OF ARTS AND SCIENCES. 47 



INVESTIOATIONS ON LlGOT AND IIeAT, PUBLISHED WITH AN APPROPRIATION FROM THE 



RuMPORD Fund. 



IV. 



CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF 

 HARVARD UNIVERSITY. 



THERMOELECTRIC LINE OF COPPER AND NICKEL 



BELOW 0°. 



Charles Bingham Penrose. 



Presented by Professor Trowbridge, June 8,1881. 



The great difficulty to be encountered in experiments on thermo- 

 electricity is the variation in the results obtained by different experi- 

 menters. There can be no comparison with previous experiments 

 when all the results are different. As an example, take the electro- 

 motive force of a junction of bismuth and copper — with one junction 

 at 0° and the other at 100° — as obtained by different experi- 

 menters : — 



Wheatstone 0.00106 



Neumann 0.00390 



J. Regnault 0.00286 



E. Becquerel 0.00483 



These results are referred to the electromotive force of a Daniell 

 element as unity. 



It will be observed that the last result is over four times as great 

 as the first. 



There are many causes which might produce this variation. Slight 

 differences in the structure of the metals often affect the results, and 

 the results obtained with the same metal, before and after it has been 

 subjected to pressure and tension, are often very different. A piece 

 of hard steel always gives different effects from a piece of soft steel. 

 But these causes must all be of minor importance ; the great trouble 

 consists in the impurity of the metals. It is well known that other 

 electrical properties of metals are greatly changed by slight differences 

 in purity. Thus the specific resistance of copper may be increased 

 fifty per cent by the presence of slight impurities. 



