Thermo-electric Action of Metals in Electrolytes. 281 



Remarks. The thermo currents produced by solutions of potassic 

 carbonate and sodic phosphate were usually strong, and so also were 

 those with aluminium, tin, and iron, and yielded many balances ; 

 whilst those from solutions of oxalic, formic, and tartaric acids were 

 too feeble to balance the corresponding voltaic ones. In obtaining 

 balances with the noble metals, the voltaic currents were so very 

 small that the Thomson's reflecting galvanometer had to be employed ; 

 these cases are indicated by an asterisk. 



By means of a separate experiment it was ascertained that the 

 amount of error caused by the voltaic current producing polarisation 

 before the thermo current had time to balance it was small, even 

 in the cases where the former current was comparatively strong, as 

 with a solution of potassic cyanide. 



A number of thermo-electric reversals occurred during the making 

 of these determinations : thus, in a solution of potassic nitrate, gold 

 was thermo-electro-positive below and negative above 200 F., and 

 platinum behaved similarly below and above 180. Aluminium in 

 solution of potassic carbonate became less thermo-positive above 

 140 F. Lead in sodic chloride and magnesium in potassic hydrate 

 was thermo-positive below 170 and negative above it. In sodic 

 sulphate nickel was thermo-negative below 110 and positive above 

 that temperature. In dilute nitric acid lead was negative below and 

 positive above 180, and tin similarly below and above 150. In 

 dilute chloric acid iron was positive up to 150 and negative above it. 

 In potassic alum nickel was positive below and feebly negative above 

 164 F. In ammonic alum aluminium was negative below and 

 positive above 150, and nickel was positive below and negative 

 above 170 F. In magnesic sulphate cadmium behaved singularly; 

 whilst it was chemico-negative to zinc and positive to aluminium at 

 60 F., by heating it alone (the zinc being in the cold part of the 

 liquid) it became more negative, but with aluminium in the place of 

 zinc, the cadmium became more positive by being heated. This expe- 

 riment was repeated and similar results obtained ; it is probably a case 

 of the kind mentioned on pp. 279 and 288. 



On comparing these reversals with the positions of the same 

 metals in the same liquids in Table I, it will be observed that some of 

 them do not agree in thermo-electric sign in the two cases. These 

 apparent discrepancies may be partly accounted for by the circum- 

 stance that in the experiments of Table I, the metals were suddenly 

 immersed in the hot and cold portions of liquid, whilst in the present 

 ones they were immersed in the two cold portions, and one of them 

 then gradually heated ; they may also be partly explained by the 

 circumstance observed that the balance obtained during a rising of 

 temperature was frequently at a somewhat higher point than one 

 obtained whilst the temperature was falling (compare also p. 272 for 



