Now we choose mass as a unit of measure of the _quantity of material_; and we are justified in doing so, because experiments have shown that material, confined in a closed space, does not appreciably alter its mass. The mass is proportional to the _weight_, generally measured as the force exerted at some definite latitude on the earth’s surface, tending to pull the body towards the centre of the earth. This force is equal to about 981 dynes at London. There is no known reason why mass and weight should be proportional to each other; for the cause of the attraction of the earth has never been satisfactorily elucidated. We therefore use weight, or the attraction of the earth, as a convenient means of determining the relative masses of two objects. Hence it appears rational to prefer the expression “atomic weight” to “atomic mass,” seeing that the former represents the actual result of experiment, and also because we are dealing in atomic weights with relative numbers. But this is really a matter of choice. The atomic weights therefore represent the relative masses in which the elements generally unite. They often, however, unite in multiples of these weights, as formulated by Dalton’s second law. The weights, arranged in numerical order in columns, give us the periodic table. Now energy can be measured in other units besides those of force and mass. Heat is one form of energy, and it is measured by an interval of temperature, and by a property which we term specific heat. It happens that the latter property varies, not with the mass of the substance heated, but with its atom, so that all elements have approximately the same atomic heat; that is, quantities of elements proportional in mass to their atomic weights require approximately equal increments of heat to raise their temperature through an equal interval, say 1°. This is the formulation of Dulong and Petit’s law previously alluded to. But here we meet with irregularities, which have up till now defied classification. The heat imparted to an aggregation of atoms is not expended solely in raising their temperature; other work is done also, as is generally supposed, in the nature of expansion, or separation of parts, as in overcoming the attraction between the atoms in the molecule, or in imparting special motion to the atoms; such work, however, involves an expenditure of energy which is either very small in proportion to the total energy imparted as heat, or is nearly the same for all elements. At present we cannot decide between these alternatives, owing to the lack of knowledge of the nature of liquids and solids. The main fact, however, is incontestable: that the heat energy required to raise different elements through the same interval of temperature is the same, not for equal masses or weights of the substances, but for their _atomic_ masses.