tion under the pencil, would trace upon the paper a line. If the pencil were stationary this line would be straight and horizontal, but if the pencil were subject to a vertical motion, the line traced on the paper moved under the pencil horizontally would be a curve, the form of which would depend on the vertical motion of the pencil. The board thus supporting the paper is put into connexion by a light cord carried over pulleys with some part of the parallel motion, by which it is alternately moved to the right and to the left. As the piston ascends or descends, the whole play of the board in the horizontal direction will therefore represent the length of the stroke, and every fractional part of that play will correspond to a proportional part of the stroke of the steam-piston.

The apparatus being thus arranged, let us suppose the steam-piston at the top of the cylinder commencing its descent. As it descends, the pencil attached to the indicator piston-rod varies its height according to the varying pressure of the steam in the cylinder. At the same time the paper is moved uniformly under the pencil, and a curved line is traced upon it from right to left. When the piston has reached the bottom of the cylinder, the upper exhausting-valve is opened, and the steam drawn off to the condenser. The indicatorpiston being immediately relieved from a part of the pressure acting upon it descends, and with it the pencil also descends; but at the same time the steam-piston has begun to ascend, and the paper to return from left to right under the pencil. While the steam-piston continues to ascend, the condensation becomes more and more perfect, and the vacuum in the cylinder, and therefore also in the indicator, being gradually increased in power, the atmospheric pressure above the indicator-piston presses it downwards and stretches the spring. The pencil meanwhile, with the paper moving under it from right to left, traces a second curve. As the former curve showed the actual pressure of the steam impelling the piston in its descent, this latter will show the pressure of the uncondensed steam raising the piston in its ascent, and a comparison of the two will exhibit the effective force on the piston. Fig. 81. represents such a diagram as would be pro

duced by this instrument. A B C is the curve traced by the pencil during the descent of the piston, and C D E that during its ascent. A is the position of the pencil at the moment the piston commences its descent, B is its position at the middle of the stroke, and c at the termination of the stroke. On closing the upper steam-valve and closing the exhaustingvalve, the indicator-piston being gradually relieved from the pressure of the steam the pencil descends, and at the same time the paper moving from left to right, the pencil traces the curve C D E, the gradual descent of this curve showing the progressive increase of the vacuum. As the atmospheric pressure constantly acts above the piston of the indicator, its position will be determined by the difference between the atmospheric pressure and the pressure of the steam below it; and therefore the difference between the heights of the pencil at corresponding points in the ascending and descending stroke, will express the difference between the pressure of the steam impelling the piston in the ascent and resisting it in the descent at these points. Thus at the middle of the stroke, the line B D will express the extent to which the spring governing the indicator-piston would be stretched by the difference between the force of steam impelling the piston at the middle of the descending stroke, and the force of steam resisting it at the middle of the ascending stroke. The force therefore measured by the line B D will be the effective force on the piston at that point; and the same may be said of every part of the diagram produced by the indicator.

The whole mechanical effect produced by the stroke of the piston being composed of the aggregate of all its varying effects throughout the stroke, the determination of its amount

is a matter of easy calculation by the measurement of the diagram supplied by the indicator. Let the horizontal play of the pencil from A to c be divided into any proposed number of equal parts, say ten: at the middle of the stroke, B D expresses the effective force on the piston, and if this be considered to be uniform through the tenth part of the stroke, as from f to g, then the number of pounds expressed by B D multiplied by the tenth part of the stroke expressed in parts of a foot, will be the mechanical effect through that part of the stroke expressed in pounds' weight raised one foot. In like manner m n will express the effective force on the piston after three fourths of the stroke have been performed, and if this be multiplied by a tenth part of the stroke as before, the mechanical effect similarly expressed will be obtained; and the same process being applied to any successive tenth part of the stroke, and the numerical results thus obtained being added together, the whole effect of the stroke will be obtained, expressed in pounds' weight raised one foot.

(162.) By means of the indicator, the actual mechanical effect produced by each stroke of the engine can be obtained, and if the actual number of strokes made in any given time be known, the whole effect of the moving power would be determined. An instrument called a counter was also contrived by Watt, to be attached either to the working beam or to any other reciprocating part of the engine. This instrument consisted of a train of wheel-work with governing hands or indices moved upon divided dials, like the hand of a clock. A record of the strokes was preserved by means precisely similar to those by which the hands of a clock or time-piece indicated and recorded the number of vibrations of the pendulum or balance-wheel.

(163.) To secure the boiler from accidents arising from the steam contained in it acquiring an undue pressure, a safetyvalve is used, similar in principle to those adopted in the early engines. This valve is represented in fig. 71. at N. It is a conical valve, kept down by a weight sliding on a rod upon it. When the pressure of the steam overcomes the force of this weight, it raises the valve and escapes, being carried off through the tube.

With a view to the economy of heat, this waste steam tube is sometimes conducted into the feeding cistern, where the steam carried off by it is condensed, and heats the feeding

water.

The magnitude of the safety-valve should be such that, when open, steam should be capable of passing through it as rapidly as it is generated in the boiler. The superficial magnitude, therefore, of such valves must be proportional to the evaporating power of the boiler. In low pressure boilers the steam is generally limited to five or six pounds' pressure per square inch, and consequently the load over the safety-valve in pounds would be found by multiplying the superficial magnitude of its smallest part by these numbers. In boilers in which the steam is maintained at a higher pressure, it would be inconvenient to place upon the safety-valve the necessary weight. In such cases a lever is used, the shorter arm of which presses down the valve, and the longer arm is held down by a weight capable of adjustment, so that the pressure on the valve may be regulated at discretion. Two safety-valves should be provided on all boilers, one of which should be locked up, so that the persons in care of the engine should have no power to increase the load upon it. In such case, however, it is necessary that a handle connected with the valve should project outside the box containing it, so that it may always be possible for the engineer to ascertain that the valve is not locked in its seat, a circumstance which is liable to happen.

Sometimes also two safety-valves are provided, one loaded a little heavier than the other. The escape of steam from the lighter valve in this case gives notice to the engine-man of the growing increase of pressure, and warns him to check the production of steam. The lever by which the safetyvalve is held down is sometimes acted on by a spiral spring, capable of being so adjusted as to produce any required pressure on the valve. This arrangement is adopted in locomotive engines, where steam of very high pressure is used; and in such cases also there are always provided two such valves, one of which cannot be increased in its pressure.

The pipe by which the boiler is fed with water will neces

sarily act as a safety-valve, for when the pressure of the steam increases in an undue degree, it will press the water in the boiler up through the feed-pipe, so as to discharge it into the feed-cistern, a circumstance which would immediately give notice of the internal state of the boiler. The steam-gauge, already described (fig. 79.), would also act as a safety-valve ; for if the pressure of steam in the boiler should be so augmented as to blow the mercury out of the steam-gauge, the steam would then issue through the gauge, and the pressure of the boiler be reduced, provided that the magnitude of the tube forming the steam-gauge were sufficient for this purpose.

(164.) In high pressure boilers which are exposed to extreme temperatures and pressures, and which are therefore subject to danger of explosion, a plug of metal is sometimes inserted, which is capable of being fused at a temperature above which the boiler should not be permitted to be raised. If the pressure of steam increase beyond the proper limit, the temperature of the water and steam will undergo a corresponding increase; and if the metal of the plug be capable of being fused at such a temperature, the plug will fall out of the boiler, and the steam and water will issue from it. Various alloys of metal are fusible at temperatures sufficiently low for this purpose. An alloy composed of one part of lead, three of tin, and five of bismuth, will fuse at the common temperature of boiling water; and alloys of the same metals, in various proportions, will fuse at different temperatures from 200° to 400°.

Although fusible plugs may be used, in addition to other means of insuring safety, they ought not to be exclusively relied on at the ordinary working pressure of the boiler. The fusible plug ought to be capable of more than resisting the pressure; but if it be so, its point of fusion would be one at which the steam would have a pressure of at least two atmospheres above its working pressure. The plug would therefore be capable of being fused only as soon as the steam would acquire a pressure of 30 lbs. per inch above its regular working pressure.

When a boiler ceases to be worked, and the furnace has been extinguished, the space within it appropriated to steam