Chapter XX., Note 20). Thus the forms PX_{3} and PX_{5} not only

exist in many solid and non-volatile substances, but also as vapours.

If a piece of phosphorus be dropped into a flask containing chlorine, it burns when touched with a red-hot wire, and combines with the chlorine. If the phosphorus be in excess, liquid _phosphorus trichloride_, PCl_{3}, is always formed, but if the chlorine be in excess the solid pentachloride is obtained. The trichloride is generally prepared in the following manner. Dry chlorine (passed through a series of Woulfe's bottles containing sulphuric acid) is led into a retort containing sand and phosphorus. The retort is heated, the phosphorus melts, spreads through the sand, and gradually forms the trichloride, which distils over into a receiver, where it condenses. _Phosphoric chloride_ or _phosphorus pentachloride_, PCl_{5}, is prepared by passing dry chlorine into a vessel containing phosphorus trichloride (purified by distillation). Phosphorous chloride combines directly with oxygen, but more rapidly with ozone or with the oxygen of potassium chlorate (3PCl_{3} + KClO_{3} = 3POCl_{3} + KCl), forming _phosphorus oxychloride_, POCl_{3} (Brodie). This compound is also formed by the first action of water on phosphoric chloride; for example, if two vessels, one containing phosphoric chloride and the other water, are placed under a bell jar, after a certain time the crystals of the chloride disappear and hydrochloric acid passes into the water. The aqueous vapour acts on the pentachloride, and the following reaction occurs: PCl_{5} + H_{2}O = POCl_{3} + 2HCl, the result being that liquid phosphorus oxychloride is found in one vessel, and a solution of hydrochloric acid in the other. However, an excess of water directly transforms phosphoric chloride into orthophosphoric acid, PCl_{5} + 4H_{2}O = PH_{3}O_{4} + 5HCl,[26] since POCl_{3} reacts with water (3H_{2}O), forming 3HCl and phosphoric acid PO(OH)_{3}.

[26] Phosphorus oxychloride is obtained by the action of phosphoric chloride on hydrates of acids (because alkalis decompose phosphorus oxychloride), according to the equation PCl_{5} + RHO = POCl_{3} + RCl + HCl, where RHO is an acid. The reaction only proceeds according to this equation with monobasic acids, but then RCl is volatile, and therefore a mixture is obtained of two volatile substances, the acid chloride and phosphorus oxychloride, which are sometimes difficult to separate; whilst if the hydrate be polybasic the reaction frequently proceeds so that an anhydride is formed: RH_{2}O_{2} + PCl_{5} = RO + POCl_{3} + 2HCl. If the anhydride be non-volatile (like boric), or easily decomposed (like oxalic), it is easy to obtain pure oxychloride. Thus phosphorus oxychloride is often prepared by acting on boric or oxalic acid with phosphoric chloride. It is also formed when the vapour of phosphoric chloride is passed over phosphoric anhydride, P_{2}O_{5} + 3PCl_{5} = 5POCl_{3}. This forms an excellent example in proof of the fact that the formation of one substance from two does not necessarily show that the resultant compound contains the molecules of these substances in its molecule. But other oxychlorides of phosphorus are also formed by the interaction of phosphoric anhydride and chloride; thus at 200° the chloranhydride, PO_{2}Cl, or chloranhydride of metaphosphoric acid, is formed (Gustavson). The chloranhydride of pyrophosphoric acid, P_{2}O_{3}Cl_{4}, was obtained (Hayter and Michaelis), together with NOCl, &c., by the action of NO upon cold PCl_{3}, as a fuming liquid boiling at 210°.

The above chlorine compounds serve not only as a type of the chloranhydrides, but also as a means for the preparation of other _acid chloranhydrides_. Thus the conversion of acids XHO into chloranhydrides, XCl, is generally accomplished by means of _phosphorus pentachloride_. This fact was discovered by Chancel, and adopted by Gerhardt as an important method for studying organic acids. By this means organic acids, containing, as we know, RCOOH (where R is a hydrocarbon group, and where carboxyl may repeat itself several times by replacing the hydrogen of hydrocarbon compounds), are converted into their chloranhydrides, RCOCl. With water they again form the acid, and resemble the chloranhydrides of mineral acids in their general properties.

Since carbonic acid, CO(OH)_{2}, contains two hydroxyl groups, its perfect chloranhydride, COCl_{2}, _carbonic oxychloride_, _carbonyl chloride_ or _phosgene gas_, contains two atoms of chlorine, and differs from the chloranhydrides of organic acids in that in them one atom of chlorine is replaced by the hydrocarbon radicle RCOCl, if R be a monatomic radicle giving a hydrocarbon RH. It is evident, on the one hand, that in RCOCl the hydrogen is replaced by the radicle COCl, which is also able to replace several atoms of hydrogen (for example, C_{2}H_{4}(COCl)_{2} corresponds with the bibasic succinic acid); and, on the other hand, that the reactions of the chloranhydrides of organic acids will answer to the reactions of carbonyl chloride, as the reactions of the acids themselves answer to those of carbonic acid. Carbonyl chloride is obtained directly from dry carbon monoxide and chlorine[27] exposed to the action of light, and forms a colourless gas, which easily condenses into a liquid, boiling at +8°, specific gravity 1·43, and having the suffocating odour belonging to all chloranhydrides. Like all chloranhydrides, it is immediately decomposed by water, forming carbonic anhydride, according to the equation COCl_{2} + H_{2}O = CO_{2} + 2HCl, and thus expresses the type proper to all chloranhydrides of both mineral and organic acids.[28]

[27] The direct action of the sun's rays, or of magnesium light, is necessary to start the reaction between carbonic oxide and chlorine, but when once started it will proceed rapidly in diffused light. An excess of chlorine (which gives its coloration to the colourless phosgene) aids the completion of the reaction, and may afterwards be removed by metallic antimony. Porous substances, like charcoal, aid the reaction. Phosgene may be prepared by passing a mixture of carbonic anhydride and chlorine over incandescent charcoal. Lead or silver chloride, when heated in a current of carbonic oxide, also partially form phosgene gas. Carbon tetrachloride, CCl_{4}, also forms it when heated with carbonic anhydride (at 400°), with phosphoric anhydride (200°), and most easily of all with sulphuric anhydride (2SO_{3} + CCl_{4} = COCl_{2} + S_{2}O_{5}Cl_{2}, this is pyrosulphuryl chloride). Chloroform, CHCl_{3}, is converted into carbonyl chloride when heated with SO_{2}(OH)Cl (the first chloranhydride of sulphuric acid); CHCl_{3} + SO_{3}HCl = COCl_{2} + SO_{2} + 2HCl (Dewar), and when oxidised by chromic acid.

Among the reactions of phosgene we may mention the formation of urea with ammonia, and of carbonic oxide when heated with metals.

[28] We are already acquainted with some of the chloranhydrides of the inorganic acids--for instance, BCl_{3}, and SiCl_{4}--and here we shall describe those which correspond with sulphuric acid in the following chapter. It may be mentioned here that when hydrochloric acts on nitric acid (aqua regia, Vol. I. p. 467) there is formed, besides chlorine, the oxychlorides NOCl and NO_{2}Cl, which may be regarded as chloranhydrides of nitric and nitrous acids (nitrogen chloride, Vol. I. p. 476). The former boils at -5°, the latter at +5°, the specific gravity of the first at -12° = 1·416, and at -18° = 1·433 (Geuther), and of the second = 1·3; the first is obtained from nitric oxide and chlorine, the second from nitric peroxide and chlorine, and also by the action of phosphoric chloride on nitric acid. If the gases evolved by aqua regia be passed into cold and strong sulphuric acid, they form crystals of the composition NHSO_{3} (like chamber crystals), which melt at 86°, and with sodium chloride form acid sodium sulphate and the oxychloride NOCl. This chloranhydride of nitric acid is termed _nitrosyl chloride_.

_Cyanogen chloride_, CNCl, is the gaseous chloranhydride of cyanic acid; it is formed by the action of chlorine on aqueous mercury cyanide, Hg(CN)_{2} + 2Cl_{2} = HgCl_{2} + 2CNCl. When chlorine acts on cyanic acid, it forms not only this cyanogen chloride, but also polymerides of it--a liquid, boiling at 18°, and a solid, boiling at 190°. The latter corresponds with cyanuric acid, and consequently contains C_{3}N_{3}Cl_{3}. Details concerning these substances must be looked for in works on organic chemistry.

In order to show the general method for the preparation of acid chloranhydrides, we will take that of acetic acid, CH_{3}·COOH, as an example. Phosphorus pentachloride is placed in a glass retort, and acetic acid poured over it; hydrochloric acid is then evolved, and the substance distilling over directly after is a very volatile liquid, boiling at 50°, and having all the properties of the chloranhydrides. With water it forms hydrochloric and acetic acids. The reaction here taking place may be explained thus: the substitution of the oxygen taken from the acetic acid (from its carboxyl) by two atoms of chlorine from the PCl_{5} should be as follows: CH_{3}·COOH + PCl_{5} = CH_{3}·COHCl_{2} + POCl_{3}. But the compound CH_{3}·COHCl_{2} does not exist in a free state (because it would indicate the possibility of the formation of compounds of the type CX_{6}, and carbon only gives those of the type CX_{4}); it therefore splits up into HCl and the chloranhydride CH_{3}·COCl. The general scheme for the reaction of phosphorus pentachloride with hydrates ROH is exactly the same as with water; namely, ROH with PCl_{5}, gives POCl_{3} + HCl + RCl--that is a chloranhydride.[28 bis]

[28 bis] This reaction indeed proceeds very easily and completely with a number of hydroxides, if they do not react on hydrochloric acid and phosphorus oxychloride, which is the case when they have alkaline properties. When the hydroxide is bibasic and is present in excess, it not unfrequently happens that the elements of water are taken up: R(OH)_{2} + PCl_{5} = RO + 2HCl + POCl_{3}. The anhydride RO may then be converted into chloranhydride, RO + PCl_{5} = RCl_{2} + POCl_{3}--that is, phosphorus pentachloride brings about the substitution of O by Cl_{2}. Thus carbonyl chloride, COCl_{2}, boron chloride, 2BCl_{3}, and succinic chloride, C_{4}H_{4}O_{2}Cl_{2}, &c., are respectively obtained by the action of phosphoric chloride on carbonic, boric, and succinic anhydrides. Phosphorus pentachloride reacts in a similar manner on the aldehydes, RCHO, forming RCHCl_{2}, and on the chloranhydrides themselves--for example, with acetic chloride, CH_{3}.COCl (when heated in a closed tube), it forms a substance having the composition CH_{3}CCl_{3}.

Phosphorus trichloride and oxychloride act in a similar manner to phosphoric chloride. When phosphorus trichloride acts on an acid, 3RHO + PCl_{3} = 3RCl + P(HO)_{3}. If a salt is taken, then by the action of phosphorus oxychloride a corresponding chloranhydride and salt of orthophosphoric acid are easily formed: 3R(KO) + POCl_{3} = 3RCl + PO(KO)_{3}. The chloranhydride RCl is always more volatile than its corresponding acid, and distils over before the hydrate RHO. Thus acetic acid boils at 117°, and its chloranhydride at 50°. Phosphoric and phosphorous acids are very slightly volatile, whilst their chloranhydrides are comparatively easily converted into vapour. The faculty of the chloranhydrides to react at the expense of their own chlorine determines their great importance in chemistry. For instance, suppose we require to know the molecular formula of some hydrate which does not pass into a state of vapour and does not give a chloranhydride with hydrochloric acid--that is, which has not any basic or alkaline properties; we must then endeavour to obtain this chloranhydride by means of phosphoric chloride, and it frequently happens that the corresponding chloranhydride is volatile. The resultant chloranhydride is then converted into vapour, and its composition is determined; and if we know its composition we are able to decide that of its corresponding hydrate. Thus, for example, from the formula of silicon chloride, SiCl_{4}, or of boron chloride, BCl_{3}, we can judge the composition of their corresponding hydrates, Si(HO)_{4}, B(HO)_{3}. Having obtained the chloranhydride RCl or RCl_{_n_}, it is possible by its means to obtain many other compounds of the same radicle R according to the equation MX + RCl = MCl + RX. M may be = H, K, Ag, or other metal. The reaction proceeds thus if M forms a stable compound with chlorine--for example, silver chloride, hydrochloric acid, and R, an unstable substance. Hence, a chloranhydride is frequently employed for the formation of other compounds of a given radicle; for instance, with ammonia they form amides RNH_{2}, and with salts ROK, with anhydrides R_{2}O, &c.

Containing, as they do, chlorine, which easily reacts with hydrogen, phosphorus pentachloride, trichloride, and oxychloride enter into reaction with ammonia, and give a series of amide and nitrile compounds of phosphorus. Thus, for example, when ammonia acts on the oxychloride we obtain sal-ammoniac (which is afterwards removed by water) and an orthophosphoric triamide, PO(NH_{2})_{3}, as a white insoluble powder on which dilute acids and alkalis do not act, but which, when fused with potassium hydroxide, gives potassium phosphate and ammonia like other amides. When ignited, the triamide liberates ammonia and forms the nitrile PON, just as urea, CO(NH_{2})_{2}, gives off ammonia and forms the nitrile CONH. This nitrile, called _monophosphamide_, PON, naturally corresponds with metaphosphoric acid, namely, with its ammonium salt. NH_{4}PO_{3} - H_{2}O = PO_{2}·NH_{2}, an as yet unknown amide, and PO_{2}·NH_{2} - H_{2}O gives the nitrile PON. This relation is confirmed by the fact that PON, moistened with water, gives metaphosphoric acid when ignited. It is the analogue of nitrous oxide, NON. It is a very stable compound, more so than the preceding.[29]

[29] The reaction of ammonia on phosphorus pentachloride is more complex than the preceding. This is readily understood: to the oxychloride, POCl_{3}, tere corresponds a hydrate PO(OH)_{3}, and a salt PO(NH_{4}O)_{3}, and consequently also an amide PO(NH_{2})_{3}, whilst the pentachloride, PCl_{5}, has no corresponding hydrate P(OH)_{5}, and therefore there is no amide P(NH_{2})_{5}. The reaction with ammonia will be of two kinds: either instead of 5 mol. NH_{3}, only 3 mol. NH_{3} or still less will act; _i.e._ PCl_{2}(NH_{2})_{3}, PCl_{3}(NH_{2})_{2}, &c. are formed; or else the pentachloride will act like a mixture of chlorine with the trichloride, and then as the result there will be obtained the products of the action of chlorine on those amides which are formed from phosphorus trichloride and ammonia. It would appear that both kinds of reaction proceed simultaneously, but both kinds of products are unstable, at all events complex, and in the result there is obtained a mixture containing sal-ammoniac, &c. The products of the first kind should react with water, and we should obtain, for example, PCl_{3}(NH_{2})_{2} + 2H_{2}O = 3HCl and PO(HO)(NH_{2})_{2}. This substance has not actually been obtained, but the compound PONH(NH_{2}) derived from it by elimination of the elements of water is known, and is termed _diphosphamide_; it is, however, more probable that it is a nitrile than an amide, because only amides contain the group NH_{2}. It is a colourless, stable, insoluble powder, which possibly corresponds with pyrophosphoric acid, more especially since when heated it evolves ammonia and gives and leaves phosphoryl nitride, PON--that is, the nitrile of metaphosphoric acid. The amide corresponding with the pyrophosphate P_{2}O_{3}(NH_{4}O)_{4} should be P_{2}O_{3}(NH_{2})_{4}, and the nitriles corresponding to the latter would be P_{2}O_{2}N(NH_{2})_{3}, P_{2}ON_{2}(NH_{2})_{2}, and P_{2}N_{3}(NH_{2}). The composition of the first is the same as that of the above diphosphamide. The third pyrophosphoric nitrile has a formula P_{2}N_{4}H_{2}, and this is the composition of the body known as _phospham_, PHN_{2} (in a certain sense this is the analogue of N_{3}H polymerised, Chapter VI.) Indeed, phospham has been obtained by heating the products of the action of ammonia on phosphoric chloride, as an insoluble and alkaline powder, which gives ammonia and phosphoric acid when subjected to the action of water. The same substance is obtained by the action of ammonium chloride on phosphoric chloride (PNCl_{2} is first formed, and reacts further with ammonia, forming phospham), and by igniting the mass which is formed by the action of ammonia on phosphorus trichloride. Formerly the composition of phospham was supposed to be PHN_{2}, now there is reason to think that its molecular weight is P_{3}H_{3}N_{6}.

The above compounds correspond with normal salts, but nitriles and amides corresponding to acid salts are also possible, and they will be acids. For example, the amide PO(HO)_{2}(NH_{2}), and its nitrile, will be either PN(HO)_{2} or PO(HO)(NH), but at all events of the composition PNH_{2}O_{2}, and having acid properties. The ammonium salt of this _phosphonitrilic acid_ (it is called phosphamic acid), PNH(NH_{4})O_{2}, is obtained by the action of ammonia on phosphoric anhydride, P_{2}O_{5} + 4NH_3 = H_{2}O + 2PNH(NH_{4})O_{2}. A non-crystalline soluble mass is thus formed, which is dissolved in a dilute solution of ammonia and precipitated with barium chloride, and the resultant barium salt is then decomposed with sulphuric acid, and thus a solution of the acid of the above composition is obtained.

It is evident from the theory of the formation of amides and nitriles (Chapter IX.) that very many compounds of this kind can correspond with the acids of phosphorus; but as yet only a few are known. The easy transitions of the ortho-, meta-, and pyrophosphoric acids, by means of the hydrogen of ammonia, into the lower acids, and conversely, tend to complicate the study of this very large class of compounds, and it is rarely that the nature of a product thus obtained can be judged from its composition; and this all the more that instances of isomerism and polymerism, of mixture between water of crystallisation and of constitution, &c., are here possible. Many data are yet needed to enable us to form a true judgment as to the composition and structure of such compounds. As the best proof of this we will describe the very interesting and most fully investigated compound of this class, PNCl_{2}, called _chlorophosphamide_, or nitrogen chlorophosphorite. It is formed in small quantities when the vapour of phosphoric chloride is passed over ignited sal-ammoniac. Besson (1892) heated the compound PCl_{5}8NH_{3} (which is easily and directly formed from PCl_{5} and NH_{3}) under a pressure of about 50 mm. (of mercury) to 200°, and obtained brilliant crystals of PNCl_{2}, which melted at 106° (in the residue after the distillation of sal-ammoniacal phospham). The chlorine in it is very stable--quite different from that in phosphoric chloride. Indeed, the resultant substance is not only insoluble in water (though soluble in alcohol and ether), but it is not even moistened by it, and distils over, together with steam, without being decomposed. In a free state it easily crystallises in colourless prisms, fuses at 114°, boils at 250° (Gladstone, Wichelhaus), and when fused with potash gives potassium chloride and the amidonitrile of phosphoric acid. Judging from its formula and the simplicity of its composition and reactions, it might be thought that the molecular weight of this substance would be expressed by the formula PCl_{2}N, that it corresponds with PON and with PCl_{5} (like POCl_3), with the substitution of Cl_3 by N, just as in POCl_3 two atoms of chlorine are replaced by oxygen; but all these surmises are incorrect, because its vapour density (referred to hydrogen--Gladstone, Wichelhaus) = 182--that is, the molecular formula must be three times greater, P_{3}N_{3}Cl_{6}. The polymerisation (tripling) is here of exactly the same kind as with the nitriles.

The most important analogue of phosphorus is _arsenic_, the metallic aspect of which and the general character of its compounds of the types AsX_{3} and AsX_{5} at once recall the metals. The hydrate of its highest oxide, arsenic acid (ortho-arsenic acid), H_{3}AsO_{4}, is an oxidising agent, and gives up a portion of its oxygen to many other substances; but, nevertheless, it is very like phosphoric acid. Mitscherlich established the conception of isomorphism by comparing the salts of these acids.[30]

[30] It is necessary to remark that, although arsenic is so closely analogous to phosphorus (especially in the higher forms of combination, RX_{3} and RX_{5}), at the same time it exhibits a certain resemblance and even isomorphism with the corresponding compounds of sulphur (especially the metallic compounds of the type MAs, corresponding with MS). Thus compounds containing metals, arsenic, and sulphur are very frequently met with in nature. Sometimes the relative amounts of arsenic and sulphur vary, so that an isomorphous substitution between the arsenides and sulphides must be recognised. Besides FeS_{2} (ordinary pyrites), and FeAs_{2}, iron forms an arsenical pyrites containing both sulphur and arsenic, which from its composition, FeAsS or FeS_{2}FeAs_{2}, resembles the two preceding.

Arsenic occurs _in nature_, not only combined with metals, but also, although rarely, native and also in combination with sulphur in two minerals--one red, _realgar_, As_{2}S_{2}, and the other yellow, _orpiment_, As_{2}S_{3} (Chapter XX., Note 29). Arsenic occurs, but more rarely, in the form of salts of arsenic acid--for instance, the so-called cobalt and nickel blooms, two minerals which are found accompanying other cobalt ores, are the arsenates of these metals. Arsenic is also found in certain clays (ochres) and has been discovered in small quantities in some mineral springs, but it is in general of rarer occurrence in nature than phosphorus. Arsenic is most frequently extracted from arsenical pyrites, FeSAs, which, when roasted without access of air, evolves the vapour of arsenic, ferrous sulphide being left behind. It is also obtained by heating arsenious anhydride with charcoal, in which case carbonic oxide is evolved. In general, the oxides and other compounds are very easily reduced. Solid _arsenic_ is a steel-grey brittle _metal_, having a bright lustre and scaly structure. Its specific gravity is 5·7. It is opaque and infusible, but volatilises as a yellow vapour which on cooling deposits rhombohedral crystals.[30 bis] The vapour density of arsenic is 150 times greater than that of hydrogen--that is, its molecule, like that of phosphorus, contains 4 atoms, As_{4}. When heated in the air, arsenic easily oxidises into white arsenious anhydride, As_{2}O_{3}, but even at the ordinary temperature it loses its lustre (becomes dull), owing to the formation of a coating of a lower oxide. The latter appears to be as volatile as arsenious anhydride, and it is probable that it is owing to the presence of this compound that the vapours of arsenious compounds, when heated with charcoal (for example, in the reducing flame of a blow-pipe), have the characteristic smell of garlic, because the vapour of arsenic itself has not this odour.

[30 bis] According to Retgers (1893) the arsenic mirror (see further on) is an unstable variety of metallic arsenic, whilst the brown product which is formed together with it in Marsh's apparatus is a lower hydride AsH. Schuller and McLeod (1894), however, recognise a peculiar yellow variety of arsenic.

Arsenic easily combines with bromine and chlorine;[31] nitric acid and aqua regia also oxidise it into the higher oxide, or rather its hydrate, arsenic acid.[32] As far as is known, it does not decompose steam, and it acts exceedingly slowly on those acids, like hydrochloric, which are not capable of oxidising.

[31] Hydrochloric acid dissolves arsenious anhydride in considerable quantities, and this is probably owing to the formation of unstable compounds in which the arsenious anhydride plays the part of a base. A compound called _arsenious oxychloride_, having the composition AsOCl, is even known. It is formed when arsenious anhydride is added little by little to boiling arsenic trichloride, As_{2}O_{3} + AsCl_{3} = 3AsOCl. It is a transparent substance, which fumes in air, and combines with water to form a crystalline mass having the composition As_{2}(OH)_{4}Cl_{2}. When heated it decomposes into arsenious chloride and a fresh oxychloride of a more complex composition, As_{6}O_{8}Cl_{2}· Arsenic trichloride, when treated with a small quantity of water, forms the crystalline compound, As_{2}(HO)_{4}Cl_{2}, mentioned above. These compounds resemble the basic salts of bismuth and aluminium. The existence of these compounds shows that arsenic is of a more metallic or basic character than phosphorus. Nevertheless _arsenic trichloride_, AsCl_{3}, resembles phosphorus trichloride in many respects. It is obtained by the direct action of chlorine on arsenic, or by distilling a mixture of common salt, sulphuric acid, and arsenious anhydride. The latter mode of preparation already indicates the basic properties of the oxide. Arsenious chloride is a colourless oily liquid, boiling at 130°, and having a sp. gr. of 2·20. It fumes in air like other chloranhydrides, but it is much more slowly and imperfectly decomposed by water than phosphorus trichloride. A considerable quantity of water is required for its complete decomposition into hydrochloric acid and arsenious anhydride. It forms an excellent example of the transition from true metallic chlorides to true chloranhydrides of the acids. It hardly combines with chlorine, _i.e._ if AsCl_{5} is formed it is very unstable. _Arsenic tribromide_, AsBr_{3}, is formed as a crystalline substance, fusing at 20° and boiling at 220°, by the direct action of metallic arsenic on a solution of bromine in carbon bisulphide, the latter being then evaporated. The specific gravity of arsenic tribromide is 3·36. Crystalline arsenic tri-iodide, AsI_{3}, having a sp. gr. 4·39, may be obtained in a like manner; it may be dissolved in water, and on evaporation separates out from the solution in an anhydrous state--that is, it is not decomposed--and consequently behaves like metallic salts. _Arsenic trifluoride_, AsF_{3}, is obtained by heating fluor spar and arsenious anhydride with sulphuric acid. It is a fuming, colourless, and very poisonous liquid, which boils at 63° and has a sp. gr. of 2·73. It is decomposed by water. It is very remarkable that fluorine forms a pentafluoride of arsenic also, although this compound has not yet been obtained in a separate state, but only in combination with potassium fluoride. This compound, K_{3}AsF_{8}, is formed as prismatic crystals when potassium arsenate, K_{3}AsO_{4}, is dissolved in hydrofluoric acid.

[32] _Arsenic acid_, H_{3}AsO_{4}, corresponding with orthophosphoric acid, is formed by oxidising arsenious anhydride with nitric acid, and evaporating the resultant solution until it attains a sp. gr. of 2·2; on cooling it separates in crystals having the above composition. This hydrate corresponds with the normal salts of arsenic acid; but on dissolving in water (without heating), and on cooling a strong solution, crystals containing a greater amount of water, namely, (AsH_{3}O_{4})_{2},H_{2}O, separate. This water, like water of crystallisation, is very easily expelled at 100°. At 120° crystals having a composition identical with that of pyrophosphoric acid, As_{2}H_{4}O_{7}, separate, but water, on dissolving this hydrate with the development of heat, forms a solution in no way differing from a solution of ordinary arsenic acid, so that it is not an independent pyroarsenic acid that is formed. Neither is there any true analogue of metaphosphoric acid, although the compound AsHO_{3} is formed at 200°, and on solidifying forms a mass having a pearly lustre and sparingly soluble in cold water; but on coming into contact with warm water it becomes very hot, and gives ordinary orthoarsenic acid in solution. Arsenic acid forms three series of salts, which are perfectly analogous to the three series of orthophosphates. Thus the normal salt, K_{3}AsO_{4}, is formed by fusing the other potassium arsenates with potassium carbonate; it is soluble in water and crystallises in needles which do not contain water. Di-potassium arsenate, K_{2}HAsO_{4}, is formed in solution by mixing potassium carbonate and arsenic acid until carbonic anhydride ceases to be evolved; it does not crystallise, and has an alkaline reaction; hence it corresponds perfectly with the sodium phosphate. As was mentioned above, arsenic acid itself acts as an oxidising agent; for example, it is used in the manufacture of aniline dyes for oxidising the aniline, and it is prepared in large quantities for this purpose. When sulphuretted hydrogen is passed through its solution, sulphuric acid and arsenious anhydride are obtained in solution. Arsenic acid is very easily soluble in water, and its solution has an exceedingly acid reaction, and when boiled with hydrochloric acid evolves chlorine, like selenic, chromic, manganic, and certain other higher metallic acids.

_Arsenic anhydride_, As_{2}O_{5}, is produced when arsenic acid is heated to redness. It must be carefully heated, as at a bright red heat it decomposes into oxygen and arsenious anhydride. Arsenic anhydride is an amorphous substance almost entirely insoluble in water, but it attracts moisture from the air, deliquesces, and passes into the acid. Hot water produces this transformation with great ease.

_Arseniuretted hydrogen_, _arsine_, AsH_{3}, resembles phosphuretted hydrogen in many respects. This colourless gas, which liquefies into a mobile liquid at -40°, has a disagreeable garlic-like odour, is only slightly soluble in water, and is exceedingly poisonous. Even in a small quantity it causes great suffering, and if present to any considerable amount in air it even causes death. The other compounds of arsenic are also poisonous, with the exception of the insoluble sulphur compound and some compounds of arsenic acid. Arseniuretted hydrogen, AsH_{3}, is obtained by the action of water on the alloy of arsenic and sodium, sodium hydroxide and arseniuretted hydrogen being formed. It is also formed by the action of sulphuric acid on the alloy of arsenic and zinc: Zn_{3}As_{2} + 3H_{2}SO_{4} = 2AsH_{3} + 3ZnSO_{4}.[33] The oxygen compounds of arsenic are very easily reduced by the action of hydrogen at the moment of its evolution from acids, and the reduced arsenic then combines with the hydrogen; hence, if a certain amount of an oxygen compound of arsenic be put into an apparatus containing zinc and sulphuric acid (and thus serving for the evolution of hydrogen), the hydrogen evolved will contain arseniuretted hydrogen. In this case it is diluted with a considerable amount of hydrogen. But its presence in the most minute quantities may be easily recognised from the fact that it is _easily decomposed_ by heat (200° according to Brunn) into metallic arsenic and hydrogen, and therefore if such impure hydrogen he passed through a moderately-heated tube metallic arsenic will be deposited as a bright layer on the part of the tube which was heated (_see_ Note 30 bis). This reaction is so sensitive that it enables the most minute traces of arsenic to be discovered; hence it is employed in medical jurisprudence, as a test in poisoning cases. It is easy to discover the presence of arsenic in common zinc, copper, sulphuric and hydrochloric acids, _&c._ by this method. It is obvious that in testing for poison by Marsh's apparatus it is necessary to take zinc and sulphuric acid quite free from arsenic. The arsenic deposited in the tube may be driven as a volatile metal from one place to another in the current of hydrogen evolved, owing to its volatility. This forms a distinction between arseniuretted and antimoniuretted hydrogen, which is decomposed by heat in just the same way as arseniuretted hydrogen, but the mirror given by Sb is not so volatile as that formed by As.

[33] The formation of arseniuretted hydrogen is accompanied by the absorption of 37,000 heat units, while phosphine evolves 18,000 (Ogier), and ammonia 27,000. Sodium (0·6 p.c.) amalgam, with a strong solution of As_{2}O_{3}, gives a gas containing 86 vols. of arsenic and 14 vols. of hydrogen (Cavazzi).

If hydrogen contains arseniuretted hydrogen, it also gives metallic arsenic when it burns, because in the reducing flame of hydrogen the oxygen attracted combines entirely with the hydrogen and not with the arsenic, so that if a cold object, such as a piece of china, be held in the hydrogen flame the arsenic will be deposited upon it as a metallic spot.[34]

[34] This spot, or the metallic ring which is deposited on the heated tube, may easily be tested as to whether it is really due to arsenic or proceeds from some other substance reduced in the hydrogen flame--for instance, carbon or antimony. The necessity for distinguishing arsenic from antimony is all the more frequently encountered in medical jurisprudence, from the fact that preparations of antimony are very frequently used as medicine, and antimony behaves in the hydrogen apparatus just like arsenic, and therefore in making an investigation for poisoning by arsenic it is easy to mistake it for antimony. The best method to distinguish between the metallic spots of arsenic and antimony is to test them with a solution of sodium hypochlorite, free from chlorine, because this will dissolve arsenic and not antimony. Such a solution is easily obtained by the double decomposition of solutions of sodium carbonate and bleaching powder. A solution of potassium chlorate acts in the same manner, only more slowly. Further particulars must be looked for in analytical works.

Arseniuretted hydrogen, like phosphuretted hydrogen, is only slightly soluble in water, has no alkaline properties--that is, it does not combine with acids--and acts as a reducing agent. When passed into a solution of silver nitrate it gives a blackish brown precipitate of metallic silver, the arsenic being oxidised. If acting on copper sulphate and similar salts, arseniuretted hydrogen sometimes forms arsenides--_i.e._ it reduces the metallic salt with its hydrogen, and is itself reduced to arsenic. Sulphuric, and even hydrochloric, acid reduces arseniuretted hydrogen to arsenic, and it is still more easily decomposed by arsenious chloride, and with phosphorous chloride it gives the compound PAs. Arseniuretted hydrogen gives metallic arsenic with an acid solution of arsenious anhydride (Tivoli).

The most common compound of arsenic is the solid and volatile _arsenious anhydride_, As_{2}O_{3}, which corresponds with phosphorous and nitrous anhydrides. This very poisonous, colourless, and sweet-tasting substance is generally known under the name of arsenic, or _white arsenic_. The corresponding hydrate is as yet unknown; its solutions, when evaporated, yield crystals of arsenious anhydride. It is chiefly prepared for the dyer, and is also used as a vermin killer, and sometimes in medicine; it is a product from which all other compounds of arsenic can be prepared. It is obtained as a by-product in roasting cobalt and other ores containing arsenic. Arsenical pyrites are sometimes purposely roasted for the extraction of arsenious anhydride. When arsenical ores are burnt in the air, the sulphur and arsenic are converted into the oxides As_{2}O_{3} and SO_{2}. The former is a solid at the ordinary temperature, and the latter gaseous, and therefore the arsenious anhydride is deposited as a sublimate in the cooler portion of the flues through which the vapours escape from the furnace. It collects in condensing chambers especially constructed in the flues. The deposit is collected, and after being distilled gives arsenious anhydride in the form of a vitreous non-crystalline mass. This is one of the varieties of arsenious anhydride, which is also known in two crystalline forms. When sublimed--_i.e._ when it rapidly passes from the state of vapour to the solid state--it appears in the regular system in the form of octahedra.[35] It is obtained in the same form when it is crystallised from acid solutions. The specific gravity of the crystals is 3·7. The other crystalline form (in prisms) belongs to the rhombohedral system, and is also formed by sublimation when the crystals are deposited on a heated surface, or when it is crystallised from alkaline solutions.[36]

[35] According to Mitscherlich's determination, the vapour density of arsenious anhydride is 199 (H = 1)--that is, it answers to the molecular formula As_{4}O_{6}. Probably this is connected with the fact that the molecule of free arsenic contains As_{4}. V. Meyer and Biltz, however, showed (1889) that at a temperature of about 1,700° the vapour density of arsenic corresponds with the molecule As_{2}, and not As_{4}, as at lower temperatures.

[36] Arsenious anhydride is obtained in an amorphous form after prolonged heating at a temperature near to that at which it volatilises, or, better still, by heating it in a closed vessel. It then fuses to a colourless liquid, which on cooling forms a transparent vitreous mass, whose specific gravity is only slightly less than that of the crystalline anhydride. On cooling, this vitreous mass undergoes an internal change, in which it crystallises and becomes opaque, and acquires the appearance of porcelain. The following difference between the vitreous and opaque varieties is very remarkable: when the vitreous variety is dissolved in strong and hot hydrochloric acid it gives crystals of the anhydride on cooling, and this crystallisation _is accompanied by the emission of light_ (which is visible in the dark), and the entire liquid glows as the crystals begin to separate. The opaque variety does not emit light when the crystals separate from its hydrochloric acid solution. It is also remarkable that the vitreous variety passes into the opaque form when it is pounded--that is, under the action of a series of blows. Thus, several varieties of arsenious anhydride are known, but as yet they are not characterised by any special chemical distinctions, and even differ but little in their specific gravities, so that it cannot be said that the above differences are due to any isomeric transformation--that is, to an arrangement of the atoms in the molecule--but probably only depend on a difference in the distribution of the molecules, or, in other terms, are physical and not chemical variations. One part of the vitreous anhydride requires twelve parts of boiling water for its solution, or twenty-five parts at the ordinary temperature. The opaque variety is less soluble, and at the ordinary temperature requires about seventy parts of water for its solution.

Solutions of arsenious anhydride have a sweet metallic taste, and give _a feeble acid reaction_. Its solubility increases with the admixture of acids and alkalis. This shows the property of arsenious anhydride of forming salts with acids and alkalis. And in fact compounds of it with hydrochloric acid (Note 31), sulphuric anhydride (_see_ further on), and with the alkali oxides are known.[37] If silver nitrate be added to a solution of arsenious anhydride, it does not give any precipitate unless a certain amount of the arsenious anhydride is saturated with an alkali--for instance, ammonia. It then gives a precipitate of silver arsenite, Ag_{3}AsO_{3}. This is yellow, soluble in an excess of ammonia, and anhydrous; it distinctly shows that arsenious acid is tribasic, and that it differs in this respect from phosphorous acid, in which only two atoms of hydrogen can be replaced by metals.[38] The feeble acid character of arsenious anhydride is confirmed by the formation of saline compounds with acids. In this respect the most remarkable example is the anhydrous compound with sulphuric acid, having the composition As_{2}O_{3},SO_{3}. It is formed in the roasting of arsenical pyrites in those spaces where the arsenious anhydride condenses, a portion of the sulphurous anhydride being converted into sulphuric anhydride, SO_{3}, at the expense of the oxygen of the air. The compound in question forms colourless tabular crystals, which are decomposed by water with formation of sulphuric acid and arsenious anhydride.[39]

[37] Arsenious anhydride does not oxidise in air, either in a dry state or in solution, but in the presence of alkalis it absorbs oxygen from the air, and acts as an excellent reducing agent. This probably is connected with the fact that arsenic acid is much more energetic than arsenious acid, and that it is arsenic acid which is formed by the oxidation of the latter in the presence of alkalis. Arsenious anhydride is easily reduced to arsenic by many metals, even by copper.

[38] The feebleness of the acid properties of arsenious anhydride is seen in the fact that if it be dissolved in ammonia water, and then a still stronger solution of ammonia be added, prismatic crystals separate having the composition of ammonium metarsenite, NH_{4}AsO_{3}. This ammonium salt deliquesces in air, and loses all its ammonia. The magnesium salt is tri-metallic, Mg_{3}(AsO_{3})_{2}; it is insoluble in water, and is formed by mixing an ammoniacal solution of arsenious anhydride with an ammoniacal solution of a magnesium salt. It is insoluble even in ammonia, although it dissolves in an excess of acids. Magnesium hydroxide gives the same salt with arsenious solutions, and hence magnesia is one of the best antidotes for arsenic poisoning. _The arsenites of copper_ are much used in the manufacture of colours, more especially of pigments. They are distinguished by their insolubility in water and by their remarkably vivid green colour, but at the same time by their poisonous character. Not only do such pigments applied to wall papers or other materials easily dust off from them, but they give exhalations containing AsH_{3}. The cupric salts, CuX_{2}, when mixed with an alkaline solution of arsenious acid, give a green precipitate of a copper salt called _Scheele's green_. Its composition is probably CuHAsO_{3}. Ammonia dissolves it, and gives a colourless solution, containing cuprous arsenate--that is, the cupric compound is reduced and the arsenic subjected to a further oxidation. The so-called _Schweinfurt green_ was still more used, especially in former times; it is an insoluble green cupric salt, which resembles the preceding in many respects, but has a different tint. It is prepared by mixing boiling solutions of arsenious acid and cupric acetate. Arsenious acid forms an insoluble compound with ferric hydroxide, resembling the phosphate; and this is the reason why freshly precipitated oxide of iron is employed as an _antidote for arsenic_. The freshly precipitated oxide of iron, taken immediately after poisoning by arsenic, converts the arsenious acid into an insoluble state, by forming a compound on which the acids of the stomach have no action, so that the poisoning cannot proceed. It is remarkable that the inhabitants of certain mountainous countries accustom themselves to taking arsenic, as a means which, according to their experience, helps to overcome the fatigue of mountain ascents. Arsenious anhydride and certain of its salts are also used in medicine, naturally only in small quantities. When taken internally arsenic passes into the blood, and is mainly excreted by the urine.

[39] Adie (1889) obtained compounds of As_{2}O_{3} with 1, 2, 4, and 8 SO_{3} by the direct action of ordinary and Nordhausen sulphuric acid upon As_{2}O_{3}. Weber had previously obtained As_{2}O_{3}SO_{3} (which disengages SO_{3} at 225°), and also other As_{2}O_{3}_n_SO_{3} (where _n_ = 3, 6, and 8), by the action of the vapours of SO_{3} upon As_{2}O_{3} at a definite temperature. The compound As_{2}O_{3},8SO_{3} loses SO_{3} at 100°. Oxide of antimony, Sb_{2}O_{3}, gives similar compounds. Adie (1891) also obtained (by the action of SO_{3} upon H_{3}PO_{4}) a compound H_{3}PO_{4}3SO_{3} in the form of a viscous liquid decomposed by water.

_Antimony_ (stibium), Sb = 120, is another analogue of phosphorus. In its external appearance and the properties of its compounds it resembles the metals still more closely than arsenic. In fact, antimony has the appearance, lustre, and many of the characteristic properties of the metals. Its oxide, Sb_{2}O_{3}, exhibits the earthy appearance of rust or of lime, and has distinctly basic properties, although it corresponds with nitrous and phosphorous anhydride, and is able, like them, to give saline compounds with bases. At the same time antimony presents, in the majority of its compounds, an entire analogy with phosphorus and arsenic. Its compounds belong to the type SbX_{3} and SbX_{5}. It is found in nature chiefly in the form of sulphide, Sb_{2}S_{3}. This substance sometimes occurs in large masses in mineral veins and is known in mineralogy under the name of antimony glance or _stibnite_, and commercially as _antimony_ (Chapter XX., Note 29). The most abundant deposits of antimony ore occur in Portugal (near Oporto on the Douro). Besides which antimony partially or totally replaces arsenic in some minerals; thus, for example, a compound of antimony sulphide and arsenic sulphide with silver sulphide is found in red silver ore. But in every case antimony is a rather rare metal found in few localities. In Russia it is known to occur in Daghestan in the Caucasus. It is extracted chiefly for the preparation of alloys with lead and tin, which are used for casting printing type.[40] Some of its compounds are also used in medicine, the most important in this respect being antimony pentasulphide, Sb_{2}S_{5} (_sulfur auratum antimonii_), and tartar emetic, which is a double salt derived from tartaric acid and has the composition C_{4}H_{4}K(SbO)O_{6}. Even the native antimony sulphide is used in large quantities as a purgative for horses and dogs. Metallic antimony is extracted from the glance, Sb_{2}S_{2}, by roasting, when the sulphur burns away and the antimony oxidises, forming the oxide Sb_{2}O_{3}, which is then heated with charcoal, and thus reduced to a _metallic state_. The reduction may be carried on in the laboratory on a small scale by fusing the sulphide with iron which takes up the sulphur.[40 bis]

[40] Printers' type consists of an alloy known as 'type-metal,' containing usually about 15 parts of antimony to 85 parts of lead; sometimes (for example, for stereotypes) from 10 to 15 per cent. Bi or 8 per cent. Sn and even Cu is added. The hardness of the alloy, which is essential for printing, evidently depends upon the presence of antimony, but an excess must be avoided, since this renders the alloy brittle, and the type after a time loses its sharpness.

[40 bis] Antimony is prepared in a state of greater purity by heating with charcoal the oxide obtained by the action of nitric acid on the impure commercial metallic antimony. This is based on the fact that by the action of the acid, antimony forms the oxide Sb_{2}O_{3}, which is but slightly soluble in water. The arsenic, which is nearly always present, forms soluble arsenious and arsenic acids, and remains in solution. The purest antimony is easily obtained from tartar emetic, by heating it with a small quantity of nitre. Metallic antimony also occurs, although rarely, native; and as it is very easily obtained, it was known to the alchemists of the fifteenth century. Very pure metallic antimony may be deposited by the electric current from a solution of antimonious sulphide in sodium sulphide after the addition of sodium chloride to the solution.

Metallic antimony has a white colour and a brilliant lustre; it remains untarnished in the air, for the metal does not oxidise at the ordinary temperature. It crystallises in rhombohedra, and always shows a distinctly crystalline structure which gives it quite a different aspect from the majority of the metals yet known. It is most like tellurium in this respect. Antimony is brittle, so that it is very easily powdered; its specific gravity is 6·7, it melts at about 432°, but only volatilises at a bright red heat. When heated in the air--for instance, before the blow-pipe--it burns and gives white odourless fumes, consisting of the oxide. This oxide is termed antimonious oxide, although it might as well be termed antimonious anhydride. It is given the first name because in the majority of cases its compounds with acids are used, but it forms compounds with the alkalis just as easily.

Antimonious oxide, like arsenious anhydride, crystallises either in regular octahedra or in rhombic prisms; its specific gravity is 5·56; when heated it becomes yellow and then fuses, and when further heated in air it oxidises, forming an oxide of the composition Sb_{2}O_{4}. Antimonious oxide is insoluble in water and in nitric acid, but it easily dissolves in strong hydrochloric acid and in alkalis, as well as in tartaric acid or solutions of its acid salts. When dissolved in the latter it forms tartar emetic. It is precipitated from its solutions in alkalis and acids (by the action of acids on the former and alkalis on the latter). It occurs native but rarely. As a base it gives salts of the type SbOX (as if the basic salts = SbX_{3}, Sb_{2}O_{3}) and hardly ever forms salts, SbX_{3}. In the antimonyl salts, SbOX, the group SbO is univalent, like potassium or silver. The oxide itself is (SbO)_{2}O, the hydroxide, SbO(OH), &c.; tartar emetic is a salt in which one hydrogen of tartaric acid is replaced by potassium and the other by antimonyl, SbO. Antimonious oxide is very easily separated from its salts by any base, but it must be observed that this separation does not take place in the presence of tartaric acid, owing to the property of tartaric acid of forming a soluble double salt--_i.e._ tartar emetic.[41]

[41] As antimonious oxide answers to the type SbX_{3}, it is evident that compounds may exist in which antimony will replace three atoms of hydrogen; such compounds have been to some extent obtained, but they are easily converted by water into substances corresponding with the ordinary formulæ of the compounds of antimony. Thus tartar emetic, C_{4}H_{4}(SbO)KO_{6}, loses water when heated, and forms C_{4}H_{2}SbKO_{6}--that is, tartaric acid, C_{2}H_{6}O_{6}, in which one atom of hydrogen is replaced by potassium and three by antimony. But this substance is reconverted into tartar emetic by the action of water.

A similar compound is seen in that _intermediate oxide of antimony_ which is formed when antimonious oxide is heated in air: its composition is SbO_{2} or Sb_{2}O_{4}. This oxide may be regarded as orthantimonic acid, SbO(HO)_{3}, in which three atoms of hydrogen are replaced by antimony in that state in which it occurs in oxide of antimony--_i.e._ SbO(SbO_{3}) = Sb_{2}O_{4}. Oxide of antimony is also formed when antimonic acid is ignited; it then loses water and oxygen, and gives this intermediate oxide as a white infusible powder, of sp. gr. 6·7. It is somewhat soluble in water, and gives a solution which turns litmus paper red.

If metallic antimony, or antimonious oxide, be oxidised by an excess of nitric acid and the resultant mass be carefully evaporated to dryness, _metantimonic acid_, SbHO_{3}, is formed. Its corresponding potassium salt, 2SbKO_{3},5H_{2}O, is prepared by fusing metallic antimony with one-fourth its weight of nitre and washing the resultant mass with cold water. This potassium salt is only slightly soluble in water (in 50 parts) and the sodium salt is still less so. An ortho-acid, SbH_{3}O_{4}, also appears to exist;[41 bis] it is obtained by the action of water on antimony pentachloride, but it is very unstable, like the pentachloride, SbCl_{5}, itself, which easily gives up Cl_{2}, leaving antimony trichloride, SbCl_{3}, and this is decomposed by water, forming an oxychloride--SbOCl, only slightly soluble in water. When antimonic acid is heated to an incipient red heat, it parts with water and forms the anhydride, Sb_{2}O_{5}, of a yellow colour and specific gravity 6·5.[42]

[41 bis] Beilstein and Blaese (1889), after preparing many salts of antimonic acid, came to the conclusion that it is monobasic, but all the salts still contain water, so that their general type is mostly: MSbO_{3}3H_{2}O, for example, M = Li, Hg (salts of the suboxide), 1/2 Pb, &c. The type of the ortho-salts, M_{2}SbO_{4}, is quite unknown, although it is reproduced in the thio-compounds, for instance, Schlippe's salt, Na_{2}SbS_{4}, but this salt also contains water of crystallisation, 9H_{2}O (Chapter XX., Note 29).

[42] Among the other compounds of antimony, _antimoniuretted hydrogen_, SbH_{3}, resembles arseniuretted hydrogen in its mode of formation and properties (it splits up at 150°, Brunn 1890; when liquified, it boils at -65° and solidifies at -92°), whilst the halogen compounds differ in many respects from those of arsenic. When chlorine is passed over an excess of antimony powder, it forms _antimony trichloride_, SbCl_{3}, but if the chlorine be in excess it forms the _pentachloride_, SbCl_{5}. The trichloride is a crystalline substance which melts at 72° and distils at 230°, whilst the pentachloride is a yellow liquid, which splits up into chlorine and the trichloride when heated; at 140° it begins to give off chlorine abundantly, carrying away the vapour of the trichloride with it; and at 200° the decomposition is complete, and pure antimonious chloride only passes over. This property of antimony pentachloride has caused it to be applied in many cases for the transference of chlorine; all the more that when it has given up its chlorine, it leaves the trichloride, which is able to absorb a fresh amount of chlorine; and therefore many substances which are unable to react directly with gaseous chlorine do so with antimony pentachloride, and in the presence of a small quantity of it chlorine will act on them, just as oxygen is able, in the presence of nitrogen oxides, to oxidise substances which could not be oxidised by means of free oxygen. Thus carbon bisulphide is not acted on by chlorine at low temperatures--this reaction requires a high temperature--but in the presence of antimony pentachloride its conversion into carbon tetrachloride takes place at low temperatures. Antimony tri- and pentachloride, having the character of chloranhydrides, fume in air, attract moisture, and are decomposed by water, forming antimonious and antimonic acids. But in the first action of water the trichloride does not evolve all its chlorine as hydrochloric acid, which is intelligible in view of the fact that antimonious anhydride is also a base, and is therefore able to react with acids; indeed antimony sulphide dissolved in an excess of hydrochloric acid (hydrogen sulphide is evolved) gives an aqueous solution of antimony trichloride, which, when carefully distilled, even gives the anhydrous compound. Antimony trichloride is only decomposed by an excess of water, and then not completely, for with a large quantity of water it forms _powder of algaroth_--_i.e._ antimony oxychloride. The first action of water consists in the formation of _oxychloride_, SbOCl--that is, a salt corresponding to oxide of antimony as a base. If antimony oxide or antimony chloride be dissolved in an excess of hydrochloric acid, and the solution diluted with a considerable amount of water, then this same powder of algaroth is precipitated. The composition varies with the relative amount of water; namely, between the limits SbOCl and Sb_{4}O_{5}Cl_{2}. The latter compound is, as it were, a basic salt of the former, because its composition = 2(SbOCl)Sb_{2}O_{3}.

With bromine and iodine, antimony forms compounds similar to those with chlorine. Antimonious bromide, SbBr_{3}, crystallises in colourless prisms, melts at 94°, and boils at 270°; antimonious iodide, SbI_{3}, forms red crystals of sp. gr. 5·0; antimony trifluoride, SbF_{3} separates from a solution of antimonious oxide in hydrofluoric acid, and SbF_{5} is formed by a similar treatment of antimonic acid. The latter gives easily-soluble double salts with the fluorides of the metals of the alkalis.

De Haën (1887) obtained very stable double soluble salts, SbF_{3},KCl (100 parts of water dissolve 57 parts of salt), SbF_{3},K_{2}SO_{4}, &c., which he proposed to make use of in the arts as very easily crystallisable and soluble salts of antimony.

Engel, by passing hydrochloric acid gas into a saturated solution of antimonious chloride at 0°, obtained a compound HCl,2SbCl_{3},2H_{2}O, and with the pentachloride a compound SbCl_{5},5HCl,10H_{2}O. Bismuth trichloride, BiCl_{3}, gives a similar compound.

Saunders (1892) obtained 5RbCl,3SbCl_{3} and RbCl,SbCl_{3}. Ditte and Metzner (1892) showed that Sb and Bi dissolve in hydrochloric acid only owing to the participation of the oxygen of the air or of that dissolved in the acid.

The heaviest analogue of nitrogen and phosphorus is _bismuth_, Bi = 208. Here, as in the other groups, the basic, metallic, properties increase with the atomic weight. Bismuth does not give any hydrogen compound and the highest oxide, Bi_{2}O_{5}, is a very feeble acid oxide. Bismuthous oxide, Bi_{2}O_{3}, is a base, and bismuth itself a perfect metal. To explain the other properties of bismuth it must further be remarked that in the eleventh series it follows mercury, thallium and lead, whose atomic weights are near to that of bismuth, and that therefore it resembles them and more especially its nearest neighbour, lead. Although PbO and PbO_{2}, represent types different from Bi_{2}O_{3} and Bi_{2}O_{5}, they resemble them in many respects, even in their external appearance, moreover the lower oxides both of Pb and Bi are basic and the higher acid, which easily evolve oxygen. But judging by the formula, Bi_{2}O_{3} is a more feeble base than PbO. They both easily give basic salts.

Bismuth forms compounds of two types, BiX_{3} and BiX_{5},[43] which entirely recall the two types we have already established for the compounds of lead. Just as in the case of lead, the type PbX_{2}, is basic, stable, easily formed, and passes with difficulty into the higher and lower types, which are unstable, so also in the case of bismuth the type of combination BiX_{3} is the usual basic form. The higher type of combination, BiX_{5},[44] in fact behaves toward this stable type, BiX_{3}, in exactly the same manner as lead dioxide does to the monoxide; and bismuthic acid is obtained by the action of chlorine on bismuth oxide suspended in water, in exactly the same way as lead dioxide is obtained from lead oxide. It is an oxidising agent like lead dioxide, and even the acid character in bismuthic acid is only slightly more developed than in lead dioxide. Here, as in the case of lead (minium), intermediate compounds are easily formed in which the bismuth of the lower oxide plays the part of a base combined with the acid which is formed by the higher form of the oxidation of bismuth.

[43] Metallic bismuth is very easily obtained when the compounds of the oxide are reduced by powerful reducing agents, but when less powerful reducing agents--for example, stannous oxide--are taken, bismuth suboxide is formed as a black crystalline powder. It is a compound of the type BiX_{2}, its composition being BiO; it is decomposed by acids into the metal and oxide, which passes into solution.

[44] The type BiX_{5} is represented by the pentoxide, Bi_{2}O_{5}, its metahydrate, Bi_{2}O_{5},H_{2}O, or BiHO_{3}, known as bismuthic acid, and the pyrohydrate, Bi_{2}H_{4}O_{7}. _Bismuth pentoxide_ is obtained by the prolonged passage of chlorine through a boiling solution of potassium hydroxide (sp. gr. 1·38), containing bismuth oxide in suspension; the precipitate is washed with water, with boiling nitric acid (but not for long, as otherwise the bismuthic acid is decomposed), then again with water, and finally the resultant bright red powder of the hydrate BiHO_{3} is dried at 125°. The prolonged action of nitric acid on bismuthic anhydride, Bi_{2}O_{5}, results in the formation of the compound Bi_{2}O_{4},H_{2}O, which decomposes in moist air, forming Bi_{2}O_{3}. The density of bismuthic anhydride is 5·10, of the tetroxide, Bi_{2}O_{4}, 3·60, and of bismuthic acid, BiHO_{3}, 5·75. _Pyrobismuthic acid_, Bi_{2}H_{4}O_{7}, forms a brown powder, which loses a portion of its water at 150°, and decomposes on further heating, with the evolution of oxygen and water. It is obtained by the action of potassium cyanide on a solution of bismuth nitrate. The meta-salts of bismuthic acid are known, for example KBiO_{3}. They generally occur, however, in combinations with metabismuthic acid itself. Thus André (1891) took a solution of the double salt of BiBr_{3} and KBr, treated it with bromine after adding ammonia, and obtained a red-brown precipitate, which after being washed (for several weeks) had the composition KBiO_{3},HBiO_{3} When washed with dilute nitric acid this salt gave bismuthic acid.

In nature, bismuth occurs in only a few localities and in small quantities, most frequently in a native state, and more rarely as oxide and as a compound of bismuth sulphide with the sulphides of other metals, and sometimes in gold ores. It is extracted from its native ores by simple fusion in the furnace shown in fig. 85. This furnace contains an inclined iron retort, into the upper extremity of which the ore is charged, and the molten _metal_ flows from the lower extremity. It is refined by re-melting, and the pure metal may be obtained by dissolving in nitric acid, decomposing the resultant salt with water, and reducing the precipitate by heating it with charcoal. Bismuth is a metal which crystallises very well from a molten state. Its specific gravity is 9·8; it melts at 269°, and if it be melted in a crucible, allowed to cool slowly, and the crust broken and the remaining molten liquid poured out, perfect rhombohedral crystals of bismuth are obtained on the sides of the crucible.[44 bis] It is brittle, has a grey-coloured fracture with a reddish lustre, is not hard, and is but very slightly ductile and malleable; it volatilises at a white heat and easily oxidises. It recalls antimony and lead in many of its properties. When oxidised in air, or when the nitrate is ignited, bismuth forms the _oxide_, Bi_{2}O_{3}, as a white powder which fuses when heated and resembles massicot. The addition of an excess of caustic potash to a solution of a bismuthous salt gives a white precipitate of the hydroxide, BiO(OH), which loses its water and gives the anhydrous oxide when boiled with a solution of caustic potash. Both the hydroxide and oxide easily dissolve in acids and form bismuthous salts.

[44 bis] Hérard (1889) obtained a peculiar variety of bismuth by heating pure crystalline bismuth to a bright red heat in a stream of nitrogen. A greenish vapour was deposited in the cooler portions of the apparatus in the form of a grey powder, which under the microscope had the appearance of minute globules. An atmosphere of nitrogen is necessary for this transformation, other gases such as hydrogen and carbonic oxide do not favour the transition. The resultant amorphous bismuth fuses at 410° (the crystalline variety at 269°), sp. gr. 9·483. (Does it not contain a nitride?)

_Bismuthous oxide_, Bi_{2}O_{3}, is a feeble and unenergetic base. The normal hydroxide of the oxide Bi_{2}O_{3} is Bi(OH)_{3}; it parts with water and forms a metahydroxide (bismuthyl hydroxide), BiO(OH). Both of these hydroxides have their corresponding saline compounds of the composition BiX_{3} and BiOX. And the form BiOX is nothing else but the type of the basic salt, because 3ROX = RX + R_{2}O_{3}. It is evident that in the type BiX_{3} the bismuth replaces three atoms of hydrogen. And indeed with phosphoric acid solutions of the bismuthous salts give a precipitate of the composition BiPO_{4}. On the other hand, in the form of compounds BiOX or Bi(OH)_{2}X, the univalent group (BiO) or (BiH_{2}O_{2}) is combined with X. Many bismuth salts are formed according to the type BiOX. For instance the carbonate, (BiO)_{2}CO_{3}, which corresponds with the other carbonates M_{2}CO_{3}. It is obtained as a white precipitate when a solution of sodium carbonate is added to a solution of a bismuth salt.[45] The compound radicle BiO is not a special natural grouping, as it was formerly represented to be; it is simply a mode of expression for showing the relation between the compound in question and the compounds of other oxides.

[45] Basic bismuth carbonate is employed for whitening the skin (veloutine, &c.)

Three _salts of nitric acid_ are known containing bismuthous oxide. If metallic bismuth or its oxide be dissolved in nitric acid, it forms a colourless transparent solution containing a salt which separates in large transparent crystals containing Bi(NO_{3})_{3},5H_{2}O. When heated at 80° these crystals melt in their water of crystallisation, and in so doing lose a portion of their nitric acid together with water, forming a salt whose empirical formula is Bi_{2}N_{2}H_{2}O_{9}. If the preceding salt belongs to the type BiX_{3}, this one should belong to the form BiOX, because it = BiO(NO_{3}) + Bi(H_{2}O_{2})(NO_{3}). This salt may be heated to 150° without change. When the first colourless crystalline salt dissolves in water _it is decomposed_. There is no decomposition if an excess of acid be added to the water--that is to say, the salt is able to exist in an acid solution without decomposing, without separation of the so-called basic salt--but by itself it cannot be kept in solution; water decomposes this salt, acting on it like an alkali. In other words the basic properties of bismuthic oxide are so feeble that even water acts by taking up a portion of the acid from it. Here we see one of the most striking facts, long since observed, confirming that action of water on salts about which we have spoken in Chapter X. and elsewhere. This action on water may be expressed thus:--BiX_{3} + 2H_{2}O = Bi(OH)_{2}X + 2XH. A salt of the type Bi(OH)_{2}X is obtained in the precipitate. But if the quantity of acid, HX, be increased, the salt BiX_{3} is again formed and passes into solution. The quantity of the salt BiOX which passes into solution on the addition of a given quantity of acid depends indisputably on the amount (mass) of water (Muir). The solution, which is perfectly transparent with a small amount of water, becomes cloudy and deposits the salt of the type BiOX, when diluted. The white flaky precipitate of Bi(OH)_{2}NO_{3} formed from the normal salt Bi(NO_{3})_{3} by mixing it with five parts of water, and in general with a small amount of water, is used in medicine under the name of magistery of bismuth.[46]

[46] With an excess of water a further quantity of acid is separated and a still more basic salt formed. The ultimate product, on which an excess of water has apparently no action whatever, is a substance having the composition BiO(NO_{3}).BiO(OH). In the latter salt we see the limit of change, and this limit appears to show that the type of the saline compounds of bismuthic oxide is of the form Bi_{2}X_{6}, and not BiX_{3}; but it is very probable, on the basis of the examples which we considered in the case of lead, that this type should be still further polymerised in order to give a correct idea of the type of the bismuthous compounds. If we refer all the bismuthous compounds to this type, Bi_{2}X_{6}, we shall obtain the following expression for the composition of the nitrates: normal salt, Bi_{2}(NO_{3})_{6}, first basic salt, Bi_{2}O(OH)_{2}(NO_{3})_{2}, magistery of bismuth, Bi_{2}(OH)_{4}(NO_{3})_{2}, and the limiting form Bi_{2}O_{2}(OH)(NO_{3}).

The general character of bismuthous oxide in its compounds is well exemplified in the nitrate; bismuthous chloride, BiCl_{3}, which is obtained by heating bismuth in chlorine, or by dissolving it in aqua regia, and then distilling without access of air, is also decomposed by water in exactly the same manner, and forms basic salts--for instance, first, BiOCl, like the above salt of nitric acid. Bismuth chloride boils at 447° and probably its formula is BiCl_{3}. Polymerisation may take place in some compounds and not in others. A volatile compound of the composition Bi(C_{2}H_{5})_{3} is also known as a liquid which is insoluble in water and decomposes with explosion when heated at 130°. Double salts containing chloride of bismuth are: 2(KCl)BiCl_{3}2H_{2}O (from a solution of Bi_{2}O_{3} and KCl in hydrochloric acid) and KClBiCl_{3}H_{2}O. Bigham (1892) also obtained KBr(SO_{4})_{2} in tabular crystals by treating the above-named double salt with strong sulphuric acid. The composition of this salt recalls that of alum.

Metallic bismuth is used in the preparation of fusible alloys. The addition of bismuth to many metals renders them very hard, and at the same time generally lowers their melting point to a considerable extent. Thus Wood's metal, which contains one part of cadmium, one part of tin, two parts of lead, and four parts of bismuth, fuses at about 60°, and in general many alloys composed of bismuth, tin, lead, and antimony melt below or about the boiling point of water.[47]

[47] As the metals contained in alloys like the above (bismuth, lead, tin, cadmium) are difficultly volatile and their alloys are fusible, they may be employed in the place of mercury in many physical experiments conducted at or above 70°, and they offer the advantage that they do not give any vapour having an appreciable tension (mercury at 100°, 0·75 mm.) Bismuth expands in passing into a molten state, but it has a temperature of maximum density. According to Luedeking the mean coefficient of expansion of liquid bismuth is 0·0000442 (between 270° and 303°), and of solid bismuth 0·0000411.

Just as in group II., side by side with the elements zinc, cadmium, and mercury in the uneven series, we found calcium, strontium, and barium in the even series; and as in group IV., parallel to silicon, germanium, tin, and lead, we noticed thallium, zirconium, cerium, and thorium; so also in group V. we find, beside those elements of the uneven series just considered by us, a series of analogues in the even series, which, with a certain degree of similarity (mainly quantitative, or relative to the atomic weights), also present a series of particular (qualitative) independent points of distinction. In the even series are known _vanadium_, which stands between titanium and chromium, _niobium_, between zirconium and molybdenum, and _tantalum_, situated near tungsten (an element of group VI. like chromium and molybdenum). Just as bismuth is similar in many respects to its neighbour lead, so also do these neighbouring elements resemble each other, even in their external appearance, not to mention the quality of their compounds, naturally taking into account the differences of type corresponding with the different groups. The occurrence in group V. determines the type of the oxides, R_{2}O_{3} and R_{2}O_{5}, and the development of an acid character in the higher oxides. The occurrence in the even series determines the absence of volatile compounds, RH_{3}, for these metals, and a more basic character of the oxides of a given composition than in the uneven series, &c.[48] Vanadium, niobium, and tantalum belong to the category of rare metals, and are exceedingly difficult to obtain pure, more especially owing to their similarity to, and occurrence with, chromium, tungsten and other metals, and also in combination among themselves; therefore it is natural that they have been far from completely studied, although since 1860 chemists have devoted not a little time to their investigation. The researches carried out by Marignac, at Geneva, on niobium, and by Sir Henry Roscoe, at Manchester, on vanadium deserve special attention. The undoubted external resemblance of the compounds of chromium and vanadium, as well as the want of completeness in the knowledge of the compounds of vanadium, long caused its oxides to be considered analogous in atomic composition to those formed by chromium. The higher oxide of vanadium was therefore supposed to have the formula VO_{3}. But the fact of the matter is, that the chemical analogy of the elements does not hold in one direction only; vanadium is at one and the same time the analogue of chromium, and consequently of the elements like sulphur of group VI, and also the analogue of phosphorus, arsenic, and antimony; just as bismuth stands in respect to lead and antimony. Investigation has shown that the compounds of vanadium are always accompanied by those of phosphorus as well as of iron, and that it is even more difficult to separate it from the compounds of phosphorus than from those of iron and tungsten. We should have to extend our description considerably if we wished to give the complete history, even of vanadium alone, not to mention niobium and tantalum, all the more as questions would not unfrequently arise concerning the compounds of these elements which have not yet been fully elucidated. We shall therefore limit ourselves to pointing out the most important features in the history of these elements, the more so since the minerals themselves in which they occur are exceedingly rare and only accessible to a few investigators.

[48] Although, guided by Brauner, who showed that didymium gives a higher oxide, Di_{2}O_{5}, I place this element in the fifth group, still I am not certain as to its position, because I consider that the questions relating to this metal are still far from being definitely answered.

An important point in the history of the members of this group is the circumstance that they form volatile compounds with chlorine, similar to the compounds of the elements of the phosphorus group, namely, of the type RX_{5}. The vapour densities of the compounds of this kind were determined, and served as the most important basis for the explanation of the atomic composition of these molecules. In this we see the power of general and fundamental laws, like the law of Avogadro-Gerhardt. An oxychloride, VOCl_{3}, is known for vanadium, which is the perfect analogue of phosphorus oxychloride. It was formerly considered to be vanadium chloride, for just as in the case of uranium (Chapter XXI.), its lower oxide, VO, was considered to be the metal, because it is exceedingly difficultly reduced--even potassium does not remove all the oxygen, besides which it has a metallic appearance, and decomposes acids like a metal; in a word, it simulates a metal in every respect. _Vanadium oxychloride_ is obtained by heating the trioxide, V_{2}O_{3}, mixed with charcoal, in a current of hydrogen; the lower oxide of vanadium is then formed, and this, when heated in a current of dry chlorine, gives the oxychloride VOCl_{3} as a reddish liquid which does not act on sodium and may be purified by distillation over this metal. It fumes in the air, giving reddish vapours; it reacts on water, forming hydrochloric and vanadic acids; hence, on the one hand it is very similar to phosphorus oxychloride, and on the other hand to chromium oxychloride, CrO_{2}Cl_{2} (Chapter XXI.). It is of a yellow colour, its specific gravity is 1·83, it boils at 120°, and its vapour density is 86 with respect to hydrogen; therefore the above formula expresses its molecular weight.[49]

[49] When the vapours of vanadium oxychloride are heated with zinc in a closed tube at 400°, they lose a portion of their chlorine and form a green crystalline mass of sp. gr. 2·88, which is deliquescent in air and has the composition VOCl_{2}. Only its vapour density is unknown, and it would be extremely important to determine whether its molecular composition is that given above, or whether it corresponds with the formula V_{2}O_{2}Cl_{4}. Another less volatile oxychloride, VOCl, is formed with it as a brown insoluble substance, which is, however, soluble in nitric acid like the preceding. Roscoe obtained a still less chlorinated substance, namely, (VO)_{2}Cl; but it may only consist of a mixture of VO and VOCl. At all events, we here find a graduated series such as is met with in the compounds of very few other elements.

_Vanadic anhydride_, V_{2}O_{5}, is obtained either in small quantities from certain clays where it accompanies the oxides of iron (hence some sorts of iron contain vanadium) and phosphoric acid, or from the rare minerals: _volborthite_, CuHVO_{4}, or basic vanadate of copper; _vanadinite_, PbCl_{2}3Pb_{3}(VO_{4})_{2}; lead vanadate, Pb_{3}(VO_{4})_{2}, &c. The latter salts are carefully ignited for some time with one-third of their weight of nitre; the fused mass thus formed is powdered and boiled in water: the yellow solution obtained contains potassium vanadate. The solution is neutralised with acid, and barium chloride added; a meta-salt, Ba(VO_{3})_{2}, is then precipitated as an almost insoluble white powder, which gives a solution of vanadic acid when boiled with sulphuric acid. (The precipitate is at first yellow, as long as it remains amorphous, but it afterwards becomes crystalline and white.) The solution thus obtained is neutralised with ammonia, which thus forms ammonium (meta) vanadate, NH_{4}VO_{3}, which, when evaporated, gives colourless crystals, insoluble in water containing sal-ammoniac; hence this salt is precipitated by adding solid sal-ammoniac to the solution. Ammonium vanadate, when ignited, leaves vanadic acid behind. In this it differs from the corresponding chromium salt, which is deoxidised into chromium oxide when ignited. In general, vanadic acid has but a small oxidising action. It is reduced with difficulty, like phosphoric or sulphuric acid, and in this differs from arsenic and chromic acids. Vanadic acid, like chromic acid, separates from its solution as the anhydride V_{2}O_{5}, and not in a hydrous state. Vanadic anhydride, V_{2}O_{5}, forms a reddish-brown mass, which easily fuses and re-solidifies into transparent crystals having a violet lustre (another point of resemblance to chromic acid); it dissolves in water, forming a yellow solution with a slightly acid reaction.[50]

[50] Strong acids and alkalis dissolve vanadic anhydride in considerable quantities, forming yellow solutions. When it is ignited, especially in a current of hydrogen, it evolves oxygen and forms the lower oxides; V_{2}O_{4} (acid solutions of a green colour, like the salts of chromic oxide), V_{2}O_{3}, and the lowest oxide, VO. The latter is the metallic powder which is obtained when the vanadium oxychloride is heated in an excess of hydrogen, and was formerly mistaken for metallic vanadium. When a solution of vanadic acid is treated with metallic zinc it forms a blue solution, which seems to contain this oxide. It acts as a reducing agent (and forms a close analogue to chromous oxide, CrO). Metallic _vanadium_ can only be obtained from vanadium chloride which is quite free from oxygen. Moissan (1893) obtained it by reducing the oxide with carbon in the electric furnace, and considered it to be most infusible of the metals in the series Pt, Cr, Mo, U, W, and V (he also obtained a compound of vanadium and carbon). The specific gravity of this metal is 5·5. It is of a grey-white colour, is not decomposed by water, and is not oxidised in air, but burns when strongly heated, and can be fused in a current of hydrogen (forming perhaps a compound with hydrogen). It is insoluble in hydrochloric acid, but easily dissolves in nitric acid, and when fused with caustic soda it forms sodium vanadate.

As regards the salts of vanadic acid, three different classes are known; the first correspond with metavanadic acid, VMO_{3} = M_{2}OV_{2}O_{5}, the second correspond with the dichromates--that is, have the composition V_{4}M_{2}O_{11}, which is equal to M_{2}O + 2V_{2}O_{5}--and the third correspond with orthovanadic acid, VM_{3}O_{4} or 3M_{2}O + V_{2}O_{5}. The latter are formed when vanadic anhydride is fused with an excess of an alkaline carbonate.

Vanadic acid gives the so-called 'complex' acids (which are considered more fully in Chapter XXI. in speaking of Mo and W)--_i.e._ acids formed of two acids assimilated into one. Thus Friedheim (1890) obtained phosphor-vanadic acid, and Schmitz-Dumont (1890) a similar arseno-vanadic acid. The former is obtained by heating V_{2}O_{5} with sirupy phosphoric acid. The resultant golden-yellow tabular crystals have the composition H_{2}OV_{2}O_{5}P_{2}O_{5}9H_{2}O, and there are corresponding salts--for example, (NH_{4})_{2}V_{2}O_{5}P_{2}O_{5} with 3 and 7H_{2}O, &c. These salts cannot be separated by crystallisation, so that there are 'complexes' of these acids in a whole series of salts (and also in nature). It may be supposed (Friedheim) that V_{2}O_{5} here, as it were, plays the part of a base, or that those acids may be looked upon as double salts. Among the true double salts of vanadium (Nb and Ta) very many are known among the fluorides, such as VF_{3}2NH_{4}F, VOF_{2}2NH_{4}F, VO_{2}F,3NH_{4}F, &c. (Pettersson, Piccini, and Georgi, 1890-92).

Vanadium was discovered at the beginning of this century by Del-Rio, and afterwards investigated by Sefström, but it was only in 1868 that Roscoe established the above formulæ of the vanadic compounds.

_Niobium and tantalum_[51] occur as acids in rare minerals, and are mainly extracted from _tantalite_ and _columbite_, which are found in Bavaria, Finland, North America, and in the Urals. These minerals are composed of the ferrous salts of niobic and tantalic acids; they contain about 15 per cent. of ferrous oxide in isomorphous mixture with manganous oxide, in combination with various proportions of tantalic and niobic anhydrides. These minerals are first fused with a considerable amount of potassium bisulphate, and the fused mass is boiled in water, which dissolves the ferrous and potassium salts and leaves an insoluble residue of impure niobic and tantalic acids. This raw product is then treated with ammonium sulphide, in order to extract the tin and tungsten, which pass into solution. The residue containing the acids (according to Marignac) is then treated with hydrofluoric acid, in which it entirely dissolves, and potassium fluoride is added to the resultant hot solution; on cooling, a sparingly soluble double fluoride of potassium and tantalum separates out in fine crystals, while the much more soluble niobium salt remains in solution. The difference in the solubility of these double salts in water acidified with hydrofluoric acid (in pure water the solution becomes cloudy after a certain time) is so great that the tantalum compound requires 150 parts of water for its solution, and the niobium compound only 13 parts. The Greenland columbite (specific gravity 5·36) only contains niobic acid, and that from Bodenmais, Bavaria (specific gravity 6·06) almost equal quantities of tantalic and niobic acids. Having isolated tantalic and niobic salts, Marignac found that the relation between the potassium and fluorine in them is very variable--that is, that there exist various double salts of fluoride of potassium, and of the fluorides of the metals of this group, but that with an excess of hydrofluoric acid both the tantalum and niobium compounds contain seven atoms of fluorine to two of potassium, whence it must be concluded that the simplest formula for these double salts will be K_{2}RF_{7} = RF_{5},2KF; that is, that the type of the higher compounds of niobium and tantalum is RX_{5}, and hence is similar to phosphoric acid. A chloride, TaCl_{5}, may be obtained from pure tantalic acid by heating it with charcoal in a current of chlorine. This is a yellow crystalline substance, which melts at 211°, and boils at 241°; its vapour density with respect to hydrogen is 180, as would follow from the formula TaCl_{5}. It is completely decomposed by water into tantalic and hydrochloric acids. _Niobium pentachloride_ may be prepared in the same manner; it fuses at 194°, and boils at 240°. When treated with water this substance gives a solution containing niobic acid, which only separates out on boiling the solution. Delafontaine and Deville found its vapour density to be 9·3 (air = 1), as is shown by its formula NbCl_{5}.[52]

[51] The researches made by Roscoe were preceded by those of Marignac in 1865, on the _compounds_ of _niobium_ and _tantalum_, to which were also ascribed different formulæ from those now recognised. Tantalum was discovered simultaneously with vanadium by Hatchett and Ekeberg, and was afterwards studied by Rose, who in 1844 discovered niobium in it. Notwithstanding the numerous researches of Hermann (in Moscow), Kobell, Rose, and Marignac, still there is not yet any certainty as to the purity of, and the properties ascribed to, the compounds of these elements. They are difficult to separate from each other, and especially from the cerite metals and titanium, &c., which accompany them. Before the investigations of Rose the highest oxide of tantalum was supposed to belong to the type TaX_{6}--that is, its composition was taken as TaO_{3}, and to the lower oxide was ascribed a formula TaO_{2}. Rose gave the formula TaO_{2} to the higher oxide, and discovered a new element called niobium in the substance previously supposed to be the lower oxide. He even admitted the existence of a third element occurring together with tantalum and niobium, which he named pelopium, but he afterwards found that pelopic acid was only another oxide of niobium, and he considered it probable that the higher oxide of this element is NbO_{2}, and the lower Nb_{2}O_{3}. Hermann found that niobic acid which was considered pure contained a considerable quantity of tantalic acid, and besides this he admitted the existence of another special metallic acid, which he called ilmenic acid, after the locality (the Ilmen mountains of the Urals) of the mineral from which he obtained it. V. Kobell recognised still another acid, which he called dianic acid, and these diverse statements were only brought into agreement in the sixties by Marignac. He first of all indicated an accurate method for the separation of tantalic and niobic compounds, which are always obtained in admixture.

[52] If niobic acid be mixed with a small quantity of charcoal and ignited in a stream of chlorine, a difficultly-fusible and difficultly-volatile oxychloride, NbOCl_{3} separates. The vapour density of this compound with respect to air is 7·5, and this vapour density perfectly confirms the accuracy of the formulæ given by Marignac, and indicates the quantitative analogy between the compounds of niobium and tantalum, and those of phosphorus and arsenic, and consequently also of vanadium. In their qualitative relations (as is evident also from the correspondence of the atomic weights), the compounds of tantalum and niobium exhibit a great analogy with the compounds of molybdenum and tungsten. Thus zinc, when acting on acid solutions of tantalic and niobic compounds, gives a blue coloration, exactly as it does with those of tungsten and molybdenum (also titanium). These acids form the same large number of salts as those of tungsten and molybdenum. The anhydrides of the acids are also insoluble in water, but as colloids are sometimes held in solution, just like those of titanic and molybdic acids. Furthermore, niobium is in every respect the nearest analogue of molybdenum, and tantalum of tungsten. _Niobium_ is obtained by reducing the double fluoride of niobium and sodium, with sodium. It is difficult to obtain in a pure state. It is a metal on which hydrochloric acid acts with some energy, as also does hydrofluoric acid mixed with nitric acid, and also a boiling solution of caustic potash. _Tantalum_, which is obtained in exactly the same way, is a much heavier metal. It is infusible, and is only acted on by a mixture of hydrofluoric and nitric acids. Rose in 1868 showed that in the reduction of the double fluoride, NbF_{5},2KF, by sodium, a greyish powder is obtained after treating with water. The specific gravity of this powder is 6·8, and he considers it to be niobium hydride, NbH. Neither did he obtain metallic niobium when he reduced with magnesium and aluminium, but an alloy, Al_{3}Nb, having a sp. gr. of 4·5.

Niobium, so far as is known, unites in three proportions with oxygen. NbO, which is formed when NbOF_{3},2KF is reduced by sodium; NbO_{2}, which is formed by igniting niobic acid in a stream of hydrogen, and niobic anhydride, Nb_{2}O_{5}, a white infusible substance, which is insoluble in acids, and has a specific gravity of 4·5. Tantalic anhydride closely resembles niobic anhydride, and has a specific gravity of 7·2. _The tantalates and niobates_ present the type of ortho-salts--for example, Na_{2}HNbO_{4},6H_{2}O, and also of pyro-salts, such as K_{3}HNb_{2}O_{7},6H_{2}O, and of meta-salts--for example, KNbO_{3},2H_{2}O. And, besides these, they give salts of a more complex type, containing a larger amount of the elements of the anhydride; thus, for instance, when niobic anhydride is fused with caustic potash it forms a salt which is soluble in water, and crystallises in monoclinic prisms, having the composition K_{8}Nb_{6}O_{19},16H_{2}O. There is a perfectly similar isomorphous salt of tantalic acid. Tantalite is a salt of the type of metatantalic acid, Fe(TaO_{3})_{2}. The composition of Yttrotantalite appears to correspond with orthotantalic acid.