the Specific Heat of Air under constant Volume. 431 



the magnitude in question is not far removed from that calcu- 

 lated by Laplace. The observers mentioned, as is well known, 

 subjected an enclosed volume of air to a sudden change of den- 

 sity by connecting it for a very short time with a large reservoir 

 of air under a known pressure (mostly that of the atmosphere), 

 and then measured the change in temperature. But as even the 

 most delicate thermometer is too slow to follow rapid alterations 

 of temperature, the enclosed air was itself used as a thermometer 

 by observing the change in pressure which it experienced when 

 the original temperature was restored. As the change in pres- 

 sure was small, it was measured by a column of water instead of 

 by one of mercury. 



Nothing can be urged against the principle of this method. 

 The doubts which might arise from the evaporating water would 

 be removed by the use of sulphuric acid. The question is 

 whether the two assumptions can in practice be simultaneously 

 realized — first, that the duration of the communication with 

 the atmosphere is sufficiently short to justify the neglect of 

 the equalization of temperature which takes place during this 

 time, and, secondly, that after so short a communication the 

 pressure in the receiver is at first exactly equal to the atmospheric 

 pressure. Doubts as to the simultaneous fulfilment of these two 

 conditions will arise at the outset; an empirical proof has not 

 been afforded by the observers. To remove this objection, a 

 knowledge of the condition immediately after the change in pres- 

 sure must be sought in some other way; and this is afforded 

 by not merely remarking the total change in pressure from the 

 moment of rarefaction or condensation of the mass of air to the 

 final restoration of the original temperature, but commencing the 

 observation very soon after the primary change in pressure. 

 From the course observed, the law of the equalization of tempe- 

 rature will be ascertained by which the condition corresponding 

 to the time zero is to be calculated. The duration of the com- 

 munication between the receiver and reservoir may be as small as 

 the mechanism of the apparatus permits. Indeed when once this 

 duration is known it can be readily allowed for in the calculation. 

 Such a method was not applicable so long as the pressure was 

 to be measured by a column of liquid ; for in order to diminish 

 capillary actions a tolerably wide tube must be used, and there- 

 fore, owing to the initial oscillations of the column of liquid, 

 the most important time for observation is lost. 



The metallic barometer now constructed in great perfection fur- 

 By a rather arbitrary correction this mean value was brought into accordance 

 with that which followed from the then known velocity of sound. The 

 experiments of Gay-Lussac and Welter, of which Laplace gives an example, 

 do not appear to have been published. Compare also Dulong, Poggen- 

 dorff's Annalen, vol. xvi. p. 454. 



