The present condition of the ruins, the walls of which nowhere rise to the full height of the chambers, does not, however, exclude the possibility of openings for light having existed just beneath the ceiling. The form of such orifices cannot surely be determined; high windows could not have existed, and there must have been low openings in the top of the wall, separated by piers, between which stood small columns, as is evident from a relief of Coyundjic, given in Fig. 48 to serve as an argument for this manner of illumination. Light and air could thus have been freely admitted, without inconvenience to the dwellers within. The high position of the apertures, immediately under the somewhat projecting roof, prevented the entrance of rain, and shut off the interior from the view of those without, just as this same manner of lighting to-day protects the harems of the East. The small shafts, which were introduced as supports between these windows, appear to have been the only representatives of columnar architecture in the Assyrian palace. If columns had been used, in their customary function, as upholders of the roof,—as members which bore an important entablature,—some traces of these would certainly have been preserved; their material could hardly have been more perishable than the sun-dried brick of the walls. The entire arrangement of plan shows that their assistance was not relied upon. The chambers were disproportionately narrow, plainly to render it possible to cover them without the introduction of intermediate supports. The beauty and fitness of the corridor-like spaces were so sacrificed to this narrowness that its universal appearance can be regarded only as a constructive necessity. It is well illustrated by the cramped principal hall of the palace of Esar-haddon at Nimrud (Fig. 49), where a greater width than that permitted by the span of ceiling timbers was only to be obtained by the erection of a division wall to provide a subsidiary support for the beams. So helpless a make-shift, destroying the unity and grandeur of the hall, could have been adopted only in entire ignorance of the opening and supporting element of the column, apparently never recognized in Assyria.
The form of the small columns, which stood in the openings allowed for light in the upper walls, can be approximately determined from the representations upon reliefs. The shafts were cylindrical, and probably without flutings; they had a roundlet, or at least a projecting fillet, at either end. The base consisted solely of a high tore, sometimes notched upon the top, or placed upon the back of a striding lion. (Fig. 50.) The most common form of the capitals was a peculiar conjunction of two spiral scrolls, similar to a doubled Ionic capital, with an echinos-like roundlet beneath and a stepped abacus above. It is hardly to be doubted that this was the prototype of the Ionic capital, although it cannot be determined from the reliefs whether a lateral roll corresponded to the volute of the front, or whether the helix was repeated upon all four sides, as is the case with the capitals of Persian columns. The small scale of the representations upon reliefs, and their careless execution, do not permit a sure understanding of any part of the capitals. A table (Fig. 51) upon a relief of Coyundjic better determines the form of the volutes; it has distinct spirals in place of the rosettes, wrongly shown by Layard’s drawing.[D] There is reason to suppose that the double helix was not the primitive and normal form of the Assyrian capital, but was rather an abbreviation of the leaved calyx so frequently met with in Phœnicia, Palestine, and Cyprus, and that the rolled ends of the leaves, shown by two of the examples in Fig. 50, originally suggested the volutes of the capital and the various spiral forms occurring upon carved Assyrian furniture, as in Fig. 81. The question will be considered more at length in the section upon Syrian architecture.
The columns of Assyria were employed only in this subordinate position, and the dimensions and shape of larger enclosed spaces were dependent upon the limited span of the wooden ceiling beams. Assyrian palaces were, in these respects, unable to fulfil the demands of a monumental architecture. It can only be surmised how roof and ceiling were constructed in detail. The beams were naturally so placed as to require the least possible length to span the clear width; the sinking in the middle, to which the elastic trunks of palm-trees so much inclined, and the accumulation of water in the hollow thereby formed, were thus avoided as well as might be. The constructive details of the roof-platform are not surely known; it is probable that a layer of clay and earth was placed upon the beams, being rolled down compactly after every rain. The exterior representation of roof and ceiling, the wall entablature, may have consisted of a painted wooden sheathing, bearing ornaments of the character displayed by the pavement. (Fig. 45.) It was divided, like the Egyptian entablature, into two parts; in neither case was there a marked distinction between roof and ceiling. The imitations of building-fronts upon reliefs make it probable that stepped battlements rose above the main cornice.
