528 



Mr. G. Gore on 



No. 110. In order to ascertain the effect of entire absence of the 

 washers with a strongly alkaline solution, I took the solntion of caustic 

 soda nsed in experiments No. 86, 87, and 88, and put it into a large 

 \/-shaped glass tube of 1-inch internal diameter, containing two large 

 platinum electrodes wholly submerged, and suspended by thin plati- 

 num wires, the ends of the tube being nearly closed by corks, and 

 heated one limb of the tube. The hot platinum became positive h\° 

 at 210° F. By allowing the liquid in the heated limb to come into 

 contact with a bung of vulcanized india-rubber, the deflection was 

 increased to 40, thus further proving that vulcanized washers were 

 liable to interfere with the results. 



Conclusions and Remarks. On classifying and comparing in different 

 ways the foregoing results obtained with the large cylinder apparatus, 

 we find 1st, that in nearly all cases where the liquid is strongly acid, 

 and chemical action is absent, the cold metal is positive to the hot one, 

 and the electric current flows upwards. And 2nd, that in nearly all 

 those where the liquid is strongly alkaline, and chemical action does 

 not occur, the hot metal is positive to the cold one, and the current 

 flows downwards ; thus affording additional evidence in support of 

 the conclusions I drew in my former investigation already referred to 

 (see p. 513). 



The nearest approach to exceptions to the above conclusions are in 

 the instance of dilute acetic acid (see No. 10) ; solution of boracic 

 acic] (No. 14) ; acid chromate of potassium (Nos. 33, 34, and 35) ; 

 acid sulphate of potassium (No. 47); chloride of potassium (Nos. 65, 

 66); (compare No. 130); and potassium alum (No. 97); (compare 

 No, 136) ; but in each of these cases the electric currents were very 

 feeble, and may have been clue to the influence of the washers. 



Nos. 3 8, 32, and 60, exhibit rather anomalous results, but in the 

 two first of these corrosion of the metal occurred ; and in each case 

 the metal plates were of palladium, and the solution contained a 

 chloride. 



It is important to notice that, provided chemical action is excluded, 

 the direction of the current in all the experiments was scarcely at all 

 dependent upon the kind of metal, but almost entirely upon that of 

 liquid; the only exception to this were Nos. 85, 96, and 102, and 

 these were but feeble ones. In the exceptional cases, Nos. 18 and 32, 

 chemical action occurred. I therefore conclude that it is the nature of 

 the liquid, and not that of the metal, which most determines the existence 

 and direction of the current. The magnitude of the current, however, 

 is largely affected in many cases by a variation in the kind of metal 

 employed, even when no chemical change takes place. 



In some cases, especially with strongly alkaline liquids, the current 

 d ue to the heat is greatly increased by keeping the hot plate at 210° 

 F. for some time ; for instance, with palladium in an aqueous solution 



