Skip to content
Ungewitter

All lessonsLesson 8

Ending the choir

Polygonal east ends: 3/6, 5/8, 7/12, and why 4/8 is awkward.

The straight bays of the choir have to end somewhere. A flat east wall is the simplest end. It is common in smaller and in Cistercian churches, but it gives up the vaulted polygon that draws the whole building towards the altar. Most Gothic choirs end in a polygon: n sides of a regular m-gon, written n/m. The choice looks like a matter of taste, yet it decides where the piers stand, whether there is an east window, and where the ribs can meet.

The usual endings

From the Lehrbuch

„Die an gotischen Werken am häufigsten vorkommenden Grundrissbildungen des Chorschlusses sind die nach 5 Seiten des Achtecks (Fig. 727) oder des Zehnecks (Fig. 728) oder nach 7 des Zwölfecks. Seltener findet sich der Abschluss nach 4 Seiten des Achtecks oder nach 6 Seiten des Zwölfecks (Fig. 729).“The most frequent plans of the east end in Gothic works are those on 5 sides of the octagon (Fig. 727) or of the decagon (Fig. 728), or on 7 of the dodecagon. Less often one finds the ending on 4 sides of the octagon or 6 sides of the dodecagon (Fig. 729).Ungewitter 1890, p. 259, Figs. 727–729
„Alle diese Polygonformen sind aus dem Halbkreis entstanden und unterscheiden sich zunächst danach, ob sie einem überhöheten Halbkreis umschrieben (Fig. 727) oder als genaue Polygonhälften dem Halbkreis einbeschrieben sind (Fig. 728).“All these polygonal forms arose from the semicircle and differ first in whether they are circumscribed about a stilted semicircle (Fig. 727) or inscribed in the semicircle as exact half polygons (Fig. 728).Ungewitter 1890, p. 259

Ungewitter sorts the endings into two families. A half polygon, such as 3/6, 4/8 or 5/10, is inscribed in a semicircle. Its centre lies on the line where the straight bays end. A polygon like 5/8 or 7/12 is wider than half. Its first sides run parallel to the axis, as if it were drawn around a stilted semicircle, and its centre lies a little inside the polygon. The difference matters for the vault. In the circumscribed family the ribs running to the centre close off on their own inside the polygon. The half polygon has its centre exactly on the last transverse arch, where no keystone can sit.

5/8
Ending

Circumscribed about a stilted semicircle: the ribs close within the polygon. A side faces the axis: the east window is kept. radius: 5.41 m; side: 4.14 m; greatest width: 10 m; angle spanned: 225°.

radius
5.41 m
side
4.14 m
greatest width
10 m
angle spanned
225°
Figure 1Choir 10 m wide with one straight bay. The polygon and its circumcircle come from the geom generator; ribs dashed, buttresses radial, the keystone marked. An accent dot is a shifted keystone, an accent ring a pier on the axis.

Moving the keystone

From the Lehrbuch

„… was bei jedem halben oder kleineren Polygonteil nur durch eine Verschiebung des Schlusssteins im Chorschluss möglich wird, wenn derselbe nämlich, wie Fig. 729 zeigt, aus dem eigentlichen Zentrum c des Polygons nach c′ gerückt wird.“… which, with any half or smaller part of a polygon, is possible only by shifting the keystone of the east end, namely when, as Fig. 729 shows, it is moved from the true centre c of the polygon to c′.Ungewitter 1890, pp. 259–260, Fig. 729
„Hierdurch erhalten aber die östlichen Rippen des Joches bei gleicher Höhe eine geringere Spannung als die westlichen und die ganze Anordnung den Charakter eines Auskunftsmittels.“As a result, however, the eastern ribs of the bay receive a smaller span than the western ones at the same height, and the whole arrangement takes on the character of a makeshift.Ungewitter 1890, p. 260

For the half polygons the book’s remedy is to move the keystone east, off the true centre. The generator puts it at the centroid of the polygon’s corners. The ribs then reach it, but the eastern ones are now shorter than the western ones while rising to the same height, and Ungewitter calls the result a makeshift. With three sides of the hexagon the shift goes so far that the ribs become those of a cross vault over a trapezoid (p. 260). Try 3/6 and 5/10 in the figure, and compare the dot with the centre ⊕ of the circle.

Why 4/8 is awkward

From the Lehrbuch

„Letztere Gestaltungen führen den Nachteil mit sich, dass in die Längenachse ein Pfeiler zu stehen kommt und das östliche Fenster verloren geht.“The latter forms bring the disadvantage that a pier comes to stand in the longitudinal axis and the east window is lost.Ungewitter 1890, p. 259

An odd number of sides puts a side across the axis, and in it the east window, the one the eye finds first from the west door. An even number puts a corner on the axis, and with it a pier and a buttress. 4/8 has both faults at once. It is a half polygon, so its keystone must be shifted. It has an even number of sides, so a pier blocks the axis and the east window is lost. 6/12 is half a dodecagon and shares both faults. That is why the book counts the two among the rarer endings.

The next figure keeps the octagon and adds sides one by one. Watch the classification change. Two or three sides are shallower than a semicircle; four are exactly half; five are circumscribed; six and more spread wider than the straight bays. Such widened endings, 6/8, 7/10 or 9/12, give the choir head a bulge that the book treats as a special case (p. 260, Figs. 731–732).

5/8
5/8

Circumscribed about a stilted semicircle: the ribs close within the polygon. A side faces the axis: the east window is kept. radius: 5.41 m; side: 4.14 m; greatest width: 10 m; angle spanned: 225°.

radius
5.41 m
side
4.14 m
greatest width
10 m
angle spanned
225°
Figure 2n sides of the octagon on the same 10 m opening. The radius grows as the polygon closes; from 6/8 the polygon is wider than the straight bay.

Where the polygon meets the straight bays

From the Lehrbuch

„Die solcher Art gefundene Länge des Strebepfeilers am Chorschluss ist auch für die geradlinige Verlängerung des Chores ausreichend, vorausgesetzt, dass die Jochlänge … nicht wesentlich grösser ist als eine Polygonseite.“The buttress length found in this way at the east end is also sufficient for the straight continuation of the choir, provided that the bay length … is not substantially greater than one side of the polygon.Ungewitter 1890, p. 274, Fig. 755

The polygon also sets the measure of the straight bays west of it. A choir bay about as long as one side of the polygon keeps the rhythm of piers and windows unbroken round the bend, and it lets the buttresses of the straight part take the same depth as those of the polygon. Ungewitter explains why this works. The straight buttresses carry a larger piece of vault than a polygon buttress, but there the thrusts of the wall arches cancel, whereas at the polygon they add up to a resultant on the buttress. For longer bays, and especially square ones, the buttresses must be strengthened (p. 274). That is why the figures draw one straight bay of about a polygon side. The Studio does the same with its straight-bay ratio of 1.

The junction itself needs care. Where the last straight bay meets the first side of a half polygon, the adjoining bay can pass almost the whole load of its vault onto one rib and the transverse rib. The pier there may then need reinforcing, as Ungewitter notes for the half decagon of Fig. 748 (p. 270). The circumscribed endings avoid this too. Their first sides run on in the line of the straight walls, and the junction pier stands like any other.

Setting it out

On site the ending is set out from the opening, not from the polygon. Draw the chord between the last two piers, find the centre of the m-gon whose n sides span it, strike the circle, and step off the sides. Then the buttresses: on each corner a radial one, on the line from the centre through the corner, and at the junction with the straight bays one square to the axis. The generator records these operations in order, and the figure replays them for 5/8.

1 / 14

Draw the opening chord of width W. Step: 1 / 14.

Step
1 / 14
Figure 3Construction of a 5/8 ending on a 10 m opening, step by step: chord, centre C, circle, sides, buttress lines. The current step is drawn in the accent colour.

Radial buttresses resist the thrust of the polygon’s ribs, which all run towards the centre. On the circumscribed 5/8 they stand evenly round the polygon. On a half polygon with its shifted keystone, the ribs no longer point along the radii, which is one more reason why Ungewitter prefers the circumscribed family.

Challenge

The Studio opens an aisleless church with a 4/8 east end. Change the polygon to an ending that keeps the east window and needs no shifted keystone. Which endings of the octagon, decagon and dodecagon pass both tests? Check your answer with Figure 1, and look at how the radial buttresses change.

Open a 4/8 choir in the Studio →