sity, soon brought it into great favour with the work-people. And its being free from the inconvenience and danger, resulting from the sparks and frequent snuffing of candles, is a circumstance of material importance, tending to diminish the hazard of fire, and lessening the high insurance premium on cotton-mills. The cost of the attendance upon candles would be fully more than upon the gas apparatus; and upon lamps greatly more, in such an establishment as Mr. Lee's. The preceding statements are of standard authority, far above the suspicion of empiricism or exaggeration, from which many subsequent statements by gas-book compilers are by no means exempt. At the same manufactory, Dr. Henry has lately made some useful experiments on the quality of the gas disengaged from the same retort at different periods of the decomposition. I have united in the following table, the chief part of his results. He collected in a bladder the gas, as it issued from an orifice in the pipe, between the retorts and the tar pit; and purified it afterwards by agitation in contact of quicklime and water. Ten cwt. or 1120 lbs. of coal were contained in the retorts. 100 measures of purified gas contain, 100 measures Hours from com of impure gas contain, gas mencement. Other Con Sulph. Carb. hydr. acid. Olef. infl. Azote sume gases. oxyg. す 54 16 64 20 180 94 225 118 Dr. Henry conceives that gas to have the greatest illuminating power, which, in a given volume, consumes the largest quantity of oxygen; and that hence the gas of cannel coal is one-third better, than the gas from common coal. 3500 cubic feet of gas were collected from 1120 ponnds of the cannel coal; and only 3000 from the same weight of the Clifton coal. From the preceding table, we see also that the gas which issues at the third hour contains, in 100 parts, of sulphuretted hy drogen and carbonic acid, each 24, of azote 44, olefiant gas 144, and of other inflammable gases 76 parts. A cubic foot of carbonic acid weighs 800 gr. A cubic foot of sulphuretted hydrogen weighs 620. The first takes about 1026 gr. of lime for its saturation; the second about 1070; and hence 1050, the quantity assigned by Dr. Henry for either, is sufficiently exact. 100 cubic feet of the above impure gas, containing 5 cubic feet of these two gases, will require at least 2100 grains What is separated by the first washing is probably vapour of bitumen or petroleum, which would injure the pipes by its deposition, more than it would profit, by any increased quantity of light. Though we thus see that the second washing in the above experiment condensed none of the olefiant gas, it is prudent not to use unnecessary agitation in a large body of water. The carbonate of lead precipitated from à cold solution of the acetate, by carbonate of ammonia, washed with water, and mixed with a little of that liquid into the consistof cream, is well adapted to the separation of sulphuretted hydrogen from coal gas. The carbonic acid may then be withdrawn from the residuary gas, by a little water of potash. We must now determine the azote present, which is easily done by firing a volume of this gas with thrice its volume of pure oxygen. What remains after agitation with water of potash, is a ence mix termine exactly the proportion of each me tal present, because the volume of the alloy is very nearly the sum of the volumes of its ingredients. I have long applied this problem to gaseous mixtures, and found it a very convenient means of verification on many occasions, particularly in examining the nature of the residuary air in the lungs of the galvanized criminal, of which an account is given in the 12th Number of the Journal of Science. PROBLEM.-In 100 measures of mixed gases, consisting, for example, of olefiant gas, carbonic oxide, and subcarburetted hydrogen, in unknown proportions, to determine the quantity of each. The first step is to find the quantity of the two denser gases, which have the same specific gravity = 0.9720. RULE.--Multiply by 100, the difference between the specific gravity of the mixture, and that of the lighter gas. Divide that number, by the sum of the differences of the sp. gr. of the mixture, and that of the denser and lighter gas; the quotient is the per-centage of the denser. See Gregory's Mechanics, vol. 1. p. 364. EXAMPLE.--A mixture of olefiant gas, carbonic oxide, and subcarburetted hydrogen, has a sp. gr. of 0.638. What is the proportion per cent of the first two? And 0.972 8.8 015 = 20 volume of the two hea. vier gases; and therefore there are 80 of the lighter gas. Hence, having fired the whole with oxygen, we must allow 160 of oxygen, for saturating the 80 measures of the subcarburetted hydrogen. Then let us suppose 35 cubic inches more oxygen to have been consumed. We know that the saturating power of olefiant gas, and of carbonic oxide with oxygen, is in the ratio of 3 to 0.5. Therefore, the quantity of olef. gas 35-(20 × 0.5) 25 3-0.5 2.5 10 measures. ture of azote and oxygen. Explode it with We see now, that a gas of sp. gr. 0.638 : do. subcarb. hydrogen For further details see Gas. 0.444 0.097 0.097 0.638 Dr. Henry gives, at the end of his experiments, (Manchester Memoirs, vol. iii. second series), some hypothetical represent ations of the constitution of coal gases, in 39 2 of carburetted hydrogen, and 15 of pure hydrogen, in 18 mea sures. With mixtures of three gaseous bodies, the problem of eliminating the proportion of the constituents, by explosion with oxygen, becomes complex, and several hypothetical proportions may be proposed. But I can hardly imagine, that pure hydrogen should be disengaged from ignited coal. There is no violation of the doctrine of multiple proportions, in conceiving a compound to exist in which three or more atoms of hydrogen may be united with one of carbon. Berthollet's experiments render this view highly probable. If the above hypothetical numbers were altered to 1.6; 2.4; and 15; their accordance with Dr. Henry's experiments would be improved. Now, this is a considerable latitude of adjust ment. The principles laid down at the commencement of this article show, that the more uniformly the coal undergoes igneous decomposition, the richer is the gas. The retorts, if cylindrical, should not exceed, therefore, 12 or 14 inches diameter, and six or seven feet in length. Compressed cylinders, whose length is 45 feet, breadth 2 feet, and inside vertical diameter about 10 inches, have been found to answer well at Glasgow. The cast iron of which they are composed, must be screened from the direct impulse of the fire, by a case of firebrick. On the maximum quantity of gas procurable from coal, it is difficult to acquire satisfactory information, at the great gas establishments. Exaggeration seems to be the prevailing foible. Mr. Accum gives the following tables, as the maximum results of his own experiments, made at the Royal Mint gas-works; Scotch cannel coal, Lancashire Wigan cannel, Yorkshire cannel, Wakefield, Cubic feet of gas. 19.890 19.608 18.860 9.748 10.223 3d do. 10.866 Staffordshire coal, 1st variety, Concerning the duration of the decomposition of a retort-charge of one cwt., various opinions are maintained. Mr. Peckston's experiments at the gas light and coak company's works, Westminster station, seem to prove, that decided advantages attend the continuance of the process for eight hours, in preference to six, or any shorter period. The average product of gas, from one chaldron of Newcastle coals, at six hours' charges, he states at 8,300 cubic feet, and at those of eight hours, at 10,000. On 76 retorts, worked for a week at the latter rate, he gives a statement to prove, that there is a saving of 771. 18s. above the former rate of working. Two men, one by day, and one by night, can attend nine or ten retorts, at eight hours charges, of 100 pounds of coal each. Scotch cannel yields its gas most rea dily, or 1.00 1.04 1.08 1.18 1.65 Newcastle coal, The following table by Mr. Peckston exhibits the ratio at which the gas is evolved from Bewicke and Crastor's Wall'send coal, when the retorts are worked at eight hours' charges: Cubic feet. Sum. During the 1st hour are ge nerated, 2000 2d, 1495 3495 3d, 1387 4882 4th, 1279 6161 5th, 1189 7350 6th, 991 8341 7th, 884 9225 8th, 775 10000 Gloucester coal, High Delph, Do. Low Delph, Do. Middle Delph, Newcastle coal, Hartley, 9.796 12.096 '16.920 Cowper's High Main, 15.876 We have already explained the princi16.584 ples of purifying gas by milk of lime. But 12.852 previous to its agitation with that liquid, it should be made to traverse a series of 16.120 refrigeratory pipes submersed under cold water. A vast variety of apparatus, some very ingenious, but many absurd, have been contrived within these few years, for exposing gas to lime in the liquid or dry state. Mr. Accum and Mr. Peckston have been at much pains in describing several of them. The gas holder is now generally preferred of a cylindrical shape, like an immense drum, open at bottom; and flat, or slightly conical at top. The diameter is from 15.112 The following varieties of coal, according to Mr. Accum, contain a less quantity of bitumen, and a larger quantity of car bon than the preceding. They soften, swell, and cake on the fire, and are well calculated for the production of coal gas: 33 to 45 feet in the large establishments, and the height from 18 to 24. The average capacity is from 15000 to 20000 cubic feet. It is suspended in a tank of water by a strong iron chain fixed to the centre of its summit, which passing round a pulley, bears the counter-weight. When totally im mersed in water, the sheet-iron, of which the gas holder is composed, loses hydrostatically about 's of its weight; or if equipoised when immersed, it becomes heavier when in air, minus the buoyancy of the included gas. The mean sp. gr. of well purified coal-gas by Dr. Henry's late experiments may be computed at 0.676, to air called 1.000; or in round numbers, its density may be reckoned two-thirds of that of air. One cubic foot of air weighs 527 gr., one cubic foot of gas weighs 351 gr.; the difference is 176 gr. Hence, 40 cubic feet have a buoyancy of one pound avoirdupois. The hydrostatic compensation is obtained by making the weight of that length of the suspending chain which is between the top of the immersed gasometer and the tangential point of the pulley-wheel, equal to one-fifteenth the weight of the gasometer in pounds, minus its capacity in cubic feet, divided by twice 40, or 80. Thus, if its weight be 4 tons, or 8960 lbs.; and its capacity 15000 cubic feet, a length of chain equal to the height of the gasometer, or to its vertical play, should weigh 597 lbs. without allowing for buoyancy. In this case, the gasometer, when out of water, would have the buoyancy of that liquid, replaced by the passage of these 597 lbs. to the opposite side of the wheel-pulley, so that twice that weight = 1194 lbs. would then be added to the constant counterpoise. When the gasometer again sinks, and loses its weight by the displacement of the liquid, successive links of the chain come over above it, augmenting its weight, and diminishing that of the counterpoise, by twofold operation, as in taking a weight out of one scale, and putting it in the other. a But we must now introduce the correction for the buoyancy of the combustible gas. In ordinary cases, we must regard it as holding a portion of petroleum vapour diffused through it, and cannot fairly estimate its specific gravity at less than 0.750; whence nearly 50 cubic feet have a buoyancy of one pound over the same bulk of atmospheric air. If we divide 15000 by 50, the quotient = 300 is the double of what must be deducted in pounds weight from the hydrostatic compensation. Thus, 597 150 447, is the weight of the above portion of chain. When the gasometer attains its greatest elevation, these 447 lbs. hang on the opposite side of the wheel, constituting an increased counterpoise of twice 447 894, to which, if we add the total buoyancy of the included gas = 300 lbs. we have the sum 1194, equal to the total increase of the weight of the iron vessel on its suspension in air. † The following plan for suspending gasometers was devised by me several years ago, and published in the Analeclic Magazine of this city for May 1817. "Account of an improved mode of suspending gasometers; by Dr. Hare. "It is well known to all who are conversant in gas light apparatus, that no mode has been heretofore devised to render gasometers accurately equiponderant at all points of their immersion in the water; a circumstance so indispensable to their action. The mode adopted in the large London establishments, and which appears to be the most approved, is that of the gasometer chain. This is costly; difficult to execute well, and not susceptible of correction, when erroneously proportioned to the desired effect; especially after the apparatus is in operation. From all these faults, the method of suspension on a beam, like that in the following cut, is entirely free. In practice it has answered perfectly: and, when we have described the mode of constructing such a beam, we think the rationale of its operation will become selfevident. Find (by trial, if possible; if not, by calculation) the weight of the gasometer when sunk so low, as that the top will be as near as possible to the water, without touching it. In the same way find the weight of the gasometer at the highest point of immersion, to which it is to rise, when in use. Then, as the weight in the last case, is to the weight in the first; so let the length of the arm A, be to the length of the arm B. From the centre D, with the radius A, describe a circle; on which set off an arch C, equal to the whole height through which the gasometer is to move. Divide this into as many parts as there are spaces in it, equal each to one-sixth of the radius or length of arm A. Through the points thus found, draw as many diameters, which will, of course, form a corresponding number of radii and divisions, on the opposite side of the circle. Divide the difference between the length of A and B, by the sum of these divisions: and let the quotient be q. From the centre D towards the side E, on radius 2, set off a distance equal to the length of the arm A, less the quotient or 9. On radius 3, set off a distance equal to A, less 2 q, or twice the quotient; and so set off distances on each of the radii; the last being always less than the preceding, by the value of q. A curve line bounding the distances thus found, will be that of the arch head E. The beam being supported on a gudgeon at D, let the gasometer be appended at G; and let a weight be appended at F, adequate to balance it at any one point of immersion. This same weight will balance it at all other points of its immersion-provided the quantity of water displaced by equal sections of the gasometer be equal. But as the weights on which A and B were predicated, may not be quite correct, and as, in the construction of large vessels, equability of thickness and shape cannot be sufficiently attained the consequent irregular buoyancy is compensated by causing the weight to hang nearer to, or farther from the centre, at any of the points taken in making the curve. This object is accomplished by varying the sliders seen opposite to the figures 1, 2, 3, 4, 5, 6. When they are properly adjusted, they are made firm by the screws of which the heads are visible in the diagram. The drawing is of a beam twelve feet in length; and of course the length of the arm A is six feet-that of B, four feettheir difference two feet; which divided by 6, the number of points taken in making the curve E, gives four inches for the quotient q. Hence the distance on radius 2, was five feet eight inches-on radius 3, five feet four inches-on radius 4, five feet-on radius 5, four feet eight inches-on radius 6, four feet four inches-and lastly four feet. The iron gudgeon, where it enters the beam, is square. The projecting parts are turned true, and should be bedded in brass or steel dies; placed, of course, on a competent frame. The sixth part of a revolution of the portions of the gudgeon thus supported, is the only source of friction to which this beam is subject during the whole period of the descent of the gasometer;-which, in large ones, does not ordinarily take place in less than six hours."† The principles of the distribution of gas are exhibited in the following table, given by Mr. Peckston. The gas holder is worked at a pressure of one vertical inch of water, and each argand burner consumes five cubic feet per hour. The following statement is given by Mr. Accum. An argand burner, which measures in the upper rim half an inch in diameter between the holes from which the gas issues, when furnished with five apertures 1-25th part of an inch diameter, consumes two cubic feet of gas in an hour, when the gas flame is one and a half inch high. The illuminating power of this burner is equal to three tallow candles eight in the pound. An argand burner three-fourths of an inch in diameter as above, and perforated with holes 1-30th of an inch diameter (what number? probably 15) consumes three cubic feet of gas in an hour when the flame is 24 inches high, giving the light of four candles eight to the pound. And an argand burner seven-eighths of an inch diameter as above, perforated with 18 holes 1-32d of an inch diameter, consumes, when the flame is three inches high, four cubic feet of gas per hour, producing the light of six tallow candles eight to the pound. Increased length of flame makes imperfect combustion, and diminished intensity of light. And if the holes be made larger than 1-25th of an inch, the gas is incompletely burnt. The height of the glass chimney should never be less than five inches. The argand burner called No. 4, when burnt in shops from sunset till nine o'clock, is charged three pounds a-year. The diameter of its circle of holes is five-eighths of an inch, and of each hole 1-32d of an inch. It is drilled with 12 holes, 5-32ds of an inch from the centre of one to the cen. |