system, is different in different directions; in fact, it appears probable that an expansion in one direction is accompanied by a contraction in another.

§ 65. Fusion and Solidification.—When heat is applied to a solid body, provided no chemical change is produced, then sooner or later the coherence of the particles is so far reduced that the solid melts; the individual particles are then able to move freely around one another, but still their coherence has not been completely overcome.

In many instances other changes of solidity precede the liquefaction, whilst in others, as soon as a definite temperature, the melting point, is attained the solids suddenly and completely liquefy. Others again soften or become pasty before melting, passing, in fact, through a state intermediate between the solid and the liquid. In this plastic condition particles can be welded together by pressure, as is the case with metals like iron and platinum. Some metals and some of the semi-metals, such as zinc, bismuth, and tellurium, before melting become brittle at a certain temperature, whilst at other temperatures they are malleable and ductile, and can then be either rolled into sheets or drawn into wire.

The change in the state of aggregation is associated with a greater or lesser absorption of heat. When the temperature of a solid is very much below its melting point, a definite amount of heat is required to produce a certain rise in temperature for each part by weight of the substance, and this is approximately the same for every degree of temperature. This amount of heat so required is styled 'the specific heat.' When the body begins to soften under the application of heat, the amount of heat required to produce a given rise in temperature increases more and more, until when the body melts the amount of heat absorbed is considerable, and is no longer perceptible as such, becoming, in fact, latent heat. The heat so absorbed serves in all probability to give an accelerated motion to the particles, and being thus converted into motion is no longer perceptible as heat. The fusion proceeds only in proportion as the heat is applied, and as this serves only for melting, the temperature remains stationary until the whole mass is fused. On the other hand, when a molten mass gives up the heat to

MELTING POINTS OF THE ELEMENTS

113

surrounding objects its temperature is not necessarily lowered below the melting point, for the part solidifying will give out its latent heat of fusion. Nor is it until the whole has solidified that the temperature begins to sink. A molten body may, however, be frequently cooled below its melting point without solidifying. In this state of superfusion the particles are in a condition of unstable equilibrium, such that the slightest change suffices to bring about solidification. This solidification is more easily produced by contact with the minutest fraction of the solid itself. At the moment of solidification the temperature rises to that of the melting point but no further: this rise in temperature is produced by the liberation of the latent heat. This acceleration in the rate of motion of the particles, corresponds to considerable increase in volume, which, as a rule, appears to take place suddenly on fusion or in part during the softening, this increase amounting in some cases to 12 or more per cent. of the volume of the solid. Yet in the case of some substances, especially water, cast iron, bismuth, and some of its compounds and alloys, and perhaps also in the case of other metals, contraction is known to attend the fusion, which can perhaps be explained as arising from an altered arrangement of the atoms in the molecules. In water this contraction amounts to nearly 10 per cent. of the volume. The change in the state of aggregation produced by pressure depends upon whether fusion be attended by an expansion or contraction, and in such a way that by sufficiently great pressure that condition is produced in which the material fills the smallest space. Ice can be liquefied by pressure, whilst by its aid the majority of other solid substances can be retained in the solid state at temperatures much above their melting points.

§ 66. Melting Points of the Elements.-The temperatures at which different substances melt are specific and characteristic for each, and serve, therefore, as important aids for their identification. In § 36 it has already been mentioned that the fusibility of the elements is a periodic function of their atomic weights. This relationship, so far as it has been in any way ascertained, is exhibited in the following table. The melting points of many elements are still unknown, because the temperature at which they melt is either too high or too low to be accurately deter

I

mined; in some other cases the rarity of the element or the difficulties surrounding its isolation have prevented the exact determination. In the following table the abbreviations used are: a= approximation, b = above, c = very low, d= very high, e = not melted, rh = red heat, drh dull red heat, brh=bright red heat, wh= white heat.

=

The elements are arranged in the horizontal lines in the order of their atomic weights. With these the melting point rises suddenly and falls suddenly; the minima of the melting points are printed in italics, the maxima in block print.

The periods of fusibility do not coincide with those of other physical properties-in fact, are less regular than these, but are nearly related, as has already been shown in § 36, to those of the atomic volumes.

It is remarkable that in every family the members of one group are difficultly fusible, whilst those of the other are easily fusible; e.g. lithium, sodium, potassium, rubidium, cæsium are easily fusible, whilst copper, silver, and gold melt at high temperatures, and similar relationships are found to exist in other families. In separate groups the melting point changes with the atomic weight, but not in the same manner. In some families. the melting point falls with increase in atomic weight, thus:

MELTING POINTS OF COMPOUNDS

Li 180°, Na 96°, K 63°, Rb 39°, Cs 26°,

Zn 433°, Cd 321°, Hg - 39°;

115

in others, again, it rises with increase in atomic weight; for example :

Ga 30°, In 176°, TI 294°,

Cl-105°, Br-7°, I + 114°;

whilst in some families it rises at first to fall again, or falls first and then rises.

§ 67. Melting Points of Compounds. In the melting points of compounds we have similar differences to those exhibited by the elements. By the introduction into a compound of certain elements the fusibility is in some cases raised, in other cases lowered. The oxides of metals, e.g., are much more difficultly fusible than the metals themselves; the majority of the oxides of the non-metals are more easily fusible than the elements; in one and the same group of elements these changes are, as a rule, found to be of the same character, but even in this case also there are exceptions. Whilst, for example, the infusible element carbon yields an oxide (CO) which melts at 60°, the corresponding oxide (SiO) of the difficultly fusible silicon is almost as difficultly fusible as the element itself. Fluorides, chlorides, bromides, iodides, melt, as a rule, much more easily than the oxides, and usually the iodide of an element is more easily fusible than the bromide, and this than the chloride, whilst the fluoride has the highest melting point. Thus, for example, the melting points of halogen compounds of the alkali metals are, according to Carnelley, as follows:

The melting point falls, therefore, with increased atomic weight of the halogen, and similar relationships are to be found in other families of the elements.

Many similar regularities are to be found amongst organic compounds; still our knowledge of the general laws in this province is much less extensive than might be imagined from the thousands of melting-point determinations which have been made.

It is, however, to be observed that in many cases the repeated introduction of a given atom or a group of atoms in an organic compound is accompanied by alternate raising and lowering of fusibility. This is the case, as was first shown by Baeyer, in the normal primary fatty acids of the general formula CH2O2. In these compounds the atom linkage is represented as follows:

n

in accordance with which, the several members of the series differ from one another only in the number of CH2 groups introduced between the carboxyl group, COOH, and hydrogen. The relationships are shown in the following table :

From the above it is seen that the first introduction of the group