PERIODIC LAW, THE. The classification of the elements, in the strict chemical sense, is a comparatively recent problem, for it was not until about the beginning of the 19th century that the two kinds of chemical substances, elements and compounds, were clearly differentiated. The early years of the century wit nessed a rapid development in analytical chemistry, the art of distinguishing different chemical substances, and the consequent building up of a vast body of knowledge of the chemical and physical properties of both elements and compounds. This rapid expansion of chemical knowledge soon necessitated classification, for on the classification of chemical knowledge are based not only the systematized literature of chemistry but also the laboratory arts by which chemistry is passed on as a living science from one generation of chemists to another. The elementary substances are few in number, rather less than a hundred, with distinctive properties, whereas compound substances are extremely numer ous, in the order of a million, with properties correspondingly di versified and often overlapping. The fact that many compounds contain elements in common renders classification easy, whereas elements, having no chemical components, are not readily ame nable to classification. It thus occurred that the classification of elements lagged many years behind that of compounds, and, in fact, no general agreement had been reached among chemists as to the classification of elements for nearly half a century after the systems of classification of compounds had become established in general use.
Probably the earliest steps towards the classification of the ele ments were taken by J. W. Dobereiner between 1817 and 1829, when he showed that certain triads of elements possessed close relationships between their members, the alkaline-earth metals, calcium, strontium, and barium, for example, having many prop erties in common, those of strontium being a mean between those of calcium and barium. Similarly in the halogen triad, the prop erties of bromine are a mean between those of chlorine and iodine, while in the alkali metal triad the properties of sodium are a mean between those of lithium and potassium. J. B. A. Dumas, L. Gmelin, E. Lenssen, Max von Pettenkofer, and J. P. Cooke, be tween 1828 and 1854, expanded Dobereiner's suggestions by showing that similar relationships were more extensive than be tween triads of elements, fluorine being added to the halogens, and magnesium to the alkaline-earth metals, while oxygen, sulphur, selenium, and tellurium were classed as one family, and nitrogen, phosphorus, arsenic, antimony and bismuth as another family of elements. • Many attempts were subsequently made to prove that the atomic weights of the elements were expressible by an arithmet ical function, and in 1862 A. E. B. de Chancourtois proposed a classification of the elements based on the new values of atomic weights (q.v.) consequent on Stanislao Cannizzaro's system of 1858. De Chancourtois plotted the atomic weights on a helical
curve such that corresponding points differed by 16, the atomic weight of oxygen. This curve brought closely related elements on to corresponding points, and he suggested in consequence that "the properties of the elements are the properties of numbers," a remarkable prediction in the light of the modern view that the elements are characterized by the natural numbers from unity on wards which express the numbers of electrons in atoms. (See ATOMIC NUMBERS.) In the following year, 1863, J. A. R. New lands proposed a system of classification of the elements in the order or atomic weights, the elements being assigned ordinal numbers from unity upwards and divided into seven groups having properties closely related to the first seven elements, hydrogen, lithium, beryllium, boron, carbon, nitrogen and oxygen. This sep tenary•relationship was termed the law of octaves, by analogy with the seven intervals of the musical scale. A somewhat similar classification was put forward by William Odling a year later, and in 1869 Dmitri I. Mendeleev, as a result of an extensive correlation of the properties and atomic weights of the elements, proposed the periodic law by which "the elements arranged ac cording to the magnitude of atomic weights show a periodic change of properties." The Periodic Classification of the Elements as proposed by Mendeleev in 1869 had few advantages over the classifications of Newlands and Odling, the families of elements in all three classi fications being arranged in horizontal rows. The principal virtue of Mendeleev's classification lay in the fact that he clearly recog nized that the outstanding periodic property of the elements is valency (q.v.). Valency is a number that expresses how many times the equivalent weight of an element is contained in its atomic weight. As the equivalent weight of an element is the weight of it which combines with, displaces from combination, or is otherwise chemically equivalent to unit weight of hydrogen, the valency of an element is simply the number of atoms of hydrogen that are equivalent to an atom of the element in chemical combining capacity. Mendeleev's classification was essentially a classification by valency, the elements falling into seven families with valency from one to seven respectively. This classification by valency was further emphasized by Mendeleev in 1871, when he put forward an improved form of the periodic table, in which the families of elements were arranged in eight vertical groups having valency from one to eight. This form of the table has continued unchanged to the present day, and on it is based prac tically the whole modern system of arrangement and classification of chemical knowledge. Moreover, on this periodic arrangement is ultimately based the bulk of scientific knowledge of the elec tronic structure of atoms.