the Carbon Cell. 185 



the alkali, shows some inclination to deviate towards the 

 hydrogen value. 



Combinations consisting of such electrically coated iron 

 wires, acting as a hydrogen electrode, and platinum, function- 

 ing as the air electrode, were finally found to be the best 

 arrangement for measuring the temperature coefficient of 

 the cell. 



The following figures were thus arrived at (values plotted 

 on diagram as curve with crosses). 



Temp. . 



. 348 D 



369° 



382° 



410° 



420° 



432° 



458° 



E.M.F. . 



1-201 



1-193 



1-191 



1-176 



1-174 



1-168 



1-159 



Temp. . 



. 482^ 



510° 



542° 



575° 



600° 



618° 





E.M.F. . 



1-144 



1-132 



1-118 



1-103 



1-096 



l-09o 





The results are in agreement with those which Haber and 

 Bruner obtained by their different methods of working up to 

 about 400° ; above this the values are higher than those of 

 Haber and Bruner, and are in closer proximity with theory. 



In connexion with this, it was thought that perhaps 

 addition of sodium oxide (Na 2 0) could render the molten 

 alkali free from water, and therefore change the E.M.F. of 

 the cell, which, as a consequence of theory, would thereby be 

 increased. This plan, however, did not bring about the 

 desired result. Adding a little oxide did not cause a marked 

 change ; on the other hand, if all the associated water was 

 eliminated, the iron, even coated electrolytically, quickly 

 assumed the passive state. 



The influence of associated water is more marked with 

 potassium hydroxide. Caustic potash freshly melted, thereby 

 containing relatively much water, <jives rise to values which 

 are much smaller than those obtained from caustic potash 

 which has been kept molten for a long time and is therefore 

 poorer in water. 



This influence of water is easily misleading. The presence 

 of water always causes iron to remain longer active and the 

 hydrogenized palladium to retain its charge. Thus it is 

 easier to work with molten alkali containing water, but the 

 results may be several centivolts too low, the equilibrium 

 pressure of the water-vapour above the molten alkali being 

 in excess of the atmospheric vapour-pressure to some extent. 



Caustic potash, moreover, from which water has been 

 eliminated by prolonge I heating, causes iron to become 

 mue.h more rapidly passive than does caustic soda. 



The following values, in the case of cau>tic potash, were, 

 Phil. Mag. S. 6. Vol. 13. No. 73. Jan. 1907. O 



