574 PROCEEDINGS OP THE AMERICAN ACADEMY. 



ice directly to the junction. In these tests the wire between the ice and 

 the junction was in free air at the temperature of the room, about 16°C. 



Experiment showed that a couple made of copper and unannealed con- 

 stantin had an electromotive force about 1.5 per cent greater than that 

 of a couple made of copper and annealed constantin, from the same 

 original piece as the unannealed. 



The electromotive force of a copper-constantin couple for a given 

 interval of temperature near the temperature of the room is two or 

 three times as great as that of a (copper)-(german-silver) couple under 

 like conditions. 



So far, then, as our tests go, it appears that in every particular con- 

 stantin is a better metal than german-silver to use thermo-electrically 

 with copper, but we have tested only one specimen of constantin. 



Our thanks are due to many persons for assistance received in the 

 prosecution of this research, but especially to Mr. John Connors, the 

 head janitor of the Jefferson Physical Laboratory, for much difficult 

 work in the construction of the Crafts heating apparatus, and to Mr. 

 John Coulson, Professor B. 0. Peirce's mechanical assistant, for the 

 skilful construction and testing of resistance coils. 



