produced by the Explosion of Hydrogen and Oxygen. 477 
clearly marked, even in the cases where special care had 
been taken to reproduce the same conditions as far as possible. 
It was thought that this might be partly due to the explosion 
of the gas taking place to a large extent outside the field of 
force, for the unexploded gases were estimated to amount to 
between 5 and 10 per cent. of the gases originally present ; 
and it will be seen, on referring to fig. 2, that the spark 
which originated the explosions was entirely outside the 
cylinder. “It also seemed possible that moisture caused by 
an explosion might affect the subsequent experiments. With 
the object of testing these points, a new apparatus was 
devised. The arrangements shown in fig. 1 were precisely 
the same as before, but the apparatus of fig. 2 was replaced 
by a brass cylinder 32°2 cm. long and 3°75 em. internal 
diameter, closed by ebonite ends, to which the wire con- 
stituting, as before, the inner cylinder was directly fixed. 
The diameter of this wire was 3°18 mm. The tubes and 
drying-bottles external to the cylinder were abolished, and 
instead new arrangements were substituted for exploding the 
gas well within the field of force and for drying the cylinder 
after explosion. The gas was exploded by means of a thin 
platinum wire placed within the cylinder and soldered to 
two brass rods which protruded through one of the ebonite 
ends of the cylinder, and which could be connected to the poles 
of a battery. The platinum wire could thus be heated to 
incandescence and the gas could be exploded. Practically 
all the gas which exploded with this arrangement lay within 
the field of force. 
It was verified that when no explosion occurred the heating 
of the wire produced no electrical effect which could be 
measured by the ballistic galvanometer. 
The cylinder was dried after an explosion by warming it 
and passing dry air through it. This was done by means of 
two glass tubes fastened by air-tight joints to the ebonite 
ends of the cylinder and communicating by means of pieces 
of pressure-tubing, which could be clipped, one with tubes 
containing calcium chloride and phosphorus pentoxide to 
dry the air as it entered, the other with a water-pump. The 
latter tube was disconnected from the pump before making 
an experiment and joined to the apparatus of fig. 1, so that by 
its means the cylinder was filled with air or hydrogen and 
oxygen as might be required in the ordinary way. It was 
hoped in this. way, by confining the explosion within the 
field of force, and by drying the cylinder after each expe- 
riment, to secure agreement between experiments made under 
