unsel together. George was in favour of pumping out the
water from the top by powerful engines erected over each shaft, until the
water was mastered. Robert concurred in that view, and although other
engineers pronounced strongly against the practicability of the scheme
and advised its abandonment, the directors authorised him to proceed; and
powerful steam-engines were ordered to be constructed and delivered
without loss of time.

In the mean time, Robert suggested to his father the expediency of
running a drift along the heading from the south end of the tunnel, with
the view of draining off the water in that way. George said he thought
it would scarcely answer, but that it was worth a trial, at all events
until the pumping-engines were got ready. Robert accordingly gave orders
for the drift to be proceeded with. The excavators were immediately set
to work; and they were very soon close upon the sand bed. One day, when
the engineer, his assistants, and the workmen were clustered about the
open entrance of the drift-way, they heard a sudden roar as of distant
thunder. It was hoped that the water had burst in--for all the workmen
were out of the drift,--and that the sand bed would now drain itself off
in a natural way. Instead of which, very little water made its
appearance; and on examining the inner end of the drift, it was found
that the loud noise had been caused by the sudden discharge into it of an
immense mass of sand, which had completely choked up the passage, and
prevented the water from flowing away.

The engineer now found that there was nothing for it but to sink numerous
additional shafts over the line of the tunnel at the points at which it
crossed the quicksand, and endeavour to master the water by sheer force
of engines and pumps. The engines erected, possessed an aggregate power
of 160 horses; and they went on pumping for eight successive months,
emptying out an almost incredible quantity of water. It was found that
the water, with which the bed of sand extending over many miles was
charged, was to a certain degree held back by the particles of the sand
itself, and that it could only percolate through at a certain average
rate. It appeared in its flow to take a slanting direction to the
suction of the pumps, the angle of inclination depending upon the
coarseness or fineness of the sand, and regulating the time of the flow.
Hence the distribution of the pumping power at short intervals along