The following was the arrangement adopted in these experiments :—A large magnet was put into a copper pot containing oil, which was heated up by a brass Bunsen, and its temperature determined by a mercurial thermometer immersed in it. The magnet was set magnetically east and west, and placed so as to act with equal force on the poles of a small magnet, whose centre was in the prolongation of its axis. This small magnet was cemented to the back of a small concave mirror, suspended by a single silk fibre, and placed in a glass case to guard it against currents of air. The deflections of the small magnet were measured exactly as in the reflecting galvanometer, and since from the nature of the arrangement, the absolute magnetism in the large magnet is directly as the tangent of the angle of deflection of the small one, its amount for any temperature was immediately measured by the reading on the scale.