Strutt[123] has recently described a simple and striking experiment to illustrate still more clearly that a radium preparation acquires a positive charge, if it is enclosed in an envelope thick enough to absorb all the α particles, but thin enough to allow most of the β particles to escape. The experimental arrangement is clearly seen in Fig. 27. A sealed tube _AA_ containing the radium, was attached at one end to a pair of thin gold leaves in metallic connection with the radium, and was insulated inside a larger tube by means of a quartz rod _B_. The inner surface of the tube was coated with tinfoil _EE_ connected to earth. The glass surface of _AA_ was made conducting by a thin coating of phosphoric acid. The air in the outer tube was exhausted as completely as possible by means of a mercury pump, in order to reduce the ionization in the gas, and consequently the loss of any charge gained by the gold leaves. After an interval of 20 hours, the gold leaves were observed to diverge to their full extent, indicating that they had acquired a large positive charge. In this experiment Strutt used ½ gram of radiferous barium of activity only 100 times that of uranium. [Illustration: Fig. 27.] If the tube is filled with 30 mgrs. of pure radium bromide, the leaves diverge to their full extent in the course of about a minute. If it is arranged that the gold leaf, at a certain angle of divergence, comes in contact with a piece of metal connected with earth, the apparatus can be made to work automatically. The leaf diverges, touches the metal, and at once collapses, and this periodic movement of the leaf will continue, if not indefinitely, at any rate as long as the radium lasts. This “radium clock” should work at a sensibly uniform rate for many years, but, from evidence considered later (Section 261), there is reason to believe that the number of β particles emitted would decrease exponentially with the time, falling to half value in about 1200 years. The period of movement of the leaf should thus gradually increase with the time, and ultimately the effect would become too small to observe. The action of this radium clock is the nearest approach to an apparent perpetual motion that has so far been observed.