152 PROCEEDINGS OF THE AMERICAN ACADEMY 



in a vulcanite frame, and were connected, one with the battery direct 

 and the other through the tangent galvanometer G, and the variable 

 resistance R ; c was merely a commutator for reversing the current 

 through the galvanometer; w' w' (Fig. 1) are two exceedingly fine 

 platinum wires welded to the main wire, about 8 mm - apart, and con- 

 nected with a second galvanometer, G'. The object of these was to 

 shunt off a small portion of the current, which, being compared with 

 the main current, gave a method of determining the temperature of 

 the large wire : this method will be further described. 



The vulcanite frame was enclosed in a box of polished tin, shaped 

 as shown in the figure. This peculiar shape was intended to reflect 

 such rays as did not fall directly upon the lens to one side, and finally 

 to the back, where they were absorbed by the V shaped partition, 

 behind which was a compartment filled with water. 



The current for heating the wire came by means of the wires to, 

 from a battery of thirty-two Bunsen cells, placed in the room below. 

 It was measured by the galvanometer, G, placed on a stone pier, and 

 its strength regulated by a mercury resistance, R. 



The part of the current shunted off by the wires w' w' was meas- 

 ured by a delicate galvanometer, G', also placed on a stone pier. In 

 this circuit there was a resistance of 1,000 ohms t, and a commutator c' 

 for reversing the current through the galvanometer. 



The rays that were given out by s passed through an opening, 

 which could be closed, if desired, by the door of double tin, a, fell 

 upon the rock-salt lens L and the rock-salt prism P, and produced a 

 spectrum at v b r about 15 cm - long. The lens and prism were both 

 of perfectly transparent cloudless rock-salt. The lens was mounted 

 in a double tube of brass, for focusing, the distance S L being about 

 14 inches. 



The prism was set on a plane surface, so as to be easily rotated or 

 moved in any direction. It was set at its angle of minimum deviation 

 by means of the reflection from the back, which, moving through twice 

 the angle that the prism was moved, and over an arc whose radius was 

 the distance from the prism to the pile, some four feet, furnished an 

 excellent method of setting at the angle of minimum deviation for 

 any part of the spectrum. After setting, this reflection was cut off by 

 means of a little screen placed at the corner of the prism. 



The thermopile T was covered with camphor lamp-black, and envel- 

 oped, with the exception of the slit, by a double casing of tin. The 

 back part of this w r as filled with water to absorb the extra rays, and to 

 keep that face of the pile at a constant temperature. This screen, 



