Let us take a flat rod of ordinary hoop iron, 30 or more centimeters in length. If, while holding this vertically, we give freedom to its molecules by torsions, vibrations, or, better still, by a few blows with a wooden mallet upon its upper extremity, we find, as is well known, that its lower portion is strongly north, and its upper south. If we reverse this rod, we now find it neutral at both extremities. We might here suppose that the earth's directing force had rotated the molecules to zero, or transversely, which in reality it has done, but only to the limit of their comparatively free motion; for if we reverse the rod to its original position, its previous strong polarity reappears at both extremities, thus the central point of its free motion is inclined to the rod, giving by its free motion great symmetrical inclination and polarity in one direction, but when reversed the inclination is reduced to zero. In Fig. 3, D shows the bar of iron when strongly polarized by earth's magnetic influence, under vibrations, with a sufficient force to have rotated its elastic center of action. C shows the same bar with its molecules at zero, or transversal, the directing force of earth being insufficient without the aid of mechanical vibration to allow them to change. The dotted lines of D suppose the molecule to be in the center of its free motion, while at C the molecules have rotated to zero, as they are prevented from further rotation by being at the extreme end of its free motion. If, now, we hold the rod vertically, as at C, giving neutrality, and give a few slight blows with a wooden mallet to its upper extremity, we can give just the amount of freedom required for it to produce evident polarity, and we then have equal polarity, no matter which end of the bar is below, the center of its free rotation here being perfect, and the rod perfectly neutral longitudinally when held east and west. If, on the other hand, we have given too much freedom by repeated blows of the mallet, its center of free motion becomes inclined with the molecules, and we arrive at its first condition, except that it is now neutral at D and polarized at C. From this it will be seen that we can adjust this center of action, by vibrations or blows, to any point within the external directing influence. [Illustration: FIG. 3.]