Page:Collected Physical Papers.djvu/190

170 rate is 1 inch in 30 seconds, and one small division of the paper measured horizontally is thus equal to 3 seconds. But very much quicker or slower rates may easily be obtained by means of the different-sized pulleys. The spot of light from the galvanometer is thrown down on the paper by an inclined mirror. The movement of the galvanometer spot takes place at right angles to the direction of motion of the paper. There is a guide rod at right angles to the motion of paper, along which the recording pencil is moved. The excursion of the galvanometer spot can thus be easily followed with a pencil, and it is quite easy to do this, when the fluctuation period is about 2 seconds. In the experiments to be described, this period varied from 2 seconds to several minutes. A curve is thus directly obtained, with conductivity or electromotive variation as ordinate, and the time as abscissa. The curves given in the accompanying papers are exact copies of the direct records.

When a substance is strained by radiation there is produced a sudden variation of conductivity. The substance automatically recovers from the strained condition (1) if it has not been overstrained by an excessive stimulus, or (2) if its electric elasticity is very great. I have found, in general, that after careful adjustment of the receiver it exhibits tendency towards self-recovery, if the intensity of incident radiation is not too strong. In the case of substances which are, electrically speaking, highly elastic, such as K, there is an automatic recovery even after strong stimulation. The difference