ing ceased, and could not be obtained again by screwing up the knobs. On the other hand, when the sheet was placed at double this distance (about 5 meters), the sparking was slightly greater than without the sheet.

Now these three observations can only be explained by the interference and re-enforcement of a direct action of the "vibrator" with one reflected from the metallic sheet, and in addition by the supposition that the action spreads out from the vibrator at a finite velocity. According to this explanation, in the first position the reflected part combines with the direct and reinforces its effects. In the second position (that of no sparking), the reflected effect in going to the sheet and returning has taken half the time of a complete vibration of the "vibrator," and so is in the phase opposite to the incident wave, and consequently interferes with it.

If it were possible to tell the direction of the current in a "resonator" at any moment, then, by employing two of them, and placing one just so much beyond the other that the currents induced in them were always in opposite directions, we would obtain directly the half-way length. Now by reflection, we virtually are put in possession of two "resonators," which we are enabled to place at this distance apart, although unable to tell more than whether there be a current or not.

The distance from the position of interference to the sheet is a quarter of the wave-length, being half the distance between these simultaneous positions of opposite effects.

In the third position, the reflected wave meets the effect of the next current but one, in the "vibrator," after the current it itself emanated from, and since these two currents are in the same direction, their effects re-enforce each other in the "resonator." This occurs at half the wavelength from the sheet.

The first two observations alone could be explained by action at a distance, by supposing the currents induced in the metallic sheet to oppose the direct action in the "resonator" everywhere, and by also supposing that in the immediate neighborhood of the sheet, the direct action is overmastered by that from the sheet, while at 2.5 meters away the two just neutralize each other.

On this explanation, at all distances further the direct action should be opposed by that from the sheet, so that the fact of being increased at 5 meters upsets this explanation. Again, behind the sheet, evidently on this supposition, the two actions should combine so as to increase the sparking, but instead of this the sparking was found to cease on placing the sheet in front of the "resonator."

In performing these experiments, the "resonator" circle was always placed in the position in which only the magnetic part of the disturbance had effect. Hertz has also used the other positions of the resonat ing circle, whereby he has observed the existence of an electric disturbance coincident with the magnetic one, the two together forming the e lete electro-magnetic wave.

Ordinary masonry walls were found to be transparent to radiation of this wave-length (that is, of about 10 meters), and some visitors to the opening meeting of the Dublin University Experimental Society, last November, were much astonished by seeing the sparking of the resonating circle out in the College Park, while the vibrator was in the laboratory. Attempts were first made last December to obtain reflection from the surface of a non-conductor, with the hope of deciding by direct experi ment whether the magnetic or electric disturbance was in the plane of polarization; that is, to find out whether the "axis of the vibrator" should be at right angles to the plane of reflection or in it, when at the polarizing angle, for obtaining a reflected radiation. It is to be ob served that in these radiations the electric vibration is parallel to the "axis of the vibrator" while the magnetic is perpendicular to it, and that they are consequently polarized in the same sense as light is said to be polarized.

Two large glass doors were taken down and used for this purpose, but without success; and until lately, when reflection from a wall was tried, the experiment seemed unlikely to be successful.

In working with the glass plate, the resonator circle was first placed so that the" vibrator" had no effect on it. Then the glass plate was carried into position for reflection, but without result, though even the reflection from the attendants moving it was amply sufficient to be easily detected.

To obviate a difficulty arising from the fact that the wave was diverg ent, we decided to try Hertz's cylindrical parabolic mirrors, for concentrating the radiation. Two of these were made with sheets of zinc nailed to wooden frames, cut to the parabolic shape required.

70 C.

In the "focal line" (which was made 12.5 centimeters from the vertex) of one of these, a "vibrator" was placed, consisting of two brass cylinders in line, each about 12 centimeters long and 3 centimeters in diameter, rounded at the sparking ends.

In order that the "resonator" wire may lie in the "focal line" of the receiving mirror, it has to be straight; this necessitates having two of them. They each consist of a thick wire 50 centimeters long, lying in the "focal line," and of a thin wire, 15 centimeters long, attached to one end at right angles, and which passes out to the back of the mirror through a hole in the zinc, where the sparking can be viewed, without obstructing the radiation in front. The total length of each "resonator" is about two wavelengths, the wave-length being about 33 centimeters, so that it may be that there are two vibrating segments in each of these "resonators."

...........

Side elevation.

FIG. 3.-Plan.

With this apparatus it is possible to deal with definite angles of incidence. No effect was obtained with glass plates using

these mirrors, whether the "vibrator" was perpendicular to the plane of reflection or in it. But with a wall 3 feet thick reflection was obtained, when the "vibrator" was perpendicular to the plane of reflection; but none, at least at the polarizing angle, when turned through 90° so as to be in it.

*

This decides the point in question, the magnetic disturbance being found to be in the plane of polarization, the electric at right angles. Why the glass did not reflect was probably due to its thinness, the reflection from the front interfering with that from the back, this latter losing half a wave-length in reflection at a surface between a dense and a rare medium; and, as Mr. Joly pointed out, is in that case like the black spot in Newton's rings, or more exactly so, the black seen in very thin soap-bubbles.

FIG. 4.

Hertz has pointed out several important things to be guarded against in making these experiments. Ultra-violet light, for example, falling on the "vibrator," prevents it working properly, the sparking in the resonator ceasing or becoming poor. Also the knobs of the "vibrator" must be cleaned of burnt metal, and polished every quarter of an hour at least, to prevent a like result.

Both these effects probably arise, as suggested by Mr. Fitzgerald, from a sort of initial brush discharging (either ultra-violet light or points being capable of doing this), which prevents the discharging impulse being sufficient sudden to start the oscillation in the "vibrator." For to start a vibration, the time of impulse must be short compared with the time of oscillation. These precautions therefore become especially needful when working with small-sized "vibrators." Possibly charging the "vibrator" very suddenly, after the manner of one of Dr. Lodge's anti-lightning-rod experiments, would save the irksome necessity of repeatedly cleaning the knobs of the "vibrator."

Several important problems seem to be quite within reach of solution by means of these Hertzian waves, such for instance as dispersion. Thus it could be tried whether placing between the reflector and the "resonator" conducting bodies of nearly the same period of vibration as the waves used would necessitate the position of the "resonator"

*Slight reflection was obtained at an incidence of 70°.

being changed so as to retain complete interference. Or again, whether interspersing throughout the mass of a large Hertzian pitchprism, conductors with nearly the same period would alter the angle of refraction. In some such way as this, anomalous dispersion, with its particular case of ordinary dispersion, may yet be successfully imitated.

The determining the rate of propagation through a large tile, or sheet of sandstone, could be easily made by means of the interference experiment, by placing it between the screen and the "resonator."

EXPERIMENTS ON ELECTRO-MAGNETIC RADIATION,

INCLUDING SOME OF THE PHASE OF SECONDARY WAVES.*

In continuation of some experiments which were described in Nature, vol. XXXIX, p. 391 (" Repetition of Hertz's Experiments and Determination of the Direction of the Vibration of Light"), attempts were made to obtain periodic reflection of electric radiation from plates of different thicknesses, analogous to Newton's rings, with the view of further identifying these radiations with "light."

It was there described how a sheet of window-glass refused to reflect the Hertzian waves, but how a masonry wall reflected them readily. The non-reflection from the thin sheet is due to the interference of the reflected waves from each side which takes place owing to a change of phase of half a period on reflection at the second surface, as in the black spot of Newton's rings.

By making the reflection plate such a thickness that the reflection from the back has to travel half a wave-length farther than that from the front, the two reflections ought to be in accordance, for they differ by a whole period, half arising from difference in path, and half from change of phase on reflection; but if the difference in paths were made a whole wave-length by doubling the thickness of the plate, there ought again to be interference, and so on.

The first plan tried with this end in view, was to fill a large wooden tank to different depths with water or other liquids. On gradually filling the tank reflection should be obtained, and at a certain depth equal to (sec r), reach a maximum; further addition of the liquid then should diminish the reflection, and at double the above depth the reflection should reach a minimum, the two waves interfering.

The mirrors for concentrating the radiation had for this purpose to

* From Nature, August 22, 1889, vol. XL, pp. 398–400.

be suspended over the tank as shown in the figure. The tank was first tried empty, but unfortunately the wooden bottom was found to reflect,

FIG. 5.

thus it was useless for the purpose intended. I then tried what ought to have been tried before constructing the tank, namely—whether ordinary boards, such as flooring, reflected. The floor was found to reflect readily. This was attributed to moisture in the wood causing it to conduct, specially as wood was found not to polarize by reflection. Experiments were then undertaken to determine if water reflected, even though in thin sheets. A large glass window was placed beneath the mirrors and flooded with water; this was found to reflect well, both when the mirrors were in the position shown and when rotated to the position "at right angles." Thus water also acts like a metal, reflecting the radiation however polarized. The glass had to be hardly more than damp to get some reflection.

The wooden tank being unsuitable, a glass tank was thought of, but was given up for solid paraffine, which, being in slabs, could be easily built up into a vertical wall of any desired thickness. Through the kindness of Mr. Rathborne a large quantity of this was lent for the purpose.

A thin sheet of paraffine about 2 centimeters thick was found not to reflect, as was expected. Next a wall 13 centimeters thick (180 centimeters long, 120 centimeters high) was tried, and found to reflect, this being the thickness required in order to add another half period to the retardation of the wave reflected from the back at an incident angle of 55°, the wave-length being taken as 66 centimeters, and the index of refraction being taken as 1.51, the square root of 2.29, the value taken as the specific inductive capacity of paraffine.

Then a wall twice the thickness was tried, but it also reflected, contrary to expectation. While in doubt as to the cause of this, it was decided to make a determination by direct experiment of the index of refraction of paraffine for these waves, by a method suggested in Nature (vol. XXXIX, p. 393), which consists in interposing a sheet or wall of paraffine between the resonator and the metallic reflection in the Hertzian experiment of loops and nodes which are formed by the interference of the reflected wave with the direct radiation; the ratio of the velocity in the wall to that in the air being easily found from the observed shifting of the loops and nodes towards the screen.