and the Economy of Fuel. 5 1 



55 parts in bulk of air, "with i part of fine raw silk, 

 formed a covering for confining heat, the conducting 

 power of which was to that of air as 576 to 1284, or 

 as 448 to 1000. Now, from the result of this last-men- 

 tioned experiment, it should seem that the introduction 

 into the space through which the heat passed of so small 

 a quantity of raw silk as 5 V part of the volume or capac- 

 ity of that space, rendered that space (which now con- 

 tained 5 5 parts of air and i part of silk) more impervious 

 to heat than even a Torricellian vacuum. The silk must 

 therefore not only have completely destroyed the con- 

 ducting power of the air, but must also at the same time 

 have very sensibly impaired that of the ethereal fluid 

 which probably occupies the interstices of air, and which 

 serves to conduct heat through a Torricellian vacuum : 

 for a Torricellian vacuum was a better conductor of 

 heat than this medium, in the proportion of 604 to 448. 

 But I forbear to. enlarge upon this subject, being sensi- 

 ble of the danger of reasoning upon the properties of 

 a fluid whose existence even is doubtful, and feeling 

 that our knowledge of the nature of heat, and of the 

 manner in which it is communicated from one body to 

 another, is much too imperfect and obscure to enable us 

 to pursue these speculations with any prospect of suc- 

 cess or advantage. 



Whatever may be the manner in which heat is com- 

 municated from one body to another, I think it has been 

 sufficiently proved that it passes with great difficulty 

 through confined air ; and the knowledge -of this fact 

 is very important, as it enables us to take our measures 

 with certainty and with facility for confining heat, and 

 directing its operations to useful purposes. 



But atmospheric air is not the only non-conductor of 



