1880.] On Heat Conduction in Highly Rarefied Air. 239 



V. "On Heat Conduction in Highly Rarefied Air." By 

 William Crookes, F.R.S. Received November 18, 1880. 



The transfer of heat across air of different densities has been 

 examined by various experimentalists, the general result being that 

 heat conduction is almost independent of pressure. Winkelmann 

 (" Pogg. Ann.," 1875-76) measured the velocity of cooling of a 

 thermometer in a vessel filled with the gas to be examined. The 

 difficulty of these experiments lies in the circumstance that the cooling 

 is caused not only by the conduction of the gas which surrounds the 

 cooling body, but that also the currents of the gas and, above all, 

 radiation play an important part. Winkelmann eliminated the action 

 of currents by altering the pressure of the gas between 760 and 

 1 millim. (with decreasing pressure the action of gas currents becomes 

 less), and he obtained data for eliminating the action of radiation by 

 varying the dimensions of the outer vessel. He found that, whereas 

 a lowering of the pressure frem 760 to 91 '4 millims. there was a 

 change of only 1*4 per cent, in the value for the velocity of cooling, 

 on further diminution of the pressure to 4*7 millims. there was a 

 further decrease of 11 per cent., and this decrease continued when the 

 pressure was further lowered to 1*92 millim. 



About the same time Kundt and Warburg (" Pogg. Ann.," 1874, 5) 

 carried out similar experiments, increasing the exhaustion to much 

 higher points, but without giving measurements of the pressure below 

 1 millim. They enclosed a thermometer in a glass bulb connected with 

 a mercury pump, and heated it to a higher temperature than the 

 highest point at which observations were to be taken ; then left it to 

 itself, and noted the time it took to fall through a certain number of 

 degrees. They found that between 10 millims. and 1 millim. the time 

 of cooling from 60° to 20° was independent of the pressure ; on the 

 contrary, at 150 millims. pressure the rate was one-and-a-half times as 

 great as at 750 millims. Many precautions were taken to secure 

 accuracy, but no measurements of higher exhaustions being given the 

 results lack quantitative value. 



It appears, therefore, that a thermometer cools slower in a so-called 

 vacuum than in air of atmospheric pressure. In dense air convection 

 currents have a considerable share in the action, but the law of cool- 

 ing in vacua so high that we may neglect convection, has not to my 

 knowledge been determined. Some years ago Professor Stokes 

 suggested to me to examine this point, but finding that Kundt and 

 Warburg were working in the same direction it was not thought 

 worth going over the same ground, and the experiments were only 



s 2 



