top board of the old breast was first removed, then the material was
carefully worked out for the length of the poling. The latter was then
placed, with the rear end resting over the hood and the forward end
forced as far as possible into the undisturbed material. When two or
three polings had been placed, a breast board was set. After several
polings were in position, their forward ends were supported by some form
a cantilever attached to the hood. Plate LXIX shows one kind of
supports. In this way all the soft material was excavated down to the
rock surface, and the roof, sides, and face were sheeted with timber. In
shoving, the polings in the roof and sides were lost. It was found that
the breast could usually be advanced 5 ft. with safety. The fixed hood
made it possible to set the face about 7 or 8 ft. in front of the
cutting edge without increasing the length of the polings. This distance
was ample for two shoves, and was generally adopted, although a great
many faces were set for one shove only.

Fixed hoods were substituted for those of the sliding type, originally
placed on Shields _B_ and _D_ at Manhattan, at about the time the latter
encountered the rock at the reef.

In placing the polings and breasting, all voids behind them were filled
as far as possible with marsh hay or bags of sawdust or clay. To prevent
loss of air in open material, the joints between the boards were
plastered with clay especially prepared for the purpose in a pug mill.
The sliding extensions to the floors of the working compartments were
often used, in the early part of the work, to support the timber face or
loose rock, as shown in Fig. 1, Plate LXVIII. At such times the front of
the extensions was held tightly against the planking by the pressure of
the floor jacks. While shoving, the pressure on the floor jacks was
gradually released, allowing the floors to slide back into the shield