called in French Soufflet d'eau ou trompe, is a machine which, by the action of falling water, supplies air to a blast furnace. It consists of an upright pipe, through which a shower of water is made to fall; and this shower carries down with it a mass of air, which is received beneath in a kind of tub, and conducted to the furnace by means of a pipe. The first idea of such a machine was doubtless suggested by those local winds, which are always produced by natural falls of water over precipices, and in the mountains (see page 278 of this Volume); but perhaps we are indebted for the first accurate theory of it to Professor Venturi.

That philosopher, in his experimental researches concerning the lateral communication of motion in fluids, proves that the water-blowing machine affords air to the furnace, by the accelerating force of gravity and the lateral communication of motion combined together. He begins with an idea, which, he candidly acknowledges, did not escape the penetration of Leonardo Da Vinci. Suppose a number of equal balls to move in contact with each other along the horizontal line AB (Plate XLVI. fig. 1.). Imagine them to pass with an uniform motion, at the rate of four balls in a second. Let us take BF, equal to 16 feet English. During each second four balls will fall from B to F, and their respective distances in falling will be nearly BC = 1, CD = 3, DE = 5, EF = 7. We have here a very evident representation of the separation, and successive elongation, which the accelerating force of gravity produces between bodies which fall after each other.

The rain water flows out of gutters by a continued current; but during its fall it separates into portions in the vertical direction, and strikes the pavement with distinct blows. The water likewise divides, and is scattered in the horizontal direction. The stream which issues out of the gutter may be one inch in diameter, and strike the pavement over the space of one foot. The air which exists between the vertical and horizontal separations of the water which falls, is impelled and carried downwards. Other air succeeds laterally; and in this manner a current of air or wind is produced round the place struck by the water. Hence the following idea of a water-blowing machine:

Let BCDE (fig. 2.) represent a pipe, through which the water of a canal AB falls into the lower receiver MN. The sides of the tube have openings all round, through which the air freely enters to supply what the water carries down in its fall. This mixture of water and and air proceeds to strike a mass of stone Q; whence rebounding through the whole width of the receiver MN, the water separates from the air, and falls to the bottom at XZ, whence it is discharged into the lower channel or drain, by one or more openings TV. The air being less heavy than the water, occupies the upper part of the receiver; whence being urged through the upper pipe O, it is conveyed to the forge.

It has been supposed by some eminent chemists, that the air which passes through the pipe O is furnished by the decomposition of water. To ascertain whether this be the case or not, our author formed a water-blowing engine of a small size. The pipe BD was two inches in diameter, and four feet in height. When the water accurately filled the section BC, and all the lateral openings of the pipe BDEC were closed, the pipe O no longer offered any wind. It is therefore evident, that in the open pipes the whole of the wind comes from the atmosphere, and no portion is afforded by the decomposition of water. It remains, therefore, to determine the circumstances proper to drive into the receiver MN the greatest quantity of air, and to measure that quantity.

1. To obtain the greatest effect from the acceleration of gravity, it is necessary that the water should begin to fall at BC, (fig. 2.) with the least possible velocity; and that the height of the water FB should be no more than is necessary to fill the section BC. Our author supposes the vertical velocity of this section to be produced by an height or head equal to BC.

2. We do not yet know, by direct experiment, the distance to which the lateral communication of motion between water and air can extend itself; but we may admit with confidence, that it can take place in a section double that of the original section with which the water enters the pipe. Let us suppose the section of the pipe BDEC to be double the section of the water at BC; and, in order that the stream of fluid may extend and divide itself through the whole double section of the pipe, some bars, or a grate, are placed in BC, to distribute and scatter the water through the whole internal part of the pipe.

3. Since the air is required to move in the pipe O with a certain velocity, it must be compressed in the receiver. This compression will be proportional to the sum of the accelerations, which shall have been delivered in the interior part KD of the pipe. Taking KD = 1.5 feet, we shall have a pressure sufficient to give the requisite velocity in the pipe O. The sides of the portion KD, as well as those of the receiver MN, must be exactly closed in every part.

4. The lateral openings in the remaining part of the pipe BR may be so disposed and multiplied, particularly at the upper part, that the air may have free access within the tube. We will suppose them to be such that 0.1 foot height of water might be sufficient to give the necessary velocity to the air at its introduction through the apertures.

All these conditions being attended to, and supposing the pipe BD to be cylindrical, it is required to determine the quantity of air which passes in a given time through the circular section KL. Let us take in fact KB = 1.5; BC = BF = a; BD = b. By the common theory of falling bodies, the velocity in KL will be

\[ \frac{7}{8} \sqrt{(a + b - 1.5)} \]

the circular section KL = 0.785 a². Admitting the air in KL to have acquired the same velocity as the water, the quantity of the mixture of the water and air which passes in a second through KL is \( \frac{6}{1} \cdot \frac{1}{\sqrt{(a + b - 1.5)}} \). We must deduct from the quantity \( (a + b - 1.5) \) that height which answers to the velocity the water must lose by that portion of velocity which it communicates to the air laterally introduced; but this quantity is so small that it may be neglected in the calculation. The water which passes in the same time of one second thro' BC is \( \frac{0.4}{a} \cdot \frac{1}{\sqrt{(a + c_1)}} \). Consequently, the quantity of air which passes in one second through KL, will be \( \frac{6}{1} \cdot \frac{1}{\sqrt{(a + b - 1.5)}} - \frac{0.4}{a} \cdot \frac{1}{\sqrt{(a + c_1)}} \), taking the air itself, even in its ordinary state of compresion, under the weight of the atmosphere. It will be proper, in practical applications, to deduct one-fourth from this quantity; 1. On account of the shocks which the scattered water sustains against the inferior part of the tube, which deprive it of part of its motion; and, 2. Because it must happen that the air in LK will not, in all its parts, have acquired the same velocity as the water.

If the pipe O do not discharge the whole quantity of air afforded by the fall, the water will descend at XZ; the point K will rise in the pipe, the influx of air will diminish, and part of the wind will issue out of the lower lateral apertures of the pipe DK.

We shall not here examine the greater or less degree of perfection of the different forms of water-blowing machines which are used at various iron forges; such as those of the Catalans, and elsewhere. These points may be easily determined from the principles here laid down, compared with those established in the articles Resistance of Fluids (Encyc.), and Dynamics (Supplement).