22 THE SCIENTIFIC PAPERS OF 



to the syphon. The oil bath is then gradually heated, and the 

 temperature observed. As soon as the boiling point of water 

 under the pressure in question is reached, the mercury piston will 

 move rapidly forward, until all the water is converted into steam. 

 The temperature continuing to increase, the piston will continue 

 its course more slowly upon the scale, where its progress is noted 

 from time to time, together with the temperature. 



The experiment is continued until the temperature reaches about 

 400", when the oil begins to boil. The gas flame is then with- 

 drawn, and the bath allowed to cool gradually. The observations 

 of the temperature and the position of the mercury piston are 

 continued until the steam contained behind it is recondensed. A 

 comparison between the two series of observations gives the 

 correct mean of the experiment, by which the effects of the 

 friction of the mercury piston, any possible slight leakage of 

 steam past it, and faults consequent on the slow transmission of 

 heat, are completely neutralized. 



The curve A on Fig. 3, Plate 4, has been drawn, expressing 

 the rate of expansion of isolated steam according to these ex- 

 periments. The results of nine separate experiments very nearly 

 coincide (as shown by the dotted lines, which give the ex- 

 treme variation in the experiments), except at the starting- 

 point, where the rate of expansion is so very great that it is 

 difficult to obtain correct observations ; changes in the barometer, 

 moreover, affect the curve in the vicinity of the boiling point. 

 To obviate the effect of these inaccuracies, the unit of volume in 

 laying down the curves from each of the nine experiments was 

 taken, not at the absolute boiling point, but at 250, where the 

 expansion had already assumed a definite course. 



The diagram also shows a straight line B, expressing the rate of 

 expansion of common air, which at first diverges greatly from the 

 hyperbolic curve of expansion of steam, although the asymptote 

 of the latter seems to run parallel to the former. The author 

 considers it therefore highly probable, " that the rate of expansion 

 of all gases may te expressed by one hyperbola, which starts from 

 the condensing point of the ffas," and that the apparently uniform 

 rate of expansion of the permanent gases may be accounted for by 

 their great elevation, at the ordinary temperature, above their 

 supposed boiling point, in consequence whereof the true curve 



