74 KOYAL SOCIETY OF CANADA 



of the battery through one megohm, and this deflection compared with 

 that obtained with the leakage currents. It was found that the in- 

 sulation was better than 100,000 ohms for the entire length of cable, 

 which was considered quite satisfactory and it was probably greater than 

 this during dry weather. 



Tilie resistance of the main leads, and compensating leads, was 

 measured on the Wheatstone bridge, and was found, to be about 40 

 ohms. A slight difference between the two pairs of leads was observ- 

 ed, anjounting to about half an oKm. To rectify this and in order 

 to bring the zero of the instrument as nearly as possible to the middle 

 of the scale, a small coil of copper wire was wound and placed in series 

 with the leads of lowest resistance. This coil was placed outside, in 

 the therm'Ometer box at the Observatory, for it was considered important 

 that this coil should undergo a.pproxim'a/tely the same variations of 

 temperature as the rest of the line. î^ew ratio coils were added to the 

 recorder, of .50 oihms each, and adjusted very carefully by means of the 

 Kelvin- Varley slide. It was found that the instrument was more 

 sensitive with these coils. OTiere was no record, of the resistances of 

 ■the ratio coils that we removed, but it is supposed they were about 25 

 ohms. 



A most important test of the accuracy of the instrument was the 

 steadiness of the zero., which gave a severe test of the compensation of 

 the line. This zero was tested in several ways. 



In the first test, the compensating leads and thermometer leads 

 were connected together, with the thermometers removed. The record- 

 er was set to trace the variations in the leads. Several traces were 

 obtained in this way and show small variations of about 1/2 a degree 

 either way from the zero, indicating that possible strains on the wire, 

 due to wind or other sources of disturbance, produced slight changes in 

 the resistance, not immediately compensated. iTihe average line traced, 

 however, over a long period, was very close to the zero mark. • 



In the second test, the two thermometers were added to the circuit 

 and placed, side by side in the Observatory, near the recorder. A simi- 

 lar line was traced as before. 



In a third test, the thermometers were put in their proper places, 

 one on the mountain and the other in the thermometetr box at the ob- 

 servatory, and the bulbs were immersed in a mixture of ice and water. 

 The zero line traced in this way was of the same character as the zero 

 traced in the other two tests, and was taken to be the true zero of the 

 instrument. In all, these tests extended over about two weeks and 

 were satisfactory. 



