"The practical difficulty of cutting these sixteen faces with precision and of equalizing the angles at which they met each other, added to the natural desire to make the division into sixteen planes clearly visible, and to give more animation to the play of light and shade, inspired the Egyptian architects with the happy notion of transforming the obtuse angles into salient ridges by hollowing out the spaces between them."[95] The highest part, however, of these pillars remained quadrangular, thus preserving a reminiscence of the original type, and supplying a connecting link between the shaft and the architrave which almost exactly corresponds to the Greek _abacus_. This quadrangular member was advantageous in two ways; it prevented any incoherence between the diameter of the shaft and the depth of the architrave, and it supplied an unchanging element to the composition.[96] The persistence of this square abacus helps to call our attention to the continual changes undergone by the shaft which it surmounts. The slight inclination of the sides gives to the latter the effect of a cone, and the contrast between its almost circular top and the right-angles of the abacus helps us to remember that the square pier was its immediate progenitor. [95] EBERS, _Ægypten_, vol. ii., p. 184. [96] CHIPIEZ, _Histoire critique des Origines et de la Formation des Ordres Grecques_, p. 44. [Illustration: FIG. 72.--Octagonal pillar; Beni-Hassan.] [Illustration: FIG. 73.--Sixteen-sided pillar; fluted.] The conical form of the pillars at Beni-Hassan, their want of a well-marked base, their sixteen flutes, the square abacus interposed between their shafts and the architrave, made, when taken together, a great impression upon the mind of Champollion. He thought that in them he had found a first sketch for the oldest of the Greek orders, and that the type brought to perfection by the builders of Corinth and Pæstum had its origin in the tombs of Beni-Hassan; he accordingly proposed to call their columns _proto-doric_. Here we shall not attempt to discuss Champollion's theory. It would be impossible to do so with advantage without having previously studied the doric column itself, and pointed out how little these resemblances amount to. The doric column had no base; the diminution of its diameter was much more rapid; its capital, which comprised an echinus as well as an abacus, was very different in importance from the little tablet which we find at Beni-Hassan. The general proportions of the Greek and Egyptian orders are, however, almost identical; the shafts are fluted in each instance, and they both have the same air of simplicity and imposing gravity.