(_b_) We may, _secondly_, regard the _Roman_ type of building as an intermediate form standing between that of Greek and Christian architecture, in so far as here we find mainly the application of arch and vaultings. It is not possible to determine with accuracy the time when the construction of arches was first discovered; it appears, however, certain that neither the Egyptians, despite the great progress they made in the arts of building, nor the Babylonians, Israelites, and Phoenicians were cognisant of the _ogive_ or the _vault._ The monuments of Egyptian architecture at any rate only show us that when it was a question of superimposing a roof over the interior of a building the one means the Egyptians had at their disposal was that of placing huge slabs of stone across like beams in horizontal position. If it was required to arch up broad entrances, or cross arches they knew of no other way of doing this than letting one stone on either side project forward, with another still more projecting one above it, so that the side walls gradually approached upwards until they reached a point where only one stone was necessary to close the remaining space between. Where such an expedient was not necessary they covered the spaces with huge slabs of stone arranged across in the manner of rafters. Among the Greeks we do, I believe, find monuments in which the arch construction has already been adopted, but they are rare; and Hirt, who has written with most authority over the building and the history of the building of antiquity, affirms that among such monuments we can rely on none with security as dating from a time previous to that of Pericles. In other words, in Greek architecture the features which are characteristic and elaborated are the column and beam in horizontal position, so that we find here the column very little used in a relation which lies apart from its true function, namely that of supporting beams. Moreover the arch that is vaulted from two piers or columns, and the knob-like formation, connotes a yet further feature, for we find here that the column already begins to forsake its determinate attribute of support. For the circular arch in its rise, its flexure and its declivity is related to a centre which has nothing to do with the column as a means of support. The separate parts of the circular arch are carried in mutual opposition; they support and prolong each other in a way that shows them far more remote from the direct assistance of the column than is the horizontally superimposed beam.