154 PROCEEDINGS OF THE AMERICAN ACADEMY 



are much more certain that the galvanometer deflections are propor- 

 tional to the heating effects. 



Of course special precautions had to be taken to keep the lens and 

 prism entirely free from moisture, a thin coating of which, such as 

 would be deposited from the open air, would entirely vitiate the 

 results. This was done by making the box one large desiccator, the 

 bottom being covered completely with lead pans filled with pumice- 

 stone and sulphuric acid, pans of these being clustered everywhere 

 about the lens and prism. The lens and prism have now remained 

 several months in the box with hardly an appreciable change. 



Determination of Temperature. 



I have described briefly the method of determining the temperature 

 of the source of heat, but it still remains to describe the process more 

 in detail, and to give the theory of the method. It was my intention, 

 in beginning the experiments, to determine the temperature by meas- 

 uring the variation of resistance of a platinum wire ; but this particular 

 arrangement of the apparatus was suggested to me by Professor 

 Rowland. 



We have seen that the current from the battery comes through the 

 wire, which was rather a strip of platinum foil, then through the 

 tangent galvanometer G, where it may be accurately measured, and 

 then through the resistance R, where its strength is regulated. By 

 setting up the cells, so that the internal resistance of the batten was 

 very small, a constant current was easily obtained. 



The two fine shuuting-wires w 1 w' took off an exceedingly small 

 portion of this current, which was measured by the sensitive sine gal- 

 vanometer G'. The theory of the method is as follows. 



Let C = whole current from battery ; G, the current through the 

 platinum wire ; G", the current through the shunt. Let E, E 1 , E" 

 be the corresponding electro-motive forces, and R, R', R" the cor- 

 responding resistances. 



Then O = G— G" E = E" = G" R" 



-r,. E> C" R" C"R" 

 and R'z=^z= C _ CII — -Q- 



(since G" is very small). 



In general, for any value of the resistance of the platinum-wire, 

 j-, kh sin UR" _ ., , , , „ , . . 



