180 DR CARGILL G. KNOTT OX THE 



palladium, we may indicate the peculiarities of the hydrogenium by means of 

 the following table of thermoelectric powers (t in centigrade degrees) : — 



Iron, 1734- 4-87^ 



Copper, 136+ -95* 



Palladium, - 625 - 3*59* 



Hydrogenium ( 0° - 200°), + 1128 - 3-59* 



(200° -300°), - 625-3-59* 



ii (300° -200°), +1578-2-55* 



(200°- 0"), + 160-3-59* 



It must be understood, of course, that these equations do not strictly hold 

 at these temperatures which separate the one group from the other ; at these 

 points there are continuous although rapid transitions which baffle an arithmetic 

 representation. 



The peculiarities here indicated may be explained as due to the escape of 

 the hydrogen during the heating and to its partial return during the cooling. 

 Only the one extremity of the hydrogen- charged wire is heated, so that only 

 from that portion will the hydrogen escape to any marked extent. Whether it 

 escapes wholly out of the wire, or is partly driven into the contiguous colder 

 portions, is not certain. The latter possibility is far from improbable, if the wire 

 is undercharged to begin with. Whatever may be the case, however, it is 

 obvious that at about 200° C. the portion of wire immersed in the hot oil begins 

 to lose its occluded hydrogen. The thermoelectric system that now exists is of 

 such a complicated nature that it would be difficult, perhaps impossible, to 

 predict what should happen. There is a pure palladium wire joined in circuit 

 with a hydrogenised palladium wire, whose charge as well as temperature 

 varies continuously from the one extremity to the other. Supposing, as I shall 

 establish later, that the thermoelectric position of hydrogenium is a function 

 of the charge, we have to do with a chain of elements of continuously varying 

 thermoelectric power. Adding to this the further complication that there is a 

 time variation of both temperature and charge, we can scarcely expect to be 

 able to prejudge the phenomenon. The mere fact of a variation of charge may 

 — for ought we know to the contrary — bring into play an electromotive force 

 of other than purely thermal origin. As the heating changes to cooling, the 

 electromotive force which had remained so constant since 200° begins to fall 

 away rapidly, but with diminishing rapidity as the temperature falls. There 

 certainly seems to be an instability in the condition of the hydrogen during this 

 cooling process, for just as the temperature below which hydrogenium was 

 stable during the heating the thermoelectric properties recover their original 

 ordinary characteristics. The final position of the hydrogenium line, after the 

 cooling, shows that there has been a loss of hydrogen at the junction ; but, as 

 was proved by resistance measurement afterwards, a very small amount of 

 hydrogen was lost to the wire as a whole — not more than 5 or 6 per cent. 



