Ruhmkorff's coil. In (a) the radiator was kept at a distance of 40 cm.; the radiation was continued for 15 seconds, after which 15 seconds was allowed for recovery. A longer time would have allowed a more complete recovery, but this would have entailed a great loss of time in the long series of experiments contemplated. The recovery is therefore partial, the return curves not exactly reaching the original starting position. The record of response to the source of radiation at a distance of 40 cm. shows that there is a partial recovery between quickly recurring flashes, observed in the fluctuating response about the balanced position. Now when the intensity of radiation is increased by decreasing the distance of the radiator to 25 cm. (see b), the strain effect persists for a longer time, and the flashes arrive before the substance can have recovered to any extent; there is thus less fluctuation in the balanced position. The intensity of radiation was further increased by placing the radiator at a distance of 15 cm. (see c) the fluctuations now disappeared; and the galvanometer deflection was held rigid as long as the radiation was kept on; in fact, we have here an effect which physiologists describe as "tetanic." On the cessation of radiation there was an immediate recovery. It will be noticed how extraordinarily consistent are the succeeding values of response. The resultant effect being due to the additive effects of numerous flashes, an occasional failure of an individual flash is of little or no importance. The series of responses in (d) was taken after half an hour; it will be noticed how very consistent they are among themselves and how similar to those in (c), showing that even after half an hour's con-
