Windsor Castle, founded by King William I, consists of a big chalk motte on which stands a round tower, with walled enclosures to the W (the Lower Ward) and E (the Upper Ward. This is no place for a full architectural description of the castle, especially since it is by no means certain that the carved Romanesque stones described here were originally from this site (see Comments and Opinions, below), but the Bibliography will be useful as a guide to further reading about the site.

The material described here is all in the form of loose or reset stones, out of context and displayed in various locations around the castle, as follows.

The Moat Garden

This is the cultivated motte on which the Round Tower stands. The earliest account we have of a garden on the mound is in King James I of Scotland’s poem,The King’s Quair, written during his period of captivity at Windsor Castle between 1413 and 1423. This describes a small and private garden alongside the wall of the Round Tower :

Now was there maid, fast by the Touris wall,

A gardyn faire, and in the corneris set,

Ane herbere grene, with wandis long and small,

Railit about, and so with treis set

Was all the place, and hawthorn hegis knet,

That lyf was non, walkyng there forbye,

That mycht within scarce any wight aspy. (Tighe and Davies (1858), 79.)

There was little or no planting on the rest of the mound then, and none is visible in Norden’s 1607 view of the castle, but Hollar’s c.1672 engraving shows the south slopes divided into a series of geometric beds, quite unlike the present arrangement. (Roberts (1997), 161).

The area of the mound is associated with the so-called Norman Gate, built alongside it by William of Wykeham c.1359. At the time of Wyatville’s restoration of the castle, the mound had been rented out to a fruit and vegetable grower who sold the produce at market. The fruit trees planted against the curtain walls below the Round Tower were apparently undermining it, and in 1836 Wyatville sent a strongly-worded letter to the then resident of the Norman Gate, Lady Mary Fox, the State Housekeeper, warning her that if any more fruit trees were planted there was a danger of it collapsing. Indeed, some years earlier part of the wall had collapsed for this very reason, and repairs at a cost of £1500 were needed to repair the damage. (Taylor (1935)). By 1844, when the mound was photographed by Fox-Talbot, it was planted with a few shrubs and small trees, but there was no longer any evidence of systematic fruit-growing. The wall that partly encircles the mound on the inner side of the middle terrace was in place by that date (figure 1).

In 1901 the Norman Gate was given as a residence to the Right Hon. Sir Dighton Probyn, on his appointment as Keeper of the Privy Purse. Probyn had been a young officer during the Indian Mutiny, when he was awarded the VC and he rose to the rank of general before he entered the service of the Prince of Wales, later King Edward VII. It was Probyn who was responsible for clearing away the remains of the Victorian design, and he transformed the mound into a richly-planted and less formal garden. He brought brown Carstone from Norfolk to build a rock garden, and introduced a water and bog garden and the present scheme of walls, terraces, borders and architectural ornaments. As we shall see, this work was in progress when Keyser photographed the Romanesque stones c.1915. The garden was neglected during the remainder of the Great War, and many plants were lost in the drought of 1921, so that Lord Wigram, Probyn’s successor as occupant of the house, had to restore the garden and added rare plants from the Far East (Taylor (1935), 326-28).

The present design is more or less Probyn’s, and consists of an outer crescent following the inside of the lower moat wall with a long bed of shrubs and herbaceous plants that runs in a three-quarter circle from the NE to the SE of the mound (figure 2). On this wall is an arched feature made of Romanesque carved stones (figure 3), and other loose stones are placed on stone shelves built onto its inner face. The bed is bounded on the inside by a low brick retaining wall, and inside this the ground is terraced to provide a flat lawn enhanced by rose beds and a small pond, with a paved terrace at the east end. From the inner edge of the lawn the mound rises steeply to the base of the Round Tower curtain wall, some 90 or 100 feet above. The lower slopes are simply grassed except towards the E, where they are planted with shrubs and trees, and at the far northern end, where the rockery descends with a series of pools and a flight of steps from the rear entrance of the Norman Gate, now the Governor’s residence, to the level of the moat wall (figure 4). At the top is the Herbere pavilion, and below it the Corinthian Fountain (figure 5), both of which include Romanesque carvings described here. Partway up the mound, and concentric with the line of the lower moat wall at the bottom and the Round Tower curtain wall at the top is a narrow terrace now known as the middle terrace (figure 6), and another arched feature made up of Romanesque carved stones is set there, with further carved stones placed on a shelf at the east end of the terrace (figures 7, 8). Above the middle terrace, up to the curtain wall of the Round Tower, the planting is denser, though still informal. At the extreme east end of the garden, at the foot of the mound, are greenhouses. The slope rising behind them has been reinforced with ashlar blocks, and among them are several stones with roll-moulded edges that show traces of possible Romanesque carving (figure 9).

The South Bowe

The South Bowe is a blocked sally port on the south side of the Upper Ward of the castle. It appears to date from the thirteenth century, although there is little in the structure that is diagnostic of date. In the passage is a rough arch constructed largely of clunch but including, in its west jamb, two moulded stones and two chevron voussoirs (figure 10).

The Carronade wall

The carronade wall is the retaining wall of the central motte on the Upper Ward side. A single stone carved with two units of diapering is set into this, described as cat. 55. It was photographed by Tsang and drawn by Brian Kerr. (figure 11).

All stones are numbered and described; the organisation of the stones being based on the type of object rather than its location. Loose stones were all photographed separately, but this was not always possible for stones built into structures, most notably the two arched structures in the Moat Garden. In these cases stones were identified by their position within the structure, and photographic references to the figures in which they appear are included in the feature descriptions.

1. The Lower Moat Wall. This structure takes the form of an arch attached to the garden side of the lower moat wall at the west (figure 12). There are two non-Romanesque stones carved with human figures in the centre, and around them is an arch made up of twelve 12th-century stones. Each jamb is made up of three stones, the arch is of five voussoirs and a niche-head in the centre forms a pseudo- tympanum. In the catalogue these stones are identified as follows:

LM1 - LM 3. Left jamb, bottom to top.

LM 4 - LM 8. Voussoirs of arch, left to right.

LM 9 - LM 11. Right jamb, top to bottom.

LM 12. Niche-head “tympanum”.

2. The Middle Terrace. The main feature is a structure in the form of an arch at the west side of this terrace, supported on a capital (R) or an impost (L) (figure 13). Five voussoirs form the arch, and five sections of label of two different designs surround the arch. Under the arch is another niche head, and below this, forming the centrepiece framed by the arch, is a post-medieval lion’s head. In the catalogue these stones are identified as follows:

MT1 - MT5. Label stones, left to right

MT6. Left hand impost

MT7 - MT11. Stones of arch, left to right

MT12. Right hand capital

MT13. Niche-head “tympanum”

Edward the Confessor held his court at his hall at Old Windsor, but this was not on the present castle site. King Edward granted Windsor with 20 hides to St Peter's Westminster on the eve of the Conquest, but William I recovered it in exchange for lands in Battersea. He began work on the castle at an unknown date in the 11thc, and work was advanced enough for it to be used as a royal residence by the 1090s. In 1095, following VCH (1923), the Earl of Northumbria was imprisoned there, and in 1100 William FitzWalter was appointed constable. From 1095, when William Rufus kept Pentecost there, until the present day, Windsor has been a regular royal residence.

The castle was certainly begun by William I, and Henry of Huntingdon records under the year 1110 that the King held his court at at Whitsun at New Windsor, which he had built himself (Greenway (2002), 52). VCH suggests that Henry I's work was confined to the royal lodging in the upper bailey. In Henry II's reign, large sums were spent on the castle in 1169-79. Later works are summarised in VCH (1923), 29-56.

The earliest reference to loose Romanesque sculpture at Windsor is found in the essay by Ambrose Poynter that is included in Wyatville’s Illustrations of Windsor Castle, published in 1841. At the end of his discussion of the Norman work at the castle, Poynter remarked that,

"a few architectural fragments in the Norman style, brought to light from the excavations during the progress of the late improvements, are perhaps the only relics of the palatial edifice of the Twelfth Century" (Wyatville (1841), 2).

He usefully included a small vignette, illustrating a group of the excavated stones (figure 23). The engraving shows five stones, from left to right:

1. A section of a three-order blind niche with a chamfered inner order, a plain roll outside it, and an outer order of sawtooth.

2. A voussoir carved on both face and soffit with the fret ornament.

3. An elaborate double-scallop capital with the shields decorated with reeded cusping and leaves. This has a hollow chamfered impost on the top and stands on a section of shaft.

4. A double composite human/ animal head corbel.

5. A section of a draped figure.

Of these, numbers 4 and 5 are certainly identifiable with stones 52 and 57, both now in the Herbere pavilion, while the fret ornament on number 2 is comparable with that on stone 27, now on a shelf at the east end of the middle terrace. This stone is now very badly worn, but it is illustrated in a photograph of c.1915, when it was in a much better state of preservation (figure 24). To the group illustrated by Poynter can be added the diapered stone re-used as masonry in the carronade wall, which must have been incorporated as part of Wyatville’s restoration, and was presumably unearthed with the rest (stone 55). This has the same design as stone 54 in the Herbere pavilion.

Although Poynter described these carvings as sculpture originating at the castle, another explanation for their presence here has been proposed, and is perhaps more commonly accepted. This is that the stones came from Reading Abbey, a major Benedictine house founded by King Henry I in 1121 as his mausoleum and demolished after the Dissolution (Baxter (2016), Kemp (1986), I, 13-19; Thurlby and Baxter (2002), 282-301; Baxter and Harrison (2002), 302-12, Hurry (1901). This was suggested in 1913 by Sir William St John Hope, who observed in his architectural history of the castle that,

"it is advisable here to add a word of warning concerning the fragments of Norman carved work that are from time to time dug up in the Castle or extracted from later work. Some of these doubtless belong to Windsor, but a large quantity of material was brought here from Reading abbey after its suppression and destruction. Much of this work was also Norman, and it is therefore difficult to assign to such fragments their proper birthplace" (Hope (1913), I, 22).

When Charles Keyser, the eminent Berkshire antiquarian, saw the Windsor stones at around the same time, they were ranged on shelves attached to the curtain wall of the Round Tower, and a tablet has been inserted recording the fact that they were originally part of the monastery of Reading founded by King Henry I in 1120 (Keyser (1916), 241).

Keyser published eight photographs of the Windsor Castle stones, and while he clearly did not photograph everything that he saw, his images are still valuable for establishing the date of the present arrangement of the stones in the Moat Garden, and for estimating losses that have taken place since 1916. From these the following conclusions can be drawn:

1. That the arched feature inside the lower moat wall was already in place then, but the uppermost of the two post-medieval stones had not yet been inserted (and furthermore that stone 8 must have been cut back to accommodate it). A label voussoir was placed above the keystone of this feature in Keyser’s photograph, but this is no longer in place and has not been discovered anywhere else in the garden (see figure 25).

2. The shelf at the east end of the middle terrace was far less overgrown in 1916, and two stones of fret voussoir type 2 were then clearly displayed side by side. Two other chevron voussoirs were then on the shelf, alongside the multi-scallop capital number 44, which is still there (see figure 24).

3. The stones in the Herbere pavilion were apparently already there in 1916. Keyser illustrated 11 stones in a row on a shelf, all of which and no others are still in the garden house today. What is missing is the tablet identifying the carvings as Reading Abbey stones, that is also shown on the shelf in Keyser’s photograph (see figure 26). In view of the chronology of the garden it seems certain that this was also added by Sir Dighton Probyn.

4. Keyser’s set of photographs suggests that the arched feature on the middle terrace had not been erected when he visited. Not only does he fail to illustrate it, but items now incorporated in it appear as loose stones among his plates, notably the two bobbin and sawtooth label voussoirs (stones 31-32, see figure 27, lower shelf).

5. Important stones illustrated by Keyser have not been discovered in the present survey. The most significant are two capitals: a second trefoil capital, similar to stone 46 but with its integral impost intact; and a double scallop capital with beading around the lower edges of the shields and furled leaves between the cones (figure 27).

It has been suggested that works and repairs undertaken by the Ministry of Works in the 1920s may have been responsible for some of these losses, and there does seem to have been a good deal of upheaval in the relevant parts of the Moat Garden at that time. In 1923-24 paths and terraces in the Moat Garden were extended and repaired; the work involving the removal of undergrowth below terrace level on the side of the motte, the removal of some modern steps, reforming the slope where necessary, repairs to the brick revetment, and the construction of ‘one or two lengths’ of retaining wall. Further repairs to the walls and path surfaces, including the reconstruction of a cracked and bulging wall had been carried out by early 1928 following a report of ‘symptoms of earth slipping’ made by Sir Lionel Earle of the Ministry of Works in April 1927 (National Archive WORK 19/738).

The evidence for Reading rather than Windsor as the source of these carved stones is strong but by no means conclusive. In 1557 Queen Mary was responsible for the reconstruction of the Poor Knights' Lodging in the Lower Ward, and the record of payments for this work indicates that the stones were brought to Windsor from Reading Abbey, which was being taken down at that time (Tighe and Davies (1858), I, 166).The accounts are reasonably full, and list payments to masons choosing the stones they wanted; taking down the stones of the doors and windows of the Lady Chapel of the abbey and digging out stones from the walls; and to labourers who dug Caen stone out of the windows for the battlements of the new lodgings at Windsor. This is the only surviving record of stones being taken to Windsor from Reading, and it is assumed that the stones now in the Moat Garden came at that time. Keyser, for example, asserted that,

"as early as the year 1557, when some repairs were needed at Windsor Castle, some barge-loads of stones were collected and removed, as being easily accessible, and suitable for the contemplated restoration. These are now arranged in groups in the curtain-wall and elsewhere at the foot of the Round Tower, and by the courtesy of the authorities and the kind assistance of Mr. Robertson, the assistant architect, I have been permitted to photograph them" (Keyser (1916), 237).

This account of the transfer of the stones from Reading does not really stand up to close scrutiny. The Lady Chapel of Reading Abbey was built by Abbot Whaplode around 1314, and would only contain Romanesque stones if they had been re-used in its construction. Even if so, it would seem unlikely that such a large number of carved stones had been recycled at Reading, and that the masons paid for choosing which stones to take would have come away with so much stone that could be of little use to them in rebuilding the Poor Knights’ Lodgings. Further, the accounts specify Caen stone, and while some of the Romanesque material in the Moat Garden is Caen, a good deal of it is not. If the stones came from Reading they must have been imported for antiquarian or aesthetic reasons; perhaps unlikely at such an early date.

Nevertheless, some of the Windsor stones are so close in design to known Reading Abbey stones as to make the possibility of a connection difficult to rule out entirely. The trefoil scallop capital, stone 46, has the horseshoe-shaped lower shield lobes and beaded border typical of many Reading Abbey cloister capitals, e.g. Reading Museum 1992-107 (figure 28). Stone 46 has lost its impost, but the similar example illustrated by Keyser clearly has the integral impost typical of Reading Abbey and uncommon elsewhere (figure 27). The type of niche head enriched with large, plain bosses seen in two examples at Windsor (stones 47 and 48), is also found in a Reading Museum stone excavated by George Zarnecki, the former Deputy Director of the Courtauld Institute of Art, at Borough Marsh in 1948, and thus certainly from the abbey (figure 29). The measurements of all three stones are comparable. Finally, the chip-carved diaper relief, stone 53, is a close match for a stone found among the Reading Abbey ruins and built into the exterior wall of Pugin’s Roman Catholic church of St James, Reading in 1840 (figure 30).

All these comparisons, of course, are with Henry I’s work at Reading (1121-30), but the king had been engaged at building work at Windsor, too, around this time. According to Henry of Huntingdon, he spent Whitsun of 1110 at New Windsor, ‘which he had built himself’ (Greenway (1996), 52). Despite the difference in dates between the campaigns at Windsor and Reading, it is fair to assume that the king used the same sculptural workshop for both, so that a similarity with Reading work does not necessarily mean that the stones came from the abbey. There are some twelfth-century stones in this collection, however, that cannot possibly date from the reign of Henry I, and while Reading Abbey church was not dedicated until 1164, there is no evidence either in the documentary record or among the known remains of the abbey of sculptural work dating from the later twelfth century. This is not true of Windsor, where Henry II spent considerable sums on building work between 1161-62 and 1173-74 (Tighe and Davis (1858), I, 31-32; VCH Berkshire (1923), 29-32). The kind of soft drapery seen on stone 57 is typical of the 1170s and ‘80s and later, most famously in the Canterbury screen fragments or the figures from St Mary’s Abbey, York, now in the Yorkshire Museum, but interestingly too at another royal site, Dunstable Priory (Beds), on the W and N doorways of c.1170, where arch voussoirs carved with fret on both face and soffit, as on stone 27, can also be seen. Other stones from this period include the chevron voussoirs in the South Bowe (stones 22 and 23), and the two impost blocked carved with fruit (stones 40 and 41).

There was certainly a tradition that the Romanesque stones at Windsor were brought from Reading, but it is impossible to trace it back any further than Sir Dighton Probyn at the very beginning of the twentieth century. Certainly Ambrose Poynter, writing sixty years earlier, considered them to be Windsor work. The few stones that closely resemble Reading sculpture date from Henry I’s reign, and seem on balance as likely to originate from the castle as the abbey. The later twelfth-century carvings must surely be from the castle, and provide evidence for sculptural work of the highest quality in the reign of Henry II.