But this initial difference can be annulled by appropriate means. One method is to stretch one of the two wires by which the initial difference may be neutralised. The particular wire which should be stretched is found by trial. When the cell is first made, there is some irregularity in the action owing to molecular instability. This can be made to disappear, if both the wires are vibrated for a time. After these precautions are taken, results are obtained which are extraordinarily consistent.
Now if one of the two wires be continuously twisted, an increasing P.D. is induced during torsion between the acted and the unacted. This may be measured by the deflection of a sensitive galvanometer. Curves could thus be obtained with electromotive force, measured by the galvanometer deflection, as ordinate, and the time during which disturbance is kept up as abscissa. Such curves were directly obtained by the recording apparatus described in my previous paper. The wire was twisted at a uniform rate. The successive dots represent the completion of 360°. To keep the deflection within the scale, a megohm was interposed in the circuit. The resistance of the cell was about 5000 ohms. The absolute values of electromotive force corresponding to the galvanometer deflections were subsequently obtained by noting the effect of a known electromotive force.
2. Effect of Torsional Disturbance.
Most of the metals—exceptions presently to be described—become negative during molecular disturbance caused by torsion, i.e., the current through the liquid is from the acted to the unacted wire. As there is a considerable vagueness in the terms positive and negative, which has led to much confusion, I would name the acted wire as becoming zincoid or Z, when under an external disturbance the current flows through the electrolyte from the acted to the unacted wire. Again, in certain cases the reverse is true; the current flows from the unacted to the acted wire; the acted wire will in that case be designated as cuproid or C.
In fig. 12 is shown the effect of twist on Zn.
The induced electromotive variation is not due to twist as such, but to molecular disturbance induced during increasing twist. For if the wire be held stationary in the twisted position, the molecular disturbance with the attendant electromotive variation will gradually disappear. (Fig. 12.) Other evidences will be brought forward to show that the effect is simply due to the molecular disturbance, and not to the twist.
The wires used in the following experiments were from commercial specimens. The length was in every case about 9 cm., but the. diameters were not the same.
