well as its change of temperature, will be subject to a variation proportional to the intervals between the successive feeds. It is manifest, therefore, that the feed should either be uninterrupted or be supplied at short intervals, so that the change of level and temperature of the water in the boiler should not be considerable.

(155.) Different methods have been, from time to time, suggested for indicating the level of the water in the boiler. We have already mentioned the two gauge-pipes used in the earlier steam-engines (31.), and which are still generally continued. There are, however, some other methods which merit our attention.

W

A

A weight F (fig.75.), half immersed in the water in the Fig. 75. boiler, is supported by a wire, which, passing steam-tight through a small hole in the top, is connected by a flexible string, or chain, passing over a wheel w, with a counterpoise A, which is just sufficient to balance F when half immersed. If F be raised above the water, A being lighter will no longer balance it, and F will descend pulling up A, and turning

F

the wheel w. If, on the other hand, F be plunged deeper in the water, a will more than balance it, and will pull it up, so that the only position in which F and A will balance each other is, when F is half immersed. The wheel w is so adjusted, that when two pins placed on its rim are in the horizontal position, the water is at its proper level. Consequently it follows, that if the water rise above this level, the weight F is lifted and A falls, so that the pins come into another position. If, on the other hand, the level of the water fall, F falls and a rises, so that the pins assume a different position. Thus, in general, the position of the pins becomes an indication of the quantity of water in the boiler.

(156.) Another method is to place a glass tube (fig.76.), with one end T entering the boiler above the proper level, and the other end T' entering it below the proper level. It must

Fig. 76.

be evident that the water in the tube will always stand at the same level as the T water in the boiler, since the lower part has a free communication with that water, while the surface is submitted to Tthe pressure of the same steam as the water in the boiler. This and the last

mentioned gauge have the advantage of addressing the eye of the engineer at once, without any adjustment; whereas the gauge-cocks must be both opened, whenever the depth is to be ascertained.

These gauges, however, require the frequent attention of the engine-man; and it becomes desirable either to find some more effectual means of awakening that attention, or to render the supply of the boiler independent of any attention. In order to enforce the attention of the engine-man to replenish the boiler when partially exhausted by evaporation, a tube was sometimes inserted at the lowest level to which it was intended that the water should be permitted to fall. This tube was conducted from the boiler into the enginehouse, where it terminated in a mouth-piece or whistle, so that whenever the water fell below the level at which this tube was inserted in the boiler, the steam would rush through it, and issuing with great velocity at the mouthpiece, would summon the engineer to his duty with a call that would rouse him even from sleep.

D

Fig. 77.

G

(157.) In the most effectual of these methods, the task of replenishing the boiler should still be executed by the engineer; and the utmost that the boiler itself was made to do, was to give due notice of the necessity for the supply of water. The consequence was, among other inconveniences, that the level of the water was subject to constant variation.

the engine is made to feed its own boiler.

The pipe G (fig. 77.), which leads from the hot water pump, terminates in a small cistern c in which the water is received. In the bottom of this cistern, a valve v is placed, which opens upwards, and communicates with a feed-pipe, which descends into the boiler below the level of the water in it. The stem of the valve v is connected with a lever turning on the centre D, and loaded with a weight F dipped in the water in the boiler in a manner similar to that described in fig. 75., and balanced by a counterpoise a in exactly the same way. When the level of the water in the boiler falls, the float F falls with it, and pulling down the arm of the lever raises the valve v, and lets the water descend into the boiler from the cistern c. When the boiler has thus been replenished, and the level raised to its former place, F will again be raised, and the valve v closed by the weight A. In practice, however, the valve v adjusts itself by means of the effect of the water on the weight F, so as to permit the water from the feeding-cistern c to flow in a continued stream, just sufficient in quantity to supply the consumption from evaporation, and to maintain the level of the water in the boiler constantly the same.

By this arrangement the boiler is made to replenish itself, or, more properly speaking, it is made to receive such a supply, as that it never wants replenishing, an effect which no effort of attention on the part of an engine-man could produce. But this is not the only good effect produced by this contrivance. A part of the steam which originally left the boiler, and having discharged its duty in moving the piston, was condensed and reconverted into water, and lodged by the air-pump in the hot well (fig. 77.), is here again restored to the source from which it came, bringing back all the unconsumed portion of its heat preparatory to being once more put in circulation through the machine.

The entire quantity of hot water pumped into the cistern c, is not always necessary for the boiler. A waste-pipe may be provided for carrying off the surplus, which may be turned to any purpose for which it may be required; or it may be discharged into a cistern to cool, preparatory to

being restored to the cold cistern, in case water for the supply of that cistern be not sufficiently abundant.

(158.) Another method of arranging a self-regulating feeder is shown in fig. 78. A is a hollow ball of metal attached to

the end of a lever, whose fulcrum is at B. The other arm of the lever c is connected with the stem of a spindle-valve, communicating with a tube which receives water from the feeding-cistern. Thus, when the level of the water in the boiler subsides, the ball a preponderating over the weight of the opposite arm, the lever falls, the arm c rises and opens the valve, and admits the feeding water. This apparatus will evidently act in the same manner and on the same prin

The mouth of the tube by which the feed is introduced should be placed at that part of the boiler which is nearest the end of the flues which issue into the chimney. By such means the temperature of the water in contact with those flues will be lowest at the place where the temperature of the heated air intended to act upon it is also lowest. The difference of the temperatures will therefore be greater than it would be if the point of the boiler containing water of a higher temperature was left in contact with this part of the flue.

(159.) It is necessary to have a ready method of ascertaining at all times the pressure of the steam which is used in working the engine. For this purpose a bent tube containing mercury is inserted into some part of the apparatus,

Fig. 79.

A

B

which has free communication with the steam. Let A B C (fig. 79.) be such a tube. The pressure of the steam forces the mercury down in the leg A B, and up in the leg B C. If the mercury in both legs be at exactly the same level, the pressure of the steam must be exactly equal to that of the atmosphere; because the steam pressure on the mercury in A B balances the atmospheric pressure on the mercury in B c. If, however, the level of the mercury in в c be above the level of the mercury in BA, the pressure of the steam will exceed that of the atmosphere.

The excess of its pressure above that of the atmosphere may be found by observing the difference of the level of the mercury in the tubes B C and в A, allowing a pressure of one pound on each square inch for every two inches in the difference of the levels.

If, on the contrary, the level of the mercury in B C should fall below its level in A B, the atmospheric pressure will