MERGE TO GYRENE. 441 concerned, were in any way materially violated, (at least, we may say, not in our opinion ;) and the eye is seldom offended by an appearance either of weakness or clumsiness in the columns, or of heaviness or insignificance in their entablatures. There is at the same time a good deal of variety in the disposition of the interiors, and the workmanship is usually very good, and occasionally, indeed very shaft. There was commonly a fillet dividing the channels, or fluting of the shaft, the proportion of which was not always the same, and we rarely saw any fluting where these were not adopted, and very seldom any columns where the shafts were left plain. The difficulty of preserving the edges of the fluting with nicety, and of keeping them from being chipped and broken, appears to have been the reason for adopting the fillet ; for as the proportions of the facades, particularly those of the interior ones, were necessarily on a small scale, the edges of the fluting, where no fillet was used, must have been nearly as sharp as the edge of a sword, and consequently very liable to accident. We may add that the width of the fillet accommodated itself to the entasis of the shaft, and was continued round the upper part of the channels, so as to form the crown of the hypotrachelium, when no annulets W'ere made use of ; for in that case the channels finished in these, forming au elegant curve from the line of the column to the lowest of the annulets, w'hich sometimes projected considerably from the upper part of the shaft. With regard to the disposition of the triglyphs with respect to the columns, we usually found them placed over the axes of the latter, with sometimes one, and sometimes two intervening, as we have already mentioned above ; with the exception, however, of those at the extremities of the zophorus, which were sometimes placed in the angle, and sometimes a little removed from it, being in the latter case placed over the joint centre of the half column and pilaster which usually terminated the fa9ade at both extremities. We must remark, with respect to the introduction of the pilaster conjointly with the columns at the angles, that the shafts and the capitals W'ere not wholly relieved from the surface, although they were more so than half their diameter. It must be recollected at the same time that the whole fa9ade was generally formed in the rock itself, and had consequently no w'eight to support, and no internal arrangements to which it was necessary that it should be accommodated. The placing of the triglyphs was therefore purely optional, and might be adapted to the taste or the fancy of the architect, who was thus enabled to follow his own ideas of proportion and arrangement, without reference to any standard but the eye.