\ 



Effects of Heat modified by Compression. 303 



the experiment i^ a good one, and that the carbonic acid 

 has been well confined, the intrusion seldom takes place in 

 any position. In whichever of the two opposite positions 

 the large tube was placed, a pyrometer was always intro- 

 duced, so as to lie as near as possible to the small tube. 

 Thus, in the first-mentioned position, the pyrometer was 

 placed immediately below the large tube, and, in the other 

 position, above it ; so that, in both cases, it was separated 

 from the carbonate by the thickness only of" the two tubes. 



Much room was unavoidably occupied by this method, 

 w;hich necessarily obliged me to use small quantities of car- 

 bonate; the subject of experiment seldom weighing more 

 than 10 or IC grains, and in others far less*. 



On the nth of April 1803, with a barrel of old sab)e 

 iron, having a bore of 0*75 of an inch, I made an experi- 

 ment in which all these arrangements were put in practice. 

 The large tube contained two small ones; one filled with 

 spar, and the other with chalk. I conceived that the heat 

 had risen to 33", or somewhat higher. On melting the 

 metals, the cradle was thrown out with considerable vio- 

 lence. The pyrometer, which, in this experiment, had been 

 placed within the barrel, to my astonishment indicated 64**. 

 Yet all was sound. The two little tubes came out quite 

 clean and unconiaminated. The spar had lost 17"0 per 

 cent.: the chalk 10"7 per cent. The spar was half sunk 

 down, and run against the side of the little tube : its surface 

 was shining, its texture spongy, and it was composed of u 

 transparent and jelly-like substance : the chalk was entirely 

 in a state of froth. This experiment extends our power of 

 action, by shov.ing, that compression, to a considerable de- 

 gree, can be carried on in so great a heat as 64*^. It seems 

 likewise to prove, that in sou)e of the late experiments wiih 



• I measured tlie capacity of the air-tubes by means uf granulated tia. 

 aclinj^ as a fine and equal sand. By comparing the weight of this tin wilK 

 an equal bulk of water, 1 found that a cubic incli of it weigtied 1:3:!0G grains, 

 and tliat each grain of it corresponded to 0-00075 of a cubic inch. From 

 these data I was able, with tokr^ble accuracy, to gMijce a ti^be by weightrg 

 the tin required to fill it. 



tl,r. 



