speed which we had reason to suppose must be infinitely accelerated. This speed had never surpassed 120—130 revolutions in a minute, on employing a pile of four pairs of plates two feet square. We must not lightly abandon conclusions founded on the nature of things, and those to which I refer are drawn solely from the integral , expressing the magnetic attraction and supposed to be independent of the speed. Besides, it rests upon the legitimate supposition that the electro-magnetic excitation of the soft iron operates instantaneously. If this were not the case, my apparatus would have shown that magnetism and electricity ought to be attributed to the motion of material particles, or to oscillations much more perceptible than are those of the propagation of sound. In short no one can deny that it is the nature of a force not to require time to act, and that, if its different effects were not instantly perceptible, it would then be some molecular motion, under the influence of mechanical laws, which takes place.
18.
At the end of my first note I said, that in using thermo-electric piles for the movement of machines, there was reason to fear the magneto-electric currents developed by magnetism in motion. The reaction which thence arises would be almost entirely destroyed in the hydro-electric pile, the liquid conductors offering too much resistance to the passage of these currents. These considerations were founded upon detached experiments. On employing a thermo-electric pile, the deviation of the needle was affected by a magnet which had been placed in a helix forming part of the circuit; this was not the case with a voltaic pair of plates of small dimension. The deviation of the galvanometer, extremely sensitive as it is, was not altered by it. This did not surprise me, since the conducting power of liquids is much below that of metals; but in making experiments on the magnetic force of a bar of soft iron, I have sometimes found considerable differences for which I could not account. I was curious to know if these differences proceeded from the weakening of the electric current produced by a pair of plates with a surface of half a square foot, or from the nature of the iron. I therefore inserted in the circuit a galvanometer at some little distance, that it might not be affected immediately by the magnetism of the bar. I was much astonished to see the needle recede upon my applying the armature, and advance as soon as it was taken off; for it was the first time that I had recognised the double office of the connecting wire, viz. that of conducting the voltaic current, and of representing at the same time a common wire subjected to the influence of a magnet in motion. The helix producing a magnet by the voltaic current is at the same time a magneto-electric
