fashioned his slide accordingly, reducing the diameter of
the tube until he conceived it was quite safe. In about a fortnight the
experiments were repeated, in a place purposely made foul as before; on
this occasion a larger number of persons ventured to witness them, and
they again proved successful. The lamp was not yet, however, so
efficient as the inventor desired. It required, he observed, to be kept
very steady when burning in the inflammable gas, otherwise it was liable
to go out, in consequence, as he imagined, of the contact of the burnt
air (as he then called it), or azotic gas, which lodged round the
exterior of the flame. If the lamp was moved horizontally, the azote
came in contact with the flame and extinguished it. "It struck me," said
he, "that if I put more tubes in, I should discharge the poisonous matter
that hung round the flame, by admitting the air to its exterior part."
Although he had then no access to scientific books, nor intercourse with
scientific men, nor anything that could assist him in his investigation,
besides his own indefatigable spirit of inquiry, he contrived a rude
apparatus by which he tested the explosive properties of the gas and the
velocity of current (for this was the direction of his inquiries)
necessary to enable the explosive gas to pass through tubes of different
diameters. In making these experiments in his humble cottage at the West
Moor, Nicholas Wood and George's son Robert usually acted as his
assistants, and sometimes the gentlemen of the neighbourhood interested
in coal-mining attended as spectators.

These experiments were not performed without risk, for on one occasion
the experimenting party had nearly blown off the roof of the cottage.
One of these "blows up" was described by Stephenson himself before the
Committee on Accidents in Coal Mines, in 1835: "I made several
experiments," said he, "as to the velocity required in tubes of different
diameters, to prevent explosion from fire-damp. We made the mixtures in
all proportions of light carburetted hydrogen with atmospheric air in the
receiver, and we found by the experiments that when a current of the most
explosive mixture that we could make was forced up a tube 4/10 of an inch
in diameter, the necessary current was 9 inches in a second to prevent
its coming down that tube. These experiments were repeated several
times. We had two or three blows up in making the experiments, by the
flame getting down in