All lessonsLesson 2
Pointed or round?
How the form of an arch changes its thrust, and why the Gothic masons chose the point.
In the twelfth century the masons of northern France gave up the round arch for the pointed one, and within a few generations it carried the whole Gothic vault. Fashion played its part, as did freedom of height: a pointed arch can rise as high as you like over a given span. Ungewitter adds a structural reason, and the line of thrust from Lesson 1 lets you see it.
One span, many arches
A two-centred pointed arch is struck from two centres on the springing line. With radius R over the span s, the ratio R : s fixes the form. At ½ both centres coincide at mid-span: the semicircle. At 1 each centre lies on the opposite springing: the equilateral arch, which circumscribes an equilateral triangle (p. 54). Above 1 the centres move outside the span and the arch becomes a lancet. The rise follows from f² = R·s − s²/4.
The figure keeps span and thickness fixed and changes only R : s. It always shows the most favourable line of thrust, with the force plan beside it.
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 = 0.38 at joint 5; thrust H = 38.4 kN, vertical load V = 97.3 kN.
- H
- 38.4 kN
- H : G
- 0.197
- largest e : d
- 0.38
Two things change. First the thrust: as the arch rises, H falls from about a fifth of the arch’s weight for the semicircle to about a twelfth for a steep lancet. In the force plan the load line moves towards the pole. This is Ungewitter’s rule again, that a high arch has a small thrust and a flat arch a large one (p. 52). For the builder it means slimmer abutments.
Second, the fit. Watch the largest e : d of the most favourable line, which measures how far it has to stray from the middle of the ring. It is worst for the semicircle and best around the equilateral arch, and it grows again for very slender lancets.
Closer to the line of thrust
Ungewitter draws the line of thrust of an arch of uniform thickness, a curve somewhat like a parabola, and lays the usual arch forms over it (Fig. 125). The semicircle and the stilted semicircle depart far from it, and even a great mass of masonry at the abutments brings the line only a little nearer. A pointed arch that is not too slender stays close everywhere: a little outside the line low down, a little inside it near the crown.
„Es steht eben keine einfache mit dem Zirkel geschlagene Bogenlinie der Stützkurve so nahe wie der Spitzbogen.“No simple arch line struck with the compass comes as close to the thrust curve as the pointed arch.Ungewitter 1890, p. 53, Fig. 125Here, for Ungewitter, lies the reason why the twelfth and thirteenth centuries took up the pointed arch so readily as they made their vaults thinner. A ring that follows its line of thrust closely can be thin; one that does not must be thick and heavy.
Slender arches want weight at the crown
The very slender lancet loses the advantage. Without a load at the crown its line of thrust is round at the top, while the arch is sharp, and the line slips out of the ring at the intrados just under the crown (p. 55). Ungewitter therefore loads slender arches at the crown: for the equilateral arch with ⅐ to ¼ of the arch’s weight, a heavy keystone or a strengthened ridge (p. 54), and for still more slender arches with even more. The right amount, he says, is easily found by trial construction of the line of thrust. Try it on a lancet:
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 = 0.38 at joint 10; thrust H = 25.3 kN, vertical load V = 152.8 kN.
- P : G
- 0
- largest e : d
- 0.38
A modest crown load, around a tenth of the arch’s weight here, pulls the line back towards the middle of the ring. More than that overdoes it, and the line bulges out at the haunches instead. In the Lab’s bare ring, without backing at the haunches, this optimum lies lower than Ungewitter’s figure for the equilateral arch, but the method is his.
Try it
In the Lab you can change the rise ratio, thickness and crown load together, and put the arch on buttresses to see how the smaller thrust of the pointed arch eases the line through the abutment.