HALL. — THERMAL AND ELECTRICAL EFFECTS IN SOFT IRON. 47 



disturbance disappeared when this commutator was opened. The storage 

 battery furnishing the main current was in the basement, and the gal- 

 vanometer rested on a brick pier which rose from this basement. The 

 disturbance was apparently not due to the storage battery, as its direction 

 was independent of the direction of the current entering the Thomson 

 effect bars from this battery. It probably was due to battery action 

 between water and metals at the cold end of the apparatus, causing a 

 slight current to creep along the varnished surface of the board B (Figure 

 6), this surface being slightly dampened by precipitation of water from 

 the air, thence along a metal path through the galvanometer, thence by 

 way of the pier surface, or some other damp non-metallic path, to the 

 basement, thence by way of the main current leads to the Thomson ap- 

 paratus again. When connection was made with the wires leading from 

 the hot end of the board B, no disturbing effect due to this connection 

 could be made out with certainty. There was, of course, no need of 

 having any metallic connection between the Thomson effect apparatus 

 and the galvanometer circuit during the measurement of the reduction 

 factor; but for convenience the one-point connection described above was 

 frequently left during the early part of the work, before the possibility 

 of trouble from it was thoroughly realized. Fortunately this connection 

 was made, as a rule, with the hot end of the Thomson effect apparatus, so 

 that when the danger was discovered it was necessary to reject on this 

 score the work of one or two days only. 



Moisture at the ends of the board B was, indeed, a frequent and 

 serious cause of trouble, especially at first. Sometimes it came from 

 leakage past the rubber gasket which separated each vessel from its 

 cover or top. To avoid this leakage the level of the water in each 

 vessel was, after a time, kept below the level of the gasket, and care was 

 taken not to let the boiling become so violent as to clog the condenser 

 return pipe and raise the pressure of steam in the boiler. At the cold 

 end, especially in warm weather, profuse precipitation of water from the 

 air was capable of drenching the cotton lying on the board. To pre- 

 vent this, the top of the cold vessel was swathed in cotton to a depth 

 of about 2 cm., and outside this cotton was placed a covering of thin 

 sheet rubber. 



Much trouble would have been saved by placing the board B consider- 

 ably higher, and by having the end posts upon it placed nearer the 

 middle ; but it is unlikely that any serious error due to the presence of 

 water on this board survived the daily reversals of current and the long 

 course of observation to which the apparatus was subjected. 



