There can be little doubt but that both the porches, and the great central pillar-cluster, were roofed in.  The double-bull and double-griffin capital are exactly suited to receive the ends of beams, which would stretch from pillar to pillar, and support a roof and an entablature. [PLATE L., Fig.1.] We may see in the entrances to the royal tombs the true use of pillars in a Persian building, and the character of the entablature which, they were intended to sustain, Assuming, then, that both the great central pillar phalanx and the three detached colonnades supported a roof, the question arises, were the colonnades in any way united with the main building, or did they stand completely detached from it?  It has been supposed that they were all porticos in antis, connected with the main building by solid walls—­that the great central column-cluster was surrounded on all sides by a wall of a very massive description, from the four corners of which similar barriers were carried down to the edge of the terrace, abutting in front upon the steps of the great sculptured staircase, and extending eastward and westward, so as to form the antce of an eastern and a western portico.  In the two corners between the northern in antae of the side porticos and the antae of the portico in front are supposed to have been large guard-rooms, entirely filling up the two angles.  The whole building is thus brought into close conformity with the “Palace of Xerxes,” from which it is distinguished only by its superior size, its use of stone pillars, and the elongation of the tetrastyle chambers at the sides of that edifice into porticos of twelve pillars each.

[Illustration:  PLATE L.]

The ingenuity of this conception is unquestionable; and one is tempted at first sight to accept a solution which removes so much that is puzzling, and establishes so remarkable a harmony between works whose outward aspect is so dissimilar.  It seems like the inspiration of genius to discern so clearly the like in the unlike, and one inclines at first to believe that what is so clever cannot but be true.  But a rigorous examination of the evidence leads to an opposite conclusion, and if it does not absolutely disprove Mr. Fergusson’s theory, at any rate shows it to be in the highest degree doubtful.  Such walls as he describes, with their antae and their many doors and windows, should have left very marked traces of their existence in great squared pillars at the sides of porticos, in huge door-frames and window-frames, or at least in the foundations of walls, or, the marks of them, on some part of the paved terrace.  Now the entire absence of squared pillars for the ends of antce, of door-frames, and window-frames, or even of such sculptured fragments as might indicate their former existence, is palpable and is admitted; nor is there any even supposed trace of the walls, excepting in one of the lines which by the hypothesis they would occupy.  In front of the building, midway between