§ II. 1. The form of the aperture: and first of doors. We will, for the present, leave out of the question doors and gates in unroofed walls, the forms of these being very arbitrary, and confine ourselves to the consideration of doors of entrance into roofed buildings. Such doors will, for the most part, be at, or near, the base of the building; except when raised for purposes of defence, as in the old Scotch border towers, and our own Martello towers, or, as in Switzerland, to permit access in deep snow, or when stairs are carried up outside the house for convenience or magnificence. But in most cases, whether high or low, a door may be assumed to be considerably lower than the apartments or buildings into which it gives admission, and therefore to have some height of wall above it, whose weight must be carried by the heading of the door. It is clear, therefore, that the best heading must be an arch, because the strongest, and that a square-headed door must be wrong, unless under Mont-Cenisian masonry; or else, unless the top of the door be the roof of the building, as in low cottages. And a square-headed door is just so much more wrong and ugly than a connexion of main shafts by lintels, as the weight of wall above the door is likely to be greater than that above the main shafts. Thus, while I admit the Greek general forms of temple to be admirable in their kind, I think the Greek door always offensive and unmanageable. § III. We have it also determined by necessity, that the apertures shall be at least above a man's height, with perpendicular sides (for sloping sides are evidently unnecessary, and even inconvenient, therefore absurd) and level threshold; and this aperture we at present suppose simply cut through the wall without any bevelling of the jambs. Such a door, wide enough for two persons to pass each other easily, and with such fillings or valves as we may hereafter find expedient, may be fit enough for any building into which entrance is required neither often, nor by many persons at a time. But when entrance and egress are constant, or required by crowds, certain further modifications must take place. [Illustration: Fig. XLIII.] § IV. When entrance and egress are constant, it may be supposed that the valves will be absent or unfastened,--that people will be passing more quickly than when the entrance and egress are unfrequent, and that the square angles of the wall will be inconvenient to such quick passers through. It is evident, therefore, that what would be done in time, for themselves, by the passing multitude, should be done for them at once by the architect; and that these angles, which would be worn away by friction, should at once be bevelled off, or, as it is called, splayed, and the most contracted part of the aperture made as short as possible, so that the plan of the entrance should become as at _a_, Fig. XLIII.