shaken by the following experiment. I had a steel cylinder constructed, bored longitudinally in the direction of the axis; it was 0ᵐ·13 in length, 6ᵐ·002 inner diameter, and 0ᵐ·005 solid thickness. This cylinder was suspended by several silk threads not twisted. Its inherent magnetism was very weak, the duration of an oscillation amounting to 103″·5. The inner surface was exposed to two magnetic bars which were only 0ᵐ·004 in thickness, and were made to slide with great caution in the direction of the inner angles, as is usually done in magnetising by the double touch. Four poles, placed symmetrically, having been rubbed each ten times, the effect of these forty frictions was to reduce the duration of an oscillation from 103″·5 to 31″·1. The operation having been repeated, but with the other poles, distanced from the first 45°, the duration of an oscillation was reduced to 26″·95, and after a third repetition, to 25″·7. Repeated frictions had no more influence on the duration of the oscillation, which might, however, still be reduced to 23″·1, by rubbing the outer surface in a similar way. This experiment proves that, in fact, the inner surface is susceptible of being rendered magnetic, but we must suppose that the separation of the magnetic fluid takes place only in the rubbed portions, as is the case in the experiments of M. Haldat. A magnetism diffused uniformly over the entire surface cannot be admitted. This experiment, simple though it be, it seems to me, may nevertheless contribute to establish our notions as to the distribution of magnetism.
As to the action of an electro-dynamic helix placed in the interior of a hollow steel cylinder, and traversed by magneto-electric discharges, no variation in the magnetic condition was perceptible. Care had been taken to isolate the helix as well as possible, and to destroy in part the magnetism which the cylinder had acquired. The duration of an oscillation amounted to 36″, and this time did not vary even when a succession of discharges in the same direction were made o pass through the helix. I was not able to try more energetic discharges: but the development of the magnetism being subject to circumstances which depend entirely on peculiar and accidental properties of the steel, it cannot be exactly predicted whether electric currents would or would not produce some effect of magnetization.
B2
