Thermo-electric Action of Metals in Electrolytes. 285 



dilute sulphuric acid at 60 F. were found to be greater than those 

 between the series with potassic bromide and chloride. 



Comparisons of the order of the corrosion series at 60 F., with the 

 thermo-electric ones of the same liquids of Table I (p. 253), and the 

 chemico-electric ones of the " cold " columns in Table X (p. 265) show 

 that the order of simple corrosion differs largely from that of chemico- 

 electric, and still more largely from that of thermo-electric action. 

 In no one pair of the twelve corresponding series of Tables X and 

 XVIII did the order of metals entirely agree, the least agreeing ones 

 were those in potassic cyanide and bromide. The degree of chetnico- 

 or thermo-electric potential therefore does not vary in the same 

 order as that of rapidity of simple corrosion in any of the liquids 

 examined. These different degrees of relative divergence of order of 

 the three tables indicate a greater degree of complexity of thermo- 

 electric than of chemico-electric action of metals in liquids in the 

 given cases, and a less dependence of the former than of the latter 

 upon simple corrosion. As real thermo-electric action of metals in 

 liquids does not necessarily include corrosion, the latter when it 

 occurs with the former is not the cause of it. 



With regard to the relation of voltaic action to corrosion, Poggen- 

 dorff published (" Phil. Mag.," 1840, vol. xvi, pp. 495 and 537) a 

 number of experiments showing that the chemico-electromotive 

 force was not the strongest in those liquids where the corrosion of 

 the metal was apparently the most rapid, but he does not appear to 

 have actually verified the relative losses of weight by weighing the 

 metals ; nor to have excluded from his tables those metals which 

 became coated with a film of insoluble matter. 



In order to ascertain whether rise of temperature caused every 

 metal to corrode faster, or had always the effect of making only 

 those metals corrode faster which by it were rendered more chemleo- 

 electro-positive, and those more slowly which were rendered negative, 

 series of corrosion experiments were made with most of the same 

 metals and liquids (see p. 252) at 160 F. as were employed in the 

 thermo-electric experiments of Table I, care been taken to keep the 

 metals wholly immersed, and the liquids from altering in strength by 

 evaporation. The length of time during which it was necessary to 

 continue the experiment varied from five minutes to eight hours. 

 Table XX (p. 286), shows the rates of loss in grains per square inch 

 per hour. The proportionate rates of increase for 100 F. rise of 

 1 temperature, as calculated from the numbers on Table XIX, p. 284, 

 are also given for comparison. 



Remarks. By comparing these results with those obtained at 60 F., 

 as shown in Table XIX, p. 284, it will be seen that in nearly every 

 case rise of temperature increased the rate of corrosion. The only 

 exception in the total fifty-seven cases was zinc in dilute nitric acid, 



u 2 



