414 Prof. H. BufF on the Thermal Conductivity 



real heating-efFect of the brass vessel upon the cooling-water. 

 This, however, was evidently not the case. 



From the small deflections of the needle, varying within 2 

 and 5, it was evident that the temperature of the junction, even 

 had it reached its maximum, could not be very different from 

 that of the surrounding air. Inasmuch as the air itself ab- 

 sorbed some of the incident rays, its temperature above the 

 junction must consequently have been gradually rising. 

 Scarcely any cooling of the junction by upward currents was 

 observable. As soon as equilibrium was restored between the 

 influx and loss of heat, the loss must have been occasioned, in 

 vacuum as well as in air, chiefly by radiation. But the results 

 of the preceding experiments show that the loss by radiation 

 is greater in vacuum than in air. The soldered junction re- 

 tains in the latter case, therefore, a comparatively higher tern- 

 perature. In other words, the same causes which retard the 

 influx of heat by radiation lessen, on the other hand, the loss. 

 Thus the diathermancy of air appears greater than it is in 

 reality. 



By wrapping the junction in paper this source of error was 

 diminished — and upon covering the interior of the cylinder 

 with double cardboard, nearly disappeared. 



The numbers obtained when this protection against the 

 disturbing influence of the cooling-water was used are as fol- 



lows 



Table YII. 



T. 



T'. 



t. 



V, 



z\ 



inillims. 



100 r 



t ' 







Dry air 









66-0 

 65-8 

 60-3 

 63-0 



46-5 

 49-0 

 440 

 460 



410 

 44-0 

 39-4 

 40-6 



70 



7-5 



11-5 



11-8 



5 

 5 

 5 

 5 



7550 



7530 



4-8 



2-5 



1710 

 17-04 

 3000 

 2900 







Dry hydrog 



en. 







70-0 

 670 



46-5 

 45-8 



40-6 

 39-7 



12-3 

 11-7 



70 

 65 



755-5 

 755-5 



303 

 29-5 



The absorptive power of the dry air is almost similar to that 

 formerly obtained when the junction was nearer the source of 

 heat. The experiments with hydrogen, quoted in Table Yll., 

 show the close agreement of its diathermancy with that of a 

 vacuum. 



The absorptive power of air for greater distances from the 



