There are two ways of rendering an electric appliance safe.  One is to use low potentials, the other is to determine the dimensions of the apparatus so that it is safe no matter how high a potential is used.  Of the two the latter seems to me the better way, for then the safety is absolute, unaffected by any possible combination of circumstances which might render even a low-potential appliance dangerous to life and property.  But the practical conditions require not only the judicious determination of the dimensions of the apparatus; they likewise necessitate the employment of energy of the proper kind.  It is easy, for instance, to construct a transformer capable of giving, when operated from an ordinary alternate current machine of low tension, say 50,000 volts, which might be required to light a highly exhausted phosphorescent tube, so that, in spite of the high potential, it is perfectly safe, the shock from it producing no inconvenience.  Still, such a transformer would be expensive, and in itself inefficient; and, besides, what energy was obtained from it would not be economically used for the production of light.  The economy demands the employment of energy in the form of extremely rapid vibrations.  The problem of producing light has been likened to that of maintaining a certain high-pitch note by means of a bell.  It should be said a barely audible note; and even these words would not express it, so wonderful is the sensitiveness of the eye.  We may deliver powerful blows at long intervals, waste a good deal of energy, and still not get what we want; or we may keep up the note by delivering frequent gentle taps, and get nearer to the object sought by the expenditure of much less energy.  In the production of light, as far as the illuminating device is concerned, there can be only one rule—­that is, to use as high frequencies as can be obtained; but the means for the production and conveyance of impulses of such character impose, at present at least, great limitations.  Once it is decided to use very high frequencies, the return wire becomes unnecessary, and all the appliances are simplified.  By the use of obvious means the same result is obtained as though the return wire were used.  It is sufficient for this purpose to bring in contact with the bulb, or merely in the vicinity of the same, an insulated body of some surface.  The surface need, of course, be the smaller, the higher the frequency and potential used, and necessarily, also, the higher the economy of the lamp or other device.

This plan of working has been resorted to on several occasions this evening.  So, for instance, when the incandescence of a button was produced by grasping the bulb with the hand, the body of the experimenter merely served to intensify the action.  The bulb used was similar to that illustrated in Fig. 19, and the coil was excited to a small potential, not sufficient to bring the button to incandescence when the bulb was hanging from the wire; and incidentally, in order