the Densities of Gases and Vapours. 343 



perature. The substances soon begin to boil, expel the air, 

 and escape by the tubes, which are very imperfectly closed by 

 the ball. When the apparatus is of the same temperature as the 

 muffle, it is withdrawn, and, after cooling, the weights of mercury 

 and of the substance respectively in the flasks A and B are 

 determined. 

 . Let 



P be the weight of mercury. 



P' that of the substance. 



8 the density of mercury, compared with air under the pres- 

 sure and at the temperature which prevailed in the muffle 

 when the apparatus was withdrawn. 



The density of vapour under the same conditions will be 



P 



. In my Cours elementaire de Chimie, 5th edit. vol. iv. p. 66, I 

 have given the arrangement of an apparatus analogous to that 

 which M. Mitscherlich has employed for substances boiling at 

 high temperatures; I simply endeavoured to obtain more equal 

 temperatures for the air-thermometer and the vapour-tube by 

 imparting a continual rotatory motion to the system of the two 

 tubes in the muffle in which it is heated, and which has several 

 metallic envelopes. This apparatus can be simplified and made 

 more convenient now that the use of gas prevails in labora- 

 tories. It consists of three tubes of wrought iron closed at one 

 end, and resembling gun-barrels • they are 50 ccntims. in length, 

 and 20 millims. internal diameter. Fig 17 represents the longi- 

 tudinal section of one of these tubes ; A B is the part which is 

 50 centims. in length. On each of two of these tubes is screwed 

 an additional piece B C of the same diameter, and on the second 

 tube a narrower tube, C D. On the third tube, which is intended 

 as a gas-thermometer, a single iron tube (fig. 18) B' C is screwed, 

 which is of almost capillary bore, and is terminated by a stopcock. 

 These three tubes, whose dimensions arc quite similar, fit upon 

 the same iron bar 1 1', of which figs. 19 and 20 give a cross 

 section, I. Fig. 19 shows by a section how the three iron 

 tubes A are arranged in reference to the central bar I. The bar 

 is longer than the iron tubes ; it is firmly fixed on two cast-iron 

 supports PP' (fig. 21), arranged so that the bar is exactly on the 

 notch. The three tubes are thus in a fixed position. This system 

 of the three tubes is surrounded by a cylinder E F of copper 

 or of sheet iron, which fits almost exactly, but so, however, 

 that the cylinder may be made to rotate rapidly about a hori- 

 zontal axis, 1 1'. A sheet-iron disc a b c (fig. 20), fitted on the 

 central bar 1 1, and almost exactly filling the cylinder E F, forms 



