are, so to speak, free to take up six atoms of chlorine. In the sesquiiodide and sesquichloride of silicon the six atoms of iodine and chlorine play the same part, and the two atoms of silicon are united together, exchanging the fourth unit of saturation, or valency, which each of them possesses :—

The sesquichloride of titanium shows an analogous composition. It must be remarked that the formulæ in question cannot be halved. The vapour density of all these bodies has been taken, and their molecular condensation must be expressed by the preceding formulæ.

The chlorides of iron and aluminium are analogous to the preceding chlorides. The result of the classical researches of H. Sainte-Claire Deville and Troost upon the vapour density of these chlorides has been to attribute to them the formulæ FeCl and Al2Cl; and we are forced to admit that the two atoms of iron and aluminium are united together in the same manner as the atoms of carbon, silicon, and titanium in the corresponding chlorides.

The couples Fe-Fe and Al-Al are, then, sexvalent. This ingenious idea is due to Friedel. Considering iron as quadrivalent in pyrites, FeS,,' the eminent chemist

'The ferric tetrachloride corresponding to pyrites does not exist. The interpretation of this want lies in the fact that in the action

regards the ferric compounds as containing two atoms of tetratomic iron united by the interchange of two units of saturation. In the couple (Fe-Fe), ferricum, there remain, therefore, only six free or disposable units of saturation. The violet chromic chloride, and perhaps the compounds which are called sesquichloride of osmium and ruthenium, have the same molecular constitution as the preceding chlorides.

These trioxides must not be confounded with the sesquioxides properly so called, which contain trivalent elements, such as arsenic, antimony, bismuth, and gold. These sesquioxides correspond to trichlorides, and the two atoms of metal which they contain are united, not directly with each other, but through an intermediary atom of oxygen.

of chlorine upon the protochloride a tetrachloride is not formed, because the affinity of iron for iron is greater than that of four atoms of chlorine for iron.

FeCl + FeCl = (Fe--Fe)"'Cl + C1C1.

We must add that important researches made by Scheurer-Kestner the ferric salt have confirmed the existence of sexvalent iron,

upon

Fe,

Iridium and rhodium also form well-characterised trichlorides and sesquioxides, which seem to belong to the preceding series; but they also form dichlorides, or rather tetrachlorides, in which we may admit the existence of couples (Ir—Ir) and (Rh—Rh) formed by the union of two atoms of iridium or two atoms of rhodium, which, having exchanged one unit of saturation, now possess only four atomicities.

As a final example of these unions which the atoms of the same element may form, by the partial exchange of their atomicities or units of saturation, we may mention the cuprous and mercurous compounds, of which the first contain two atoms of copper, the second two atoms of mercury, united together.

The formula which is here attributed to mercurous chloride has been amply justified (p. 115), whence it seems allowable to attribute an analogous composition to cuprous chloride, though here there is some

certainty.

un

CHAPTER II.

I.

Affinity and Atomicity, two Distinct Properties of

Atoms.

WE have in the preceding pages defined atomicity by regarding it as the saturating capacity of atoms, or as their valency in combinations. It is, then, a property inherent in the nature of atoms.

show how it differs from affinity.

We must proceed to

Affinity is the force of combination, chemical energy. It determines the intensity and the direction of chemical reactions, and is estimated by the thermal effects which these reactions produce. It varies essentially with different atoms. In combining with atoms of hydrogen, atoms of chlorine, iodine, and bromine liberate very different quantities of heat; their affinity for hydrogen is very different, and is proportional to the quantities of heat liberated. But if we consider the combinations of the same elements with oxygen we shall find the order of affinities reversed. Chlorine is the element which possesses the weakest affinity for this

body. The compounds of chlorine and oxygen are very unstable; some decompose with explosion-that is to say, are formed with absorption of heat. The affinity or chemical energy of a given body must therefore be considered as a relative property. It depends upon the nature of the element with which the one in question combines.

It depends also upon the conditions under which the bodies are placed. Berthollet long ago showed the influence which is exercised upon affinity by physical conditions, such as the degree of cohesion and the insolubility of bodies. This fact is too well known to require further remark (see p. 4); but we must remember how physical agents, such as heat, light, or electricity, can augment or diminish chemical energy, stimulate or retard the exercise of affinity. If mercury is heated to a certain temperature its atoms are in a condition capable of attracting atoms of oxygen. If the heat is increased the atoms of mercury and oxygen will be separated again. The affinity of mercury for oxygen is therefore subordinate to the temperature. It is a relative and not an absolute property, like the atomic weight. In the same manner a stream of electric sparks or the silent electric discharge can determine combinations between atoms which would have no action upon each other under ordinary conditions. Inversely, the same influences can produce decomposition, as is the case with the battery current. Here, again, the conditions in which the atoms are placed exercise a visible influence upon their affinities.

Atomicity is the capacity of saturation, or the value