406 Prof. H. BufF on the Thermal Conductivity 



Table I. 





T. 



t. 



r. 



z. 



T- 



lOOr 



Dry air 



ri9-8 



\ 56-0 

 47-7 



27-9 







50-5 



(99-25) 





 22 



760 



356-6 





Air- vac\ium(approx. ) 



[18-4 

 \ 54-6 



[47-0 



28-6 







480 

 (88-5) 





 15 



15 



309-5 



Hvdroffen . . 



r22'5 

 J57-5 

 [ 50-5 



28-0 







57-8 

 (130-6) 





 17 



760 



466-4 





Rarefied hydrogen. . . 



f23-0 

 ^58-0 

 [521 



291 







530 



(110-0) 





 14 



5 



3780 



Air as well as hydrogen, before being introduced into the 

 apparatus, was passed slowly through a tube 40 centims. long, 

 filled with well-dried calcium chloride ; and a beaker filled with 

 calcium chloride was placed in the glass cylinder. 



The results of these experiments are, as far as hydrogen is 

 concerned, the same as those obtained by Magnus, viz. that 

 the denser hydrogen is more diathermanous than the rarefied. 

 Air, however, shows the same property, the cause of which 

 becomes clear on studying the course of the experiments. 



After the greatest deflection of the needle had been reached, 

 it remained almost constant for 30 minutes or more, although 

 the temperature of the hot water continually decreased. The 

 decreasing supply of heat from this source had therefore to be 

 made up in some other way. By suddenly cooling the contents 

 of the brass vessel by means of ice, the influence on the needle 

 was perceptible only after the lapse of several minutes. Evi- 

 dently the temperature of the junction had become more and 

 more dependent on the thermal condition of the sides of the 

 cylinder. Inasmuch as the conductivity of glass is not great, 

 it is easily conceivable how the temperature of those parts of 

 the cylinder at a distance from the heating vessel continued 

 to rise, although the temperature of the water had already 

 considerably decreased. Under these conditions the thermo- 

 meter must have received numerous heat-rays from the sides 

 of the cylinder, and upward currents of gases nmst have been 

 produced whenever the glass cylinder was filled with gas. 



The results of the experiment must therefore remain doubt- 

 ful as long as no care is taken to prevent the heating of the 

 walls of the vessel intended to receive the thermopile. 



