» 



Vapour-Densities of some Carbon Compounds. 497 



carot'ully fractionated so as to ensure a constant boilino-point, 

 l)oilino; under correspondingly reduced pressures. The ar- 

 ranoement did not materially differ from that described in 

 the Trans. Cheni. Soc. 1885 ; all the joints were sealed so 

 that no leakage occurred, and teuiperature could be recovered 

 from day to day within 1 100 of a degree. 



As the standard of eom])arison for the densities of the sub- 

 stances mentioned above was oxvgen, its density at these 

 temperatures was required. Instead of a direct determination 

 of the densities of oxygen at these temperatures, however, 

 the coefficient of expansion of the gas between 0° and 130° 

 was directly measured with an imknown weight ; it was found 

 to be 1/272*52 = 0*00o6694. The mean of the determinations 

 of the weight of one litre of oxygen at 0^ by Regnault, Jolly, 

 Leduc, and Ravleigh was accepted as the basis of calculation, 

 namely, 1-42961. 



As the coefficient of expansion of oxygen was determined 

 by a method practically identical with that emploj^ed for the 

 measurements of densities, it will be more convenient to defer 

 the description of these experiments until an account of the 

 apparatus has been given. 



Description of Density-appjaratus, 



The apparatus consists essentially of the volume-tube and 

 pressure-gauge A and B (fig. 1, p. 498), which are connected by 

 a short piece of thick-walled india-rub ber tubing. The volume- 

 tube A consists of a glass tube of about 20 mms. diameter 

 and about 780 mms. in length, constricted at its lower end, 

 near which is attached the side-tube a. The top is blown 

 into a cap h of the form shown in the figure ; this serves the 

 double purpose of trapping the weighing-bulb after its inser- 

 tion, and also of retainino- a small globule of mercurv. The 

 bulb is thus prevented from floating on the surface of the 

 mercury in A, and from breaking off the glass points 1, 2, 

 3, (fcc. The globule of mercury gives off' vapour when the 

 tube is heated, and the space rapidly becomes saturated with 

 mercury -vapour : experiments have shown that if this pre- 

 caution be omitted the pressure of mercury-vapour takes 

 many hours to rise to its maximum, on account of its great 

 density, and consequent slow rate of diffusion. 



The whole tube is inclosed in a jacket D, which contains 

 during the experiment the vapour of cblorobenzene ; and D 

 is attached to the apparatus for regulating the pressure under 

 which the cblorobenzene boils. The whole system of tubes 

 rests on its cork g in mercury, contained in a v'essel E, 



Phil. Mag. IS. 6. YoL 6. No. 34. Oct. 1903. 2 K 



