Page:Collected Physical Papers.djvu/362

342 of the galvanometer index on a moving piece of paper attached to the plate M.

The successive closure and opening of the keys are made automatically and in proper sequence by means of a clock work, the whole process being repeated at intervals of 15 minutes.

The most difficult problem is the automatic record of galvanometric deflections. This can be secured without any difficulty by means of photography. A spot of light reflected from the galvanometer mirror is in this case allowed to fall on a photographic plate which descends at an uniform rate by clockwork. This, however, entails the use of a dark room and subsequent development of the plate. The trouble was avoided by the device of direct record of the galvanometer deflection by means of electric sparks.

The sparking method has been previously employed in which the deflected index of the galvanometer in connection with one electrode of an induction coil leaves a spark record on a moving piece of paper. Several difficulties are, however, encountered in employing this method for record with a highly sensitive galvanometer. There is a liability of leakage of the high tension current into the galvanometer circuit. The discharge of the spark gives moreover a backward kick to the index by which the normal deflection undergoes an unknown variation.

The above difficulties have been removed in the following manner. The moving coil of the sensitive D'Arsonval galvanometer, has a long glass index I, at right angles to the plane of the coil. The glass index is