," I cried as he rushed up to us, "and a deadlier
scheme of Dame Nature's to frighten and murder poor humanity I have
never seen."

As I spoke I lit the lamp again and held it just above a tiny fissure in
the rock. It was at once extinguished.

"What is it?" said Bainbridge, panting with excitement.

"Something that nearly finished _me_," I replied. "Why, this is a
natural escape of choke damp. Carbonic acid gas--the deadliest gas
imaginable, because it gives no warning of its presence, and it has no
smell. It must have collected here during the hours of the night when no
train was passing, and gradually rising put out the signal light. The
constant rushing of the trains through the cutting all day would
temporarily disperse it."

As I made this explanation Bainbridge stood like one electrified, while
a curious expression of mingled relief and horror swept over Henderson's
face.

"An escape of carbonic acid gas is not an uncommon phenomenon in
volcanic districts," I continued, "as I take this to be; but it is odd
what should have started it. It has sometimes been known to follow
earthquake shocks, when there is a profound disturbance of the deep
strata."

"It is strange that you should have said that," said Bainbridge, when he
could find his voice.

"What do you mean?"

"Why, that about the earthquake. Don't you remember, Henderson," he
added, turning to the Inspector, "we had felt a slight shock all over
South Wales about three weeks back?"

"Then that, I think, explains it," I said. "It is evident that Pritchard
really did climb the rocks in a frantic attempt to escape from the gas
and fell back on to these boulders. The other man was cut down at once,
before he had time to fly."

"But what is to happen now?" asked Bainbridge. "Will it go on for ever?
How are we to stop it?"

"The fissure ought to be drenched with lime water, and then filled up;
but all really depends on what is the size of the supply and also the
depth. It is an extremely heavy gas, and would lie at the bottom of a
cutting like water. I think there is more here just now than is good for
us," I added.

"But how," continued Bainbridge, as we moved a few steps from the fatal
spot, "do you account for the interval between the first death and the
second?"

"The escape must have been intermittent. If wind blew down the cutting,
as probably was the case before this frost set in, it would keep the gas
so diluted that its effects would not