methods of ascertaining whether a particular substance be simple or complex, and, if complex, of determining its constituents. See Vol. V. Before treating of the differences of the several groups, certain preliminary analytical operations which are in constant use will be mentioned. As general chemistry is divided into inorganic and 1. Solution. Some solids, when brought in contact organic, so analytical chemistry is divided with water or other liquid, gradually pass into the liquid p. 397 Am. into inorganic analysis and organic analysis; state, or dissolve; such are said to be "soluble.' The ed. (p. 489 and both of these are again divided into liquid which effects the solution is "the solvent," and qualitative and quantitative analysis. The the liquid obtained is termed the "solution." If the first of these latter branches ascertains the elemental substance can be recovered without material alteration, chemical composition of any substance or mixture, the solution is said to be "simple;" but if a chemical and the second the relative proportion in which change has taken place, the product is a "chemical sothese constituents are present. Of course the suc-lution. 2. Evaporation is the driving off of the solcessful application of the second method of analysis presupposes an acquaintance with the results of the first in any particular case. Quantitative chemical analysis is classified, moreover, under the headings of gravimetric analysis, in which the material to be analyzed is weighed, and its component parts are likewise determined by weight; volumetric analysis, in which the composition of a substance is determined by measuring the volumes of solutions of known strength which are used in the reactions incident to the analysis; and eudiometric, or volumetric, gas analysis, which determines both the qualitative and the quantitative composition of gaseous mixtures. Organic quantitative analysis, lastly, is divided into proximate and ultimate analysis, according as the complex vegetable or animal compound is resolved into simpler compounds only, or at once into the few chemical elements which in organic compounds unite in so many different forms of molecular combination. The methods of ultimate analysis dealing with some three or four elements only are quite simple, while proximate analysis involves the use of many reagents and methods, as the various organic principles contained in plant and animal tissues have very different physical and chemical properties. The space at our disposal will allow of only a cursory view of the most important of these several branches of chemical analysis. The methods of inorganic qualitative analysis will first be passed in review; of the two branches of inorganic quantitative analysis, gravimetric and volumetric, as the methods of the former correspond in principle almost uniformly with those of qualitative analysis, they will not be noted here, and for a description of the latter the reader is referred to manuals on the subject; eudiometric, or gas, analysis will then be noted, especially in its application to industrial gases. The subject of ORGANIC ANALYSIS will be treated under that title. vent by boiling. 3. Precipitation occurs when, on mix- If The First Group.-We have no group reagent for the alkali metals, but special tests enable us to recognize them without difficulty. Any salts of ammonium, when heated in a dry test-tube, will volatilize, condensing again, in most cases, upon the cooler portion of the tube. When, in concentrated solution, in the presence of a little free hydrochloric acid, platinic chloride is added, a yellow crystalline precipitate is formed of the composition (NH),PtCl. This double chloride is inINORGANIC QUALITATIVE ANALYSIS.-In qualita- soluble in alcohol; it resembles that of potassium and tive analysis the student attempts to determine the com- platinum, but can be distinguished by the fact that position of the different forms of matter by applying to upon ignition it leaves only spongy platinum. On them reagents-bodies of known properties which when warming ammonium salts with potassium hydrate or applied properly give him information in the form of lime-water, ammonia gas is liberated, and is at once recreactions. The reactions yield new products, with ac-ognized by its pungent odor. The most delicate test for companying changes in appearance and properties. ammonium salts is their action with Nessler's solution When the reaction takes place between two liquids, a not (a potassium hydrate solution of potassio-mercuric infrequent result is the formation of an insoluble com- iodide). This reagent produces in dilute solutions a pound, a precipitate. Or the result of the reaction may yellowish coloration, and in concentrated solutions a be the escape of a gas with what is called effervescence. reddish-brown precipitate. Potassium salts require an According to the action of reagents we can divide the intense heat for volatilization. Their concentrated solumetals or basic elements into six groups, the members tions, in the presence of a little free acid, yield with of which are in many respects similarly affected by re- platinic chloride a double salt, K,PtCl,, analogous to agents. These six groups are: that of ammonium and resembling it. This salt is somewhat soluble in water, but insoluble in alcohol. ignition it leaves metallic platinum and potassium chloride. A ready test for potassium compounds is to dip a clean platinum wire into the solution, and then to hold it in the flame of a Bunsen burner, when we have a violet-colored flame. Sodium salts obscure the violet of the potassium; but if a piece of blue glass or a prism filled with an indigo solution be held between the eye and the flame, the yellow of the sodium flame is absorbed, and the violet color due to potassium becomes visible. The intense yellow color imparted to flame by sodium compounds is the best means of identifying sodium. When to a solution of sodium chloride etc. I. Potassium, Sodium, III. Ammonium, Cobalt, Lithium, Nickel, etc. II. IV. Barium, Iron, Strontium, Aluminium, Calcium, Chromium, V. VI. etc. On platinic chloride is added and the solution allowed to evaporate, a double chloride, Na,PtCl, results, which separates in aurora-red needles. This color and its solubility in water and alcohol distinguish it from the corresponding potassium and ammonium salts. Potassium pyroantimoniate (K2H„Sb,O1), in concentrated neutral solutions of sodium salts, throws down a crystalline white precipitate of Na2H,Sb,O,. The double chloride of platinum and lithium, like that of sodium, is soluble. In concentrated lithium solutions sodium carbonate produces, after some time, a white granular precipitate of lithium carbonate. With hydro-sodium phosphate, HNA,PO, no precipitate is obtained in the cold; but when the liquid is boiled, there separates a white crystalline compound, lithium phosphate, soluble in hydrochloric acid, but not reprecipitated from a cold solution on the addition of ammonium hydrate. Lithium salts give to the colorless gas or alcoholic flame a rich carmine color. sulphide. The fixed alkalies, as KOH, precipitate from manganese solutions white manganous hydrate, Mn(OH)2, rapidly turning brown, owing to oxidation. With zinc the hydrate is white in color and readily soluble in an excess of the reagent; with nickel the hydrate is apple-green; with cobalt is formed a blue basic hydrate which changes on boiling to a pink. Both nickel and cobalt hydrates are insoluble in excess of the reagent. The fixed alkaline carbonates precipitate the entire group; ammonium hydrate and carbonate only precipitate manganese completely. With the group reagent, ammonium sulphide, manganese yields a fleshcolored sulphide soluble in acetic and the mineral acids; while nickel and cobalt furnish black sulphides insoluble in the dilute mineral acids, but soluble in concentrated nitric acid or aqua regia. Of these four sulphides, nickel is the only one appreciably soluble in the group reagent. Its solubility is shown by the dark color of the supernatant liquid. Concentrating the solution and adding acetic acid will cause it to separate out. With a bead or globule of fused borax three of the four elements in this group give characteristic color-manganese, an amethyst color; nickel, a reddish-brown; and cobalt, a deep blue. Manganese compounds, also, when heated on platinum foil with sodium carbonate and nitrate. yield a bright-green mass of manganate of soda. The different effect of potassium cyanide upon nickel and cobalt solutions gives us a means of separating these elements. Potassium cyanide precipitates from nickel a dirty-green cyanide of nickel, which dissolves in excess of cyanide of potassium. From this solution the nickel may be precipitated as cyanide by careful neutralization with hydrochloric acid or as hydrated sesquioxide on addition of a hypochlorite. Boiling does not alter the double nickel salt. From cobalt solutions potassium cyanide precipitates a chocolate-colored cyanide also soluble in excess. But by boiling the solution the potassio-cobaltous cyanide is oxidized to potassio-cobaltic cyanide, which is no longer precipitable. From cobalt solutions nitrite of potash precipitates, in the presence of acetic acid, a yellow crystalline precipitate of double nitrite of cobalt and potash, while nickel gives no such result. If the four elements of this group be precipitated together as sulphides, this mixture, after washing, is to be treated with warm dilute HCl, which will dissolve MnS and ZnS, but leave CoS and NiS. On adding excess of sodium hydrate to the solution containing manganese and zinc the former only is precipitated, the latter remaining in solution. The mixture of insoluble NiS and CoS is dissolved in concentrated HNO3 evaporated to dryness and the two elements separated by the nitrite-of-potash method given The Second Group is known as the group of the alkaline earths; its reagent is ammonium carbonate. Magnesium may be considered as a separate section of this group, being distinguished from the other members by the insolubility of its hydrate, the ready solubility of its sulphate, and the non-precipitation of its carbonate, in the presence of ammoniacal salts. Again, the alkaline hydrates have no effect upon barium, strontium, or calcium salts, but precipitate magnesium hydrate from the solutions of that metal. Sulphuric acid or soluble sulphates precipitate barium readily, strontium more slowly, and calcium only in concentrated solutions or upon addition of alcohol. The three sulphates are all white in color, and differ chiefly in relative solubility. The best precipitant for calcium is ammonium oxalate. Both bichromate and neutral chromate of potassium precipitate barium chromate, insoluble in acetic acid, but do not precipitate strontium chromate from either acid or neutral solutions. All the carbonates are white and soluble in dilute acids; the sulphates of strontium and calcium are soluble in dilute acids, but that of barium is not. Barium salts impart a yellowish-green color to flame, calcium salts an orangered, and strontium salts a crimson color. Magnesium solutions, in the presence of ammoniacal salts, are precipitated on the addition of hydro-sodium phosphate, yielding a white crystalline precipitate of MgNH,PO, When this is ignited, it loses water and ammonia and leaves Mg,P,O,, pyrophosphate of magnesia. From solutions of magnesium free from ammoniacal salts a concentrated solution of ammonium carbonate will, after standing some time, precipitate a double carbonate, (NH)2CO3. MgCO3. This furnishes a means of separating magnesium from the alkalies. Or if the mag-above. nesium solution be free from ammonium salts, it can be boiled with barium hydrate, and magnesium hydrate will be precipitated. The excess of barium hydrate can then be removed by adding dilute sulphuric acid, and the filtrate from the barium sulphate which separates will contain only the alkalies. The following method is used in separating the members of this group when together in solution: First add some NHCl, and then the group reagent, (NH),CO,. A precipitate forms, which is filtered, and to the filtrate add HNa, PO,, when, if magnesium be present, a precipitate of MgNH,PO, is obtained. Dissolve the original carbonate precipitate in dilute HCl, evaporate the solution to dryness, and treat the dry residue with alcohol. If insoluble in alcohol, dissolve in water, adding H2SO,, when, if barium be present, insoluble BaSO, will be found; but if soluble, boil off the alcohol, add ammonium carbonate, filter, and dissolve the precipitate in HNO3. Evaporate to dryness, and treat the dry residue with alcohol. If soluble, expel the alcohol, add ammonium oxalate, and calcium oxalate will be precipitated; but if insoluble, test in the Bunsen flame, when the intense red color will prove the presence of strontium. 4 The Third Group includes manganese, zinc, nickel, and cobalt, and has as a group reagent ammonium The Fourth Group includes aluminium, chromium, and iron, and has for its reagent ammonium hydrate, NH.OH. This reagent precipitates them as hydrates, that of aluminium being white and gelatinous, only slightly soluble in excess of precipitate; that of chromium, bluish-green, soluble in excess of ammonium hydrate. With iron in the ferric condition is gotten a reddish-brown precipitate insoluble in excess of the ammonium hydrate. Ferrous solutions, if pure, yield a white precipitate of ferrous hydrate, but this very speedily oxidizes and becomes dirty-green or brown in color. The fixed alkalies precipitate the entire group, the hydrates of aluminium and chromium dissolving in excess of the reagent, Boiling will, however, cause the reprecipitation of the chromium hydrate. The alkaline carbonates behave like their hydrates except in the case of ferrous salts, where white ferrous carbonate is precipitated. The group reagent of the preceding group, (NH)2S, precipitates aluminium and chromium as hydrates, and iron as black ferrous sulphide, FeS. The blow-pipe tests are among the most characteristic with the metals of this group. Aluminium compounds heated upon charcoal, and then moistened with cobaltous nitrate and reheated, show a blue color. Chromium compounds impart a beautiful emerald-green color to the borax glass, and, fused with sodium carbonate and nitrate on platinum foil, yield a yellow mass of sodium chromate. Iron imparts to the borax glass a color yellowish while hot and light-green when cold. The ferric salts are distinguished from ferrous by the following reactions: With ferric salts, the addition of H&S or (NH4)2S reduces them to ferrous salts with separation of sulphur, the latter reagent then precipitating FeS. Ferrocyanide (yellow prussiate) of potash with ferric salts yields a deep-blue precipitate of ferric ferrocyanide (Prussian blue), with ferrous salts a greenish-white precipitate of ferrous ferrocyanide. With ferricyanide (red prussiate) of potash ferrous salts yield Turnbull's blue, very similar to Prussian blue in appearance, while ferric salts yield only a green-brown coloration, but no precipitate. Potassium sulphocyanate gives a deep blood-red color with ferric salts even in dilute solution, and no color at all with ferrous salts. Salicylic acid is also a very sensitive reagent for ferric salts, yielding a fine purple color. Ferrous compounds are converted into ferric by heating the solutions with a little nitric acid. The reduction of ferric to ferrous compounds is best effected by the aid of hydrogen sulphide or sulphurous acid. The original solution is to be tested as to whether the iron existed in the ferrous or ferric condition, as after precipitation by the group reagent it is always in the ferric state. The members of this group, if precipitated together, may be identified as follows: The well-washed precipitate is placed in a porcelain dish with sodium hydrate and some bromine water, and then warmed. The chromium and aluminium are dissolved, while the ferric hydrate remains insoluble. This latter is filtered off, dissolved in HCl, and tested with potassium ferrocyanide. The filtrate containing aluminium and chromium is tested with excess of NH4Cl and warmed until ammonia no longer escapes. Aluminium, if present, would have been thrown down. The filtrate from the latter or the liquid free from aluminium is acidified with hydrochloric acid, hydrogen sulphide gas conducted through the solution to reduce the chromic oxide, which is afterward precipitated by ammonia hydrate. It should be said that ammonium hydrate, in addition to the hydrates of this group, precipitates the phosphates of this group and the phosphates, oxalates, borates, and fluorides of the alkaline earths. The Fifth Group, consisting of tin, arsenic, antimony, gold, and platinum, yield insoluble sulphides when treated with hydrogen sulphide in acid solution. It is true that the succeeding, or sixth, group show the same behavior, but from the latter the fifth group is distinguished by the solubility of its sulphides in ammonium or potassium sulphides or the fixed alkaline hydrates. Tin, the first element of the group, may exist in either the stannous or the stannic state. Stannous salts form purple of Cassius with chloride of gold, reduce mercuric salts to mercurous, and ultimately separate fine gray metallic mercury, form a brownish-black sulphide, not very soluble in alkaline hydrates, but easily in alkaline sulphides, with which it forms double sulphides. From this solution, however, dilute acids precipitate yellow stannic sulphide. Any tin compound heated on charcoal together with sodium carbonate and some potassium cyanide is reduced to the metallic condition. Antimony is characterized by the following tests: On adding water to the acid solution of its chloride, a white precipitate of oxychloride forms, soluble in tartaric acid. The group reagent precipitates orangeyellow antimony trisulphide. Antimony compounds, placed in an apparatus in which hydrogen is being generated, yield stibine, H,Sb. If this gas be conducted through glass tube, narrowed at intervals, and heat be applied to the tube near the narrowed portions, mirror-like deposits of metallic antimony will be formed; the escaping gas, when ignited, burns, and on holding a cold object in the flame a black spot of metallic antimony deposits. This metallic deposit is not soluble in sodium hypochlorite (chlorinated soda), and can in this way be distinguished from the deposit of metallic arsenic. Compounds of antimony heated on charcoal with sodium carbonate and potassium cyanide in the reducingflame yield a white and Brittle globule of metal, while a white coating of antimonious oxide forms on the charcoal. Arsenic, similarly, forms a hydride, H.As, which serves for the detection of minute traces of this element (Marsh's test). The metallic mirror or spot gotten in this case, however, is more lustrous than that of antimony, is readily soluble in sodium hypochlorite, and if evaporated with ammonium sulphide leaves a yellowish residue. Arsenic compounds heated on charcoal give off whitish fumes of arsenious oxide, accompanied by a garlic-like odor. The group reagent precipitates from acid solution of arsenious compounds yellow trisulphide; from arsenic acid solutions sulphur separates, and finally the solution is reduced to the arsenious state, when the yellow trisulphide separates. Silver nitrate precipitates yellow silver arsenite or chocolate-brown silver arseniate, according to the nature of the solution. A dry alkaline acetate heated with arsenious oxide (white arsenic) gives rise to a very disagreeable-smelling compound, alkarsine. Platinum is recognized by the yellow precipitate gotten in hydrochloric acid solution with potassium or ammonium chlorides. The precipitate in the latter case, when ignited, leaves spongy platinum as the only residue. Gold salts in solution are precipitated on the addition of ferrous sulphate. Stannous chloride containing some stannic salt produces a purple precipitate (purple of Cassius) even in dilute gold solutions. The group precipitate, consisting usually of stannic sulphide, antimony sulphide, and arsenic sulphide, may be separated as follows: Treat in a dish with concentrated HCl containing some bromine. Everything except some sulphur will be dissolved. Filter, if necessary, and to cold filtrate add an excess of concentrated sodium hydrate and one-fifth its volume of alcohol, A white flocculent precipitate will indicate antimony. Filter it off, dissolve in HCI, apply some confirmatory test. The filtrate from the antimoniate is boiled, to expel alcohol. Tartaric acid is then added to acid reaction, magnesium chloride, and finally ammonium hydrate to alkaline reaction. Arsenic, if present, will be precipitated as ammonium magnesium arseniate. Filter; acidify with HCl, and add hydrogen sulphide, which will precipitate any tin as yellow sulphide. The Sixth Group includes mercury, silver, lead, copper, cadmium, and bismuth. A subdivision of this group into two sections is found convenient, as silver, lead, and mercurous chlorides are relatively insoluble in water, and hence are precipitated on acidifying the solution with hydrochloric acid, while mercuric, copper, cadmium, and bismuth salts remain in solution. The sulphides of this group are precipitated by HS in acid solution, but differ from the sulphides of the fifth group in being insoluble in alkaline sulphides. Silver with NaOH precipitates brown oxide; with ammonium hydrate, the same precipitate, soluble in excess of the reagent. Hydrogen sulphide forms black silver sulphide insoluble in dilute acids; hydrochloric acid or chlorides form a white curdy precipitate (AgCl) soluble in ammonium hydrate. Silver compounds heated with sodium carbonate on charcoal before the blowpipe are reduced to metal yielding a white ductile globule. Lead salts with NaOH yield a white hydrate soluble in excess of reagent. Sodium carbonate precipitates a basic carbonate. Hydrochloric acid and chlorides throw down a white curdy precipitate of lead chloride soluble in large quantity of water when heated. HS precipitates black-lead sulphide soluble in warm nitric acid. Sulphuric acid produces a white precipitate of lead sulphate readily soluble in a solution of ammonium acetate. Lead compounds heated on charcoal with sodium carbonate yield soft malleable globules of metal, the charcoal at the same time receiving a slight yellow incrustation of lead oxide. The salts of mercury must be distinguished as mer curous or mercuric. Mercurous salts with sodium hydrate yield black mercurous oxide, while with ammonium hydrate a black mercur-ammonium compound is produced. With HS a black sulphide is at once produced, insoluble in nitric acid; with potassium iodide a green iodide is formed; with hydrochloric acid a white curdy precipitate of mercurous chloride (calomel) is formed. Mercuric salts with sodium hydrate yield a reddish-brown precipitate, which becomes yellow if the reagent is added in excess; with ammonium hydrate a white mercur-ammonium compound (white precipitate) is formed. HS produces a precipitate turning from white to yellowish, brownish, and with excess of the reagent to black, color. Stannous chloride reduces the mercuric to mercurous salt and then precipitates calomel, an excess of the reagent reducing the latter to metallic mercury. Potassium iodide added to mercuric solutions produces a precipitate of scarlet mercuric iodide. Mercury compounds in general, when heated with sodium carbonate in a narrow glass tube, give a sublimate of metallic mercury. A piece of clean copper immersed in a mercury-salt solution becomes coated with gray mercury, which, rubbed, yields a bright amalgam surface. Bismuth solutions yield with alkaline hydrates a white hydrate insoluble in excess, and with sodium carbonate a basic carbonate. A characteristic reaction for bismuth is the decomposition of its salts by water; this is most sensitive with the chloride, which, poured into water, yields a white precipitate of oxychloride of bismuth. This oxychloride, unlike that of antimony, is not soluble in tartaric acid. Hydrogen sulphide precipitates black sulphide of bismuth. Before the blowpipe bismuth compounds mixed with sodium carbonate yield brittle reddish-white globules of metal. The incrustation of oxide is orange while hot and yellow when cold. Copper salts yield a bluish hydrate with potassium hydrate, turning black on boiling; ammonium hydrate gives a greenish-blue precipitate soluble in excess of the reagent with azure-blue color. Hydrogen sulphide gives a brownish black precipitate of sulphide soluble in warm nitric acid. A piece of bright metallic iron immersed in copper solutions is rapidly coated with red copper. Copper compounds heated on charcoal with sodium carbonate yield a button of metallic copper. Borax glass is colored by copper compounds, when heated in the oxidizing flame, green while hot, blue when cold. Cadmium salts are precipitated white by sodium or ammonium hydrates; the precipitate in the first case is not soluble, but in the second case is soluble in excess. Hydrogen sulphide precipitates a yellow sulphide soluble in warm nitric acid and in boiling HCl and H,SO,, thus distinguishing it from copper. When cadmium compounds are heated with sodium carbonate on charcoal, the latter becomes covered with reddish-brown cadmium oxide. Acids do not admit of the same exact classification as bases, as there are no general group-reagents which serve to distinguish them with much accuracy. For individual tests the reader must consult the manuals of analysis. GAS ANALYSIS.-In the analysis of gaseous mixtures volumes only can be determined, as the weights of the gases concerned are relatively small, and the experimental difficulties connected with the weighing would be so great as almost certainly to vitiate the results. In measuring the volumes, on the other hand, after allowing for the change in volume due to the varying conditions of temperature and pressure, the measurements can be made with reasonable accuracy. No uniform course of procedure can be laid down for all cases of gas analysis, as the mixtures that arise in practice are of very different character, and the procedure must be based largely upon the physical and chemical properties of the gases concerned. Thus we have to deal with quite a complex mixture in illuminating gas, where hy, drogen, carbonous oxide, numerous hydrocarbons, and free nitrogen are all to be determined; with a much simpler mixture of some of these constituents in the VOL. II.--D FIG. 1.-Bunsen's Apparatus. either by absorption, with the aid of some reagent introduced into the tube containing the gaseous mixture, or by what is termed "combustion," which may be a slow chemical combination of the combustible gas with oxygen or air, or an instantaneous one brought about by the aid of an electric spark caused to flash through the tube containing the gas. We shall mention briefly the methods most generally practised, and show how they are applied to the case of the most commonlyoccurring mixtures. The first carefully-elaborated method, and one still in many respects the most perfect, is that of Bunsen, in which the gases are measured in closed tubes over mercury, the absorption reagents are as far as possible used in the solid form, and the combination of the combustible gas with oxygen effected by the passage of the electric spark between platinum wires fused in the upper end of the tubes. Observations of the thermometer and barometer are made with each reading of gaseous volume, in order to apply to the reading of volume the corrections for varying temperature and pressure. These reading are all made through a telescope, called a cathetometer," placed on the opposite side of the room, so that the heat of the body may not affect the confined gas or gases, as it would if the observer came sufficiently close to make the reading with the unaided eye. Fig. 1 shows the arrangement of the apparatus 66 spectively, and allowing for the tension of aqueous vapor in the gas being noted. To absorb the carbon dioxide, the first constituent to be removed, a small pellet of fused potassium hydrate cast on the end of a platinum wire is introduced and allowed to remain in the gas for several hours. After its withdrawal, and after the tube has taken the temperature of the room, a reading is made of the residual gas, reckoned as dry, inasmuch as the potassium hydrate has absorbed moisture as well as carbon dioxide. A small coke-ball saturated with fuming sulphuric acid is next introduced, to absorb the olefines or illuminating constituents of the gas. After its withdrawal and the reading of the residual gas, a similar ball soaked with alkaline solution of pyrogallic acid is introduced, to absorb the free oxygen. This is then followed in proper time by the introduction of a coke-ball saturated with an acid solution of cuprous chloride, which absorbs the carbon monoxide. With this absorption the withdrawal of constitu in the case of an analys's by Bunsen's method; m is here the eudiometer-tube in which the volume of gas is being measured, the tube being supported firmly in an erect position in a mercury-trough; d is a syphonbarometer, and is a thermometer; ss represents the wires from an induction-coil, B, which are made to terminate in two platinum wires fused in the top of the eudiometer-tube from opposite sides. These platinum wires do not join, but are separated by an interval of a fraction of an inch, through which interval the spark must flash. All the readings are made, as before stated, through the cathetometer, g, placed at some distance. The chief objection to Bunsen's method is the great length of time occupied in the absorptions and combustions, as after each handling of the tube some time must be given for the apparatus to take the temperature of the room before a reading can be made that will be accurate. To obviate this difficulty, a number of methods have been proposed, and several introduced into practice, in which the absorption is effected by liquid reagents, which, of course, act much more rapidly as the surface of contact with the gases is much greater; and the residual gas is then measured either at once in the absorption-tube or after transferral to another tube. The first of these liquid-absorption methods was proposed by Doyère, and subsequently improved by Regnault and Reiset, by Frankland and Ward, and by Russell. All these methods involve the use of apparatus made with accuracy, and can be used for exact scientific investigation. Much simpler in construction are what are called gas-burettes," which serve for the analysis of industrial gases. Apparatus of this kind has been devised by Stammer, Winckler, Raoult, and Bunte. These gas-burettes allow of the absorption of one gaseous constituent of a mixture after another by using successively the appropriate absorption reagent solutions and after each absorption measuring residual gases. The most convenient form of absorption apparatus for the analysis of furnace-gases and all similar mixtures is that of Orsat, which, after several modifications, has assumed the shape shown in fig. 2. In the apparatus figured here the measuring-tube M, surrounded by a cold-water mantle, communicates at its lower end with a flask of acidulated water A, by the raising or lowering of which the gas is drawn into or expelled from the tube. At the other end it is connected by a capillary tube with a very narrow tube of block tin, which in turn connects by means of the metallic stopcocks a, b, and c with the three absorption-vessels N, P, and K. It also connects by means of a three-way cock d with the rubber tubes S and J. Through the former the gas to be analyzed is drawn into the measuring-tube M by lowering the bottle A, the stopcocks a, b, and c being closed and the tube M full of water; through the latter, by means of the rubber compression-bulb, the connecting tubes are emptied of air, and the absorption-liquid is drawn up to the same level, m', m', m'", in all three of the absorption-vessels before the gas to be analyzed is drawn in. The absorption-vessels consist of wide tubes filled with fragments of narrower glass tubes, so as to furnish a large surface for absorption, and connecting by a narrow-drawn-out portion with the liquid in the tubulated bottles below. The tubulures of these absorption-vessels and of the flask A are of course open during the analysis and closed with corks when not in use. This form of apparatus permits of the absorption of three gaseous constituents one after the other, and the measuring of the residual gas after such absorption. с The method of analysis of a gaseous mixture may now be indicated. With the more complete apparatus of Bunsen (see fig. 1) the analysis of quite complex mixtures may be made. If we take the case of illuminating gas, which contains a number of constituent gases to be determined, we proceed first to the absorptions. A portion of the gas is first measured in the eudiometer over mercury corrections reducing the observed temperature and pressure to 0° C. and 760 mm. re ents is complete so far as reagents can effect it, and the residual gas, which contains hydrogen, marsh-gas, and nitrogen, is ready for the combustion, which is effected as shown in the cut. Before passing the spark there must be added a quantity of pure oxygen, the volume of which is measured, and of air, which is likewise meas-. ured. The oxygen is to unite with the hydrogen to form water, and with the carbon of the marsh-gas to form carbon dioxide; the air is to dilute the explosive mixture and lessen the violence of the combination. After the spark has passed a decided contraction of volume ensues, which, when the temperature has become constant, is measured. This contraction gives the volume of aqueous vapor formed, allowance being made for the tension of aqueous vapor. A fused ball of potassium hydrate is then introduced, to absorb the carbon dioxide, the other product of the combustion. After its absorption and the reading of the residual gas, the only constituent of the original gas remaining is the nitrogen. To determine it, a measured amount of pure hydrogen, sufficient to combine with the oxygen which remains after the previous combustion, is introduced, and then the spark passed again. this combustion the contraction is read, and then the residue can be only nitrogen and the excess of hydrogen just before introduced. This excess is, of course, known after the reading of the contraction last ensuing; so After |