Comparison of Formulmfor Total Radiation. 439 



indicated by the thermometer as the temperature of the radia- 

 ting surface. 



The method of work and the degree of accuracy attained 

 may now be further illustrated by tabulating some of the 

 results. 



Tables I and II exhibit two independent series of measure- 

 ments, on a fall of temperature through 40° C, the initial tem- 

 peratures being approximately 60° C. as shown. For con- 

 venience the tables are arranged in the order of the intervals of 

 time included in each, rather than in that of increasing tem- 

 perature-difference which would on some accounts be prefer- 

 able. The measurements were on the radiation of the iron 

 disk. For the series indicated by Table I the initial deflection 

 obtained before heating the disk was o x = +03 scale division. 

 The disk was then heated up to a temperature a little in excess 

 of 60° and hung in position. While it cooled down to 21° - 6 a 

 series of 57 observations were made at intervals of two 

 minutes. The residual correction was then found to be 

 <? 2 = — 2 - scale divisions. The sensitiveness of the galvanome- 

 ter remained unchanged during the 114 minutes of work, while 

 the temperature within the box was raised from 16°' 16 to 

 16°*72. For the purpose of tabulation the absolute tempera- 

 tures, time intervals, and corresponding galvanometer deflec- 

 tions were obtained by interpolation for intervals of 5° in suc- 

 cession, after all necessary corrections had been applied to the 

 recorded readings. The horizontal column H gives the hour and 

 minute at which the record thus obtained should have been 

 made. Column T gives the absolute temperatures of the 

 disk, T those of the thermopile, and T — T the temperature 

 differences. Column E gives the experimental result in radia- 

 tion, as obtained in scale divisions of deflection. Column W 



gives the numbers fori™ e —1), which is the factor 



involving the variable, T, in Weber's formula. Column S 



gives the numbers for I™ 4 — 1 ), involving the same variable 



in Stefan's formula. Each number thus obtained by formula 

 is then divided by the corresponding values of E in the same 

 vertical column obtained by experiment. The quotients are 

 arranged in the horizontal columns W -~ E and S -=- E respec- 

 tively, and multiplied by a power of 10 for the purpose of 

 avoiding inconvenient decimals. This has no influence since 

 our object is only that of comparison. If a formula be cor- 

 rect, and if the measurements be free from error, the numbers 

 contained in the corresponding horizontal column of quotients 

 should be identical. If they progressively increase or decrease 



