176 Mr. T. K. Chinmayanandam. 5. Saturated vapours, even when very dense, show no increase of scattering power beyond what the density would lead one to expect. If molecular aggregates are formed, they are not numerous enough to show by this method. 6. Liquid ether apparently scatters about seven times less light than a corresponding mass of ether vapour. DESCRIPTION OF PLATE 2. 1. Photographs of the two polarisations in ether vapour. Upper image lias vibration vertical, lower horizontal. Screens over the upper image of opacity increasing from left to right. Equality at No. 5. 2. Similar photographs for nitrous oxide, using same opaque screens. Equality falls between Nos. 1 and 2. 3. Ether vapour. Nos. 1 and 5 at low pressure ; 2, 3, and 4 at saturation pressure, but with successively reduced lens aperture (see p. 172). Note that No. 3 approximately equals Nos. 1 and 5. On Haidinger's Rings in Mica. By T. K. Chinmayanandam, M.A. (Hons.). (Communicated by Dr. Gilbert T. Walker, C.S.I., D.Sc, F.R.S. Received July 10, 1918.) 1. Introduction. The interference rings observed between plane parallel surfaces under diffuse monochromatic illumination have recently acquired much importance, in view of their practical application in the construction of spectroscopes of high resolving power. These rings were first observed by Haidinger in mica. This circumstance and the fact that mica is readily available, and by its natural cleavage furnishes plates with absolutely parallel faces, make the study of the rings in that substance one of considerable interest. In a paper published in the ' Philosophical Magazine ? for November, 1906, Lord Rayleigh described some interesting features of the rings observed in natural light. Mica being a doubly refracting substance, there are in fact two systems of rings super- posed upon each other, being each due to one of the beams polarised in planes at right angles to each other. It may be expected, therefore, that there must be regions of maximum and minimum visibility in the field of view. This is exactly what happens, and Lord Rayleigh has noted that in a plate of mica On Haidingers Rings in Mica. 177 0*213 mm. thick, observed in the light of a salted Bunsen flame, the rings were indistinct along four directions apparently at right angles, radiating from the centre of the field. He has remarked in conclusion that it would be of interest to undertake a more detailed discussion of the subject, especially in comparison with observations on a plate whose optical constants are known. It is proposed in this paper to give an account of some further observations made by the writer on the rings in mica, and also to attempt a theoretical discussion of the phenomenon. Mica, as is generally known, occurs in many different varieties with different optical constants. In muscovite, which is the most common variety, the apparent angle between the optic axes is about 70°. Most of the observa- tions of the present writer were made with this variety of mica, of which sheets of different thicknesses were available. In some other varieties, such as biotite and phlogopite, the axial angle is much smaller, less than 10°, so that these varieties are nearly uniaxial. Of these, biotite, being strongly absorbent to light, is not very suitable for observation of Haidinger's rings ; but by the courtesy of the Director of the Geological Survey of India, the writer was furnished with a few good specimens of phlogopite from Burma, with which some observations, to be mentioned later in the paper, were also made. The phenomena observed differ, as may be expected, with different varieties of mica. As regards the source of light, Lord Eayleigh used an ordinary sodium flame. It was found by the writer that a much more convenient source is a Cooper-Hewitt lamp with some suitable ray filter ; the light being powerful, the observations are very easily made. 2. Experimented Methods and Results. The methods of observing and photographing the phenomena in transmitted light as well as in reflected light may now be described. All the observations were made in natural light. Transmitted System. — The rings formed by the transmitted light are very easily observed. The mica plate is simply held close to the eye, which is then directed towards the Cooper-Hewitt lamp, a ground glass plate and a green ray filter being interposed between the mica plate and the lamp, to diffuse the light and to render it monochromatic. If a photograph is to be taken, the camera lens is brought close to the mica, the best lens to use being a short-focus one, with as large an aperture as possible, consistent with good definition. This arrangement, though suitable for visual observa- tion, does not give a good photograph, on account of the feebleness of contrast between the bright and the dark rings. The contrast is very much greater in the rings formed by reflection from the surface of the mica, and it Q A 178 Mr. T. K. Chinmayanandam. was therefore desirable to devise some simple means of observing and photographing the reflected systemof rings. Reflected System of Rings. — The chief difficulty in the observation of this system is that the head of the observer screens the light off from the mica plate, so that the centre of the system never comes into the field of view. The usual way of getting over this difficulty is to use a plate of glass at an inclination of 45° to the surface of the mica, and to have the source of light on one side. This arrangement is, however, not quite suitable for the present purpose, for the eye, or the camera lens in case a photograph is to be taken, cannot be brought quite close to the mica plate, and has to be kept at a distance. The result is not only that the field of view becomes corre- spondingly limited, but also that, a larger area of the crystal being used, the smallest defect or non-uniformity, or want of flatness in the mica, results in the rings becoming distorted or irregular. There is also a loss of light in reflection at the glass plate. To get over these difficulties, the following simple arrangement has been used in the present investigation, to observe and photograph the reflected rings. A white sheet of cardboard is taken, and a small hole bored in it in the centre. This is mounted vertically on a stand, and the mica plate is fixed immediately in front of it, with its face parallel to the cardboard. The side of the cardboard facing the mica is lighted by one or more Cooper-Hewitt lamps, and the rings are viewed from the other side through the hole in the cardboard against a dark background. The reflected rings can be seen in this way very easily and clearly. To prevent extraneous scattered light from entering the eye or the camera lens, the edges of the hole and the back of the cardboard may be blackened. For visual observation, a green ray filter may be fixed behind the cardboard, though the effects described in this paper are, indeed, seen very well without it. For photographing the rings the green light is not very suitable, but a cell containing a solution of sulphate of quinine may be used with great advantage. The function of this cell is to cut of! the ultra-violet rays and the two lines at 4047 and 4078 A.TJ. without in any o way affecting the bright violet line at 4358 A. IT., and thus to secure the homogeneity of the violet light. The green and the yellow rays are, indeed, transmitted freely by the cell, but they do not affect the photographic plate appreciably. The cell may be fixed behind the camera lens. Process plates, with an exposure of 10-15 minutes, give very satisfactory pictures, one of which is reproduced in fig. 1. Some of the experimental results obtained with mica may now be described. A number of plates of muscovite mica of different thicknesses were examined. The dimensions of the rings vary, of course, with the On Haidinger's Rings in Mica. 179 thickness of the plate and the wave-length of the light. The general form of the curves of minimum visibility is, as seen in the figure, roughly a system of hyperbolas. As the entire field that could be visually \ observed Fig. l. was not covered by a single negative, a diagram (fig. 2) has been drawn (copied from two or three negatives taken in different positions) to show the curves observed. It will be seen that there are two series of curves with their axes nearly at right angles. In each set, at least two curves, sometimes 180 Mr. T. K. Chinmayanandam. more, could be observed on either side of the centre. The plane containing the optic axes of the crystal was parallel to the direction AB. The curves are fixed relatively to the crystal, and rotate with the latter as it is turned round. The general form of the curves does not vary much with the wave- length of the light or with the thickness of the crystal, but their actual dimensions and position relatively to the centre of the system do depend upon both. As the thickness of the mica plate is varied, one of the pairs of hyperbolas moves towards the centre, and for certain thicknesses meets at that point, forming a cross in the field, as shown by the dotted lines in Fig. 2. fig. 2. The appearance is then somewhat similar to that described by Lord Eayleigh, except that the other hyperbolas remain in the field of view, and a single cross is not observed in any case. A slight asymmetry may be noticed with regard to the curves on the same axes situated on either side of the centre; this is apparently due to the fact that the acute bisectrix of the angle between the optic axes is not exactly perpendicular to the plane of cleavage. The rings do not actually vanish at any point of the field, though the lines of minimum visibility are seen well defined. A scrutiny of On Haidingers Rings in Mica. 181 the photograph reproduced in the figure will show that the rings appear doubled where a line of minimum visibility meets them at a small angle, as is the case near the vertices of the hyperbolas, but where it cuts across them nearly at right angles, we find the bright bands on one side running into and meeting the dark bands on the other side, producing a pretty ( system of dislocated rings. It can easily be verified that these phenomena are due to the superposition of two independent sets of rings, by looking at the ring system through a nicol. As the latter is rotated, the lines of minimum visibility disappear absolutely in four positions of the nicol at right angles to one another. 3. Theory. The fact that in mica the acute bisectrix of the angle between the optic axes is very nearly normal to the plane of cleavage simplifies the theory considerably. The rings observed by the light reflected from the crystal are, of course, due to the interference of the rays reflected from the first surface of the crystal with those that enter the plate and, after undergoing one or more internal reflections at the second surface, emerge from the first surface in a direction parallel to the former. Now since any ray incident on the crystal is generally split up into two rays polarised in and perpendicular respectively to the principal plane, and travelling in slightly different directions with different velocities, there will be two rays, polarised in planes perpendicular to each other, emerging in any given direction from the first surface after one internal reflection. Suppose Si, 82 are the respective path differences in the two cases between the directly reflected and the internally reflected rays. Then the dark rings of the two sets are given respectively by Si = n\, $ 2 = m\, where n, m are integers. A point of minimum visibility will correspond to the condition that 81 = NX, and 8 2 = (M + |) X, where (N—M) is an integer. A line of minimum visibility evidently satisfies the condition that, at every point on it, the respective orders of the rings of the two sets have a constant difference equal to N— (M + |), i.e., the condition that the relative path difference of the two rays polarised in perpendicular planes and emerging in the same direction is constant over all points on the line. The problem of determining the form of such a line, therefore, reduces to finding (Si— S 2 ) corresponding to different directions. For an isotropic medium, the path difference S between the interfering rays is simply 2 pe cos r, where jn is the refractive index and e the thickness of the plate, and r is the angle of refrac- tion. Since, in mica, the normal to the plate is practically the bisectrix of 182 Mr. T. K. Chinmayanandam. the angle between the optic axes, the velocities of each of the two rays into which the incident ray is split up is, by symmetry, the same before and after internal reflection at the rear surface of the plate, and we may to a sufficient degree of approximation write 81 = 2e ~ cos ri, v S 2 = 2e — cos r 2 , v where % is the velocity in air, and v f and v" are the velocities of the refracted waves in the crystal. Thus we have s ^ o fcosri cosr 2 L v v Even without evaluating this expression in terms of the angle of emergence of the rays, it is seen that it is identical with the well-known formula giving the form of the isochromatic lines in convergent polarised light for a plate of thickness 2e* Since mica is biaxial, the optic figure should resemble a series of Cassini's ovals and leminiscates. The angle between the optic axes in muscovite is, however, fairly large, and in the centre of the field the isochromatic lines look very much like hyperbolas. The analogy between the form of the lines of minimum visibility and of the isochromatic lines in convergent polarised light pointed out here is strikingly verified by an examination of the rings in phlogopite.f In this variety of mica the optic axes are not widely, separated, and the lines of minimum visibility in Haidinger ? s rings are found to be closed curves of the form of Cassini's ovals, exactly as indicated by theory. It is interesting to consider the actual form of the rings, and see exactly how the two systems of rings give rise to the observed lines of minimum visibility. For simplicity, the difference between the wave velocities and the corresponding ray velocities may be assumed to be negligible. Let <f> be the azimuth and £0 the angle of incidence of any ray falling on the plate, and <£', K > V* ?"> similar co-ordinates for the corresponding refracted wave- normals. Also let v > v'> v"> be the velocities of the two waves in air and in the crystal respectively. Then 4> = <f>' = 4>" (1) , sin f __ sin £" __ sin £ /r>\ and. — -. .. — . ("") * See Drude's i Optics,' English translation by Millikan, p. 350. t This analogy between the lines of minimum visibility and the isochromatic lines in polarised light is not applicable to all crystals cut in any manner, but only to those which cleave or which are cut perpendicular to one of the axes of optic symmetry in the crystal. On Haidingers Rings in Mica. 183 The wave-velocities in the direction <j>, f, are given by the equation cos 2 <j> sin 2 cj> cot 2 £ _ r. ,ox v 2 — a 2 v 2 — b 2 v* — e 2 the solution of which is 2^ = a 2 + b 2 - sin 2 £ ( Ai cos 2 <£ + A 2 sin 2 <f>)±* (4) where Ai = a 2 — c 2 , A 2 = b 2 -~c 2 , and A 3 = & 2 — & 2 ; and a: 2 = A 3 2 - 2A 8 sin 2 f (A x cos 2 0- A 2 sin 2 <-&) + sin 4 £ (A x cos 2 <£ + A 2 sin 2 <j>) 2 . (5) Cte ^Aere ^6 Angle between the Optic Axes is Large. — In this case, if £ be small, we can neglect the fourth and higher powers of sin £ and (5) reduces to k = A 3 -- sin 2 f(AiCos 2 ^> — A 2 sin 2 $). (6) Prom (4) and (6) we get for v' and v" 2v' 2 = ^ 2 + & 2 + A3-2AiCos 2 <^sin 2 r J or v' 2 = a 2 — Ai cos 2 <£ sin 2 f ; "1 •2 w r* and similarly v" 2 = b 2 —A 2 sin 2 $ sin 2 f But from (2), v' 2 = -7-0-t- ^ and ^*" 2 — ~^V ^> 2 , sm J £0 sm J 5) so that on eliminating v', v" ', we get from (7) sin 2 P = ^ 2 sin2 6) ^o 2 -f- Ai cos 2 <£ sin 2 £0 and sin* f" = ^^"'f . ,,. ^o •+■ A 2 sm J <p sm J £ (7) }>■ (8) If now Si, S 2 , be the path differences between the rays reflected at the first surface and the rays that emerge from the same surface after refraction and one internal reflection, ^0 Si = 2e cos f ' ~ = 2e sin f cot f '. v Similarly S 2 = 2e sin £ cot f ". From (8) cot* £' = V + AiCOS 2 <ftsin 2 g- -aW£ S 2 1 Hence ~ = — {-y 2 — sin 2 g, (a 2 sin 2 <£ -f c 2 cos 2 <£)}, (9) tx.e Lb • and similarly S 2 1 -^ = 75 {% 2 — sin 2 go (& 2 cos 2 <£ + c 2 sin 2 <£)}. (10) 184 Mr. T. K. Chinmayanandam. The forms of the two sets of Haidinger's rings observed in the crystal are thus determined respectively by sin 2 £n — — -(& 2 sin 2 <£ + c 2 cos 2 <£) = const. (11) ci and — -~^ (b 2 cos 2 $ + e 2 sin 2 (j>) = const., (12) or, in Cartesian co-ordinates, by a 2 y 2 +b 2 x 2 = const. and b 2 x 2 -\-c 2 y 2 — const. Both sets of rings are hence, to the first order of approximation, ellipses. The lines of minimum visibility are given by 81 — S2 = const., or - {^o 2 — sin 2 £0 (a 2 sin 2 <£ -h c 2 cos 2 <£)p a ■—• - {^o 2 — sin 2 £ (b 2 cos 2 $ + c 2 sin 2 <£)p = const., or, approximately, sin 2 f ( - (a 2 sin 2 ^ + c 2 cos 2 <£) — - (6 2 cos 2 <£ + c 2 sin 2 <£) = const., which in Cartesian co-ordinates becomes - (a 2 y 2 4- c 2 x 2 ) — - (& 2 # 2 -f (?y 2 ) = const., (13) representing a series of hyperbolas, whose asymptotes make an angle 2 a with each other, where tan 2 a= b/a. Since a and b are nearly equal in mica, the asymptotes are practically at right angles. This is not, however, exactly what is observed in the photo- graph, nor are the curves strictly hyperbolas. The deviation must be ascribed to the fact that the fourth power of sin f in equation (5) is not entirely negligible. Case where the Angle betiveen the Optic Axes is Very Small.- — This is the case with the phlogopite variety of mica, which is practically uniaxial. In this case, A3 becomes so small that in equation (5) the term A3 2 becomes less important than that involving sin 4 f, so that the approximation made in the previous case would no longer be valid. We can, however, write equation (5) in the approximate form K 2 = A3 2 - A 3 {a 2 + b 2 ) sin 2 & ( Ai cos 2 <£ - A 2 sin 2 <£) + ^!±ZZ sin 4 f (Ai cos 2 4> + A 2 sin 2 cf>) 2 , (14) 4: which has been obtained by writing sin 2 £ = (a? + b 2 ) sin 2 £o/2 in equation (5). On Haidingers Rings in Mica. 185 Proceeding exactly as in the previous section, the form of the rings is found to be given by the equations B 2 _ 2 v 2 + (Ai cos 2 6 + A 2 sin 2 </>) sin 2 £ • 2 y &2- a 2 + b 2 ±* -sin ft, that is, sin 2 f ^ 1 — — — — -,o 2 ^ r = const., L a 2 + 6 2 4 : /c J or, since k is small compared with <x 2 +6 2 , If we neglect the squares of A 3 in (14), we have & 2 +& 2 . 2!//A a -l • a • 2 i\ a AiCos 2 <f>— A 2 sm 2 <f> * = — ~— sm 2 fo (Ai cos 2 + A 2 sin 2 <£)-A 3 . -~ 2 T , A • 2 T 2 ^ ^ Ai cos J <£ +- A 2 sm J cf> so, that equation (15) becomes on being transformed to Cartesian co-ordinates, x 2 + y 2 A,x 2 + A 2 y 2 f ( A 1 x 2 + A 2 y 2 A 3 Aj^-A^n _ p . /a /a ( a a + b 2 ) \ x + ^ 2/ 2 a 2 + Z> 2 * A^ + A 2 */ 2 / J ~" ' which reduces to where / denotes the focal length of the lens when the rings are photo- graphed. When A 3 vanishes absolutely, i.e., in the case of a uniaxial crystal cut perpendicular to its optic axis, a = b, Ai = A 2 , and hence equation represents two sets of circles of slightly different radii. In the general case, the equation represents two series of curves of the fourth degree. The lines of minimum visibility are given approximately by , [*+*p> {l + AlCOs2 t + £ 2Sin2< H l = const., (17) _a 2 + b 2 2v I a 2 + b 2 J J K ' which becomes, on squaring and reducing, B 2^k (a 2 + W) 2 ( Ai cos 2 <£ + A 2 sin 2 0) 2 -sin 2 ^ • A 3 (a 2 + & 2 )(AiCos 2 ^-A 2 sin 2 <^>) = const., (18) where squares of A 3 have been neglected. In Cartesian co-ordinates equation (18) may be written, if x 2 + y 2 =/ 2 sin 2 g), (»■ + Vf <^+^g -(»■ + y) A, A -y '£ = const, 186 Mr. T. K. Chinmayanandam. This is of the form {(oc-ct) 2 + (l + /3)y 2 } {(x + *Y + (l + lS)y*} = const which is the equation of an oval, similar to that of Cassini. 4. Numerical Besults. If we confine ourselves to the plane containing the optic axes and the plane perpendicular to it, the positions of successive rings can be easily worked out directly without any approximations. Also, the observation of the number of rings between the intersections of the two successive lines of minimum visibility with those planes enables us to calculate with very fair accuracy the ratios of the principal refractive indices, if we only know roughly the mean refractive index of the specimen and the angle between the optic axes. Consider first the plane containing the optic axes. The section of the wave surface by that plane is a circle of radius b and an ellipse whose axes are a and c. Bays incident in this plane are refracted and reflected, and finally emerge in the same plane ; and if 8 be the path difference for a ray emerging at an angle ft, we have, for the refracted ray corresponding to the circular section, 4yu,Veos 2 r = 46 2 (/* 2 2 -sin 2 ft) = S 2 =' n 2 \ 2 , (20) or sin 2 ft = A*a 2 --*-?■• ( 21 ) 4er Corresponding to the elliptic section, we have ix 2 = fjL ± 2 cos 2 r -f- fis 2 sin 2 r. (22) It can easily be shown also that sin ft = fjij sin r fi' Eliminating r from (22) and (23), we get (23) a 2 = . f^J^ (24) ^ ^ 4 + (^i 2 -^3 2 )sin 2 ft- V ; Also 8 = 2e seer {//,-- sin ft sin r}, (25) which after reduction becomes sin* £•„ = ^-^ . ^. (26) It may be noted that equations (21) and (26) are identical with the general equations (9) and (10) when cf> = 0. For the plane at right angles to that containing the optic axes, it can be similarly shown that n 2 X 2 t . o o o n 2 X 2 2 sir ft = /if-77 and sir ft = ^ — — . ^— respectively. fl2 4e» iU ~ ^ 4e* * •■-» On Haidingers Rings in Mica. 187 We proceed to show how these simple formulae can be made use of to determine the ratios of the principal refractive indices of mica. Suppose the first line of minimum visibility intersects the plane <fi = at a point which corresponds to order n of one set and order (m + |) of the other (m^>n). Let the next line of minimum visibility lie on orders (n — r) and [m -f- \ — (r + 1)] respectively of the two sets of rings. Then, . 2 . o n 2 X 2 o (m-f4) 2 \ 2 u$ sm 2 Jb = /* 2 2 — ^ = /*3 2 - v A % } . ^ 2 4^2 A* 4^2 * f ._2> sin 2 ft/ - 2_ 0-Q 2 ^ 2 _ 2 _[m + l-(^+l) ] 2 X 2 M3 2 so that 2m - r 2 = { 2 (m + i) (r + 1) - (r + 1) 2 } ^, fix or since, in a fairly thick plate, r is small compared with n or m, /i3 2 __ W, r 2- • ^t (2"7) Similarly for the plane perpendicular to that containing the optic axes we get ^i_rn/_ r' /*2 2 " n f * 7+T {M) In equations (27) and (28), the quantities r, r' are known, being simply counted off on the ground glass plate of the camera or the photographic negative ; m, n, m', and n\ are all of the order 2fie/\, where jjl is the mean refractive index of the mica, e its thickness, and X the wave-length of the incident light. In the actual specimen used, //, was about 1*62, e = 0*0144 cm., and X = 4*36xl0~ 5 cm., so that m, n, etc., were of the order 1050. The differences m—n, w! —n f , being very small compared with m, n, it is sufficient for our purpose to know the absolute values of the latter roughly, provided the differences m—n and m f — n\ which are small and integral, are known accurately. Now, in the direction f = in the plane <£ = which corre- sponds to that of the single ray velocity in the crystal, the respective orders of the two sets of rings are the same. Further, the difference between the respective orders of the rings will increase by unity each time as we pass successive lines of minimum visibility. Thus if z be the number of lines of minimum visibility between the directions § i> = and f = #, we have m — n = z — 1 and m'— n' = z — 1 + 1 = z. If the field of view covers the direction f = 0, z can be directly counted off. Otherwise, we can calculate the same by a simple method. It can easily be shown that the number of rings between successive lines of minimum 188 Mr. T. K. Chinmayanandam. visibility is nearly the same, so that if we know the number N" of rings lying between the first and the last of those lines in the region 0<£o<# 5 we have r{z—l) = N". N is determined by the equation XT 2e f 1 /- sin 2 0\*"l r N = T /i i 1 -( 1 — W-lS~~2~ no ' where n is the number of rings from the centre up to the first line of minimum visibility. It is sufficient if we use the mean value for /x and an approximate value for 0. In the specimen used by the writer, the angle between the optic axes was about 73°, so that 6 = 37° nearly. The values of the other quantities are as follows :— r = 21*5, r f = 20-5, n = 4, N = 61, 3—1 == m—n = 3, m' —n' = 4, fisffjLi, fi^j fX2y can now be calculated from equations (27) and (28). The refractive indices were also measured with a ref ractometer, and the calculated and observed values of their ratios are given below. Observed. Calculated. 0-9761 0-9760 0-9790 0-9785 5. Summary and Conclusion, This paper is an attempt at a fuller study of Haidinger's rings in mica, some special features of which due to double refraction were described by Lord Eayleigh in a paper in the 'Philosophical Magazine' for November, 1906. The principal results obtained here are as follows : — (1) The regions of minimum visibility in the field, due to the superposition of two independent sets of rings, lie along a series of curves which, in muscovite, are approximately hyperbolas. Lord Kayleigh's observation that the lines of minimum visibility are crosses traversing the centre of the field is true only for particular thicknesses, and, even then, is not a complete description of the phenomena, as two series of hyperbolas are seen in addition to the cross. In phlogopite, the rings are indistinct along closed curves of oval shape. (2) The usual method of observing reflected systems of rings, using a plate of glass inclined at 45° to the plate under observation, is not satisfactory for the observation of Haidinger's rings in mica. The effects of a want of flatness, or other defect, in the mica plate, were considerably minimised in a new method devised in the present investigation for observing and photographing the On Haidinger's Rings in Mica. 189 reflected system of rings. A diffusing screen is placed close to the mica with its surface parallel to it, and the rings are observed through a hole in the centre of the former. (3) In mica, as in all other crystals which cleave or which are cut perpen- dicular to one of the axes of optic symmetry, theoretical considerations indicate that the lines of minimum visibility in Haidinger's rings should be practically the same as the isochromatic lines observed in a plate of twice the thickness in convergent polarised light. This conclusion is in agreement with the observations described above. (4) In such crystals, if the angle between the optic axes is also large, the rings are shown to be approximately two sets of ellipses, given respectively by the equations a 2 y 2j tc 2 x 2 = const., c 2 y 2 + b 2 x 2 = const., where a, h, c are the principal velocities in the crystal. The major axes of one set are in the same direction as the minor axes of the other. These results are not far from the truth in the case of muscovite. (5) The ratios of the principal refractive indices in mica can be found with fair accuracy by observing the number of rings lying between successive lines of minimum visibility in the plane containing the optic axes and in a perpen- dicular plane. The investigation was carried out in the Physical Laboratory of the Indian Association for the Cultivation of Science, Calcutta, and the writer's best thanks are due to Prof. C. V. Raman, who suggested the investigation and took much interest in its progress. He has also much pleasure in acknowledging the kindness of Dr. H. H. Hayden, F.R.S., who put some specimens of phlogopite mica at the author's disposal, and of Dr. G. T. Walker, F.R.S., whose interest in the work and encouragement have been of the greatest value to the author of this paper.