396 
the boiling point, the authors employ Gay Lussac’s process for 
taking the specific gravity of vapours. A slight modification is, how- 
ever, necessary. Previous to the introduction of the bulb contain- 
ing the weighed substance, dry hydrogen is introduced into the 
graduated tube and measured with all the precautions belonging to 
a gas analysis. It will be obvious that in the subsequent calculation 
the volume of hydrogen corrected at standard temperature and pres- 
sure must be subtracted from the volume of mixed gas and vapour, 
also corrected at standard temperature and pressure. 
When the substance will not bear heating to its boiling point, the 
authors employ a process resembling that of Dumas in principle, 
but differing very widely from it in detail. Dumas’ flask with 
drawn-out neck is replaced by two bulbs, together of about 300 cub. 
cent, capacity, joined by a neck, and terminating on either side in a 
narrow tube. One of the narrow tubes has some very small dilata- 
tions blown upon it (6), the other is merely bent (D). (See the 
drawing.) The apparatus, whose weight should not exceed 70grm., 
is weighed in dry air, then placed in a bath, being secured by a 
retort-holder grasping the neck joining the large bulbs C and C. 
The end A, projecting over the one side of the bath, is made to 
communicate with a hydrogen apparatus ; the end D passes through 
a hole in the opposite side of the bath, which is plugged up water- 
tight by means of putty. Dry hydrogen is transmitted through 
the whole arrangement, and escapes at D through a long narrow 
tube joined to it by a caoutchouc connecter. 
The bath is next filled with warm water until the bends a and a 
are covered. The connection with the hydrogen apparatus is then for 
a moment interrupted, to allow of the introduction of a small quantity 
of the substance at A. The substance, which should not more than 
half-fill the small bulb 6, is partially vaporised in the stream of hydro- 
gen, and in that state passes into the part GC. All the while, the 
