360 PROCEEDINGS OF THE AMERICAN ACADEMY. 



ductivity at the point of the bar under consideration, it is possible to 

 calculate the change in the amount of heat conducted past any point 

 on reversal of the heating current. In this way the change in the 

 amount of heat conducted out at No. 8 and No. 9 would be determined. 



In practice it is advisable to use more than one short section in 

 determining the Thomson Effect. In the experiment as set up the 

 sections were made about 1 cm. in length, with the exception of some 

 half as long near the ends. A part of the wire comprising at least 

 four sections was used in calculating the value of the Thomson Effect. 



For convenience of reference, let the quantities used in calculating 

 the Thomson Effect be designated as follows : — 



Let Xi and x^ be any two points on the experimental wire, whose tem- 

 peratures are dx-^ and 6x2- 



Let o- be the Thomson Effect coefficient defined by the relation 

 o-(9xi — 0x.2) = heat developed in Xi to Xo when one c. g. s. unit of elec- 

 tricity passes from Xi to Xo. 



Let Hx^x^ — resistance of section (x^ toa-j), current entering B2. 



Let Hx^xo, = heat dissipated from (xi to X2), current entering B2. 



(r7 B\ 

 — ) = heat conducted past Xi out from section, current enter- 

 dxj . T, 

 ing r>2. 



Let kx.y ( — ) = heat conducted past Xo into section, current entering Bo. 

 " \dx ' 



Let the same letters with dashes represent similar quantities when the 



current enters Bj. 



Let Y = the electric current in amperes. 



Then : — 



2 a- (6x2 - ^^-i) ~ = (^\ X2 - ffr, X.,) — Y^ (R'x, X2 - Rx, x.3) 



+ h 



■Xj, 





The Curve of Distribution of Temperature. 



This curve has been referred to as a e, Fig. 3, and is obtained from a 

 separate experiment with the same apparatus, with the exception that a 

 standard low resistance is required in addition. The determination of 

 this curve involved the measurement of the resistance of the short sections 

 of the experimental wire, (1) when a current small enough not to heat 

 the wire appreciably was flowing, (2) when the heating current Y was 

 flowinsr. From the change of resistance due to the heatino; current the 



