t number of posts had been placed, the blocking on which the
caisson had rested was knocked or blasted out, and the rock underneath
was excavated. The blocking was then re-set at a lower elevation. The
posts under the cutting edge were then chopped part way through and the
air pressure was lowered about 10 lb., which increased the net weight to
more than 4,000,000 lb. The posts then gradually crushed and the
caissons settled to the new blocking. The tilt or level of the caisson
was controlled by chopping the posts more on the side which was desired
to move first.

The caisson nearly always carried a very large net weight, usually about
870 tons. The concrete in the walls, which was added as the caisson was
being sunk, was kept at about the elevation of the ground. There was
generally a depth of from 5 to 20 ft. of water ballast on top of the
roof of the working chamber. The air pressure in the working chamber was
usually much less than the hydrostatic head outside the caisson. For
example, the average air pressure in the south caisson during January,
1906, was 16-1/2 lb., while the average head was 62.5 ft., equivalent to
27 lb. per sq. in. Under these conditions, there was a continued but
small leakage into the caisson of from 15,000 to 20,000 gal. per day.

In the rock the excavation was always carried from 2 to 5 in. outside
the cutting edge. As soon as the cutting edge was cleared, bags of clay
were placed under it in a well-tiered, solid pile, so that when the
caisson was lowered the bags were cut through and most of the clay, bags
and all, was squeezed back of the cutting edge between the rock and the
caisson.

Table 1 shows the relation of the final position of the caissons to that
designed.

The cost of rock excavation in the caisson was $4.48 per cu. yd. for
labor and $10.54 for top charges.

The bottom of the shaft is an inverted concrete arch, 4 ft. thick,
water-proofed with 6-ply felt and pitch. As soon as the caisson was down
to its final position and the excavation was completed, concrete was
deposited on the uneven rock surfaces, brought up to the line of the
water-proofing, and given a smooth 1-in. mortar coat. The felt was stuck
together in 3-ply mats on the surface with hot coal-tar pitch. These
were rolled and sent down into the working chamber, where they were put
down with cold pitch liquid at 60 deg. Fahr. Each sheet of felt overlapped
the one below 6 in. The water-proofing was covered by a 1-