GRAVIMETRIC DETERMINATION.

The solution containing the molybdate is neutralised and treated with an excess of mercurous nitrate. The precipitate is allowed to settle for some time, filtered, and washed with a dilute solution of mercurous nitrate. Then it is dried, transferred to a weighed Berlin crucible containing ignited oxide of lead, mixed, ignited, and weighed. The increase in weight gives the amount of trioxide, MoO_{3}. This contains 66.7 per cent. of molybdenum.

URANIUM.

Uranium occurs chiefly as pitchblende, which is an impure oxide (U{3}O{8}). It is also found as sulphate in uranochre, johannite, &c.; and as phosphate in the uranites, torbernite (hydrated phosphate of uranium and copper), and autunite (hydrated phosphate of uranium and lime). It also occurs in some rarer minerals.

The oxide is used for colouring glass; and the nitrate and acetate are used as reagents. "Uranium yellow," used for enamel painting, is sodium uranate. The uranates, in which the oxide of uranium acts as an acid, are mostly insoluble and of secondary importance.

Uranium forms two families of salts, uranous and uranic; corresponding to the oxides UO{2} and UO{3} respectively. The former are generally green and the latter yellow. Uranous salts are converted into uranic by boiling with nitric acid or other oxidising agents. Uranic salts, on the other hand, are easily reduced by sulphuretted hydrogen, stannous chloride or zinc. This property is made use of in determining the quantity of uranium in pure solutions by titrating with permanganate of potassium solution as in the case with iron.

Detection.—The most characteristic reaction of the uranium compounds is their behaviour in the presence of alkaline carbonates in which they are freely soluble; even ammonium sulphide will not precipitate uranium from these solutions. On neutralising the carbonate with an acid a uranate of the alkali is precipitated. Ammonia or sodic hydrate (free from carbonates) give yellow precipitates, which are insoluble in excess of the reagent, but are soluble in acids. Ferrocyanide of potassium gives a reddish-brown precipitate. Uranium colours the borax-bead yellowish-green in the oxidising, and green in the reducing, flame.

Solution and Separation.—The compounds of uranium are soluble in acids. Powder the substance and evaporate with an excess of nitric acid. Take up with hydrochloric acid, dilute, pass sulphuretted hydrogen, and filter. Peroxidise the filtrate with a little nitric acid, add an excess of ammonic carbonate and some ammonium sulphide, and filter. Render the solution acid, boil; and precipitate the uranium by means of ammonia. Filter off, and wash it with dilute ammonic chloride. Ignite, and weigh as protosesqui-oxide, U{3}O{8}.

GRAVIMETRIC DETERMINATION.

The solution containing the uranium free from other metals is, if required, first peroxidised by boiling with nitric acid. Ammonia in slight excess is added to the nearly-boiling solution. A yellow precipitate is formed, which is filtered off hot and washed with a dilute solution of ammonium chloride. The precipitate is dried and ignited; and weighed as U{3}O{8}, which contains 84.8 per cent. of uranium.

VOLUMETRIC METHOD.

This is based on the precipitation of uranium as phosphate from acetic acid solutions and the recognition of complete precipitation by testing with potassic ferrocyanide; it is the converse of the process for the volumetric determination of phosphate.

_The standard solution of phosphate_ is prepared by dissolving 29.835 grams of hydric sodic phosphate (Na_{2}HPO_{4}.12H_{2}O) in water and diluting to 1 litre. 100 c.c. will be equivalent to 2 grams of uranium.

Take 1 gram of the sample (or, if poor in uranium, 2 grams) and separate the uranium as described. Dissolve the precipitate in nitric acid and evaporate to a small bulk, add 2 grams of sodium acetate, dilute with water to 100 c.c., and boil. Titrate the boiling solution with the sodium phosphate till it ceases to give a brown colouration with potassium ferrocyanide. Calculate the percentage in the usual way.

FOOTNOTES:

[78] MnO{2} + 4HCl = MnCl{2} + Cl{2} + 2H{2}O.

[79] Provided a sufficiency of ammonic chloride is present.

[80] With some silicates, &c., a preliminary fusion with sodium carbonate will be necessary.

[81] Instead of sodium acetate, ammonium succinate can be used.

[82] Journ. Soc. Chem. Industry, vol. x. p. 333.

[83] MnO_{2} + 2FeSO_{4} + 2H_{2}SO_{4} = Fe_{2}(SO_{4})_{3} + MnSO_{4} + 2H_{2}O.

[84] If the ore is very rich, a smaller quantity (0.75 or 1.5 gram) must be taken; otherwise the iron will be insufficient.

[85] MnO{2} + 4HCl = MnCl{2} + 2H{2}O + Cl{2}. Cl{2} + 2KI = 2KCl + I{2}.

[86] Iodine probably lost by volatilisation.

[87] Obtained as a brown powder by digesting red lead with nitric acid and filtering.

[88] The water for dilution and the dilute sulphuric acid used for washing should be previously tested, to see they have no reducing action, with dilute permanganate of potassium solution.

[89] Arnold and Hardy, Chemical News, vol. lvii. p. 153.



CHAPTER XIV.

EARTHS, ALKALINE EARTHS, ALKALIES.

ALUMINA.

Alumina, the oxide of aluminium (Al{2}O{3}), is found in nature fairly pure in the mineral corundum; transparent and coloured varieties of which form the gems sapphire and ruby. A coarser compact variety contaminated with oxide of iron constitutes emery. Compounded with silica, alumina forms the base of clays and many rock-forming minerals. China clay (or kaolin) is used as a source of alumina. Bauxite, hydrated alumina, is also used for the same purpose—that is, for the preparation of sulphate of alumina. The mineral cryolite is a fluoride of aluminium and sodium.

Corundum is characterised by a high specific gravity (4.0) and extreme hardness. By these it is distinguished from felspar and similar minerals, which it somewhat resembles in general appearance.

Aluminium is used for a variety of small purposes: it is white, light, and very tenacious; but owing to the difficulty of its reduction it is expensive.

Aluminium forms one series of salts which closely resemble those of ferric iron. It forms an interesting series of double sulphates, known as the alums. Common potash alum is Al{2}(SO{4}){3},K{2}SO{4},24H{2}O.

Detection.—Alumina is not precipitated from its acid solution by sulphuretted hydrogen, but it is thrown down by ammonia (with the other earths) as a white hydrate, soluble in soda and insoluble in ammonic carbonate. Filtered off and ignited, it assumes, after treatment with nitrate of cobalt before the blowpipe, a blue colour which is characteristic. With natural compounds containing metallic oxides this colour is masked. It is more satisfactory to make a separation in the wet way and to test the ignited oxide.

Separation and Solution.—If the substance is insoluble in hydrochloric acid it is finely powdered and fused with "fusion mixture" with the help, in the case of corundum (which is very refractory) of a little caustic soda or potash. The method of working is the same as that described for opening up silicates. See under Silica. Corundum cannot be powdered in Wedgwood, or even agate, mortars; since it rapidly wears these away and becomes contaminated with their powder. It is best to use a hard steel mortar and to extract the metallic particles from the bruised sample with a magnet or dilute acid.

When the substance has been completely attacked and dissolved, it is evaporated to dryness with an excess of hydrochloric acid on the water-bath to render any silica present insoluble. The residue is extracted with hydrochloric acid and freed from the second group of metals by means of sulphuretted hydrogen. The filtrate from this (after removing the sulphuretted hydrogen by boiling) is nearly neutralised, and treated with 8 or 10 grams of hyposulphite of soda[90] in solution. It is then boiled till the sulphurous oxide is driven off. The precipitate is filtered off, ignited, and weighed as alumina.

It is sometimes more convenient to proceed as follows:—After boiling off the sulphuretted hydrogen peroxidise the iron with a little nitric acid, add a solution of ammonic chloride, and then ammonia in very slight excess; boil, filter, wash, ignite, and weigh the oxides. These generally consist of ferric oxide and alumina. It is a common practice to determine the iron, calculate it to ferric oxide, and so to estimate the alumina indirectly. This may be done either by igniting in a current of hydrogen and estimating the iron by the weight of oxygen lost; or, by dissolving with sulphuric and hydrochloric acids, and determining the iron volumetrically. It should be borne in mind that these oxides will also contain any phosphoric oxide that happened to be in the mineral.

In general analyses of samples containing alumina, it may be contained in both the soluble and insoluble portions. In these cases it is better to fuse the sample with "fusion mixture" before treatment with acids. The alumina in the fused mass will exist in a state soluble in acids.

GRAVIMETRIC DETERMINATION.

Solutions containing alumina free from the other metals are diluted to a convenient bulk and heated nearly to boiling. Add chloride of ammonium, and then ammonia in slight excess; boil, allow to settle, filter, and wash with hot water. Dry the precipitate, and ignite in a platinum or porcelain crucible at the strongest heat. Cool, and weigh. The substance is alumina, Al{2}O{3}, which contains 52.94 per cent. of aluminium. It is only in special cases, such as the analysis of metals and alloys, that it is reported as aluminium. The percentage of alumina is generally given.

Ignited alumina is difficultly soluble in acids; it is not reduced by hydrogen at a red heat. Ignited with ammonium chloride portions are volatilised.

Direct Determination of Alumina in the Presence of Iron.—The iron and alumina are precipitated as hydrates by ammonia. The precipitate is dissolved in hydrochloric acid and the iron reduced to the ferrous state. It is then added to a hot solution of potash or soda. The solution is boiled till the precipitate settles readily, filtered, and washed with hot water. The alumina is contained in the filtrate, which is acidified with hydrochloric acid and the alumina precipitated therefrom as hydrate with ammonia, as just described.

Determination of Alumina in the Presence of Phosphates and Iron.—For details, see a paper by R.T. Thomson in the "Journal of the Society of Chemical Industry," v. p. 152. The principles of the method are as follows:—If the substance does not already contain sufficient phosphoric oxide to saturate the alumina, some phosphate is added. The iron is reduced to the ferrous state and phosphate of alumina precipitated in an acetic acid solution. It is purified by reprecipitation, ignited, and weighed as phosphate (Al{2}O{3},P{2}O{5}), which contains 41.8 per cent. of alumina, Al{2}O{3}.

EXAMINATION OF CLAYS.

Moisture.—Take 5 grams of the carefully-prepared sample and dry in the water-oven till the weight is constant.

Loss on Ignition.—Weigh up 2 grams of the sample used for the moisture determination, and ignite in a platinum-crucible to redness, cool, and weigh.

Silica and Insoluble Silicates.—Weigh up another 2 grams of the dried sample, and place them in a platinum dish; moisten with water, and cover with 20 c.c. of sulphuric acid. Evaporate and heat gently to drive off the greater portion of the free acid. Allow to cool; and repeat the operation. Extract by boiling with dilute hydrochloric acid, filter, wash, dry, ignite, and weigh. The quantity of insoluble silicates is determined by dissolving out the separated silica with a strong boiling solution of sodium carbonate. The residue (washed, dried, and ignited) is weighed, and reported as "sand."

Alumina and Ferrous Oxide.—To the filtrate from the silica add "soda" solution till nearly neutral, and then sodium acetate. Boil and filter off the precipitate. Reserve the filtrate. Dissolve the precipitate in hydrochloric acid, and dilute to exactly 200 c.c. Divide into two parts of 100 c.c. each. In one determine the iron by reducing and titrating in the way described under volumetric iron. Calculate the percentage as ferrous oxide, unless there are reasons to the contrary, also calculate its weight as ferric oxide. To the other portion add ammonia in slight excess, and boil. Filter, wash with hot water, dry, ignite, and weigh as mixed alumina and ferric oxide. The weight of the ferric oxide has already been determined in the first portion: deduct it, and the difference is the weight of alumina.

Lime.—To the reserved filtrate, concentrated by evaporation, add ammonium oxalate and ammonia; boil, filter, ignite strongly, and weigh as lime.

Magnesia is separated from the filtrate by adding sodium phosphate. It is weighed as magnesium pyrophosphate.

Potash and Soda.—These are determined in a fresh portion of the sample by Lawrence Smith's method, as described on page 333.

THORIA.

This is an oxide of thorium, ThO{2}. It is only found in a few rare minerals. It is a heavy oxide, having, when strongly ignited, a specific gravity of 9.2. In the ordinary course of analysis it will be separated and weighed as alumina. It is separated from this and other earths by the following method. The solution in hydrochloric acid is nearly neutralised and then boiled with sodium hyposulphite. The thoria will be in the precipitate. It is dissolved, and the solution heated with ammonium oxalate in excess. The precipitate is thorium oxalate, which is washed with hot water, dried, and ignited. It is then weighed as thoria, ThO{2}. Thoria which has been ignited is not readily soluble in acids.

ZIRCONIA.

The oxide of zirconium, ZrO{2}, is found in the mineral zircon, a silicate of zirconia, ZrSiO{4}. When heated intensely it becomes very luminous, and is used on this account for incandescent lights.

In the ordinary course it is thrown down by ammonia with the other earths, from which it is thus separated:—The hydrates precipitated in the cold, and washed with cold water, are dissolved in hydrochloric acid, nearly neutralised with soda, and precipitated by boiling with hyposulphite of soda. Dissolve; and from the hydrochloric acid solution precipitate the thoria (if any) with ammonium oxalate. To the filtrate add carbonate of ammonia, which will precipitate any titanium present. The zirconia will be in solution, and is recovered by precipitating with potassium sulphate, or by evaporating the solution and igniting. It is separated from alumina by taking advantage of its insolubility in potassic hydrate.

It is estimated in zircons in the following way:—The powdered substance is fused with bisulphate of potash, and extracted with dilute sulphuric acid. The residue is fused with caustic soda and extracted with water. The portion not dissolved, consisting of zirconate of soda, is dissolved in hydrochloric acid. The solution is diluted, filtered if necessary, and treated with ammonia in excess. The precipitate is filtered off, washed with hot water, dried, ignited, and weighed as zirconia, ZrO_{2}. This is a white powder, which is insoluble in acids; even in hydrofluoric acid it is only slightly attacked.

CERIUM.

Cerium occurs as silicate (together with the oxides of lanthanum, didymium, iron and calcium) in the mineral cerite, which is its chief source. It also occurs as phosphate in monazite, and as fluoride in fluocerite. The oxalate is used in medicine. Cerium forms two classes of salts corresponding to the oxides, cerous oxide (Ce_{2}O_{3}) and ceric oxide (CeO_{2}). Compounds of cerium with volatile acids yield dioxide on ignition; and this, on solution in hydrochloric acid, yields cerous chloride and chlorine.

In the ordinary course cerium is thrown down along with alumina and the other earths by ammonia. It is separated by dissolving the hydrates in hydrochloric acid, and oxidizing with chlorine water. On treating with oxalic acid, cerium, lanthanum, and didymium are precipitated as oxalates, which on ignition are converted into oxides. These are soluble in acids. Their solution in hydrochloric acid is nearly neutralised; acetate of soda is then added, and an excess of sodium hypochlorite. On boiling, the cerium is precipitated as dioxide, which is filtered off, ignited, and weighed.

Cerium is detected by giving with borax a bead which is yellow in the oxidising, and colourless in the reducing flame. Traces of cerium compounds boiled with dioxide of lead and nitric acid will give a yellow solution.

LANTHANUM AND DIDYMIUM

occur together with cerium in cerite, and are separated with that metal as oxalates, as described under Cerium.

Didymium salts have a rose or violet colour, and impart (when in sufficient quantity) the same colour to the borax bead. Solutions have a characteristic absorption-spectrum.

The separation of lanthanum and didymium in the solution from which the cerium has been precipitated is effected by precipitating them together as oxalates, igniting, and dissolving in dilute nitric acid. This solution is then evaporated to dryness and ignited, for a few minutes, just below redness. A subnitrate of didymium is formed, and remains as an insoluble residue on extracting with hot water. The separated salts are treated with ammonia and ignited, and weighed as oxides (La{2}O{3} and Di{2}O{3}).

YTTRIA.

Yttria is found in gadolinite and some other rare minerals. It is precipitated along with the other earths by ammonia. It is distinguished by the insolubility of its hydrate in potash, by the insolubility of its oxalate in oxalic acid, and by not being precipitated by hyposulphite of soda or potassium sulphate. Further, it is precipitated by potash in the presence of tartaric acid as an insoluble tartrate. This reaction distinguishes the members of the yttria group from most of the other earths. The other members of the group closely resemble it, and amongst them are erbia, terbia, ytterbia, scandia, &c.

BERYLLIA.

The oxide of beryllium, BeO (also known as glucina), occurs in nature mainly as silicate. Beryl, the green transparent variety of which is the emerald, is the best known of these. It is a silicate of alumina and beryllia.[91] Some other minerals in which it occurs are phenakite, euclase, and chrysoberyl.

In the ordinary course of analysis, beryllia will be precipitated with alumina, &c., by ammonic hydrate. It is distinguished by the solubility of its hydrate in ammonic carbonate, by not being precipitated by boiling with sodium hyposulphite, and by not being precipitated by ammonic sulphide from an ammonic carbonate solution.

The analysis of silicates containing beryllia is thus effected. The finely powdered substance is fused with twice its weight of potassium carbonate; and the "melt" is extracted with water, and evaporated with a slight excess of sulphuric acid to render the silica insoluble. Treat with water, filter, and evaporate the filtrate until a crust is formed. Potash alum crystallises out. The liquor is poured off into a warm strong solution of ammonium carbonate. Ferric hydrate and alumina will be precipitated. They are filtered off, re-dissolved, and again precipitated in ammonic carbonate solution; the combined filtrates are boiled for some time, and acidified slightly with hydrochloric acid. The carbon dioxide is boiled off, and the beryllia is then precipitated as hydrate with ammonia. The hydrate is washed with hot water, dried, ignited, and weighed as beryllia, BeO.

Beryllia has a specific gravity of 3.08. It is white, infusible, and insoluble in water. After ignition, it is insoluble in acids, except sulphuric, but is rendered soluble by fusion with alkalies.

Beryllia, in a solution of carbonate of ammonia, is precipitated as carbonate on boiling in proportion as the carbonate of ammonia is volatilised. The hydrate is dissolved by a boiling solution of ammonic chloride, ammonia being evolved.

THE ALKALINE EARTHS.

LIME.

Lime is an oxide of calcium, CaO. It occurs abundantly in nature, but only in a state of combination. The carbonate (CaCO_{3}), found as limestone, chalk, and other rocks, and as the minerals calcite and arragonite, is the most commonly occurring compound. The hydrated sulphate, gypsum (CaSO_{4}.2H_{2}O), is common, and is used in making "plaster of Paris." Anhydrite (CaSO_{4}) also occurs in rock masses, and is often associated with rock salt. Phosphate of lime, in the forms of apatite, phosphorite, coprolite, &c., is largely mined. Lime is a component of most natural silicates. Calcium also occurs, combined with fluorine, in the mineral fluor (CaF_{2}). In most of these the acid is the important part of the mineral; it is only the carbonate which is used as a source of lime.

Lime, in addition to its use in mortars and cements, is valuable as a flux in metallurgical operations, and as a base in chemical work on a large scale. A mixture of lime and magnesia is used in the manufacture of basic fire-bricks.

Carbonate of lime on ignition, especially when in contact with reducing substances, loses carbonic acid, and becomes lime. This is known as "quicklime"; on treatment with water it becomes hot, expands, and falls to a powder of "slaked lime" or calcium hydrate (CaH{2}O{2}). The hydrate is slightly soluble in water (0.1368 gram in 100 c.c.), forming an alkaline solution known as lime-water. Calcium hydrate is more generally used suspended in water as "milk of lime."

As a flux it is used either as limestone or as quicklime. Silica forms with lime a compound, calcium silicate, which is not very fusible; but when alumina and other oxides are present, as in clays and in most rocky substances, the addition of lime gives a very fusible slag.

Detection.—Calcium is detected by the reddish colour which its salts impart to the flame. It is best to moisten with hydrochloric acid (or, in the case of some silicates, to treat with ammonium fluoride) before bringing the substance into the flame. When seen through a spectroscope, it shows a large number of lines, of which a green and an orange are most intense and characteristic. Calcium is detected in solution (after removal of the metals by treatment with sulphuretted hydrogen and ammonium sulphide) by boiling with ammonium oxalate and ammonia. The lime is completely thrown down as a white precipitate. Lime is distinguished from the other alkaline earths by forming a sulphate insoluble in dilute alcohol, but completely soluble in a boiling solution of ammonium sulphate.

Lime compounds are for the most part soluble in water or in dilute hydrochloric acid. Calcium fluoride must be first converted into sulphate by evaporation in a platinum dish with sulphuric acid. Insoluble silicates are opened up by fusion with "fusion mixture," as described under Silica.

Separation.—The separation of lime is effected by evaporating with hydrochloric acid, to separate silica; and by treating with sulphuretted hydrogen, to remove the second group of metals. If the substance contains much iron, the solution is next oxidised by boiling with a little nitric acid; and the iron, alumina, &c., are removed as basic acetates. The filtrate is treated with ammonia and sulphuretted hydrogen, and allowed to settle. The filtrate from this is heated to boiling, treated with a solution of ammonium oxalate in excess, boiled for five or ten minutes, allowed to settle for half an hour, and filtered. The precipitate contains all the lime as calcium oxalate.

GRAVIMETRIC DETERMINATION.

The precipitate of calcium oxalate is washed with hot water, dried, transferred to a weighed platinum crucible, and ignited at a temperature not above incipient redness. This ignition converts the oxalate into carbonate, with evolution of carbonic oxide, which burns at the mouth of the crucible with a blue flame.[92] Generally a small quantity of the carbonate is at the same time converted into lime. To reconvert it into carbonate, moisten with a few drops of ammonic carbonate solution, and dry in a water-oven. Heat gently over a Bunsen burner, cool, and weigh. The substance is calcium carbonate (CaCO_{3}), and contains 56 per cent. of lime (CaO). It is a white powder, and should show no alkaline reaction with moistened litmus-paper.

Where the precipitate is small, it is better to ignite strongly over the blowpipe, and weigh directly as lime. With larger quantities, and when many determinations have to be made, it is easier to make the determination volumetrically.

VOLUMETRIC METHODS.

These are carried out either by dissolving the oxalate at once in dilute sulphuric acid, and titrating with permanganate of potassium solution; or by calcining it to a mixture of lime and carbonate, and determining its neutralising power with the standard solutions of acid and alkali.

Titration with Permanganate of Potassium Solution.—This solution is made by dissolving 5.643 grams of the salt in water, and by diluting to 1 litre; 100 c.c. are equivalent to 0.5 gram of lime. The solution is standardised by titrating a quantity of oxalic acid about equivalent to the lime present in the assay; 0.5 gram of lime is equivalent to 1.125 gram of crystallised oxalic acid. The standardising may be done with iron. The standard found for iron multiplied by 0.5 gives that for lime.

The process is as follows:—The calcium oxalate (having been precipitated and washed, as in the gravimetric process) is washed through the funnel into a flask with hot dilute sulphuric acid, boiled till dissolved, diluted to 200 c.c. with water, and heated to about 80 C. The standard solution of "permanganate" is then run in, (not too quickly, and with constant shaking) until a permanent pink tinge is produced. The c.c. used multiplied by the standard, and divided by the weight of the substance taken, will give the percentage of lime.

Estimation of Lime by Alkalimetry.—The methods of determining the amount of an alkali or base by means of a standard acid solution, or, conversely, of determining an acid by means of a standard alkaline solution, are so closely related that they are best considered under one head. The same standard solution is applicable for many purposes, and, consequently, it is convenient to make it of such strength that one litre of it shall equal an equivalent in grams of any of the substances to be determined. Such solutions are termed _normal_. For example, a solution of hydrochloric acid (HCl = 36.5) containing 36.5 grams of real acid per litre, would be normal and of equivalent strength to a solution containing either 17 grams of ammonia (NH_{3} = 17) or 40 grams of sodic hydrate (NaHO = 40) per litre. It will be seen in these cases that the normal solution contains the molecular weight in grams per litre; and, if solutions of these strengths be made, it will be found that they possess equal neutralising value.

If, now, a solution containing 98 grams of sulphuric acid (H{2}SO{4} = 98) per litre be made, it will be found to have twice the strength of the above solution, that is, 100 c.c. of the soda would only require 50 c.c. of the acid to neutralise it. The reason for this will be seen on inspecting the equations:—

NaHO + HCl = NaCl + H{2}O. 2NaHO + H{2}SO{4} = Na{2}SO{4} + 2H{2}O.

Acids like sulphuric acid are termed bibasic, and their equivalent is only half the molecular weight. Thus, a normal solution of sulphuric acid would contain 49 grams (98/2) of real acid per litre. Similarly, lime and most of the bases are bibasic, as may be seen from the following equations; hence their equivalent will be half the molecular weight.

2HCl + CaO = CaCl{2} + H{2}O. 2HCl + MgO = MgCl{2} + H{2}O.

The standard normal solution of hydrochloric acid is made by diluting 100 c.c. of the strong acid to one litre with water. This will be approximately normal. In order to determine its exact strength, weigh up 3 grams of recently ignited pure sodium carbonate or of the ignited bicarbonate. Transfer to a flask and dissolve in 200 c.c. of water; when dissolved, cool, tint faintly yellow with a few drops of a solution of methyl orange, and run in the standard "acid " from a burette till the yellow changes to a pink. Read off the number of c.c. used, and calculate to how much sodium carbonate 100 c.c. of the "acid" are equivalent. If the "acid" is strictly normal, this will be 5.3 grams. It will probably be equivalent to more than this. Now calculate how much strictly normal "acid" would be equivalent to the standard found. For example: suppose the standard found is 5.5 gram of sodium carbonate, then—

5.3 : 5.5 :: 100 : x (where x is the quantity of normal "acid" required). x = 103.8 c.c.

To get the "acid" of normal strength, we should then add 3.8 c.c. of water to each 100 c.c. of the standard solution remaining. Suppose there were left 930 c.c. of the approximate "acid," 35.3 c.c. of water must be added and mixed. It should then be checked by another titration with pure sodium carbonate.

_The standard solution of semi-normal "alkali."_ The best alkali for general purposes is ammonia, but, since it is volatile (especially in strong solutions), it is best to make it of half the usual strength, or _semi-normal_. One litre of this will contain 8.5 grams of ammonia (NH_{3}), and 100 c.c. of it will just neutralise 50 c.c. of the normal "acid." Take 100 c.c. of dilute ammonia and dilute with water to one litre. Run into a flask 50 c.c. of the standard "acid," tint with methyl orange, and run in from a burette the solution of ammonia till neutralised. Less than 100 c.c. will probably be used. Suppose 95 c.c. were required, there should have been 100, hence there is a deficiency of five. Then, for each 95 c.c. of standard "ammonia" left, add 5 c.c. of water, and mix well. 100 c.c. will now be equivalent to 50 c.c. of the "acid."

As an example of the application of this method, we may take the determination of lime in limestone, marble, and similar substances.

Determination of Lime in Limestone.—Weigh up 1 gram of the dried sample, and dissolve in 25 c.c. of normal acid, cool, dilute to 100 c.c., and titrate with the semi-normal solution of alkali (using methyl-orange as an indicator). Divide the c.c. of alkali used by 2, subtract from 25, and multiply by 0.028 to find the weight of lime. This method is not applicable in the presence of other carbonates or oxides, unless the weight of these substances be afterwards determined and due correction be made.

STRONTIA.

Strontia, the oxide of strontium (SrO), occurs in nature as sulphate, in the mineral celestine (SrSO{4}), and as carbonate in strontianite (SrCO{3}). It is found in small quantities in limestones, chalk, &c.

Strontia is used in sugar-refining, and for the preparation of coloured lights.

Detection.—It is detected by the crimson colour which its compounds (when moistened with hydrochloric acid) impart to the flame. The spectrum shows a large number of lines, of which a red, an orange, and a blue are most characteristic.

It resembles lime in many of its compounds, but is distinguished by the insolubility of its sulphate in a boiling solution of ammonium sulphate, and by the insolubility of its nitrate in alcohol. From baryta, which it also resembles, it is distinguished by not yielding an insoluble chromate in an acetic acid solution, by the solubility of its chloride in alcohol, and by the fact that its sulphate is converted into carbonate on boiling with a solution formed of 3 parts of potassium carbonate and 1 of potassium sulphate.

It is got into solution in the same manner as lime. The sulphate should be fused with "fusion mixture," extracted with water, and thoroughly washed. The residue will contain the strontia as carbonate, which is readily soluble in dilute hydrochloric or nitric acid.

Separation.—It is separated (after removal of the silica and metals, as described under Lime) by adding ammonia and ammonia carbonate, and allowing to stand for some hours in a warm place. In the absence of baryta or lime it is filtered off, and weighed as strontium carbonate, which contains 70.17 per cent. of strontia. It is separated from baryta by dissolving in a little hydrochloric acid, adding ammonia in excess, and then acidifying with acetic acid, and precipitating the baryta with potassium bichromate, as described under Baryta. The strontia is precipitated from the filtrate by boiling for some time with a strong solution of ammonic sulphate and a little ammonia. Fifty parts of ammonic sulphate are required for each part of strontia or lime present. The precipitate is filtered off, and washed first with a solution of ammonic sulphate, and then with alcohol. It is dried, ignited and weighed as strontium sulphate.

GRAVIMETRIC DETERMINATION.

The determination of strontia in pure solutions is best made by adding sulphuric acid in excess and alcohol in volume equal to that of the solution. Allow to stand overnight, filter, wash with dilute alcohol, dry, ignite at a red heat, and weigh as sulphate (SrSO_{4}). This contains 56.4 per cent. of strontia (SrO); or 47.7 per cent. of strontium.

BARYTA.

Baryta, oxide of barium (BaO), commonly occurs in combination with sulphuric oxide in the mineral barytes or heavy spar (BaSO{4}), and in combination with carbon dioxide in witherite (BaCO{3}). These minerals are not unfrequently found in large quantity (associated with galena and other metallic sulphides) in lodes. Small isolated crystals of these are frequently found in mining districts. Barium is a constituent of certain mineral waters. The minerals are recognised by their high specific gravity and their crystalline form.

Compounds of barium are often used by the assayer, more especially the chloride and hydrate. The salts are, with the exception of the sulphate, generally soluble in water or hydrochloric acid. In such solutions sulphuric acid produces a white precipitate of baric sulphate, which is practically insoluble in all acids.

The dioxide (BaO_{2}) is used for the preparation of oxygen. On strong ignition it gives up oxygen, and is converted into baryta (BaO), which, at a lower temperature, takes up oxygen from the air, re-forming the dioxide.

Detection.—Barium is detected by the green colour its salts, especially the chloride, give to the flame. This, viewed through the spectroscope, shows a complicated spectrum, of which two lines in the green are most easily recognised and characteristic. The salts of barium give no precipitate with sulphuretted hydrogen in either acid or alkaline solution, but with sulphuric acid they at once give a precipitate, which is insoluble in acetate of soda. In solutions rendered faintly acid with acetic acid, they give a yellow precipitate with bichromate of potash. These reactions are characteristic of barium.

Baryta is got into solution in the manner described under Lime; but in the case of the sulphate the substance is fused with three or four times its weight of "fusion mixture." The "melt" is extracted with water, washed, and the residue dissolved in dilute hydrochloric acid.

Separation.—The separation is thus effected:—The solution in hydrochloric acid is evaporated to dryness, re-dissolved in hot dilute hydrochloric acid, and sulphuric acid is added to the solution till no further precipitate is formed. The precipitate is filtered off, and digested with a solution of ammonium acetate or of sodium hyposulphite at 50 or 60 C. to dissolve out any lead sulphate. The residue is filtered off, washed, dried, and ignited. The ignited substance is mixed with four or five times its weight of "fusion mixture," and fused in a platinum-dish over the blowpipe for a few minutes. When cold, it is extracted with cold water, filtered, and washed. The residue is dissolved in dilute hydrochloric acid, and (if necessary) filtered. The solution contains the barium as baric chloride mixed, perhaps, with salts of strontium or lime. To separate these, ammonia is added till the solution is alkaline, and then acetic acid in slight excess. Chromate of baryta is then thrown down, by the addition of bichromate of potash, as a yellow precipitate. It is allowed to settle, filtered and washed with a solution of acetate or of nitrate of ammonia. It is dried, ignited gently, and weighed. It is BaCrO_{4}, and contains 60.47 per cent. of baryta.

GRAVIMETRIC DETERMINATION.

The gravimetric determination of baryta, when lime and strontia are absent, is as follows:—The solution, if it contains much free acid, is nearly neutralised with ammonia, and then diluted to 100 or 200 c.c. It is heated to boiling, and dilute sulphuric acid is added till no further precipitation takes place. The precipitate is allowed to settle for a few minutes, decanted through a filter, and washed with hot water; and, afterwards, dried, transferred to a porcelain crucible, and strongly ignited in the muffle or over the blowpipe for a few minutes. It is then cooled, and weighed as sulphate of baryta (BaSO_{4}). It contains 65.67 per cent. of baryta (BaO).

In determining the baryta in minerals which are soluble in acid, it is precipitated direct from the hydrochloric acid solution (nearly neutralised with ammonia) by means of sulphuric acid. The precipitated baric sulphate is digested with a solution of ammonic acetate; and filtered, washed, ignited, and weighed.

VOLUMETRIC DETERMINATION.

The principle and mode of working of this is the same as that given under the Sulphur Assay; but using a standard solution of sulphuric acid instead of one of barium chloride. The standard solution of sulphuric acid is made to contain 32.02 grams of sulphuric acid (H{2}SO{4}), or an equivalent of a soluble alkaline sulphate, per litre. 100 c.c. will be equal to 5 grams of baryta.

Five grams of the substance are taken, and the baryta they contain converted into carbonate (if necessary). The carbonate is dissolved in dilute hydrochloric acid. Ten grams of sodium acetate are added, and the solution, diluted to 500 c.c., is boiled, and titrated in the manner described.

Lead salts must be absent in the titration, and so must strontia and lime. Ferrous salts should be peroxidised by means of permanganate or chlorate of potash. Other salts do not interfere.

MAGNESIA.

Magnesia, the oxide of magnesium (MgO) occurs in nature in the rare mineral periclase (MgO); and hydrated, as brucite (MgH_{2}O_{2}). As carbonate it occurs in large quantity as magnesite (MgCO_{3}), which is the chief source of magnesia. Mixed with carbonate of lime, it forms magnesian limestone and dolomite. It is present in larger or smaller quantity in most silicates; and the minerals, serpentine, talc, steatite and meerschaum are essentially hydrated silicates of magnesia. Soluble magnesian salts occur in many natural waters; more especially the sulphate and the chloride. Kieserite (MgSO_{4}.H_{2}O) occurs in quantity at Stassfurt, and is used in the manufacture of Epsom salts.

Detection.—Magnesia is best detected in the wet way. Its compounds give no colour to the flame, and the only characteristic dry reaction is its yielding a pink mass when ignited before the blowpipe (after treatment with a solution of cobalt nitrate). In solution, it is recognised by giving no precipitate with ammonia or ammonic carbonate in the presence of ammonic chloride, and by giving a white crystalline precipitate on adding sodium phosphate or arsenate to the ammoniacal solution.

Magnesia differs from the other alkaline earths by the solubility of its sulphate in water.

Magnesia is dissolved by boiling with moderately strong acids; the insoluble compounds are fused with "fusion mixture," and treated as described under Silicates.

Separation.—It is separated by evaporating the acid solution to dryness to render silica insoluble, and by taking up with dilute hydrochloric acid. The solution is freed from the second group of metals by means of sulphuretted hydrogen, and the iron, alumina, &c., are removed with ammonic chloride, ammonia, and ammonic sulphide. The somewhat diluted filtrate is treated, first, with ammonia, and then with carbonate of ammonia in slight excess. It is allowed to stand for an hour in a warm place, and then filtered. The magnesia is precipitated from the filtrate by the addition of an excess of sodium phosphate and ammonia. It is allowed to stand overnight, filtered, and washed with dilute ammonia. The precipitate contains the magnesia as ammonic-magnesic phosphate.

In cases where it is not desirable to introduce sodium salts or phosphoric acid into the assay solution, the following method is adopted. The solution (freed from the other alkaline earths by ammonium carbonate) is evaporated in a small porcelain dish with nitric acid. The residue (after removing the ammonic salts by ignition) is taken up with a little water and a few crystals of oxalic acid, transferred to a platinum dish, evaporated to dryness, and ignited. The residue is extracted with small quantities of boiling water and filtered off; while the insoluble magnesia is washed. The filtrate contains the alkalies. The residue is ignited, and weighed as magnesia. It is MgO.

GRAVIMETRIC DETERMINATION.

The solution containing the magnesia is mixed with chloride of ammonium and ammonia in excess. If a precipitate should form, more ammonic chloride is required. Add sodium phosphate solution in excess, stir and allow to stand overnight. Filter and wash the precipitate with dilute ammonia. Dry, transfer to a platinum or porcelain crucible, and ignite (finally at intense redness); cool, and weigh. The substance is magnesic pyrophosphate (Mg_{2}P_{2}O_{7}), and contains 36.04 per cent. of magnesia.

VOLUMETRIC METHOD.

The magnesia having been precipitated as ammonic-magnesic phosphate, which is the usual separation, its weight can be determined volumetrically by the method of titration described under Phosphates.

The same standard solution of uranium acetate is used. Its standard for magnesia is got by multiplying the standard for phosphoric oxide by 0.5493. For example, if one hundred c.c. are equivalent to 0.5 gram of phosphoric oxide, they will be equivalent to (0.5 .5493) 0.2746 gram of magnesia. The method of working and the conditions of the titration are the same as for the phosphate titration. The quantity of substance taken for assay must not contain more than 0.1 or 0.2 gram of magnesia. After precipitating as ammonic-magnesic phosphate with sodium phosphate, and well washing with ammonia, it is dissolved in dilute hydrochloric acid, neutralised with ammonia, and sodic acetate and acetic acid are added in the usual quantity. The solution is boiled and titrated.

EXAMINATION OF A LIMESTONE.

Silica and Insoluble Silicates.—Take one gram of the dried sample and dissolve it in 10 c.c. of dilute hydrochloric acid; filter; wash, dry, and ignite the residue.

Organic Matter.—If the residue insoluble in hydrochloric acid shows the presence of organic matter, it must be collected on a weighed filter and dried at 100. On weighing, it gives the combined weights of organic and insoluble matter. The latter is determined by igniting and weighing again. The organic matter is calculated by difference.

Lime.—Where but little magnesia is present, this is determined by titration with standard acid. Take one gram, and dissolve it in 25 c.c. of normal hydrochloric acid. Tint with methyl-orange and titrate with semi-normal ammonia. Divide the quantity of ammonia used by 2, deduct this from 25, and multiply the remainder by 2.8. This gives the percentage of lime. Where magnesia is present, the same method is adopted, and the magnesia (which is separately determined) is afterwards deducted. The percentage of magnesia found is multiplied by 1.4, and the result is deducted from the apparent percentage of lime got by titrating.

Magnesia.—Dissolve 2 grams of the limestone in hydrochloric acid, and separate the lime with ammonia and ammonium oxalate. The filtrate is treated with sodium phosphate, and the magnesia is weighed as pyrophosphate, or titrated with uranium acetate.

Iron.—Dissolve 2 grams in hydrochloric acid, reduce, and titrate with standard permanganate of potassium solution. This gives the total iron. The ferrous iron is determined by dissolving another 2 grams in hydrochloric acid and at once titrating with the permanganate of potassium solution.

Manganese.—Dissolve 20 grams in hydrochloric acid, nearly neutralise with soda, add sodium acetate, boil, and filter. To the filtrate add bromine; boil, and determine the manganese in the precipitate. See page 300.

Phosphoric Oxide.—This is determined by dissolving the ferric acetate precipitate from the manganese separation in hydrochloric acid, adding ammonia in excess, and passing sulphuretted hydrogen. Filter and add to the filtrate "magnesia mixture." The precipitate is collected, washed with ammonia, ignited, and weighed as pyrophosphate.

THE ALKALIES.

The oxides of sodium, potassium, lithium, csium, and rubidium and ammonia are grouped under this head. Of these csia and rubidia are rare, and lithia comparatively so. They are easily distinguished by their spectra. They are characterised by the solubility of almost all their salts in water, and, consequently, are found in the solutions from which the earths and oxides of the metals have been separated by the usual group re-agents.

The solution from which the other substances have been separated is evaporated to dryness, and the product ignited to remove the ammonic salts added for the purpose of separation. The residue contains the alkali metals generally, as chlorides or sulphates. Before determining the quantities of the particular alkali metals present, it is best to convert them altogether, either into chloride or sulphate, and to take the weight of the mixed salts. It is generally more convenient to weigh them as chlorides. They are converted into this form, if none of the stronger acids are present, by simply evaporating with an excess of hydrochloric acid. Nitrates are converted into chlorides by this treatment. When sulphates or phosphates are present, the substance is dissolved in a little water, and the sulphuric or phosphoric acid precipitated with a slight excess of acetate of lead in the presence of alcohol. The solution is filtered, and the excess of lead precipitated with sulphuretted hydrogen. The filtrate from this is evaporated to dryness with an excess of hydrochloric acid, and the residue, consisting of the mixed chlorides, is gently ignited and weighed. In many cases (such as the analysis of slags and of some natural silicates where the percentage of alkalies is small) the percentage of soda and potash (which most commonly occur) need not be separately determined. It is sufficient to report the proportion of mixed alkalies; which is thus ascertained:—Dissolve the ignited and weighed chlorides in 100 c.c. of distilled water, and titrate with the standard solution of silver nitrate (using potassic chromate as indicator) in the manner described under Chlorine. The c.c. of silver nitrate used gives the weight in milligrams of the chlorine present. Multiply this by 0.775, and deduct the product from the weight of the mixed chlorides. This will give the combined weight of the alkalies (Na{2}O and K{2}O) present. For example, 0.0266 gram of mixed chlorides required on titrating 14.2 c.c. of silver nitrate, which is equivalent to 0.0142 gram of chlorine. This multiplied by 0.775 gives 0.0110 to be deducted from the weight of the mixed chlorides.

Mixed chlorides 0.0266 gram Deduction 0.0110 " ——— Mixed alkalies 0.0156 "

Assuming this to have been got from 1 gram of a rock, it would amount to 1.56 per cent. of "potash and soda."

The relative proportions of the potash and soda can be ascertained from the same determination. Sodium and potassium chlorides have the following composition:—

Sodium 39.38 Potassium 52.46 Chlorine 60.62 Chlorine 47.54 ——- ——- 100.00 100.00

The percentage of chlorine in the mixed chlorides is calculated. It necessarily falls somewhere between 47.5 and 60.6 per cent., and approaches the one or the other of these numbers as the proportion of the sodium or potassium preponderates. Each per cent. of chlorine in excess of 47.5 represents 7.63 per cent. of sodium chloride in the mixed chlorides. The percentage of potash and soda in the substance can be calculated in the usual way. Sodium chloride multiplied by 0.5302 gives its equivalent of soda (Na{2}O), and potassium chloride multiplied by 0.6317 gives its equivalent of potash (K{2}O).

The weight of sodium chloride in the mixed chlorides is also calculated thus:—Take the same example for illustration. Multiply the chlorine found by 2.103. This gives—

(0.01422.103) = 0.02987.

From the product deduct the weight of the mixed chlorides found—

Product 0.02987 Mixed chlorides 0.02660 ———- Difference 0.00327

The difference multiplied by 3.6288 gives the weight of sodium chloride in the mixture. In this case it equals 0.0118 gram. The potassium chloride is indicated by the difference between this and the weight of the mixed chlorides. It equals 0.0148 gram. We have now got—

Sodium chloride 0.0118 gram Potassium chloride 0.0148 "

from 1 gram of the rock taken. Multiplying these by their factors we have (Soda = 0.01180.5302; Potash 0.01480.6317)—

Soda = 0.625 per cent. Potash = 0.935 "

Concentration of the Alkalies.—With the exception of magnesia, all the other bases are separated from the alkalies in the ordinary course of work without the addition of any re-agent which cannot be removed by simple evaporation and ignition. Consequently, with substances soluble in acids, successive treatment of the solution with sulphuretted hydrogen, ammonia, ammonic sulphide, and ammonic carbonate, filtering, where necessary, will yield a filtrate containing the whole of the alkalies with ammonic salts and, perhaps, magnesia.

The filtrate is evaporated in a small porcelain dish, with the addition of nitric acid towards the finish. It is carried to dryness and ignited. The residue is taken up with a little water, treated with a few crystals of oxalic acid, and again evaporated and ignited. The alkaline salts are extracted with water, and filtered from the magnesia into a weighed platinum dish. The solution is then evaporated with an excess of hydrochloric acid, ignited at a low red heat, and weighed. The residue consists of the mixed alkaline chlorides.

For substances (such as most silicates and similar bodies) not completely decomposed by acids, Lawrence Smith's method is generally used. This is as follows:—Take from 0.5 to 1 gram of the finely powdered mineral, and mix, by rubbing in the mortar, with an equal weight of ammonium chloride. Then mix with eight times as much pure calcium carbonate, using a part of it to rinse out the mortar. Transfer to a platinum crucible, and heat gently over a Bunsen burner until the ammonic chloride is decomposed (five or ten minutes). Raise the heat to redness, and continue at this temperature for about three quarters of an hour. The crucible must be kept covered. Cool, and turn out the mass into a 4-inch evaporating dish; wash the crucible and cover with distilled water, and add the washings to the dish; dilute to 60 or 80 c.c., and heat to boiling. Filter and wash. Add to the filtrate about 1.5 gram of ammonium carbonate; evaporate to about 40 c.c., and add a little more ammonic carbonate and some ammonia. Filter into a weighed platinum dish, and evaporate to dryness. Heat gently, to drive off the ammonic chloride, and ignite to a little below redness. Cool and weigh. The residue consists of the mixed alkaline chlorides.

Separation of the Alkali-Metals from each other.—Sodium and lithium are separated from the other alkali-metals by taking advantage of the solubility of their chlorides in the presence of platinic chloride; and from one another by the formation of an almost insoluble lithic phosphate on boiling with a solution of sodium phosphate in a slightly alkaline solution. Csium, rubidium, and potassium yield precipitates with platinic chloride, which are somewhat soluble, and must be precipitated from concentrated solutions. Csium and rubidium are separated from potassium by fractional precipitation with platinum chloride. Their platino-chlorides, being less soluble than that of potassium, are precipitated first. One hundred parts of boiling water dissolve 5.18 of the potassium platino-chloride, 0.634 of the rubidium salt, and 0.377 of the corresponding csium compound. The separation of lithium, csium, and rubidium is seldom called for, owing to their rarity. The details of the separation of potassium from sodium are described under Potassium. Ammonia compounds are sharply marked off from the rest by their volatility, and it is always assumed that they have been removed by ignition; if left in the solution, they would count as potassium compounds. They will be considered under Ammonia.

SODIUM.

Sodium is the commonest of the alkali metals. It is found in nature chiefly combined with chlorine as "common salt" (NaCl). This mineral is the source from which the various compounds of sodium in use are prepared. Sodium occurs abundantly as nitrate (NaNO_{3}) in Chili saltpetre, and as silicate in various minerals, such as albite (or soda-felspar).

It occurs as fluoride in cryolite (Na{3}AlF{6}), and as carbonate in natron, &c. Sulphates are also found. Sodium is very widely diffused, few substances being free from it.

The detection of sodium is easy and certain, owing to the strong yellow colour its salts impart to the flame; this, when viewed by the spectroscope, shows a single yellow line.[93] The extreme delicacy of this test limits its value, because of the wide diffusion of sodium salts. It is more satisfactory to separate the chloride, which may be recognised by its taste, flame coloration, fusibility, and negative action with reagents. The chloride dissolved in a few drops of water gives with potassium metantimoniate, a white precipitate of the corresponding sodium salt.

Sodium salts are dissolved out from most compounds on treatment with water or dilute acids. Insoluble silicates are decomposed and the alkali rendered soluble by Lawrence Smith's method, which has just been described. The separation of the sodium from the mixed chlorides is effected in the following way:—The chlorides are dissolved in a little water and the potassium separated as platino-chloride. The soluble sodium platino-chloride, with the excess of platinum, is boiled, mixed with sulphuric acid, evaporated to dryness, and ignited. On extracting with water, filtering, evaporating, and igniting, sodium sulphate is left, and is weighed as such.

It is more usual, and quite as satisfactory, to calculate the weight of the sodium chloride by difference from that of the mixed chlorides, by subtracting that of the potassium chloride, which is separately determined. For example, 1 gram of a rock gave—Mixed chlorides, 0.0266 gram, and 0.0486 gram of potassic platino-chloride. This last is equivalent to 0.0149 gram of potassium chloride.

Mixed chlorides found 0.0266 Deduct potassium chloride 0.0149 ——— Leaves sodium chloride 0.0117

The weight of sodium chloride found, multiplied by 0.5302, gives the weight of the soda (Na_{2}O).

GRAVIMETRIC DETERMINATION.

The solution, which must contain no other metal than sodium, is evaporated in a weighed platinum crucible or dish. Towards the finish an excess, not too great, of sulphuric acid is added, and the evaporation is continued under a loosely fitting cover. The residue is ignited over the blowpipe, a fragment of ammonic carbonate being added towards the end, when fumes of sulphuric acid cease to be evolved. This ensures the removal of the excess of acid. The crucible is cooled in the desiccator, and weighed. The substance is sulphate of soda (Na_{2}SO_{4}), and contains 43.66 per cent. of soda (Na_{2}O), or 32.38 per cent. of sodium (Na).

VOLUMETRIC METHODS.

There are various methods used for the different compounds of sodium. There is no one method of general application. Thus with "common salt" the chlorine is determined volumetrically; and the sodium, after deducting for the other impurities, is estimated by difference.

With sodic carbonate and caustic soda, a given weight of the sample is titrated with standard acid, and the equivalent of soda estimated from the alkalinity of the solution.

With sodium sulphate, a modification of the same method is used. To a solution of 3.55 grams of the salt contained in a half-litre flask, 250 c.c. of a solution of baryta water is added. The volume is made up to 500 c.c. with water. The solution is mixed and filtered. Half of the filtrate is measured off, treated with a current of carbonic acid, and then boiled. It is transferred to a half-litre flask, diluted to the mark, shaken up, and filtered. 250 c.c. of the filtrate, representing a quarter of the sample taken, is then titrated with standard acid. The standard acid is made by diluting 250 c.c. of the normal acid to 1 litre. The c.c. of acid used multiplied by 2 gives the percentage. A correction must be made to counteract the effect of impurities in the baryta as well as errors inherent in the process. This is small, and its amount is determined by an experiment with 3.55 grams of pure sodium sulphate.

EXAMINATION OF COMMON SALT.

Moisture.—Powder and weigh up 10 grams of the sample into a platinum dish. Dry in a water oven for an hour, and afterwards heat to bare redness over a Bunsen burner. Cool, and weigh. The loss gives the water.

Chlorine.—Weigh up two separate lots of 1 gram each; dissolve in 100 c.c. of water, and determine the chlorine by titrating with the standard silver nitrate solution, using chromate of potash as indicator. See Chlorine.

Insoluble Matter.—Dissolve 10 grams of the salt in water with the help of a little hydrochloric acid. Filter off the sediment, wash, ignite, and weigh. This residue is chiefly sand. Dilute the nitrate to 500 c.c.

Lime.—Take 250 c.c. of the filtrate, render ammoniacal and add ammonium oxalate; wash, dry, and ignite the precipitate. Weigh as lime (CaO).

Magnesia.—To the filtrate from the lime add phosphate of soda. Allow to stand overnight, filter, wash with dilute ammonia, dry, ignite, and weigh as pyrophosphate.

Sulphuric Oxide.—To the remaining 250 c.c. of the filtrate from the "insoluble," add an excess of barium chloride. Collect, wash, dry, ignite, and weigh the barium sulphate.

Sodium.—It is estimated by difference.

The following may be taken as an example:—

Moisture 0.35 Insoluble matter 0.40 Lime 0.40 Magnesia 0.05 Sulphuric oxide 0.60 Chlorine 59.60 Sodium 38.60 ——— 100.00

POTASSIUM.

Potassium occurs in nature as chloride, in the mineral sylvine (KCl), and more abundantly combined with magnesium chloride, in earnallite (KCl.MgCl_{2}.6H_{2}O). It occurs as nitrate in nitre (KNO_{3}), and as silicate in many minerals, such as orthoclase (or potash-felspar) and muscovite (or potash-mica).

Potassium compounds are detected by the characteristic violet colour they impart to the flame. The presence of sodium salts masks this tint, but the interference can be neutralised by viewing the flame through a piece of blue glass. Viewed through the spectroscope, it shows a characteristic line in the red and another in the violet. These, however, are not so easy to recognise or obtain as the sodium one. Concentrated solutions of potassium salts give a yellow crystalline precipitate with platinum chloride, and a white crystalline one with the acid tartrate of soda. For these tests the solution is best neutral. These tests are only applicable in the absence of compounds other than those of potassium and sodium.

GRAVIMETRIC DETERMINATION.

This process serves for its separation from sodium. Take 1 gram of the sample and dissolve it in an evaporating dish with 50 c.c. of water. Acidify with hydrochloric acid in quantity sufficient (if the metals are present as chlorides) to make it acid, or, if other acids are present, in at least such quantity as will provide the equivalent of chlorine. Add 3 grams of platinum, in solution as platinum chloride, and evaporate on a water-bath to a stiff paste, but not to dryness. Moisten with a few drops of platinic chloride solution without breaking up the paste by stirring. Cover with 20 c.c. of strong alcohol, and wash the crystals as much as possible by rotating the dish. Allow to settle for a few moments, and decant through a filter. Wash in the same way two or three times until the colour of the filtrate shows that the excess of the platinum chloride used is removed. Wash the precipitate on to the filter with a jet of alcohol from the wash-bottle; clean the filter-paper, using as little alcohol as possible. Dry in the water-oven for an hour. Brush the precipitate into a weighed dish, and weigh it. It is potassium platino-chloride (K_{2}PtCl_{6}), and contains 16.03 per cent. of potassium, or 30.56 per cent. of potassium chloride (KCl), which is equivalent to 19.3 per cent. of potash (K_{2}O).

If the filter-paper is not free from precipitate, burn it and weigh separately. The excess of weight over that of the ash will be due to platinum and potassic chloride (Pt and 2KCl). This multiplied by 1.413 will give the weight of the potassic platino-chloride from which it was formed. It must be added to the weight of the main precipitate.

The mixed alkaline chlorides obtained in the usual course of analysis are treated in this manner; the quantity of platinum added must be about three times as much as the mixed chlorides weigh.

VOLUMETRIC METHODS.

These are the same as with soda.

Examination of Commercial Carbonate of Potash.—The impurities to be determined are moisture, silica, and insoluble matter, chlorine, sulphuric oxide, and oxide of iron. These determinations are made in the ways described under the examination of common salt.

The potassium is determined by converting it into chloride and precipitating with platinum chloride, &c., as just described.

Available Alkali.—Weigh up 23.5 grams of the sample, dissolve in water, and make up to 500 c.c. Take 50 c.c., tint with methyl orange, and titrate with the normal solution of acid. The c.c. of acid used multiplied by 2 gives the percentage of available alkali calculated as potash (K_{2}O).

Soda.—This is calculated indirectly in the following way:—Deduct from the potassium found the quantity required for combination with the chlorine and sulphuric oxide present, and calculate the remainder to potash (K_{2}O). The apparent surplus excess of available alkali is the measure of the soda present.

Carbon Dioxide.—The c.c. of acid used in the available alkali determination, multiplied by 2.2 and divided by 2.35, gives the percentage of carbon dioxide.

LITHIUM.

Lithia, the oxide of lithium (Li_{2}O), occurs in quantities of 3 or 4 per cent. in various silicates, such as lepidolite (or lithia-mica), spodumene, and petalite. It also occurs as phosphate in triphyline. It is a constituent of the water of certain mineral springs. A spring at Wheal Clifford contained as much as 0.372 gram of lithium chloride per litre. In small quantities, lithia is very widely diffused.

The Detection of lithia is rendered easy by the spectroscope; its spectrum shows a red line lying about midway between the yellow sodium line and the red one of potassium. It also shows a faint yellow line. The colour of the flame (a crimson) is characteristic.

The reactions of the lithium compounds lie between those of the alkalies and of the alkaline earths. Solutions are not precipitated by tartaric acid nor by platinic chloride. The oxide is slowly soluble in water. The carbonate is not freely soluble. Lithia is completely precipitated by sodic phosphate, especially in hot alkaline solutions.

In its determination the mixed alkaline chlorides obtained in the separation of the alkalies are dissolved in water, a solution of soda is added in slight excess, and the lithia precipitated with _sodic_ phosphate. Before filtering, it is evaporated to dryness and extracted with hot water rendered slightly ammoniacal. The residue is transferred to a filter, dried, ignited, and weighed. The precipitate is lithium phosphate (3Li_{2}O, P_{2}O_{5}), and contains 38.8 per cent. of lithia. The separation of lithia from magnesia is not given by the usual authorities. Wohler recommends evaporating the solution to dryness with carbonate of soda. On extracting the residue with water, the lithia dissolves out and is determined in the filtrate. One hundred parts of water dissolve, at the ordinary temperature, 0.769 parts of lithium carbonate (Li_{2}CO_{3}); the basic magnesia compound is almost insoluble in the absence of carbon dioxide and ammonium salts.

CAESIUM.

The oxide of caesium, caesia (Cs_{2}O), is found associated with lithia in lepidolite, &c., and, together with rubidium, in many mineral waters. The mineral pollux is essentially a silicate of alumina and caesia; it contains 34.0 per cent. of the latter oxide.

Caesium is best detected by the spectroscope, its spectrum being characterised by two lines in the blue and one in the red; the latter is about midway between the lithium and sodium lines.

If not detected by the spectroscope, or specially looked for, caesia would, in the ordinary course of work, be separated with the potash and weighed as potassium platino-chloride.

Caesia is separated from all the other alkalies by adding to the acid solution of the mixed chlorides a strongly acid cold solution of antimonious chloride. The acid used must be hydrochloric. The caesium is precipitated as a white crystalline precipitate (CsCl.SbCl_{3}), which is filtered off, and washed, when cold, with strong hydrochloric acid; since it is decomposed by water or on warming. The precipitate is washed into a beaker, and treated with sulphuretted hydrogen; after filtering off the sulphide of antimony, the solution leaves, on evaporation, the caesium as chloride.

RUBIDIUM.

Rubidium occurs widely diffused in nature, but in very small quantities. It is generally associated with caesium.

It is detected by the spectroscope, which shows two violet lines and two dark red ones. Like caesium, it is precipitated with platinic chloride, and in the ordinary course of work would be weighed as potassium. It is separated from potassium by fractional precipitation with platinic chloride. Rubidium platino-chloride is much less soluble than the potassium salt.

AMMONIUM.

It is usual to look upon the salts of ammonia as containing a compound radical (NH_{4} = Am), which resembles in many respects the metals of the alkalies. Ammonium occurs in nature as chloride in sal ammoniac (AmCl), as sulphate in mascagnine (Am_{2}SO_{4}), as phosphate in struvite (AmMgPO_{4}.12H_{2}O). Minerals containing ammonium are rare, and are chiefly found either in volcanic districts or associated with guano. Ammonia and ammonium sulphide occur in the waters of certain Tuscan lagoons, which are largely worked for the boracic acid they contain. The crude boracic acid from this source contains from 5 to 10 per cent. of ammonium salts. It is from these that the purer forms of ammonium compounds of commerce known as "from volcanic ammonia" are derived. But the bulk of the ammonia of commerce is prepared from the ammoniacal liquors obtained as bye-products in the working of certain forms of blast furnaces and coke ovens, and more especially in gas-making.

Ammonia hardly comes within the objects of assaying; but it is largely used in the laboratory, and the assayer is not unfrequently called on to determine it. Ammonium salts are mostly soluble in water. In strong solutions they give a yellow precipitate of ammonium platino-chloride on the addition of chloride of platinum; and with the acid tartrate of soda yield a white precipitate of hydric ammonic tartrate. These reactions are similar to those produced with potassium compounds.

Heated with a base, such as lime or sodic hydrate, ammonium salts are decomposed, yielding ammonia gas (NH_{3}), which is readily soluble in water. The solution of this substance is known as ammonic hydrate or "ammonia."

They are volatilised on ignition; either with, or without, decomposition according to the acid present. This fact is of importance in analytical work; since it allows of the use of alkaline solutions and reagents which leave nothing behind on heating. It must be remembered, however, that, although ammonic chloride is volatile, it cannot be volatilised in the presence of substances which form volatile chlorides without loss of the latter. For example: ferric oxide and alumina are thus lost, volatilising as chlorides; and there are some other compounds (notably ammonic magnesic arsenate) which on heating to redness suffer reduction. The presence of ammonic chloride in such cases must be avoided.

Detection.—Compounds of ammonium are detected by their evolving ammonia when mixed or heated with any of the stronger bases. The ammonia is recognised by its odour, by its alkaline reaction with litmus paper, and by yielding white fumes, when brought in contact with fuming acid. In consequence of the use of ammonium salts and ammonia as reagents, it is necessary to make a special test for and determination of ammonium.[94] In the ordinary course of work it will be "lost on ignition." The determination presents little difficulty, and is based on the method used for its detection.



Solution and Separation.—Although ammonium salts are soluble in water, there is no necessity for dissolving them. The compound containing the ammonia is boiled with an alkaline solution; and the liberated ammonia condensed and collected. The substance is weighed out into a flask of about 200 c.c. capacity. The flask is closed with a rubber cork perforated to carry a 20 c.c. pipette and a bulb exit tube. The latter is connected with a receiver, which is a small flask containing dilute hydrochloric acid (fig. 61). The flask containing the substance is corked, and the greater part of the soda solution is run in from the pipette. The solution is then boiled. The ammonia volatilises, and is carried over into the hydrochloric acid, with which it combines to form ammonic chloride. The distillation is carried on gently until the bulk of the liquid is driven over. The ammonia in the receiver will be mixed only with the excess of hydrochloric acid. This separation is used in all determinations.

GRAVIMETRIC DETERMINATION.

The contents of the flask are transferred to a weighed platinum dish, and evaporated on the water-bath. It is dried until the weight is constant. The chloride of ammonium remains as a white mass which, after cooling in a desiccator, is weighed. It contains 33.72 per cent. of ammonium (NH{4}), or 31.85 per cent. of ammonia (NH{3}). On heating over the Bunsen burner it is completely volatilised, leaving no residue.

VOLUMETRIC DETERMINATION.

Weigh up 1.7 gram of the substance and place it in the flask. Measure off 50 c.c. of the normal solution of acid, place them in the receiver, and dilute with an equal volume of water. Run in through the pipette (by opening the clip) 20 c.c. of a strong solution of soda, boil until the ammonia has passed over, and then aspirate a current of air through the apparatus. Disconnect the receiver, and tint its contents with methyl orange. Titrate the residual acid with a semi-normal solution of alkali. Divide the c.c. of the "alkali" solution used by 2, and deduct from the 50 c.c. The difference will give the number of c.c. of the normal acid solution neutralised by the ammonia distilled over. Each c.c. of "acid" so neutralised, represents 1 per cent. of ammonia in the sample. If the results are to be reported as ammonium, 1.8 gram of the sample is taken instead of 1.7 gram.

COLORIMETRIC DETERMINATION.

This is effected by means of "Nessler's" reagent, which strikes a brown colour with traces of ammonia, even with a few hundredths of a milligram in 100 c.c. of liquid. With larger quantities of ammonia the reagent gives a precipitate. This reagent is a strongly alkaline solution of potassic mercuric iodide; and is thus made:—

Nessler's solution: Dissolve 17 grams of mercuric chloride in 300 c.c. of water; and add the solution to one of 35 grams of potassium iodide in 100 c.c. of water until a permanent precipitate is produced. Both solutions must be cold. Then make up to a litre by adding a 20 per cent. solution of potash. Add more of the mercuric chloride (a little at a time) until a permanent precipitate is again formed. Allow to settle, decant, and use the clear liquor. Four or five c.c. are used for each 100 c.c. of liquid to be tested.

_A Standard Solution of Ammonia_ is made by dissolving 0.315 gram of ammonic chloride in water, and diluting to 100 c.c. Ten c.c. of this are taken and diluted to 1 litre. One c.c. contains 0.01 milligram of ammonia (NH_{3}).

In working, the solution containing the ammonia is diluted to a definite volume, and to such an extent that 50 c.c. of it shall not contain more than 0.02 or 0.03 milligram of ammonia. Fifty c.c. of it are transferred to a Nessler glass and mixed with 2 c.c. of Nessler's reagent. The colour is noted, and an estimate made as to the amount of ammonia it indicates. A measured quantity of the standard ammonia, judged to contain about as much ammonia as that in the assay, is then put into another Nessler glass. It is diluted to 50 c.c. with water, and mixed with 2 c.c. of "Nessler." After standing a minute or two, the colours in the two glasses are compared. If the tints are equal, the assay is finished; but if the standard is weaker or stronger than the assay, another standard, containing more or less ammonia, as the case may be, must be prepared and compared with the assay. Two such experiments will generally be sufficient; but, if not, a third must be made. The addition of more standard ammonia to the solution to which the "Nessler" has already been added does not give a satisfactory result.

When the ammonia in 50 c.c. has been determined, that in the whole solution is ascertained by a suitable multiplication. By 10, for example, if the bulk was 500 c.c., or by 20 if it was a litre.

Distilled water is used throughout. It must be free from ammonia; and is best prepared by distilling an ammonia-free spring water.

FOOTNOTES:

[90] Al_{2}Cl_{6} + 3Na_{2}S_{2}O_{3} + 3H_{2}O = Al_{2}(HO)_{6} + 6NaCl + 3S + 3SO_{2}

[91] 3BeO,Al_{2}O_{3},6SiO_{2}

[92] CaC_{2}O_{4} = CaCO_{3}+CO.

[93] Resolved into two with a powerful spectroscope.

[94] Ammonium compounds are frequently produced when dissolving metals in nitric acid; or when nitrates are heated in the presence of the metals.



PART III—NON-METALS.



CHAPTER XV.

OXYGEN AND OXIDES.—THE HALOGENS.

OXYGEN.

Oxygen occurs in nature in the free state, forming 23 per cent. by weight, or 21 per cent. by volume of the atmosphere; but, since it is a gas, its presence is easily overlooked and its importance underestimated. Except in the examination of furnace-gases, &c., the assayer is not often called upon to determine its quantity, but it forms one of his most useful reagents, and there are many cases where he cannot afford to disregard its presence. It occurs not only in the air, but also dissolved in water; ordinary waters containing on an average 0.00085 per cent. by weight, or 0.85 parts per 100,000.

Chemically, it is characterised by its power of combining, especially at high temperatures, with the other elements, forming an important class of compounds called oxides. This combination, when rapid, is accompanied by the evolution of light and heat; hence oxygen is generally called the supporter of combustion. This property is taken advantage of in the operation of calcining, scorifying, cupelling, &c. The importance of a free access of air in all such work is seen when it is remembered that 1 litre of air contains 0.2975 gram of oxygen, and this quantity will only oxidise 0.1115 gram of carbon, 0.2975 gram of sulphur, or 3.849 grams of lead.

Oxidation takes place at the ordinary temperature with many substances. Examples of such action are seen in the weathering of pyrites, rusting of iron, and (in the assay office) the weakening of solutions of many reducing agents.

For methods of determining the percentage of oxygen in gases, for technical purposes, the student is referred to Winkler & Lunge's "Technical Gas Analysis."

OXIDES.

Oxides are abundant in nature, almost all the commonly occurring bodies being oxidised. Water (H_{2}O) contains 88.8 per cent. of oxygen; silica, lime, alumina, magnesia, and the other earths are oxides, and the oxides of the heavier metals are in many cases important ores; as, for example, cassiterite (SnO_{2}), hmatite (Fe_{2}O_{3}), magnetite (Fe_{3}O_{4}), and pyrolusite (MnO_{2}). In fact, the last-named mineral owes its value to the excess of oxygen it contains, and may be regarded as an ore of oxygen rather than of manganese.

Most of the metals, when heated to redness in contact with air, lose their metallic lustre and become coated with, or (if the heating be prolonged) altogether converted into, oxide. This oxide was formerly termed a "calx," and has long been known to weigh more than the metal from which it was obtained. For example, one part by weight of tin becomes, on calcining, 1.271 parts of oxide (putty powder). The student will do well to try the following experiments:—Take 20 grams of tin and heat them in a muffle on a scorifier, scraping back the dross as it forms, and continuing the operation until the whole of the metal is burnt to a white powder and ceases to increase in weight.[95] Take care to avoid loss, and, when cold, weigh the oxide formed. The oxide should weigh 25.42 grams, which increase in weight is due to the oxygen absorbed from the air and combined with the metal. It can be calculated from this experiment (if there has been no loss) that oxide of tin contains 21.33 per cent. of oxygen and 78.67 per cent. of tin. Oxidation is performed with greater convenience by wet methods, using reagents, such as nitric acid, which contain a large proportion of oxygen loosely held. Such reagents are termed oxidising agents. Besides nitric acid, permanganate of potash, bichromate of potash, and peroxide of hydrogen are largely used for this purpose. One c.c. of nitric acid contains as much oxygen as 2.56 litres of air, and the greater part of this is available for oxidising purposes. Try the following experiment:—Take 2 grams of tin and cover in a weighed Berlin dish with 20 c.c. of dilute nitric acid, heat till decomposed, evaporate to dryness, ignite, and weigh. The 2 grams of tin should yield 2.542 grams of oxide. The increase in weight will be proportionally the same as in the previous experiment by calcination, and is due to oxygen, which in this case has been derived from the nitric acid.

The percentage of oxygen in this oxide of tin (or in any of the oxides of the heavier metals) may be directly determined by heating such oxides in a current of hydrogen, and collecting and weighing the water formed.

It is found by experiment that 88.86 parts by weight of oxygen, combining with 11.14 parts of hydrogen, form 100 parts of water; so that from the weight of water formed it is easy to calculate the amount of oxygen the oxide contained.



Take 1 gram of the dried and powdered oxide and place it in a warm dry combustion tube. Place the tube in a furnace, and connect at one end with a hydrogen apparatus provided with a sulphuric acid bulb for drying the gas, and at the other with a weighed sulphuric acid tube for collecting the water formed. The apparatus required is shown in fig. 62. Pass hydrogen through the apparatus, and, when the air has been cleared out, light the furnace. Continue the heat and current of hydrogen for half an hour (or longer, if necessary). Allow to cool. Draw a current of dry air through the weighed tube. Weigh. The increase in weight gives the amount of water formed, and this, multiplied by 0.8886, gives the weight of the oxygen. The percentage of oxygen thus determined should be compared with that got by the oxidation of the metal. It will be practically the same. The following results can be taken as examples:—

Twenty grams of tin, calcined as described, gave 25.37 grams of oxide.

Two grams of tin, oxidised with nitric acid and ignited, gave 2.551 grams of oxide.

One gram of the oxide of tin, on reduction in a current of hydrogen, gave 0.2360 gram of water (equivalent to 0.2098 gram of oxygen), and left 0.7900 gram of metal.

Ten grams of ferrous sulphate gave, on strong ignition, 2.898 grams of ferric oxide (Fe{2}O{3})[96] instead of 2.877.

The student should similarly determine the percentage of oxygen in oxides of copper and iron. The former oxide may be prepared by dissolving 5 grams of copper in 50 c.c. of dilute nitric acid, evaporating to dryness, and strongly igniting the residue. The oxide of iron may be made by weighing up 10 grams of powdered ferrous sulphate (= to 2.014 grams of iron) and heating, at first gently, to drive off the water, and then at a red heat, until completely decomposed. The weight of oxide, in each case, should be determined; and the percentage of oxygen calculated. Compare the figures arrived at with those calculated from the formula of the oxides, CuO and Fe{2}O{3}.

It would be found in a more extended series of experiments that the same metal will, under certain conditions, form two or more oxides differing among themselves in the amount of oxygen they contain. These oxides are distinguished from one another by such names as "higher" and "lower oxides," "peroxides," "protoxides," "dioxides," &c.

The oxides may be conveniently classified under three heads:—

(1) Those that are reduced to metal by heat alone, such as the oxides of mercury, silver, platinum, gold, &c.;

(2) Those which are reduced by hydrogen at a red heat, which includes the oxides of the heavy metals;

(3) Those which are not reduced by these means, good examples of which are silica, alumina, the alkalies, and the alkaline earths.

Another important classification is into acid, basic and neutral oxides. The oxides of the non-metallic elements, such as sulphur, carbon, phosphorus, &c., are, as a rule, acid; and the more oxygen they contain, the more distinctly acid they are. The oxides of the metals are nearly all basic; and, as a rule, the less oxygen they contain, the more distinctly basic they are.

The basic oxides, which are soluble in acids, give rise to the formation of salts when dissolved therein. During the solution, water is formed, but no gas is evolved. The oxide dissolved in each case neutralizes an equivalent of the acid used for solution.[97] The basic properties of many of these can be taken advantage of for their determination. This is done in the case of soda, potash, lime, &c., by finding the quantity of acid required to neutralize a given weight of the substance.

There are some oxides which, under certain conditions, are acid to one substance (a stronger base) and basic to another (a stronger acid). For example, the oxides of lead and of tin, as also alumina, dissolve in caustic soda, acting as acids; whilst, on the other hand, they combine with sulphuric or hydrochloric acid, playing the part of bases.

The oxides known as "earths," when ignited, are many of them insoluble in acids, although easily dissolved before ignition.

It is common in complete analyses of minerals to meet with cases in which the sum total of the elements found falls short of the amount of ore taken; and here oxygen must be looked for. For example, this occurs in the case of a mixture of pyrites with oxide of iron, or in a mixture of sulphides and sulphates. The state in which the elements are present, and the percentage (say of sulphides and sulphates) can in many cases be determined; but this is not always required. When the difference between the sum total and the elements found is small, it is reported as "oxygen and loss." When, however, it is considerable, the oxygen may be reported as such; and its amount be either determined directly in the way already described, or calculated from the best determination that can be made of the relative amounts of oxides, sulphides, sulphates, &c., present. Such cases require a careful qualitative analysis to find out that the substance is present; and then the separation of each constituent is made as strictly as possible. These remarks apply especially to ores of the heavy metals. The separation of the constituents is effected with suitable solvents applied in proper order. The soluble sulphates, for example, are extracted with water; the oxides by the dilute acids or alkalies in which they are known to be soluble. The oxygen in the sulphates and oxides thus obtained is estimated by determining the sulphur and metals in the solutions, and calculating the amount of oxygen with which they combine. The metals of the earths and alkalies are almost invariably present as oxides, and are reported as such; except it is known that they are present in some other form, such as fluoride or chloride. Thus, silica, alumina, lime, water, &c., appear in an analysis; even in those cases where "oxygen and loss" is also mentioned. As an example of such a report, take the following analysis of Spanish pyrites:—

Sulphur 49.00 Iron 43.55 Copper 3.20 Arsenic 0.47 Lead 0.93 Zinc 0.35 Lime 0.10 Silica, &c. 0.63 Water 0.70 Oxygen and loss 1.07 ——- 100.00

The following example will illustrate the mode of calculating and reporting. A mineral, occurring as blue crystals soluble in water, and found on testing to be a mixed sulphate of iron and copper, gave on analysis the following results:—

Water 44.51 per cent. Sulphuric oxide 28.82 " Copper 8.44 " Ferrous iron 11.81 " Ferric iron 0.38 " Zinc 0.28 " ——- 94.24

There is here a deficiency of 5.76 per cent. due to oxygen. Nothing else could be found, and it is known that in the sulphates the metals exist as oxides. By multiplying the weight of the copper by 1.252, the weight of copper oxide (CuO) will be ascertained; in this case it equals 10.57 per cent. The ferrous iron multiplied by 1.286 will give the ferrous oxide (FeO); in this case 15.19 per cent. The ferric iron multiplied by 1.428 will give the ferric oxide (Fe{2}O{3}); in this case 0.54 per cent. The zinc multiplied by 1.246 will give the zinc oxide (ZnO); in this case it equals 0.35 per cent. The analysis will be reported as—

Water 44.51 Sulphuric oxide 28.82 Copper oxide 10.57 equal to copper 8.44% Ferrous oxide 15.19 Ferric oxide 0.54 Zinc oxide 0.35 ——- 99.98

The following (A) is an analysis of a sample of South American copper ore, which will serve as a further illustration. The analysis showed the presence of 6.89 per cent. of ferrous oxide, and some oxide of copper.

The analysis (B) is that of an ore from the same mine after an imperfect roasting. It will be seen that the carbonates have been converted into sulphates. If the total sulphur simply had been determined, and the sulphate overlooked, the "oxygen and loss" would have been 5.65 per cent., an amount which would obviously require an explanation.

A. B. Water 0.25 0.59 Organic matter 0.54 — Sulphur 29.50 21.33 Copper 10.92 9.80 {Copper 9.57 {Copper oxide 0.28 Iron 32.09 39.73 {Iron 34.32 {Ferric oxide 7.73 Lead 0.35 0.12 Zinc 0.86 0.69 Cobalt 0.06 0.11 Lime 5.25 7.69 Magnesia 2.33 2.55 Sulphuric oxide 1.00 5.30 Carbon dioxide 8.87 — "Insoluble silicates" 5.12 8.38 Oxygen and loss 2.86 2.47 ——- Potash 0.15 100.00 Soda 1.09 ——- 100.00

WATER.

Water occurs in minerals in two forms, free and combined. The term "moisture" ought, strictly, to be limited to the first, although, as has already been explained, it is more convenient in assaying to apply the term to all water which is driven off on drying at 100 C. The combined water is really a part of the mineral itself, although it may be driven off at a high temperature, which varies with the base. In some cases a prolonged red heat is required; whilst with crystallised salts it is sometimes given off at the ordinary temperatures. This latter phenomenon, known as efflorescence, is mostly confined to artificial salts.