In addition to being called upon to produce a large variety of different tints, the glass-maker is also called upon to produce various depths of the same tint. In many cases this can be readily done by the simple means of varying the amount of colouring material added to the glass. Where the colouring effect of small quantities of these substances is not excessively powerful there is no very great difficulty in doing this, but in certain cases this mode of regulating the intensity of the colour is not available. Thus copper-ruby glass cannot readily be made of so light a tint as to appear of reasonable depth when used in sheets of the thickness of ordinary sheet-glass. As has already been indicated, the desired tint is obtained by the process of “flashing,” _i.e._, of placing a very thin layer of deep ruby-coloured glass upon the surface of a sheet of ordinary more or less colourless glass of the usual thickness. This is generally accomplished by having a pot of molten ruby glass available close to a pot from which colourless glass is being gathered. A small gathering of ruby glass is first taken up on the pipe, and the remaining gatherings required for the production of the sheet are taken from the pot of colourless glass. When such a composite gathering is blown into a cylinder in the manner described in the previous chapter, the ruby glass lies as a thin layer over the inner face of the cylinder, but special care and skill on the part of the gatherer and blower is required to ensure that this layer shall be evenly distributed and of the right thickness to produce just the tint of ruby required. Since the whole layer of red glass is so thin, a very slight want of uniformity in its distribution leads to wide variations of tint, and in practice these are often seen in the less successful cylinders of such glass. The chemical composition of the ruby and the colourless glass which are to be employed for this purpose must also be properly adapted to one another in order to produce two glasses which shall have as nearly the same coefficient of thermal expansion as possible. If this requirement is not met, the resulting glass is subjected to internal strains which may lead to fracture, while, if the ruby glass has the higher coefficient of expansion, the sheet after flattening tends to draw itself up on the “flashed” side and cannot be passed out of the annealing kiln in a properly flat condition.