis / Lb\#183 ; After this the difference begins to decrease at first slowly throughout Table VII .
which refers to the other prism P Q , so that the arc of section has changed slightly between Tables VI .
and VII .
, both arcs passing through the point P. Table VII .
covers an arc of 22 ' ( Table III .
) .
The first two lines in Table VII .
refer to a wave normal lying on the same side of P as L. In the rest of the table the wave normal is on the side of P remote from L. This arc is continued throughout Table VIII .
; the theoretical values of \#181 ; , are still greater than the experimental , but the difference diminishes as we proceed along the arc .
The last three observations in Table VII .
and the first three in Table VIII .
overlap .
Table VIII .
covers an arc of 20 ' .
Thus the observations extend over two arcs , each passing through the point P inclined to one another at that point at so small an angle as to be almost continuous , the one arc being 35 ' 40 ' , and the other 40 ' 40 ' in length .
the results of theory and experiment agree at the extremity of the first arc , .
differ most widely in the neighbourhood of the point where the two arcs meet , and tend towards equality again throughout the second arc .
It is worth noticing that the experimental results at the end of Table VI .
, the encl of the fin-it arc , agree closely with those at the beginning of Table VII .
, the beginning of the second arc .
The experimental results for the inner sheet are therefore represented by a curve which coincides with FRESNEL 's curve at the extremity of one axis , and lies inside it throughout the next 7 5 ' of its length .
For the outer sheet the differences are less than for the inner sheet .
0-0 ' varies so slowly that it seemed sufficient for the purposes of comparison to calculate every fifth theoretical value , with the exception of that portion of Table V. for which the variations of 0-0 ' are sensible .
The section cuts the principal section A O C hetween lines -1 and 5 , Table V. Just MR. R. T. GLAZEBROOK ON PLANE W .A .
VES IN .A .
BI .
A.X.A.L CRYST .
A.L. 359 at that point the theoretical value is greater than the experimental .
It continues so until we reach line 18 , Table V. , that is for an arc of about 9 ' .
From that point onwards the theoretical value of \#181 ; 2 is uniformly less than the experimental Thus the result of experiment for the outer sheet is represented by a curve which coincides with the theoretical section at the point in which it cuts the plane A O C , then lies inside it for an arc of about 9 ' .
At this point the two again coincide , and for the rest of the arc experimented on , the section given by experiment lies outside that given by theory ; the difference between the two increasing throughout Tables VI .
and VII .
, and diminishing again in Table VIII .
Section VII .
---Effect of possible Errors in the Valites of JJ .
, a , \#181 ; 0 , Jl , c di : : ; cussed .
The next task will be to determine the effect of any small errors which may have occurred : ( 1 .
) In the determination of JJ .
, a , \#181 ; 0 , JJ .
, c ; ( 2 .
) In the angles which fix the position of the planes P R , P Q relatively to the axes of the surface .
( 1 .
) .
Jl , a\#183 ; To determine Jl , a we considered the length of the radius vector common to the principal section A O C and the section by the plane P R. This radius vector is fixed in length , and is on FRESNEL 's theory the radius vector of an ellipse , axes JJ .
, a and Jl , c , inclined at a small angle to Jl , a\#183 ; The increase in the value of Jl , a will therefore increase the value of this angle , that is , it will alter the position of the plane P L R ' and be more properly con sidered under ( 2 ) .
Jl , b is determined by the intersection of the section of the inner sheet by the plane P R , with a circle of radius \#181 ; 0 , and is therefore fixed at least within the limits of experimental error .
Jl , c is found from an independent and , on the whole , less trustworthy series of observations-less trustworthy because the faces were much less plane than P , Q , R. Let us consider the effect of decreasing the value of Jl , c\#183 ; If E be \#183 ; the angle between the optic axes .
As JJ .
, c decreases tan E decreases ; therefore E decreases .
Let us consider how thiR affects 0 and 0 ' .
Taking the same figure as previously .
Fig.12 .
Let O ' K be an arc perpendicular to A C. Let 0 , O , ' be the new positions of O O ' .
Then 0 0=0 '0'=8 " I I ' Let 0 , m , O ' n be perpendicular on N 0 , N 0/ respectively .
O'L= 7 ' approximately O'LP= 60 ' approximately .
\#183 ; .
LK= 14 ' approximately Let N fall between P and K 80=-O11i=-8E cos NOL 80'=O , 'n=8E cos NO , ' A 8( 0+ 0 ' ) = -8 " ( cos NOL- cos NO , ' A ) Thus 8( 0+ 0 ' ) is negative and decreases numerically as we approach P ; for N O L , N 0/ A become more nearly equal .
Thus between Kand P 0+0 ' decreases with E. Betw8en L and K both 0 and 0 ' decrease , therefore 0 + 0 ' decreases et fortio1\#183 ; i ; but 8( 0+ 0 ' ) is always numerically less than 28E .
Let us see how these changes affect / Li\#183 ; We have [ Section IV .
( l)J -2 =-1 +1-- 8-1 - -1 ) cos(0+0 ' ) .
\#181 ; 12 \#181 ; i \#181 ; ?
c2 \#181 ; a2 If / Le is decreased , - ; +-\ is increased , ~-~ is increased , 0+0 ' is decreased .
\#181 ; c \#181 ; , t \#181 ; c \#181 ; a Therefore cos ( 0+0 ' ) is increased so long as 0+0 ' is less than 90 ' .
:MR .
R. T. GLAZEBROOK ON PLANE WAVES JN A BIAXAL CRYSTAL .
361 Therefore c:,2- \#181 ; :2 ) cos ( 0+0 ' ) is increased so long as 0+0 ' is less than 90 ' .
And the increase in ( ; :z-~2 ) cos ( 0+0 ' ) is greatest when 0+0 ' is least .
Also the increase in ( -:\- 1 2 ) cos ( 0 + 0 ' ) is less than the increase in ( \ + -1 0 ) .
\#181 ; c \#181 ; a \#181 ; c~ \#181 ; " ~ Therefore ---\ + ~ - ( _ !
o - -1 o ) cos ( 0+ 0 ' ) is increased and this increase is greatest \#181 ; c~ \#181 ; a~ \Jl , c~ \#181 ; " ~ ' when 0+0 ' is greatest .
Therefore , as 0+0 ' increases up to 90 ' \#181 ; ,1 decreases .
Wl1en 0+0 ' is greater than 90 ' , cos ( 0+0 ' ) is negative .
As 0+ 0 ' increases , cos ( 0+ 0 ' ) increases numerically .
Therefore ( \-22 ) cos ( 0+0 ' ) increases numerically and is negative .
\#181 ; c \#181 ; a ~ Therefore 1 2 + 2-2 - ( \- \ ) cos ( 0 + 0 ' ) increases a fortiori .
\#181 ; c \#181 ; a , \#181 ; c \#181 ; a Therefore \#181 ; ,1 decreases .
Thus the effect of decreasing / Le is to produce a decrease in the value of \#181 ; ,1 , and this decrease continually increases as we go from L to Q. Now if we compare the differences between the theoretical and experimental values of \#181 ; ,1 , given in Tables V. and VI .
, we see that they increase with 0+ 0 ' , the theoretical value being in excess of the experimental .
The effect of the assumed alteration in the value of / Le , vould be to decrease the theoretical value of \#181 ; ,1 by an a1~ount which constantly increases , that is , to bring it more nearly into agreement with experiment .
But the effect in Tables VII .
and VIII .
would be contrary .
On referring to them we see that the differences between theory and experiment continually decrease .
The effect of the proposed alteration would be to subtract from the theoretical value of \#181 ; ,1 given in those tables : a quantity which increases rnpidly as we get further from L. The result would be that the theoretical value would soon become less than the experimental , and the difference between the two would continually increase .
We mjght , it is true , choose the decrement of / Le so as to make the extremities of the arcs considered coincide , and then the differences would be diminished throughout .
The effect of this would be to make the theoretical section lie outside of the expeiimental for about 75 ' , then cut it and lie inside it .
Whether the two could again be made to coincide when cutting the plane A O B experiment alone could settle .
Let us now consider the effect of the supposed change on / L:1 , \#183 ; We have ~I DCCCLXX IX .
3 A 362 l\IR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
2 _ _l_+~ -(J__ _ __ !
_ ) cos ( 0-0 ' ) o- 2 { ) 0 ~ .
fl , ~ " fl , a fl , c " fl , c " fl , a " We have seen that 0 is decreaFJed by the change .
0 ' is at first decr'eased , though by less than 0 , afterwards it is increased .
Therefore , for any point on the curve L P the value of 0-0 ' is decreased by this alteration .
Therefore cos ( 0-0 ' ) is increased .
( \- \ ) cos ( 0-0 ' ) is increased , though by less than(- ; + \ ) .
~ ~ ~ ~ Therefore __ !
_ + __ !
_ - ( __ !
_ - __ !
_ ) cos ( 0-0 ' ) is increased fl , c 2 \#181 ; , u_2 \#181 ; , / ' fl , tt2 .
Therefore p.,2 is decreased .
Or the variation in p. , z is in the same direction aFJ in p.,1 .
But since 0-0 ' is \lt ; 0+0 ' ( \-~)cos ( 0-0 ' ) is \gt ; ( _ !
_2 - \ ) ' cos ( 0+0 ' ) .
fl , c / La / Le / La And the whole change in -+-o - -o --0 COS ( 0-0 ) 1 1 ( 1 1 ) , \#181 ; , l fl , a " fl , cN fl , a " is much less than the whole change in Or the decrease in p.,2 is sn1all compared with that in p.,1 .
Again , as the variation in 0-0 ' varies from zero at L to a maximum about K , while cos ( 0-0 ' ) also varies considerably , the variation in p.,2 will be considerably different for different points between L and K. After passing K , however , 0-0 ' is nearly constant all along the curve .
The change in the value of p.,2 , therefore , will not differ much for different points , that is , for different waves .
The whole effect of the alteration , therefore , will be to decrease throughout the theoretical value of p. , z , this decrease being greatest about Table V. , line 23 .
In the greater part of Table V. the result will be to produce a closer agreement between theory and experiment .
But in the rest of the work , since the theoretical value of p.,2 is less tha , n the experimental , the effect will be to increase the difference between the two and so widen the discrepancy between theory and experiment .
Thus ta.king both sheetF ; we cannot produce on the whole between theory and experiment by decreasing the value of / 1-c\#183 ; increasing it .
a closer agreement Neither can we by For though we might produce closer agreement for the outer sheet by thus increasing I\IR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAX .
A.L CRYSTAL .
3G3 the theoretical value of JJ,2 , we should add greatly to the error in the case of the inner sheet .
Section VIII.-EJfect of Change in Position ef Planes of Prisms with reference to the Principal Planes ef the Crystal .
We have now to discuss the effect of small variations in the position of the planes PR and P Q. We take P R first , and we consider its position as determined by C L and the angle A L P , while the position of the normals to P and R is fixed by P L ( vide Section II .
) .
Fig. 13 .
Now variations in C L and A L P produce variations in O L , O ' L , X. X. ' , \amp ; c. , and hence in 0 0 ' , while variations in L P only change 0 and 0 ' , 0 L , O ' L , X. and X. ' remaining constant .
Let us consider , therefore , the variations of L P first .
A reference to our tables shows us that through the greater part of the arcs L P , P Q ' , 0-0 ' is constant , so that 80 considered as dependent on L P is equal to 80 ' nearly .
Again , and since the change in ( 0-0 ' ) due to a small increment of L P is very small , such an increment will not alter the value of f-Lz\#183 ; But if 0-0 ' is nearly constant for a small motion along L P , 0+0 ' will be nearly so for a small motion perpendicular to L P. Thus a small motion.perpendicular to P L will not alter appreciably the vnlue of JJ , 1 ; :3 A 2 3G,1 l\1R .
R. 'l ' .
GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
we ma , y therefore obtain corrections to \#181 ; ,1 \#181 ; ,2 independently , to \#181 ; ,1 by supposing an increase in the value of L P , to \#181 ; ,2 by supposing the planes PR , P Q to turn through a small angle about some line in them .
Moreover , we shall consider that line to be in each case the line of intersection with the plane AOC , and treat this line as the same for both planes , the two lines thus considered as coincident are really inclined at an angle of 3 ' .
To find then the variation which we must assume in L P to bring the results of theory and experiment into closer agreement for the inner sheet , we proceed as followsIf in the equation 22=(~+ 12)-(~- 12 ) cos ( 0+0 ' ) J , l,1 J , l,1 , J , l , c J , l , c J , l , a we substitute an experimental value for \#181 ; ,1 , we can determine a value of ( 0+0 ' ) .
Subtracting this from the calculated value , we get 8( 0+ 0 ' ) .
But for the displacement considered 80=80 ' .
cos A , Hence we have found 80 ; but cos ( LN -X.)= OL cos 0 .
0 Land X. are unaltered .
cos by the displacement supposed ; therefore we get 8(L N ) , and L P=L N +cf\gt ; ' .
~ , is constant , therefore we have found 8 L P. To put this plan into execution , I chose as the experimental value for \#181 ; ,1 the 21st in Table VI .
the differences between theory and experiment having at that point a maximum value and being fairly regular .
The resulting value for ( 0+ 0 ' ) is 70 ' 57 ' 20 " The calculated value is 70 ' 16 ' 40 ' ' The difference is oo 40 ' 40 " .\#183 ; .
80=0 ' 20 ' 20 " This gives for 8(L P ) 8(L P)=0 ' 21 ' 20 " Thus at this point the result of theory and experiment could be brought into coincidence by supposing the line OP normal to the face P to ttll'n through an angle of 0 ' 21 ' 20 " in the plane L P away from L. Before discussiug the effect of this change on the values of \#181 ; ,1 in other directions , let us consider the variation it will be necessary to make in the angle AL N to bring the theoretical results for the outer sheet into accordance with experiment .
Taking :MR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
365 the experimental value for \#181 ; ,2 corresponding to the incident wave which gave rise to the value of \#181 ; ,1 we have just been discussing , we have \#181 ; ,2= l \#183 ; 68447 Whence by means of the formula -o= - 0+- 0 - - - - COS ( 0-0 ) 2 ( 1 1 ) ( 1 1 ) , \#181 ; 2~ \#181 ; a~ \#181 ; c~ \#181 ; ?
\#181 ; i we find 0-0'=9 ' 15 ' The value of 0-0 ' given by theory is Table VI .
, line 21 , 0-0'=9 ' 30 ' We have to find the change in the value of X or AL N which will produce this .
In this variation 0+ 0 ' is constant .\#183 ; .
80=-80 ' but 8(0-0')=-15 ' .
\#183 ; .
80= - 7 ' 30 " 80 ' = 7 ' 30 " From these results we can find the increment in X from the equations cos OL cos 0= --cos ( LN -A ) cos A , tan A= tan OL cos X cos 0= cos OL cos LN + sin OL sin LN cos x .
8 _ sin 080 \#183 ; \#183 ; X- sin OL sin LN sin X Substituting values we are led to a value of 8 x , nearly equal to 1 ' .
So that the new value of x= 60 ' 20 ' [ Section II .
( l 7 ) ] .
This change in the value of X produces a corresponding change in the values of A , A1 .
The new values are for the prism P R A= 5 ' 12 ' 20 ' ' ( I ) ( 2 ) 366 MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BfAXAL CRYSTAL .
while for the prism P Q cos OL - Joo\#183 ; = 1 \#183 ; 9945594 b cos A. \gt ; -. .
= 5 ' 10 ' 20 " I \gt ; -. .
=3 ' 54 ' .
I l cos OL ' 1 \#183 ; 9946764 og cos ; \.1 cos O'L ' log-- 1 \#183 ; 9969629 cos A. , ' ( 3 ) ( 4 ) ( 5 ) ( 6 ) ( 7 ) ( 8 ) The ensuing Table gives the values of p , 1 , p,2 for nine different incident waves taken from the previous work , calculated on the supposition that both these changes have been made in the position of the plane P R. The column headed Difference gives the excess of theory over experiment ; that headed a gives the excess of the theory before it was modified ; / 3 gives the change in the theoretical value of p,1 due to this displacement .
Outer sheet .
Where taken 8-81 .
LN +X .
from Experiment Table .
0 I II 0 I II Table V. , line 8 4 25 20 7 3 30 " " " 18 8 20 50 14 12 40 " " " 28 9 12 0 24 6 10 " VI .
" 11 9 15 20 22 4 20 " " " 21 9 14 30 39 14 0 " VII .
" 11 9 12 30 46 31 0 " " " 21 9 9 50 56 16 30 I " VIII .
" 11 9 5 40 71 3 0 " " " 26 9 4 20 80 39 20 i Difference .
, ,~ ~cw Theory .
\#181 ; 2 Experiment \#181 ; 2 Former Value .
Theory .
+ 7 1\#183 ; 68533 1\#183 ; 68526 1\#183 ; 68-537 +11 I 1\#183 ; 68465 1\#183 ; 68454 1\#183 ; 68462 - 3 1\#183 ; 68445 1\#183 ; 68448 1\#183 ; 68437 - 9 1\#183 ; 68443 1\#183 ; 68452 1\#183 ; 68437 - 4 1\#183 ; 684i3 1\#183 ; 68447 1\#183 ; 68437 - 9 1\#183 ; 68444 1\#183 ; 68453 1\#183 ; 68438 -12 1\#183 ; 68445 1\#183 ; 68457 1\#183 ; 68439 I - 5 I 1\#183 ; 68447 1\#183 ; 68452 1\#183 ; 68440 -4 I 1\#183 ; 68448 1\#183 ; 68444 1\#183 ; 68442 TABLE IX .
Inner sheet .
LN- X ' .
8+8 ' .
0 I II 0 I II 47 12 20 90 ' +14 18 50 \#181 ; 1 Former 111 Experiment Value .
I N ; h Theory .
ew eory .
1\#183 ; 68106 1\#183 ; 68099 1\#183 ; 68012 1\#183 ; 67743 1\#183 ; 67721 1 \#183 ; 67707 1\#183 ; 6635::l 1\#183 ; 66300 1\#183 ; 66285 1\#183 ; 64674 1\#183 ; 64603 1\#183 ; 64590 1\#183 ; 62897 1\#183 ; 62807 1\#183 ; 62806 1\#183 ; 60974 1\#183 ; 60897 1\#183 ; 60882 1\#183 ; 58432 1\#183 ; 58365 1\#183 ; 58346 1\#183 ; 55216 1\#183 ; 55157 1\#183 ; 55138 1\#183 ; 53793 1\#183 ; 53774 1\#183 ; 53779 Difference .
+ 3 1 - 14 2 - 15 3 - 13 4 - 1 5 - 15 6 -19 7 -19 8 + 5 9 ~ ~ ~ ~ P\gt ; N t::J to ~ 0 0 Pi 0 z 'tl t , ~ t::J ~ ~ t::J a. I { 3 .
I W. I H z + 7 4 +22 37 P\gt ; +52 67 t:d +71 86 ~ +90 91 ~ +77 92 I ~ +67 86 +59 78 I @ +29 24 I r\lt ; j [ / ) .
~ I !
; !
) C , --.J It will be noticed that the differences for this modified theory are much less than for the original , but they have nearly all changed sign .
In fact , the correction applied , calculated so as to produce the closest possible agreement in line 5 , has been too great , and the differences can be reduced by applying a similar but smaller correction in the opposite direction .
Now this variation in the value of \#181 ; ,1 has been caused almost entirely by the increase of the angle L P ; we must therefore decrease this angle again , and as the numbers in the difference column are each about one-fifth of the corresponding number in the column / 3 , we must decrease L P by one-fifth of the amount by which it was previously increased .
But this incren , se was o(LP)=0 ' 21 ' 20 " we must therefore decrease this by 0 ' 4 ' 15 ' leaving the total increase 0 ' 17 ' 5 ' In consequence of this , we must add to the modified theoretical values of \#181 ; ,1 respectively , quantities equal to one-fifth of the corresponding numbers in column / 3 .
The result gives us a new set of theoretical values , which agree with experiment throughout to a remarkable degree .
These results are expressed in Table X. Column 1 gives the values of \#181 ; ,1 already modified .
Column 2 gives one-fifth of quantities in / 3 , which , when added to the values of \#181 ; ,1 in Column 1 respectively , give Column 3 the resulting theoretical values .
Column 4 gives the experimental values , and Column 5 the excess of theory over experiment .
TABLE X. 'Modified value I \#190 ; ~\#183 ; Final vttlue of \#181 ; 1 .
I Experimental I Difference .
Of \#181 ; I. value of \#181 ; 1 .
1\#183 ; 68102 1 1\#183 ; 681U3 1\#183 ; 68099 4 1 1\#183 ; 67707 7 1\#183 ; 67714 1\#183 ; 67721 -7 2 1\#183 ; 66285 13 1\#183 ; 66298 1\#183 ; 66300 -2 3 1\#183 ; 64590 17 1\#183 ; 64G07 1\#183 ; 64603 4 4 H\gt ; 2806 18 1\#183 ; 62824 1\#183 ; 62807 17 5 1\#183 ; 60882 18 1\#183 ; 60900 1\#183 ; 60897 a 6 l\#183 ; ,58 ; 146 17 1\#183 ; 58363 1\#183 ; 58:365 -2 7 1\#183 ; 551:38 16 1\#183 ; 55154 1\#183 ; 58157 -3 8 1\#183 ; 53779 5 1\#183 ; 53784 l"/ 53774 10 9 \#183 ; ----- ~------\#183 ; These differences are exceedingly small-sometimes positive , sometimes negativebut rarely greater tha , n the possible error of experiment as appears from the tables of error in I. , II .
, III .
, and IV .
:M:R .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
369 Thus we have found a plane section of the surface which agrees closely with the results of experiment .
Thjs section passes through L , a point on the principal section A C at an angular distance of 1 ' 21 ' 4211 from C , is inclined to that section at an angle of 60 ' 20 ' , while the normal to the face Plies at an angular distance from L equal to 35 ' 0 ' 1911+0 ' 17 ' 15"=35 ' 17 ' 3411 Under these circumstances it will be necessary to consider the probability of making such an error as this result indicates in the position of the normal to P with reference to the principal axes .
But before doing this we must say a few words about the variations in the values of p.,2 produced by the correction now considered .
The final correction will not affect any of them except the first , whjch will be slightly increased by it .
This follows from the fact that between lines 1 and 2 of Table IX .
0-0 ' varies considerably , and increases or decreases with 0+ 0 ' , so that p.,1 , p.,2 increase or decrease together .
The change considered has increased p.,1 , it will therefore increase p.,2 .
The difference in line 2 is considerable , but it may be noticed that the experimental value is clearly too small , falling as it does ( Table I. , line 18 ) between two considerably greater values .
In the other cases the differences between theory and experiment ltave been diminished but remain of the same sign as before .
A still further displacement in the same direction would therefore produce still further agreement .
We have now to consider the effect of varying CL .
Let us trace the effect of increasing C L by 1 0 ' .
The value of JJ , a depends on that of C L being found from it by the formula ( Section V. ) where and r is given from experiment r=l\#183 ; 68550 The new value of a will be found to be instead of p. , " = 1\#183 ; G8571 l \#183 ; G85G0 This increase in the value of a will increase the i=mgle between the optic axes from 9 ' 4 ' 5 ' ' , the value already fonnd , to 9 ' 9 ' 2011 .
We have , therefore , to discuss the effect of these changes in the theoretical values of \#181 ; 1 , \#181 ; z , We will further suppose that the angle x is so altered that the section still passes through the former position of P-i.e .
, so that the only error made in the prev10us determination of its position has been in the position of R ' .
This is probable , for O R ' is small and A R ' , BR ' , nearly right angles .
The new value of x will be x=59 ' 32 ' These changes will produce variations in the values of O L , O'L , A 'A .
' , for P R ' , and O L ' , O ' L ' , 'A , \ ' , for P Q ( we assume that L L ' is unaltered ) .
The complexity of the changes rendered it very difficult to discuss the effects on \#181 ; 1 , \#181 ; 2 from general considerations .
I have therefore calculated the new values in five different positions taken from the previous tables .
The table following gives the general results .
TABLE XI .
Whence taken .
Difference .
/ 11 / 11 / 11 Modified Theory .
Experiment .
Former Theory .
Difference .
'rable v. , line 28 36 1\#183 ; 66386 1\#183 ; 66300 1\#183 ; 66352 52 1 " VI .
, " 11 49 1\#183 ; 64652 1\#183 ; 64603 1\#183 ; 64674 71 2 " VI .
, " 21 65 1\#183 ; 62872 1\#183 ; 62807 1\#183 ; 62897 90 3 , , VII .
, , , 21 41 1\#183 ; 58406 1\#183 ; 58365 1\#183 ; 58432 67 4 .
, VITI .
, " 11 43 1\#183 ; 55200 1\#183 ; 55157 1\#183 ; 55216 59 5 Whence taken .
Difference .
/ 12 / 12 / 12 Difference .
Modified Theory .
Experimeet .
Former Theory .
Table V. , line 28 -1 1\#183 ; 68447 1\#183 ; 68448 1\#183 ; 68437 -11 1 " VI .
, " 11 -6 1\#183 ; 68446 1\#183 ; 68452 1\#183 ; 68437 -15 2 " VI .
, " 21 +l 1\#183 ; 68448 1\#183 ; 68447 1\#183 ; 68437 -10 3 , , VII .
, " 21 -7 1\#183 ; 68450 1\#183 ; 68457 1\#183 ; 68439 -18 4 " VIII .
, " 11 0 1\#183 ; 68452 1\#183 ; 68452 1\#183 ; 68440 -12 5 The difference table on the left is between the modified theory and experiment , that on the right between the former theory and experiment .
The upper table refers to the inner , the lower to the outer sheet .
It will be seen that the differences for \#181 ; 1 , \#181 ; 2 are both decreased , though still remaining of the same sign .
The differences for \#181 ; 1 are about two-thirds of their former value .
If we then increase O L still further by twice as much as previously , so that the total increase is 30 ' , we shall get still closer agreement for \#181 ; 1 .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
371 TABLE XII .
-- a. ( 3 .
/ J,1 / J- " Final Theory .
Experiment .
Difference .
1G 48 1\#183 ; 66304 1\#183 ; 66300 4 22 66 1\#183 ; 64608 1\#183 ; 64603 5 25 75 1\#183 ; 62822 1\#183 ; 62807 15 26 78 1\#183 ; 58354 1\#183 ; 58365 -11 16 48 1\#183 ; 55268 1\#183 ; 55157 11 y. o. Final / TJ-2h eory .
Exper/ Ji,2m ent .
Difference .
10 30 1\#183 ; 68467 1\#183 ; 68448 19 9 27 1\#183 ; 68464 1\#183 ; 68452 12 11 33 1\#183 ; 68470 1\#183 ; 68447 23 11 33 1\#183 ; 68472 1\#183 ; 68457 15 12 36 1\#183 ; 68475 1\#183 ; 68452 23 Column a gives the change in / Li due to a change of 10 ' in CL .
Column / 3 gives three times that change which , assuming the theory of proportional parts , will be the change in / Li due to a change of 30 ' in C L. Subtracting this from the values of / Li given by the original theory , we get the final theoretical values of / Li\#183 ; These agree much more closely with experiment than the original values .
But the effect of this further change will be to over correct / Lz\#183 ; The first change has reduced the differences for / Lz by the amounts given by y , Table XII .
Multiplying these by 3 we get 8 , the changes produced in / Lz by a change of 30 ' in C L , and adding these values to the original theoretical values of / L : !
- we get the final theory , which differs from experiment by rather more than the original theory and in the other direction .
Thus an increase of C L will not , on the whole , produce the required effect .
Similarly a decrease will not produce it either .
Section IX.-Possibility of an Error in the Positions of the Faces discussed .
We have seen , however , that by increasing L P by 1 7 ' 5 ' ' and A L P by 1 ' , we can make the agreement between theory and experiment extremely close .
This alteration in the values of L P and ALP may be effected in two ways .
We may suppose either that , the axes of elasticity retaining their position with reference to the planes rn , rn ' , we have made an error in the observed angles PM , p M ' ( Section II .
) , and hence in the position of P referred to the axes .
Or , secondly , that the position of P relatively to rn m ' has been accurately found , but that the axes of elasticity differ slightly in direction from their assumed positions , so that O C is not parallel to the intersection of mm ' , and O B does not bisect the angle MOM ' .
We will take the first hypothesis first , and consider what must be the error in p M , P M ' to give rise to the required change in the position of P. Fig. 14 .
Let P ' be the new position of P , we require M P ' M ' P ' .
Let P K be perpendicular on L P ' LK= LP approximately KP ' = 0 ' 17 ' approximately RP=PLK sin 35 ' approximately =60'x\#183 ; 57 =34 ' approximately Now P K P ' is approximately a small plane triangle with a right angle at K. whence Again whence PP'= yiPK2+P'K2 =J342+172 17 ' yi5 PP'=38 ' cos CP ' = cos CL cos LP ' - sin CL sin LP ' cos ALP ' CP'=35 ' 58 ' CP =35 ' 43 ' [ Section II .
( 12 ) ] PP'= 0 ' 38 ' 1\IR .
R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
373 Hence from the triangle PC P ' we get POP'=0 ' 58 But from the triangle A C P but ACP = 57 ' 42 ' 30 " [ Section II .
] .
\#183 ; .
AOP ' = 58 ' 40 ' 30 " AOM=58 ' 7 ' 30 " [ Section II .
( 7 ) ] .
\#183 ; .
MOP= 0 ' 33 ' cos MP ' = cos MOP sin OP ' whence MP'=54 ' 58 ' 50 " The original value of M P ' was as given by the mean of a large number of observations MP ' = 54 ' 17 ' 6 " [ Section II .
( 3 ) ] .
The difference , amounting as it does to over 40 ' , is far beyond any possible error of experiment .
Let us find further the change in M ' P , whence The observed value was cos M'P ' = sin OP ' cos M'OP ' M'CP'=63 ' 47 ' -33 ' =63 ' 14 ' M'P'=74 ' 39 ' 40 " M'P=75 ' 16 ' 20 " [ Section III .
( 4 ) ] .
The difference is 0 ' 36 ' 40 ' ' .
This , again , is far beyond a.ny possible experimental error .
Thus , so long as we assume the position of the axes of elasticity to be definitely fixed in the crystal so that O O is parallel to the intersection of m m1 while O B bisects the angle between them , the displacement of the plane P R ' necessary to brjng the results of experiment and theory into agreement is far too great to be possible .
But there remains the supposition that the axes of elasticity have not exactly the same position in all crystals of arragonite .
So that the displacement of P R relatively to the axes might be effected by changing slightly their position with reference to the faces of the crystal , P R retaining their position , relatively to those faces , unchanged .
The possibility of this is a question for the mineralogist .
I have been as yet unable to find data for ~ satisfactory answer .
It seems , however , plausible to suppose that in a substance like arragonite which is not chemically pure , but contains foreign substances to a variable degree in different specimens , some slight variation such as that indicated might occur .
We can , without much difficulty , calculate the amount of the change .
Fig. 15 .
We may very approximately treat the rotation round O L as if it were about O C , C L being very small .
The effect of a rotation of 1 ' about O C will be to bring A and B to A , B , respectively in the same plane , where AA =BB =1 ' I I A , B , being to the right of A , and B. The increment of 17 ' in L P may be effected by turning the axes through an angle of 17 ' about a normal to the plane L P passing through 0 , let O S T be this normal , let A ' B ' C ' be the final positions of A BC .
A , A ' , B , B ' are arcs of small circles whose centres lie on OST at T and S respectively , so that A TA'=B SB'=l7 ' I , CC ' is an arc of a great circle , and is = 17 ' .
B , B'=SB , x B'SB , SOB , =60 ' approximately SB , .
0 , / :1 .\#183 ; .
= sm60 = 2 .
B B'=17\/ 3 . .
, 2 BB ' = v'(BB,2+B ' 8,2 ) =v(6o'2+ix ; 7~ ) = 1 ' 2 ' approximately A , A'=ATA ' sin 30 =8 ' 30 " .
/ ( 17'2 AA'2= V G0'2+ 4 ) approxi .
mately = 1 0 1 , approxi .
mate1 y .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAXAL CRYSTAL .
375 Thus the new positions of the axes are inclined to the old at angles of 1 ' l ' , 1 ' 2 ' , and 17 ' respectively for O A , 0 B , 0 C. So that the axis of C in its new position would be inclined to the line of junction of m ni ' at an angle of 17 ' , while that of b would be inclined to the bisector of these planes at an angle of 1 ' 2 ' .
If variations such as here considered be possible , so that we cannot be certain a priori of the position of the axes of elasticity , the only method of testing FRESNEL 's or any other theory will be by trial .
Section X.-General Results of the Investigation .
Combining the results of the two series of experiments , it seems to me most probable that FRESNEL 's theory is only true as a first approximation .
Both series of observations have led to the discovery of considerable deviation from the theory , unless we assume errors in the experiments , especially in the determination of the position of the planes of the prisms with reference to the axes of the crystals , which are greatly in excess of the amounts we can reasonably expect .
It will be noticed that the values taken for the constants fLa and / 1-b in the two parts of the paper differ by \#183 ; 0002 and \#183 ; 000l 7 respectively .
This is due to the fact that the two crystals used were different .
RuDBERG ( PoGG .
Annalen , vol. xvii .
, p. 1 ) found differences of as much as \#183 ; 0004 in the values of / 1-b deduced from two specimens of arragoni te .
It may be objected that the variations between theory and experiment are not in exactly the same direction in the two Parts ; but we must remember ( 1 ) that the arcs investigated are taken from entirely different portions of the surface ; ( 2 ) in the first Part , the approximately elliptic section belongs to the outer sheet ; in the second , to the inner sheet of the wave surface .
Fig. 16 .
The accompanying figure will illustrate the results of the investigation .
M O N is the principal plane of the first prism nearly coincident with the plane through the optic axes A O C. N O L the principal plane of the second prism nearly coincident with the plane BO A , so that O N is nearly coincident with O A. M O L the principal planes of the two prisms in the second crystal here treated as coincident , inclined at about 60 ' to B O C , and cutting it in a line nearly coincident with O C. Hence OM is not far removed from O C. 0 Lis nearly in the plane AO Band inclined at about 60 ' to O A or O N. The strong lines give approximately the form of the sections of FRESNEL 's surface by these planes , the dotted lines the results of experiment .
In the case of the arc N ' L ' the results of theory and experiment agreed closely .
For the arcs M N , M ' N ' the experiments covered an arc of about 16 ' from M and M ' .
For N L the experiments covered an arc extending from L to about 10 ' on the side nearei ; .t to N of the point where the two arcs intersect .
For ML the experiments extended over an arc of about 70 ' measured from M. Section XI.--Ejfect of Dispersion considered .
The theory of dispersion appears to me to afford a more probable explanation of these small variations from FRESNEL 's construction .
FRESNEL himself remarked ( ' Second Supplement au premier Memoir sir la double refraction , ' CEuvres Completes de FRESNEL , tom .
ii .
) that in the case of the vibrations which constitute light the radius of the sphere of action of the molecular forces brought into play by the vibration is not necessarily very small compared with the wave length .
And , consequently , it is incorrect to suppose that the propagation of each of the disturbances of which a vibration is composed is uninfluenced by the distm\#183 ; bances which precede and follow it , and that the velocity of propagation is independent of the manner in which they proceed and follow it .
This supposition is the basis of FRESNEL 's work on double refraction .
Let us consider the effect of dispersion in a doubly refracting medium .
In an isotropic medium the relation between V , the velocity of wave propagation , and A. , the wave length , is generally allowed to be l b C =y=a+,._2+)..i+ , \amp ; c. , a , b , c , \amp ; c. , being constant , the values of the terms continually decreasing , so that except in highly dispersive media we may put t=a+:2 with sufficient exactness .
MR. R. T. GLAZEBROOK ON PLANE WAVES IN A BIAX .
A.L CRYSTAL .
377 Let us suppose that an equation of this form holds in crystalline media also , only that a , b , c , \amp ; c. , instead of being constants are functions of the directions of propagation and vibration ; and , further , let us suppose that FRESNEL 's construction is true for waves of infinite length , so that the equation . !
.=a gives us a FRESNEL 's wave surface .
V From the known values of the constants of the wave surface for diflerent rays of the spectrum the constants of the surface for infinite wave lengths can be found , and hence the value of a calculated in any given direction .
If from experiment we find the value of . !
_ or p , for any wave length , the difference V p , -a ought on this theory to be equal to : 2 .
And if we find the values of p , for different wave lengths ( p,1 , p,2 , p,3 say ) in the same direction , we have whence b 11-a-a= ) , , .
, 2 To verify this I observed the values of p , in two directions for the rays C , D , and F , with the following results :- ( ::r = 1 \#183 ; 2403 \#181 ; u-a=l\#183 ; 2875 ( first experiment ) \#181 ; c-a l \#183 ; 2770 ( second experiment ) ( ~:r = 1 \#183 ; 46978 \#181 ; F-a=l\#183 ; 47208 ( first experiment ) \#181 ; u-a =1\#183 ; 47348 ( second experiment ) .
The numbers , especially in the last case , are sufficiently close to make it seem worth while continuing the investigations .

