suppose a hole to be broken through the skin under water. The water at once passes into the interior in quantities depending upon the area of the hole and the depth it is below the water-level. A very simple rule expresses the initial rate of inflow.

Let A = area of the hole (in square feet).

d

the depth below water in feet (taken about the centre of the hole will be near enough for practical purposes).

Then, if v = velocity of inflow of the water in feet per second,

v2 = 64d (approximately); and v = 8 √d;

so that, immediately after an accident, the volume of water passing into the vessel in each second

= 8√d × A (cubic feet).

Suppose, for example, the hole is 2 square feet in area, and has its centre 12 feet under water:

v = 8 √12 = 273
273 feet per second.

Water flowing in per second = 273 × 2 = 55 cubic feet.
If the vessel floats in sea-water,

Tons of water flowing in per second = 55 ÷ 35 = 1.58. Similarly, for any other depth or area of hole in the

bottom of a ship, this rule will enable the rate of inflow to be determined very nearly.

Reverting to Fig. 10, it is obvious that, if the water can find free access to every part of the interior-which would be true if there were no partitions forming watertight compartments the ship must sink unless the power of her pumps is sufficient to overcome the leak; or some means is devised for checking the inflow, by employing a sail, or a mat, or some other "leak-stopper"; or the total unoccupied space in the interior is less than the reserve of buoyancy, a condition not commonly fulfilled. Very little reflection will show that it is hopeless to look alone to the pumps to overcome leaks that may be caused by collision, ram attacks, or torpedo explosions; the area of the holes broken in the skin admitting quantities of water far too large to be thus dealt with. Hence attention is directed to two other means of safety: the first, minute watertight subdivision of the interior of the ship, to limit the space to which water can find access; the second, the employment of leakstoppers, which can be hauled over the damaged part, and made to stop or greatly reduce the rate of inflow. This latter is a very old remedy, Captain Cook having used a sail as a leak-stopper during his voyages, and many ships having been saved by similar means. It has acquired renewed importance of late, and various inventors have proposed modifications of the original plan, but all these are based upon the old principle of "stopping" the leak. Such devices are not embodied in the structure or design of the ship, but form simply part of her equipment; whereas watertight subdivision is a prominent feature in the structure of a properly constructed modern iron ship. It will be well, therefore, to sketch some of its leading principles. In doing so, we shall, for the sake of simplicity, make use of boxshaped vessels for purposes of illustration; but the conclusions arrived at will, in principle, be equally applicable to less simple forms, like those of ships.

There are three main systems of watertight subdivision: (1) by vertical athwartship bulkheads; (2) by longitudinal bulkheads; (3) by horizontal decks or platforms. Besides these there is the very important feature of construction known as the "double bottom," the uses of which will be described further on. In Figs. 10 and 11 the hole in the skin, admitting water to the hold, is supposed to lie between two transverse bulkheads (marked ab and ce) which cross the ship and form watertight partitions rising to some height above the load-draught line (WL) and terminating at a deck marked "Main Deck." The great use of these bulkheads will be seen if attention is turned to Fig. 11, which represents the condition of the box-shaped vessel after her side has been broken through. The vessel has sunk deeper in the water than when her side was intact; and it is easy to determine what the increase in draught has been when one knows the volume (fgeb, in Fig. 10) of the damaged compartment, as well as the volume in that space which is occupied by cargo, or machinery, or other substances. To simplify matters, suppose this compartment to be empty; and assume the length ac to be one-seventh of the total length AA: then the volume fgeb will be about oneseventh of the total displacement; and when this compartment is bilged and filled with water up to the height of the original water-line WL, one-seventh of the original buoyancy will be lost. In fact, the compartment between the bulkheads no longer displaces water; in it the water-level will stand at the height of the surface of the surrounding water; and since the weight of the ship remains constant, the lost buoyancy must be supplied by the parts of the ship lying before and abaft the damaged compartment. For this

reason we must have

original water-line area × increase in draught
=x displacement

=

× original water-line area × original draught.
Increase in draught = original draught.

This very simple example has been worked out in detail because it illustrates the general case for ship-shape forms. The steps in any case are:—

(1) The estimate of loss of buoyancy due to water entering a compartment; this loss being equal to the part of the original displacement which the damaged compartment contributed, less the volume in the compartment occupied by cargo, &c.

(2) The estimate of the increased draught which would enable the still buoyant portions of the vessel to restore the lost buoyancy if the entry of water were confined to the damaged compartment.

And to these, in practice, must be added—

(3) The change of trim (if any) resulting from filling the damaged compartment.

Reverting to Figs. 10 and 11, it will be obvious that, if the transverse bulkheads ab and ce did not rise above the original water-line WL, more than one-sixth of the original draught, they would be useless as watertight partitions; because, when the compartment was bilged, their tops would be under water before the increase of draught had sufficed to restore the lost buoyancy. And when their tops are under water (unless the deck at which the bulkheads end forms a watertight cover to the compartment), the water is free to pass over the tops, or through hatchways and openings in the deck, into the adjacent compartments, thus depriving them also of buoyancy, and reducing the ship to a condition but little better than if she had no watertight partitions in the hold. Fig. 12 illustrates this serious defect. The main deck at which the transverse bulkheads ab and ce end is lower than in Figs. 10 and 11, all other conditions remaining unchanged; and consequently, when the compartment is bilged, the water can pour over the tops of the bulkheads into the spaces before and abaft.

Hence this practical deduction. Watertight transverse bulkheads can only be efficient safeguards against foundering when care is taken to proportion the heights of their tops

above the normal load-line to the volumes of the compartments; or else to make special provisions for preventing water from passing into adjacent compartments by means of watertight plating on the decks at which the bulkheads end, in association with watertight covers or casings to all hatchways and openings in the decks. Unfortunately, in iron merchant ships, anything approaching to efficient subdivision by transverse bulkheads is commonly, wanting; and in some cases, where the number of the compartments has been sufficiently great to secure a fair degree of safety, the heights to which the bulkheads have been carried have

not been sufficient to ensure efficiency when the vessels were fully laden.

A vessel would ordinarily be considered well subdivided if she would keep afloat with any two compartments filled simultaneously. This was the recommendation of the council of the Institution of Naval Architects in 1867; but in the vessels of the Royal Navy it is not unusual to find the subdivision so minute that from three to six of the largest compartments may be simultaneously filled, without bringing the tops of the bulkheads under water, or allowing water to pass into compartments adjacent to those filled.

The midship compartments of a ship are usually the largest, and claim most attention; but those near the extremities are also important, because, although their