In considering the development of Dumas's substitution theory, we found that Laurent retained this conception of equivalency when he spoke of an equivalent of hydrogen being replaced by an equivalent of chlorine (see p. 272). A certain weight of chlorine was able to take the place and play the part of a certain weight of hydrogen in a compound; these weights, of hydrogen and chlorine, were therefore equivalent. This conception has been much used since Laurent's time, but it has for the most part been applied to the atoms of the elements. Hydrogen being taken as the standard substance, the elements have been divided into groups, in accordance with the number of hydrogen atoms with which one atom of each element is found to combine. Thus certain elements combine with hydrogen only in the proportion of one atom with one atom; others combine in the proportion of one atom with two atoms of hydrogen; others in the proportion of one atom with three atoms of hydrogen, and so on. The adjective _monovalent_, _divalent_, _trivalent_, etc., is prefixed to an element to denote that the atom of this element combines with one, or two, or three, etc., atoms of hydrogen to form a compound molecule. Let us consider what is implied in this statement--"The nitrogen atom is trivalent." This statement, if amplified, would run thus: "One atom of nitrogen combines with three atoms of hydrogen to form a compound molecule." Now, this implies (1) that the atomic weight of nitrogen is known, and (2) that the molecular weight, and the number of nitrogen and hydrogen atoms in the molecule, of a compound of nitrogen and hydrogen are also known. But before the atomic weight of an element can be determined, it is necessary (as we found on p. 146) to obtain, analyze, and take the specific gravities of a series of gaseous compounds of that element. The smallest amount of the element (referred to hydrogen as unity) in the molecule of any one of these gases will then be the atomic weight of the element. When it is said that "the molecular weight, and the number of nitrogen and hydrogen atoms in the molecule, of a compound of nitrogen and hydrogen are known," the statement implies that the compound in question has been obtained in a pure state, has been analyzed carefully, has been gasefied, and that a known volume of the gas has been weighed. When therefore we say that "the nitrogen atom is trivalent," we sum up a large amount of knowledge which has been gained by laborious experiment. This classification of the elements into groups of equivalent atoms--which we owe to Frankland, Williamson, Odling, and especially to Kekule--has been of much service especially in advancing the systematic study of the compounds of carbon. It helps to render more precise the conception which has so long been gaining ground of the molecule as a definite structure.