2 feet.

5 centim. long and 4 centim. mean diameter, the winding at the ends being exposed so that it could be put up close to the iron. A search-coil of 100 turns, about 1 centim. in diameter, was mounted so that it could be fixed at different positions relatively to one of the solenoids, which was then placed upon the iron plate* and upon the other solenoid alternately. A current of constant value was turned on and off in each position of the solenoid, and the throw on a ballistic galvanometer in the circuit of the search-coil observed. The following table, giving the throw of the needle for different positions of the search-coil, shows how perfectly the iron plate mirrored the magnetism of the solenoid. The discrepancy in the figures arises more from errors of observation than anything else.

*The iron plate was ordinary boiler-plate inch thick, size 3 feet by

The first position in the above table is the most important, as the coil is placed in a part of the field that experiences the greatest change by the presence or absence of the iron plate, the difference being that between 236 and 129. If the searchcoil could have been placed so that its centre was at the very end of the solenoid, the throw ought to be just one half of what it would be with the coil continued to double its length. This was found to be so when the solenoid was placed on the other solenoid, but could not be tried with the iron as part of the search-coil protruded.

The experiment was then made of lifting the solenoid step by step a little distance from the iron plate, when it was found that the presence of the iron plate had the effect of producing an image of the coil receding an equal amount behind the plate.

A larger search-coil was placed in different positions outside the solenoid; and in all positions on which any readable throw of the needle could be obtained, it was found that the iron plate gave an effect almost exactly the same as an equal solenoid in the position of its image.

In order to try the effect of tilting the solenoid, it was mounted on a strip of wood hinged to another, the angle

Fig. 4.

between them being fixed by a clamp (fig. 4.) This was then placed alternately on the iron plate and against a similar

hinged strip upon which the other solenoid was placed, the angle of inclination being the same for both solenoids (fig. 5).

Fig. 5.

The angle was altered step by step from 0 to 90 degrees, and in all cases the iron plate produced the same effect as its image within limits not greater than those shown in the previous table. It was, however, only when the solenoid was placed at the end of the hinged strip nearest the iron plate that readings of any value could be obtained, as small dislocations in the position of the coil laterally were apt to produce changes of reading comparable with those due to considerable changes in angle. At distances greater than 5 centim. the iron had very little effect. The search-coil during these experiments was moved about to different positions both inside and outside the solenoid, preference being given to the most sensitive positions.

Then the effect of differently-shaped solenoids was tried. A long narrow one and a wide flat one were used with similar effects, so far as they could be measured.

There is a difference between the magnetic reflexion of an iron plate and the optical reflexion of a mirror in the following respect an optical image has those of its lines which are at right angles to the mirror reversed while those parallel to the mirror are not reversed. For instance, if we hold a magnet with its north end pointing from us and towards a mirror, the optical image of it will have its North end pointing towards us (fig. 6). But in the case of a magnetic image this is inversed: the magnetic image has its South end pointing towards us. In fact the reflexion of a North end is a South end. If, however, we look at the reflexion of the currents flowing in the solenoid instead of at the sign of the pole of the solenoid, we see that laws of the optical case hold good

for the magnetic ease. The image (fig. 7) of a current flowing in any direction in a plane parallel to the mirror is another current flowing parallel to the first and towards the same part.

A North pole viewed from the front is equivalent to a lævocheiral circulation of current. The image of this pole turned towards the mirror is, to the spectator, a dextro-cheiral circulation, or a South pole. The same is true if, as in fig. 8, the solenoid has its axis parallel to the plane of the mirror.

In order to try some effects with an isolated pole, a solenoid,

2 metres long and 1.5 centim. mean diameter, was uniformly wound with twelve turns per centimetre of a suitable wire. When a current of 15 amperes was passed through this a pole of a strength of about 40 units (C.G.S.) was obtained, and this might be regarded as an isolated pole at distances not greater than 30 centim. from it. It was found that bringing an iron plate within a few centimetres of this pole produced very nearly the same effect upon the field as would have been produced by bringing up a pole of opposite sign to a point at double the distance.

The field close to the iron plate ought to have been exactly twice as strong as if no iron plate were there. This was found to be so within the limits of error of observation (about 3 per cent.) when the pole was within 6 centim. of the iron plate. As the distance was increased the throw of the needle became too small to be accurately read, and the field being extremely weak the susceptibility of the iron would be very low.

It was this weakness of the field (for one could not consider the pole as a point except at distances exceeding 10 centim.) that spoilt some experiments on kaleidoscope effects. One can conceive that if the laws of reflexion hold good between a magnetic pole and a sheet of iron, it ought to be possible, by means of several sheets of iron, to build up a magnetic kaleidoscope by which the effect of a great number of magnetic poles could be produced from a single pole. The idea of a magnetic kaleidoscope was suggested by Lord Kelvin many years ago, but it is doubtful whether it has ever been successfully carried out. If, for instance, two plates of susceptible