over the centres of their bases, and thus give unequal curvature to their sides. This would hardly have been tolerated had it been merely a preference for the form of the pointed arch which determined its use. These awkward arches are so plainly the result of a groping struggle with the difficulties of vaulting a curved oblong space, that they seem to show beyond question that the pointed arch was not introduced as an admired form, but that it came unsought
FIG. 11.
in the course of constructive experiment. The interest of this vault lies chiefly in what it exhibits of experiment in the application of new principles as yet but feebly apprehended. This groping procedure is shown again in the forms given to the plans of the diagonal ribs which are curved instead of being straight, in order, in some measure, to avoid the extreme one-sided position of the ridges of the longitudinal cells, which must pass through their intersection, and also to avoid, in some measure, the excessive inequality of the sides of the transverse arches. By this means, too, the inequality of magnitude in the transverse cells, e and f, is in some
