HEAT.—States of Matter. 183
rate of expansion is different for different solids,and the co-efficient of dilatation usually increaseswith the temperature. Among the metals, platinumexpands the least, and its rate of expansion is nearlythe same as that of glass ; and hence a junctionalmost permanent may be made by fusing a fineplatinum wire into a glass tube. Senarmont hasshown that in crystals not belonging to the regularsystem, the conductivity or conducting power forheat is not the same in all directions. In a slice ofquartz cut at right angles to the axis of the prism,the conductivity will be the same from the centreto the circumference; but this will not be the casein a plate of quartz which has been cut parallel tothe same axis. The fundamental experiment ofSenarmont can be easily repeated with a slice ofselenite covered with wax, and having a perforationin the centre for the introduction of a heated wire.The wax on melting forms a well-defined ellipsefrom the unequal transmission of the heat.
When solids are heated to a certain temperaturedepending on the body itself, they pass into theliquid state with the absorption of a definite amountof heat. In the case of many bodies, such as iron,glass, boracic acid, &c., this change is preceded bya well-marked viscid condition. In all bodieshitherto examined there is a change of volume inthe conversion of the solid into the liquid, the liquidusually occupying a greater volume than the solid;but in some cases, as water, bismuth, and cast iron,a diminution of volume occurs in liquefaction. J.Thomson drew the important inference on theoreticalgrounds that the melting point of bodies whichexpand in liquefying is raised by pressure, while itis, on the other hand, lowered if the body contractsin liquefying. He calculated in the case of waterthat the freezing point would be reduced 0°-0076 C.for each atmosphere of additional pressure. Thisresult was experimentally verified by Sir W. Thom-son ; and Bunsen has shown that paraffin andspermaceti, which expand in liquefying, have theirmelting points raised by pressure. The remarkablephenomenon of regelation, first discovered by Fara-day, has been successfully explained by J. Thomsonfrom the lowering of the freezing point of water bypressure. When two pieces of ice at the meltingpoint are pressed together, the pressure causesmelting to take place at the portions of the surface incontact. The temperature of liquefaction is lowered,while the water so formed escapes out of the way.Hence, as soon as the pressure diminishes, the twoparts are frozen together at a temperature below0°. On the same principle is explained the follow-ing beautiful experiment, which is due to J. Bottom-ley :—A fine metallic wire passing over a block ofice has a weight attached to it, so as to exertpressure upon the ice. .The wire will graduallyenter the ice, melting the surface on which it presses,while the water thus formed immediately freezesagain behind the wire, so that the block remains asolid mass after having been traversed by the wire.
Water may be cooled several degrees below thefreezing point without changing into ice, particularly
if it is deprived of air, and inclosed in an exhaustedhermetically sealed vessel. Other liquids, such asbromine and glacial acetic acid, exhibit this pheno-menon, so that the same body may be preserved ina closed vessel for any length of time, within acertain range of temperature, either as a solid ora liquid.
The determination of the compressibility of liquidsis a problem of great difficulty, involving for itscomplete solution some mechanical questions whichare not definitely settled. Regnault has attemptedto ascertain by experiment the absolute compressi-bility of mercury and water, his method of operatingbeing founded on equations supplied by Lame.
For one atmosphere, the compressibility of mercuryis 0-000003517, or -s-g-jVjTirib of the original volume;for water the corresponding number is 0-00004746, orTTfftnr^ original volume. Within the limits
of his experiments, Regnault has confirmed theobservation, first made by (Erstedt, that the co-efficient of compressibility does not sensibly diminishfor these liquids when the pressure augments. Butliquids near the “ critical temperature ” differ greatlyin compressibility from mercury or water. Theyyield much more to pressure than ordinary liquids,and the compressibility diminishes as the pressureincreases. The average compressibility for oneatmosphere of liquid carbonic acid between 55 and
125 atmospheres is 0-00133, or yTgth of the volumeunder 54 atmospheres; while from 125 to 220atmospheres, the average compressibility for eachatmosphere is 0-00071, or -j-yg-g-th of the volumeunder 125 atmospheres (Andrews).
As a general rule, liquids dilate when heated, theonly exception yet known being water, which con-tracts when heated from 0° to 4°. The dilatation ofliquids by heat augments with the temperature, butnot according to any simple law. For this reasonthere is an error of nearly 0°-2 in the mercurial ther-mometer at temperatures midway between 0° and
100°. At higher temperatures the error arising fromthis cause amounts to several degrees, and requiresto be carefully corrected. The absolute dilatationof a liquid by heat, that is, its change of volume,independent of the change of capacity of the contain-ing vessel, can be determined by the method ofDulong and Petit, which depends essentially on thefact that the heights of two columns of liquid com-municating below by a horizontal tube are inverselyas the densities of the liquids. Hence, if two columnsof the same liquid, at different temperatures, becompared, the absolute change of volume due to thedifference of temperature will be obtained. Tiiil-ORIER observed that the dilatation by heat of liquidcarbonic acid is greater than that of the gases, andDrion has found that the co-efficient of dilatation ofliquid sulphurous add increases rapidly with thetemperature, being for 1° C., 0-00173 at 0° and0-00486 at 100°. In the case of certain homologousliquids, such as the methyl and ethyl alcohols, theco-efficient of dilatation is nearly the same, if theliquids are examined through corresponding rangesof temperature from their boiling points.