supported by the circle; the circle is provided with a pivot which passes through the alidade, and has, at its upper extremity, two hooks to receive the pin fixed to the thread. With this arrangement of the stirrup, it is important that the alidade in which the magnet bar lies should rest on the rim of the circle; otherwise, the friction, taking place near the axis of rotation, would produce a displacement of the parts relatively to each other, in consequence of the impulse arising from the vibrating bar. Moreover, the stirrup is so constructed that the magnet bar fits in either on its broad or narrow side. This is done for the purpose of determining accurately, by observations of declination in any of the various positions of the magnet bar in the stirrup, the position of the mirror relatively to the magnetic axis of the bar.
10. The box and the measuring scales.—The box which protects the magnetometer from the influence of currents of air is constructed so as to afford ready access to the instrument within. It forms a cylinder of 800 millimetres in diameter, and 300 in height. The cylindrical form is given to it for this reason; in the measurement of intensity, in order to ascertain the moment of inertia, a wooden rod 700 millimetres in length is placed at right angles on the magnet bar of 600 millimetres in length, and this rod, to which weights are suspended, must find a place in the box along with the magnet bar, and must vibrate freely. In order to perform these experiments with convenience, it is also requisite that the box should admit of being entirely opened at the top, and of being tightly closed again, so that there should only remain an aperture at the top for the suspension thread, and one for the mirror at the side. The latter may be closed with a small wooden slider, to exclude air when not observing. The box is closed above by two semicircular lids, which must fit exactly, one of which is provided with a small aperture for the thread. This aperture is not situated in the centre of the circle formed by the two semicircular lids, but is so placed that the thread passing freely through it, the mirror of the magnet bar may hang close before the aperture in the side of the box. This arrangement is necessary, in order that a small aperture may suffice to allow the light to pass from the scale to the mirror, and from that to the telescope. Around the case are fixed the measuring scales on which may be placed a second magnet bar to the south or north, to the east or west of the magnetometer, at prescribed distances
