Apart from this feature, the buildings do not affect much regularity. In
courts and facades, to a certain extent, there is correspondence; but in
the internal arrangements, regularity is decidedly the exception. The
two sides of an edifice never correspond; room never answers to room;
doorways are rarely in the middle of walls; where a rooms has several
doorways, they are seldom opposite to one another, or in situations at
all corresponding.

There is a great awkwardness in the communications. Very few corridors
or passages exist in any of the buildings. Groups of rooms, often
amounting to ten or twelve, open into one another; and we find
comparatively few rooms to which there is any access except through some
other room. Again, whole sets of apartments are sometimes found, between
which and the rest of the palace all communication is cut off by thick
walls. Another peculiarity in the internal arrangements is the number of
doorways in the larger apartments, and their apparently needless
multiplication. We constantly find two or even three doorways leading
from a court into a hall, or from one hall into a second. It is
difficult to see what could be gained by such an arrangement.

The disposition of the various parts of a palace will probably be better
apprehended from an exact account of a single building than from any
further general statements. For this purpose it is necessary to select a
specimen from among the various edifices that have been disentombed by
the labors of recent excavators. The specimen should be, if possible,
complete; it should have been accurately surveyed, and the survey should
have been scientifically recorded; it should further stand single and
separate, that there may be no danger of confusion between its remains
and those of adjacent edifices. These requirements, though nowhere
exactly met, are very nearly met by the building at Khorsabad, which