412 Prof. H. BufF on the Thermal Conductivity/ 



of calculation taken as the basis of my experiments. The 

 starting-point of this calculation is the assumption that the 

 maximum deflection is in every case proportional to the differ- 

 ence of temperature t. If this were correct within the range 

 of the temperatures selected, the first five figures in the last 

 column ought to be equal, or nearly so. A more careful con- 

 sideration of my experiments, however, leads me to the con- 

 clusion that the differences there apparent do not arise solely 

 from errors of observation, but chiefly from the fact of its 

 being impossible, notwithstanding all precautions, to overcome 

 the influence of heat from the sides of the glass cylinder. 

 The differences between the numbers found and their average 



value, — 7 — =43*4, however, are much too small to leave the 



o 



main question doubtful. 



To show the difficulties of maintaining the initial tempera- 

 tures of the glass cylinder, I give here an experiment in full : — 



Table YI. 



T. 



r. 



z. 



T'. 







h 





71 







9-55 



7-34 



84-0 



230 



9-56 



7-34 



74-0 



26 5 



9-58 



7-34 



71-4 



270 



9-59 



7-34 



703 



271 



9-60 



7-34 



68-8 



270 



9-61 



7-34 



68-0 



27-0 



9-62 



7-35 



66-4 



26-7 



9-63 



7-35 



64-5 



26-0 



9-65 



7-33 



The figures of the first three columns require no further ex- 

 planation ; those in the fourth column, under T^, give the 

 temperatures of the cooling-water at the height of the junc- 

 tion — ?'. e. 38 millims. under the surface of the water. Here, 

 as is apparent, the temperature could be kept very uniform by 

 means of the regular flow of the water from the surface. But 

 it was impossible to avoid a gradual increase of temperature 

 above that point and towards the inside of the vessel. This 

 is shown by the still rather slow return of the needle after it 

 reached the maximum. It is clear that the maximum was 

 thus somewhat altered, though to a very trifling degree, 

 which in some cases, however, may not have been exactly 

 proportional to the differences of temperature. But, as already 

 stated, the main question at issue, that of the great absorptive 

 power of air for rays of low temperatures could only slightly 

 have been affected thereby. 



