All lessonsLesson 10
Rules of thumb across the centuries
Lechler, Hoffstadt, Viollet-le-Duc, Hase and Ungewitter on one span.
Lessons 5 to 7 sized abutments by drawing the line of thrust through them. That is Ungewitter’s own method, and it is laborious. For centuries builders used something quicker: a rule tying the buttress to the span. The Lehrbuch collects several, from a late Gothic master’s instructions to the author’s own, and compares them. This lesson puts them side by side on one span and looks at what each takes into account.
Rules instead of statics
From the Lehrbuch
„So lange die Ergebnisse der Statik dem Praktiker nicht brauchbar oder handlich genug sind, müssen für ihn derartige Anhalte in der That als Ersatz dienen, mit Recht haben es daher auch neuere Meister für wichtig genug gehalten, geeignete Regeln aufzustellen.“As long as the results of statics are not usable or handy enough for the practitioner, such guides must indeed serve him as a substitute; recent masters have therefore rightly thought it important enough to set up suitable rules.Ungewitter 1890, p. 273Ungewitter does not despise rules of thumb. As long as statics gives the practitioner no handy results, he writes, such rules must serve as its substitute (p. 273). He reports five, pp. 273–275, and mentions a sixth in passing:
- Lechler (1516). Lorenz Lechler’s Unterweisungen derive a whole choir from its clear width. The Lehrbuch quotes only the buttress. Its projection beyond the wall is made by doubling the thickness, a tenth of the width, so the whole buttress comes to 0.3 L (p. 273).
- Hoffstadt (1840), from late Gothic manuscripts. Wall and buttress are each L/10 thick, and the buttress projects by the diagonal of the square on that thickness, nearly ¼ L in all, more exactly 0.2414 L (p. 273).
- Viollet-le-Duc, a method said to have been used still in the sixteenth century. Divide the intrados into three equal parts; the horizontal distance from the upper division point to the springing gives the depth (p. 273).
- Hase: the same thirds, plus 15 cm for every 4½ m of abutment height (pp. 273–274).
- The earlier editions of the Lehrbuch: fold the diagonal rib into the plan, add the thickness of rib and web, and draw a tangent (p. 274, Fig. 754).
- Ungewitter 1890, Mohrmann’s fifth rule, grounded on the statics of the book: ¼ L for a semicircular vault, down to ⅙ L for a steep one, plus 5 cm per metre of abutment height (p. 275).
The rules give 2.15 to 3 m: the deepest is 1.39 × the shallowest. Lechler: 3 m · 0.3 L; Hoffstadt: 2.41 m · 0.241 L; Viollet-le-Duc: 2.4 m · 0.24 L; Hase: 2.74 m · 0.274 L; Earlier editions: 2.15 m · 0.215 L; Ungewitter 1890: 2.72 m · 0.272 L.
- Lechler
- 3 m · 0.3 L
- Hoffstadt
- 2.41 m · 0.241 L
- Viollet-le-Duc
- 2.4 m · 0.24 L
- Hase
- 2.74 m · 0.274 L
- Earlier editions
- 2.15 m · 0.215 L
- Ungewitter 1890
- 2.72 m · 0.272 L
For a medium case the rules agree well, and Ungewitter says so. Compared with one another they show quite close agreement, and tested against old buildings or by calculation they fit well for average, “not too unfavourable” conditions (p. 274). Watch the shaded bands. Ungewitter compiled historical examples and found buttress lengths between 1/7 and 1/2 of the clear span, most often between ¼ and ⅓ (p. 276). For medium spans the rules land in or near the common band.
The spread grows at the ends of the range. Rules made of pure fractions, like Lechler’s and Hoffstadt’s, scale the buttress with the span. Hase’s rule and Ungewitter’s add a fixed amount for height, so for small spans they give relatively deep buttresses, as the book itself requires: small vaults under 5 m need 20 to 30 cm extra (p. 275).
What each rule sees
From the Lehrbuch
„Immer dürfen sie aber, wie ihre Urheber mit grosser Entschiedenheit aussprechen, nur als ungefähre Anhalte dienen …“But they may always serve only as approximate guides, as their authors state with great emphasis …Ungewitter 1890, p. 274„Als das beste der angegebenen Verfahren muss das dritte von C. W. Hase bezeichnet werden.“The third of the methods given, that of C. W. Hase, must be called the best.Ungewitter 1890, pp. 274–275A rule can only respond to what it measures. Lechler and Hoffstadt look at the span alone. Viollet-le-Duc’s thirds also see the form of the vault, because a steeper arch puts its upper third point closer to the springing. Hase adds height. Ungewitter’s rule sees span, rise and height, and its conditions add the material, the bay proportion and whether the wall helps. The next two figures change rise and height at a fixed span of 10 m.
From 2.15 to 3 m. Hoffstadt and Lechler ignore the rise altogether. Lechler: 3 m · 0.3 L; Hoffstadt: 2.41 m · 0.241 L; Viollet-le-Duc: 2.4 m · 0.24 L; Hase: 2.74 m · 0.274 L; Earlier editions: 2.15 m · 0.215 L; Ungewitter 1890: 2.72 m · 0.272 L.
- Lechler
- 3 m · 0.3 L
- Hoffstadt
- 2.41 m · 0.241 L
- Viollet-le-Duc
- 2.4 m · 0.24 L
- Hase
- 2.74 m · 0.274 L
- Earlier editions
- 2.15 m · 0.215 L
- Ungewitter 1890
- 2.72 m · 0.272 L
As the vault grows steeper its thrust falls (Lesson 5), and the rules that see the rise follow. Viollet-le-Duc’s thirds go from ¼ L for the semicircle to 2/9, ⅕ or ⅙ for a pointed arch, depending on its steepness (p. 273); Ungewitter’s fifth rule interpolates between ¼, 2/9 and ⅙. Lechler and Hoffstadt stay where they are. Above a rise of 2 : 3 the construction of the earlier editions drops out: the book itself warns that it would make the abutments too thin (p. 274).
From 2.15 to 3 m. Only Hase and Ungewitter add depth for height. Lechler: 3 m · 0.3 L; Hoffstadt: 2.41 m · 0.241 L; Viollet-le-Duc: 2.4 m · 0.24 L; Hase: 2.74 m · 0.274 L; Earlier editions: 2.15 m · 0.215 L; Ungewitter 1890: 2.72 m · 0.272 L.
- Lechler
- 3 m · 0.3 L
- Hoffstadt
- 2.41 m · 0.241 L
- Viollet-le-Duc
- 2.4 m · 0.24 L
- Hase
- 2.74 m · 0.274 L
- Earlier editions
- 2.15 m · 0.215 L
- Ungewitter 1890
- 2.72 m · 0.272 L
Height is where the rules part company most clearly. A tall wall gives the thrust a long lever, and a buttress that suffices at 6 m is too shallow at 25 m. Only Hase’s rule and Ungewitter’s grow with height. Viollet-le-Duc’s thirds are said to hold up to an abutment height of 1½ spans (p. 273); beyond that the figure warns. This is why Ungewitter rates Hase’s method the best of the older ones, and why his own rule is a refinement of it.
Rules and their limits
From the Lehrbuch
„Für mittlere Verhältnisse sind dieselben meist gut zutreffend, sie verlieren aber naturgemäss ihre Geltung, sobald besondere Fälle vorliegen, sie können dann sogar zu bedenklichen Irrungen führen.“For average conditions they mostly fit well, but they naturally lose their validity as soon as special cases arise, and can then even lead to serious errors.Ungewitter 1890, p. 128„Die richtige Feststellung der Widerlager ist wohl als die wichtigste Frage der ganzen mittelalterlichen Konstruktionslehre anzusehen.“The correct determination of the abutments is probably to be regarded as the most important question of the whole of medieval construction.Ungewitter 1890, p. 275The book’s verdict is balanced. Rules of thumb are useful for average cases and dangerous in special ones: very high walls, flat vaults, heavy webs, the pier buttresses of flying buttresses. For those Ungewitter sends the reader back to the line of thrust (p. 275). He adds a historical warning too. Overly rigid prescriptions for such dimensions are pointless, and they were unknown to the earlier Middle Ages (p. 125). The rules collected here are late: Lechler wrote in 1516, at the end of the Gothic, and Hoffstadt in the nineteenth century. They record experience; they did not create it.
This is why the Studio has a rule-set switch. The default is Ungewitter 1890. Switch to Hoffstadt, Viollet-le-Duc, Hase or the earlier editions, and every buttress of the plan is sized again by that rule.
Challenge
The Studio opens an aisleless church of 10 m span, sized by Hase’s rule. Switch between the rule sets and note the buttress depth each time. Then raise the abutment height to 25 m and repeat. Which rules would you trust for the tall church, and which would you check by tracing the line of thrust in the Lab?