We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 11 April 2016
Many attempts have been made to determine how the Westminster Hall roof works, and nearly as many different conclusions have been reached (Figs 1, 3, and 4). Some architects, engineers, and architectural historians have concluded that the principal support of the roof is provided by its great wooden arch. Some have argued that the roof functions like other hammerbeam roofs, but they have not been able to agree on how a hammerbeam roof works. Still others have held that so unprecedented a space could only have been spanned with wood by combining two different types of roofs — the hammerbeam roof and the arched roof. The Westminster Hall roof is so complex that it appears as if it might function in any one of these ways, but it cannot function in more than one.
1 Attempts by engineers to determine how the roof works are evaluated in Heyman, Jacques, ‘Westminster Hall Roof’, Proceedings of the Institution of Civil Engineers, 37 (1967), pp. 137–62 CrossRefGoogle Scholar; in Morris, E. Toby, Black, R. Gary, and Tobriner, Stephen O., ‘Report on the Application of Finite Element Analysis to Historic Structures’, Journal of the Society of Architectural Historians, 54 (1995), pp. 336–47 CrossRefGoogle Scholar; and in Mainstone, Rowland J., ‘Structural Analysis, Structural Insights, and Historical Interpretation’, J.S.A.H., 56 (1997), pp. 316–40 Google Scholar.
2 Herland’s career is outlined in Harvey, John, English Mediaeval Architects (Gloucester, 1984), pp. 137–41 Google Scholar, and in the same author’s ‘The Mediaeval Carpenter and His Work as an Architect’, Journal of the Royal Institute of British Architects,3rd Series, 45 (13 June 1938), pp. 736-39 and 743.
3 Smirke, Sydney, ‘Remarks on the Architectural History of Westminster Hall…’, Archaeologia 26 (1836), pp. 406–14 CrossRefGoogle Scholar; and ‘Second Letter from Sydney Smirke…’, Archaeologia, 27 (1837), pp. 415-21. The quotations are from the ‘Second Letter’, pp. 417 and 418. Sydney Smirke was the architect and engineer who later designed and constructed the dome of the reading room of the British Museum.
4 Baines, F., Westminster Hall: Report to the First Commissioner of H. M. Works, &c., on the Condition of the Roof Timbers of Westminster Hall, with Suggestions for Maintaining the Stability of the Roof (London, 1914), pp. 22 and 48 Google Scholar.
5 These three examples are illustrated in Howard, F. E. and Crossley, F. H., English Church Woodwork: A Study in Craftsmanship During the Medieval Period, A.D. 12–1500 (London, 1919), pp. 96, 97, and 100Google Scholar. Similar triangular feet were widely used in continental Europe during the eleventh and twelfth centuries ( Courtenay, Lynn T., ‘Where Roof Meets Wall: Structural Innovations and Hammer-Beam Antecedents, 1150-1250’, Annals of the New York Academy of Sciences, 441 [1985], pp. 89–124)CrossRefGoogle Scholar.
6 Pearson, Sarah, Barnwell, P. S., and Adams, A. T., A Gazetteer of Medieval Houses in Kent (London, 1994), fig. 97, section B-B1Google Scholar.
7 Brandon, Raphael and Brandon, J. Arthur, The Open Timber Roofs of the Middle Ages (London, 1849 Google Scholar; reprinted, Prince George, BC, 1977), pp. 22-23. The Brandons also pointed out that the wallpost functions to prevent spreading rather than being a loadbearing member (pp. 20-22).
Harvey, John concluded that with the hammerbeam roof, ‘a system of triangulation made it possible for principal rafters, thus stiffened, to bridge large spans without intermediate supports’ (The Medieval Architect [London, 1972], p. 135 Google Scholar).
8 Villard de Honnecourt understood the potential of this triangular foot and showed how it could be cantilevered into an interior ( Bowie, Theodore (ed.), The Sketchbook of Villard de Honnecourt [Bloomington, Indiana, and London, 1959], pl. 58 Google Scholar).
9 In the King’s Works, for example, it is stated, ‘actually the Westminster roof incorporates two principles of timber construction — the arch-brace and the hammer-beam, which here interpenetrate…’ (Brown, R. Allen, Colvin, H. M., and Taylor, A.J., History of the King’s Works: the Middle Ages [London, 1963], 1. p. 530 Google Scholar).
10 Courtenay, Lynn T., ‘The Westminster Hall Roof and Its 14th-century Sources’, J.S.A.H. 43 (1984), pp. 308–09 Google Scholar and fig. 18. Courtenay, L. T. and Mark, R., ‘The Westminster Hall Roof: A Historiographic and Structural Study’, J.S.A.H., 46 (1987), pp. 374–93 Google Scholar.
11 Occasionally, a triangular configuration has its interior space filled solid as for Tunstead Church, Norfolk (Brandon and Brandon, Open Timber Roofs, p. 21).
12 Baines noted that the Eltham Hall roof had little in common structurally with the roof of Westminster Hall (Westminster Hall, p. 26, fig. 1). A. C. Pugin showed in measured drawings that the roof of the Great Hall at Hampton Court Palace likewise has very little in common (Specimens of Gothic Architecture Selected from Various Ancient Edifices in England… (London, 1821-23), vol. 2, pl. 8).
13 The surviving building accounts are summarized in Brown et al, King’s Works, 1, pp. 529-33.
14 The new masonry was designed by the architect Yevele, Henry (Harvey, John H., Henry Yevele, c. 1320 to 1400: the Life of an English Architect [London, 2nd ed.; 1946], pp. 47–49 Google Scholar; figs 23 and 57). The flying buttresses show in photographs reproduced in Miele, Chris, ‘The Battle for Westminster Hall’, Architectural History, 41 (1998), figs 9 and 10Google Scholar.
15 About one-third of the roof was replicated between 1913 and 1922 under the direction of the architect Frank Baines, whose report furnishes the most complete and reliable data on the roof ( Baines, , Westminster Hall, p. 6 Google Scholar). Baines estimated that a minimum of about 35-40 per cent of the roof needed to be replaced in 1914 and that to make the roof entirely sound without steel reinforcement would require the replacement of 70-80 per cent of all timbers (mainly because their joints had deteriorated so completely; ibid., p. 36). Steel reinforcement was used, and all parts of the exterior surfaces of timbers were saved whenever possible. For a summary of his unpublished records of the restoration, see Courtenay, Lynn T., ‘The Westminster Hall Roof: a New Archaeological Source’, Journal ofthe British Archaeological Association, 143 (1990), pp. 95–111. Google Scholar.
The most extensive replacement of timbers known to have been made before Baines’s work was by Soane, John in c. 1820 (Britton, J[ohn] and Brayley, [Edward Wedlake], Palatial and Parliamentary Edifices of Westminster [London, 1835], p. 441)Google Scholar. Soane added massive struts from the lower principal rafter to the wallpost, and Baines removed these struts. Baines also removed the extensive metal reinforcements which had probably been added by Barry, Charles in c. 1850 (Baines, , Westminster Hall, p. 68 and Drawing no. 1Google Scholar). While they existed, the struts and straps enabled various parts of the continually deteriorating roof to function as needed in unintended ways. With Baines’s own extensive, but better concealed, reinforcement, any single element could fail entirely without endangering the roof.
16 Smith, J. T., ‘Medieval Roofs: a Classification’, in Smith, J. T., Faulkner, P. A., and Emery, Anthony, Studies in Medieval Domestic Architecture, ed. by Swanton, M. J. (London, 1975), p. 72 and 74Google Scholar.
17 A good idea of the maximum lengths of timber available in England from the eleventh through the fourteenth centuries can be gained from the spans indicated for cathedral naves and choirs in Gwilt, Joseph, An Encyclopaedia of Architecture: Historical, Theoretical, & Practical, rev. by Papworth, Wyatt (London, 1903), pp. 187–97 Google Scholar. The average of all examples cited is about 34 ft, and roughly another 3 ft of length was needed to rest on top of the walls to each side. Timbers of about 40 ft were thus about the maximum used for most buildings constructed throughout the entire period. There was little variation from century to century, and very few spans were wider.
18 Harvey, W., ‘Westminster Hall Roof and the Woodman’, ‘Westminster Hall Roofs, Old and New’, and ‘Westminster Hall Roof and Mechanical Science’, The Builder, 121 (1921), pp. 220–21, 374-75 and 440-41Google Scholar respectively.
19 For an example of a grid used by a medieval architect, see Branner, Robert, ‘Drawings from a Thirteenth-Century Architect’s Shop: the Reims Palimpsest’, J.S.A.H., 17 (1958), fig. 3Google Scholar.
20 Baines, Westminster Hall, Drawing no. 4.
21 The inside surface of the Great Arch is 41 ft 11½ in high and 64 ft 67/8 in wide (Baines, Westminster Hall, Drawing no. 1). The height is thus 65 per cent of the width, which is sufficiently close (within 2 per cent) to two-thirds to be intentional. The Arch was probably still closer to two-thirds before some settling took place.
22 To construct a third-point arch, a Gothic architect divided a base line into thirds, and including each end, this created four points. He used two-thirds of the length of the base line (three points) for a radius. He constructed the left-hand segment of the arch by placing the point of a compass on the base line two-thirds of the way from its left side, and he constructed the right-hand segment by using a point two-thirds of the way from the right side. For a fourteenth-century design showing the construction of a three-point arch, see Harvey, John, Yevele, fig. 7 (the uppermost curves)Google Scholar.
23 The present length of the English foot was standardized at 30.48 cm in the reign of Henry I (1100-35). The Norman foot had previously been 29.78 cm (about a quarter-inch less; Harvey, , The Medieval Architect, 108 Google Scholar).
24 Baines shows the length of the hammerbeams as just over 20 ft (measured using his scale; Westminster Hall, Drawing no. 1), yet he provided Cescinsky and Gribble with a length of 17 ft 9 in ( Cescinsky, Herbert and Gribble, Ernest R., ‘Westminster Hall and Its Roof’, The Burlington Magazine, 40 [1922], p. 83 Google Scholar).
25 In 1396 Herland was granted a small house near the Palace of Westminster ‘for keeping his tools and for making his models (formae, formulae) and moulds for his carpentry work, which house was delivered to him by order of the late King 30 years ago…’ ( Baines, , Westminster Hall, p. 5 Google Scholar). Although Baines translated the word ‘formae’ and ‘formulae’ as ‘models’ in this case, he translated ‘forme’ as ‘patterns’ in a 1395 document (his p. 4). Salzman translates the entire phrase forme et molde from the 1395 document simply as ‘design’, and this legal phrase is all but certain to have meant an essentially two-dimensional pattern rather than a three-dimensional model ( Salzman, L. F., Building in England Down to 1540: a Documentary History [Millwood, 1979 reprint], p. 472 Google Scholar). For other examples of the use of form and mould, see Shelby, L. R., ‘Role of the Master Mason in Medieval English Building’, Speculum, 39 (1964), pp. 393–94 CrossRefGoogle Scholar.
26 Timoshenko, Stephen P., History of Strength of Materials (New York, 1953 Google Scholar; 1983 reprint), p. 6. Mainstone, Rowland J., Developments in Structural Form (Harmondsworth [1975], p. 284)Google Scholar.
27 For clear diagrams of this structure, see Hewett, C. A., English Cathedral Carpentry (London, 1974), figs 74–76 Google Scholar. Hewett shows snub tenons on the upper ends of timbers expected solely to bear weight (fig. 74).
28 The evidence for both buildings is summarized in Emery, Anthony, Darlington Hall (Oxford, 1970), 237, 241, and 142Google Scholar; fns 15 and 19. A measured drawing of the destroyed roof at Dartington Hall shows that its structure closely resembled the roof of Pilgrims’ Hall (Courtenay, ‘14th-century Sources’, fig. 12). The roof was reconstructed with thinner elements than it originally had.
The dorter of Westminster Abbey was later used for Westminster School, and although it was destroyed in 1941, it is documented in measured drawings and photographs. Drawings made in 1716 are reproduced in the Wren Society, vol. XI (Oxford, 1934), pl. 8. The measurements indicated correspond to the scale on the drawings, but the drawings incorrectly omit wallposts, which show clearly in photographs.
29 Baines, Westminster Hall, p. 21.
30 As Mainstone emphasized, ‘…the manner of jointing is crucial’ (‘Structural Analysis’, p. 319).
31 Salzman, Building in England, p. 211.
32 Baines, Westminster Hall, p. 48.
33 If the hammerbeams were originally horizontal, they sank 23/8 in to 53/8 in. Harvey argued that the hammerbeams were originally angled upwards by about 9 in, and his evidence seems convincing, but that would still mean that their ends sank by a total of only about 1 ft (‘Woodman’, p. 221 and fig. 2).
34 On aesthetics as a functional requirement of medieval architecture, see Harvey, John, ‘Mediaeval Design’, Transactions of the Ancient Monuments Society, n. s. 6 (1958), pp. 69–71 Google Scholar.
