Measurement of the Peltier Ejfect. 461 



effective resistances of the two junctions are nearly the same, 

 r 2 may be written = r 3 -f- r±. Then 



(Ci + C 2 ) 



Taking the results from the first part of the table, 



_ 0-0159 x 0-035 

 7 ' 3 ~ 0-125x2-9 



r s = 0*0015 ohm nearly. 



The maximum possible resistance of each junction will be 

 r 3 together with half the resistance of the bismuth rod. 

 In the complete equation for P (5), 



? . 1 = r 2 = 0-0015 + 0009 = 0-0024 ohm nearly. 



The approximate value of ~r as obtained by substitution 



in (1) and (2) is 1*08. Substituting these values in equation 

 (5) the second term 



(C2 - cy {(S)^A ( "-^} 



= -(l-45-l-35){(^i)0-0024 + 0} 



= -9'2x 10~ 6 volt nearly. 



Since P = 15900xl0 -6 volt by equation (6), it follows 

 that the second term in (5) is much less than the errors of 

 observation and can be neglected. Thus the calculation of 

 P by equation (6) is correct to 1 part in 1000, even when Q 2 

 differs from Q x by 4 per cent. When errors of observation 

 are taken into account the final result is probably correct to 

 1 part in 200, and this order of accuracy was aimed at in 

 designing the apparatus. 



The bismuth, which was supplied by Griffin and known 

 as Kahlbaum's pure bismuth, was carefully crystallized by 

 Dr. Lownds in the following way. A crucible was over- 

 wound with a coil of eureka wire and surrounded on all 

 sides with sand and asbestos. About 1000 gms. of bismuth 

 were melted in the covered crucible, and the whole mass was 

 very slowly cooled by gradually diminishing the current 

 through the coil. The whole operation of cooling lasted 

 3ti hours. The crucible was broken and the lump of bismuth 

 removed. A blow with a hammer near the 



