THE 73° CALORIE. 



385 



The first column of figures gives tlie difference between the two read- 

 ings of a Carey Foster bridge. The second gives similar differences 

 added to the small auxiliary resistances employed. These data are 

 plotted in figure 7, using A2 which was measured, instead of Ai. 



A2 A2 



Correction for this is made by multiplying the slopes by — = - — ^— 



Ai A2— Ao, 



which amounts to the addition of .0004 to the numerical values of the 

 slopes. Thus corrected, these slopes are —.0219 for the slow flow, and 

 — .0228 for the fast flow, giving —.0237 when the correction for Bjj 

 has been applied. A further correction is applied to reduce this 

 result to the scale of a single thermometer, — it thus becomes —.0198, 

 the value which appears in its proper place in the table below. The 

 slopes have in all cases been obtained arithmetically, and may not be 

 in exact agreement with the plotted lines. 



In the following table, a "set" consists of at least four points, — 

 one in each group. The first series of four runs, though comprising a 

 considerable number of points, is not entitled to as much consideration 

 as the last, because the points were not taken in regular order, and 

 often lie too close together. If the last six runs alone be considered, 

 and if they be weighted according to the number of sets in each 

 run, the mean is —.0201, which perhaps best represents the data. 



Had the previous data been included, counting each as a single set, 

 they would have reduced the>esult by two parts in ten thousand. 

 This figure needs but one further correction, that to the gas scale. 

 For this purpose use was made of Callendar's well-known formula 



