330 PROCEEDINGS OF THE AMERICAN ACADEMY. 



close of one of the regular experiments as described below, the charging 

 current was allowed to continue for a number of hours (twelve in one 

 case, thirty in another) after the wire had apparently reached a state of 

 saturation. At the end of this time, the palladium being unquestionably 

 saturated, the volume of hydrogen liberated per minute was carefully 

 measured and compared with the average ammeter reading. Two 

 independent trials of this sort gave 6.973 and 6.975 cu. cm. as the volume 

 (reduced to 0°C. and 760 mm. pressure) set free per minute by one 

 ampere of current — as measured by the ammeter employed in this work. 

 In the final experiment the method of procedure was as follows. A 

 length of palladium wire, freshly cut from the original coil, was cleaned 

 by washing in caustic potash and then in distilled water. It was annealed 

 by bringing it to a white heat in the open air and allowing it to cool 

 quickly, the heating being accomplished by passing a current of electric- 

 ity through it. This process left no visible tarnish, though the surface 

 of the wire appeared less lustrous than before heating.* After the 

 annealing the palladium was mounted in the voltameter, as already ex- 

 plained, and the latter was placed in its water bath, and this was stirred 

 continuously. This bath was, approximately, at the room temperature, 

 and by keeping the latter fairly constant the temperature of the water 

 bath did not change during the progress of the experiment by more than 

 1°C. The resistance of the wire was then carefully measured. In this 

 work, the commutator, M, in Fig. 7, was set so that the circuit of the 

 charging battery, Z>j, was broken, while the current from another battery, 

 B 2 , was allowed to flow through the palladium wire and the resistances 

 i? x , 7? 2 , i? 3 , i? 4 , etc. The wheel rheostat was included between _R„ and 

 i? 4 , and on this rheostat were found balancing points (see Fig. 6) for the 

 two ends of the palladium wire, i. e., points so situated that when one of 

 them was connected, through a sensitive galvanometer, with the appro- 

 priate end of the palladium wire, no deflection was observed. To guard 

 against any errors from thermo-electric effects, these balancing points 

 were located four times in each measurement of resistance, the commu- 

 tator A" (in Fig. 7) being used to reverse the current through all the 

 resistances, while, for each position of A^ reversal of the commutator 

 M, by turning it through 90°, had the effect of turning the palladium 

 wire end for end with respect to the remainder of the circuit. 



* In a preliminary experiment a piece of palladium cut from the same coil had 

 been annealed in an atmosphere of nitrogen, with the result that the wire retained 

 its high polish, but refused, when exposed to the nascent hydrogen, to absorb the 

 latter in any considerable quantity. 



