348 Dr. C. Barns on the Relation between the Thermoelectric 



tion-currcnts, which may possibly be generated in the rheostat, 

 are without disturbing effect. The commutator merely serves 

 the purpose of a double key. 



As the electromotive forces measured were all very small, 

 the large resistance W could be left unaltered, so that 

 e= const, x. Now the resistance « was so chosen that the 

 intensity of the current from e exceeded that of the partial 

 current from E by the minimum possible. The thermoelectric 

 force, however, decreasing with the temperature (T — t = r) 

 of the ends, a moment soon arrives at which the intensities of 

 the two currents are equal, and the deflection of the needle =0 

 in consequence. At this point the thermometers are read off. 



Thermoelement. — Instead of measuring the electromotive 

 force of soft and hard steel directly, it was found expedient to 

 compare all the rods with one and the same piece of copper 

 wire. By this means the apparatus could be considerably 

 simplified and many practical difficulties avoided. The con- 

 struction of the copper-steel couple is given in vertical section 

 in fig. 2. 



To raise the ends of the steel rod to different temperatures, 

 two doubly tubulated spherical receivers, each about 1 decim. 

 in diameter, were used. These, held in position by movable 

 supports of poorly conducting material, and so placed that the 

 tubulures A and B were horizontal, the other two vertical, 

 were connected by a glass rod c cl fitting water-tight in the 

 perforated corks adapted to the horizontal tubulures. This 

 rod served a double purpose : by uniting the receivers as one, 

 it prevented breakage of the very brittle steel rods, at the 

 same time allowing them to be easily adjusted and removed; 

 on the other hand, the receivers could by means of it be placed 

 at any distance apart, this being necessary, as the rods to be 

 examined were of very different lengths. 



On one side of the glass rod the copper wires which acted 

 as poles of the instrument were inserted once for all ; on 

 the other two appropriate holes served for the introduction 

 of the steel rods s s to be tested. The ends of the latter were 

 connected with the corresponding ends of the copper wires by 

 small flat clamp-screws of brass. 



The apparatus being thus ready for experiment, the two 

 receivers were filled with distilled water at T and t degrees 

 respectively, where t was so chosen as to differ but slightly 

 from the temperature of the room. The thermometers (intro- 

 duced through the vertical tubulures) were read off with a 

 telescope as follows : the deflection of the needle having 

 become very small, t was determined, after which, when the 

 current in G was =0, T, whereupon another check-reading 



