320 



NATURE 



{Jan. 31, 1889 



The Hall effect observed in conductors follows at once ; 

 for the rotation of a displacement is equivalent to com- 

 bining it with a small perpendicular displacement ; and 

 it is this perpendicular or transverse E.M.F. exerted by 

 a magnetic field which Hall discovered. At the same 

 time, there are one or two facts which militate against 

 this view of the Hall effect, chief among which is the 

 singular behaviour of nickel, which rotates light one way 

 and electric displacement the other way. For some time 

 it v/as possible to hope for a way out of this through the 

 usual convenient avenue of " impurity " ; but now that 

 both experiments have been performed on the same 

 identical piece of metal, still with opposite results, this 

 exit is closed. In this unsettled state, so far as I know, 

 the connection between the rotation of light and the Hall 

 effect at present stands. 



It may be well here to repeat the caution, appended as 

 a footnote to the last article, not to assume that this 

 account of the magnetic rotation of light and the Hall effect 

 is true. If true, however, it is convenient as linking the 

 phenomena on to hysteresis, and the direction of the 

 effect in iron is correctly given — namely, a rotation 

 against the magnetizing current. 



Prof. Ewing has since pointed out, in a letter to me, 

 that, attending more precisely to the instruction of his 

 curves, we find the difference in ^ for positive and nega- 

 tive magnetizing forces only lasts through a number of 

 cycles for the time during which the final state has been 

 approached, and does not persist after a steady state has 

 been reached. This would make the magnetic rotation 

 of light a function of time ; and certain experiments by 

 Villari on spinning a glass disk between the poles of 

 a magnet, so that fresh and fresh portions of glass were 

 continually exposed to the magnetic field, showed a 

 marked falling off in the amount of rotation as soon as 

 high speeds were obtained ; thus proving, apparently, 

 that a certain short time was necessary to set up the 

 effect. This experiment, and other modifications of it, 

 want repeating, however. 



Prof. Ewing has subsequently expressed a doubt as to 

 whether the kinematic resolution of a displacement into 

 two equal opposite circular components is, under the 

 circumstances, legitimate. 



Prof. Fitzgerald has further pointed out that, although 

 when attending to one element only the theory might 

 possibly work, yet, as soon as one take? into account the 

 whole wave-front, it breaks down ; for all the main mag- 

 netic disturbance lies in the wave-front, as is well known, 

 and the extra magnetic disturbance which I have postu- 

 lated as a consequence of electrostatic displacement is 

 annulled by interference of adjacent elements. 



If I were quite sure that there were no vestige of truth 

 in the suggestion I have made, I should, of course, with- 

 draw it ; but, as I do not feel perfectly sure either way, I 

 leave it in a dilapidated condition for the present. 



Another and apparently distinct accomit of the mag- 

 netic rotation can also be hinted at, which links the 

 phenomenon on to the facts of thermo-electricity. It 

 labours under worse disadvantages than the preceding, 

 being more hazy. 



Referring back to Nature, vol. xxxvii. p. 12, we find 

 that, to explain what is called the "Thomson effect " in 

 metals, we were led to suppose a connection between one 

 kind of electricity and some kinds of matter more intimate 

 than between the other kind of electricity and the same 

 matter. Thus, the atoms of iron were said to have a 

 better grip of positive electricity than of negative ; while 

 copper, on the other hand, had a better grip of negative 

 than of positive. All metals could be arranged in one or 

 other of the two classes, with the exception of lead, which 

 appears to grip both equally. It is the same phenomenon 

 as was originally named by Sir W. Thomson, " the specific 

 heat of electricity in a substance." Certain it is that 

 vibrating atoms of iron push positive electricity from the 



more rapid to the less rapid places of vibration — that is, 

 from hot to cold— and a whole class of the metals do the 

 same ; while another class, like copper, push it from cold 

 to hot. 



Permitting ourselves to picture this effect as a direct 

 consequence of the Ohm's law relation between electricity 

 and matter, combined with a special relationship between 

 certain kinds of matter and one or other kind of electricity, 

 a relationship which can exhibit itself in other ways also, 

 we get a possible though rather hazy notion of a Faraday 

 rotation in a magnetic field by supposing that the 

 Amperian molecular currents in these substances consist 

 not of precisely equal positive and negative currents, but 

 of opposite currents slightly unequal ; say, for instance, 

 that the density of the positive constituent of the bound 

 ether of a substance is slightly different from that of the 

 negative constituent, so that on the whole the bound ether 

 in a magnetized molecule is slowly rotating one way or 

 the other, at a pace equal to the resultant rotation 

 of its constituents. Suppose that in iron the positive 

 Amperian electric current is the weaker of the two, then 

 the ether, as a whole, will be rotating with the negative 

 current, and accordingly an ethereal vibration entering 

 such a medium will begin to screw itself round in a 

 direction opposite to that of the magnetizing current. 

 Whereas in copper or other such substance it would be 

 rotated the other way. 



According to this (admittedly indistinct) view, lead 

 ought to show no rotatory effect at all ; and of course, 

 therefore, no Hall effect either. And the classes into 

 which metals are divided by the sign of their Hall effect 

 should coincide with the classes into which the sign of 

 their Thomson effect throws them. 



Hall finds that, of the metals he examined, iron, cobalt, 

 and zinc fall into one class, while gold, silver, tin, copper, 

 brass, platinum, nickel, aluminium, and magnesium fall 

 into the other. Now, referring to the thermo-electric 

 results of Prof. Tait, we find iron, cobalt, platinum, and 

 magnesium with a negative sign to their Thomson-effect- 

 coefficient, or with lines in the thermo-electric diagram 

 sloping downwards ; while gold, silver, tin, copper, 

 aluminium, and zinc slope upwards, or have a positive 

 sign to their " specific heat of electricity." 



According to this, therefore, the discordant metals are 

 zinc, platinum, and magnesium. The proper thing to 

 say under these circumstances is that the metals used 

 in the very different experiments were not pure. They 

 certainly were not ; but I do not feel able to con- 

 scientiously bolster up so inadequate a theory by help 

 of this convenient fact. 



In the Philosophical Magazine for May 1885, Mr. Hall 

 gives some more measurements, showing that in bismuth 

 the effect is enormous, and in the same direction as in 

 copper, whereas in antimony it is also great, and in the 

 same direction as in iron. All these things seem to point to 

 some thermo-electric connection — whether it be of the sort 

 I have vaguely tried to indicate, or some other. 



Other Outstanding Problems. 



Outstanding problems bristle all over the subject, and 

 if I pick out any for special mention it will only be because 

 I happen to have made some experiments in their direc- 

 tion myself, or otherwise have had my thoughts directed 

 to them, and because they have not been so directly called 

 attention to in the body of the articles. 



Referring back to the end of Part II., "a current re- 

 garded as a moving charge," it is natural to ask. Is this 

 motion to be absolute, or relative to the ether only, or 

 must it be relative to the indicating magnetometer } In 

 other words, if a charged body and a magnetic needle are 

 flying through space together, as, for instance, by reason 

 of the orbital motion of the earth, will the needle experience 

 any deflecting couple ? 



It is one of many problems connected with the ether 



