256 Dr. G. Gore. Relations of Heat to Voltaic and 



With regard to the influence of composition of the liquid, the salts 

 of simple chemical character, such as potassic cyanide, fluoride, 

 bromide, and nitrate, gave the largest proportion of thermo-electro- 

 positive couples ; the more complex ones, such as the double sulphates, 

 gave the largest number of negative ones. I have not examined 

 whether this is related to differences of specific heat of the salts. Even 

 the highly thermo-positive metal nickel was rendered thermo-negative 

 in solutions of potassic hydrate and chloric acid, and similarly with 

 the thermo-positive metals platinum and palladium, each in eleven 

 different liquids, and with aluminium in solution of potassic hydrate, 

 sodic sulphate, and ammonia alum. The usually thermo-negative 

 metal iron was more frequently positive than any other metal, it was 

 strongly so in solutions of potassic bromide, hydrochloric acid, caustic 

 potash, and sodic sulphate, and so also was the thermo-negative metal 

 cadmium in solution of caustic potash. 



The liquids in which the hot metal was thermo-electro-positive in 

 the largest proportion of cases were those containing highly electro- 

 positive bases, such as the alkali metals ; thus, the highly chemico- 

 electro- and thermo-electro-positive body potassium, forms the highly 

 thermo-electro-negative compounds potassic cyanide, fluoride, bromide, 

 nitrate, carbonate, and iodide. From a much smaller number of facts, 

 I had previously drawn the conclusion that, provided ordinary 

 chemical action is absent, liquids of alkaline reaction are usually 

 thermo-electro-negative, and acid ones, thermo-electro-positive. This 

 is largely confirmed in the present table in the cases of the least 

 corrodible metals, gold and platinum. 



In every experiment of this kind, the two pieces of metal were 

 simultaneously immersed in the hot and cold portions of liquid respec- 

 tively, and no cases of reversal of direction of current therefore 

 occurred ; many would have happened had the metals been slowly 

 heated in the liquids. In a subsequent series of similar experiments, 

 in which the one limb of the vessel was gradually heated after immer- 

 sion of the electrodes, a number of such instances were observed (see 

 p. 281). Bleekrode also (" Poggendorff's Annalen," 1870, vol. cxxxviii, 

 p. 571 ; " Annales de Chimie," April, 1870 ; also " Philosophical 

 Magazine," 1870, vol. xl, p. 311) has recorded a few of such cases. 



In order to obtain a thermo-electric series of all the strongest 

 thermo-positive and negative couples in Table I, the electric potentials 

 of most of the top and bottom couples were measured. The remaining 

 members were omitted because their currents were very variable in 

 potential, probably in consequence of polarisation, and of alterations 

 of rapidity of diffusion of the layers of liquid in contact with the 

 electrodes. The electric potential usually declines quickly (but 

 occasionally only a little), hence the numbers obtained depend upon 

 the degree of rapidity with which the determinations are made. The 



