in various S^lbstances. 457 



ing from 70 to 10 was 576 seconds; but in the experi- 

 ment No. 4, when this same space was filled with 54 

 parts air, and I part raw silk, the time of cooling was 

 1284 seconds. 



Now, supposing that the silk had been. totally in- 

 capable of conducting any Heat at all, if we suppose, 

 at the same time, that it had no power to prevent the 

 air remaining in the globe from conducting it, in that 

 case its presence in the globe could only have prolonged 

 the time of cooling in proportion to the quantity of the 

 air it had displaced to the quantity remaining, that is to 

 say, as i is to 54, or a little more than 10 seconds. But 

 the time of cooling was actually prolonged 708 seconds 

 (for in the experiment No. i it was 576 seconds, and in 

 the experiment No. 4 it was 1284 seconds, as has just 

 been observed) ; and this shows that the silk not only 

 did not conduct the Heat itself, but that it prevented 

 the air by which its interstices were filled from conduct- 

 ing it ; or, at least, it greatly weakened its power of 

 conducting it. 



The next question which arises is, how air can be 

 prevented from conducting Heat ? and this necessarily 

 involves another, which is, How does air conduct Heat? 



If air conducted Heat, as it is probable that the 

 metals and water, and all other solid bodies and unelastic 

 fluids, conduct it, that is to say, if, its particles remain- 

 ing in their places, the Heat passed from one particle to 

 another, through the whole mass, as there is no reason to 

 suppose that the propagation of Heat is necessarily in 

 right lines, I cannot conceive hovv the interposition of so 

 small a quantity of any solid body as ^ part of the 

 volume of the air could have effected so remarkable a 

 diminution of the conducting power of the air, as ap- 



