The preparation of the surface about to be decorated demanded much time and care.  Seeing how imperfect were the methods of construction, and how impossible it was for the architect to ensure a perfectly level surface for the facing stones of his temple-walls and pylons, the decorator had perforce to accommodate himself to a surface slightly rounded in some places and slightly hollowed in others.  Even the blocks of which it was formed were scarcely homogeneous in texture.  The limestone strata in which the Theban catacombs were excavated were almost always interspersed with flint nodules, fossils, and petrified shells.  These faults were variously remedied according as the decoration was to be sculptured or painted.  If painted, the wall was first roughly levelled, and then overlaid with a coat of black clay and chopped straw, similar to the mixture used for brick-making.  If sculptured, then the artist had to arrange his subject so as to avoid the inequalities of the stone as much as possible.  When these occurred in the midst of the figure subjects, and if they did not offer too stubborn a resistance to the chisel, they were simply worked over; otherwise the piece was cut out and a new piece fitted in, or the hole was filled up with white cement.  This mending process was no trifling matter.  We could point to tomb-chambers where every wall is thus inlaid to the extent of one quarter of its surface.  The preliminary work being done, the whole was covered with a thin coat of fine plaster mixed with white of egg, which hid the mud-wash or the piecing, and prepared a level and polished surface for the pencil of the artist.  In chambers, or parts of chambers, which have been left unfinished, and even in the quarries, we constantly find sketches of intended bas-reliefs, outlined in red or black ink.  The copy was generally executed upon a small scale, then squared off, and transferred to the wall by the pupils and assistants of the master.  As in certain scenes carefully copied by Prisse from the walls of Theban tombs, the subject is occasionally indicated by only two or three rapid strokes of the reed (fig. 178).  Elsewhere, the outline is fully made out, and the figures only await the arrival of the sculptor.  Some designers took pains to determine the position of the shoulders, and the centre of gravity of the bodies, by vertical and horizontal lines, upon which, by means of a dot, they noted the height of the knee, the hips, and other parts (fig. 179).  Others again, more self-reliant, attacked their subject at once, and drew in the figures without the aid of guiding points.  Such were the artists who decorated the catacomb of Seti I., and the southern walls of the temple of Abydos.  Their outlines are so firm, and their facility is so surprising, that they have been suspected of stencilling; but no one who has closely examined their figures, or who has taken the trouble to measure them with a compass, can maintain that opinion.  The forms of some are slighter