accompanied by the rumbling noises which are heard in a telephonic circuit which contains a microphone.

Now the most striking peculiarity of such a column is that, while it is in the non-conducting state, if a spark is allowed to pass anywhere in its neighbourhood between the knobs of a Leyden jar or the poles of an induction-coil, the waves sent out from the spark at once render the column a conductor.

This experiment was reproduced at the meeting of the Physical Society on October 27 by Mr. Croft, and it at once struck me as bearing a strong resemblance to the effects produced by electromagnetic waves on photoelectric impulsion-cells. These cells I have already fully described (see 'Philosophical Magazine,' March 1891); and in addition to what I showed with regard to them, I may here add that the effects of the sparks of an induction-coil on an impulsion-cell were observed when the cell (enclosed in a dark box) was over 140 feet distant from the coil and from everything connected with it. Moreover, it was found that, in most cases, the power of electromagnetic radiations to change the state of the cell from one of insensitiveness to one of sensitiveness to light depends on the length of the electromagnetic waves; for, by adding capacity (by means of Leyden jars) to the sparking circuit, the waves ceased to affect the cell, and, on removing this capacity, their ability to effect the change was restored.

I have always held that this action of the cell is due to a rearrangement of the molecules on the sensitized surface of the plate in the cell by the electrical disturbances produced in the wires connected with its poles, these disturbances being due to the radiations emanating from the sparking circuit; and I found that by completely enclosing the cell, together with its terminal wires, in a metal box, the ability of the radiations to act on the cell was destroyed.

After having seen the experiment shown by Mr. Croft, I filled several glass tubes with the filings of copper, tin, zinc, bismuth, antimony, &c., and found M. Branly's result in every case. It seems to me that if the filings are extremely fine (powders, in fact) the result is much more difficult to obtain. Powders used in this way are very great insulators, and they must be strongly pressed together by the corks which close the glass tubes in order to conduct; but then it becomes difficult to alter their conductivity by mechanical disturbances. On the other hand, a tube filled with very coarse iron filings was found to conduct under all circumstances; so that, apparently, the filings must be neither very fine nor very coarse.

As in the case of the impulsion-cells, the effect of the electromagnetic disturbances is cut off by completely enclosing the tubes and their terminal wires in a metal box. If the terminal wires are allowed to project outside the boxwhether they are connected with the battery and galvanometer or not the tubes are affected, which clearly shows that the result is due to the electrical disturbances set up in the leading wires of the tube.

In order to establish a closer connexion with the impulsioncells, it seemed to me that films containing metallic particles almost in mathematical contact, but yet free to undergo very minute displacements of rotation, should replace the tubes of filings. Accordingly, I have formed such films by imbedding fine metallic powders in layers of gelatine and of collodion. The powders which I employ in these films are very much finer than the filings which I have used in the glass tubes.

The media, or bases, in which these powders are embedded are (so far as I have gone) gelatine and collodion.

To form a gelatine film proceed thus :-Pour a very thin layer of dissolved gelatine over a glass or ebonite plate; let this layer get almost quite dry; if it gets quite dry, hold it over a beaker of boiling water until it absorbs a sufficient quantity of steam to render it very slightly plastic: take a test-tube with fine metallic powder immersed in alcohol; shake this up well, and rapidly pour some of it over the gelatine surface, securing, to all appearance, both continuity and uniformity of the metallic layer. The alcohol soon evaporates, leaving the metallic layer embedded in the gelatine, but not wholly sunk below it.

In the case of a collodion film, the metallic powder is shaken up in a test-tube and then rapidly poured over a glass plate. After some time the film dries and peels off the plate. The surface of the film which was in contact with the glass is the metallic surface sought; the other surface of the film probably contains no metallic particles, and is simply a continuous surface of dry collodion which is an absolute nonconductor.

If the layer of gelatine above described is too thick, it will probably happen that its upper surface (that remote from the glass or ebonite plate) does not contain metallic particles, and is a non-conductor. In this case, if the film is peeled off the plate there remains on the plate a metallic surface which will answer our purpose.

Such a film may now replace the tube of filings in M. Branly's experiment.

Let ABCD represent the film, L a battery of one or two voltaic cells, G a galvanometer, and K a key which can be connected with the film by means of the wires s, w, the

former clamped to the film at a point P, while the latter ends in a stout platinum piece the end of which is a round knob touching the film at a movable point Q.

Now it is found that when the contacts are all made no current passes, and we have to begin by making the platinum knob touch the film very close to P-perhaps millim. from P. Still no current passes, even when a spark of an induction-coil is allowed to pass a few feet. from the film. But by touching either of the wires s, w with an electrified body (that which I commonly use is a common gas-lighting electrical machine), the resistance of the film and the contacts P, Q is overcome, and a strong current is shown by the galvanometer. We can now gradually increase the distance between P and Q each time that Q is removed from P, the wire s or w being touched by the electrified body; and by these successive steps the whole of the film is finally rendered conducting.

The special characteristic of such a film which I wish to point out is this: if, while the film is conducting, the current is stopped by breaking the contact at Q, the film will be found to be a non-conductor if the contact is almost instantly re-made at the same point Q; while, if the contact is broken anywhere else in the circuit, as at K, and then re-made, the film will still be found to conduct.

Again, if when the film has been rendered a conductor, the circuit is broken at K, and left broken for (apparently) any length of time, and then re-made, the film will be found to be still a conductor.

The breaking of the contact at Q is instantly fatal to the conductivity of the film, if the film is not many days old; but, after some days, it will be found that on breaking the contact at Q and then re-making it, the film remains a conductor, unless we delay the re-making of the contact for half

a minute or so. Doubtless this is due to the fact that, as the film gets more hard and less plastic, the metallic particles find it more difficult to move, by rotation or otherwise, in their environment.

Finally, these films are very much less sensitive to mechanical disturbances given to their supports than the tubes of filings. They do not appear to be susceptible to the action. of heat; but in many instances their conductivity was destroyed by breathing upon them, or by allowing a stream of steam to strike their surfaces, the conductivity being always restored by the electromagnetic radiations. The more rapid vibrations of light have not produced any effect so far as my observations have gone.

The prime cause of the action is to be sought in the electrical surgings produced in the leads s and w; but it is clear that the state of affairs at the places P, Q of contact of the film with the electrodes is a most important matter.

It seems clear, therefore, that the sensitized surface in an impulsion-cell has a close analogue in a slightly plastic film filled with almost mathematical completeness by fine metallic particles.

VI. On the Sudden Acquisition of Conducting-Power by a Series of Discrete Metallic Particles. By Prof. OLIVER J. LODGE*.

THE

HE recent experiments of Mr. Croft and Prof. Minchin remind me of an observation I frequently made when engaged with the syntonic arrangement of Leyden-jar circuits, or sympathetic electric resonance. I found, if the knobs of the receiver were very close together, a weak battery and bell being in circuit, that the occurrence of a scintilla at the receiver frequently caused the bell to ring for some time, and in general to show signs that the knobs were in a state of feebly adhesive contact. It was just as if their surface-layers or skins had been broken through, or opened out, in such a manner as to increase the molecular range of a few of the closest superficial molecules and thereby to cause cohesion to set in at a distance considerably greater than the ordinary distance.

A phenomenon which may be similarly caused is that discovered by Lord Rayleigh, with regard to the effect of electrified sealing-wax near a vertical water-jet: drops which would otherwise have rebounded being thereby caused to cohere. And, by the method of two impinging jets, the necessary

* Communicated by the Physical Society: read November 24, 1893, as a contribution to the discussion on Prof. Minchin's paper.

difference of potential was shown to be only a volt or two, since a pair of impinging jets ceased to rebound if connected to the opposite terminals of a Grove cell.

Once more, the effect first observed by Mr. Guitard (1850), and rediscovered by myself and the late J. W. Clark, concerning the adhesion of dust-particles or mist-globules in electrified air, is but a more violent variety of the same sort of effect. And the action of electricity on a steam-jet discovered by Robert von Helmholtz and afterwards by Shelford Bidwell, and worked at by Richarz, Aitken, J. J. Thomson, and others, may not be very different.

Thinking of the Lord Rayleigh variety of experiment as in many respects the most definite, it is natural to explain it electrolytically as due to the polarisation of the water-drops, or the formation in each drop of molecular chains each with a negatively charged oxygen atom at its termination on one hemisphere and a positively charged hydrogen atom at its opposite extremity.

Such drops then colliding about the region of their poles would be attracted not only by their ordinary cohesive forces, but by electrical force also, and thus the effective molecular range would be increased and cohesion might set in over an unusual distance.

In some such way I have allowed myself to fancy that the adhesion of my knobs might be explained, and I suggest that the conductivity of a chain of metallic filings under an electric polarising influence may be due to something of the same

cause.

That a tap should break the minute points of contact is likely enough, though I see no reason why another tap should restore communication,-if so it does.

At the last moment I write this and send it off in the hope that it may arrive in time to be read at the Meeting as a contribution to the discussion on these interesting experiments.

Nov. 23, 1893.

VII. On the Magnetic Shielding of Concentric Spherical Shells. By A. W. RÜCKER, F.R.S.*

SPE PECIAL interest has of late attached to the problem of shielding magnetic forces by means of iron screens, and recently the Astronomer Royal has described the arrangement (designed by Messrs. Johnson and Phillips) by which

* Communicated by the Physical Society: read November 24, 1893.