true through a limited range. From the results of various experiments, into the detail of which I can not at present enter, it appears that, generally speaking, the curve of response (with molecular effects as ordinates, and the intensities of disturbance as abscissæ) is not a straight line. It is at first slightly convex, then straight, and in the last part concave. It is only in the second part that the curve is approximately straight.
In considering the effect of electric radiation in varying the conductivity of the particles, we have to bear in mind that no explanation can be regarded as complete, unless it explains not only the diminution, but also the increase of resistance; also the phenomenon of automatic recovery and of the opposite effects which are exhibited by the same receiver under different molecular conditions. The increase of resistance of the Ag′ receiver and its instantaneous recovery are directly opposed to the theory of coherence.
The state of balance between the distortion produced by radiation and the force of restitution on the one hand, and the different equilibrium positions with different radiation intensities on the other, point to the effect being due to some strain produced by radiation.
Fatigue of the Receiver.—I wished to trace the gradual appearance of fatigue in the Ag′ receiver, and for this purpose kept it acted on with slight intermissions for nearly 3 hours. At the end of that time it began to show unmistakable signs of fatigue. Fig. 6 shows the
effect when the radiator was at a distance of 20 cm.; the deflections were now only twenty-one divisions, whereas before this the deflection was thirty-three divisions with the radiator at the increased distance of 25 cm. Formerly the recovery commenced immediately on the cessa-
