Glass, Merret tells us, is ‘a concrete of salt and sand or stones.’ This, in modern scientific language, we should express by saying that it is a combination of silica with an alkali. But these substances alone are not enough. You cannot make a glass fit for practical use from a pure quartz sand with the addition of nothing else than a salt of potash or soda. Such a glass—a simple alkaline silicate—would indeed be transparent, but it would be difficult to work and very fragile. In all cases there is need of a second base, and this, to speak generally, should be either lime or oxide of lead. The latter base we may for the present neglect; speaking generally, it is the presence of lime that gives the working qualities and the requisite toughness. These, then, are the essential materials for the preparation of glass. Other substances may be present; alumina, for example, or one or other of the oxides of iron, but as a rule the presence of these latter bases is not desired—the glass would be better without them. Putting aside, then, for the present the glass in which lead is a constituent, as well as that in which the soda is replaced by potash, it is remarkable how little difference of composition we find in examples of glass of the most divergent origin. Let us compare the composition of a Roman ‘lachrymatory’ with that of a piece of modern English plate-glass. In a hundred parts we find— Silica. Soda. Lime. Iron Oxide. Alumina. Roman lachrymatory 71·5 16·5 8 1 2 English Plate-glass 72 17 6 2 2 These examples are indeed two extreme terms of a long but continuous series. A sample of Saracenic glass of the fourteenth or of Venetian glass of the sixteenth century, would yield on analysis much the same result.[1] This, then, may be regarded as the normal composition of such glass as I shall have to deal with in this history. The main question has generally been—How can the sand or silica, the premier element in glass, be best converted into a substance which shall in external aspect resemble as closely as possible the native rock crystal (itself pure silica), but which at the same time shall be not only fusible, but after fusing pass on cooling through a plastic condition when it may be expanded into a vesicle and otherwise worked up into various shapes? Long practical experience has shown that this can be best effected by adding to the sand materials containing both soda and lime, and as far as possible nothing beyond these bases. A glass thus compounded we may take as our normal type, but, as I have said, the soda may in certain cases be replaced by potash and the lime by lead oxide.