McKAY. — HALL EFFECT AND CURRENT DENSITY IN GOLD. 363 



The ammeters were also compared directly with copper voltameters 

 for positions on the scale in the neighborhood of 1.2 amperes, this por- 

 tion of the scale being very frequently used. In preparing the vol- 

 tameters, the directions followed were in general those found desirable 

 by Gray, as given in Holman's " Laboratory Notes for Use in the 

 IVIassachusetts Institute of Technology." In accordance with the advice 

 of Ostwald and Luther, a stream of carbon dioxide was kept passing 

 through the solutions during the run. The weights used in measuring 

 the gain of copper were compared with each other and the value of one 

 of them was kindly determined by Professor T. "W. Eichards in terms 

 of the "Washington standard. A table of corrections was made out, 

 these corrections not amounting to more than four hundredths of a 

 milligram, except in the case of the five and ten gram weights, for 

 which the corrections were 0.1 and 0.2 milligrams respectively. 



The following results of direct comparison were obtained in the 

 neighborhood of 1.20 amperes. 



For the Siemens and Halske instrument this was about three-fourths 

 of the full scale reading, for the Weston about one-eighth of the full 

 scale reading. 



Primary currents below 0.010 ampere were measured by a poten- 

 tiometer method. (See Figure 4.) The same Carhart-Clark cell and 

 German silver potentiometer wire were used as in the determination of 

 the sensitiveness of the astatic galvanometer. 



The ammeters were also calibrated in the neighborhood of 0.01 am- 

 pere to 0.02 ampere by the potentiometer method, as they were some- 

 times used for measuring primary currents of about this strength. For 

 this calibration they were placed in series in the position corresponding 

 to that of the gold plate in the figure above. 



