obliged to consider the values of l alone in relation to the different amalgams. As we have already seen from the values obtained by the present authors for copper and gold amalgams, and those obtained by Johns and Evans for silver amalgams (as seen in Table X.), it can be concluded that the value of l, at the same temperature, is the same, within experimental error, for amalgams of all the metals in Group 1 (b) in the Periodic Table. This means that, atom for atom, the effect of copper, silver, and gold on the conductivity of mercury is the same, provided the concentration is small.

*

A further point in Skaupy's theory is that, neglecting the variation of viscosity with concentration, the value of

should be constant. At 11.5° C. and 100° C. there

() should be

was very little change in the value of l with concentration,

and, consequently, the values of (영)

at those temperatures

are not discussed. However, at 300° C., as seen in Table IX., the value of (-1) increased with concentration in the case of both copper and gold amalgams. The variation of (1-1) with "C" is shown in Graph V., and the relation between the two quantities is approximately a linear one,

thus

showing that () is approximately constant. It is impor

tant to point out here that an accurate experimental determination of is very difficult for low concentrations and that any error in its value involves a much greater percentage error in the value of (-1).

*Loc. cit.

A possible explanation of the increase in the value of

1 AL

1 = c I with temperature is that the electron concenCL tration in the amalgam as compared with mercury is increased with increase of temperature; but the question of whether viscosity enters into the problem in any form

must be left until experiments have been carried out on alkali amalgams.

SUMMARY.

1. The conductivities of dilute amalgams of copper and gold have been determined at 11.5°, 100°, and 300° C. In the case of gold amalgams, at each temperature the conductivity was determined over a range of concentrations equal

to the maximum range possible at 11.5° C. ; whilst for copper amalgams, at each temperature the determinations were made for increasing concentrations up to the limit of solubility of copper in mercury at that temperature. For gold amalgams of concentrations "C"=1221 and "C"=2441, the conductivity was also measured at 217.3° and 257.5° C.

2. The average temperature coefficients of resistivity between 11.5° and t° C. of gold amalgam of concentration "C"=1221, and of copper amalgam of concentration "C" =0322, were measured, and in both cases were found to increase as the temperature difference increased. However, the values were less than the corresponding ones in the case of pure mercury.

3. The values of the average temperature coefficients of resistivity of gold amalgams of various concentrations, between 11 5° and 100° C., showed a diminution as the concentration increased.

4. In the case of gold amalgams, at 11.5° and 100° C. the increase of conductivity relative to the conductivity of mercury at the same temperature was practically proportional to the concentration, but at 300° C., in the case of both copper and gold amalgams, this was found not to be the case.

1 ΔΙ

5. For copper and gold amalgams, the value of CL

(i. e., the ratio of the increase of conductivity relative to mercury to the concentration) was determined at each temperature and for each concentration, and the values compared with the corresponding ones obtained by Johns and Evans* for silver amalgams. It was found that the value of AL

1 AL

at infinite dilution (i. e., for extremely small concentrations) was practically the same at the same temperature for amalgams of copper, silver, and gold. This means that, atom for atom, the effect on the conductivity of mercury of the different metals in Group I. (b) in the Periodic Table is

the same.

October 1st, 1928.

* Loc. cit.

CXXI. Electronic Waves and the Electron. By Sir J. J. THOMSON, O.M., F.R.S., Master of Trinity College, Cam bridge*.

SUMMARY.

RECENT experiments have shown that a moving electron is accompanied by a train of waves. No such waves would be produced by the motion of an electron if, as hitherto assumed, it consisted solely of a point charge of electricity. The electric and magnetic forces round a moving electron of this type can be calculated from Maxwell's equations and have long been known, and there is nothing in the distribution of these forces approaching that in a train of waves. In this paper it is shown that if the structure of the electron were such that this point charge or something analogous to it formed a nucleus which was surrounded by a system such as we shall proceed to describe, then the motion of the electron would from the ordinary laws of electro-dynamics give rise to a train of waves, and, moreover, that the relation between the wave-length of this train and the velocity of the electron is exactly that indicated by the experiments of G. P. Thomson. The electronic waves on this view are electrical waves, but they do not travel through the normal ether, but through an ether modified by the system which envelops the nucleus of the electron. These waves would be produced if the system enveloping the nucleus, and which we shall call the "sphere" of the electron, were made up of parts which can be set in motion by electric forces, and when in motion produce the effects of electric currents. Such a structure might consist either of a distribution of discrete lines of force, or of a number of positively- and negatively-electrified particles distributed through the sphere of the electron. These would behave like free particles, even though the opposite charges were bound together in doublets, if the frequency of the forces acting upon them were large compared with the natural frequencies of the doublets. The properties of a structure of this kind are discussed, and it is shown that the sphere of the electron would have a definite period of vibration, the frequency of the vibration being proportional to the square root of the number of electrified systems per unit volume. These vibrations are of a peculiarly interesting kind, inasmuch as though they are electrical vibrations they are not accompanied by any radiation of energy, so that when once started

Communicated by the Author.

they are maintained for an indefinite time. The vibrations consist of an oscillating electric field which is not accompanied by a magnetic one, the Poynting vector vanishes, and there is no transmission of energy. The sphere of the electron can thus vibrate, and so also can the nucleus, the time of vibration for the nucleus being proportional to the time light takes to travel round its circumference. Thus the two parts of the electron, the nucleus and the sphere, are each capable of vibration, and when the electron is in a steady state the vibrations of the two parts will be in resonance. The electron has thus, in addition to the steady electric field due to the negative charge on the nucleus, an alternating field in which the energy remains constant since there is no radiation. The oscillating field is the seat of energy, and thus the total energy of the electron is that due to the charge on the nucleus plus that due to the oscillating field. This is important in connexion with the calculation of the size of the electron. The usual estimate 14 x 10-13 cm. for the radius, a, is deduced on the assumption that the energy due to the charge on the nucleus e2/2a accounts for the whole of the energy of the electron; if e2/2a represents but a part, and it may be a small part, of the total energy, the corresponding value of a would be much larger.

Stationary Electron.

When the nucleus is not in motion it is shown that the components of the oscillating electric and magnetic forces are represented by equations of the form

where Po is the period of vibration of the electron and fa function such that

An interesting special case representing a symmetrical electron is when

(X, Y, Z) (a, B, y) are the components of the electric and magnetic forces respectively and the distance from the centre of the nucleus. Here the oscillating electric force is radial and always proportional to the steady force due to the