Intelligence and Miscellaneous Articles. 323 



from the value p, to the value p 2 . The quantity p, is calculated from 

 p lt and p 2 from p 3 . 



But it is necessary to determine whether the closing of the stop- 

 cock has been instantaneously effected at a given moment ; this is 

 the essential point of the method. For this purpose a voltaic circuit 

 is arranged, containing an electromagnet ; and the movement of the 

 stopcock determines the closure of this circuit at the moment when 

 the orifice is opened, and afterwards its rupture at the moment when 

 the orifice is closed. The electromagnet moves a pencil which leaves 

 a trace upon a sheet of paper which moves at a known rate ; from 

 the length of this trace is deduced the duration T of the opening of 

 the stopcock. A series of experiments comprises those in which we 

 make T vary without changing either p x or p 2 . This series is re- 

 presented by a line having for abscissae the values of T, and for ordi- 



nates the values of— 2 . The ordinates vary according to a cer- 



tain law as long as T is below the duration 6 which corresponds to 

 the instant sought, and according to a different law when T is above 

 that duration. The curve is then formed of two very different 

 branches, the point of intersection of which is determined graphi- 

 cally. The abscissa and the ordinate of that point give the duration 

 6 of the complete flow and the value of p 3 which we want. 



The lower branch was virtually a right line, nearly parallel to 

 the axis of abscissas, which indicated a very slow heating-action on 

 the part of the sides. Hence is deduced a correction giving the supe- 

 rior limit of the value that p 3 would have assumed if the sides had 

 been impervious to heat. The feebleness of the thermal action of 

 the sides is remarkable ; we may attribute it to the formation of a 

 gaseous sheath varnishing the sides. 



First mode of observation. — p x — p 2 is small; it is measured by 

 means of an oil manometer, whose branches communicate respec- 

 tively with the reservoir A (29 litres) and the reservoir B (520 

 litres), and by an open-air manometer communicating with one of 

 the reservoirs. Similarly p 3 —p 2 is measured. All necessary pre- 

 cautions are taken so that the gas enclosed in the manometers 

 may not by its motion disturb the expansion. In this way I 

 found that the quantity 



log/?,— log p 

 was constant for air and carbonic acid when p 2 varied from 1 to 5 

 atmospheres. I did not raise the pressure higher, because the resist- 

 ance of the sheet-iron reservoir B imposed this limit. Carbonic 

 acid presented the oscillation that I described in 1862. 



I concluded from this that, if one of these gases expanded in a 

 space impervious to heat without acquiring an appreciable velocity, 

 the law of expansion would be represented by the known formula 

 of Laplace and Poisson, »=Ap m 



A and m being two constants for the same gas; m = l # 41 for air, 

 and 1"29 for carbonic acid. 



This result is interesting as regards the mechanical theory of heat, 



