INSTRUMENTS AND iMETHODS. 



303 



During blizzards fine driven snow entered this can and settled over the insulator and 

 destroyed the insulation. There is no means of getting over this diffi- 

 culty, for the drift during blizzards enters every small crevice and no 

 matter how the protecting device is arranged it soon becomes full of 

 snow which connects the insulated parts to earth. 



This, however, was not a serious difficulty for one is chiefly concerned 

 with the potential gradient only during fine weather, so the loss of part 

 of the record during blizzards did not affect the main results of the 

 work. Also during blizzards the potential was far above the limits of 

 the recording electrometer and so could not have been measured if the 

 insulator had remained perfect. The chief practical effect was the 

 necessity to clean the insulator at the end of each blizzard and to put 

 it in good order for the subsequent fine weather. 



The Benndoif self-recording electrometer had of necessity to be within 

 the living hut, the temperature of which was kept, except occasionally, 

 well above the freezing point. A very liberal share of the hut was 

 apportioned to the physical laboratory but unfortunately it was in the 

 coldest corner of the hut. In the h\it twenty-five men lived in one large 

 room in which all the cooking was done. In consequence the hiimidity 

 was always high and water vapour literally distilled into the physical 

 laboratory. This vapour condensed on to the walls which were generally 

 covered with a coating of ice or water. It was through this wall that 

 the wire to the potential gradient instrument had to pass, and the 

 method of doing this was the first difficulty which had to be solved. 



In ordinary observatories, the connecting wire is usually passed 

 through a hole in the wall sufficiently large for the wire not to touch 

 it when slightly displaced by the wind. Such a hole was quite impos- 

 sible in our hut, for it would have let in cold air, but much more 

 serious it would have been the entrance for drift snow during the 

 blizzards. The wall of the hut consisted of two thicknesses of wood, 

 then an air space of three inches, and then two more thicfaiesses of 

 wood. The first attempt made was to pass a metal tube through the wall, plug it up 

 at each end with sulphur and pass a thin wire through the centre. The sulphm' then 

 not only closed the tube but acted as supporting insulators for the wire. This method 

 soon proved unusable, for the water vapour which was deposited on the well-insulated 

 wooden wall was deposited still more rapidly both inside and outside the metal tube. The 

 water was in itself sufficient to destroy all insulation, but on account of the metal tube 

 projecting into the cold air outside the hut, the end within the hut was so cold that the 

 water froze and a large mass of ice crystals formed upon it which were pretty to look at 

 but totally destructive to the insulating properties of the sulphur. Some method had to be 

 found which would not only protect the insulator from the water deposited all over the wall 

 but would also not conduct cold from tho outside, and so cause an intensified deposition on 

 the insulator. 



There is no need to go into all the experiments made ; but the final solution of the 

 problem will be given. As is usually the case, the solution is so simple that one wonders why 

 it was not the first thought of. Figure 90 shows the completed insulator. A and B are 

 the two separate parts of the wall through which the insulator was to convey the conductor. 

 C C is a brass rod about one-eighth inch thick. Around this rod a coating of sulphur D was cast 



Fig. 89. Insulator for 

 collector rod. 



