experienced by Air in rushing through small Apertures. 491 



decreases as the temperature is augmented. Now according to the 

 theoretical views on this subject brought forward in the papers on 

 " Camot's Theoiy/' and " On the Dynamical Theory of Heat," 

 already referred to, a cooling effect was expected for low tempe- 

 ratures ; and the amount of this effect was expected to be the 

 less the higher the temperature ; expectations which have there- 

 fore been perfectly confirmed by experiment. But since the excess 

 of the heatof compression above the thermal equivalent of the work 

 was, in the theoretical investigation, found to diminish to zero* 

 as the temperature is raised to about 33° Cent., or 92° Fahr., 

 and to be negative for all higher temperatures, a heating instead 

 of a cooling effect would be found for such a temperature as 

 171° r., if the data regarding saturated steam used in obtaining 

 numerical results from the theory were correct. All of these 

 data except the density had been obtained from Regnault's very 

 exact experimental determinations ; and we may consequently 

 consider it as nearly certain, that the true values of the density 

 of saturated aqueous vapour differ considerably from those which 

 were assumed. Thus, if the error is to be accounted for by the 

 density alone, the fact of there being any cooling effect in the air 

 experiments at 171° Pahr. {77° Cent.) shows that the density of 

 saturated aqueous vapour at that temperature must be greater 

 than it was assumed to be in the ratio of something more than 

 1416 to 1390, or must be more than 1*019 of what it was assumed 

 to be : and, since the experiments render it almost if not abso- 

 lutely certain, that even at 100° Cent, air rushing through a 

 small aperture would produce a final cooling effect, it is probable 

 that the density of steam at the ordinary boiling-point, instead 



of being about .. „^„ >- ? as it is generally supposed to be, must 



Ibyo'o 1430'6 

 be something more than -. o^.^ - of this ; that is, must exceed 



l645' 



With a view to ascertain what effect would be produced in the 

 case of the air rushing violently against the thermometer-bulb, 

 the leather diaphragm was now perforated with a fine needle, and 

 the bulb ])laced on the orifice so as to cause the air to rush 

 between the leather and the sides of the bulb. With this 

 arrangement the following results were obtained : — 



* Sec the table in § 51 of the Account of Carnot's Theory, from which 

 it appears that the clement tabulated woulil have the value 1390, or that 

 of the meehauical equivalent of the thermal unit, at about 'S6° Cent. 



