the temperature of the air surrounding the glass be considerably increased. For this purpose the samples of glass are exposed to a current of air saturated with moisture at a temperature of about 80° C. in a specially arranged incubator for one or more days, means being provided for securing a constant stream of moist air during the whole time. On examining the glass surfaces after this exposure--any wiping or other cleaning of the surfaces being avoided--various qualities of glass are found to show widely varying appearances. The best and most stable glasses remain entirely unaffected; less stable kinds show small specks, which merge into a generally dulled surface in unstable kinds. There is no doubt that this test gives a sharp classification of glasses, but it yet remains to be proved that this classification agrees with their true relative durability in practice; the writer is inclined to doubt whether this is really the case, since certain glasses that have proved very satisfactory in this respect in practical use all over the world were classed among the less stable kinds by this test. Before leaving the subject of the chemical behaviour of glass, a reference should be made to the changes which glass undergoes when acted upon by light and other radiations. Under the influence of prolonged exposure to strong light, particularly to sunlight, and still more so to ultra-violet light, or the light of the sun at high altitudes, practically all kinds of glass undergo changes which generally take the form of changes of colour. Glasses containing manganese especially are apt to assume a purple or brown tinge under such circumstances, although the powerful action of radium radiations is capable of producing similar discoloration in glasses free from manganese. Apart from these latter effects, of which very little is known as yet, there can be no doubt that the action of light brings about chemical changes within the glass, but it is by no means easy to ascertain the true nature of these changes, although they most probably consist in a transfer of oxygen from one to another of the oxides present in the glass. Although it has not been definitely proved, it seems very unlikely that the glass either loses or gains in any constituent during these changes. Good examples of the changes undergone by glass under the action of sunlight are frequently found in skylights, where the oldest panes sometimes show a decided purple tint which they did not possess when first put in place. The glass spheres of the instruments used for obtaining records of the duration of sunshine at meteorological stations also show signs of the changes due to light--the glass of these spheres when new has a light greenish tint, but after prolonged use the colour changes to a decided yellow. The coloured glass in stained-glass windows also shows signs of having undergone changes of tint in consequence of prolonged exposure to