retorts, which may be of cast iron or of clay, the first substances evolved are steam and atmospheric air. The former is an objectionable product for several reasons, and the coal should consequently be kept as free as possible from moisture before it is used. The conversion of water into steam absorbs a portion of heat, and serves to keep down the temperature in the retort, causing the evolution of tarry matters, which at a higher temperature would have been converted into gas. The steam also is liable to be decomposed in coming in contact with portions of the fuel heated to redness, and its oxygen thus unites with proportionate quantities of carbon, producing carbonic acid and carbonic oxide, and thus robs the coal of a part of its chief gas-producing elements. The hydrogen set free passes over uncombined with carbon. While the retorts are at a low red heat the ingredients of the coal tend to form chiefly tarry matters, and with them come over aqueous and ammoniacal vapors and sulphurous acid. The ammoniacal product derives its nitrogen from the atmospheric air contained in the retorts, and as it comes over unites with the sulphurous acid. A portion of the sulphur of the pyrites present in the coal unites with hydrogen, forming the gas sulphuretted hydrogen, and a portion uniting with carbon forms the very injurious product sulphuret of carbon. The numerous oily compounds, exceedingly rich in carbon, are also produced, which remain in part with the tar, and in the form of vapor also mix with the carburetted hydrogen and olefiant gas, which are at the same time evolved, adding largely to the illuminating property of the mixture. At a full red heat the gaseous mixture sought for is almost exclusively produced; and it is therefore an object to bring the contents of the retorts as rapidly as possible into this condition to avoid their being wasted by their elements forming comparatively useless compounds. The retorts are for this reason not allowed to cool when discharged of the coke residuum and recharged with fresh coal. In the old practice of 8-hour charges it was estimated that in the first hour of the operation 20 per cent. of the whole product of illuminating gas was obtained, and in each succeeding hour the following proportion: 2d, 15 per cent.; 3d, 14; 4th, nearly 13; 5th, 12; 6th, 10; 7th, 9; 8th, 8. The gas at first has but little illuminating power. It is the best in the early part of the process, while the coal is becoming thoroughly heated, its specific gravity then being 0.650, after which it gradually falls to about 0.345 at the close of the operation. At 3 different stages of the distillation continued for 10 hours the mixture has been found to be in 100 measures as follows: 1st, carburetted hydrogen 82.5, olefiant gas 13, carbonic oxide 3.2, nitrogen 1.3-sp. gr. 0.650; 2d, at 5 hours, carburetted hydrogen 56, olefiant gas 7, carbonic oxide 11, hydrogen 21.3, nitrogen 4.7 -sp. gr. 0.5; and at the close of the operation olefiant gas had disappeared, carburetted hydro

gen was 20, carbonic oxide 10, hydrogen 60, nitrogen 10-sp. gr. 0.345. Though the presence of any of these ingredients is not regarded as impairing the technical purity of the gas, its quality for illuminating purposes is greatly affected by the different proportions in which they are found; and by continuing the operation too far it is obvious that the proportion of non-illuminating compounds rapidly increases. Sulphuretted hydrogen is especially apt to make its appearance, and become intermixed in the last stage of the process, seriously affecting the purity of the mixture. Other substances which contaminate it are the vapors of water and ammonia, formed in the first part of the operation, and which should be condensed and separated in the tar receiver, and also carbonic acid and sulphurous acid gas. At high temperatures the retorts become coated with a dense carbon of excessive hardness, which is supposed to be deposited from gaseous carbon separated from light carburetted hydrogen, hydrocarbon vapors, or from olefiant gas, as these are decomposed in passing over very hot surfaces. The latter gas may be made to yield up a portion of its carbon in this form by conducting it through red-hot tubes, and the volatile hydrocarbons are known to be similarly affected. High temperature is therefore injurious from its tendency to decompose the most valuable products of the distillation, and reduce the substances to their elementary condition or simple combinations nearly approaching it. The methods in use, however, for effecting the decomposition of carbonaceous compounds, which produce illuminating gas, do not admit of the complete regulation of the temperature which a perfect operation would require; and there is hence a necessary waste of gas in the manufacture and uncertainty as to its exact composition. The duration of the process is now generally reduced to from 4 to 6 hours, depending upon the kind of coal and size of the retorts, and the richest gases are evolved during the first hour. In Scotland the rich cannels are sometimes carbonized in 3 hours. By practice and careful attention to the qualities of the materials employed, manufacturers succeed in producing at the same works an article not varying greatly in composition; but the gas of different establishments is found to be of very different qualities. Its value is estimated by referring it to some standard with which the illuminating power of a given quantity is compared by actual trial. Five cubic feet of gas are commonly consumed in an hour by an Argand burner, and the light is compared by means of suitable instruments with that of spermaceti candles, each burning 120 grains in an hour. The gas of the London works varies with the coals employed, from an illuminating power of 11.5 to 18 candles, the best results being with varieties of cannel, which also commonly afford a few hundred feet more of gas per ton. The range at these works in variable amount of product rarely exceeds 1,000 cubic

England, and by many are considered decidedly superior to those of metal.* The iron retorts are made of various forms-plain cylinders 7 feet long and 1 foot diameter, also of oval cross section; and again of the same modified by the bottom being bent in toward the centre, giving to the cross section a kidney or ear shape; but the last, which were formerly in high favor, are now entirely out of use. The kind generally preferred at present is the so called D-shaped retort, its cross section having the shape of this letter, with the flat side for the bottom. They are from 7 to 9 feet in length, from 1 to 2 feet in width, and from 12 to 15 inches in height. Many works use what are called "through retorts," which are from 16 to 18 feet in length. These are charged at each end. Experienced engineers regard them as the most economical. Retorts are set in groups of 3 to 13 in each furnace. In large establishments the furnaces are double stacks, 2 single retorts placed end to end or one through retort reaching through, so that the neck or mouthpiece of each projects through the exterior wall; the fire places under the retorts also open out on either side, and the flues from all of them unite in one along the central line of the stack. The retorts are set in the brick work in 2 or 3 horizontal rows, 5 for each furnace with a fire under each of the 2 or 3 retorts of the lower row. The flame from the fires can pass among them all before escaping by the flue. The furnaces succeed each other in the same stack through the length of a large building, the retorts in large establisments numbering many hundred, of which all may be kept in operation in the winter season, and part only in the summer. At the great central gas works in London a new form and arrangement of retorts have been tried within a few years, by which it was thought great economy in consumption of fuel for distilling the coal was effected, together with increased production of gas and other advantages. But after the most flattering reports, the trial does not seem to have been satisfactory, as the company are said to have returned to the old plan. Thirteen retorts were arranged over one furnace fire, set in 2 series, the upper of 6 retorts, 18 feet in length, made of fire clay, only their necks of cast iron pro

A late paper on clay retorts by Mr. J. Church, read to the London institution of civil engineers, presents the following facts relating to the comparative quantities of gas made by the two kinds of retorts of the form, of 15 inches by 18 inches in section, and 74 feet in length. Iron retorts lasting 365 days, and working off 1 cwt. of coal for each charge, effected the carbonization of 2,190 cwt. of coal, which at 9,000 cubic feet of gas per ton gave a total quantity of 985,500 cubic feet of gas per retort; while the clay retorts lasted 912 days, carbonized 5,472 cwt. of coal, which at 9,000 cubic feet of gas per ton gave 2,462,000 cubic feet of gas per retort. It will thus be seen that the clay retorts yielded a greater quantity of gas from the same weight of coal than the iron retorts; but

the specific gravity of the gas thus made was less, and its illuminating power was diminished in consequence of the increased temperature of the clay retorts, which caused the last portions of the gas to be decomposed. Clay retorts in large establishments are best worked with the addition of an exhauster. The pressure is thus reduced, and, the gas not escaping through the pores, the quantity is increased about 200 cubic feet per ton of coal. In small works, the exhauster

would not save the expense of its introduction, requiring as

does steam power to work it.

it

jecting beyond the brick work. The lower set were of iron 19 feet long, going like those above completely through the stack, and like them they were charged and discharged at both ends. They were heated by the descending current of flame, which had first been led around the earthern retorts. These bear a higher heat, and they also retain it much better, than the iron retorts; they were charged every 4 hours, while the others were charged every 6 hours. The average yield of each retort was 18,500 feet of gas per day. By this arrangement a large number of retorts of great capacity is concentrated in comparatively little room. Clay retorts accumulate little of the solid carbon, which is apt to form a hard lining in the iron ones, requiring them to be frequently cleaned out with great labor. That which collects in the former is easily removed in flakes with a bar on leaving them open at both ends so that a current of air can pass through. An iron retort is estimated to last through the production of 700,000 feet of gas; 6 times this amount has been obtained from a "through" clay retort. All forms of retorts are provided with a mouthpiece, commonly of the same size with the body, to which it is attached by bolts, if the parts are not both in the same piece. This mouthpiece projects in front of the brick work, and the pipe called the stand pipe, usually about 4 inches in diameter, for conveying off the gases, enters through its upper side, and rises 4 or 5 feet up the front of the brick work, then bending back passes down by what is called the dip pipe into the large horizontal pipe laid on the top of the brick work, which is called the hydraulic main, and receives the products from all the retorts in the furnaces beneath it. The retorts are charged with coal by means of a long semi-cylindrical scoop of sheet iron, which being filled, and the old charge being raked out, is pushed into the red-hot retort; by means of the handle it is twisted over so as to empty itself, and is then withdrawn. A charge of 120 to 200 lbs. of coal, about half filling a single retort, is thus introduced at once. Before much gas can escape, the cast iron cover, coated with suitable luting, is brought into place and screwed fast. In discharging a retort the cover is first loosened, and the gas which escapes is ignited. By this means an explosion is avoided, which would be likely to occur if the interior were suddenly exposed to the atmosphere. The coke, as it is withdrawn, bringing with it a large amount of heat, is sometimes used in the furnaces at once with great economy, no time being allowed for the heat to escape. The hydraulic main is about half filled with water or fluid tar, which is made to retain its level by a discharge pipe suitably placed. Into this the dip pipes penetrate about 3 inches; and the aëriform fluids, pressed forward from the retorts, readily pass up through the denser fluid to the reservoir above, but are prevented by it from turning back through the pipes when hydraulic main is a very strong pipe of east or the flow from the retorts is interrupted. The

of boiler plate iron, often a foot and a half in diameter, placed exactly level, and extending the whole length of the furnace stacks. It serves to collect a large portion of the tar and ammonia; but a part is kept by the heat in the state of vapor, and is carried along with the gas. To separate this, the mixture is cooled by passing it through a series of tall condensing pipes of iron placed in the open air, and sometimes kept cool by water made to trickle down their outside surface. The gas goes up one and down another in succession. At the bottom of each pair a pipe communicating with a cistern in the ground conveys to this the tar and ammonia which separate from the gas. From the condensers the gas in some works passes into the bottom of a large tower nearly filled with paving stones or coke, and having a contrivance at the top for sprinkling a fine shower of water in the space above the contents. After rising up into this space, the gas, deprived of still more tar and ammonia, passes down by a pipe which leads under the tower and terminates in the bottom of another similar one close by, discharging the gas among the stones or coke. From the top of this it is conveyed down again by a large pipe which passes under ground to the next apparatus. The towers are called scrubbers; the sprinklers at the top consist of hollow cross-arms perforated with small holes and kept revolving by steam power. The passage of the gas through different parts of the apparatus is continually subjected to increased resistance, which is all turned back upon the retorts. The weight of the gas holders at the end of the series adds largely to this pressure. The effect of it is found to be a proportional increase of the carbonaceous deposit in the retorts; and this causes them to burn out rapidly, and moreover is produced, it is believed, at the expense of the most valuable light-producing vapors. In the use of clay retorts the gas is even forced by the pressure through the porous material. The gas obtained is also of poorer quality by reason of the pressure; and the materials tend to produce a large proportion of tar. To obviate these injurious effects an apparatus has been introduced within a few years, called the exhauster, which, placed next to the scrubber, pumps up the gas and forces it along. A highly approved form in use at Liverpool is an oscillating machine of the plan of the ventilateur du Hartz, described at the end of the article BLOWING MACHINES, to which reference may be made for its description. In New York a revolving exhauster is employed on the plan of that of Jones of Birmingham or Beales of London. The next operation is to separate from the gas the mixed sulphuretted hydrogen and carbonic acid; also any ammonia that may be present (if the scrubber apparatus has not been used), combined with either carbonic, hydrochloric, sulphurous, or sulphuric acids. For this purpose the mixture is passed into either the wet or dry lime purifier. In the former it is distributed under a surface of lime water, or milk

of lime, so as to bubble up through the liquid, which at the same time is stirred by a rapidly revolving arm. Provision is made in some machines for drawing off the old liquid and supplying fresh portions, as the lime is exhausted by combining with the chlorine and hydrosulphuric acids. But a better method is to employ two pairs of machines, and, as the lime in one becomes spent, replace them with the other. The proportion of lime required varies of course with that of the sulphuretted hydrogen and carbonic acid present. As a general allowance, a bushel of quick lime mixed with 48 gallons of water should, if properly applied, purify 10,000 cubic feet of gas. The product should be frequently tested to determine its purity. This is done by presenting a piece of paper moistened with solution of acetate of lead to a jet of the escaping gas. If it becomes discolored at all, the gas is impure. The dry lime purifier consists of an iron box, until recently about 6 feet in length, 5 in breadth, and 3 in depth, but now sometimes 25 feet square, arranged with several shelves, one above another, of iron grating, which are covered with lime to the depth of about 3 inches. Movable trays are conveniently substituted for the shelves. The lime is sprinkled with water to slake it. The gas, let in below through a large pipe, passes up through the lime, which combines with and retains its impurities, and the gas is then conveyed away by another pipe. The French arrange the slaked lime upon trays in layers with moss, which serves to divide the gas into minute jets and multiply the points of contact with the lime. They spread among the moss a little less than a bushel of lime to the square yard of surface; and they use about 24 bushels to a ton of coal, the average product of which is less than 7,000 cubic feet of gas to the ton. The purifier is of circular form, and the rim of the cover shuts down into a ring containing water, thus making a gas-tight joint, while the cover can be moved on and off with perfect freedom. The dry lime purifier is generally preferred to the wet lime; the passage of the gas is less obstructed, and the product is less offensive to the neighborhood. By connecting the interior of the purifier by means of a large pipe with the flue of the chimney, the current of air allowed to enter draws up the offensive gases from the lime, and they are decomposed in the hot air of the chimney. This would seem to be a hazardous process, endangering explosion. Tomlinson refers to it, but we have only known the foul gases to be conducted to flues unconnected with fires, or pumped out and passed through copperas to decompose them. The cover of the purifier being raised, the lime can be taken out without causing any annoyance; and it is then sometimes burned in ovens and afterward used a second time, but the common practice is to throw it away after once using. From the purifier the gas is conducted to the station meter, the instrument for meas uring and registering the quantity produced

during any period of time. It is constructed upon the principle of the meters used by each consumer, but on a much larger scale. A cylindrical box more than half filled with water contains a drum supported upon an axis which passes from one end to the other of the outer case. This drum is divided into 4 hollow compartments, each of which has an opening into the central space around the axis, and also slits in the rim communicating with the space next the outer case. The gas is conducted into the central space and flows into one of the partially submerged compartments. As this fills, the pressure of the gas causes the drum to begin to revolve, and as it passes round the next compartment comes in position to be filled, and so on, each chamber as it is submerged in its revolution discharging its measure of gas into the space above the water next to the outer case, from which it is conveyed away to the gas holder. The capacity of the chambers being known, and the number of revolutions recorded by a train of wheel work connected with dials on the outside of the case, the whole quantity of gas passed through is at any time seen. But to give uniform results the water must always be at the same level; as it rises it diminishes the gas capacity of the compartments, and as it sinks it enlarges this, and more gas is passed with each revolution. The station meters are also provided with a clock, to the minute hand of which is attached an index carrying a pencil. By the lines this traces upon a disk of paper attached to a plate which is affixed to and revolves with the axis of the drum, any irregularity in the supply, and the time of its occurrence, are registered. The dials are provided each with a single pointer; one marks the number of tens of cubic feet, another the number of hundreds, another thousands, &c., up to tens of millions. The largest instruments of the kind pass about 650 cubic feet in one revolution of the drum, and register in an hour about 70,000 cubic feet. The next apparatus is the gas holder or reservoir. It is a cylinder of plate iron suspended or floating with its open lower end in a cistern of water. As the gas is admitted beneath, the cylinder is lifted up by it; and as it is required for consumption, the pres sure of the gas holder forces it through the pipes laid for its conveyance. These receptacles are of immense size for the supply of large cities. One in Philadelphia is of the capacity of more than 1,000,000 cubic feet, being 140 feet in diameter and 70 feet high. The imperial gas company of London have one 200 feet in diameter and 100 feet high. To obtain room for gas holders of great height to descend, so that their tops may come down to the level of the water, and the gas be all expelled from them, they are made in sections which shut one within another like the parts of a telescope. The lower edge of each section except the lowest curves up in a flange directed outward, and forming a ring. As the section rises out of the water this ring comes up filled with it, and

1

catches the flange of the next section below, which is bent inward for this purpose. The water has the effect of making the joint proof against the escape of gas.-Between the gas holder and the main pipe is placed an apparatus called the governor, contrived to regulate the pressure of the gas as it is admitted into the main; and it is found useful to repeat these machines wherever the gas is distributed at points varying considerably in elevation, as from its low specific gravity the pressure increases with the elevation, the rate being of an inch for every foot of difference of height. It has even been recommended to use one for every 30 feet of rise in the ground through which the pipes are carried. In construction it resembles a gas holder, an inverted cylinder being suspended with its lower part in water so that as the gas comes in from below the upper portion is lifted. This carries suspended from its centre a cone of cast iron, turned true, and adapted when raised sufficiently high to fit closely the bevelled opening by which the gas enters the central part of the apparatus. The outlet pipe commences in the upper portion and passes down through the bottom. The cone and cover are nearly counterbalanced by a weight, which passes over a pulley outside of the machine. When the pressure of gas is moderate the cover and cone descend, leaving a wide opening for the gas to enter. When the pressure increases the cone rises and checks the flow.-The gas, being now delivered into the main, is distributed to the various points where it is consumed. Each customer is furnished with his own meter, which registers on the principle explained the quantity he consumes. Small pipes convey the gas from the meter to the various burners affixed at their terminating points. Upon the form and condition of these the economical consumption of the gas in proportion to the light produced in great measure depends. Each one is furnished with a stopcock upon its own supply pipe, and by means of this the quantity admitted to the burner is regulated. The opening at the end for the gas to escape is often in the form of a narrow slit. This gives to the flame the form of a thin sheet known as the bat's wing; and a similar effect is produced by two small round holes in the end of the burner inclined toward each other. The principle of the Argand burner is explained in the article ARGAND LAMP. In applying this principle to gas, the burner is a hollow ring perforated with holes about of an inch apart, and measuring of an inch in diameter, bored with the greatest accuracy. These are large consumers, but give a brilliant light; and in most burners this is attained at the expense of an unnecessary quantity of gas. Economy in this respect and a flame free from flickering have been most important desiderata in the construction of all new burners, of which there is the greatest variety. No burner in use appears to combine these advantages so admirably as that known as "Gleeson's American gas burner,"