The next important modification in form was that of making the boiler entirely cylindrical with hemispherical ends, which is probably the form best fitted to combine as many requisites as possible. With this form the furnace is often placed in a second cylinder within the boiler, and forming the first part of the flue: thus the fuel beiug entirely surrounded by the water, little or no heat is lost by radiation ; but there are serious objections to this practice on the score of acci dents, as well as the deficiency in draught, owing to the coufined space in which the combustion takes place.
When an engine is of such a size as to require more steam than one boiler of the ordinary magnitude can supply, it has two or more, set aide by side, communicating with a common steam-pipe. Since the extent of surface exposed to the pressure of the steam,•and therefore the liability to bursting, increases in a greater ratio than the capacity, there is obviously a limit to size, which can never safely be surpassed, while the security is proportionally increased by diminishing that capa city : hence the necessity for using two or more small boilers instead of one large one ; and the principle, carried to its limit, constitutes that of the tubular boiler, in which the steam is generated in a series of independent metal pipes of small diameter, all communicating with a common steam-chamber or reservoir, itself sthall, and strong enough to resist great pressures.
These tubular boilers, however, have not come into general use, not only on account of their complexity, and consequent liability to derange ment, but because, unless in the case of locomotives, or steam-boat engines, there does not exist any demand for steam of such high pressure as they are intended to generate.
Since, in accordance with the general hydrostatic law, every unit of surface of the boiler has to sustain the same pressure, if a small portion of that surface can be opened by the pressure of the steam, when it has attained that which the boiler was only intended to withstand, by the escape of a quantity of the vapour at this orifice, the elasticity of the remainder is again reduced below the limit. This is the object of the safety-valve, which is such an aperture, kept closed by a valve retained in its scat by a weight calculated to yield to, or be raised by, the pressure the moment the steam exceeds its proper elastic force.
The safety-valve was first contrived by l'apin, and used in his digester and boilers, and has ever since constituted en essential appendage to every boiler. In its simplest form it is an obtuse conical valve, kept in its seat (which is at the end of a short collar, standing up from the surface of the boiler) by a weight acting at the end of a lever, resting on the spindle of the valve, and having its fulcrum or hinge at the other extremity. The effective weight by which the valve is kept
down may be varied by shifting the position of the weight on the arm of the lever ; and as this alteration might be unintentionally made by carelessness or accident, the valve should be inclosed in a box under lock and key, to prevent its being tampered with. A chain attached to the valve, and accessible to the eugineer, should be provided, to enable him to raise the valve, to ascertain that it is in efficient order, and has not become fixed in its socket. But the best mode of applying the weight Is directly on the valve, so that it cannot be increased, as long as it is inclosed, by any accidental alteration in its position. In loco motive engines the weight would be liable to derangement by the motion, and a spiral spring is employed to keep down the valve.
The safety-valve is perfectly effective as long as it is free to rise in its seat, and is loaded with a constant weight, which ought never to be more than one-third the pressure the boiler is just capable of with standing. The rapid diminution in the number of accidents from explosions, notwithstanding the increased employment of steam, suffi ciently proves that they are nine times out of ten caused by gross negligence or culpable recklessness; but to obviate as much as possible the recurrence of explosions, every boiler should have at least two safety-valves, both secured from access, and yet both capable of being raised by band from time to time : one should be loaded with a less weight than the other, that by the escape of the steam from it the engineer may be warned to reduce the quantity of steam generated, by "damping" his draught ; and the other safety-valve should be only loaded with a weight equivalent to one-third the pressure which, by computation founded on actual experiments, would burst the boiler, if made with metal of a given thickness; for, however carefully the boiler may have been made, it is impossible to ensure equal strength of it in every A originally suggested by Trevithick for insuring safety from a boiler by inserting in it a plug of metal, which, melting at the temperature attained by the steam when its tension became dangerous, might open an exit for it. This plan is adopted in France, but besides that it is repugnant to our ideas of mechanical fitness, it is liable to many objections ; none of the pure metals melt at a temperature suffi ciently low to be available, and all the fusible alloys soften long before they melt, and vary in these respects with minute differences iu the proportion of their ingredients, so that the plug might be driven out before the proper time.