The doctrine of _specific heat_ was taught by Black in his lectures at Glasgow between 1761 and 1765. The subject was subsequently investigated experimentally by Irvine between 1765 and 1770, and by Crawford in 1779. A series of determinations was published in 1781 by Wilcke, in the _Transactions_ of the Swedish Academy. In these the term _specific caloric_, since changed to _specific heat_, was first used. About this time the determination of the amount of heat required to raise substances through a definite interval of temperature was made the subject of experiment by many observers, notably by Lavoisier and Laplace, who greatly improved the calorimetric arrangements. The values they obtained long remained the most trustworthy estimations of the specific heats of substances. Their joint research had a further influence on the development of thermo-chemistry by indicating the general experimental conditions which were needed to ensure accuracy in such determinations. Lavoisier and Laplace also measured, in 1782–1783, the heat disengaged by the combustion of substances, and that evolved during respiration. In 1819 Dulong and Petit pointed out that the specific heat of a number of substances, more particularly the metals, were inversely proportional to their atomic weights; or, in other words, the product of the specific heat into the atomic weight was a constant. The nature of the relation will be seen from the following table of certain of the results obtained by Dulong and Petit:— Element. At. wt. Spec. heat. Atomic heat. Bismuth 208 0.0288 6.0 Lead 207 0.0293 6.0 Gold 197 0.0298 5.8 Platinum 195 0.0314 6.1 Silver 108 0.0570 6.1 Copper 63 0.0952 6.0 Iron 56 0.1138 6.4 It will be seen that these various elements have an uniform, or nearly uniform, atomic heat—approximately 6.2 on the average. This would appear to prove that, as Dulong and Petit expressed it, “the atoms of simple substances have equal capacities for heat.” The variations from a constant value are due partly to errors of observation, but more particularly to the circumstance that the substances compared are not all in a strictly comparable condition—_e.g._, they are not all equally remote from their melting points. It was shown, moreover, that the amount of heat needed to raise a substance through a definite interval of temperature increased with the temperature. The range of temperature through which a determination was made in a particular instance affected, therefore, the value of the specific heat. The most noteworthy departures from a uniform value were observed to occur among the metalloids—_e.g._, carbon, the various modifications of which had different specific heats—and generally among elements of low atomic weight, in which the variation of specific heat with temperature was particularly rapid.