It will render the description which follows easier, to suppose that the pole of the Od-subject which causes clock-rotation be turned directly from the experimenter; for example, that an egg be placed longways to the experimenter with its small end from him, or a bar magnet with its northward pole from him. In the case of a horse-shoe magnet, both poles are then turned from you. So, too, in the case of the hand, the fingers and thumbs are both to be turned away.
_a._ Odometer held immediately before, and a quarter of an inch from, the small end of an egg. _Result_—Distad motion, or motion in the direction of the long axis of the egg, from the egg.
_b._ Odometer similarly held to the great end. _Result_—Proximad motion of the odometer—that is, again, motion directly from the axis of the egg.
_c._ Hold the odometer near either side of the egg, one-fifth of the distance from either end. _Results_—Transverse sinistrad oscillations. The same current may be detected above the egg on the same parallels.
The effects described are given in fig. 17. Then here are, at either end of the egg, two rectilinear currents acting at right angles to each other. Fig. 18 represents the complementary motions to the above, which are obtained by touching the odometer finger with the thumb. A parallel combination of rectilinear motions is produced, but in another way.
The next three figures exemplify the composition of forces necessary to produce the rotatory motion of the odometer.
Figs. 19 and 20 are intended to represent the large ends of two eggs, so placed that the axial currents of the two shall cross at right angles, at a point equidistant, let us say at half an inch exactly, from the end of each egg. If the ring be suspended exactly at the point of meeting of the two forces, it will be driven off in the diagonal, and continue simply to oscillate in the line A B. But if either of the eggs is moved back to double its former distance from the point of intersection of the two forces, the forces will be rendered unequal, and new results will ensue. The two experiments by which the results of this arrangement may be tried are represented in figs. 21, 22, and figs. 23, 24. The longer current is necessarily thereby the weaker of the two in each combination. Accordingly, rotatory motion supervenes in each case instead of diagonal oscillation; and the direction of the rotation is from the stronger towards the weaker current.