#Considerations showing that electric response is due to molecular disturbance.#--The electromotive variation varies with the substance. With superposition of stimuli, a relatively high value is obtained in tin, amounting sometimes to nearly half a volt, whereas in silver the electromotive variation is only about .01 of this value. The intensity of the response, however, does not depend on the chemical activity of the substance, for the electromotive variation in the relatively chemically inactive tin is greater than that of zinc. Again, the sign of response, positive or negative, is sometimes modified by the molecular condition of the wire (see Chap. XII). As regards the electrolyte, dilute NaCl solution, dilute solution of bichromate of potash &c. are normal in their action, that is to say, the electric response in such electrolytes is practically the same as with water. Ordinarily I use tap-water as the electrolyte. Zinc wires in ZnSO_4 solution give responses similar in character to those given by, for example, Pt or Sn in water. #Test experiment.#--It may be urged that the E.M. effect is due in some way (1) to the friction of the vibrating wire against the liquid; or (2) to some unknown surface action, at the point in the wire of the contact of liquid and air surfaces. This second objection has already been completely met in experimental modification, fig. 55, _b_, where the wire was shown to give response when kept completely immersed in water, variation of surface being thus entirely eliminated. Both these questions may, however, be subjected to a definite and final test. When the wire to be acted on is clamped below, and vibration is imparted to it, a strong molecular disturbance is produced. If now it be carefully released from the clamp, and the wire rotated backwards and forwards, there could be little molecular disturbance, but the liquid friction and surface variation, if any, would remain. The effect of any slight disturbance outstanding owing to shaking of the wire would be relatively very small. We can thus determine the effect of liquid friction and surface action by repeating an experiment with and without clamping. In a tin wire cell, with interposed external resistance equal to one million ohms, the wire A was subjected to a series of vibrations through 180 deg., and a deflection of 210 divisions was obtained. A corresponding negative deflection resulted on vibrating the wire B. Now A was released from the clamp, so that it could be rotated backwards and forwards in the water by means of the handle. On vibrating the wire A no measurable deflection was produced, thus showing that neither water friction nor surface variation had anything to do with the electric action. The vibration of the still clamped B gave rise to the normal strong deflection.