436 TT. LeConte Stevens — Experimental 



a lower fixed temperature. The temperature of the wire was 

 deduced from the variation in its resistance. The result was 

 to show that for low temperatures Stefan's formula gives 

 results that are too great, and for high temperatures they are 

 too small. The application of Weber's formula to Schleier- 

 macher's data yields results that are more consistent. 



The following experiments have been made with a view to 

 comparing the formulas of Dulong and Petit, Rosetti, Stefan 

 and Weber in their application to temperatures but little 

 removed from that of the atmosphere, not employing the 

 indirect method of Schleiermacher, but measuring tempera- 

 tures directly with the thermometer and radiation with the 

 thermopile. 



The radiating body was a metallic disk instead of the Leslie 

 cube. A hole was drilled into it, parallel to one of its flat 

 surfaces and extending nearly to the center. This was for the 

 reception of a thermometer. The disk could therefore be 

 raised to any desired temperature within the range of the ther- 

 mometer, and its temperature be noted at any moment while 

 it was cooling. One face was kept always scrupulously clean 

 and smooth so as to prevent variations in emissive power. This 

 face was turned toward a thermopile at a fixed distance from 

 it. The thermopile was enclosed in a double -walled box of 

 sheet brass for the purpose of preventing draughts and other 

 disturbances that might affect its temperature. Through the 

 cover of this box a delicate thermometer was passed, so that 

 its bulb rested almost in contact with the thermopile. Varia- 

 tions of temperature within the box were thus measurable to 

 within a hundredth of 1° C. In the end of the box turned 

 toward the disk was a circular opening, 7 cm in diameter, pro- 

 vided, with a double-walled adjustable brass cover. This was 

 controlled by means of a string so that the thermopile could 

 be exposed to radiation for a brief interval and then shut off 

 from this. The thermopile was connected with a very delicate 

 mirror galvanometer of low resistance, and all readings were 

 taken by means of telescope and scale. The sensitiveness of 

 this galvanometer was tested from time to time by putting it 

 in circuit with a standard Clark cell and a resistance of 72,000 

 ohms. Around the suspended disk, except on the side toward 

 the thermopile, double brass screens were placed, and a screen 

 also above the space between it and the box. Disturbing air 

 currents were thus avoided, not completely, but as far as pos- 

 sible. The entire apparatus was in a basement room whose 

 temperature remained very nearly constant from day to day. 

 A thermometer, hung outside of the brass screens, permitted 

 comparison between the temperature of the room and that 

 within the box containing the thermopile. 



