104 PROFESSOR H. B. DIXON ON THE RATE OF EXPLOSION IN GASES. 
the gas was driven from the mercury holder through one washing bottle containing 
strong aqueous potash and through two towers packed with lumps of caustic potash. 
The cyanogen mixtures were passed direct from the mercury holder into the explosion 
tube, the gases having been dried in preparation. 
The explosion tube was, in most of the experiments, a leaden pipe, 100 metres long 
and 9 mm. in diameter. With some of the more violent explosives— e.g., cyanogen 
and acetylene, a leaden tube of 6'5 mm. was used; and in the chlorine experiments 
straight wrought-iron pipes lined with glass were employed. 
The leaden tube was, in most cases, coiled on a drum, 2 feet in diameter, which 
was immersed in an iron vessel containing water. By heating the water and pumping 
dry air through the tube, the ap])aratus could be quickly dried after an explosion in 
which water was formed. For determinations of the rate of explosion of gases at 
100° C, this arrangement was also suitable. Two inconv-eniences arose from this dis¬ 
position of the pipe ; (1) the difficulty of determining its exact length, and (2) the 
uncertainty whether the explosion-wave was propagated along the axis of the coiled 
pipe, or whether it followed a shorter path nearer to the inside wall. 
It was found impossible to coil a small leaden pipe without lengthening it appreci¬ 
ably ; the outside of the tube was therefore measured after each coil was wound on 
the drum, and the length of the axis of the pipe calculated. After a series of experi¬ 
ments the coils were then unwound, by rolling the drum along a corridor, and the 
length was measured directly. The length so obtained did not vary more than an 
inch, or at most two, from that calculated on winding. 
To determine whether the flame was propagated centrally, or whether it took a 
shorter cut, measurements of the rate of explosion of samples of oxygen and hydrogen 
from the same mixture were made alternately, in a leaden pipe 9 mm. in diameter 
(1) coiled on the drum, and (2) lying straight on the floor. The experiments were 
thrice repeated, and no appreciable difference in the rate could be detected ; from 
which it may be concluded that in a pipe of this bore the wave-front travels quickest 
along the axis. Possibly in the experiments with tubes of a larger diameter the 
curvature may have slightly affected the results. 
In the experiments made under reduced pressure, the tube was filled from the holder 
in the ordinary way. The steel stopcock* near the second bridge was closed, and the 
gas was then sucked from the tube by a powerful pump, maintaining a vacuum of about 
26 inches of mercury. A gauge showed when the desired reduction of pressure was 
readied. The other stopcock was then closed, and the spark immediately passed. In 
the experiments made under increased pressure, the gases were driven into the tube 
from a strong reservoir, either by water or by mercury, according to the nature of the 
gaseous mixture. In all cases the gases were fired by an electric spark near one end 
* These steel stopcocks, made for me by Mr. J. J. Hicks, of Hatton Garden, have stood many hundreds 
of explosions witliont leaking. The gun-metal stopcocks employed at first were indented by the explo¬ 
sions and soon leaked. 
