TRANSACTIONS OF SECTION B. 305 



allowed the method to retain all its accuracy and simplicity, when he proposed to 

 heat a bulb containing air in the same bath, and of the same size, as that con- 

 taining the substance. After being heated, the two bulbs were then sealed at the 

 same temperature. Bunsen, in 1867, employed an air-bath, similar in principle to 

 those of Mitscherlich and Regnault, but heated it by a very elaborate arrano-e- 

 nient of gas burners. He also simplified the calculation by taking care that all the 

 tubes were of exactly the same weight and same size. He did not seal the tubes, 

 but closed them by glass caps, lined with india-rubber and fitted with glass 

 plugs. Dumas, in cases where the vapour rendered the outlet difficult to seal, 

 used globes fitted with ground stoppers. 



For the Dumas process at high temperatures, Deville and Troost, in 1860, 

 recommended heating the bulb in a specially constructed furnace, in the vapours 

 of substances having high but definite boiling-points, such as mercury, sulphur, 

 zinc, or cadmium. For temperatures above the boiling-point of sulphur, they used 

 porcelain globes. For temperatures up to that point, the smaller and more com- 

 pact apparatus devised by Greville Williams answers admirably. 



Roscoe, last year, in determining the vapour-densities of the chlorides of lead 

 and thallium, used porcelain globes of 300 c.c. capacity, heated in a muffle, but 

 determined the temperature by the method of specific heat, a large piece of 

 platinum being employed for the purpose, and checked the result by the simul- 

 taneous determination of the vapour density of mercury. 



For working at a reduced pressure, Regnault proposed partially exhausting 

 the bulb by means of an air-pump during the experiment ; when the desired tem- 

 perature was reached, it was sealed off at a point where the neck had been nar- 

 rowed to a convenient size. In 1876, Habermann gave a complete diagram of the 

 apparatus, replacing the air-pump by a Bunsen pump ; but although he made no 

 alteration in the method, still it was referred to by Sommaruga as Habermann's. 



Various experiments had been performed on vapours mixed with air, but the 

 main point in Playfair and Wanklyn's method (1861) consisted in stopping the 

 supply of vapour before the bath in which the bulb was being heated had attained 

 its maximum temperature. 



Natanson, in 1855, in order to use the Gay-Lussac method up to a temperature 

 of 300°, heated the upper part of the tube by means *of charcoal in a cylindrical 

 furnace, and determined the temperature by thermometers suspended in the air- 

 space between the graduated tube and the inner tube of the heating apparatus. In 

 correcting for the tension of mercury-vapour he used Avogadro's tables. 



Greville Williams, in 1857, wishing to make some determinations at varying 

 pressures, devised the following method : — The graduated tube is, after it has been 

 filled and the bulb has been inserted, screwed by means of a nipple cemented to 

 the bottom into an orifice in the top of a small metallic cistern into a second 

 orifice in which a long open glass tube is fitted. Into this tube mercury is poured 

 until the required pressure is obtained. To reduce the pressure the excess of 

 mercury is allowed to escape by a tap in the side of the cistern. The. whole is 

 heated in a water- or oil-bath. 



In Regnault's apparatus for the same purpose, the two tubes are fastened to 

 the bottom of the water-bath and are connected by a T piece, which is closed by a 

 three-way cock of special construction. 



For determinations up to 150° Greville Williams's compact modification consists 

 in replacing the large vessel of mercury, and the open glass cylinder by a cylinder 

 closed and rounded at the lower extremity, so as to resemble a large test-tube. 

 This is then filled to a depth of 50-60 m. m. with mercury, and above that with 

 water or oil to a convenient height. The graduated tube is filled and the bidb 

 inserted over the mercurial trough ; it is then immersed' in the large tube by 

 means of a rod having at the end a small cup containing mercury. The large tube 

 may be supported on wire gauze and heated by a Bunsen burner, or may be 

 placed in a shallow oil-bath. 



Schiff, in 1862, proposed steadying and manipulating the graduated tube by 

 means of a loaded handle, which was secured to its upper extremity by spring clips. 



Grabowski, in 1866, replaced the charcoal 'furnace of Natanson by a very much 



1879. x 



