The final condition is represented at E, in its relation to that at D. It will be observed that though each circle projects a little beyond the side of the square out of which it is formed, the space cut off at the angles is greater than that added at the sides; for, having our materials in a more concentrated arrangement, we can afford to part with some of them in this last transformation, as in all the rest. § V. And now, what have the base and the cornice of the wall been doing while we have been cutting the veil to pieces and gathering it together? The base is also cut to pieces, gathered together, and becomes the base of the column. The cornice is cut to pieces, gathered together, and becomes the capital of the column. Do not be alarmed at the new word, it does not mean a new thing; a capital is only the cornice of a column, and you may, if you like, call a cornice the capital of a wall. We have now, therefore, to examine these three concentrated forms of the base, veil, and cornice: first, the concentrated base, still called the BASE of the column; then the concentrated veil, called the SHAFT of the column; then the concentrated cornice, called the CAPITAL of the column. And first the Base:-- [Illustration: Fig. X.] § VI. Look back to the main type, Fig. II., page 55, and apply its profiles in due proportion to the feet of the pillars at E in Fig. IX. p. 72: If each step in Fig. II. were gathered accurately, the projection of the entire circular base would be less in proportion to its height than it is in Fig. II.; but the approximation to the result in Fig. X. is quite accurate enough for our purposes. (I pray the reader to observe that I have not made the smallest change, except this necessary expression of a reduction in diameter, in Fig. II. as it is applied in Fig. X., only I have not drawn the joints of the stones because these would confuse the outlines of the bases; and I have not represented the rounding of the shafts, because it does not bear at present on the argument.) Now it would hardly be convenient, if we had to pass between the pillars, to have to squeeze ourselves through one of those angular gaps or brêches de Roland in Fig. X. Our first impulse would be to cut them open; but we cannot do this, or our piers are unsafe. We have but one other resource, to fill them up until we have a floor wide enough to let us pass easily: this we may perhaps obtain at the first ledge, we are nearly sure to get it at the second, and we may then obtain access to the raised interval, either by raising the earth over the lower courses of foundation, or by steps round the entire building. Fig. XI. is the arrangement of Fig. X. so treated. [Illustration: Fig. XI.]