May II, 1 893 J 



NATURE 



35 



of physical optics, namely, the action of an electromag- 

 netic field upon light. The discoveries which have 

 hitherto been inade may be classified under four heads : 

 (1) Faraday's experiments, which show that when plane 

 polarised light is transmitted through a transparent mag- 

 netised medium, a rotation of the plane of polarisation is 

 produced ; (2) Kerr's experiments, which show that the 

 effect of electrostatic force on a transparent medium is 

 to convert it into one which is optically equivalent to a 

 uniaxal crystal whose axis is in the direction of the force ; 

 (3) Kerr's experiments on the reflection of plane polarised 

 light at the surface of a magnetised iron reflector, which 

 show that a rotation of the plane of polarisation of the 

 reflected light takes place, which in certain cases is in the 

 same and in others in the contrary direction to that of 

 the amperean current which may be conceived to produce 

 the magnetic force ; (4) Kundt's experiments on the re- 

 flection of light from magnetised iron, cobalt, and nickel, 

 and also on the transmission of light through thin mag- 

 netised films of these metals. There is also another series 

 of experiments by Kundt, in which polarised light is 

 refracted at the upper surface of a plate of glass, is then 

 reflected at the lower surface, and again refracted at the 

 upper surface. The results of these experiments show that 

 the plane of polarisation of the ultimately emergent light 

 is rotated in the contrary direction to that produced by 

 an iron reflector. 



There seems to be a fair amount of evidence to lead to 

 the conclusion that Hall's effect is intimately connected 

 with the action of a magnetic field upon light, but further 

 evidence is required before it can be asserted that both 

 phenomena are due to the same ultimate cause. Up to 

 the present time Hall's effect has, 1 believe, only been 

 ■detected in conducting media ; but if it be assumed to be 

 capable of existing in transparent media, theory furnishes 

 results which, as far as they have been worked out, are 

 in agreement with experiment. HalFs effect is capable of 

 explaining the experiments of Faraday, and it also gives 

 a result in accordance with Kundt's experiments on re- 

 flection and refraction from a plate of magnetised glass 

 in the case in which the magnetisation and incidence are 

 normal. It would be quite possible to apply this theory 

 to the case of oblique incidence, but the work would be 

 laborious and the final results complicated. The experi- 

 ments of Prof Dewar on liquid oxygen would seem to 

 provide a more promising way of testing this theory, for, 

 on account of the high susceptibility of this substance to 

 magnetic action, it is possible that an effect might be 

 observed in the case of direct reflection. ^ According to 

 theory, Hall's effect ought to be positive in the case of 

 glass and gaseous oxygen, and negative in the case of a 

 solution of perchloride of iron ; and a repetition of 

 Kundt's experiments, in which the latter liquid is em- 

 ployed in the place of glass, ought to show that the rota- 

 tion takes place in the same direction as that produced 

 by metallic iron. Such experiments would be valuable as 

 a further test of the theory, but they do not appear to 

 have been made. 



A paper recently communicated to the Cambridge Philo- 

 sophicalSociety(Mayi)stillfurtherconfirmstheview which 

 I have put forward. In this paper I have transformed 

 the formuliE for reflection at a magnetised transparent 

 medium by assuming that the refractive index is a com- 

 plex quantity. The resulting formulae for the amplitudes 

 of the reflected vibrations agree very well with Kerr's 

 experiments so far as qualitative results are concerned, 

 provided the values and signs of certain quantities are sup- 

 posed to be determined by optical, as distinguished from 

 electromagnetic methods. They are, moreover, the same 



" I'he effect produced hy a. single reflection from magnetised glass would 

 be too feeble to be detected ; but Dr. Kerr suggested to me that an effect 

 might possibly be observed by employing the method of multiple 

 refleciions. *^ 



NO. 1228, VOL. 48] 



as regards their form as those deducible from Maxwell's 

 theory by taking into account the conductivity combined 

 with Hall's effect ; but unfortunately the values of certain 

 constants, when expressed in terms of electrical quantities, 

 differ from the values which are required by optical ex- 

 periments, in a manner which prevents a perfectly satis- 

 factory electromagnetic theory being constructed in this 

 way, and 1 doubt whether it will be possible to attain the 

 j end in view until a theory based upon the mutual reaction 

 I of ether and matter has been discovered in which quanti- 

 ties, upon which the motion of matter depends, enter into 

 combination with electromagnetic quantities. 



Although the sign of Kerr's effect m nickel is the same 

 : as in iron and cobalt, the sign of Hall's effect is different. 

 This difficulty is apparent rather than real, for a theory 

 ' based upon the mutual reaction of ether and matter might 

 I very well introduce a factor containing the free periods of 

 j the vibrations of the matter which would change the sign 

 I of the magnetic terms. Some light might be thrown 

 \ on this point by determining the principal incidence and 

 azimuth for nickel and cobalt. 



The generally received theory, that reflection and re- 

 fraction are materially influenced when any of the free 

 periods of the vibrations of the matter fall within the 

 limits of the visible spectrum, suggests that the sign of 

 Kerr's effect may be different in the case of the ultra-violet 

 and the infra-red portions of the spectrum from what it is 

 in the luminous portion. Experiments on this branch of 

 the subject are needed, and possibly the employment of 

 a fluorescent substance, such as quinine, in the case of 

 the ultra-violet waves, or of a solution of iodine in disul- 

 phide of carbon, in conjunction with Prof Langley's 

 bolometer,'- when the infra-red waves are experimented 

 upon, might furnish important information on this point. 

 The experiments of Kerr on the effect of electrostatic 

 force suggest that if light were reflected from a strongly 

 electrified metallic conductor, certain peculiarities would 

 be observed. In the absence of experiments, which do 

 not appear to have been made, it would be impossible to 

 predict with certainty what these effects are likely to be ; 

 but it would seem probable that an electrified metallic 

 reflector would behave like a doubly-refracting metallic 

 medium having a jzVi^/^ optic axis which is perpendicular 

 to the reflecting surface. When light is reflected from 

 the surface of a uniaxal crystal which is cut perpendicu- 

 larly to the axis, the component vibration at right 

 angles to the plane of incidence is reflected in the same 

 manner as if the medium were isotropic. Under these 

 circumstances we should anticipate that in the case of an 

 electrified metallic reflector, the component vibration in 

 the plane of incidence would be much more strongly 

 affected by electrification than the component at right 

 angles to this plane. If this speculation should be verified 

 by experiment, it would follow that the principal inci- 

 dence and azimuth, and also the difference between the 

 changes of phase of the two components, would be 

 affected by electrification in a manner which could be 

 observed. 



In conclusion I would point out that further experi- 

 ments are required of the following nature : — 



(1) Experiments on the reflection of light from mag- 

 netised transparent media, such as glass, perchloride of 

 iron, and also if possible from liquid oxygen. 



(2) Experiments on reflection from and transmission 

 through magnetised metals, special attention being paid 

 to the effects produced by the n6n-luminous portion of 

 the spectrum. 



(3) Experiments on reflection from electrified metallic 

 reflectors. 



A. B. Basset. 



' I do not know whether the bolometer is more sensitive to heat than a 

 pair of average eyes are to light ; if it is, experiments on the infra-red waves 

 ought to be easier than experiments on luminous waves. 



