If, these objects being removed, I lay a horse-shoe magnet on the table before me, with its poles turned directly _from_ me, the northward limb being on my left hand, the southward limb-pole on my right, and experiments parallel to those just described are made, the results remain the same. If, near one side of the horse-shoe magnet, I lay my left hand on the table, the palm downwards, the thumb held wide of the fingers, the ring, if suspended half an inch from and above either of the finger-points, displays clock-rotation; suspended similarly before and above the point of the thumb, versed-rotation. Or the fingers of the left hand, so disposed, may be compared, in reference to Od, to the northward pole of a horse-shoe magnet, while the thumb corresponds with its southward pole. If, removing my left hand, I turn the horse-shoe magnet, without altering the side on which it rests, half round, so that the poles point directly _towards_ me, the northward pole being now, of course, on my right, the southward pole on my left, the ring held as before over either of the two poles, displays the same results. If I now move the magnet still nearer to me, so that its two poles are an inch beyond the edge of the table, I can obtain results which furnish a more precise explanation of the two rotatory movements already described, than I had before arrived at. If I now suspend the ring, with its lowest part on a level with the magnet, and half an inch from its northward pole—that is, half an inch nearer me—it begins to oscillate longitudinally, with a bias towards me, as if it were repelled from the pole of the magnet. If I then suspend the ring an inch vertically above the first point of suspension, it begins to oscillate transversely, with a bias towards the right, or as if impelled by a dextrad current. If I then lower the ring half an inch, the first effect observed is, that it oscillates obliquely, being evidently impelled at once to the right and towards me—that is, in the diagonal of the two forces, of each of which I had before obtained the separate influence. In this third variation of the experiments, I have brought the ring to the limit of the two currents, where both tell upon it. This oblique oscillation soon, however, undergoes a change: it changes into clock-rotation, showing that the transverse or dextrad current is stronger than the longitudinal or proximad current. If parallel experiments be made at levels below that of the pole of the magnet, corresponding but opposite results ensue. If the whole series be repeated upon the south pole of the magnet, opposite but perfectly corresponding results are again obtained: and similar results may be obtained with the two poles of an egg.