trate of tin, Sn"Cl, and Sn"(NO3)2, are derived from two molecules of hydrochloric and nitric acid, H,Cl, and H,(NO3)2, respectively; trichloride and trinitrate of bismuth, Bi"Cl, and Bi""(NO3)3, derived from three molecules of hydrochloric and nitric acid, H,Cl, and H1(NO,), respectively. But the formulæ of the salts of dibasic acids with diquivalent metals and of tribasic acids with triquivalent metals are very simple. Thus lead sulphate Pb"SO,, is derived from sulphuric acid H2SO4, and bismuth phosphate Bi""PO4 from phosphoric acid H,PO,, &c. On the other hand the salts of dibasic acids with triquivalent metals, and of tribasic acids with diquivalent metals, are highly complex. Sulphate of bismuth, for example, must be represented by the formula Bi""2(SO4)3, derived from three atoms of sulphuric acid H¿(SO4)3, and so in other instances.

(8.) It will be perceived from the above tables and examples, that a salt is usually derived from its corresponding acid by a substitution of metal for hydrogen. Many salts, however, known as salts of the alkaloids, are formed in a different way—namely, by a direct union of the acid with ammonia or some other alkaloidal base; as exemplified by hydrochloride of ammonia NH2HCI, nitrate of ammonia NH,HNO3, &c. But salts of this character may also be considered to contain a composite metal, or rather metalloid, in place of the hydrogen of the acid. Thus by associating with each atom of ammonia in the salt an atom of hydrogen from the acid, each such atom of ammonia NH ̧, becomes converted into an atom of ammonium NH, a pseudo-metallic grouping, which in its combinations presents a most marked analogy to potassium, as illustrated below:

Without assuming any knowledge of the actual molecular arrangement of ammoniacal salts, it is found most convenient in

practice, especially when comparing them with metallic salts, to regard them as salts of ammonium rather than as salts of ammonia. But the salts of the correlated alkaloids aniline, morphia, strychnia, &c., are usually represented after the manner of ammonia salts, thus:

(9.) It was formerly the custom to regard ternary acids as compounds of one or more equivalents of water with a special oxidised grouping, and the corresponding salts as compounds of one or more equivalents of metallic oxide with the same oxidised grouping. Thus the formulæ of chloric acid, and chlorate of potassium were written H2O.Cl2O, and K,O.Cl,O,, corresponding to 2HC10, and 2KC10, respectively. But this custom, which was based on assumptions since shown to be erroneous, is now falling gradually into disuse. Thus the acids of chlorine form the following series, the ternary members of which may be obtained by direct and successive oxidation of the binary member, hydrochloric acid, the hydrogen of which cannot possibly exist in the state of water:

It is true that many ternary acids and salts may be directly or indirectly formed from, or separated into, a special oxidised grouping and water or metallic oxide, but the same acids and salts may also be formed from, or separated into, a variety of other sub-compounds, and the one mode of composition or decomposition has no more right than has each of the others to set up for itself a rational formula. Thus if we act upon sulphate of

copper by metallic iron, magnesia, peroxide of barium, and sulphide of sodium respectively, we have the following reactions:

FeSO4 or Fe SO 4

+ Cu

MgSO4 or MgO . SO, + CuO
BaSO4 or BaO2. SO2 +

2

Na2SO4 or Na2S. 04 +

CuO2

CuS

From each of these reactions there might, with equal reason, be inferred the pre-existence in sulphate of copper of the respective groupings SO4, SO3, SO2, and CuS,-and the correctness of the respective rational formulæ CuSO4, CuO.SO3, CuO,.SO2, and CuS.O47 in accordance with the theories of Dulong, Berzelius, Longchamps, and Laurent respectively. Moreover, sulphate of copper may be electrolysed into Cu and SO, +0, while it may be formed by combining CuO with SO,, or CuO, with SO,, or Cus with 04 From considerations of this kind chemists have thought it better to employ, as much as possible, what are called synoptic formula, which express only the composition of acids and salts, and not their internal molecular arrangement.

(10.) The anhydrous acids assumed to pre-exist in ternary acids and salts are mostly hypothetical, but those which have an actual independent existence are found to be quite devoid of acid properties, and in their reactions upon various classes of bodies to differ greatly from the corresponding acids. Hence the appellation acid is altogether inapplicable to them, and consequently the phrase anhydrous acid is become gradually disused, and the word anhydride adopted in its stead. The only anbydrides often concerned in chemical reactions are the carbonic, silicic, stannic, sulphurous, and arsenious. With the carbonic and sulphurous anhydrides may be associated carbonic oxide and sulphuric anhydride respectively, as in the following table:

The carbonic, silicic, stannic, sulphurous, and arsenious acids are unstable ill-defined bodies, which readily break up into water and the respective anhydrides.

(11.) The general term salt is often taken to include the acid, or salt of hydrogen, as well as the salt of a true metal, such as sodium, or of a quasi-metal, such as ammonium. Using the term in this broad sense, it may be said that whenever different salts of different bases or basylides occur in solution, they undergo mutual decomposition to a greater or less extent according to circumstances. Thus when solutions of chloride of hydrogen and nitrate of sodium are mixed together in equivalent proportions, we have produced some chloride of sodium and nitrate of hydrogen together with some unaltered chloride of hydrogen and nitrate of sodium, or the two salts become four salts, thus:

xHCl + xNaNO3

=

yNaCl + HNO3 + (x − y)HCl + (x − y)NaNO3.

But whenever any one of the freshly formed salts is removed from the sphere of chemical action by precipitation or volatilisation, there is a complete instead of a partial decomposition, thus:

xHCI + xAgNO3

=

AgCl + HNO3.

Hence we arrive at the following general law. Any two salts which, by an exchange of their respective basylides, can, under the conditions of the experiment, form an insoluble or volatile compound, undergo a complete double decomposition with precipitation of the insoluble or evolution of the volatile compound. Oxalate of calcium, for instance, being insoluble in water, we know that when chloride of calcium solution is mixed with excess of oxalate of ammonium solution, the whole of the calcium will be precipitated in the form of oxalate of calcium, thus:

CaCl2 + (NH4)2C2O4

=

2NH CI + CaC204.

Again, acetic acid or acetate of hydrogen being volatile at a moderate temperature, we know that when acetate of sodium is

heated with sulphate of hydrogen a double decomposition will take place, and acetic acid be liberated, thus:

NaH3C2O2 + H2SO4

=

H1C2O2 + NaHSO4.

As will hereafter be seen, the deposition of characteristic precipitates and evolution of characteristic gases or vapours constitute the most general means by which the presence of particular bodies can be analytically established.

§ II.-CHEMICAL MANIPULATION.

(12.) The spirit-lamp is very useful for many operations, especially when a small, smokeless, not over hot flame is required. The charcoal burner is scarcely necessary in a laboratory furnished with gas, but is otherwise almost indispensable. In a chauffer of the kind figured on page 38, one or two pieces of charcoal may be kept slowly burning, by occasionally blowing off the ash; or a large brisk fire may be kept up, capable of boiling a gallon or more of water. A few pieces of charcoal may also be readily burnt on a coarse wire grating or trellis, resting by its edges on a couple of bricks or other suitable support. But gas is by far the most convenient fuel for ordinary laboratory work. Among burners which are very generally useful may be mentioned the bat's-wing-gauze burner. This consists of a large bat's-wing nipple, screwed into an elbow of brass tube standing on a flat iron foot, and provided with a gallery of some kind to support a brass chimney covered at the top with wire gauze. The bat's-wing flame burnt without the chimney is convenient for bending glass tube, and when reduced by partially turning off the gas, is well fitted for blowpipe testing (fig. 4). With the chimney on, and the ascending mixture of gas and air burnt at the top of the gauze, as shown on page 34, a large smokeless flame is obtained suitable for heating sand-baths, flasks, test-tubes, &c. An argand burner, screwed into a flat iron foot and provided with a short brass chimney, is also convenient for