Demons and Discoveries in a Miracle Window of Canterbury Cathedral
Rachel Koopmans, York University, Toronto, Canada
Canterbury Cathedral’s ‘miracle windows’ are a triumph of early Gothic pictorial narrative. Originally encompassing all twelve of the ambulatory windows in the cathedral’s Trinity Chapel, the series portrayed scenes from the life of Thomas Becket (d. 1170) and many dozens of his posthumous miracles. Eight of the twelve windows now retain at least some medieval glass, with the best survivals in the north aisle windows. Famed for their images of men, women and children suffering from ailments such as blindness, leprosy, drospy, wounds, and multiple forms of disability, the miracle windows are an unparalleled storehouse of medieval pilgrimage imagery .
Last summer, the Friends of Canterbury Cathedral provided the funds for a new investigation of window nV, the third of the six miracle windows on the Chapel’s north aisle (Fig. 1) . This investigation was stimulated by questions about the amount of original glass still in the window. Early photographs and archival documents suggested that two of nV’s panels that had been long dismissed as late nineteenth-century creations could be medieval compositions (nos. 5 and 6 on Fig. 1). There were also questions about the extent of modern restorations in nV’s four known medieval panels. nV’s glass was removed in July 2018 and brought to the cathedral’s conservation studio, headed by Leonie Seliger. There the glass was analyzed and photographed before being returned to the window in early September.
In the course of this examination, undertaken by Seliger with input from the author, it became clear that the two queried panels were in fact medieval compositions with much of their original glass intact. The first of these panels pictures pilgrims on the road to Canterbury, the earliest known representation of this subject, while the second depicts pilgrims queuing up within the cathedral to have their ampullae filled with the famous Becket blood and water relic. For more information on these discoveries, see http://www.medievalists.net/2018/09/discovery-earliest-known-image-pilgrims-canterbury/ and https://www.bbc.co.uk/news/av/uk-england-kent-45422965/medieval-glass-discovery-in-canterbury-cathedral-window). The four other medieval panels within the window were also examined (nos. 4, 7, 8, and 9 on Fig. 1), as were the four panels made for the window by Samuel Caldwell Sr in 1894, glass that had never been closely inspected before (nos. 1, 2, 3 and 10 on Fig. 1) .
This article will concentrate on the new assessment of one of the most eye-catching panels of nV, a diamond panel that depicts three demons, a man lying on a bed, and a man (wrongly restored with a female head) rushing to the sleeper’s aid (Fig 2, no. 4 on Fig. 1). The 2018 examination revealed that there was more medieval glass in this panel than is indicated in Canterbury’s Corpus Vitrearum Medii Aevi (CVMA) catalogue . Published in 1981, this catalogue was compiled without the opportunity to see nV’s panels out of their armature. With the glass out on the bench, the 2018 team could determine that the modern repair work was less extensive than had been thought. While the triangular shaped sides and the heads are modern replacements, nearly all of the rest of the glass is intact (see Fig. 3 for the updated assessment of the panel’s medieval content). The panel’s inscription, listed in the CVMA catalogue as mostly illegible, could be read in full. The close scrutiny of nV’s medieval panels also revealed evidence of strengthening strap-lead repairs likely dating to the seventeenth century. Such discoveries have provided the basis for a clearer understanding of nV’s restoration history, a better grasp of how much medieval glass survives in the window, and a heightened appreciation for how skilfully Canterbury’s glaziers transformed their written sources into striking visual images.
The miracle of the knight Stephen of Hoyland
The demons panel in nV was inspired by a miracle story from a very early phase of Becket’s cult – so early, in fact, that the tale begins before Becket’s death. For some thirty years, a knight named Stephen of Hoyland had been suffering visitations from a demon that were so severe he feared for his life. The writer of Stephen’s story, Benedict of Peterborough, states that the knight “was so pressed down and suffocated that unless he woke up quickly he would die” . Though the knight “spoke to many doctors, increased his gifts to them and offered them still greater ones,” this “brought him no benefit at all.”
Stephen first found relief on a night when Thomas Becket had returned from the exile imposed by Henry II but was still living (i.e., the short period between Becket’s return to England on 1 December and his death on 29 December 1170). The knight had been invoking the names of many saints when he said, “‘Lord, free me for the love of the archbishop of Canterbury who was exiled for you and the freedom of your church!’” This didn’t make the demon go away, so the knight, thinking that “many of the archbishops of Canterbury had been banished for the same cause, added, ‘Lord, for the love of the archbishop Thomas!’” This also didn’t work, and so the knight clarified that he meant “‘The Thomas who was last exiled for you!’” He had at last hit on the right formula: “immediately, that which had harassed him disappeared and he was freed.” That Benedict thought it worthy to report all of this shows the importance of establishing that it really was the living archbishop of Canterbury whose name dispelled the demon.
The demon came back again another night while Becket still lived, and the same invocation released the knight a second time. Then, sometime shortly after Becket’s death, the demon appeared yet again, this time in the form of a bird: “it seemed that it was unable to approach him in his bed, and so circled him from afar. The knight spoke to and taunted it, saying, ‘I trust in the merits of the blessed martyr Thomas, and I do not fear you: his grace will protect me, and will recover me from your control.’” This spelled the demon’s end: “from this hour,” Benedict writes, “no phantom troubled the knight in this way.”
Window nV was dedicated to events from the very beginning of Becket’s cult. Stephen of Hoyland’s miracle showed Becket to have had miraculous powers even before his death, and as such had obvious appeal for such a window. Canon Arthur Mason, who published his Guide to the Ancient Glass in Canterbury Cathedral in 1925, was the first to tentatively suggest a link between nV panel and the knight’s miracle . He looked for matches for panels with stories in the miracle collections written by Benedict of Peterborough and his colleague at Christ Church, William of Canterbury, in the 1170s. Though these two collections contain over seven hundred miracle stories, stories of demon possession are not numerous, and there are only two that describe a demon attacking a sleeper. One of these, in the middle of William’s collection, concerns a young monk in a Cluniac monastery in Yorkshire; the other is Stephen’s miracle, found very near the beginning of Benedict’s collection . It is not clear whether Mason considered the story of the Cluniac monk as a possible source for this panel, but as this was not an early miracle, and neither the sleeper nor the man coming to the sleeper’s aid are dressed as monks, the link he suggested to Stephen’s story is surely correct.
The three demons
Perhaps the most notable aspect of the nV panel is that it features three demons rather than the single demon afflicting Stephen in Benedict’s story. The glaziers rendered the demons in different colors, using a bluish white glass for the one on the right, a light tan glass for the one on the left, and a striking deep red glass for the one at the top of the image. They also portrayed them as doing different things. The bluish-white demon is pushing down on the sleeper’s legs; the light tan demon has his hands cupped over the sleeper’s head; and the red one, rather terrifyingly, is flying down from the ceiling.
The glaziers likely meant us to read not one but two things in this: first, the different ways in which the single demon afflicted the knight, and second, the three times that an invocation of Becket’s name worked to dispel it. The two standing demons would seem to be inspired by Benedict’s description of the knight being “pressed down and suffocated [vel oppressus vel suffocatus].” The bluish-white demon presses down onto the sleeper’s legs, so much so that the glaziers cut into his legs to make room for the demon’s hands. The unusual posture of the light tan demon with the cupped hands must be meant to evoke suffocation. The head of the sleeper is a modern restoration; if the original had survived, it might be clearer that he is being suffocated. The demon flying from the ceiling was evidently inspired by Benedict’s description of the final manifestation of the demon as a circling bird. At the close of the story, Benedict underlines the knight’s threefold liberation, saying that when the knight was telling the monks at Canterbury about his miracle, he praised “the martyr who, while living, had delivered him from this terrible distress once and then twice. After his death, he delivered him once more, and there was no need for another time.”
In other panels in nV, the glaziers utilized different colours of glass to help viewers distinguish between people or animals in a crowd. With the contrasting colours of the demons, they used the same device to convey narrative information and compress three different events into a single image. In the CVMA catalogue, the torsos of the demons are all marked as restored or repainted pieces, a judgement that may have been made in reaction to the stark painting of the standing demons’ arms and hands and the flying demon’s navel, ribs, elbow and knees. The hands of the standing demons look like crude paws when compared to the more delicately painted human hands. This differing painting style would seem to be intended as a visual distinction between the demonic and the human figures.
All of the heads of the demons are modern replacements, but the surrounding ground outlining these heads is intact. The original heads were clearly quite distorted. The elongated shape of the red demon’s head may indicate that it once portrayed an open-mouthed demon, such as the type seen in the Romanesque sculpture of Autun cathedral (see, for instance, the falling figure of Simon Magnus https://indrasmusings.wordpress.com/2016/02/26/inferno-xix-simon-the-magician/). The nV panel retains one foot of the light tan and one foot of the bluish-white demon. Both have the long-pronged toes and spike-like heel familiar from medieval demon imagery.
All three demons are naked, another marker of their non-human status. They are also hairless. Contemporary medieval images of demons typically portray demons as quite hairy, as one can see as the stained glass panel now in the Victoria and Albert museum depicting Christ’s temptation in the wilderness (see https://collections.vam.ac.uk/item/O8576) and the imagery of demons on the closely contemporary Guthlac Roll (see https://blogs.bl.uk/digitisedmanuscripts/2014/04/on-a-roll.html: the uppermost demon on roundel 8, the only one that is hair-free, is strikingly similar to the red demon on the nV panel). The Canterbury glaziers may have chosen to leave their demons hairless so that the differing colors of their skin, so important to the narrative design of the panel, would stand out more clearly. The corrosion on the red demon’s muscular torso has rather splendidly revealed the original painted shading along its ribs (Fig. 4). The strongly-painted navel and lines of the demon’s belly, elbow, and knee also give a convey of the demon’s strength, and the angle at which it is set, hurtling down towards the sleeper, visually conveys its threat.
The sleeping Stephen and his servants
With our improved knowledge of the condition of this panel, it can now take its place as a key medieval image of demonic attack . Most demon imagery from the Middle Ages is connected to stories of Biblical or saintly figures battling demons or performing exorcisms. What makes the nV panel stand out is that it is linked to a very particular experience of a named human being. Stephen told the Christ Church monks that “he would have preferred to have renounced all of his possessions than to continue to be vexed by this terrible phantom,” and that doctors told him “it was an ephialtes, what we would term, in Latin, an incubus,” though he disagreed, “constantly assert[ing] that it was a demon” .
The glaziers portrayed the sleeping Stephen in a contorted position. His left arm twists down across his body as if he is attempting to protect himself and turn away from the demon’s attack. His bedding, too, is disordered, with the brownish-pink blanket and the green sheet sliding down off the bed. That the glaziers were deliberately working to evoke the torment Stephen felt in his sleep is clear when one compares this image with another nV panel that also depicts a sleeper (Fig. 5, panel no. 8 on Fig. 1), in this case a priest having a vision in which an illustrious man conveys a message about the miraculous power of the blood and water relic . The priest’s body and bedding are far more composed, and he is clearly having a much more serene night. He even has a fancy pillow under his head (the head is a restoration).
Benedict reported that when Stephen was being attacked by the demon, “he used to cry out in his sleep and call to his servants by name, ordering them to run to his aid. Those who had been called, and also those around him who had been woken by his order, shook the shouting man roughly, and they raised him up so that he sat or stood erect. He had told them many times that he should be made to stand on his feet, and he asked them to wake him by pulling hard on his hair.” The figure in the centre coming to Stephen’s aid now has the head of a female, but his exposed leg and feet underneath the bed show clearly that he is a male servant. Though feet and the bottom of his cloak are marked as restored pieces in the CVMA catalogue, these are original. His boots are elaborately decorated with painted parallel lines round the tops and polka dots along the sides and bases. Similar decorative elements on the boots of the walking pilgrims on the pilgrims on the road panel (Fig. 1 no. 5).
One of the servant’s hands is on Stephen’s torso, while the other must have been pictured on the same piece of glass that showed Stephen’s head. The replacement piece shows this hand as lying flat alongside Stephen’s face, but since Benedict’s story states explicitly that the servants would “pull hard on his hair” to wake the knight, the original hand might well have been pictured as grabbing hair and pulling. If Stephen’s head was pictured as facing downwards, his hair would have been at just the point the sleeve of the servant ends now. Though the drama of the moment is now obscured by the disappointing replacement heads, the forty-five degree angle at which the bodies of the servant and the demon are depicted was no doubt deliberately designed to convey the rushing speed of both figures, one attempting to harm the knight and the other to rescue him. The bent elbow on the flying demon’s torso is in the right location for part of his hand to be visible at the servant’s shoulder. A hand attempting to hold back the servant’s approach might well have been depicted on the piece of glass depicting the demon’s head.
Benedict’s description of Stephen’s miracle makes it clear that servants, in the plural, came to his aid. A small piece of curved yellow glass located beneath the white demon’s rear might be the remains of a second servant figure. It is the right shape and angle to portray a knee and upper shin of a person positioned just behind the demon. An alternative explanation is that this yellow glass was part of another piece of bedroom furniture or of drapery, but if so, it is difficult to explain why it would be curved and without any signs of painted decoration. There are two pieces of replacement blue ground behind the white demon’s back and head in which the shoulder and head of such a figure could easily fit. If a head or shoulder of this figure were still present on the panel when it was restored, the choice may have been made to erase it with replacement blue ground. This theory gains some support by the fact that the two blue pieces behind the white demon are the only pieces of replacement ground in the entire panel.
nV’s inscriptions and the dating of nV to the mid 1180s
The inscription to the demons panel is listed as VEXA | T | . . . E | IO | NI | O . . in the CVMA catalogue. Much of the paint is gone from the middle part of the inscription, and the addition of backplating, though well intentioned, has obscured the remains of the original lettering . It was only with the close examination possible in 2018 that the inscription could confidently be read in full: VEXAT | USDE | MO | NIO :’ , or vexatus demonio, “vexed by a demon.” The U in VEXATUS, which currently looks like an O with the backplating, is formed in an unusual way. The first minimum is upright while the second is curved (see Fig. 6, which shows this part of the inscription in raking light). The letter U is formed in the same way on another panel in nV (Fig. 1 no. 9, in the word SANGUIS).
This short and straightforward inscription, “vexed by a demon,” is typical of nV’s panels. The panel picturing pilgrims on the road, for instance, reads PERE[G]RINI S[T], “pilgrims of the saint.” The longest inscription in nV is found on the scroll held by the visionary figure in the panel discussed above (Fig. 5). With the abbreviations expanded, it reads Cantuar[iam] haustu sanguinis s[anc]t[i] [con]valeat, or “At Canterbury, with a drink of the blood of the saint, he would be healed.” The nV inscriptions are strikingly different from those found in the other surviving miracle windows on the north and south aisles. There, every panel has an inscription, and they are all in the form of rhyming Latin hexameter verses. The 2018 investigation determined not only that there are no hexameter verses in nV’s panels, but also that two of nV’s panels, nos. 6 and 7 on Figure 1, were not given an inscription at all .
The ways in which inscriptions appear – or don’t – on nV’s medieval panels are much more typical of the use of inscriptions on the typological windows in the cathedral’s choir than the other miracle windows. Other aspects of nV evocative of the typological windows are the tall, stately, and rather stiff figures, the layering of spatial fields, the ways in which groupings of figures and animals have stacked-up heads, the treatment of decorative elements such as altars and drapery, the representation of complex narratives or events in single scenes, and the use of distinctive types of glass. The deeply rich blue ground of the typological windows, for instance, was used for the two openings on Becket’s tomb on panel 9 on Figure 1, and it is no-where found in the other miracle windows.
It looks highly likely that a workshop related to the typological windows was responsible for the glazing of nV. The typological windows are usually dated ca.1178-80 . The vaulting of the Trinity Chapel was completed ca. 1182, but glazing could not have progressed for too many years. In the later 1180s, a major dispute broke out between the monks of Christ Church and the archbishop of Canterbury. This dispute persisted for well over a decade and was so severe that it must have resulted in a lengthy halt to the glazing and the completion of other decorative elements of the Trinity Chapel . Window nV was very likely glazed before this interruption. Its medieval panels, including the demons panel, are best dated to the mid 1180s. This is slightly earlier than the dating suggested in the CVMA catalogue (ca.1190-1206) .
Evidence of strap lead repairs from the early 1660s
An unexpected element of the 2018 analysis was the discovery of numerous raised strips on the exterior faces of all of nV’s medieval panels and occasionally on the interior faces as well. The white arrows on Figure 7 show where these strips are found on a portion of the demons panel. These strips appear where strap leads, 8-10mm in width, were soldered onto the medieval lead matrix, most likely with putty placed beneath them. These straps were not applied to cover breaks or fractures in the glass, which is where one would typically expect to find them. Instead, the purpose of these strap leads must have been to enhance structural integrity. They had the unintended consequence of protecting the strips of glass beneath them from wear and corrosion. The buttressing straps must have remained on the panels for many years, because on some types of glass, particularly certain whites, the former locations of the straps are so pronounced that they appear as significantly raised welts. The straps were applied with little concern for the panels’ readability. They were placed directly through the middle of heads and the bodies of figures, as well as across inscriptions. When the straps were in place, the panels would have looked far more chopped up than they did originally or do now.
The team concluded that these strap leads were most likely placed on the panels in the 1660s . After the dilapidations of the Interregnum, the cathedral’s windows were the subject of repair en masse. We know this from a rich set of glaziers’ bills surviving in the cathedral’s archives. One, dated 8 May 1661, concerns the six windows of the north aisle of the Trinity Chapel ambulatory. The bill lists the footage of each window in turn, starting with nII and working back to nVII. It is prorated in terms of size, the repairs costing 6 pence per square foot . This is highly suggestive of the kind of expeditious repairs indicated by the placement of strap leads on nV’s panels. Very similar ghosts of strap leads were found on the sole surviving panel from nVI, now held by the Fogg Museum in Cambridge, Massachusetts. It is not yet known whether they are also to be found on the other medieval panels in the miracle windows. Seliger and the author would be very pleased to hear from readers of Vidimus if they have seen signs of similar repair work.
Samuel Caldwell Jr’s restoration of nV in 1919-1920
It is highly likely that nV’s medieval panels were still in their original lead matrices with the attached strap leads into the early twentieth century. The glazier George Austin Jr releaded nIV, nIII, and nII in the miracle window sequence in the mid 1850s, but though he added two panels to fill in gaps in nV, he did not relead nV’s medieval panels. Nor did his son, Samuel Caldwell Sr., despite the fact that he removed one of Austin Jr’s replacement panels in nV and put in five new ones of his own (four of these are still in the window, Fig. 1, nos. 1, 2, 3, and 10). It was left to Caldwell Sr’s son, Samuel Caldwell Jr, to perform the first major overhaul and releading of nV’s medieval panels .
Caldwell Jr submitted an estimate for “repairing and releading” nV to the cathedral’s Chapter in May 1919 . Though he does not mention it in his estimate, this work would include the return of two panels of nV that had long resided in the cathedral’s Great South Window (window no. SXXVIII). One of these was the demons panel. In his 1925 guidebook, Mason noted that the demons panel and the one immediately beneath it (i.e., no. 7 on Fig. 1), “were placed at the end of the 18th century in the South Transept window,” and “were brought back to their original place in 1920” . Though the CVMA catalogue casts some suspicion on Mason’s statement, the demons panel and its companion are indeed discernible in the centre of the Great South Window (positions 7e and 8e) in a pre WWI photograph . The Chapter Act Book for 1919 notes that the Dean and Chapter had decided to take the opportunity of the glass to “correct mistakes etc. that had occurred when the glass was placed in some of the windows on former occasions” .
One item on the agenda was evidently returning the demons panel and its companion to nV. The Great South Window had been arranged with glass collected from around the cathedral in 1792 . Most of the glass put there were ancestor figures from the clerestory, and it is not clear why the two nV panels were moved there unless it was simply to create a contrasting central focal point. The two panels were too broad to fit into the transept window’s lights, and so both of them lost their triangular sides at this point. It does not appear, though, that they were releaded: they were simply cut down and moved.
When Caldwell Jr worked on nV in 1919/1920, he dismantled its medieval panels, including the two from the transept window. He discarded their matrices and strap leads, put in replacement pieces, and reconstructed parts of the panels that he found confused. He did not make changes to three of the four panels that had been made and placed in the window by his father aside from moving them to a less visible position at the top of the window (now panels no. 1, 2, and 3, Fig. 1). He acid-etched and altered pieces in the fourth of Caldwell Sr’s creations (Fig. 1 no. 10, actually made up of three separate panels), most likely with the intent to make them blend in better with the medieval panels placed immediately above.
Caldwell Jr did more work on some of nV’s medieval panels than others. For the demons panel, his work was relatively light. He supplied new triangular sides to make good what the panel had lost when it was moved into the transept. He replaced all of the heads. This would be an unfortunate loss under any circumstances, but is particularly so as his replacements are awkward and unsatisfying. He also put in two pieces of replacement ground. As discussed above, this may have been in the place of a head and upper torso of the figure of a second servant, which perhaps Caldwell Jr felt were too degraded or cut down to be readable.
Conclusion: A proposed reconstruction of the demons panel
Figure 8 is Seliger’s proposed reconstruction of the original appearance of the demons’ panel. A male head has been provided for the servant rushing to Stephen’s aid, and Stephen himself has the downward turned head that is more common for sleeping figures in Canterbury’s medieval glass. The reconstruction includes a head and part of the torso of a second servant behind the bluish-white demon. New heads for the demons have been supplied that are based on medieval models. The sides of the panel are the most speculative parts of the reconstruction. It is possible that more of the bodies of the demon on the left and the (proposed) servant on the right would have been pictured on these sides. Rather than attempting to reconstruct such figures, columns have been placed on the panel like to those on the vision panel (Fig. 5).
This reconstruction, made with Photoshop, is hypothetical. We have no record of the appearance of the demons panel before Caldwell Jr restored it. Nevertheless, such a reconstruction is helpful in suggesting the panel’s original vitality. The heads supplied by Caldwell Jr for the red and the bluish-white demons are particularly inept. Replacements based on medieval models amp up the frightening energy of the image considerably, and proper heads on the standing servant and the sleeper make the emergency depicted on the panel more readable. In the reconstruction, too, it is more noticeable how the glaziers juxtaposed human heads and demon heads on the left, centre, and right of the panel. Such juxtaposition heightens the menacing threat those demons pose to the sleeping Stephen.
This is the only surviving panel in the entire miracle window sequence that depicts demons. Though panels could, of course, have been lost, it is nevertheless striking that glaziers in the later workshop decided not to portray demons in their depiction of the cure of an insane young man from Fordwich (nIV 14 and 15). A fleeing demon from a person’s mouth or sometimes his or her rear was a standard formula for representing the cure of the insane, but here, in the panel representing his cure (Fig. 9), the young man simply kneels, his bonds removed from his hands. Importantly, no demons are mentioned in Benedict’s story of the young man’s cure.
Here again the glaziers seem to have closely inspected the contents of their written source when making decisions about the panel’s design. Further investigation of the miracle windows is planned to coincide with the 800th anniversary of Becket’s translation and the 850th anniversary of his martyrdom in 2020. This work will no doubt reveal further evidence of the extraordinary care and skill the Canterbury’s glaziers applied to the miracle windows’ composition and narrative design.
Acknowledgements: I am most grateful to the glaziers at Canterbury’s Conservation Studios, especially to Leonie Seliger, for the opportunity to work alongside them in the examination of the glass of nV. The project would not have been possible without the generous financial support of the Friends of Canterbury Cathedral. Nick Teed of the York Glaziers Trust, headed by Sarah Brown, kindly supplied his photographic expertise. A Visiting Fellowship from the British Academy at the University of Kent from July-December 2018 funded my residence in Canterbury of the duration of the project. I would also like to extend my gratitude to the editors of Vidimus for the invitation to submit this article.
 The fundamental works on Canterbury’s medieval glass are M. Caviness, The Windows of Christ Church Cathedral, Canterbury, Corpus Vitrearum Medii Aevi: Great Britain (London, 1981), and her earlier study, idem., The Early Stained Glass of Canterbury Cathedral (Princeton, 1977).
 For the purposes of this article, the numbering of the figural panels of nV has been simplified from the CVMA catalogue. The CVMA numbering is found in parentheses below the revised numbering.
 The author is preparing a revised CVMA volume for Canterbury’s miracle windows which will provide new descriptions and restoration charts for each panel.
 See Windows of Christ Church, pp. 178-179 and fig. 246a.
 Stephen’s story is found in Benedict of Peterborough, Miracula S. Thomae, ed. J. C. Robertson, Materials for the History of Thomas Becket vol. 2 (London, 1876), pp. 21-281, I.13, pp. 44-45. All translations of passages from Benedict’s story in this article are my own.
 A. J. Mason, A Guide to the Ancient Glass in Canterbury Cathedral (Canterbury, 1925), p. 29.
 For the story about the Cluniac monk, see William of Canterbury, Miracula S. Thomae, ed. By J. C. Robertson in Materials for the History of Thomas Becket vol. I (London, 1875), III.8, pp. 380-381.
 For comparative imagery, see Debra Higgs Strickland, Saracens, Demons, and Jews: Making Monsters in Medieval Art (Princeton, 2003), esp. pp. 61-78, and D. A. Sprunger, “Depicting the Insane: A Thirteenth-Century Case Study,” Marvels, Monsters, and Miracles: Studies in the Medieval and Early Modern Imaginations, ed. By Timothy S. Jones and David A. Sprunger (Kalamazoo, MI: Medieval Institute Publications, 2002), pp. 223-241. See also Marek Tamm, “Saints and the Demoniac: Exorcising Rites in Medieval Europe,” Folklore: Electronic Journal of Folklore 23-24 (2003): 7-24.
 On this condition, see the notable articles of William F. MacLehose, “Fear, Fantasy and Sleep in Medieval Medicine,” in Emotions and Health, 1200-1700, ed. Elena Carrera (Leiden, 2013): 67-94, and Stephen Gordon, “Medical Condition, Demon or Undead Corpse? Sleep Paralysis and the Nightmare in Medieval Europe,” in Social History of Medicine (2015): 425-444.
 This image was thought to be linked to a story about a man with a toothache (see Windows of Christ Church, p. 179), but the inscription on the panel, read in full for the first time in the 2018 investigation, definitively ties it to the story of William of London, a miracle that is also portrayed in the panel now to its right (Fig. 1 no. 9). For the William of London story and further bibliography on the Becket blood and water relic, see Rachel Koopmans, “Water Mixed with the Blood of Thomas: Contact Relic Manufacture Pictured in Canterbury Cathedral’s Stained Glass,” Journal of Medieval History 42:5 (2016): 535-558.
 Derek White, who placed the backplating on the inscription, described his intentions in his conservation report: ‘The inscription did not relate to this panel and also made no sense in its Latin text. It was decided however to enhance the letters which were recognizable, for future scholars to make their own judgement’.
 On panel no. 6, the white line underneath the figures is painted with pseudo-cufic ornamentation.
 See Windows of Christ Church, p. 79.
 The best and fullest description of this dispute is, despite its age, William Stubbs’s introduction to Epistolae Cantuarienses: Chronicles and Memorials of the Reign of Richard I, vol. 2 (London, 1865).
 Windows of Christ Church, p. 178.
 On repairs to the Cathedral Fabric at the Restoration, see Jeremy Gregory, “Canterbury at the Ancien Régime: The Dean and Chapter, 1660-1828,” in The History of Canterbury Cathedral … , pp. 208-214, and Margaret Sparks, “The Refitting of the Quire of Canterbury Cathedral, 1660-1716: Pictorial and Documentary Evidence,” Journal of the British Archaeological Association 154 (2001), 170-190.
 See Canterbury Cathedral Archives, DCc-TV8-1-26.
 This is evident both from early photographs and from inspection of the panels themselves. The restorations are all those characteristic of Caldwell Jr.
 CCA DDc-CA/16, p. 439, entry for 31 May 1919.
 Mason, A Guide to the Ancient Glass, p. 29.
 There is a payment “To Simmonds the Vesturer for his Care in Arranging the painted Glass for the Great South Window by Order” in the Treasurer’s Book for 1791-92: see CCA DDc-TB-127, p. 62.
 Chapter Act Book for 1910-1919, CCA DDc-CA/16, p. 418, entry for 4 January 1919.
 Positions 7e and 8e. See again the payment “To Simmonds the Vesturer for his Care in Arranging the painted Glass for the Great South Window by Order” in the Treasurer’s Book for 1791-92: see CCA DDc-TB-127, p. 62.