ELECTRICAL PROPERTIES OE HYDROGENISED PALLADIUM. 183 



measurements, which were usually the same, and differed only in one case by 

 as much as 1 per cent., may be regarded as a fair indicator of the amount of 

 hydrogen present in the particular specimen. The same palladium wire was 

 used throughout as the basis for the successive hydrogeniums. 



The resistances were the same in both circuits, so that the galvanometer 

 deflections were proportional to the electromotive forces. The circuits could 

 be thrown on to the galvanometer in either direction and in rapid alternation. 

 The sequence in which the readings were taken was as follows : — Let a, b, sym- 

 bolise the readings of the two circuits, and let -I- or — be prefixed to indicate 

 the direction in which the current flowed through the galvanometer. Then the 

 order in which the observations were made was 



+a— a— 6+6— b—a'+a—a—b+b— 6&c. 



From + a to the next + a represents a complete series, from which two cor- 

 responding readings may be obtained ; and similarly from + b to the next + b. 

 An example, taken at random from the pages of the experimental book, will 

 make the method clear : — 



+a 



161 



(308) 

 (325) 



— a 

 147 



155 



-6 



164-5 

 164-5 



+ 6 

 (344-5) 



180 



(344-5) 



171 



(324-5) 



153-5 



170 



(362 ) 



192 



The zero line of the scale lay at the centre, and to this zero the zero position 

 of the galvanometer spot of light was roughly adjusted. Hence the sum 

 161 + 147 ( = 308) is to a first approximation double the deflection due to the 

 average current during the time taken to make the two readings; and similarly 

 for all the other successive pairs of readings belonging to either circuit. The 

 bracketed numbers show these sums, no two of which are of course simul- 

 taneous or correspond exactly to the same temperature. But let us take any 

 four in chronological order, such that the first and last belong to one circuit! 

 and the two intermediate ones to the other. Then we may assume, if the rate 

 of change is slow, that the sums of these pairs are proportional to electromotive 

 forces which correspond to the same temperature. In the example given, two 

 pairs of corresponding numbers may be obtained by simply adding successive 

 pairs of bracketed numbers of the same name, namely, 634, 689, and 650-5, 

 706-5. Every number so finally obtained depends on at least three readings. 

 I shall give here a portion of one of the tables containing these reduced num- 



