lessen the risk of its getting out of hand and ramming the lock-gate, an accident which has occurred on the big locks that connect Lake Superior with Lake Huron.  So catastrophic would be such a mishap, releasing as it might this immense accumulation of water, that it seemed desirable at whatever expense to provide additional safeguards against it.  There are in the first place cross-chains, tightening under pressure, which may be drawn across the bows of a ship that threatens to become unmanageable.  Secondly, the lock-gates are doubled at the entrance to all the locks, and at the lower end of the upper lock in each flight.  And, thirdly, each flight of locks can be cut off from the lake by an “emergency dam” of peculiar construction.  It is essentially a skeleton gate, which ordinarily lies uplifted along the top of the lock-wall, but can be swung across, lowered, and gradually closed against the water by letting down panels.  In its ordinary position it lies high above the masonry—­conspicuous from some distance out at sea as a large cantilever bridge, swung in air.

Peculiar difficulties have been encountered in establishing the foundations of the locks.  The lowest of each flight are planted in deep morasses, and could only be settled by removing vast masses of estuary slime to a depth of 80 feet below sea-level.  The sea was cut off and a dredger introduced, which gradually cleared its way down to the bottom rock.  But the troubles which the American engineers will remember are those which have presented themselves in the Culebra cutting.  The channel is nine miles long.  Its average depth is between 100 and 200 feet, but at one point it reaches 490 feet.  The formation of the ground varies extraordinarily.  At some points it is rock; at others rock gives place to contorted layers of brilliantly colored earth which is almost as restless as quicksand.  Unfortunately, it is at places where the cutting is deepest that its banks are most unstable.  The sides of the lowest 40 feet of the excavation—­the actual water channel—­are cut vertically and not to a slope; in a firm formation this reduces the amount of excavation, but in loose material it must apparently have increased the risk of slides.  But, however this may be, slips on a gigantic scale were inevitable.  The cutting is an endeavor to form precipitous slopes of crumbling material under a tropical rain-fall:  it may be likened to molding in brown sugar under the rose of a watering-pot.  The banks have been in a state of constant movement, and are broken up into irregular shelves and chasms, so that at some points the channel resembles a natural ravine rather than an artificial cutting.  One thing is certain,—­that for some years to come the channel will only be kept open by constant assiduous dredging.  But it is, of course, easier to dredge out of water than to excavate in the dry.  The material excavated from the Culebra channel will aggregate nearly one hundred million cubic yards.  Some of it has been utilized in