60 MM. A. Kundt and E. Warburg on Friction 



with Maxwell's theory ; that of carbonic acid is, according to our 

 experiments, sensibly less than the theoretic one. 



The absolute conduction-coefficients of gases could only be 

 approximately calculated, as we could procure only an approxi- 

 mate knowledge of the water-value of our thermometer. Using 

 this approximate value, we find the conduction-coefficient of air 

 equal to \^ of that obtained by M. Stefan. 



We will now describe how we succeeded in making the space 

 of the cooling-apparatus employed approximately a vacuum with 

 respect to the conduction of heat. 



It appeared at first that the time of cooling of a thermometer, 

 if a vacuum is produced in the well-dried cooling-space, dimi- 

 nishes with time when the apparatus is left to itself. For ex- 

 ample, in one experiment, in twelve hours it fell from 351 

 seconds to 307 (in which times the thermometer sank from 60° 

 to 20°). The traces of aqueous vapour which cause this may 

 partly proceed from the fat of the cock ; but they are partly 

 discharged from the glass sides ; for the discharge can be acce- 

 lerated by heating parts of the latter. By such heating we 

 were able to lessen the time of cooling in another exhausted ap- 

 paratus from 282 to 264 seconds. 



The quantities of aqueous vapour in question are extraordi- 

 narily small, and not to be indicated by the manometer. The 

 velocity of cooling of a thermometer, which can be measured 

 with great nicety and facility, consequently proves to be an ex- 

 tremely fine test for the quality of a vacuum ; we doubt if a finer 

 can be found. 



We obtained the best vacua by drying the apparatus in the 

 oil-bath at 200°, and shutting it off at that temperature. The 

 following Table shows the gradual augmentation of the time of 



cooling. 



Pressure. Time of cooling (60°-20°). 



millim. seconds. 



760 225 



154 313 



9-8 363 



4 369 



1-3 364 



Vacuum 1 444 



„ II 555 



„ III 602 



„ IV. (at 200°, dried, 

 exhausted, and shut off) 712*5 



By the last manipulation the time of cooling was increased in 

 the ratio of 6 : 7. 



In order now to prove whether a vacuum thus produced should 



