184 ANNUAL KECORD OF SCIENCE AND INDUSTRY. 



The fullest series of experiments were made with air and 

 hydrogen, and the numbers obtained for these gases showed 

 that in air, down to a pressure of one millimeter, the conduc- 

 tion of heat is independent of the pressure. Hydrogen, on 

 the other hand, showed a quite divergent and jjitherto unex- 

 plained behavior with reference to pressure, the changes of 

 the currents in this gas with different pressures by no means 

 affording an explanation of the observed differences in the ve- 

 locity of cooling. In proof of this, for example, he states that 

 whereas with a lowering of the pressure from '750 millimeters 

 to 91.4 millimeters, there was a lowering of the value for the 

 velocity of cooling of only 1.4 per cent.; yet that on fur- 

 ther diminishing the pressure to 4.7 millimeters, there was 

 a further decrease of 11 per cent., the decrease continuing 

 when the pressure was still further lowered to 1.92 milli- 

 meters. Whether the circumstances operating in this case 

 were accidental, or whether these anomalies are due to the 

 properties of the gas itself, can be decided only by further 

 and more exact researches. Besides the above question of 

 simple heat-conduction in gases, Winkelmann also undertook 

 to determine the relation in which heat-conduction stands to 

 the temperature employed. In this investigation he had to 

 employ new apparatus made of glass. He was also obliged 

 to effect the separation of the conduction from the radiation 

 on a different principle from that used in the former measure- 

 ments. The observations were so arranged that first the 

 time of cooling from 18 to 8 was determined, and then 

 from 118 to 108. With three apparatuses, very different 

 in their dimensions, the author obtained the temperature co- 

 efficients 1.3661, 1.3429, 1.3644, referring to the tempera- 

 tures 7.4 to 7.6, and 107.7 to 109. That is to say, if the 

 heat-conduction at the lower temperature be put equal to 

 one, then at the higher temperature it has the value just 

 given. Besides air and hydrogen, carbonic acid was also ex- 

 amined in these researches. If the latter changes its heat- 

 conduction with the temperature in the same w^ay as air and 

 hydrogen, it is obvious, by combination of the values of hy- 

 drogen and carbonic acid, that the same relative numbers 

 should be obtained as those given by hydrogen and air. The 

 values so obtained, however, are altogether smaller, whence 

 it appears that the conduction of carbonic acid is not de- 



