222 Dr. F. Auerbach on the Passage of 



1. In a thin iron wire, Bunsen elements being employed, 

 the following were found as the means of many experiments, .of 

 which two consecutive ones only were constantly used for the 

 formation of the differences of resistance: — 



ic 2 -i Vl = 0-0037, u? 3 -ir 2 = 0-0049, '^iZ^L 1 =0-0041, 



Mean 0*0042; 

 (mean of w = 6'83) therefore 



8 = 0-00061. 



If this 8 were solely a consequence of the heating, it could 

 not but be much greater in comparison with the 8 of experi- 

 ment-series 1. 



The results of another series of experiments, in which the 

 thermal influences were at most very slight, are graphically 

 represented in fig. 7. 



5. The share of the temperature-influences in the value of 

 8 was compelled to represent itself isolatedly in a copper wire. 

 The copper was pure, 1= 18000, <r/ = 0'41, and the mean value 

 of w = 4'164. I found, on employing the same Bunsen ele- 

 ments as in 4, 



103-11'! = 0-00090, 

 therefore 



^P^ =o = 0-00011. 



Z\L\ 



This value agrees with the probable value of the thermal 

 influence for this special case, so far as agreement is possible 

 in rough calculations of this sort. 



Of various artifices which I employed in order to exclude 

 the influence of heating, the following finally proved to be the 

 most effectual, at least with thin wires. The resistance of the 

 iron wire was determined approximately, to one or two places 

 of decimals. On closing the bridge, there then resulted a still 

 smaller deflection toward one side, perhaps toward that to 

 which corresponds too small a measuring resistance. The last 

 figure of this measuring resistance was then made 1 higher by 

 insertion in the resistance-case, so that now, on the closing of 

 the bridge, a deflection resulted toward the opposite side. Let 

 these two deflections be equal a n and b n for the case in which 

 n elements are used : if these are small quantities (in the ex- 

 periments their angles never exceeded 1 5'), and if induction- 

 phenomena of every sort are excluded, any observed deflection 

 s n can be reduced to an additive or subtractive resistance by 

 dividing it by a n + b n , according as this deflection is observed 

 at the taking-out or insertion of a unit in the last place of de- 



