182 DR CARGILL G. KNOTT ON THE 



value 25. The curves marked B in Plate XI. bring out the peculiarities to the 

 eye. 



If we try to trace the hydrogenium line on the thermoelectric diagram from 

 this curve, we shall get rather a curious result, which is shown roughly in the 

 small diagram (C) in the upper left-hand corner of Plate XI. A glance at the 

 electromotive force curve indicates, indeed, that the ratio of the thermoelectric 

 powers of hydrogenium and palladium referred to platinum begins with a 

 negative value greater than unity, passes through zero at about 150° C, becomes 

 equal to positive unity at about 200° C, and attains a value greater than unity 

 as the temperature passes through 300° C. This may be explained in two 

 ways. It may be due to an electromotive force other than thermal, brought 

 into existence by the variation of charge or a possible convection of hydrogen 

 along the wire. Or, it may be due to the integral electromotive force of a row 

 of hydrogenium elements of varying charge and temperature being a function 

 of the space-rate of either variation. Ultimately the ratio of the thermoelectric 

 powers becomes a little less than unity, and continues so until the temperature 

 reaches its highest value. The cooling curve belongs to a diagram line which 

 lies between palladium and platinum, dividing the space in the ratio of 

 1 to 4. The main features are identical with those of the first experiment, 

 the differences in detail being the result of the much higher temperatures 

 employed, in virtue of which more hydrogen is lost to the wire. In the 

 small thermoelectric diagram the palladium line is drawn parallel to the 

 platinum line, although they are really inclined at an angle which would 

 make them meet at a point corresponding to - 600° C. The irregularities 

 due to the instability of the hydrogen appear in the thermoelectric ex- 

 periments at a lower temperature than in the resistance experiments. This 

 may well be due to the inequality of temperature distribution along the wire 

 in the former case. 



It now remained to make a study of the thermoelectric properties of 

 hydrogeniums of intermediate charges. The object aimed at was to obtain 

 careful determinations at temperatures below those at which arise the irregu- 

 larities due to the instability of the hydrogen. I confined myself, therefore, to 

 temperatures below 100° C. The arrangement and reduction of results, which 

 present some novelties of operation, were as follows : — 



A triple junction of platinum, pure palladium, and hydrogenium was im- 

 mersed in oil, along with a centigrade thermometer, which, however, was used 

 rather as an indicator than an accurate measurer. The platinum-palladium 

 circuit was indeed the real thermometer ; and in terms of its electromotive 

 force the electromotive force of the platinum-hydrogenium circuit was obtained. 

 The resistance of each specimen of hydrogenium was measured before and 

 after the electromotive force measurements. The mean of these two resistance 



