450 Prof. Tyndall on the Action of Free Molecules on 



revealed the cause of its transporting more heat than the 

 vacuum in the first experiments, and less heat, not only than 

 the vacuum, but than air or oxygen, in the second experi- 

 ments. In the one case the thermometer, being close to the 

 source, came within the range of the convection-currents of 

 this mobile gas, the heat being transported to it by these 

 currents. In the other case a considerable distance intervened 

 between the source and the pile, which was further effectually 

 protected by the narrow tubulure. Through it the currents 

 could not pass ; but they nevertheless existed in the recipient 

 A B, lowering the temperature of the source without heating 

 the thermoscopic instrument. 



The experimental resources of Magnus were great, and he 

 here applied them ; but the defects of his method were radical 

 and irremovable. These defects reached their culmination in 

 the subsequent researches of Professor Buff*, who, by pur- 

 suing substantially the same method, arrived at the result that 

 a stratum of air 2^ inches thick absorbed 60 per cent, of the 

 radiation from a source of 100° C.f Buff also found olefiant 

 gas to be more diathermanous than air ; whereas at atmo- 

 spheric pressures it is many hundred, and at a pressure of ^q 

 of an atmosphere many thousand times more opaque to heat. 



This is the point at which aqueous vapour enters into the 

 experiments of Magnus. When dry and humid air were 

 compared together in his apparatus, no difference between them 

 was observed. But, apart from all disturbance, it would 

 require an instrumental arrangement far more delicate and 

 powerful than that here employed, to bring into view the 

 action of a stratum of mixed air and aqueous vapour 11 inches 

 deep, and having a temperature of only 15° C. Disturbances, 

 however, were not absent. In the first place, the convection 

 currents which enabled dry air to reduce the radiation by 

 11*12 per cent., were more than sufficient to mask the action 

 of the vapour. Secondly, dry and humid air were brought in 

 succession into direct contact with the face of the thermopile. 

 The pile was therefore affected by any difference of tempe- 

 rature between it and the air, and it could scarcely be sup- 

 posed that these temperatures were always alike. It was also 

 affected by the condensation and evaporation which occurred 

 when humid air and dry air were brought successively into 

 contact with its lampblack-coated face. To ** vapour-hesion " 

 Magnus subsequently ascribed very large effects. Here we 

 have the conditions specially suited to the development of the 



* Phil. Mag. oth ser. vol. iv. p. 401. For my reply see Proc. Roy. 

 Soc. vol. xxx. p. 10. t IWd. 1877, vol. iv. p. 424. 



