So4 Liquefaction and Solidijicatio7i of Carho?iic Acidi 



As below 32^, or at reduced pressures, the augmentation of 

 temperature is productive of much less expansive influence, we 

 may infer that under the weight of a few atmospheres, as when 

 near to its freezing point, liquid carbonic acid is scarcely more 

 dilateable by heat than water. Between — 4° and +32°, its 

 expansion is 0.053 while that of air is 0.069. These facts suggest 

 the inquiry how far water at very high temperature and pressure 

 may be obedient to the same expansive influence, and thus by 

 suddenly filling the whole interior of boiler, sometimes cause ex- 

 plosions. 



The pressure of carbonic acid gas, when placed over its liquid^ 

 is given by Thilorier at 32° and 86°, as 36 and 73 atmospheres 

 respectively. By means of the gauge S, M, R, — I found the 

 pressure as follows : 



32° - - - - 36 atmospheres. 



45° - - - - 45 do. 



66° - - - - 6.0 do. 



86° - - - - 72 do. 



The principle of the gauge renders it capable of registering the 

 pressure with great accuracy ; — for as one tube, M, begins to mark 

 the pressure from the commencement of an experiment, and the 

 mercury in the other, R, does not reach a visible point until the 

 first has shown a pressure of several atmospheres, the second tube 

 is equivalent in eff'ect to one of several times its length. The 

 first determines the amount of pressure, at which the mercury 

 reaches the initial point on the 2nd, and the 2nd, subsequently, 

 exhibits the multiplicators of that initial quality. Thus, if when 

 the mercury is at five atmospheres in M, it is at the unit mark 

 in R, the value of that unit will be five, and the numbers repre- 

 sentative of the pressure on R, must be multiplied by five ; or R 

 is equal in effect to a tube five times its length. By these means 

 very short tubes may be used to determine very high pressures. 

 Inequalities in temperature, irregularities in the cement, and 

 other causes, may vary the capacity of the socket T, W, but as 

 M always signifies the unit for R, in each case, no error can arise 

 from these causes. There must, of course, be a correction for 

 the weight of the mercurial column in R, which is to be added 

 to the product. Care must be taken to keep the temperature of 

 the vessel which holds the liquid below that of the gauge and 

 tubes, otherwise the hquid will be formed by condensation in 



