Molten glass may be regarded as a mutual solution of a number of chemical substances--usually silicates and borates. When cooled in the ordinary way these bodies remain mutually dissolved, and ordinary glass is thus simply a congealed solution. The dissolved substances have, however, natural freezing-points of their own, and if the molten mass be kept for any length of time at a temperature a little below one of these freezing-points, that particular substance will begin to solidify separately in the form of crystals. The facility with which this will occur depends upon the properties of the ingredients and upon the proportions in which they are present in the glass. In some cases this devitrification sets in so readily that it can scarcely be prevented at all, while in other cases the glass must be maintained at the proper temperature for hours before crystallisation can be induced to set in. In either of these cases, provided that the glass is cooled sufficiently rapidly to prevent crystallisation, the sequence of events during the subsequent cooling of the mass is this: as the temperature falls further and further below the natural freezing-point of one or other of the dissolved bodies, the tendency of that body to crystallise out at first rapidly increases; as the temperature falls, however, the resistance which the liquid presents to the motion of the molecules increases at a still greater rate, so that two opposing forces are at work, one of them an increasing tendency towards crystallisation, the other a still more rapidly increasing resistance to any change. There is thus for every glass a certain critical range of temperature during which the greatest tendency exists for the crystallising forces to overcome the internal resistance; through this range the glass must be cooled at a relatively rapid rate if devitrification is to be avoided; at lower temperatures the crystallising forces require increasingly longer periods of time to produce any sensible effect, until, as the ordinary temperature is approached, the forces of internal resistance entirely prevent all tendency to crystallisation. The phenomena just described in reality constitute the natural limit to the range of bodies which can be obtained in the vitreous state: as we approach this limit the glass requires more and more rapid cooling through the critical range of temperature, and is thus more and more liable to devitrify during the manufacturing processes, until finally the limit is set when no industrially feasible rapidity of cooling suffices to retain the mass in the vitreous state.