having a close analogy to the former: In such a case we have in factA: AD: D';-thus for

Chloride of sodium........ A = 1.169, D= 2.24

The density of the latter salt by experiment is 1.994 (vid. p. 80); according to Karsten it is 1.915, according to Kopp, 1.945.-Again for Sulphate of baryta........ ▲ 3393, D = 4.44

=

strontia .... A' = 2.901

and 3.393 : 2.901 :: 4.44 : 3.796

The experimental result is 3.77; Karsten makes it 3-59, Breithaupt 3.95.

Other examples might be adduced, but the relation thus developed is not sufficiently general to constitute a law.

In conclusion, M. Filhol remarks that the subject of the relation between atomic weight and density is not so far advanced as has been imagined, that it is easy by slightly altering the numbers which form the bases of these calculations to obtain results of very fascinating uniformity: but this very circumstance ought to create distrust of such results; for it is extremely difficult-considering the great number of causes of error, often inevitable, which beset these calculations-to attain that degree of exactness which apparently belongs to many of the results which have been put forth in connection with this subject.

A series of elaborate memoirs on atomic volume and specific gravity has been published by Messrs. Playfair and Joule in the Memoirs and Quarterly Journal of the Chemical Society of London. (Chem. Mem. 2, 477; 3, 54 and 199; Qu. J. of Chem. Soc. 1, 121.) The principal results contained in these memoirs are as follows.

I. The volumes of salts in solution are either equal to each other, or are multiples one of the other.

If 9, the atomic volume of water (atomic weight of hydrogen = 1) be assumed as the standard of comparison, this law may also be thus expressed: Compounds dissolved in water increase its volume either by 9 or by multiples of 9.

a. Certain salts, such as the magnesian sulphates, the alums, &c., dissolve in water without increasing its bulk more than is due to the liquefaction of the water which they themselves contain.

b. Anhydrous salts, or salts containing a small proportion of water occupy in solution a number of volumes, which pass along with them unchanged into their union with other salts.

c. The volume occupied by double salts when dissolved is (with certain exceptions) the sum of the volumes occupied by their constituents when separate.

II. The volumes of solid bodies bear a simple relation to each other, being multiples of a certain submultiple of the volume of ice.

=

The density of ice as determined by Playfair and Joule is 0.9184, which agrees almost exactly with the number 0.9180 determined by Brunner (N. Ann. Chim. Phys. 14, 369). Hence the atomic volume of ice = 9.8. This being determined, the submultiple in question adopted by the authors as the unit of volume for solids is = 1.225; so that the atom of ice, considered as a globe or cube, will have twice the linear dimensions of the atom possessing this unit-volume. In some cases the volumes of solid bodies correspond more nearly to multiples of

122506125 than of 1.225 itself: but these cases being comparatively few are regarded as exceptional. The following table gives the atomic volumes, &c. of some of the metals in a finely divided state'; also of flowers of sulphur.

The volumes of the metals in their more compact state do not accord very well with the law above stated, probably from the effect of cohesion.

=

The volume of the magnesian oxides in their most compact state, is found to be very nearly 5 × 1·225 6·12. Hence that of oxygen in these oxides must be 2 x 1.225 = 2·45*.-Now the volume of sulphur is 8.57; consequently, that of sulphuric acid, S O3, will be 8:57 + 3 x 2·45 = 15.92. Again the mean sp. gr. of sulphate of soda, as determined by various experimenters, is 2.562: this gives 27.9 for the atomic volume of that salt; and deducting 15.9 the volume of sulphuric acid, we have 12 for the atomic volume of soda: this is nearly 10 x 1·225 12 25.—The volume of potash (deduced from the sp. gr. 2756 determined by Karsten) is 17.75, which is not far from 14 x 1225 = 17·15.— The volume of anhydrous sulphate of ammonia is 39.2 (Chem Mem. 2, 428), from which deducting 15.9 for the acid, we have left 23-3 as the volume of oxide of ammonium. This is very nearly 18 × 1.225 23.27.-Aluminum has a density of 2.5 (Wöhler): this gives 5.47 for its atomic volume, making it nearly equal to 4x 1.225 or 5.5. Chromium has the density 5.1 which makes its atomic volume = 5.5: and admitting that in the sesqui-oxides oxygen enters with the volume 3 x 1.225 = 3.675, we shall have for the volumes of alumina and oxide of chronium 5.51 x 2 + 3.675 × 3 = 22.05.-Sesqui-oxide of iron being isomorphous with alumina and oxide of chromium its volume in combination may likewise be assigned as 22.05. The reason for selecting the particular results detailed in this paragraph will be seen immediately. The volumes assigned to alumina and the sesqui-oxides of iron and chromium, must be regarded as to a considerable extent hypothetical.

III. In highly hydrated salts the water of crystallization always occupies the volume of ice.

a. In the class of hydrated arseniates and phosphates with 24 atoms of water of crystallization, and in carbonate of soda with 10 atoms of water, neither acid nor base occupies any appreciable space, the volume of

*This agrees very nearly with the number 32 assigned by Kopp (whose numbers are based on the oxygen scale of atomic weights) for the atomic volume of oxygen: for 32 X 12.56. (W.)

the salt being the same as that of the water of crystallization rozen into ice.

b. In cane and milk-sugar the carbon ceases to occupy space, the hydrogen and oxygen taking up exactly the space of the corresponding quantity of water frozen into ice.

These results are exhibited in the following table. It is especially remarkable that in the ordinary phosphate and arseniate of soda, the atom of basic water disappears as well as the two atoms of soda.

IV. Another class of salts including all the hydrated magnesian sulphates, sulphate of alumina, borax, pyrophosphate of soda, and the alums, possess a volume made up of that of their bases and of their solid water, their acids ceasing to be recognizable in space.

It will be seen from these tables that many salts contain one atom of water of crystallization for every unit-volume in their base. Thus sulphate of alumina possesses 18 atoms of water, and the volume of its base is 18 x 1.225. Sulphate of soda has ten atoms of water, and the volume of its base is 10 x 1.225. Biborate of soda also crystallizes with 10 atoms of water. The magnesian sulphates generally crystallize with 7 atoms of water of these however 2 atoms are united by a much less powerful affinity than the rest, being driven off by a heat of 212°, and even under certain circumstances escaping in dry air at ordinary temperatures. The number of atoms of water essential to the crystallized salt may therefore be estimated at 5, which is the actual number contained in the ordinary crystals of sulphate of copper. The volume of the base of these salts is 5 x 1.225. Now the volume of solid water being 9.8 and the unit-volume of the base 1.225, it follows that the volume of the salt (in which the acid does not appear) must be a multiple of 9.8 + 1.225, that is of 11.025. In the first series of researches by Messrs. Playfair and Joule (Chem. Mem. vol. II. p. 401) numerous tables are given showing that in many classes of salts-sulphates, chlorides, oxalates, &c. the volume in the solid state is a multiple of 11 or of some number very near it. The explanation of this fact is contained in what has just been stated. T

b. State of Aggregation.

A compound is at ordinary temperatures either solid, liquid or gaseous. I. A solid compound may be formed:

1. From two gases. Condensation.-Hydrochloric acid gas forms a solid compound with ammoniacal gas, viz. sal-ammoniac.

2. From a gaseous and a liquid body. Absorption.-Mercury absorbs chlorine and oxygen gases, forming solid compounds.

3. From a gaseous and a solid body. Absorption again.-Iron and other solid metals absorb oxygen gas, and hydrate of soda absorbs carbonic acid gas.

4. From two liquids.-Mercury and bromine.

5. From a liquid and a solid.-Mercury forms solid amalgams with several metals; burnt lime mixed with its weight of water crumbles to solid hydrate of lime; burnt gypsum mixed with water hardens into the state of gypsum combined with water of crystallization.

6. From two solids, generally by fusion.-The combinations of metals with one another or with sulphur.

II. A liquid compound may be formed:

1. From two gases.

bine and form water.

Condensation.-Hydrogen and oxygen gases com

2. From a gas and a liquid. Absorption.-Water absorbs hydrochloric acid gas forming solution of hydrochloric acid.

3. From a gas and a solid. Absorption.-Arsenic, antimony, or tin absorbs chlorine gas, forming a liquid metallic chloride.

4. From two liquids. Mixture in its most confined sense.-Water and alcohol; sulphuret of carbon and chloride of sulphur.

5. From a solid substance and one that is liquid at the ordinary or a somewhat higher temperature. Solution in the wet way.-Salts and water, camphor and spirit of wine, sulphur and fatty matters, lead and

mercury.

6. From two solids.-Sometimes in the cold, as common salt and ice,

bismuth-amalgam and lead-amalgam; sometimes not below a red heat, as carbon and sulphur.

III. A compound gaseous at the ordinary temperature and pressure of the air arises only:

1. From two permanent gases.-Hydrogen and chlorine.

2. From a permanent gas and a liquid.--Hydrogen gas and bromine. 3. From a permanent gas and a solid.-Hydrogen gas and sulphur; oxygen gas and carbon.

Since no compound which is gaseous at ordinary pressures and temperatures is ever formed by the combination of two liquids or two solids or a solid and a liquid, while on the contrary solid and liquid compounds are formed by the union of two permanent gases, it may be surmised that if any of the hitherto undecomposed bodies are really compound, such will probably be found among the solid and liquid classes.

The less completely the mutual affinity of ponderable bodies is satisfied, or in other words, the less complicated the combinations which they form, the stronger is their attraction for heat and the greater their elasticity; those elements which are gaseous under ordinary circumstances have also on an average the smallest atomic weights.

c. Crystalline Form.

The crystalline form of a compound probably bears a definite relation to that of its elements. Such a relation however has not yet been completely traced out, partly because the crystalline forms of many important elements, oxygen for instance, are unknown,-partly because one and the same substance, simple or compound, often assumes one or another crystalline form according to circumstances, i.e. exhibits Dimorphism. (q. v.) The existence of such a relation is however apparent from the facts by which Mitscherlich has established his important theory of Isomorphism. The term Isomorphous in its widest sense applies to those bodies which can replace one another in a compound without producing any alteration in the crystalline form of that compound, except small angular differences. Such bodies may be divided into the following groups.

A. Substances which are isomorphous both in the separate state and in combination.-Substances possessing the same crystalline form and replacing one another in combination according to equal numbers of atoms without alteration of crystalline form. Arsenic and antimony crystallize in acute rhombohedrons: As O3 exhibits the same crystalline form as Sb O3, and many double salts containing As O3 as one base, present, according to Mitscherlich, the same crystalline form as the corresponding salts in which As O3 is replaced by Sb 03.

B. Substances which replace one another in compounds according to equal number of atoms. The crystalline form of such substances in the separate state is either different or else unknown; but they replace one another in combinations according to equal numbers of atoms and without alteration of crystalline form. Titanium crystallizes in cubes, tin in regular six-sided prisms; but both Ti O2 and Sn O crystallize in square prisms. The crystalline forms of lime and magnesia are unknown, but Ca O, CO and MgO, CO2 crystallize in obtuse rhombohedrons. This group of substances may perhaps be hereafter shown to be identical with the first, when we shall have become acquainted with the crystalline forms of these bodies in the separate state and, possibly have discovered that their difference of form may be referred to Dimorphism.

C. Substances which replace one another in combination according to