Such was the state of the air-pump when it engaged the attention of Mr. Smeaton. The imperfections of Alr. Hawksbee's pump, which he endeavoured to remedy, arose from the difficulty in opening the valves at the bottom of the barrels, and from the pistons not fitting exactly when put close down to the bottom, which left a lodgment of air which could not be got out of the barrel. The first of these defects was owing to the small orifice through which the air passed in raising the slip of bladder, and which op posed a considerable resistance, from the circumstance of its having been necessary to keep it moist with oil or wa ter. The diameter of the hole was about of an inch. Now, suppose the air to have been rarefied 140 times, which was nearly the maximum effect of the ma chine, and 40 inches of mercury will represent its elas ticity, or, in other words, a column of mercury of this height would have prevented the air from expanding. But, as the specific gravity of mercury is about t2i, and the 3 1 weight of the cubic inch of water is 252 grains, - 14 X 131 X 252 X •7854= 5.5 grains nearly, is the pressure which the air could exert in this attenuated state. This force was incapable of overcoming the cohesion of the bladder to the plate, by which means the process of ex haustion was discontinued. To obviate this, Mr. Smeaton resolved to expose a greater surface to the action of the air ; with this view, instead of one hole, he "made use of seven, all of equal size and shape, one being in the cen tre, and the other six round it, so that the valve was sup ported at proper distances by a kind of grating made by the solid parts between these holes." To expose a suffi ciently great surface, the holes were made hexagonal as represented, (Fig. 11. Plate CCCCLXIV.) and the parti tions filed almost to an edge. The letters E, F, G, H, show where the metal was a little protuberant, to hinder the pis ton from striking against the bladder. To prevent any lodgment of air in the lower part of the barrel, he remov ed the external pressure from the piston valve, by making the piston move through a collar of leather, and forced the air out by a valve applied to the plate at the top of the barrel which opened outwards.
The latest improvement which has been made on this machine is due to Cu thbertson, a philosophical instrument maker of Amsterdam, who has ingeniously contrived to allow the air to escape, by opening a passage to it mecha nically, without the assistance of its elastic force. We shall give a description of a machine of this contrivance as constructed by Miller and Adie, instrument makers, Edinburgh. GB (Fig. 12.) is a section of the barrel of the pump, where 1' represents a tube connected with an orifice in the plate of the pump, and screwed to the bar rel at S. C represents a frustum of a cone which is made to fit exactly an excavation in the piece HH. At R' this conical valve is connected to a solid rod R'R, having a loop at its lower extremity, which is made to embrace a horizontal bar in such a manner as to allow it a little play. When the cone is placed in its receptacle, a screw A is driven home from below, whose head is large enough to prevent the cone from being drawn out by pressing against the plate HH. By drawing the rod the cone can be ed a certain degree out of its cavity, as represented in Fig.
11. by which means a communication is opened between the interior of the receiver and the barrel. On pressing down the rod, on the other hand, the cone is restored to its place, and, being ground quite tight, it completely shuts up all communication. DD represents another cone, with an aperture about half an inch in diameter, in order to re ceive collars of leather which are represented in Figs. 12 and 14, as shaded. These collars occupy about two-thirds of the channel, and are pierced through so as to embrace very lightly the rod R'R, on purpose to raise the cone C. This cone is received inverted into the piston LL, which is hollowed out into a conical form so as to fit it exactly. When so placed the rod PSE, which is hollow in order to contain R, is screwed into the cone at E. The rod PSE is provided with the flanch SS, to prevent the cone, when pressed down, from being driven from its cavity altogether. By depressing the rod the cone DD is at first lowered so as to open a communication between the part of the bar rel which is above and that which is below the piston ; and then the flanch SS, acting upon the piston LL, forces it down also. F and K are collars of leather contained in nuts which are screwed upon the plates MN, QK. V is a conical valve ground into an aperture in the plate MN; it is raised up by the pressure of the part V' of the piston, and is prevented from escaping by means of the contri vance represented in Fig. 12, which Mr. Adie has the me rit of inventing. 0 is an aperture in the side of the bar rel, fitted with an exterior screw, to which a tube may be attached for the purpose of condensing air. flaying thus explained the parts of the pump, let us now attend to the manner in which it operates. On raising the rod G (which is done precisely in the same manner as in FIawksbee's pump,) the collars of leather DD, which closely embrace the rod RR', raise it, and consequently lift the cone C from its cavity, so as to allow the air confined in the tube 1' and the receiver to expand itself into the lower part of the bar rel. The cone D is also raised to fill its cavity and shut up all communication between the upper and lower part of the barrel. In this manner the air contained in the part of the barrel above the piston is condensed, and is driven through the valve V, which is acted upon by the elasticity of the air, and completely raised by the action of V' against it. On depressing the piston rod, the valve V' falls back into its former place, and prevents the air from re-enter ing into the barrel; by continuing the descent of the rod, the cone C is also replaced in its socket, and thus prevents the air contained in the lower part of the barrel from rush ing back into the receiver ; the cone D is now displaced, however, and opens a passage by which this air can ascend so as to occupy the upper part of the barrel. On raising the piston a second time, the same movements take place, by which means another portion of air is expelled through the passage V. It is obvious that this machine may also serve for condensing air into a vessel connected by a tube to the orifice 0, by allowing the orifice S to communicate with the atmosphere, or the gas which it may be required to condense.