Properties, Specific Resistance, and Hardness of Steel. 347 



wards, however, it was found expedient to measure these 

 forces (as Kohlrausch and Ammann * had done in similar 

 experiments before) by a method of compensation, the object 

 being to avoid the difficulties from the species of polarization 

 due to Peltier's phenomena. The method can be easily de- 

 duced from that proposed by Bosscha, the latter, in the case 

 where small electromotive forces are to be measured, admitting 

 of simplification. 



In the diagram (fig. 4), E denotes the compensating element 

 (1 Daniell element, = 11*7 Weber-Siemens units), e the ther- 

 moelectric couple whose electromotive force is to be deter- 

 mined, both acting as shown in the figure, C a Weber's 

 commutator (employed for reasons given below), G the gal- 

 vanoscope. Let us represent the resistance of the branch a b 

 by x, that of the branch a E b by W + k, where W represents 

 the large resistance of a rheostat interposed, h that of the re- 

 mainder of the branch (about equal to 1 Siemens unit) inclu- 

 ding E. 



When the current in Gr is zero, we have 



e 



E W + k + x 



e 

 But as the ratio ^ is small, and therefore of necessity also x 



(maximum value = 10 Siemens units), in comparison with 

 W + h + x (about 20,000 Siemens units), we may with suffi- 

 cient approximation put 



E 



the experimental accuracy obtainable allowing us to neglect 

 k + x in comparison with W. In the experiments the branch 

 a b was a small Siemens rheostat. 



The precise moment in which the current in the galvanoscope 

 is equal to zero can be best determined by observing whether 

 the needle on closing and opening the circuit remains at rest. 

 This, however, is only possible when the opposed currents 

 from e and E which pass through the galvanoscope are closed 

 simultaneously. To accomplish this, the little cups at the end 

 of the rods 1 and 2 of the commutator were quite filled with 

 mercury, those of 3 and 4 only partially. By this device, on 

 closing, (j\ and C 2 are first brought into contact, and the cur- 

 rent E C 2 C 1 6aWE, not passing through the galvanometer, 

 comes into action ; in the next moment (C 3 and C 4 being 

 joined) the current from e and the partial current from E re- 

 ferred to are closed simultaneously. In this way also induc- 

 * Pogg. Aim.QTL\\. p. 459, 1870. 



