6o: 



NA TURE 



[April 30, 1885 



of speaking disrespectfully of anomalous dispersion, which 

 he says is quite as important as double refraction. I grant 

 this, but my interpretation of his language when I heard 

 the lecture was that so many possible ways had been 

 shown of explaining anomalous dispersion that it was 

 mere child's play (or fiddle-playing) to discuss it while 

 the burning question of double refraction awaited ex- 

 planation, upon which question seems to depend the 

 whole safety of the wave-theory of light, that theory being 

 in imminent danger of destruction therefrom. 



I shall now give a brief account of the gyrostatic mole- 

 cules, crude and improved. The crude one is a fly-wheel 

 inside a massless shell. Here there is no gyrostatic 

 action opposing a motion of translation, but only op- 

 posing a motion of rotation. This is the molecule 

 which was stated to give the wrong kind of variation 

 of magneto-optic rotation with variation of wave-length. 

 The improved gyrostatic molecule (p. 320) consists of two 

 fly-wheels on one axis. But the axis is cut in two in the 

 middle between them, and the parts fitted together by a 

 ball and cylinder joint. The other ends of the half axes 

 are supported in ball-and-socket joints in the massless 

 shell. So far as rotation of the shell is concerned, this 

 acts like one gyrostat, the axis always remaining in one 

 line. But if the shell be frictionless, the ether can only 

 give translational movement to it, and the double gyrostat 

 produces a gyrostatic effect when the molecule is accele- 

 rated in any direction except along the axis. 



The special function of this molecule is to explain 

 magneto-optic rotation of the plane of polarisation. The 

 axis of the molecule is supposed to be the direction of the 

 lines of force. It is required to be proved that, gyrostatic 

 molecules being imbedded in the ether with their axes 

 parallel and their directions of rotation the same, the 

 velocity of propagation of a circular disturbance going 

 with the gyrostat is greater than that of a circular dis- 

 turbance in the opposite direction. With a steady propa- 

 gation of circularly polarised light, the gyrostats will 

 clearly execute a precessional motion. The theory of this 

 motion is examined after the manner of Thomson and 

 Tait's 'Natural Philosophy" for a ray along the axes, 

 and the gyrostatic effect is found to be equivalent to 

 altering the effective density of the molecule, and so 

 altering the velocity of propagation. Thus if v and V are 

 the velocities of propagation along the axis of rays 

 polarised circularly in the two directions, it comes out that 

 approximately 



-, = !+// -, 



V y 



where // is a constant depending on the form of the gyro- 

 stats, co is the angular velocity of the precessional rotation 

 of the gyrostats, and y is the velocity of rotation of the 

 gyrostats. This is a totally different law to the action of 

 the crude gyrostatic molecule, and is in accordance with 

 experiment. 



If now we have improved gyrostatic molecules im- 

 bedded in the ether, their minute rotations will affect the 

 velocity of propagation in the manner of crude molecules, 

 but their translations will affect the velocity in the manner 

 now elucidated. But observe that by diminishing the size 

 of the molecules the influence of the rotational motion 

 diminishes, but the influence of the translational motion 

 remains the same (on the assumption that the angular 

 gyrostatic velocity is kept the same and the ratio of mass 

 of gyrostats to mass of molecule remains the same). 

 Hence, if we have small enough molecules, the law which 

 agrees with experiment alone holds. This is a very satis- 

 factory state of affairs, and I believe it is the first time 

 that Sir Willi im Thomson's hint about this phenomenon, 

 so long ago thrown out, has been developed. 



There is still so much matter in the lectures that I have 

 not touched upon that I am in some difficulty as to what 

 to omit. But I certainly should like to transcribe nearly 



the whole of the last lecture. This is of course imposs- 

 ible, but I will claim a little space for some remarks on 

 Rankine's beautiful but futile attempt to get over the 

 fatal difficulty of double refraction (p. 271) : — 



" Suppose here a massless rigid lining of our ideal 

 cavity in the luminiferous ether. Let there be a massive, 

 heavy molecule inside, with fluid around it. The main 

 thing is that this molecule, which only affects the effective 

 inertia of the ether by adding its own mass to the moving 

 mass of the ether, has solotropy of inertia. Imagine this 

 spherule (drawing on the board an oblate spheroid with 

 axis vertical) moving first in a horizontal direction. The 

 effective inertia of this sheath will be altered if it moves 

 to and fro in a vertical direction, there being, by hypo- 

 thesis, liquid between it and the ether. The density of 

 this mass must be greater than the density of the liquid, 

 that is all. If there is danger of its coming to the sides of 

 the cavity, let there be springs to keep it in place, if you 

 like, but let its connection with the lining of the cavity 

 be in the main through fluid pressure. Then its effective 

 inertia is different in different directions. This fluid 

 lining seems to hit off the very thing we wanted. Now 

 comes Rankine's want of strength. He cut around the 

 edges of it, and, I think, rather jumped at it, and put 

 down a wave-surface the same as Fresnel's, and said that 

 it came to that. But, alas 1 Stokes (long before Lord 

 Rayleigh suggested it) showed that it would give a differ- 

 ent surface from Fresnel's. Lord Rayleigh, in repeating 

 Rankine's suggestion, showed his strength where Rankine 

 was not so strong in mathematical powers of grappling 

 with a difficult mathematical problem. Lord Rayleigh is 

 a man who grapples with a difficulty and sees how much 

 he can do with it. He puts it aside if he cannot solve it, 

 but he never shirks it. Rankine was not a mathematician 

 in that sense at all. Lord Rayleigh finds, not Fresnel's 

 wave-surface, but a wave-surface differing from Fresnel's 

 by certain terms appearing in reciprocals instead of 

 directly." 



Now Stokes has shown that Huyghen's construction 

 satisfies experiment with great accuracy, and hence Ran- 

 kine's effort fails. The desperate condition of the wave- 

 theory is shown by the words penned by Lord Rayleigh 

 before he knew of Stokes's experiments (p. 272) : " Should 

 the verdict go against the view of the present paper, it is 

 hard to see how any consistent theory is possible which 

 shall embrace at once the laws of scattering, regular 

 reflection and double refraction." 



It appears, then, that after all the labour which has 

 been expended upon the wave-theory of light, it fails 

 absolutely, and, as it seems, hopelessly, in two points of 

 primary importance. One is the extinction of the ray 

 polarised by reflection ; the other is double refraction. 

 In other matters we have difficulties, but we can see a 

 possible means of escape. Here there seems to be none. 



Before concluding this series of articles I wish to say a 

 little more about the manner of their delivery. It is a 

 rare experience for students to have the opportunity of 

 studying the workings of a great mind while grappling 

 with a problem. This is what occurred during the three 

 weeks of the Baltimore lectures. During the whole of 

 this period one or two ardent students were hunting up 

 references in the Peabody Library, &c, and literally filled 

 Sir William Thomson's rooms with the results of their 

 searches, and Sir William generally read these books. 

 He says (p. 76) : — " An interminable number of books 

 have been brought to me, and in every one of them I 

 have found something very important." But at p. 9S he 

 says : — " I got another quarter-hundredweight of books 

 on the subject. I have not yet read them all through." 

 In this way he often came for the first time upon re- 

 searches bearing on the question in hand. Thus (p. 77) : 

 •' I only found this morning that Lommel also goes on 

 to double refraction of light in crystals [with imbedded 

 molecules]. The very problem I am breaking my head 



