The composition of the window-glass of the thirteenth century is in some ways remarkable. It contained as much as from 8 to 10 per cent. of alumina, which we must regard as replacing in a measure the silica, for this constituent falls to as low as 56 per cent., and we can hardly otherwise account for the high percentage of the other bases—14 per cent. of lime, 17 per cent. of potash, and often 3 or 4 per cent. of iron. The result was a tough, somewhat horny glass, hard to work in consequence of the short duration of the viscous stage during the cooling. This was one reason for the smallness of the gatherings, and the modest dimensions of the resultant discs. On the other hand, such glass resists the action of the atmosphere better than any made nowadays, and the large amount of potash present probably promoted the brilliancy of the colours. From the earliest times the blue colouring was given by cobalt, and this was never of a richer and purer tint than in the twelfth century; already in the thirteenth copper was added to correct a tendency to purple. The famous ruby red, which became rarer after the thirteenth century until in the seventeenth the secret was entirely lost, was produced by the partial reduction of a small quantity of suboxide of copper, but in this case the colour is only developed on reheating the glass. The more purplish tint given by a somewhat similar treatment with gold was not known to the mediæval glass-maker.[95] Manganese was of course the source of the purple—the colour was used for flesh-tints in the twelfth century! The green was made by a mixture of the _æs ustum_ or copper scale with a native oxide of iron, the latter often known as _ferretto_—of this the best came from Spain. Finally, the yellow was given either by the sesqui-oxide of iron kept well oxidised by the presence of bin-oxide of manganese, or (where the surroundings favoured a reducing action) by a mixture of sulphur and some sooty material which probably yielded an alkaline sulphide. But in the older glass the yellow colour was never very brilliant; at a later time a fine yellow was obtained by a cementation process from silver, which was applied as a chloride or a sulphide to the surface of the glass. If I trespass beyond my limits to give this rapid summary of what is known of the colours of mediæval window-glass, it is because much of it will be found applicable to the contemporary Oriental enamelled ware and to the later Venetian glass.