Skip to content
Ungewitter

All lessonsLesson 1

Why arches stand

The line of thrust, as Ungewitter draws it, and the middle third.

An arch is a pile of wedge-shaped stones. The mortar between them holds almost nothing in tension, and yet a well-built arch stands for centuries. Why? Georg Ungewitter’s Lehrbuch der gotischen Konstruktionen (third edition, 1890) answers with a single drawing: the line of thrust, in German the Stützlinie. This lesson follows his construction step by step.

A hanging cord, turned upside down

Hang a cord under an arch and load it at every point exactly as the arch is loaded. The cord sags into a polygon, with a corner under each weight. Ungewitter calls it the Seilpolygon, the funicular polygon. Turned upside down it gives the arch form that suits those loads: the cord is all tension, so the inverted form is all compression.

„Wird unter einem Bogen ein Seil ausgespannt, das man genau so belastet wie den Bogen selbst, so wird es die theoretisch richtige Bogenlinie im Spiegelbild zeigen.“If a cord is stretched under an arch and loaded exactly like the arch itself, it will show the theoretically correct arch line as a mirror image.Ungewitter 1890, p. 45

This inverted cord is the line of thrust. It follows the force from stone to stone through the arch, down to the springing and into the abutment.

Drawing it: the force plan

Ungewitter does not hang cords: he draws. Divide the arch into voussoirs and find each one’s weight, which acts vertically at its centroid. In a separate figure, the Kräfteplan or force plan, lay off the weights one below the other to scale: the load line AB. Since the arch and its load are symmetric, draw a horizontal through the midpoint C of AB. Mark the horizontal thrust H on it, as a trial length, and its end is the pole P. Draw rays from the pole to the division points: each ray gives the force in one joint, in size and in direction. Draw a parallel to each ray between neighbouring weight lines, and the funicular polygon appears in the arch.

„Da H von vornherein nicht genau zu ermitteln ist, giebt man ihm zunächst probeweis eine Länge CP.“Since H cannot be determined exactly beforehand, it is first given a trial length CP.Ungewitter 1890, p. 52, Fig. 124

The trial is the point. In the figure below the line starts too steep and leaves the ring. Increase H, which moves the pole away from the load line, until the line fits. Push on, and it becomes too flat and leaves the ring again, low down near the springing.

▲▲▲▲▲✕✕✕✕✕✕✕✕✕✕✕✕▲▲▲▲W 138.4 kNW 138.4 kNPABCH = 39.2 kN
39 kN

H is too small: the line is too steep and leaves the ring. Line of thrust outside the ring at 6 of 13 ring joints: the arch cannot stand on this line; largest e : d = 0.72 at joint 4; thrust H = 39.2 kN, vertical load V = 132.7 kN.

H
39.2 kN
largest e : d
0.72
Figure 1Semicircular arch, span 8 m, ring 0.8 m. The crown point is fixed; only the pole distance H changes. The red dots mark where the line crosses each joint.

Only a narrow range of H keeps the line in the ring. Ungewitter puts it briefly: a high arch has a small thrust and a flat arch a large one (p. 52). A steeper line means a smaller thrust, a flatter line a larger one.

An arch stands while a line fits

The figure shows that an arch does not admit just one line of thrust but a whole family of them. Which one does the arch actually use? Ungewitter does not need to know. It is enough that one of them fits:

„… überhaupt hält man einen Bogen oder ein Gewölbe noch für haltbar, so lange noch eine befriedigende Drucklinie in ihm möglich ist.“… in general an arch or vault is still considered sound as long as a satisfactory line of pressure is possible within it.Ungewitter 1890, p. 52

If no line fits, the arch falls. It opens at its Bruchfugen, the joints of rupture, alternately at the intrados and the extrados, and always on the side away from the line (p. 53, Fig. 124c).

The middle third

Where the line crosses a joint decides how the joint is pressed. Through the middle, the pressure is the same everywhere: the force divided by the area. Through the third point, the edge nearer the line carries twice the average and the other edge nothing at all (p. 52). Beyond that the far edge would have to pull, which mortar cannot do, so the joint begins to open. Hence the rule of the kern: the line should stay in the middle third of every joint.

„Soll nach der üblichen Annahme die Mittellinie des Druckes überall in dem mittleren Drittel der Wölbdicke liegen, so bleibt kein grosser Spielraum …“If, by the usual assumption, the centre line of pressure is to lie everywhere in the middle third of the vault thickness, little latitude remains …Ungewitter 1890, p. 51

Now thin the ring. The figure always shows the most favourable line, the one that departs least from the middle line of the ring (p. 52). Watch it leave the kern, then the ring.

▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲▲W 142.6 kNW 142.6 kN
0.8 m

The most favourable line stays inside the ring but leaves the middle third: some joints open at one edge. Line of thrust outside the kern at 9 of 13 ring joints, but inside the ring; largest e : d = 0.3 at joint 5; thrust H = 52.3 kN, vertical load V = 132.7 kN.

d : R
0.182
largest e : d
0.3
Figure 2Semicircular arch of 8 m span, most favourable line of thrust. The tinted band is the kern, the middle third of the ring. d : R relates the thickness to the radius of the ring’s middle line.

At about d : R = 0.11 no line fits any more: the bare semicircle cannot be thinner. Engineers have known this limit since Milankovitch (1907), at 0.1075. Here it drops out of Ungewitter’s construction, applied with many voussoirs. To keep the line in the kern the bare ring needs about three times that thickness. That is why Ungewitter backs round arches with masonry above half their height (p. 54); the Lab leaves this backing out, so its rings must be thicker.

Try it

The Lab draws the same construction for any span, rise, thickness and crown load, and follows the line down through the buttresses. It checks every joint against the kern and shows the force plan beside the section. Try making the ring as thin as it can be while the buttress still holds.