KING. 



MEASUREMENT OP THOMSON EFFECT IN COPPER. 



363 



A mercury trough resistance, a rough tin plate resistance and a 

 Kelvin current balance were included in the circuit of the heatinof cur- 

 rent. The mercury trough resistance was capable of being varied con- 

 tinuously by means of amalgamated sliding copper contacts. The slider 

 on the Kelvin balance being set to read the desired current, the pointer 

 on the balance arm was kept exactly at the middle mark throughout the 

 experiment, by adjusting the sliding contacts of the mercury trough 

 resistance. The current was drawn from ten one-hundred-ampere hour 

 accumulators in parallel, specially set up for the experiment. 



The Standard Low Resistance. 



This was specially made for the experiment and consisted of four sheets 

 of manganine in parallel between heavy copper terminals. The resist- 

 ance of the standard was .000733 ohm, so that witli the largest current 

 used, that is, 35 amperes, heat was generated at the rate of about one 

 watt. For the dissipation of this energy there was a surface of about 

 275 square cm. immersed in oil through which there was a good water 

 circulation. 



TTie Experimental Bar and Potential Leads. 



The bar was of best quality copper magnet wire. Chemical analysis 

 showed a percentage of 99.3 copper. 



The length between the blocks Bj and Bo was 15.6 cm. The diameter 

 was somewhat irregular, as shown by the following : — 



The bar was annealed and its surface blackened by dipping, while hot, 

 in an acid solution of silver nitrate. 



Nineteen potential leads of no. 40 B. & S. copper wire were copper 

 plated to the bar. Their positions were as below : — 



