After the formulation of the periodic law and classification by Mendeleev, numerous attempts were made to express the law in rigid terms and the classification in a completely unambiguous form. Most of these, which have no longer even historical impor tance, served merely to demonstrate that the octet arrangement by valency is indispensable to the classification of the elements. The discovery of the inert gases, argon, etc., by Lord Rayleigh and Sir William Ramsay from 1894 onwards, introduced a new com plication into the periodic classification. The properties of argon indicated that it should be placed between the highly electro negative element, chlorine, and the highly electropositive element, potassium, that is, between a halogen element and an alkali metal. As the inert gases have no valency, Mendeleev proposed to create a new "zero" group to accommodate them, thus increasing the number of groups or families from eight to nine, despite the fact that several places in group VIII. were unoccupied. For example, neon is the sole element having atomic weight between that of fluorine in group VII. and that of sodium in group I., and could thus have been accommodated in one of the vacant spaces of group VIII. of Mendeleev's table. The other inert gases could also be fitted into the remaining vacancies of group VIII.
This classification of the elements into eight families or groups, having valency from one to eight respectively, presents certain anomalies. For example, only one element, hydrogen, precedes the inert gas helium of group VIII., though there are seven other groups. The inert gas argon has atomic weight greater than potas sium in group I., though it should have smaller atomic weight according to its position in the group VIII. preceding; the proper ties of cobalt indicate that it should precede nickel in the classi fication, though its atomic weight is slightly greater ; the atomic weight of iodine in group VII. is less instead of greater than that of tellurium in group VI. ; the triad of elements, iron, cobalt and nickel, require to be placed in group VIII., in order that univalent copper shall fall into group I., and similarly with the triads ruthenium, rhodium, palladium, and osmium, iridium, platinum. The anomalies relating to the position of hydrogen and that of these triads are inherent in the periodic classification. The anoma lies relating to atomic weight are, however, only apparent.
Mendeleev assumed that the periodicity of properties was a function which depended primarily on the magnitude of the atomic weights. It is certain that this is not strictly true, and that the order of the elements follows another function, the numbers of electrons in atoms, known as the atomic numbers (q.v.). These numbers are based on the theories of Sir Ernest Rutherford and Niels Bohr as to the structure of atoms. In Rutherford's theory
an atom consists of a central massive nucleus surrounded by nearly massless negatively charged electrons, sufficient in number to neutralize the positive nuclear charge. Bohr's theory involved a quantitative explanation of the lines in the spectra of atoms in terms of the nuclear charges, and was utilized by H. G. J. Moseley to determine the nuclear charges from X-ray spectra. Moseley's work showed that atoms, adjacent in the periodic classification, differ by unity in nuclear charge and therefore by one electron. Assuming that iron had 26 electrons as the 26th element in the chemical order, Moseley showed that cobalt had 27 and nickel 28 electrons, despite the fact that nickel has a smaller atomic weight than cobalt. The method was successfully applied to the other periodic anomalies, and it was proved that the true order of the elements in the classification is that of atomic number, not atomic weight. The periodic law consequently states that the properties of the elements are periodic functions of their atomic numbers.
The Abridged or Typical Classification.—Examination of the properties of the elements in the order of atomic number shows that the elements fall into series or periods of eight families having valency from one to eight respectively. The first series, consisting of hydrogen and helium, may be regarded as a rudi mentary or primitive period, hydrogen being the prototype of all the valent elements and helium the prototype of all the non-valent or inert elements. The eight elements succeeding helium, lithium to neon, form a period of eight groups from I. to VIII. The next eight elements, from sodium to argon, form a similar period of eight groups. These two octet periods may be regarded as gener alized types for the classification of the remaining elements. In stead of eight, however, there are 18 elements up to the next inert gas, krypton. Of these 18 only eight exhibit close resem blance to the preceding elements of the two short typical periods. These eight elements are potassium, calcium, gallium, germanium, arsenic, selenium, bromine and krypton—exactly sufficient to fill the eight groups of this 4th period. This arrangement leaves a gap of ten in atomic number between calcium (20) and gallium (31), by excluding the ten elements from scandium to zinc. These ten elements exhibit a gradual transition in properties from close resemblance to calcium on the part of scandium to close resemblance to gallium on the part of zinc. The ten excluded elements may consequently be regarded as a first transition series bridging a gap in atomic number and in properties between more typical elements of the period.