462 



PROCEEDINGS OP THE AMERICAN ACADEMY. 



and the base 1 \ u long. Each end of this is driven into a small hole in 

 the iron of the hemisphere, thus insuring perfect metallic contact. 



The lead wires from each hemisphere are bound together, and their 

 ends inserted in separate holes on the wood-fibre block used for the 

 terminals in taking measurements. 



The following procedure is used in making the calibration. With the 

 baths at the same temperature, tests are made to determine the existence 

 of any outside E. M. F. The temperatures are then adjusted and the 

 electromotive force determined. Next the temperatures are reversed 

 and the electromotive force again determined. 



! ! ! ; '''i ; ^ r :>I^ :T ^: :r ^ ! '~ : '" : !~^ T ~"T~-^r-~-^--':— i-TH 



■"" rr n^ 



Figure 3. 



The E. M. F. at the contact of the U piece and the hemispheres was 

 at first very large, due to the fact that the U piece was simply pressed 

 down on the iron, but after the small holes were made in the sphere and 

 the U piece was driven in place, it was impossible to detect any extra- 

 neous E. M. F. 



The following curve shows the results of the calibration as carried out 

 in August, 1902. 



The method of calibration of course assumes that the metal employed 

 is of uniform thermo-electric properties, aud so is applicable in strictness 

 only to pure metals. 



In computing results from the curves, the E. M. F. is taken at the 

 temperature of the lower junction and the difference of temperature 



