negative, is sometimes modified by the molecular condition of the wire (see below). In the modified form of apparatus the wires in the cell are immersed to a definite depth in the electrolyte; there is thus a perfect and invariable contact between the wire and the electrolyte. The wire in the cell is clamped below, and torsional vibration gives rise to a strong electrical response. If the wire be now carefully unclamped, and the vibration repeated as before, there will now be found no electrical effect. As all the rest of the circuit was kept absolutely the same in the two different sets of experiments, these results conclusively prove that the responsive electromotive variation is solely due to the mechanical stimulation of the acted wire. The excitatory effect due to the disturbance persists for a time. This may be shown by keeping the galvanometer circuit open during the application of vibration, and completing it at various short intervals after the cessation, when a persisting electrical effect diminishing rapidly with time will be observed. When the wire is brought to the normal condition, successive responses to uniform stimuli are, in the case of metals which, like tin, show no fatigue, exactly the same. I usually interpose a high external resistance, varying from 1 to 5 megohms, so that the galvanometer deflections may be proportional to the electromotive variations;
