now given near A, a transitory electrical current in response to the disturbance will be found to flow round the circuit, in a definite direction. Disturbance of B gives rise to a current in the reverse direction. For quantitative measurement it is necessary to have the intensity of stimulus maintained uniform, or increased or decreased in a definite manner. Instead of a tap, the stimulus of torsional vibration is more satisfactory. By maintaining the amplitude of vibration constant, or increasing or decreasing the amplitude, we can either keep the stimulus constant, or increase or decrease it in a quantitative manner. I shall first describe some of the typical results which may be obtained with the simple "straight wire form" of the apparatus. Worked with care it gives consistent and good results. For quantitative measurements requiring the greatest exactitude the "cell form," to be presently described, will be found more satisfactory.
Recording Apparatus.—The galvanometer used is a sensitive dead-beat D'Arsonval. The records are taken by means of a cylindrical modification of the response recorder described in a previous paper, or by means of photography. In the latter method, a clockwork moves the photographic plate at a uniform rate and a curve is traced on the plate by the moving galvanometer spot of light. The disturbance of molecular equilibrium caused by the stimulus is attended by an electromotive variation, which gradually disappears on the restoration of molecular equilibrium. The rising portion of the response curve shows the electromotive effect, due to stimulus, and the falling portion the, recovery. The ordinate represents the electromotive variation, and the abscissa the time.
