The Dean and Canons’ Houses of St George’s Chapel, Windsor: An Architectural History

By John Crook

The College of St George at Windsor Castle was founded by Edward III in 1348 to support the newly created Order of the Garter, and to this day fulfills the same primary purpose. The domestic buildings provided for the Warden, Canons and Priest-Vicars – now the Deanery and Canons Cloister – are an astonishing survival: despite enlargement and alteration over the centuries, a significant amount of the mid-fourteenth-century fabric survives, though often hidden from view. A recent program of refurbishment and conservation revealed much hitherto unknown evidence for the way the buildings were constructed, their fittings and decoration and their subsequent evolution. The author maintained a continuous ‘watching brief’ throughout the refurbishment works, the results of which are published here for the first time.

The archaeological evidence is supplemented by the excellent survival of documentation, both for the initial construction of the buildings and their subsequent development: we know the precise date of each stage of construction, the cost and even the names of the workmen involved. The post-medieval history of the buildings is also highly significant, and for this period we have the benefit of knowing more about the deans and canons who influenced the ways their dwellings developed, and of a continued wealth of documentary evidence.

• Language: English
• Publisher: Oxbow Books
• Release date: Dec 2, 2022
• ISBN: 9781789258660

Book preview

1

The castle down to the reign of Henry III

---

The site

Approaching Windsor Castle from the east, both the road from the M4 via Datchet and the railway from London run through the flood plain of the Thames. The castle forms an elevated backdrop, with the defensive towers of the Lower Ward rising above the trees that hug the castle’s lower slopes (Fig. 1.1).

The hill on which the castle stands is an isolated inlier of the Upper Chalk (Seaford and Newhaven Chalk formations) created by local folding of the strata, so that a small dome of chalk, capped with clay-with-flint deposits, rises like an island about 33.5m (110ft) above the Shepperton gravel of the flood plain and the Holocene alluvium of the riverbed. North of this hill the shifting meanders of the Thames laid down gravel terraces; at one stage the river ran close to the chalk outcrop, cutting a cliff whose steep, tree-clad remnants overlook the valley bottom. To the south the inlier is surrounded by the silts, clays, and sands of the Lambeth Group, and further south still, in Windsor Great Park, increasingly thick deposits of the London Clay Formation are capped by the sands of the Bagshot Formation, which give the heathland of the southern park its distinctive character. This heathland provided the nearest source of building stone: the scattered sarsen boulders known as ‘heathstone’, a hard siliceous sandstone that was used as the principal facing material for the medieval castle.¹

Fig. 1.1. Impression of Windsor Castle from the north-east (Google Earth/John Crook).

Although some prehistoric, Romano-British, Roman, and Saxon pottery fragments were discovered during the investigation of the Norman motte,² they may well have been imported with the soils of which the motte is composed; no certain evidence for pre-Conquest occupation of the wider castle area has yet been found.

Windsor takes its name from ‘Old Windsor’, two miles or so down the river, a mid- to late Saxon settlement, which, as shown by excavations in the 1950s, reached a considerable size in the eighth century.³ The etymology of ‘Windsor’ is debated, but the most probable lexical elements are ‘winding’ plus ora, ‘bank’ or ‘shore’, referring to the sinuous local course of the Thames.

It is thought that this settlement, with its high-status buildings, was a royal estate, strategically situated on the borders of Wessex and Mercia. The site appears to have been abandoned for a time after destruction by fire c. 900, probably by Viking raiders, but in the eleventh century it resumed importance, perhaps as a royal residence, though Derek Keene has questioned whether Edward the Confessor in fact held many councils at Windsor.⁴ That he had a palace there is suggested by a number of historical anecdotes,⁵ for example, Henry of Huntingdon’s account of a quarrel between Earl Godwin’s two sons in aula regia in 1063, though that was written 60 years after the event.⁶ Whether any of Edward’s charters were issued at Windsor seems debatable. Perhaps in 1058 or 1061, and certainly in 1065, Edward held his Whitsun court there, but whether he ceremonially ‘wore his crown’ is debated; that may have been a post-Conquest feature of the seasonal feasts.⁷ At his death, Edward gave Old Windsor to the monks of Westminster Abbey,⁸ perhaps an indication that he viewed Windsor as a secondary royal residence compared with Westminster. Old Windsor was reacquired by William I after the Conquest, and he and William Rufus also celebrated the Whitsun feast there on four occasions. By 1107, however, Westminster was the favoured location for the Whitsun crown-wearing.

In 1170 William ordered a castle to be built on the hill at the place that would soon take the name ‘New Windsor’, then part of the estate of ‘Clewer’. These immediate post-Conquest years were times of unrest as the Normans established their hold on England, and William also built castles at Wallingford and Oxford, part of a line of defences along the Thames valley. The hill at Clewer must have been regarded as of strategic importance, with its commanding views of the valley, and possibly there was already a bridge there, though the earliest reference to any bridge dates from 1191. The entry for Clewer in Domesday Book relates how one half-hyde had been appropriated from the estate where the ‘castle of Windsor’ was located: its first mention. The entry regarding the 1110 crown-wearing in MS E of the Anglo-Saxon Chronicle states indeed that Henry held his court ‘on Þam niwan Windlesoran’ (‘at the New Windsor’) – a clear distinction from the old site.⁹ By 1110, when Henry I held his Whitsun court there, the town of New Windsor was beginning to be established below the castle.¹⁰

The development of the Lower Ward

The development of Windsor Castle was traced in great detail by William St John Hope over a century ago in his monumental Windsor Castle, An Architectural History,¹¹ the fruit of many years of painstaking research that was finally published in 1913, making copious use of the Pipe Rolls and other royal documents. His work remains a significant source, but many of the findings have been corrected or amplified by Howard Colvin and his collaborators in The History of the King’s Works,¹² and Steven Brindle’s comprehensive Windsor Castle.¹³ In its earliest days the castle simply comprised a motte and bailey, the usual form of Norman castle construction, though on an unusually large scale. The bailey was initially established to the west of the motte, enclosed by a ditch, and this was the area later known as the Middle Ward. West of the bailey and separated from that core area by a ditch (soon reinforced by a wall on the upper side) was the area that would eventually become the outer bailey, sloping gently downwards, and initially somewhat ill-defined and unprotected.¹⁴ By the mid-twelfth century the present broad division of the castle precincts into the Upper Ward, Middle Ward (including the motte), and Lower Ward had been established. The north curtain wall of the Lower Ward was built around 1180 but the western defences were fully established only in 1227–30 when the wall and three towers were built.

However, a hall and chamber block in the Lower Ward appear to have been constructed during the reign of Henry II, before the circuit of curtain walls was fully complete. The area was therefore vulnerable to attack, and the buildings in question suffered during the siege of 1193, when William Marshall’s knights surrounded the castle, which Prince John had taken over as a stronghold during his attempt to supplant his brother King Richard. Thus, the earliest mention of these buildings of Henry II in the royal accounts relates to their repair following the siege, in Richard I’s reign. A chamber ‘outside the royal castle of Windsor’ (i.e. in the Lower rather than the Middle Ward) was repaired in 1194–5 together with other buildings;¹⁵ and, less ambiguously, ‘the King’s Hall at Windsor’ was repaired in 1197–8.¹⁶ These seem to be the first references to a group of buildings that is increasingly well documented throughout its long history. Furthermore, elements of the medieval buildings survive, as discussed below, and the site has twice been archaeologically excavated.

The buildings again suffered in 1216 when Louis the Dauphin, the future Louis VIII of France – already acclaimed as King of England by the rebel barons – besieged the castle, and early in the reign of Henry III work started on improving the defences.¹⁷ St John Hope equates this with unspecified major expenditure on the castle in 1223–6.¹⁸ By about 1230 the curtain wall on the south side of the motte and Lower Ward had been constructed, continuing north-west up the end of the Lower Ward and including the three towers that remain dominant features of that elevation today, albeit much restored, currently known as the Salisbury, Garter, and Curfew towers.¹⁹ They are well documented in an account roll between September 1227 and August 1230, mentioning ‘the work of a wall with three towers by the King’s brief’.²⁰

As we shall see, the Lower Ward was much favoured by Henry III, and the reconstruction and repair of the hall and chamber seems to have been in progress by January 1223 when the king ordered payment to be made to his constable Engelard de Cigogny ‘for the rebuilding of our hall at Windsor’, followed by further payments for ‘the repair of our hall and houses at Windsor’.²¹ The term ‘rebuilding’ and ‘repair’ emphasise that the buildings in question were of earlier date, probably of the time of Henry II. Then in 1224, Thomas, the king’s carpenter, was granted ‘one hundred of fir for the doors and windows of our hall at Windsor’.²² The work on repairs to the hall was still in progress in 1227–8, together with the construction of a new kitchen.²³ In 1233–4, payments were made ‘for repairing the windows of the great hall and the great chamber of the same hall, and for certain glass windows in the middle of the windows of the same hall on either side of the king’s dais’.²⁴

True, in none of these documents is the location of the hall precisely stated, but the archaeological evidence points to a major reconstruction in the thirteenth century, and given the subsequent creation of royal apartments close to the hall, as described below, it is unlikely that these references are to a different ‘king’s hall’ in the Upper Ward. The documentation, though sparse, points to a complex typical of the architectural layout of the period: a great hall with a dais at the ‘high’ end, and an adjacent chamber block.

Fig. 1.2. No. 25 The Cloisters from the south. The first-floor arch is identified as a thirteenth-century doorway, the opening on the ground floor as a window. The projecting wing to the right is late nineteenth-century (part of ‘Lord Normanby’s Residence’ in 1895).

Fig. 1.3. Ground floor of No. 25, view south-west. A plain, round-headed twelfth-century arch, into whose south jamb a thirteenth-century doorway has been inserted.

No. 25 The Cloisters and Denton’s Commons

Some elements of this complex survive in No. 25 The Cloisters – not of course part of Canons’ Cloister, but well to the west (Fig. 1.2). Parts of the external walls show the wide joints and rough diagonal tooling typical of twelfth-century work; and some twelfth-century masonry and features are identifiable internally. Archaeological investigations in 2006, discussed more fully below, suggested that there were two twelfth-century wings corner to corner: one alongside the curtain wall and the other projecting south, with a rectangular space between the two blocks. However, much of the building may be attributed to Henry III’s rebuild as we shall see.

Important medieval remains within the house were revealed when it was converted to provide accommodation for St George’s House in 1964–5 and were recorded by Peter Curnow of the Historic Buildings Inspectorate of the then Ministry of Public Buildings and Works.²⁵ At ground level a round-headed opening is clearly twelfth-century, but a thirteenth-century doorway with deeply moulded arch and foliate capital is inserted into the south jamb at right-angles to the earlier arch, evidence for the upgrading of the building by Henry III (Fig. 1.3). On the first floor are remains of a chamber, with a wide window embrasure through the curtain wall at the north end. On the east side is another window embrasure flanked by slender shafts with typical thirteenth-century capitals and bases (Figs 1.4 and 1.5). A doorway next to it is a later insertion. Remnants of a thirteenth-century decorative scheme were discovered on the west wall, featuring signs of the Zodiac.²⁶ Another survival was a medieval hearth (behind an eighteenth-century surround) at the south end, whose splendid circular chimney – probably also thirteenth-century – may be seen through a glazed observation panel at second floor level (Fig. 1.6). Also in No. 25, visible within a guest bedroom of St George’s House on the second floor, are four surviving trusses of a scissor-braced roof (Fig. 1.7), one of a pair of parallel roofs as investigated below. Although the timbers have not been tree-ring dated, this type of roof is normally attributed to the thirteenth century, and it is therefore part of Henry III’s works. It is the earliest timber roof within the castle.

More of this wing survived indeed until 1859, having been incorporated into the house that Canon Denton adapted from the medieval building in 1519; it is convenient to move ahead two and a half centuries and discuss that building here, because understanding its original structure has important implications for the layout of this part of the thirteenth-century castle.

James Denton, a Canon of the Ninth Stall from 1509 to 1533, had followed a well-trodden ecclesiastical path: scholar at Eton, then of King’s College Cambridge where he became a Fellow then Bursar, and holder of numerous church preferments – he was, it would seem, a wealthy man. In 1519 he funded the construction of the ‘New Commons’ as accommodation and a common hall for the choristers and chantry priests.²⁷ The choristers probably slept on the first floor in the southern two-thirds of the building; the thirteenth-century first-floor chamber at the north end was perhaps where they were taught, given the survival of painted musical notation on the walls: seemingly for educational purposes (Fig. 1.8). A study by Dom Anselm Hughes suggested, however, that the music dated from ‘the decade 1470–80’, and he noted at least one earlier musical scheme beneath the one now displayed.²⁸ He did, however, speculate that the music might constantly have been revised over successive coats of obliterating whitewash – rather like wiping previous writing on a blackboard. The function of the room as the choir school seems therefore to pre-date Canon Denton’s modifications – he might of course have adapted the building precisely because it was already in use by the choristers. A ‘room where the choristers are taught’ is mentioned in the treasurer’s accounts for 1489–90, when a ‘lattice’ was fixed in the window.²⁹

Fig. 1.4. No. 25 The Cloisters. East wall of first-floor conference room, formerly an upper chamber, with splay of thirteenth-century window (and later medieval inserted doorway).

Fig. 1.5. No. 25 The Cloisters. Thirteenth-century capital on nook shaft of Henry III window in the first-floor conference room.

Fig. 1.6. No. 25 The Cloisters. Probably thirteenth-century chimney serving the chamber at the north end of the chamber block.

Fig. 1.7. No. 25 The Cloisters. Four trusses of a thirteenth-century scissor-braced roof surviving at the north end of the chamber block.

Denton’s ‘New Commons’ were regrettably short-lived, for in 1550 the house was allocated to the college’s Reader in Divinity. From the early 1570s the house became the residence of one of the canons.

Photographs of the building’s demolition in 1859 (which St John Hope qualified as ‘needless’)³⁰ show a massive wall of chalk rubble, presumably faced externally, on the east side of the house, rising through three floors, beneath the truncated scissor-braced roof (Fig. 1.9). The east wall was visible after the demolition of what was then a quite recent two-storey addition on the east side, incorporating a castellated wall around what seems previously to have been a courtyard. The west wall of the house was narrower, but also formed of chalk rubble into which sixteenth-century brick window embrasures had been inserted. The function of this wall was revealed by an archaeological examination in 1895 during the excavation of service trenches, when a much more substantial wall, similar in thickness to the east wall, was discovered. The narrower wall appears therefore to have been a spine wall running down the centre of the medieval building; it still continues northwards as a partition wall within No. 25 and, like the main west wall of the building, ends in a buttress on the curtain wall. The plan of the 1895 works, now attributed to the Chapter Surveyor A. Y. Nutt (Fig. 1.10),³¹ shows that the excavators thought this wall was post-medieval, built when Denton’s Commons was constructed, but the final photographs clearly indicate that the fabric is of rubble of similar construction to the east wall; he was perhaps confused by the brick window embrasures. Regrettably the footings of the narrower wall were not re-excavated during an investigation of the area undertaken by Cambrian Archaeology in 2006 as part of a Time Team Live Big Royal Dig, which would have confirmed the point. Furthermore, the narrower wall evidently supported the foot of the west slope of the thirteenth-century scissor-braced roof, and an engraving by Wenceslaus Hollar of c. 1670 (see Fig. 4.18) suggests that there was a similar roof parallel to it to the west, spanning the western half of the building.

Fig. 1.8. No. 25 The Cloisters. Musical notation on south wall of first-floor chamber.

Thus, this part of the building, with its important thirteenth-century survivals, appears to have consisted of two halves, independently roofed, and separated by a spine wall rising up to roof level. Nevertheless, for St John Hope and, indeed, the authors of the History of the King’s Works, the north wing was the great hall mentioned in the Pipe Rolls. The fact that it was sub-divided and possibly of two storeys surely disqualifies this possibility. In an article written in 2000, Tim Tatton-Brown was perhaps the first to question the identification.³² He proposed that the north wing was the chamber block, and that the hall was aligned east–west with its north-west corner corresponding to the south-east corner of No. 25. His theory was tested by the Time Team excavation previously mentioned. Although the trenches were small, it was felt there was sufficient evidence to indicate that the hall ran east–west against the curtain wall rather than to the south of it.³³ The building thus postulated was 8.5m (28ft) wide. It was therefore too narrow to be aisled, but it might have compensated for this by its length – up to about 25m (80ft) if it extended as far as the medieval wing of the building that is now St George’s House. One parallel for a hall of this size is the thirteenth-century ‘Arthur’s Hall’ against the north wall of Dover Castle, which measures about 23m (75ft) long internally by 9m (30ft) wide.

Fig. 1.9. Demolition of Denton’s Commons, 1859 (SGC M.175/226).

The entrance to the hall would have been at the east end, and it may be significant that in 1250 the king issued a writ that a wall 3.05m (10ft) high should be built from the door of the great hall to the Galilee (the porch at the west end of the chapel), which again supports the notion that the hall extended eastwards from the chamber block now represented by No. 25.³⁴

The possibility that the twelfth-century north wing was originally the hall, but that it was converted into a chamber block by Henry III seems most obviously to be excluded by the fact that the works on the hall during his reign are said to be ‘repairs’ rather than a reconstruction, with the caveat that the word reparatio was used a century later in the Windsor surveyors’ accounts and the Pipe Rolls to mean ‘preparation’ rather than ‘reparation’. At the end of Henry’s works the north wing was undoubtedly a chamber block, so it could never have been the repaired hall, which has to be located elsewhere.

Fig. 1.10. Detail from plan attributed to A. Y. Nutt, showing investigation of the site of Denton’s Commons in 1895 (SGC P/75). The colour phasing should be ignored.

Fig. 1.11. No. 7 Canons’ Cloister. Blocked round-headed window opening in the north face of the Latrine Tower, at ground-floor level.

The documentary evidence for this interpretation is somewhat ambiguous – and amongst the difficulties is the fact that when a ‘hall’ or ‘chamber’ are mentioned it is not always certain that the buildings in question are in the Lower Ward rather than comprising part of the ‘king’s houses’ in the Upper Bailey. One reference, discussed more fully below, was seized on by St John Hope as proving the north–south wing was the hall: in 1255–6, Henry III ordered that a lavatorium (wash place) be created ‘at the head of our hall’ (a parte orientali).³⁵ Hope’s translation of the last phrase as ‘on the east side’ reflects his identification, but the Latin might merely mean in the area east of the hall.

The towers of the north curtain wall

On the north side of the Lower Ward are a number of towers. At the east end, where the wall and ditch separating the Middle and Upper Wards met the north curtain wall, was the Winchester Tower, rebuilt and heightened by William of Wykeham in 1357–8 and subsequently known as the ‘Wykhamtour’, but greatly rebuilt by Wyatville in the 1820s.³⁶ This tower, however, forms part of the Middle Ward and need not concern us here.

The next tower has no historical name, but because of its nineteenth-century bay windows has been dubbed the ‘Bay Window Tower’. This tower was raised higher during the creation of Canons’ Cloister in the 1350s, when the curtain wall was heightened. Like most of this elevation, archaeological evidence is almost totally obscured by Anthony Salvin’s refacing in the 1860s. During conservation works in 2011, however, a narrow (461mm) Romanesque window opening was discovered on the east face at modern second-floor level. The voussoirs of its round-headed arch were of Wheatley limestone, and the jambs of Taynton stone. Disturbance to the masonry suggested there had been another similar window opposite it on the west side.³⁷

Third in the series of towers is the ‘Latrine Tower’, which housed the communal latrine of the dean and canons from the 1350s. It, too, appears to be of earlier date, as was shown in 2011 when the walls of the ground-floor drawing room of No. 7 were replastered. Above the fireplace in the north wall jambs and a few voussoirs of the arch of another round-headed window were discovered (Fig. 1.11). The opening appears to have been about 775mm (31in) wide internally, tapering to an external width of 550mm (22in). Stylistically this Romanesque window would suggest that the tower formed part of Henry II’s fortification of the Lower Ward. However, the stone used for this opening was Reigate, more commonly associated with Henry