record friction of the writer against the glass surface becomes a source of error. This difficulty I have been able to overcome by the two independent devices, the Resonant Recorder and the Oscillating Recorder. In the former the writer is maintained by electric means in a state of continuous to and fro vibration, about ten times in a second. There is thus no continuous contact between the writer and the smoked glass surface, friction being thereby practically eliminated. The writer in this case taps a record, the successive dots occurring at intervals of 0.1 second. The responsive fall of the leaf is rapid, hence the successive dots in this part of the record are widely spaced; but the erection of the leaf during recovery takes place slowly, hence the recovery part of the curve appears continuous on account of the superposition of the successive dots. The advantage of the Resonant Recorder is that the curve exhibits both response and recovery. This apparatus is admirably suited for experiments which last for a few hours. There is, however, some drawback to its use in experiments which are continued for days together. This will be understood when we remember that for the maintenance of 10 vibrations of the writer in a second, 10 electric contacts have to be made; in other words, 36,000 intermittent electric currents have to be kept up per hour. This necessitates the employment of an electric accumulator having a very large capacity.
In the Oscillating Recorder the recording plate itself moves to and fro, making intermittent contact with the writer about once in a minute. The recording smoked glass plate is allowed to fall at a definite rate by the unwinding of a clock wheel. By an electromagnetic arrangement the holder of the smoked glass plate is made to oscillate to and fro, causing periodic contact with the writer.
