232 MR. W. F. G. SWANN ON THE SPECIFIC HEATS OF AIR AND 



directly from the heater. Thus we should expect that the term A, which REGNAULT 

 found to represent 5 per cent, of the total rise in temperature of the calorimeter per 

 minute, might be practically zero when the gas was flowing through the pipe. 

 Further, the passage of the gas would increase the radiation loss from the pipe, since 

 it would increase its mean temperature. Both of these effects tend in the same 

 direction, but the second is probably small. The form of connecting pipe used by 

 REGNAULT would tend to minimise the above effects, but it is impossible to say by 

 how much. 



It is probable that the true value of A lies somewhere between zero and the value 

 found by REGNAULT. If we take A equal to zero, REGNAULT'S experiments give 

 0'248 for the specific heat. If we take for A the mean between the two extreme 

 values, zero and REGNAULT'S value, the specific heat works out to 0'242, which is in 

 close agreement with my own value, 0'2422. The above criticism applies not only to 

 the work of REGNAULT, but also to most of the more recent measurements of the 

 specific heats of gases at constant pressure, and it is possibly owing to this fact that 

 these measurements give values for the specific heats which are so much lower than 

 my own values and the theoretical values. 



(24) The Specific Heat of Carbon Dioxide. The carbon dioxide was supplied by 

 the manufacturers in iron bottles, each bottle containing about 20 Ib. of liquid CO 2 . 

 A preliminary test showed that the gas contained about 0'5 to 0*7 per cent, of air and 

 about 0'5 to 07 per cent, of water vapour. It was consequently necessary to pass it 

 through the drying apparatus used in tho experiments on air, the KOH tower was, 

 however, replaced by a tower containing crystals of calc. spar, which served to collect 

 any acid fumes from the sulphuric acid rocker, without introducing any gas other 

 than carbon dioxide. The iron bottles when supplying the gas were fitted with an 

 automatic pressure reducer, which prevented the pressure in any part of the apparatus 

 attaining a dangerously high value. The T-piece A (fig. 2) was replaced by a large 

 glass tap, followed by a piece of rubber tubing which could be partially constricted 

 by a clip, thus all possibility of air diffusing back into the carbon dioxide was 

 prevented. The flow of gas into the pressure regulator could be roughly adjusted by 

 means of the tap, and more finely by means of the clip. It was impossible to get rid 

 of all the air in the C0 2 , but I found that, after a large quantity of gas had been taken 

 from a bottle, the remaining gas was practically pure, the percentage of air falling 

 from about - 6 per cent, in the case of a full bottle to about O'Ol per cent, in the case 

 of a bottle which was nearly empty. Unfortunately I had used a large quantity of 

 the gas before I discovered this fact, but I was able, however, to calibrate the fine 

 tubes with pure CO 2 taken from bottles which were nearly empty, and in the later 

 calorimetric experiments I adopted the following device. Two bottles, A and B, were 

 arranged so as to be able to supply gas in parallel. The automatic pressure reducer 

 was fixed to the tube joining the bottles. A was half empty and B was full. Gas 

 was passed through the apparatus from B until the temperature conditions were 



