remarkable fact, that the air which escapes from a vessel by blowing through an aperture under any pressure whatever, does not alter in temperature, although it expands on issuing from the vessel.' From this they have deduced the explanation of two other facts, well known, observed at Schemnitz in Hungary, and at Chaillot, near Paris. The wind or blast of the columnar machine at Schemnitz produces a degree of cold which freezes water even in summer; whilst the blast from the reservoir of air of the Chaillot engine, which is obtained under a constant pressure of two atmospheres and a half, scarcely affects the most sensible thermometer in the same

season.

91. This explanation, still but little known, would perhaps be contested if it were not confirmed by the experiments of the artificial congelation of water, the analysis of which we are now about to offer, and which is obtained by a current of condensed air. Let A B C'D', plate II. fig. 1, be a cylindrical vessel, in which moves a piston C D, and having a cock E, which can be opened or shut at pleasure. We suppose that the part A B C D of this vessel contains atmospheric air, more or less compressed than exterior atmospheric air. The cock E being shut, and the interior air A B C D no longer communicating with the outer air, let us suppose the piston C D lowered to C' D'. The included air will be expanded, and this dilatation will produce a fall of temperature so much the more sensible as the dilatation is greater. Let us now suppose that the piston is fixed to CD, and that ABCD is a vessel filled, wholly or in part, with condensed atmospheric air, and maintained by any means whatever at the same degree of compression. In this hypothesis, the air, at the moment of opening the cock E, will blow through E with a constant force, and the thermometer, having its bulb placed at E, will indicate no perceptible change of temperature.

92. In the first hypothesis the whole volume of air contained in the vessel expands within the vessel, and the temperature falls in the second hypothesis only that portion of air which issues from the vessel expands outside, and the temperature of that issuing air does not alter perceptibly. Such are the facts observed by Gay Lussac and Welter. Let us examine what passes in the experiment which is repeated in all courses of natural philosophy, to show the artificial congelation of water. The receiver of a condenser is filled with air compressed to the force of several atmospheres. This receiver carries at its upper end a capillary tube, through which the air of the receiver may be allowed to escape. To this current of air a glass bulb, similar to that of a thermometer, is represented; and small crystals, scarcely visible to the naked eye, are soon formed on the outside surface of the bulb. Although the time for the formation of these crystals is very short, it is necessary to divide it by imagination into several periods. In the first period the compressed air dilates in the whole capacity of the receiver, and cools: in the following periods the air, more and more dilated, falls to a very low temperature; and, finally, in the last period, the current of air attains the maximum of

cold. From this observation it results, that the small crystals obtained on the glass bulb do not proceed from the cooling of the air outside of the receiver of the condenser, but from the fall of the temperature within the receiver. This fall inside is not instantaneous. It increases by the dilatation of the air of the receiver. This air, although subject to a progressive dilatation in the whole capacity of the receiver, preserves, during its flow, an elastic force greater than that of the atmosphere, strikes against the glass bulb, and cools it. What proves that this cooling takes place is, that the atmospheric air surrounding the bulb deposits on it a slight stratum of liquid water, on which the small crystals are formed, produced by the current of air previously cooled within the receiver.

93. In July 1826 M. Doubuisson, chief engineer at the corps royal des mines, published experiments on the flow of atmospheric air compressed in a gasometer, from which it issued into the atmosphere. He found that the quantity of air expelled from the gasometer through an orifice in a thin plate, and under a determinate pressure, is to the quantity which issues through short cylindrical or conical tubes, having the aperture where the air issues of the same diameter as the orifice in a thin plate, in the ratio of 1000 to 1427. In giving an account of these experiments in the Bulletin de la Société Philomatique, for September 1826, M. Hachette remarked that M. Doubuisson had not made the air flow through the adjutage known by the name of the tube of Venturi, which is only a common bellows pipe reversed, the section of its greatest diameter being taken from the orifice. The air being expansible would fill that tube, and the experiment would have shown the increase of expenditure due to the acceleration of the velocity of the air on the section of the adjutage of least diameter.'

94. In October, 1826, Messrs. Thenard and Clement visited the iron foundries of Fourchambault (department of La Nierre), and the following experiment was made under their eyes :—A workman presented a piece of deal board against the blast of a pair of bellows set in motion by a steam-engine. When the board was at a certain distance from the orifice of the pipe it was strongly repelled: if it was put nearer to the plane of the orifice it was carried towards the plane, as though the repulsion had changed to attraction. This effect takes place only so far as the end of the pipe is in the plane of some object surrounding it.

95. M. Clement first discovered that the atmospheric air acted, in this case, on the board as it does on the outside of a conical pipe through which water is flowing. This philosopher returned to Paris, showed, by means of a boiler which he had at his disposal, that steam, at a pressure of two or three atmospheres, produced an effect similar to that of a blast of a forge bellows. He adapted to the boiler a vertical cylindrical pipe, terminated by a circular plate about a decimetre (3.9 inches) in diameter, in the centre of which was a circular orifice of a smaller diameter.

96. When the steam issues through that orifice, a circular disk of the same diameter as the

plate is brought near to it, and the disk carried towards the plate is observed to adhere to it, as if it were attracted by a force acting in the direction contrary to that of gravity. Points projecting more or less from the face of the disk or of the plate opposite, determined the distance of these faces apart. M. Clement stated that the facts we have just related, in a memoir which he read to the Royal Academy of Sciences December 6th, 1826, and which is referred to the examination of commissioners.

97. On the 11th April, 1827,' M. Hachette observes, I repeated M. Clement's principal experiment, at the sitting of the society for the encouragement of arts, making use of a chamber pair of double bellows, which had its nozzle terminating in a plate of copper. I announced on the same day that the adherence of a disk opposed to the plate did not depend essentially on the expansibility of the air from the bellows, and that I had obtained effects similar to those obtained by M. Clement, by making water flow between the approximated disks, the curvatures of which I had varied.

98. At the sitting of the Société Philomatique, on the 13th of April 1827, I presented a bent tube, by means of which, blowing with the mouth, the same phenomena are produced as by the blowing machine of Fourchambault, or by the steam boiler of M. Clement.'

99. The study of these phenomena leads to this question:-To determine the pressure in every point of the exterior and interior surfaces of a vessel which is filled with a liquid or a gas, supposing this vessel to empty itself into the atmosphere,-1st, by an orifice in a thin plate, 2nd, by an adjutage or pipe, and 3d, by a zone comprised between two surfaces brought nearly together. It is to arrive at the solution of this question, that I have endeavoured to simplify the apparatus previously employed, and have made several experiments, an account of which I have given in the following notes, which were read at the sitting of the Société Philomatique on the 28th of April :

100. Some experiments on the flow of the gases between two surfaces brought nearly together elicited a variety of very singular phenomena. The principal fact observed by Messrs Thenard and Clement results from the combined action of the shock of the air against a plate, and of the atmospheric pressure on the same plate. All the circumstances of this action are rendered evident by means of the very simple instrument represented, figs. 7, 8. ABCD, fig. 2, is a bent tube of tin plate or glass, terminated by a circular plate, CD, of tin. In the centre of the plate is an orifice E, of three or four millimetres (12 or 157 of an inch) in diameter. Three or four small strips of tin are soldered on the edges of the plate, and are intended to retain opposite to the plate a disk of the same diameter, and of any material we like.

101. The instrument may yet be reduced, fig. 2, to a single plate C D of tin plate, in the centre of which is a small orifice covered by a straight tube A E, soldered on the plate. For the plate, ECD, of tin, any other metal, or a slice of a large cock may be substituted. The bent tube,

is seen in such a position that the plate CD may be nearly horizontal. On this plate the disk C'D' of any material, flexible or inflexible, is placed: now, on blowing at A with considerable force, the disk, however tight it may be, will not quit the plate. Inverting the tube, as shown in fig. 4, and adding at A a second tube A a, fitting close into the extremity A of the first tube A B, on blowing at A, the wind passes through the orifice E and enters into the atmosphere by the cylindrical zone comprised between the edges of the plate C D, and of the disk C' D'. Not only the disk does not fall, but it is driven towards the plate CD by a force greater than is necessary to be in equilibrium with its weight.

102. The tin plate strips soldered on the edge of the plate C Ď, fig. 4, end in a ring G H. A support, GʻII', of cork or other materials, slides and is held between the strips, and supports a disk of paper or pasteboard C" D”, at any desirable distance from the plate C D. Regulating this distance suitably, and blowing at A', the disk C"D" will be seen to approach the plate CD, and to take the position C' D', very near to the plate CD.

103. The same effects will take place with the disk C'D', fig. 3, on blowing up the extremity A of the straight pipe A E, held in a position nearly vertical.

104. When the disk C'D' is flexible and a little elastic, and the experimenter blows into A, figs. 2 and 3, or into A', fig. 4, a sound is produced, resulting from the successive beatings of the disk on the plate CD.

105. The air is driven from the mouth hole A of the tube towards the orifice E of the plate CD, and strikes the part of the disk opposite that orifice, and the mean pressure on that part of the disk is greater than the atmospheric pressure. The blast of air takes the place of the air included between the plate and the disk opposite to it. It moves in that interval with a velocity which decreases from the edges of the orifice. The elastic force of that air decreases at the same time, so that the mean pressure between, and the inner face of the disk, becomes less than the atmospheric pressure, and, as the latter pressure is exerted on the whole of the outer face of the disk C'D', this disk, subject to two contrary pressures on opposite faces, obeys the greater; whence it follows that the disk C'D' must be driven towards the plate CD.

106. It is necessary for the disk C'D' to be near the orifice E of the pipe A E, in order that the shock of the air may be modified by the atmospheric pressure.

107. Let C D' CD, fig. 5, be a vessel in the form of a cymbal, composed of a hollow cylinder, CDEF, and of a flat rim, in width equal to CF or GD". Having fitted to the bottom, CD, a pipe, A E, which covers the orifice E of the diameter of three millimetres (= 0·118 inches), on blowing into A against the disk C'D' near the flat rim C" D", this disk is impelled towards the orifice E.

108. The vessel and pipe is represented, fig. 5, of about half the real size. The weight of the disk, increased by that of the bodies attached to it at P, amounts to above twelve grammes (=

185 English grains). This weight is the measure of the pressure resulting from a common blast at A in the upper extremity of the pipe A E.

109. When we have blown repeatedly on the disk C'D', it becomes covered with moisture, and we can see the furrows of the threads of air in the direction of radii, and ending at a little circumference nearly of the same diameter as the orifice E.

110. The disk C'D' being fifty-four millimetres (2.13 inches) in diameter, the pressure of the atmosphere on that disk is equal to a weight of twenty-three killogrammes (50-74lbs. avoirdupois); whence it follows that, in this experiment, the pressure of the air blown against the inner face of the disk, and the atmospheric pressure exerted on the exterior face of the same disk, differ from each other only about one half a thousandth part of the second pressure. All other circumstances remaining the same, the form of the orifice of the plate modifies the phenomena. When that orifice is a lengthened rectangle, or a cross, fig. 3, the difference of the pressures on the opposite faces of the disk diminishes considerably. The following experiments have for their object to measure these pressures, in the situation in which, the plate and the disk being circles of the diameter, the orifice of the plate is also a circle.

111. The following experiments on the motion of air between two plane surfaces, are no less curious and important. A bent tube, B B', fig. 6, was fitted on the sides of the chest of a smith's Distances between the disks C D, C' D'.

113. It is evident from this table, that when the distance from the disk to the plate is only a millimetre (0-039 inch) the wind from the bellows enters into the atmosphere, through a cylindrical zone of 312 square millemetres, and its height one millimetre. When the distance is thirteen millimetres (= half an inch), the surface of the cylindrical zone is 4082 square millimetres. For the first distance of a millimetre, the zone through which the air passes is smaller in surface than the orifice; and, for the second distance of thirteen millimetres, it is ten times larger. In both cases, the action of the shock of the air blown against the disk by the bellows, is diminished by the atmospheric pressure.

114. The combination of the shock of the air and the pressure of the atmosphere, does not take place between two plane surfaces only. The plate being supposed terminated by a plane face, the opposite face of the disk may be slightly convex. A great convexity would too far remove the disk from the plate, and, if the face of the disk were concave, the shock of the air on that face would no longer be counterbalanced by the opposite atmospherical pressure.

115. The metallic plate soldered to the extremity of the hellows-pipe mentioned before, is 125 millimetres (4.9 inches). M. Hachette placed on this plate a disk of smooth cardboard, and stuck on successively a series of pieces of paper, in order to get the number of pieces which would by their weight be equal to

forge-bellows; the bellows being put in motion, by means of the lever, the air in the chest was kept at a constant pressure, which was measured by a column of water by means of a tube with a double bend, having one end fixed in the chest of the bellows. The air arrived through rightangled pipes B, B, B', and issued from the orifice E, made in the centre of a wooden disk CDcd; another disk C' D'H', fig. 11, carried a solid rod or tail H' H, passing through a plate G G', and sliding in the case K K', several upright, CG, DG', are fixed in the parallel disks CD, GG. At the distance of thirteen millimetres, the pressure of air from the bellows against the interior face of the disk becomes equal to the atmospherical pressure against the face opposite. In this first series of experiments, the rod HH' was supported by a line HQ P, passing over a pulley having its axis of rotation at R, and a weight to balance the friction, the weight of the disk C'D' and its rod H H' was previously put on the disk B.

112. In order to continue these experiments, the line HPQ of the rod H H' was detached, and the weights were placed on the platform of that rod. When the distance between the disks exceeds thirteen millimetres (half an inch), the shock of the air predominates over the atmospheric presure, the disk is raised, the weights which maintained it at the distances stated in the first column of this table, were observed to be as follows:

22

Weights in equilibrium with the shock of the air of the bellows against the disk C' D'. 35 grammes 540 grains E. = 340. the atmospheric pressure, while the bellows were in motion. This number of pieces increased considerably when the disk of card-board presented a slight degree of convexity towards the plate. This effect of the change of curvature was also confirmed by experiments on the flowing of water.

116. We must now examine the motion of air between a circular plate and a flexible and slightly elastic disk of the same diameter placed upon the plate. Fig. 7 represents a disk of smooth paper rather thin, C'D' is put upon the plate CD, fig. 2; having wetted this disk about its centre, by touching it with a drop of water carried at the end of the finger, a gentle blast is to be given at A, the extremity of the bent pipe ABCD. The paper being a little transparent in the moistened part, the orifice E in the plate may be seen, and, during the blast at A, the wetted part swells up from the inside outwards, opposite the orifice E, and continues that curvature; the rest of the disk trembles, and a whistling or humming noise is heard; by blowing with more force the shock of the air overpowers the atmospheric pressure, and the paper disk flies away. These phenomena become more perceptible on a paper disk of greater diameter. I have put upon the metallic pipe of 124 millimetres in diameter, soldered to the extremity of the pipe of a pair of chamber bellows, a disk of brown paper, rather thick and moistened; on working the bellows the disk swells up, as in

the preceding experiment, opposite the orifice; at a certain distance from that orifice the disk is depressed, and it separates from the edges of the plate in order to allow a passage for the air. The depression stops, momentarily, the communication of the air between the centre and edges of the plate; the air, the flowing of which is interrupted, increases in elastic force, and opens itself a new passage. The preceding depression and inflections of the disk are renewed, producing very intense irregular sounds, which mix with those of the metallic plate.

117. The motion of a liquid between two surfaces must now be compared with the motion of an aëriform fluid between the same surfaces. 118. The motion of an aeriform fluid or a liquid, which we compare, takes place between the surfaces, S, S', sufficiently near together to prevent the atmospheric air from penetrating the space between the two surfaces.

119. When the aëriform fluid contained in a vessel passes under a given pressure into that space, the fluid fills it by its expansibility, and it enters into the atmosphere through a zone which has for its limit the edges of the two surfaces S, S', or of only one of them; the perimeter of this zone being greater than that of the orifice made in the surface S, through which the fluid issues from the vessel containing it, it follows that the velocity of the fluid is decreasing from the orifice to the edges of the zone through which it flows into the atmosphere; and, as the fluid in motion fills the whole space comprised between the zone and the orifice, it loses a very considerable part of the elastic force which it had in the interior of the vessel, in order that its mean pressure against the surface S' may be less than the atmospherical pressure. The expansibility of the fluid is not a necessary element of the difference of the pressure exerted on the opposite sides of the surface S'. By substituting a liquid, in place of an aëriform fluid, the adherence of the liquid to the surfaces S, S', answers instead of the expansibility: these surfaces being sufficiently near together, the atmospheric air does not introduce itself into the space which separates them; the liquid fills that space, from which it issues to enter into the atmosphere The velocity of the fluid decreases as with the aëriform fluid, from the orifice made in the surface S, to the edges of the surface S', and the mean pressure exerted by the liquid inside, against one side of the surface S', is less than the atmospheric pressure against the surface opposite.

120. We have hitherto considered the nature

of air with reference to its motion and elasticity, without examining the effects which result from a change of temperature. Now the air that we breathe is continually operated upon in the great laboratory of nature by the action of solar and other heat.

121. The following table contains the expansion of 1000 parts of air, nearly of the common density, by heating it from 0° to 212°. The first column contains the height of the barometer; the second contains this height augmented by the small column of mercury in the tube of the manometer, and therefore expresses the density of the air examined; the third contains the total

124. If we would have a mean expansion for any particular range, as between 12° and 92°, which is the most likely to comprehend all the geodetical observations, we need only take the difference of the bulks 26-038 and 222.006= 195-968, and divide this by the interval of temperature 80°, and we obtain 2-4496, or 2.45, for the mean expansion of 1°.

125. It would perhaps be better to adapt the table to a mass of 1000 parts of air of the standard temperature 32°; for in its present form it shows the expansibility of air originally of the temperature 0°. This will be done with sufficient accuracy by saying (for 212°) 1071-718 : 1484-210 1000: 13849, and so of the rest.

126. This will give for the mean expansion of 1000 parts of air between 12° and 92 2.29. 127. Although it cannot happen that in measuring the differences of elevation near the earth's surface, we shall have occasion to employ air greatly exceeding the common density, we may insert the experiments made by general Roy on such airs. They are expressed in the following table; where column first contains the densities measured by the inches of mercury that they will support when of the temperature 32°; column second is the expansion of 1000 parts of such air by being heated from 0° to 212°; and column third is the mean expansion for 1°.

128. We have much more frequent occasion to operate in air that is rarer than the ordinary state of the superficial atmosphere. General Roy accordingly made many experiments on such airs. He found in general that their expansibility by heat was analogous to that of air in its ordinary density, being greatest about the temperature 60°. He found, too, that its expansibility by heat diminished with its density, but he could not determine the law of gradation. When reduced to about one-fifth of the density of common air, its expansion was as follows:

From this very extensive and judicious range of experiments it is evident that the expansibility of air by heat is greatest when the air is about its ordinary density, and that in small densities it is greatly diminished.

129. The facts we have now been examining lead us to a very important part of domestic economy, namely, the best mode of heating and ventilating houses by reference to a change in the atmospheric density, but new and more valuable information has since been furnished by the researches of Mr. Tredgold.

130. In winter we require artificial heat, and during a short part of summer we seek for coolness, but at all times we need pure and wholesome air. These, however, are comforts which are not always to be commanded, and particularly where we desire to join economy with healthiness and comfort. The principles concerned in the movement of invisible elastic fluids are seldom understood by those who engage in the management of ventilation; and, on the still more recondite subject of heat, we too often find that the most absurd opinions are entertained. On the other hand, persons rarely take the trouble to think for themselves; and, most likely, because very little pains have been taken to reduce the subject to principles, or to render it accessible to those who would wish to be acquainted with it; and more especially those who would wish to be able to distinguish quackery from science.

131. This seems an anomaly that can be explained only by the powerful influence of habit, which leads us in the steps of our forefathers, while in other arts changes have been made lation of our dwellings. For in their large which render it necessary to improve the ventimansions the wind was suffered to blow freely through them, and a current of air to circulate wainscot and the wall. It must be habit also through the wide space between the pannelled that renders the constant attendance at the bench or the bar supportable in the noxious atmosphere and elevated temperature of a court of justice. It must be habit which causes the offensive effluvia of a hospital to be disregarded by medical men, for surely these are not necessary evils; but before I visited hospitals, courts, manufactories, and poor-houses, for the express purpose of seeing how they were ventilated, I had no idea of the magnitude of these evils. All places are not equally ill ventilated, for there are some