exceed that of the steam, and the quantity of the excess may be ascertained exactly in the same way.

If the tube be glass, the difference of levels of the mercury would be visible; but it is most commonly made of iron; and in order to ascertain the level, a thin wooden rod with a float is inserted in the open end of B C, so that the portion of the stick within the tube indicates the distance of the level of the mercury from its mouth. A bulb or cistern of mercury might be substituted for the leg A B, as in the common baroThis instrument is called the steam-gauge.

meter.

If the steam-gauge be used as a measure of the strength of the steam which presses on the piston, it ought to be on the same side of the throttle-valve (which is regulated by the governor) as the cylinder; for if it were on the same side of the throttle-valve with the boiler, it would not be affected by the changes which the steam may undergo in passing through the throttle-valve, when partially closed by the agency of the governor.

For boilers in which steam of very high pressure is used, as in those of locomotive engines, a steam-gauge, constructed on the above principle, would have inconvenient or impracticable length. In such boilers the pressure of the steam is equal to four or five times that of the atmosphere, to indicate which the column of mercury in the steam-gauge would be four or five feet in height. In such cases a thermometer-gauge may be used with advantage. The principle of this gauge is founded on the fact, that between the pressure and temperature of steam produced in contact with water there is a fixed relation, the same temperature always corresponding to the same pressure. If, therefore, a thermometer be immersed in the boiler which shall show the temperature of the steam, a scale may be attached to it, on which shall be engraved the corresponding pressures. Such gauges are now very generally used on locomotive engines.

(160.) The force with which the piston is pressed depends on two things, 1st, the actual strength of the steam which presses on it; and, 2dly, on the actual strength of the vapour which resists it. For although the vacuum produced by the

what had been produced in the atmospheric engines, yet still some vapour of a small degree of elasticity is found to be raised from the hot water in the bottom of the condenser before it can be extracted by the air-pump. One of these pressures is indicated by the steam-gauge already described; but still, before we can estimate the force with which the piston descends, it is necessary to ascertain the force of the vapour which remains uncondensed, and resists the motion of the piston. Another gauge, called the barometer-gauge, is provided for this purpose. A glass tube A B (fig. 80.), more than thirty inches long and open at both ends, is placed in an upright or vertical position, having the A lower end в immersed in a cistern of mercury c. To the upper end is attached a metal tube, which communicates with the condenser, in which a constant vacuum, or rather high degree of rarefaction, is sustained. The same vacuum must therefore exist in the tube A B, above the level of the mercury, and the atmospheric pressure on the surface of the mercury in the cistern c will force the

Fig. 80.

mercury up in the tube A B, until the column which is suspended in it is equal to the difference between the atmospheric pressure and the pressure of the uncondensed steam. The difference between the column of mercury sustained in this instrument and in the common barometer, will determine the strength of the uncondensed steam, allowing a force proportional to one pound per square inch for every two inches of mercury in the difference of the two columns. In a well-constructed engine which is in good order, there is very little difference between the altitude in the barometergauge and the common barometer.

To compute the force with which the piston descends, thus becomes a very simple arithmetical process. First, ascertain the difference of the levels of the mercury in the steamgauge; this gives the excess of the steam pressure above the atmospheric pressure. Then find the height of the mercury in the barometer-gauge; this gives the excess of the atmospheric pressure above the uncondensed steam. Hence, if these two heights be added together, we shall obtain the

excess of the impelling force of the steam from the boiler, on the one side of the piston, above the resistance of the uncondensed steam on the other side: this will give the effective impelling force. Now, if one pound be allowed for every two inches of mercury in the two columns just mentioned, we shall have the number of pounds of impelling pressure on every square inch of the piston. Then, if the number of square inches in the section of the piston be found, and multiplied by the number of pounds on each square inch, the force with which it moves will be obtained.

From what we have stated it appears that, in order to estimate the force with which the piston is urged, it is necessary to refer to both the barometer and the steam-gauge. This double computation may be obviated by making one gauge serve both purposes. If the end c of the steam-gauge (fig. 79.), instead of communicating with the atmosphere were continued to the condenser, we should have the pressure of the steam acting upon the mercury in the tube BA, and the pressure of the uncondensed vapour which resists the piston acting on the mercury in the tube BC. Hence the difference of the levels of the mercury in the tubes would at once indicate the difference between the force of the steam and that of the uncondensed vapour, which is the effective force with which the piston is urged.

(161.) But these methods of determining the effective force by which the piston is urged, can only be regarded as approximations, and not very perfect ones. If the condensation of steam on one side of the piston were instantaneously effected, or the uncondensed vapour were of the same tension during the whole stroke; and if, besides this, the pressure of steam on the piston were of uniform intensity from the beginning to the end of the stroke, then the steam and barometer gauges taken together would become an accurate index of the effective force of steam on the piston: but such is not the case. When the steam is first admitted through the steam-valve it acts on the piston with a pressure which is first slightly diminished, and afterwards a little increased, until it arrives at that part of the stroke at which the steamvalve is closed, after which the pressure is diminished. The

T

pressure, therefore, urging the piston is subject to variation; but the pressure of the uncondensed vapour on the other side of the piston is subject to still greater change. At the moment the exhausting-valve is opened, the piston is relieved from the pressure upon it by the commencement of the condensation; but this process during the descent of the piston is gradual, and the vacuum is rendered more and more perfect, until the piston has nearly attained the limit of its play. These variations, both as well of the force urging the piston as of the force resisting it, are such as not to be capable of being accurately measured by a mercurial column, since they would produce oscillations in such a column, which would render any observations of its mean height impracticable.

To measure the mean efficient force of the piston, taking into account these circumstances, Mr. Watt invented an instrument, which, like all his mechanical inventions, has answered its purpose perfectly, and is still in general use. This instrument, called an indicator, consists of a cylinder of about 1 inch in diameter, and 8 inches in length. It is bored with great accuracy, and fitted with a solid piston moving steam-tight in it with very little friction. The rod of this piston is guided in the direction of the axis of the cylinder through a collar in the top, so as not to be subject to friction in any part of its play. At the bottom of the cylinder is a pipe governed by a stop-cock and turned in a screw, by which the instrument may be screwed on the top of the steam-cylinder of the engine. In this position, if the stop-cock of the indicator be opened, a free communication will be made between the cylinder of the indicator and that of the engine. The piston-rod of the indicator is attached to a spiral spring, which is capable of extension and compression, and which by its elasticity is capable of measuring the force which extends or compresses it in the same manner as a spring steel-yard or balance. If a scale be attached to the instrument at any point on the piston-rod to which an index might be attached, then the position of that index upon the scale would be governed by the position of the indicator-piston in its cylinder. If any force pressed the indicator-piston upwards, so as to compress the spring,

the index would rise upon the scale; and if, on the other hand, a force pressed the indicator-piston downwards, then the spiral spring would be extended, and the index on the piston-rod descend upon the scale. In each case the force of the spring, whether compressed or extended, would be equal to the force urging the indicator-piston, and the scale might be so divided as to show the amount of this force.

Now, let the instrument be supposed to be screwed upon the top of the cylinder of a steam-engine, and the stop-cock opened so as to leave a free communication between the cylinder of the indicator below its piston and the cylinder of the steam-engine above the steam-piston. At the moment the upper steam-valve is opened, the steam rushing in upon the steam-piston will also pass into the indicator, and press the indicator-piston upwards: the index upon its piston-rod will point upon the scale to the amount of pressure thus exerted. As the steam-piston descends, the indicator-piston will vary its position with the varying pressure of the steam in the cylinder, and the index on the piston-rod will play upon the scale, so as to show the pressure of the steam at each point during the descent of the piston.

If it were possible to observe and record the varying position of the index on the piston-rod of the indicator, and to refer each of these varying positions to the corresponding point of the descending stroke, we should then be able to declare the actual pressure of the steam at every point of the stroke. But it is evident that such an observation would not be practicable. A method, however, was contrived by Mr. Southern, an assistant of Messrs. Boulton and Watt, by which this is perfectly effected. A square piece of paper, or card, is stretched upon a board, which slides in grooves formed in a frame. This frame is placed in a vertical position near the indicator, so that the paper may be moved in a horizontal direction backwards and forwards, through a space of fourteen or fifteen inches. Instead of an index a pencil is attached to the indicator of the piston-rod: this pencil is lightly pressed by a spring against the paper above mentioned, and as the paper is moved in a horizontal direc