is evenly balanced (taking the level of the outside water to be the natural level of the water inside), as the pressure upon the water exposed to the full atmosphere is 15 pounds upon each square inch of its surface, while that upon the same plane, but within the pipe, will sustain a column of water 2 feet high (weighing one pound) and 14 pounds pressure of air, making a total of 15 pounds, which is, therefore, an equilibrium of pressure over the whole surface of the water at its natural level.

If, in consequence of a second stroke of the pump, the air pressure in the pipe is reduced to 13 pounds per inch, the water will rise another 24 feet. This rule is uniform, and shows that the rise of a column of water within the pipe is equal in weight to the pressure of the air upon the surface of the water without; hence it is only necessary, to determine the height of a column of water that will weigh 15 pounds per square inch of area at the base, to ascertain how far a suction-pump will cause the water to rise. It must be understood, that the distance varies with the height above sea level, and also with the pressure of the atmosphere. At our level of the sea, the column of water that the atmosphere will support is about 33 feet in height, and a pump will "draw water" (as it is called) this distance; but the force which sends the water into the pump at this height is so diminished as to be almost balanced by its own weight; hence a liftingpump will deliver water very slowly, drawing it this distance.

To be reliable, the cylinder and piston should be in good order, all the joints perfectly air-tight, a check-valve be placed in the lower end of the suction-pipe; and even then the pumps should be run at a high speed. Pumps will give more satisfactory results when the lift is from 22 to 25 feet. There is hardly any limit to the distance a pump will draw water through a horizontal suctionpipe, provided the pipe is perfectly tight, and everything is so proportioned as not to cause undue friction.

The capacity of any pump may be determined by multiplying the area of the piston in inches by its stroke in inches, giving the number of cubic inches per single stroke; this divided by 231 (the

number of cubic inches in a standard gallon) will give the number of gallons per single stroke; but it must be remembered that all pumps throw less water than their capacity, the deficiency ranging from 20 to 40 per cent., according to the quality of the pump. This loss arises from the lift and fall of the valves, from inaccuracy of fit or leakage, and in many cases from there being too much space between the valves and piston, or plunger. The higher the valves of any pump have to lift to give the necessary opening, the less efficient the pump will be.

The power required to raise a given quantity of water a certain height may be computed by the following rule: Multiply the amount of water in gallons to be raised per minute by 8.35 lbs. (the weight of a gallon of water), and this product by the height, in feet, of the discharge from the point of suction; divide the result by 33,000, which will give the theoretical horse-power required to raise the amount of water to a certain distance. See table on page 522.

The quantity of water which any pump will lift, or discharge, may be estimated by multiplying the area of the piston by the speed; but this rule infers that the pump is fully supplied, and the water thoroughly discharged at every stroke.

Rule for finding the diameter of pump-plunger for any engine.When the pump-stroke is the stroke of the engine, the diameter of the steam-cylinder multiplied by 0-3 will give the proper diameter of pump-plunger.

Another rule. When the pump-stroke is of the stroke of the engine, the diameter of the cylinder multiplied by 42 will give the proper diameter of pump-plunger.

Diameter of pump-plunger should be equal to the diameter of the cylinder when the pump-stroke is the engine-stroke.

Diameter of pump-plunger should be equal to of the diameter of the cylinder when the pump-stroke is the engine-stroke. The velocity of water in pump-passages should not exceed 500 feet per minute. Pump-valves should have an area of the area of the pump.

Feed-pumps for condensing engines. For condensing engines, the diameter of the pump-plunger should equal 1·11 the diameter

of the steam-cylinder when the pump-stroke is half the enginestroke, and the diameter of steam-cylinder when the pump-stroke is the stroke of the engine.

Rule to find the diameter of the feed-pump ram.— Multiply the square of the diameter of the cylinder in inches by 0083. The product is the diameter of the ram in inches. All boiler feed-pumps, when working at ordinary speed, should be capable of discharging one cubic foot of water per horse-power per hour.

Rule for finding the necessary quantity of water per minute for any engine. Multiply the cubic space in the cylinder in inches, to which steam is admitted before being cut off, by twice the number of revolutions per minute, and divide the product by the comparative volume of steam at the pressure used; the quotient will be the cubic inches of water required per minute.

A circulating-pump is used to lift water from the sea and force it through the condenser. Such pumps are not always worked by the main engines, but sometimes are independent or worked by an independent auxiliary engine. See cut on page 352.

Although a pump will require to be in good condition to lift water 33 feet, it will with ease draw water on a level at 1000 feet (providing the pipes are all tight), and force it to any height that the machinery of the pump is capable of bearing.

The reason why pumps do not work is, either that the watersupply is exhausted, the pipes or pistons leak, or the valves prevented from seating. If the valves and connections of a pump are tight and in good order, and it is not located too high above the supply, there is no reason why it should not work.

Pumps become hot from two reasons, either they are placed too near the boiler, or the pump and check-valves leak, and allow the hot water to escape back from the boiler into the barrel of the pump, which has the effect of expanding the valves and preventing them from doing their work.

A boiler feed-pump, or injector, for any engine should be capable of supplying one cubic foot of water per horse-power per hour. Engines, in general, do not use that amount; in fact, the better

class of automatic cut-off engines will develop a horse-power with a water-consumption of from 25 to 30 lbs.; but it is always best to have the pump or injector sufficiently large, so that, in case the power should be increased, it may be equal to the demand.

An air-chamber is placed on a pump to cushion the water-piston, and relieve the jar that would be induced by the pump-piston striking against a solid column of water; but, to produce the desired effect, it must be perfectly air-tight, otherwise the air will escape. Even when the air-chamber is perfectly air-tight, they require to be frequently refilled, as in fast-running pumps and fire-engines the air becomes condensed. This may be done by stopping the engine or pump, opening a cock or valve that connects with it, and allowing the air to rush in. There is a very general impression among engineers and those having charge of fire-engines, that there is a vacuum in the air-chamber, and the remark is often heard that the pump or engine has lost its vacuum. This is a mistake, as there is no such thing as a vacuum in the air-chamber of a steam-pump or fire-engine. The air-chamber has lost its supply of air either by leakage or condensation. The result is the pump commences to work and labor.

A feed-pump pet-cock, or valve, is a small cock, generally placed on the barrel of the pump above the suction-valve, for the purpose of ascertaining whether the pump is working right or not.

Mud-boxes, strainers, or arresters should be attached to the extreme end of all lift-, suction,- bilge-, or circulating-pumps, for the purpose of arresting any matter that would be liable to choke the pump or prevent the valve from seating.

How to keep pipes and pumps from freezing. The only certain preventive is the removal of the water from them; consequently, in all cases provision should be made for turning it off during very severe nights. It must be observed, however, that merely shutting off the water is not sufficient; it must all be let out of the pipes. For this purpose a small tap or pet-cock should be placed above the main stop-cock, or the latter should be made with a vent, to allow the water to flow out when it is turned off.

Injectors.

The injector, though simple in design, modest in appearance, and diminutive in size, is, nevertheless, one of the most wonderful, important, and useful machines which the mechanical arts have ever presented to man. It consists of a slender tube, called the steam-tube, through which steam from the boiler passes to another or inner tube, called the receiving-tube. The latter tube conducts a current of water from the pipe into the body of the injector. Opposite the mouth of this second tube, and detached from it, is a third fixed tube, called the delivery-tube. This tube is open at the end facing the water-supply and leading from the injector to the boiler.

Its action is identical to that of the steam-jet, or blower-pipe in the chimney of the locomotive. The principle is, that steam being admitted to the inner tube of the injector, enters the mouth of a combining-tube in the form of a jet, near the top of the inlet waterpipe. If the level of the water be below the injector, the escaping jet of steam, by its superficial action (or friction) upon the air around it, forms a partial vacuum in the combining-tube and inletpipe, and the water then rises by virtue of the external pressure of the atmosphere. Once risen to the jet, the water is acted upon by the steam in the same manner as the air has been seized and acted upon in first forming the partial vacuum into which the water rose. Giffard was the first to make a practical application of the principles embodied in the injector; in fact, when he invented his injector, he may be said to have invented them all. His discovery was, that the motion imparted by a jet of steam to a surrounding column of water was sufficient to force it into the boiler from which the steam was taken, and, indeed, into a boiler working at even a higher pressure. It is not at all extraordinary to see injectors, attached to boilers carrying a pressure of 70 or 80 lbs. per square inch, forcing water into other boilers under a pressure of 250 lbs. per square inch. This extraordinary accumulation of power may be explained as follows: the velocity with which steam