_d._ Repeat the experiment, with the thumb applied. _Results_—Oscillations ensue of the same length, and they are again unequal, but in the contrary direction. Now they do not reach the pile of sovereigns by an inch, but they pass three inches in the opposite direction. [Illustration: Diagram showing figs. 15-24] [Illustration: Diagram showing fisg 25-29] Thus a force is brought into view having this new quality: when the Od-relations of the experimenter are versed, a change ensues, not into motion in a plane transverse to the former one, but the direction of the new motion is simply the opposite of the first, or the odometer appears to be attracted or repelled towards the Od-subject alternately. _e._ Try the same experiment with a single sovereign, or with the zinc disc. _Result_—The odometer held at four inches distance is attracted and repelled just as in the preceding instance. Then an irregular form of the Od-subject, or its unequal mass at different parts, have nothing to do with this new motion; and it is evident that the relation of the latter to the former class of oscillatory motions will be easily determinable. _f._ Lay the proper disc before you (see fig. 15,) and hold the odometer over the production in each direction of its transversal line beyond the limits of the disc. _Results_—When held near the right edge of the disc, as before mentioned, a dextrad motion is developed; that is to say, the odometer moves off from the dextral cardinal point of the disc, oscillatively. This movement, or those oscillations outward, are fainter and fainter, as the odometer is held over points more and more remote from the disc. At length, at the distance of an inch and a half, the odometer becomes absolutely stationary. When moved, however, still farther off, motion begins again, which is very lively at four to five inches distance from the disc, its direction being sensibly toward the disc. Moved farther off, still the same motion continues, and is detectable ten to twelve inches off the Od-subject. When the same experiments are made on the left edge of the Od-subject, phenomena just the reverse are manifested for the same distance. The extra-marginal dextrad motion is transverse for an inch and a half. Then there occurs a point of quiescence; on the other side of which the odometer swings in free and long sinistrad or repelled oscillations. _g._ Repeat these experiments (fig. 16,) with the thumb applied. _Result_—On the left side the near extra-marginal dextrad motion is replaced by a tangential proximad motion; and the centrifugal oscillations beyond the point of quiescence are replaced by centripetal oscillations. On the right side again, the near dextrad extra-marginal oscillations are replaced by a proximad tangential current: while beyond the point of quiescence, the remote centripetal oscillations are reversed into centrifugal ones.