hese structures
evidently consist of two semi-domes, unequal in size and separated by an
uncovered space. If such an arrangement was found convenient for a portable
and temporary dwelling like a tent, why should it not have been applied to
the permanent homes both of the king and his people?

Arches still standing in the city gates, fragments of vaults found within
the chambers of Sargon's palace at Khorsabad, the evidence of the
bas-reliefs and the existing methods of building in Mesopotamia--all concur
in persuading us that the vault played an important part in the
constructions of Assyria, and consequently in those of Chaldaea; but we
should not go so far as to say that all the rooms in the palace at
Khorsabad and elsewhere were covered with barrel vaults, domes, or
semi-domes. Our chosen guides, have, we think, allowed themselves to be a
little too absolute in this particular; it is quite possible that by the
side of the vaulted chambers there were others with wooden roofs. This
conclusion is suggested partly by Sir H. Layard's discovery of considerable
quantities of wood ashes in the palaces he excavated, partly by the
evidence of ancient texts that wood was often used throughout this region
to support the roofs at least of private houses. We may quote, in the first
place, some remarks in Strabo's account of Susiana, which the Greek
geographer borrowed from one of his original authorities: "In order to
prevent the houses from becoming too hot, their roofs are covered with two
cubits of earth, the weight of which compels them to make their dwellings
long and narrow, because although they had only short beams, they had to
have large rooms, so as to avoid being suffocated." This same writer, in
speaking of these roofs, describes a singular property of the palm-tree
beams. The densest and most solid of them, he says, instead of yielding
with age and sinking under the weight they have to support, take a gentle
upward curve so as to become better fitted than at first for the support of
the heavy roof.[214]

The necessity for the presence of a thick roof between the sun and the
inside of the rooms is here very clearly affirmed. It will also be noticed
that the general form of apartments in Susiana and Assyria did not escape
the observer in question. As he saw very clearly, the great disproportion
between their length and their width was to be explained as easily by the
requirements of a wooden roof as by those of a clay v