51 



ordinary conclusion then announced, that an electric current in an un- 

 equally heated conductor, if its nominal direction be from hot to cold 

 through the metal, causes a cooling effect in iron, and a heating effect in 

 copper, the author describes new experiments which he has recently 

 made, and which are as decisive in leading to the same conclusion as 

 those by which he had first established it. He also describes ex- 

 periments by which he had recently given an independent demon- 

 stration that brass has the same property as copper, and platinum 

 the same quality as iron, with reference to electric convection of heat ; 

 results anticipated*, one as certain, and the other as highly probable, 

 from the previous results regarding electric convection in copper and 

 iron, and from the known thermo-electric relations between these 

 metals and the others. 



2. The phenomenon of thermo-electric inversion between metals, 

 discovered by Gumming, forms the subject of the second part. A 

 mode of experimenting is described, by which inversions may be 

 readily detected when they exist between any two metals, and, when 

 thermometers are available, the temperature of neutrality determined 

 with precision. Various results of its application are mentioned, of 

 which some are shown in the following Table : 



The number at the head of each column expresses the temperature Centigrade 

 by mercurial thermometers, at which the two metals written below it are thermo- 

 electrically neutral to one another ; and the lower metal in each column is that 

 which passes the other from bismuth towards antimony as the temperature rises. 

 PI> P 2 > PS denote three particular specimens of platinum wire, used by the author 

 as standards. 



* See Proceedings, May 4, 1854 ; also " Dynamical Theory of Heat," Part VI. 

 135 ; Transactions R.S.E., May 1854, p. 146. 



VOL. VIII. F 



