air may move in the pipe IK with sufficient velocity. The part of the tube between the lowest perforations and B, and also the vessel CD, must be completely closed, to prevent the escape of the internal air. The water escapes at an aperture above F, a foot in diameter, and sometimes sluices m n are used to regulate its discharge. the atmo sphere. The wind is Fabri and Dietrich imagined that the wind is ocsupplied from casioned by the decomposition of the water, or its transformation into gas, in consequence of the agitation and percussion of its parts. But M. Venturi,2 to whom we are indebted for the first philosophical account of this machine, has shewn that this opinion is erroneous, and that the wind is supplied from the atmosphere; for when the lateral perforations were shut, no wind was generated. Hence the principal object in the construction of these machines is to combine as much air as possible with the descending current. With this view the water is often made to pass through a kind of cullender placed in the open air, and perforated with a great number of small triangular holes. Through these apertures the water descends in many small streams, and by exposing a greater surface to the atmosphere, it carries along with it an immense quantity of air, and is conveyed to the pedestal G by a tube A B, open and enlarged at ɑ, so as to be considerably wider than the end b of the pipe which holds the cullender. According to Venturi, the diameter of the aperture at C should not exceed one half of the diameter of the tube C H. It has been generally supposed that the waterfall should be very high;3 but Dr. Lewis has shewn, by a variety of experiments, that a fall of 4 or 5 feet is sufficient, and that when the height is greater than this, two or more blowing machines may be erected, by conducting the water from which the air is extricated into another reservoir, from which it again descends and generates air as formerly. That the air, which is necessarily loaded with moisture, may arrive at the furnace in as dry a state as possible, the condensing vessel CD should be made as high as circumstances will permit; and in order to * Experimental Enquiry concerning the Lateral Communication of Motion in Fluids. Prop. 8. 3 Wolfius makes the length of the tube CH five or six feet. Opera Mathematica, tom. i, p. 830. determine the strength of the blast, it should be furnished filled with water. with 4 a gauge Franciscus Tertius de Lanis observes, that he has seen a greater wind generated by a blowing machine of this kind, than could be produced by bellows 10 or 12 feet long.s The rain wind is produced in the same way as Cause of the the blast of air in water-blowing machines. When rain wind. the drops of rain impinge upon the surface of the sea, the air which they drag along with them often produces a heavy squall, which is sufficiently strong to carry away the mast of a ship. The same phenomenon happens at land, when the clouds empty themselves in alternate showers. In this case, the wind proceeds from that quarter of the horizon where the shower is falling. The common method of accounting for the origin of winds by local rarefactions of the air, appears to be pregnant with insuperable difficulties; and there is reason to think that these agitations in our atmosphere ought rather to be referred to the principle which we have now been considering. The Ventaroli or natural blasts which sometimes issue from volcanic mountains, arise from the air carried down the crevices by the falls of water. At the foot of the cascades which descend from the glaciers of Roche Melon upon the naked rock of La Novalese, Venturi found that the force of the wind arising from the air dragged down by the water could scarcely be withstood." 2. Description of Whitehurst's Machine for raising Water. Mr. Whitehurst, an ingenious watchmaker of Derby, appears to have been the first who entertained the ingenious idea of raising water by means of its momentum. A machine upon this principle was erected at Oulton in Cheshire, and was described in the Transactions of the Royal Society for 1755. It was intended for the service of a brewhouse and other offices, and was found to answer effectually the purpose of its erection. This 4 In Magisterio Naturæ et Artis, lib. v, cap. 3. 5 If we call a the diameter of the aperture at b, and d the diameter of the tube A B, then, according to Venturi, the quantity of air which passes into the tube in one second will be 6.1 a2 √ a + d — 1.4 0.4 a2 / 0.1 a. 6 Those who wish for more information upon the subject of water-blowing machines may consult Lewis's Commerce of Arts; the Journal des Mines, No. 91; or Nicholson's Journal, vol. i, 4to, vol. xii, p. 48. machine is represented in Plate IX, Fig. 9, where A M is the spring or original reservoir of water, whose surface at Mis on a level with B, the bottom of the reservoir B N. The main pipe A E is about 200 yards long, and 14 inches in diameter; the branch pipe E F is of the same diameter, and for the service of the kitchen, offices, &c. situated at least 18 or 20 feet below the surface of the reservoir A M; the cock Fis about 16 feet below the surface of the water at M. A valve box, with its valve a, is shewn at D, and C is an air vessel into which are inserted the extremities m, n of the main pipe, which are bent downwards for the purpose of preventing the air from being driven out when the water is forced into it. Now, when the cock F is opened, the water will rush out with a velocity of nearly 32 feet per second, corresponding to a pressure of 16 feet perpendicular height. A column of water, therefore, 200 yards long and 1 inch diameter, is now put in motion, and must have a considerable momentum. Hence, if the cock F is suddenly shut, the water will rush through the valve a into the air vessel C, and condense the included air. This condensation will take place every time that the cock is opened, so that the included air being compressed, will press upon the water in the air vessel, and raise it into the reservoir B N. Mr. Whitehurst observes, that the condensation of the air was strong enough to burst the vessel C in a few months after it was first constructed, though it was made of sheet lead of about 9 or 10 lbs. to a square foot. Hence he concluded that the momentum of the water was much superior to the simple pressure of the column. This simple and ingenious machine is obviously the same in principle as the hydraulic ram invented by Montgolfier, and which differs from it only in this, that the operation analogous to that of opening the cock Fis produced by the motion of the water itself, as will be seen in the following description of this ingenious contrivance. 3. Description of Montgolfier's Hydraulic Ram. This interesting machine was first constructed by Montgolfier about the year 1797, and has been brought to a very perfect state by a series of improvements which he has succesively made upon it. The rams which we have represented in Plate IX, Fig. 11, 12, 13, 14, 15, contain the improvements which have been made so late as 1816. The large pipe AB (Fig. 11 and Fig. 14), called the body of the ram, passes through the side of the reservoir P Q, from which the fall of water is obtained. It has a trumpet mouth at one end A, and at the other end an opening HH, which can be closed by valves C or D. When these valves are open, the water will issue at HH with a velocity due to the height AP; but when the internal valve C is closed, as in the figure, the water is prevented from issuing. When the valve C opens, it descends into the position shewn by the dotted lines G G (Fig. 14) being guided between three or four stems g, g, which have hooks at the lower ends for supporting the valves. In this case the water has a free passage between these stems, and the width of the passage can be increased or diminished by the screws with which the stems are fixed. The valve C is made of metal, and has a hollow cup or dish of metal attached to its lower surface. The seat H H of the valve is wider than the diameter of the pipe A B. It consists of a short cylinder or pipe screwed by its flanch h h (Fig. 14) into the opening of the upper surface of the head R of the ram; and the cylinder is so formed as to have an inverted cup or annular space i i round the upper part of it for containing air, which cannot escape when it is compressed by the water. A small pipe kl, leading from this annular space to the open air, is furnished with small valves kl, one of which, k, opens inwards to admit the air into i i, but to prevent its return, while the other valve I admits a certain quantity of air, and then shuts and prevents any farther entrance. The valve D is exactly the same as C, only it descends as in the figure when it shuts, and rises when it opens. The upper part of the head of the ram at E (Fig. 11) is made flat, and has several valves which allow the water to pass freely from the pipe AB, but prevent, its return. On each side of the head of the ram, at the part opposite to these valves, is a hollow enlargement, shewn by the dotted lines K, forming a circular bason, through the centre of which the pipe ABR passes. This part of the construction is shewn more distinctly in Fig. 12, which is a transverse section through LEZ in a plane perpendicular to that of the paper. The pipe is here made flat instead of circular, as seen at E, Fig. 12, for forming the seats of the valves, and the bason KK is covered with an air vessel FF. This air vessel communi VOL. II. U cates all round the pipe B, with the bason K K, and with the vertical pipe M. The machine being thus constructed, let us suppose the pipe ABR (Fig. 11 and 14) full of water, and the valve C to be opened, the water will lift the valve D, and escape with a velocity due to the height of the reservoir. In a short time, the water having acquired an additional velocity, raises the valve G, which shuts the passage, and prevents the escape of the water. The consequence of this is, that all the included water exerts suddenly a hydrostatical pressure on every part of the pipe, compressing at the same time the air in the annular space i i, which by its elacity diminishes the violence of the shock. This hydrostatical pressure opens the valves at E, and a portion of the water flows into the air vessel F, and condenses the air which it contains. The valves at E now close, preventing the return of the water into the pipe, and the water recoils a little in the tube with a slight motion from B to A, in consequence of the reaction or elasticity of the compressed air in i i, and also of the metal of the pipe, which must have yielded a little to the force exerted upon it in every direction. The recoil of the water towards A produces a slight aspiration within the head R of the ram, which causes the valve D to descend by its own weight, and prevent the water X which covers it from descending into the tube. The air, however, passes through the pipe lk, opens the valve k, and a small quantity is sucked into the annular space ii; but the quantity is very small, as the valve k closes as soon as the current of air becomes rapid. During the recoil towards A, the valve C, being unsupported, falls by its own weight; and when the force of recoil is expended by acting on the water in the reservoir PQ, the water begins again to flow along AB R, and the very same operation which we have described is repeated without end, a portion of water being driven into the air vessel F at every ascent of the valve C. The air in this vessel being thus highly compressed, will exert a force due to its elasticity upon the surface of the water in the vessel F, and will force it up through the pipe M to a height which is sufficient to balance the elasticity of the included air. The small quantity of air which is drawn into the annular space ii through the air tube lk at each aspiration, causes an accumulation of air in the space ii; and when the aspiration |