In 1850 the number of substances generally recognised as chemical elements, in the sense in which that term was first employed by Boyle, was sixty-two. Two members—viz., the pelopium of Rose and the ilmenium of Hermann—were, however, subsequently shown to be identical with metals already known. At the present time (1910) the number of the chemical elements definitely recognised as such is eighty-two. In 1850, as now, they were broadly classified as metals and non-metals, although it was felt then, no less strongly than now, that no very clear line of demarcation was traceable between the two groups. Sixty years ago the elements usually styled non-metals were thirteen in number; to-day the number is nineteen—the increase being due to the inclusion of arsenic and the discovery of the so-called inactive elements, helium, argon, krypton, neon, and xenon. In 1850 there were forty-seven elements definitely classed as metals; in 1910 the number is sixty-three.

At all periods in the history of chemistry as a science the general tendency has been to name substances, whenever possible, in accordance with the theoretical conceptions of the time, and hence it has happened that the same body at successive periods has had very dissimilar names. But in naming the substances we term elements, theoretical conceptions are not usually applicable. Oxygen, it is true, derives its name from such a conception; and, etymologically, the name connotes an error. Hydrogen, too, has no more right to be called the water former than oxygen. Davy, who invented the term chlorine, advocated that the chemical elements should be named from some distinguishing peculiarity, either of origin or of physical property. In the main this principle has been adopted especially in later years although there are numerous instances of names derived from pure arbitrary sources. It is largely for the reason that the names of the elements are, with rare exceptions, unconnected with theories that they have remained unchanged, whereas names of compounds, which are far more frequently dependent upon speculative ideas, have constantly been altered in order to comply with the prevailing hypotheses of the period. At the same time it is not always clear that the etymology of certain of the elements is well ascertained. It has been recently shown, for example that the commonly accepted origin of the word “antimony” from antimoine, based on the alleged experiences of mediæval ecclesiastics has no valid foundation. The word is, in reality, derived from the Arabic alhmoud: this became latinised to althimodium and eventually to antimonium.

By the middle of the nineteenth century the system of symbolical notation suggested by Berzelius was everywhere current; and, stripped largely of its dualistic associations, this system still remains the most generally convenient method of expressing the composition, analogies, and numerical relations of substances. During the middle of the last century philosophic chemists, although subscribing, with hardly an exception, to the doctrine of definite combining proportions, were by no means agreed as to the sufficiency of Dalton’s explanation of the experimental laws of chemical combination; and the hypothesis of atoms in the Daltonian sense was not universally accepted. To some the atomic theory of Dalton, which assumed that the combining proportion was identical with the relative weight of the atom, was unnecessary as an explanation of the laws of combination. Or at most it was only one out of a variety of molecular conditions in which matter might exist. Consequently some chemists were in the habit of drawing a distinction between chemical atoms and physical atoms. The chemical atom was identical with the Daltonian atom but this was by no means the same as the physical atom of Democritus or Leucippus. The view in 1850, in fact, was not very dissimilar from that to which recent experimental inquiry has led. But it can hardly be said that the doubts were dependent upon valid experimental evidence; they arose rather from the erroneous interpretation of imperfectly ascertained facts—upon the supposed inconsistencies of the law of Gay Lussac with the hypotheses of Avogadro and Ampère. As soon as the facts were clearly perceived and the inconsistencies reconciled we heard less of the supposed distinction between the chemical and the physical atom. It is only within quite recent time, and as the result of entirely new lines of inquiry, that the distinction has been revived.

In the early part of the last century attempts were made by Berzelius to classify the chemical elements according to their electro-chemical relations, and by Thomson according as they were “supporters” or “non-supporters of combustion.” It was soon perceived that Thomson’s system had no philosophical basis, and it quickly fell into disuse. After the discovery of isomorphism, an endeavour was made by Graham to arrange the simple bodies in accordance with their natural relations, and even before 1850 the various elements were grouped by him very much as now.

This scheme of classification, somewhat modified by considerations of valency, and occasionally corrected by more accurate information concerning true analogies (as when vanadium was transferred by Roscoe to the nitrogen group), was in general use for practically a quarter of a century—in fact, until it was superseded by the gradual adoption of Mendeléeff’s arrangement based on periodicity. There can, however, be little doubt that this attempt by Graham at a natural classification paved the way along which Newlands and eventually Mendeléeff were led to devise our present rational system of grouping the chemical elements.

The numerical relationships existing among the equivalents and atomic weights of the elements of certain of these groups, pointed out by Dumas, Pettenkofer, Odling, Gladstone, and others, gave rise to much speculation. The values of the gradational differences, of course, depended upon whether equivalents or atomic weights were employed; but the immediate point is that, whichever basis was adopted, definite numerical relations were to be perceived. Thus, in the case of the group of the halogens, it was pointed out that the individual members are connected together as follows:

where a = 19; d = 16.5; d´ = 28.

Thus, too, in the case of the nitrogen group:

where a = 14; d = 17; d´ = 44.

On the basis of these and similar numerical relationships it was surmised that, just as the successive members of a group of homologous organic radicals are formed by increments of CH₂, so the substances in the several groups of the elements may be produced by successive additions of some form of matter common to them all. This has its counterpart, somewhat modified, in the modern hypothesis of the disintegration of the elements. Dumas conceived the elements in any particular group to be built up by successive accretions of particular forms of matter; Rutherford and Soddy suppose them to be derived by the successive elimination of matter from some unstable parent substance.

Since 1850 the existence of at least twenty-two new elements may be said to have been established. Of course, many more than this number have been announced, more or less tentatively; but subsequent investigation has either not confirmed their existence, or has definitely disproved it. The names, symbols, and atomic weights of the twenty-two, arranged in alphabetical order, are as follows:

The additions have been due, to some extent, to the refinement of processes of analysis already in use, but more especially to the employment of new analytical methods; or, lastly, to the application of a generalisation concerning the mutual relations of the elements which has served to indicate not only the existence of new and specific members of families of elements already known, but to point out the probable mode of their occurrence.2

Although the existence of the element fluorine was surmised as far back as 1771, when Scheele first recognised that the product of the action of oil of vitriol upon fluor-spar contained a hitherto unknown substance, it was not until 1886 that this substance was definitely isolated by Moissan by the electrolysis of the acid potassium fluoride in solution in hydrogen fluoride. Cerium tetrafluoride, CeF₄, and lead tetrafluoride, PbF₄, when heated, were observed by Brauner to evolve a gas having a smell resembling that of hypochlorous acid, which was probably free fluorine. Certain violet-coloured varieties of fluor-spar, when powdered, emit a peculiar smell, which has been attributed to free fluorine.

Gore observed that anhydrous hydrogen fluoride would not conduct electricity—a fact confirmed by Moissan. Moissan found, however, that on adding potassium fluoride to the liquid it readily suffered electrolysis with the liberation of free fluorine as a light greenish yellow gas with a pungent, irritating smell resembling that of hypochlorous acid. It has a vapour density corresponding with an atomic weight 19. By the application of cold and pressure it may be liquefied. At still lower temperatures it may be frozen to a white solid. Fluorine is characterised by an extraordinary chemical activity, and combines, even at ordinary temperatures, with a large number of substances. Sulphur, phosphorus, arsenic, antimony, boron, iodine, and silicon inflame or become incandescent in contact with it. It combines with hydrogen with explosive violence, even in the dark and at the lowest temperature. It unites also with the metals, occasionally with incandescence, and decomposes water with liberation of oxygen.

The application, by Bunsen, of the spectroscope to chemical analysis almost immediately resulted in his discovery, in 1860, of cæsium, and, in 1861, of rubidium. Cæsium was first detected in the mineral water of Dürkheim in the Palatinate and in the mineral petalite, by the two blue lines it forms in the spectrum, whence its name from the Latin cæsius, used to designate the blue of the clear sky. Rubidium was found in a lepidolite by means of a number of lines in different parts of the spectrum not previously observed, two being especially remarkable in the outermost region of the visible red portion—whence the name of the element from the Latin rubidus, used to designate the darkest red colour. The new metals were found to have the closest analogies to potassium, with which they usually occur associated in nature. Rubidium is found in a number of lepidolites, leucite, spodumene, triphylite, mica, and orthoclase, and in the Stassfurt carnallite; in sea-water and in many mineral waters. It occurs also in the ashes of many plants such as those of beetroot, tobacco, tea, coffee, etc. It is doubtful if it is a normal constituent of plant food, attempts to introduce it in place of potash having failed. It is not improbable that these elements would have remained unknown except for spectrum analysis. At all events, one of them—cæsium—was missed in 1846 by Plattner, in the course of the analysis of the mineral pollucite, in which it occurs to the extent of one third of its weight. After the discovery of cæsium by Bunsen, this mineral was again analysed by Pisani, when it was found that the alkali which Plattner had mistaken for potassium was in reality cæsium. Cæsium is found to a very small extent in many mineral waters, in a variety of minerals, and in the ashes of plants.

In 1861 Sir William Crookes made known the existence of a new element which he called thallium. He found it in a seleniferous deposit obtained from an oil of vitriol factory in the Harz. It was characterised by giving a bright green line in the spectroscope—whence its name from θαλλός, a green or budding twig. The discovery was confirmed in the following year by Lamy. Thallium, in its general chemical relations, has many analogies to the metals of the alkalis although in the metallic state it has the closest resemblance to lead. It occurs in many varieties of pyrites, in a few minerals, such as crookesite, lorandite, zinc-blende and copper pyrites, etc., and in certain mineral waters.

In 1863 Reich and Richter, by means of the spectroscope, detected the presence of a new element in the zinc-blende of Freiberg. The observation that it afforded two indigo-blue lines in the spark-spectrum led them to give it the name indium. It has since been found in numerous blendes, in various zinc and tungsten ores, and in many iron ores. It is a silver-white, ductile, and malleable metal, melting at 174°, and burning when heated with a violet flame. It is related in chemical characters to aluminium and zinc. Its true place in the natural scheme of classification of the elements was indicated by Mendeléeff.

In 1875 Lecoq de Boisbaudran discovered a new element in the zinc-blende of Pierrefitte in the Pyrenees, also by means of spectrum analysis. The spark-spectrum of its salts affords two characteristic violet lines quite different in position from those given by indium. To the new element its discoverer gave the name of gallium. It has been found in very small amounts in other blendes, but is still one of the rarest of the chemical elements. It is a bluish-white, hard, and slightly malleable metal fusing at a temperature not much higher than that of a hot summer day. Its existence and main properties, as well as its more significant chemical relationships, were predicted by Mendeléeff in 1869 from considerations based upon his periodic law. (See ante.)

In the same year Mendeléeff also predicted the existence of a new element belonging to the group of which boron is the first member, which he provisionally termed eka-boron, and described its main properties. Mendeléeff’s prediction was verified in 1879 by Nilson’s discovery of the element scandium. Scandium occurs associated with yttrium, ytterbium, etc., in many Swedish minerals, such as euxenite, gadolinite, yttrotitanite, etc. The metal itself has not been isolated, but the properties of its compounds correspond closely with those of the corresponding ekaboron compounds, as predicted by Mendeléeff.

A further illustration of the value of the principle of periodicity, as developed by Mendeléeff, in indicating the existence of new elements, is seen in the discovery of germanium. In 1885 Weisbach discovered a new Freiberg silver mineral, to which he gave the name argyrodite. This on analysis by Winkler was found to contain a new element to the extent of about seven per cent. with properties identical with those predicted by Mendeléeff for a missing element in the fourth group of the periodic series, consisting of silicon, tin, and lead, and which he had provisionally termed eka-silicon. Argyrodite, in fact, is a double sulphide of silver and germanium, 2Ag₂S.GeS₂. Germanium is a greyish-white, lustrous metal of sp.gr. 5.5., melting at about 900°, and resembling silicon and tin in its general chemical relations.

Dysprosium, europium, gadolinium, lutecium, neodymium, praseodymium, samarium, thulium, and ytterbium (neoytterbium) belong, like scandium, to the group of the so-called rare earth metals. These substances have been detected in a great variety of minerals, many of which are extremely rare. The elements most frequently occur in nature associated with yttrium, cerium, thorium, and zirconium.

Dysprosium was first detected, in 1886, by Lecoq de Boisbaudran in the so-called erbium earth of Mosander, in which Cleve had previously (1880) announced the existence of two other elements, holmium and thulium. There is some reason to believe that the holmium of Cleve is identical with dysprosium. Ytterbium was discovered by Marignac, in 1878, in the mineral gadolinite. In 1906 Auer von Welsbach announced that Marignac’s ytterbia was a mixture, which was confirmed in the following year by Urbain, who separated it into two elements, which he named neoytterbium and lutecium. Europium was discovered by Demarçay in 1901. All these earths are met with in small quantities associated with yttria in gadolinite, euxenite, samarskite, xenotime, cerite, orthite, and other similar minerals. Their compounds, or such of them as have been described, resemble the corresponding compounds of yttria. They are recognised by differences in their spectroscopic behaviour. Gadolinium was detected, independently, in 1886, by Marignac and Lecoq de Boisbaudran in the terbium earth of Mosander.

What was long known as didymium (διδυμος = a twin) was discovered by Mosander in 1841. It owes its name to its close chemical relationship to, and almost constant association with, lanthanum—both elements occurring in many minerals, more particularly in cerite, allanite, and monazite. In 1885 Auer von Welsbach announced that the didymium of Mosander was, in reality, a mixture of two elements which could be separated by the systematic fractional crystallisation of the double ammonium nitrates; to these elements he gave the names praseodymium (πράσινος, leek-green) and neodymium (νέος, new). Neodymium salts are rose-coloured, whereas those of praseodymium are green, and the elements are further characterised by differences in their absorption and spark-spectra. When mixed, the substances give the spectrum originally considered to be characteristic of didymium.

Samarium was discovered in 1879 by Lecoq de Boisbaudran in samarskite. Its salts are yellow, and afford in solution characteristic absorption bands.

It is not improbable that many of the minerals from which the so-called rare earths are obtained contain elements hitherto unrecognised, and it is possible that certain of the substances now assumed to be elements may, like didymium, turn out to be mixtures. In fact, additional elements have from time to time been announced, as for example, the decipium of Delafontaine (1878) and the monium or victorium of Crookes (1899), pronounced by Urbain to be identical with gadolinium: their individuality cannot as yet be said to be established. Didymium itself was stated by Krüss and Nilson (1888) to be even more complicated than the work of Auer von Welsbach would seem to indicate, and to contain no fewer than eight elementary substances. As yet, however, no confirmation of this surmise has been obtained.

The chemistry of the rare earths has of late years been greatly extended owing to the employment of certain of the members of the group in the manufacture of the “mantles” used in gas-lighting, and which consist substantially of thoria, mixed with about one per cent. of ceria. Large quantities of monazite, thorianite, thorite, cerite, and other minerals, are now worked up for the sake of the thoria and ceria they contain, and considerable amounts of residual products, consisting largely of other members of the family, are now available for investigation. It is reasonably certain, therefore, that our knowledge of this section of inorganic chemistry will be largely augmented in the immediate future. Indeed, the application of thoria to the construction of gas-mantles may be said to have removed that substance from the category of the rare elements. No sooner was it discovered that it was capable of useful application than unexpected sources of supply were found.

The same result has followed in other cases. One of the most significant developments of modern chemistry is seen in the efforts which are constantly being made to turn the so-called rare elements to useful account; and when they are found to be technically valuable it is generally observed that hitherto unknown sources of supply are soon available. Cerium salts have been found to be useful in the colouring of glass and porcelain, as mordants in dyeing, in photography, and in medicine. Zirconium has been used in incandescent electric lighting, and thallium has been employed in the manufacture of highly refractive optical glass. Titanium, molybdenum, and vanadium are used in the manufacture of steel of high tensile strength. Tantalum and tungsten are employed in the construction of filaments in incandescent electric lighting. Tantalum, indeed, has been found to occur in considerable quantities, and to be more largely distributed than was hitherto supposed. Alloys of tungsten and aluminium are used in automobile construction, and alloys of tungsten, aluminium, and copper in the manufacture of propeller blades. Tungsten steel is used in armour plates, and to stiffen the springs of cars; in the manufacture of piano-wire, and to increase the permanency of magnets. Even the rarer metals of the platinum group are finding many important applications. Osmium-iridium is used for the bearings of compasses, for the tips of gold pens, and in the construction of standard weights. Osmium and ruthenium enter into the composition of filaments for electric lighting. The extraordinary influence of light on the electric conductivity of selenium has been made use of in the transmission of photographs by telegraph and telephone wires, and for measuring the light intensity of the Röntgen rays in clinical work.