(1.) Introduction.—In previous papers I have had occasion to discuss at some length the relations between Joule's discovery of the elongations accompanying magnetization in iron and Wiedemainn's discovery of the twist produced in an iron wire under the combined influence of longitudinal and circular magnetizing forces. Following out a suggestion of Maxwell's as to the intimate connection between these two phenomena, I turned my attention to nickel, in the hope of finding the Wiedemann effect in it opposite to what it is in iron in low magnetic fields. The fact that, as Barrett ‡ had shown, nickel contracted in length when longitudinally magnetized, whereas, in low fields, iron elongated wds the ground for this hope, which experiment fully justified.

In a recent paper on the strains produced in iron, steel, nickel, and cobalt tubes in the magnetic field (Trans. Boy. Soc. Edin., vol. xxxix., 1898), I have obtained, on sufficiently reasonable assumptions, values for the strain coefficients at the inner and outer walls of these tubes. Hitherto, work by other experimenters on similar lines had been on ellipsoids, rods, or wires, solid through-out. The theoretical discussion of the Wiedemann Effect (as I have called it) was rendered the more difficult on this account, and also because the current producing the circular magnetization flowed through the magnetized material. In my paper of 1888 referred to above, I make an attempt to get a formula by means of which the Wiedemann effect may be calculated from the Joule effect.