transition as an active and formative member. The apse of the lady chapel (Ill. 172) is trilobed, and the ribs are no longer purely decorative, but have really become structural members, since they are made to carry the three compartments of the vault which bulge upwards from them. The step to the fully developed chevet vault with ribs completely structural was an easy one, and was perhaps taken at Largny (c. 1140), or in the chevet of St. Germer (Ill. 173), which is of about the same date. A completely logical rib vault had now been evolved to replace the old half-dome, although many adjustments and perfections still remained to be made.

The chevet vault, as thus evolved in the apse, was immediately applied to the choirs of churches with ambulatory, which indeed offered precisely the same problem on a slightly larger scale. An immediate consequence of the introduction of the chevet vault in this position was the stilting of the wall rib. The width of the bays in the chevet was always less than in the straight portions of the church. Where the ambulatory existed, this was necessary in order to avoid making the outer wall arch of the ambulatory vault unduly wide, and in simple apses esthetic reasons seem to have led to the same result. Now from such narrow bays, to raise the crown of the narrow wall arch to the same level as the crown of the greatly wider transverse arch, there was need not only of pointing, but also of stilting. Thus the wall arch in the chevet came to be always highly stilted, and so generated those singularly graceful twisted surfaces that characterize the developed Gothic chevet vault (Ill. 173).

Stilting the wall rib, indeed, was no new idea, and had been employed in connection with vaults on a rectangular plan as early as c. 1125 at Dhuizel and elsewhere. In the fully developed Gothic, this construction was destined to become one of the most typical and strongly accentuated characteristics, not only of the chevet vaults, but of the nave vaults as well, and more than any other single feature to give rise to the peculiar form of the French Gothic vault. Furthermore, it was through this stilting that the structural usefulness of the wall rib in determining the form and shape of the vault was first fully demon

strated. From the moment when it became usual to stilt the wall rib, this member was seldom omitted.

There seems, therefore, no reason to doubt that the wall rib was stilted in the nave for the same reason that it was stilted in the chevet, i.e., to raise it to the desired level. Mr. Moore,1 however, believes that this rib was stilted in order to concentrate thrusts. If the wall ribs be stilted, it is obvious that the thrust of the vault will be brought to bear on a much smaller portion of the outside wall, than if the wall ribs be not stilted, since the conoid formed by the five converging ribs is much narrower in the first case than in the second. Thus by stilting the wall rib, the entire thrust of the vault is gathered on a single vertical line, where it can easily be neutralized by a flying buttress. No single feature was, therefore, more essential to the stability of the Gothic skeleton structure than this same stilted wall rib. I believe, however, that Mr. Moore, is mistaken in thinking that stilted wall ribs were developed in order to meet this structural necessity. Historical evidence shows that this stilting, which later became so integral a part of the structure of the building, grew up long before the flying buttress had been dreamed of, and was developed without other notion of structural importance than the raising of the crown of the wall rib. The apse of Largny, for example, where the problem of buttressing could as yet have had no weight, is supplied with fully stilted wall ribs.

A direct consequence of the development of the chevet vault was the substitution of a polygonal for a semicircular plan in the apses and chevets. It is true that this polygonal plan had appeared sporadically before the chevet vault was perfected. Outside the limits of the Ile de France it had already been employed in Limousin, Velay, Provence, and even Artois, where the polygonal apse of St. Ulmer of Boulogne is said to date from as early as the XI century. But in the Ile de France itself the earliest instance known is the church of Coudun (c. 1125), whose apse is polygonal internally, but semicircular externally. At Auvers, .however, side by side with one of the earliest examples of the ribbed half-dome (c. 1131) there

1 Gothic Architecture, p. 133.

is an apse polygonal externally (Ill. 174), and at Ciry is another example about contemporary. Nevertheless it was only in the second half of the century that this motive became thoroughly established.1

ILL. 174.

- Plan of Auvers. (From Arch. de la Com. des Mon. Hist.)

The lack of enthusiasm with which the polygonal apse was at first received must be attributed to the strength of the tradition that the east end of the church should be semicircular, for the As at Azy-Bonneil, Bussiares, Chassemy, Hautevesnes, Marigny-en-Orxois, Marisy-St.Mard, etc.