through T, has blown the air out of v through B, the communication between T and the boiler is closed. Now the same motion of the regulator which closes this, opens the communication belween T' and the boiler; for the sliding plate R (fig. 11.) is moved from the one tube to the other, and at the same time, as we have already stated, the condensing pipe is brought to play on v. While, therefore, a vacuum is being formed in v by condensation, the steam, flowing through T', blows out the air through B', as already described in the other vessel v; and while the air in s is rushing up through a into v, followed by the water raised in s by the atmospheric pressure, the vessel v' is being filled with steam, and the air is completely expelled from it.

The communication between T and the boiler is now again opened, and the communication between T' and the boiler closed by moving the regulator R (fig. 11.) from the tube T to T'; at the same time the condensing pipe is removed from over v, and brought to play upon v'. While the steam once more expels the air from v through B, a vacuum is formed by condensation in v', into which the water in s rushes through the valve A'. In the mean time v is again filled with steam. The communication between T and the boiler is now closed, and that between T' and the boiler is opened, and the condensing pipe removed from v', and brought to play on v. While the steam from the boiler forces the water in v' through B' into the force-pipe F, a vacuum is being. produced in v, into which water is raised by the atmospheric pressure.

Thus each of the vessels v v' is alternately filled from s, and the water thence forced into F. The same steam which forces the water from the vessels into F, having done its duty, is condensed, and brings up the water from s, by giving effect to the atmospheric pressure.

During this process, two alternate motions or adjustments must be constantly made; the communication between T and the boiler must be opened, and that between T' and the boiler closed, which is done by one motion of the regulator. The condensing pipe at the same time must be brought from v to play on v', which is done by the lever placed upon it. Again

the communication between T' and the boiler is to be opened, and that between T and the boiler closed; this is done by moving back the regulator. The condensing pipe is brought from v to v by moving back the other lever, and so on alternately.

For the clearness and convenience of description, some slight and otherwise unimportant changes have been made in the position of the parts. A perspective view of this engine is represented at the head of this chapter. The different parts already described will easily be recognised.

The engine of Savery was very clearly described in a small work published in London in 1702, entitled, The Miner's Friend, or an Engine to raise Water by Fire described, and the Manner of Fixing it in Mines; with an Account of the several Uses it is applicable unto, and an Answer to the Objection made against it; by Thomas Savery, Gentleman. This volume was dedicated to William III. (to whom the engine had been exhibited at Hampton Court palace), to the Royal Society, and to the mining adventurers of England. The following are the uses to which Savery proposed the engine should be applied: First, to raise water for turning all sorts of mills; second, supplying palaces and houses with water, and supplying means of extinguishing fire therein by the water so raised; third, the supplying cities and towns with water; fourth, draining fens or marshes; fifth, for ships; sixth, the drainage of mines.

Dr. Harris, in his Lexicon Technicum, or Dictionary of Arts and Sciences, mentions a machine of Savery's for propelling a vessel in a calm, by paddle-wheels placed at the side; but it does not appear that Savery contemplated the application of a steam engine to work these wheels.

It is only from scattered passages in publications of the day that it can be ascertained to what extent the engines of Savery were practically applied. In his address to the Royal Society, he speaks of the "difficulties and expense which he encountered in instructing artisans to make engines according to his wish; but that after much experience the workmen had become such masters of the thing, that they bound themselves to deliver the engines' exactly tight and fit for

service, and such as he (Savery) dare warrant them to every one that has occasion for them.'

In his address to the miners of England he also says, "that the frequent disorders and cumbersomeness of water engines then in use encouraged him to invent engines to work by this new force; that though they were obliged to encounter the oddest and almost insuperable difficulties, yet he spared neither time, pains, nor money, till he had conquered them."

In Bradley's Improvements of Planting and Gardening, 1718, the author thus speaks of an engine erected by Savery: —

"Supposing the situation of a house or garden to be a considerable height above any pond, river, or spring, and that it has at present no other conveniency of water than what is brought continually by men or horses to it. In this case, the wonderful invention of the late Mr. Savery, F.R.S., for raising water by fire, will not only supply the defect, by flinging up as much water as may be desired, but may be maintained with very little trouble and very small expense.

"It is now about six years since Mr. Savery set up one of them for that curious gentleman Mr. Balle, at Cambden House, Kensington, near London, which has succeeded so well that there has not been any want of water since it has been built; and, with the improvements since made to it, I am apt to believe will be less subject to be out of order than any engine whatever."

It is remarkable that, notwithstanding the high pressure steam necessary for the operation of Savery's engine, he does not appear to have adopted the obvious expedient of a safety valve. The safety valve had been previously known, having been invented about the year 1681, by Papin, for his digester, which was a close boiler, contrived by him for stewing meat and digesting bones, by submitting them to a higher temperature than that of water boiling in an open vessel.

The safety valve which has ever since been used for steam boilers of every kind is a valve which opens outwards, and is fitted to an aperture in the boiler, so as to be steam tight. It is pressed down by a weight, the amount of which is regulated

shall be limited. Thus, if the magnitude of the valve be a square inch, and the pressure of the steam be limited to 10 lbs. per square inch above the pressure of the atmosphere, then the valve would be loaded with a weight of 10 lbs. ; but as it was found necessary to vary from time to time the limiting pressure of the steam, or the load of the safety valve, these valves were usually constructed so as to be held down by the pressure of a lever having a sliding weight upon it. By moving the weight on the arm of the lever, the pressure on the valve could be increased or diminished at the discretion of the engineer. This contrivance was first applied to Savery's engines, by Desaguliers, about the year 1717, before which year Savery died.

It is justly observed by Mr. Farey, in his treatise on the steam engine, that, "when a comparison is made between Captain Savery's engine and those of his predecessors, the result will be in every respect favourable to his character as an inventor, and as a practical engineer; all the details of his invention are made out in a masterly style, and accidents and contingencies are provided for, so as to render it a real working engine; whereas De Caus, the Marquis of Worcester, Sir Samuel Morland, and Papin, though ingenious philosophers, only produced mere outlines, which required great labour and skill of subsequent inventors to fill up, and make them sufficiently complete to be put in execution.”

About the year 1718 further improvements were made in the construction of Savery's engine, by Dr. Desaguliers; but it is probable that some of these were suggested by the proceedings of the inventors of the atmospheric engine, which shall presently describe.

(33.) In order duly to appreciate the value of improvements, it is necessary first to perceive the defects which these improvements are designed to remove. Savery's steam engine, considering how little was known of the value and properties of steam, and how low the general standard of mechanical knowledge was in his day, is certainly highly

"Captain" is a title given in Cornwall to the superintendent of the works connected with a mine.

creditable to his genius. Nevertheless it had very considerable defects, and was finally found to be inefficient for the most important purposes to which he proposed applying it.

At the time of this invention, the mines in England had greatly increased in depth, and the process of draining them had become both expensive and difficult; so much so, that it was found in many instances that their produce did not cover the cost of working them. The drainage of these mines was the most important purpose to which Savery proposed to apply his steam engine.

It has been already stated that the pressure of the atmosphere amounts to about fifteen pounds on every square inch. Now, a column of water, whose base is one square inch, and whose height is thirty-four feet, weighs about fifteen pounds. If we suppose that a perfect vacuum were produced in the steam vessels v v' (fig. 12.) by condensation, the atmospheric pressure would fail to force up the water, if the height of the top of these vessels above the water to be raised exceeded thirty-four feet. It is plain, therefore, that the engine cannot be more than thirty-four feet above the water which it is intended to elevate. But in fact it cannot be so much; for the vacuum produced in the steam vessels v v′ is never perfect. Water, when not submitted to the pressure of the atmosphere, will vaporise at a very low temperature, as we shall hereafter explain; and it was found that a vapour possessing a considerable elasticity would, notwithstanding the condensation, remain in the vessels v v' and the pipe s, and would oppose the ascent of the water. In consequence of this, the engine could never be placed with practical advantage at a greater height than twenty-six feet above the level of the water to be raised.

(34.) When the water is elevated to the engine, and the steam vessels filled, if steam be introduced above the water in v, it must first balance the atmospheric pressure, before it can force the water through the valve B. Here, then, is a mechanical pressure of fifteen pounds per square inch expended, without any water being raised by it. If steam of twice that elastic force be used, it will elevate a column in F of thirty-four feet in height; and if steam of triple the force