92 M. Achille Cazin on Internal Work in Gases. 



In fig. 4 we Lave only shown six curves, to avoid confusion. 

 In figs. 4, 6, 7, and 8 a division represents a second in the di- 

 rection of the abscissa?, and 1 centim. in the direction of the 

 ordinates. 



It is plainly seen from the figure how the curves run as the 

 time 6 increases. For a small value of 6 (curve A) the line falls 

 rapidly for three seconds until h— — 136 millims. ; then it rises 

 more slowly, passes the abscissse-axis until h = + 16 millims., and 

 again descends very slowly till A=0. 



For a greater value of 6 (curve E) the line falls for two se- 

 conds, until h= — 41 millims.; then it rises at first parallel to 

 the preceding, attains h — + 30 millims., and further on coincides 

 with the curve A. 



The separation of these curves evidently depends on the liquid 

 in the manometer not instantly indicating the pressure of the 

 gas, and on the level being more or less behind according as it 

 must run a greater or less length before reaching the position of 

 equilibrium. 



Let us suppose the variation of the real pressure of the gas 

 to be represented by the line g hi (fig. 5). When the valve of 

 the manometer is opened the liquid would indicate h= — fg if 

 it could move instantly, and if the pressure of the gas remained 

 invariable. But whilst the liquid moves, the pressure of the gas 

 increases; and when the level indicates h=—kb, the pressure 

 corresponds to the point /. The point b is evidently higher 

 than the point g, and it is also lower than the point / on account 

 of the impulse the liquid receives. The level afterwards rises until 

 the manometric column and the pressure of the gas are in equi- 

 librium, which can only take place after a certain time. 



Let the valve be opened after a time of > of. Instead of 

 instantly indicating h— —f ] g\ it only indicates h = —Mb'; and 

 the distance of the point b' from g 1 and /' is less than before, 

 because the impulse and the initial rarefaction f'g 1 have been 

 less ; moreover the duration f ] k' of the depression is also less 

 than fk, because k' b' <kb. Hence the curve b'c 1 approaches 

 closer tog hi than the curve b c and will join it sooner. 



In the following paragraph several observations will confirm 

 this interpretation. In fig. 4 the line X X is traced, with which 

 the experiments of Series I. agree as well as possible. It repre- 

 sents approximately the law of the variable pressure of the gas 

 during the expansion from 0*5 second. Several other experi- 

 ments, in which the value of 6 was less, have shown that the 

 curves differ very little from the curve A until 0=O'l second. 

 I have not made observations of shorter duration. 



If follows from this, that if op } (fig. 5) denote the positive 

 excess of the pressure of the compressed gas in reservoir A above 



