ece or Egypt are in question. We find none of
those well defined elements, those clear and precise pieces of information
which elsewhere allow us to obliterate the injuries worked by time and
human enemies. The foot of every wall is heaped about with such formless
masses of brick and brick dust, that it is almost impossible to make full
explorations or to take exact measurements. One must be content with an
approximation to the truth.

With the one exception of the staged tower at Khorsabad, we shall not
attempt to give a single restoration in the proper sense of the word. Not
that we mean to say that the different temple models given in our Plates
II., III., and IV., and in our Fig. 173, are creations of our fancy. No one
of the four pretends to reconstruct one famous building more than another.
They are abstract types, each representing, in its general features, one of
the varieties into which Assyro-Chaldaean temples may be divided. The
arrangements in which the originality of each type consists were only fixed
by M. Chipiez after long researches. In each case he has taken for his
point of departure either a Greek or Assyrian text, a sculptured relief, or
facts gleaned by the examination of original sites; in most cases he has
been able to supplement and correct the information gained from one of
these sources by that from another. He has thus entered into the spirit of
Mesopotamian architecture, and restored the chief forms it put on in its
religious buildings according to time and district. He cannot say that all
the details figured were found united, as they may be here, on a single
building; but they are not inventions, no one of them is without authority,
and the use to which they are put has been decided by the examination of
actual remains. We may say the same of proportions. These are the result of
study and of the collation of one ruin and one piece of evidence with
another; they have not been taken from any single building. Finally there
were certain details, such as the trace and elevation of the ramps, that
were full of difficulty. M. Chipiez arrived at the solution finally adopted
by an inductive process, by carefully weighing the obvious conditions of
the problem and choosing those arrangements by which its requirements
seemed most simply and conveniently met. In virtue of their general
character M. Chipiez's restorations reach a high degree of probability.
They may be compared, if we may use the expressi