Such a combination could not long exist in architectonic systems in which the stone column played its true part. It is a survival from the use of wood. Another characteristic feature is the complete absence both from this fragment and from the columns in the sculptured reliefs of vertical lines or divisions of any kind, no trace of a fluted or polygonal shaft has been found.[259] In writing the history of the Egyptian column we explained how the natural desire for as much light as possible led the architects of Beni-Hassan to transform the square pier, first into an octagonal prism, secondly into one with sixteen sides.[260] And to this progressive elaboration of the polyhedric shaft the flutes seemed to us to owe their origin. On the other hand, with tall and slender supports such as those afforded by palm trunks no necessity for reduction and for the shaving of angles would arise, and those flutes whose peculiar section is owing to the desire for a happy play of light and shadow, would never have been thought of. If we imitate a natural timber shaft in stone we have a smooth cylindrical column like that seen in Fig. 74. Again, the shafts of the columns in the bas-reliefs, appear slender in comparison with those of Egypt, or with the doric shafts of the oldest Greek temples (see Fig. 41 and 42). In the fragmentary column from Khorsabad (Fig. 74) we have only a small part of the shaft but if we may judge from the feeble salience of the capital, its proportions must have been slender rather than heavy and massive. Wherever the stone column has been used in buildings of mediocre size, the architect seems to have been driven by some optical necessity to make his angle columns more thickset than the other supports. Thus it was in Assyria, in the little temple at Kouyundjik (Fig. 42), where the outer columns are sensibly thicker than those between them; at Khorsabad (Fig. 41) the same result was obtained by rather different means. The edifice represented in this bas-relief bears no little similarity to certain Egyptian temples and to the Greek temple _in antis_.[261] The strength of these angular piers contrasts happily with the elegance of the columns between them. The latter are widely spaced, and, as in some Egyptian buildings, the architrave is but a horizontal continuation of the corner piers. If we analyse the column and examine its three parts separately we shall be led to similar conclusions. The stone column no doubt bore the architrave upon its capital wherever it was used, and both in Chaldaea and Assyria we find the same arrangement in those light structures which we have classed as belonging to the architecture of the tent (Figs. 70 and 72). The origin of the forms employed in stone buildings is most clearly shewn by the frequent occurrence of the volute, a curvilinear element suggested by the use and peculiar properties of metal. [Illustration: FIG. 75.--Capital; from a small temple.]