of this kind the two ends repulse and attract the liquid at the same time.  Their motions are of the same phase; if it were desired that one should repulse while the other was attracting, it would be necessary to place two drums back to back, separated by a stiff partition, and put them in connection with two distinct pump chambers whose movements were so arranged that one should be forcing in while the other was exhausting.  A system of this nature is shown to the right in Fig. 1.

The vibrations are obtained by the aid of small metal spheres fixed in tubular supports by movable levers to which are communicated the motions of compression and dilatation of the air in the pump chamber.  They oscillate in a plane whose direction may be varied according to the arrangement of the sphere, as seen in the two apparatus of this kind shown in Fig. 1.  Fig. 2 will give an idea of the general arrangement.  The two pistons of the air-pumps are connected to cranks that may be fixed in such a way as to regulate the phases as may be desired, either in coincidence or opposition.  The entire affair is put in motion by a wheel and cord permitting of rapid vibrations being obtained.  The air is let into the apparatus by rubber tubing without interfering with their motions.

[Illustration:  FIG. 2.]

We may now enter into the details of the experiments: 

The first is represented in Fig. 2.  In a basin of water there is placed a small frame carrying a drum fixed on an axle and capable of revolving.  It also communicates with one of the air cylinders.  The operator holds in his hand a second drum which communicates with the other cylinder.  The pistons are adjusted in such a way that they shall move parallel with each other; then the ends of the drums inflate and collapse at the same time; the motions are of the same phase; but if the drums are brought near each other a very marked attraction occurs, the revolving drum follows the other.  If the cranks are so adjusted that the pistons move in an opposite direction, the phases are discordant—­there is a repulsion, and the movable drum moves away from the other.  The effect, then, is analogous to that of two magnets, with about this difference, that here it is the like phases that attract and the different phases that repel each other, while in magnets like poles repel and unlike poles attract each other.

It is necessary to remark that it is indifferent which face of the drum is presented, since both possess the same phase.  The drum behaves, then, like an insulated pole of a magnet, or, better, like a magnet having in its middle a succeeding point.  In order to have two poles a double drum must be employed.  The experiment then becomes more complicated; for it is necessary to have two pump chambers with opposite phases for this drum alone, and one or two others for the revolving drum.  The effects, as we shall see, are more easily shown with the vibrating spheres.