[Illustration: FIG. 55.--SUCCESSIVE MODIFICATIONS OF THE BLOCK METHOD FROM THE 'STRAIGHT WIRE' (_a_) TO 'CELL FORM' (_e_) When A is excited, current of response _in the wire_ is from less excited B to more excited A. Note that though the current of response is constant in direction, the galvanometer deflection in (_d_) will be opposite to that in (_b_).] It will be seen that the two limbs of the tube filled with water serve the purpose of the strip of moistened cloth used in the last experiment to make electric connections with the leading-out electrodes--with the advantage that we have here no chance of any shifting of contact or variation of surface, the contact between the wire and the surrounding liquid being perfect and invariable. On now vibrating the end A of the tin wire by means of the ebonite clip holder, a current will be found to flow from B to A through the wire--that is to say, towards the excited--and from A to B in the galvanometer. The next modification (_c_) is to transfer the galvanometer from the electrolytic to the metallic part of the circuit, that is to say, it is interposed in a gap made by cutting the wire A B, the upper part of the circuit being directly connected by the electrolyte. Vibration of A will now give rise to a current of response which flows in the metallic part of the circuit with the interposed galvanometer from B to A. We see that though the direction of the current in this is the same as in the last case, yet the galvanometer deflection is now reversed, for the evident reason that we have it interposed in the metallic and not in the electrolytic part of the circuit. The next arrangement (_d_) consists simply of the preceding placed upside down. Here A and B are held parallel to each other in an electrolytic bath (water). Mechanical vibration may now be applied to A without affecting B, and _vice versa_. The actual apparatus, of which this is a diagrammatic representation, is seen in (_e_). Two pieces, from the same specimen of wire, are clamped separately at their lower ends by means of ebonite screws, in an L-shaped piece of ebonite. The wires are fixed at their upper ends to two electrodes--leading to the galvanometer--and kept moderately and uniformly stretched by spiral springs. The handle, by which a torsional vibration is imparted to the wire, may be slipped over either electrode. The amplitude of vibration is measured by means of a graduated circle. It will be seen from these arrangements: (1) That the cell depicted in (_e_) is essentially the same as that in (_a_). (2) That the wires in the cell being immersed to a definite depth in the electrolyte there is always a perfect and invariable contact between the wire and the electrolyte. The difficulty as regards variation of contact is thus eliminated.