ON GASEOUS EXPLOSIONS. 967 
a sharp detonation. I collected and analysed some of ib: ib was 0°7 
oxyhydrogen gas, the residue nitrogen, with a trace of oxygen.’ 
This experiment, and another one of Grove’s, in which the decompo- 
sition of water was produced by dropping molten platinum into it, 
seemed to have served as the starting-point of Deville’s experiments. 
Deville published many of his results for the first time in two 
lectures delivered before the Paris Chemical Society on March 18 and 
April 1, 1864,’ though much of his work had been done considerably 
earlier. The papers do not appear in the ordinary publications of the 
Society, but in a supplementary set of volumes of ‘ Lecons,’ which seem 
to be comparatively rare, neither the Royal Society’s nor the Chemical 
Society’s libraries possessing a copy. 
Deville appears to have realised the necessity for removing the dis- 
sociated products from further reaction, and in his experiments three 
methods were employed for the separation of the dissociated material from 
the unaltered gas, viz. :— 
~ (1) By means of diffusion. 
(2) By admixture with an inert gas. 
(3) By very rapid cooling from a high temperature. 
Preliminary Diffusion Experiments.—Before proceeding to describe 
the first experiments made by the diffusion method it will be well to 
mention some results obtained by Deville on the flow of gases through 
porous walls at ordinary temperatures, as they have considerable 
bearing on the later experiments. . 
(1) If a fairly rapid current of hydrogen be passed into a tube of 
unglazed porous ware, and thence to a collecting vessel, it is found that 
instead of hydrogen practically pure air is collected. Analyses gave the 
following as an average composition :— 
Oxygen . : : 4 d ae 
Nitrogen . ; $ , mo 
Hydrogen. ; : , . (trace) 
Therefore hydrogen passes completely out through the walls of the tube, 
and air is absorbed in spite even of an excess of pressure amounting in 
some cases to several centimetres of mercury. 
Diffusion Apparatus.—Deville then proceeds to describe his cele- 
brated diffusion apparatus— which is shown diagrammatically, 
Fig. 9—which he used in so many experiments. An internal tube of 
unglazed biscuit—or earthenware—passes centrally through a larger and 
shorter tube of glazed porcelain, the ends of the latter being closed by 
rubber stoppers or corks. The gas to be studied is passed through the 
annular space between the tubes. The apparatus can be heated along 
its middle portion by a coke furnace to a temperature of from 1100° to 
1300° C.? 
Hydrogen and CO: Experiments.—Employing this apparatus cold, 
Deville makes the following experiment: A fairly rapid current of CO, 
34 Lecons sur la Dissociation professées devant la Société Chimique de Paris. 
2 No data are given by which this temperature estimate can be even approxi- 
mately verified, but no blast appears to have been used, hence 1500° C. would bea 
probable maximum. ; 
