Oxide of load was added in some cases, not as a coloring matter, but as a flux, to facilitate the fusion of the glaze.

In other cases the pigment used was covered with a vitreous coat of an alkaline silicate of alumina.
The pigments were not applied to an entirely flat surface.

Prior to the reception of the coloring matter and the glaze, each brick was modelled by the hand, the figures being carefully traced out, and a slight elevation given to the more important objects.

A very low bas-relief was thus produced, to which the colors were subsequently applied, and the brick was then baked in the furnace.
It is conjectured that the bricks were not modelled singly and separately.

A large mass of clay was (it is thought) taken, sufficient to contain a whole subject, or at any rate a considerable portion of a subject.