44 PROCEEDINGS OP THE AMERICAN ACADEMY 



To prevent thermo-electric troubles I used continuous wires from 

 this key to the distant room where the astatic galvanometer was 

 placed. This was an instrument of tolerable, but not extraordinary, 

 sensitiveness. Its resistance was probably about five ohms, and 

 the time of a single vibration about six seconds. I have spoken 

 as if the electric circuit were closed at every stroke of the engine 

 by the key just described. In fact, the circuit was closed only 

 when this key and another under the hand of the observer at the 

 galvanometer were in operation simultaneously. Tbe method of 

 procedure did not require measurement of the galvanometer deflec- 

 tions. It merely required the observation of that temperature in 

 the paraffin pot which should make the galvanometer deflections 

 zero. The period of vibration of the needle was so much greater 



than the time of a stroke of the engine, that, by holding his key 

 down for several strokes in succession, the observer at the galva- 

 nometer could magnify any effect produced upon the instrument, 

 and therefore determine more sharply when the desired condition 

 of no effect was reached. 



The method of work was somewhat as follows. Everything being 

 in readiness, the observer at the galvanometer would satisfy himself 

 by trial that the closed circuit, with the iron-nickel thermopile 

 omitted, had no perceptible effect upon the galvanometer. He 

 would then signal for the thermopile to be brought into action, 

 and, watching the galvanometer, would determine whether the 

 paraffin was too hot or too cold for equilibrium, and signal the 

 other experimenters accordingly. If it was too cold, finely di- 

 vided paraffin was thrown into the pot, while some of the over- 

 heated liquid was drawn off through a cock at the bottom. The 

 following set of observations was made on April 17, 1801, with 

 the 2 mm. thermopile, contact occurring near the point of cut-off 

 (Figure 5, C). 



