The Devil’s Looking Glass

A black stone in the British Museum carries a formidable reputation: John Dee supposedly used it to call spirits. In 2021, an investigation established where its obsidian came from. The harder question is how a Mexican mirror became an English magician’s relic.¹²

A scholar with too many instruments

The mirror is roughly nineteen centimetres across, almost circular, with a short rectangular tab pierced by a hole. Its two faces have been ground and polished; its edges show the work of someone turning volcanic glass into a finished object. Without the label and the name attached to it, a visitor could recognise a remarkable piece of craftsmanship without knowing anything of Elizabethan magic.²

The fitted case changes the encounter. A paper label in Horace Walpole’s handwriting identifies the stone with Dee’s spirits, then draws on Samuel Butler’s satirical poem Hudibras to call it the Devil’s Looking-glass. Walpole acquired it in 1771, long after Dee’s death. His inscription made a claim about the past, and helped give the object its future.¹

Dee had lived from 1527 until 1608 or 1609. He advised on navigation, studied mathematics and astronomy, collected books and equipment, and belonged to the intellectual world around Elizabeth I. His interests also embraced astrology and alchemy. The surviving books at the Royal College of Physicians contain his signatures, annotations and drawings: direct physical evidence of an active reader, following subjects far beyond the supernatural.³

The habit of presenting him as the queen’s magician can make all those activities look like preparations for a séance. His library suggests something larger. A reader could move between geometry, ancient history, natural knowledge and alchemy without accepting the boundaries a modern library would put between them. Those connections made sense within his pursuit of knowledge, even when the conclusions he hoped to reach no longer persuade us.³⁵

Alchemy, in particular, involved more than the promise of gold. Historian Jennifer Rampling has traced Dee’s reading, copying and circulation of alchemical works, alongside the materials and apparatus kept at Mortlake. Texts supplied recipes and authority; practitioners tested procedures and exchanged information. Surviving marginal notes and correspondence make this labour visible, although much of its physical equipment has vanished.⁵

During the 1580s Dee also tried to obtain knowledge from angels. He employed intermediaries, or scryers, who looked into stones or crystals and described what they said they saw. Edward Kelley became the most famous of them. Dee recorded the communications, leaving a body of writing that allows historians to examine their activities without accepting their explanation of the source.²⁴

Scrying means seeking visions or information through concentrated attention to a reflective or translucent surface. The object might be a crystal, a mirror or another suitable medium. Calling something a shew-stone tells us how a person understood its purpose; it does not give us a modern material description precise enough to identify one surviving object. That ambiguity would become a central problem for the black mirror.⁴⁶

There were several instruments to confuse. Christopher Whitby’s study of the early spiritual records identified references to different crystals, including a stone in a frame. He also distinguished a concave optical glass, capable of producing unusual reflected images, from the flat obsidian disc. His conclusion about the museum mirror was cautious: the connection was possible, but circumstantial.⁶

The black surface nevertheless seems to suit the familiar Dee so well that the attribution can feel self-evident. An Elizabethan occultist, an unreadable stone and a promise of hidden knowledge appear to complete one another. The mirror’s material history offers a firmer starting point. Before anyone in England could give it an angelic or diabolical identity, someone in Mexico had to make it.²

Before the English magician

Obsidian is volcanic glass. Its usefulness in Mesoamerica extended across everyday work, warfare and religious life. A sharp implement and a polished mirror could begin with the same broad class of material, then demand different skills and acquire different purposes. The finished disc belongs to that history of manufacture, in which stone was selected, shaped and worked until it could return an image.²

The archaeological study published in Antiquity in 2021 was led by Stuart Campbell, with Elizabeth Healey, Yaroslav Kuzmin and Michael D. Glascock. Their investigation examined the mirror associated with Dee alongside three other British Museum objects. They also placed those pieces within a wider group of Mexican mirrors and polished slabs in collections around the world.²

Several details make the comparison useful. Two of the circular mirrors have square tabs, including the Dee example. Another has a tab shaped more like a teardrop. The rectangular slab has one polished face, a rougher underside and deliberately worked edges. These are variations in the objects themselves, surviving beneath the names of European collectors through whose hands they later passed.²

The tabs could have allowed circular mirrors to be attached to a person or a sculpture. Chipping around the hole in the Dee mirror may have resulted from suspension. Neither observation tells us exactly where this individual object once hung. It could be tempting to place it on a priest or a deity, but its original setting has not survived in the records consulted for this article.²

The researchers turned to images and descriptions as well as the stone. Indigenous depictions show circular mirrors in the cultural world from which these objects emerged. An early colonial account compiled by the Franciscan Bernardino de Sahagún describes specialist mirror-makers cutting, abrading and polishing. The activities recorded there help explain the skilled work embodied in a finished reflective surface.²

The most evocative association is with Tezcatlipoca, whose name is commonly translated as Smoking Mirror. In representations of the deity, mirrors appear at the head, chest or back, and can replace his missing foot. The 2021 study discusses their relationship to revelation, prediction and power. Such images help recover what a mirror could signify within Mexica religious thought.²

Those associations do not identify the owner of this particular disc. Nor do they make every obsidian object a sacrificial implement. A material used across many parts of life cannot be assigned one function merely because its most dramatic religious meaning is well known. The mirror’s shape and cultural comparisons support a Mexican attribution; its lost archaeological context still limits the story we can tell about its use.²

Dating brings the same difficulty. Campbell and his colleagues considered most circular examples likely to belong to the Late Postclassic period, while allowing that manufacture continued after the Spanish conquest. An object surviving in a collection without a secure excavated context cannot easily be placed on one side of that historical boundary. The chemistry would help locate its stone, but would not date the hands that polished it.²

What survives from Mexico? The material, the worked form and comparisons with other objects and images. The maker’s name, workshop and original owner remain unknown.²

A crossing without a passenger list

The Spanish conquest supplies the broad historical route by which Mexican objects entered European collections. It does not provide a passenger list for this mirror. After the destruction of the Mexica capital and the consolidation of Spanish rule, artefacts moved through networks of tribute, seizure, gifts and collecting. Their arrival in Europe could preserve a striking object while stripping away the circumstances that had explained it.²

Lists of early shipments included mirrors. The Antiquity paper notes one consignment containing eight mirrors of different kinds, entrusted to Diego de Soto. That establishes that such objects could travel into Habsburg Europe in the period relevant to Dee. No mark on the British Museum disc identifies it as one of those eight, and no receipt connects it with that shipment.²

Even the apparent dividing line between conquest loot and later export is difficult to apply to an unprovenanced object. The study notes that some artefacts sent to Europe were commissioned by Hernán Cortés. Indigenous craftspeople continued making things in altered political and economic conditions. An unfamiliar object in a European cabinet might therefore embody an older technique, a new commission, or both.²

No surviving record identifies this mirror’s Atlantic crossing. Mirrors travelled to Europe in the years after the conquest, but the ship, year and person responsible for carrying this one remain unknown.²

Dee had opportunities to encounter unusual objects on the Continent. The researchers discuss his studies at Louvain in 1548–1550 as one possible occasion, and his later connections with central Europe as another. These are proposed settings for an acquisition, based on his movements and contacts. They are not two competing eyewitness accounts of a purchase.²

His continental life is better documented in other respects. In 1583, Dee and Kelley left England for east central Europe with books from Dee’s collection. Rampling’s research follows the alchemical texts, experiments and patronage that developed around them. The men’s interests overlapped, but Kelley’s growing reputation as a practical alchemist increasingly offered patrons something different from Dee’s spiritual ambitions.⁵

Such evidence makes a mobile scholarly world tangible. Books could be copied, translated, lent and presented to prospective patrons. Recipes and reputations travelled together. It is easy to see how a rare Mexican mirror might also circulate through courts and collections. The ease of imagining a transaction, though, cannot substitute for the document that would establish it.²⁵

Would Dee have understood its Mexican meanings? Campbell and his colleagues suggest that his interest in the Americas could have made the object’s associations attractive to him. They also point out that Europe already had traditions of magical mirrors. An imported black surface could enter an existing practice even if its new owner understood little of the beliefs connected with it in Mexico.²

The possibilities differ. Dee might have acquired the mirror because he knew something of its foreign religious significance; he might have valued an unusual material within familiar European practices; or the connection with him might have been attached later. None requires us to imagine an unbroken occult tradition passing intact across the Atlantic. The surviving evidence records an object moving between societies whose expectations of it could be very different.²

The collector and the missing years

By the eighteenth century, a story had settled around the stone. Horace Walpole’s description of Strawberry Hill placed it in the Great North Bedchamber. He called it a speculum of kennel-coal, using a material identification that the modern obsidian analysis would supersede. He also described it as an instrument with which Dee had deceived the public.⁷

Walpole’s account carried a named chain of owners. The mirror had been in the collection of the Mordaunts, earls of Peterborough, then passed to Lady Elizabeth Germain and the duke of Argyll. The duke’s son, Lord Frederic Campbell, gave it to Walpole. The Lewis Walpole Library’s catalogue records the gift in 1771, and identifies the surviving object by its British Museum registration number.⁷

The fitted case and Walpole’s label provide physical anchors for that part of the history. Their evidence is valuable because it locates the attribution in a known collector’s hands. It also fixes its distance from the events it claims to explain. Walpole could report a tradition about Dee; he had been born more than a century after the Elizabethan scholar.¹⁷

In a letter preserved in Yale’s edition of his correspondence, Walpole described the stone’s polished surface and its strong leather case. He referred to the earlier Peterborough catalogue and repeated the succession of owners. That was enough for him to declare:

The confidence is revealing. Walpole’s evidence was an inherited description supported by a history of possession. The editors’ note on the same page adds an obstacle for later researchers: the relevant catalogues of the Peterborough and Germain collections had not been traced. A document invoked as proof could no longer be put beside the mirror and examined.⁸

There is an earlier clue, and it prevents the story from collapsing into a simple eighteenth-century invention. Campbell and his colleagues discuss testimony from a lawsuit in 1624. Thomas Hawes recalled seeing a round, flat, crystal-like stone in the house of John Pontois before Pontois left to serve the Virginia Company at the end of 1618. Pontois said it had been brought to Dee by an angel and had enabled him to know strange things.²

The testimony matters because Pontois received books and possessions from Dee. It places a claim about a remarkable flat stone within the circle of his material heirs much closer to his lifetime. Hawes was reporting what Pontois told him about the stone’s history. The angelic delivery remains a belief within the testimony, not a fact established by the lawsuit.²

Could that stone be the mirror now in the museum? Its round, flat form makes the suggestion worth taking seriously. The identification remains incomplete: the description lacks measurements or another feature unique to this object. Pontois’s collection was dispersed in 1625 and 1626, and the researchers suggest the first earl of Peterborough may then have acquired it. No surviving transfer is supplied to close that link.²

After Walpole, the record becomes a history of collecting and sales. The Strawberry Hill dispersal in 1842 sent it into another series of private hands. The British Museum bought it from the Reverend R. W. Stannard in 1966. Those later records allow the museum object to be followed backwards through the market, but cannot by themselves bridge the earlier gap to Dee’s lifetime.⁷⁸

Reading the stone with X-rays

The scientific investigation approached a problem that paper records could not settle. What if the stone itself had come from somewhere quite different? A European obsidian source would have complicated the Mexican interpretation, and strengthened the possibility that an object resembling a Mexican mirror had been made or substituted in Europe. Its black appearance alone was insufficient to decide.²

The team used portable X-ray fluorescence, usually shortened to pXRF. The instrument measures the elemental composition of a surface. Geological obsidian sources can have distinctive combinations of elements, allowing an artefact’s composition to be compared with material collected from known deposits. The purpose was to identify where the raw material originated, without cutting a sample from the mirror.²

They took repeated ninety-second readings at two locations on each museum object. Repeating measurements allowed them to check whether different spots gave consistent results. Reference standards helped calibrate the instrument. The researchers then compared the artefacts with geological samples from four Mexican source areas: Otumba, Pachuca, Ucareo and Zaragoza.²

The paper records a practical limitation. The museum surfaces could not be cleaned, leaving some possibility that centuries of handling had introduced contamination. Readings from different positions were consistent, which reassured the researchers. They did not pretend that an old museum object could be handled like a newly prepared laboratory specimen, untouched by everything that had happened since its manufacture.²

Across several elements, the four objects separated into two groups. The mirror associated with Dee matched the Pachuca source, as did the other circular mirror with a square tab. The teardrop-tab mirror and the rectangular slab matched Ucareo. The correspondence between shape and source was suggestive, although four objects could not establish a universal manufacturing rule.²

Pachuca lay within Aztec territory and was heavily exploited. Ucareo lay in the territory of the Tarascan state, to the west. The latter source sometimes occurs in slabs suited to making mirrors. Taken together, these results put the museum pieces within the material geography of central Mexico, while allowing more than one source and manufacturing tradition to emerge.²

The conclusion was a real advance. A Mexican origin had long been proposed from the objects’ appearance and comparisons; the chemical evidence supplied an independent way to test it. For the Dee mirror, Pachuca replaced a broad cultural attribution with a geological source assignment. Walpole’s label had supplied a famous English name. The instrument supplied evidence of a Mexican beginning.²

It supplied no manufacture date. A chemical match cannot say whether the mirror was polished before the conquest or during the first colonial decades. It cannot identify a workshop, name a craftsperson or tell us who bought the finished object. Nor can it find a missing transaction between Pontois and the Peterborough family. Those questions require evidence of another kind.²

Two different histories: Geological provenance asks where the stone formed. Ownership provenance asks who held the object. The first can be tested chemically; the second still depends on records and identification.²

Was it really Dee’s mirror?

The strongest broad account brings several forms of evidence together. The object resembles Mexican mirrors; its obsidian matches Pachuca; an early witness described a round, flat stone associated with Dee among Pontois’s possessions; and later collectors preserved a Dee attribution. Campbell and his colleagues concluded that the historical work strengthened the association. That is more persuasive than a story resting on Walpole alone.²

It is still possible to separate the object’s Mexican identity from its English owner. A genuine imported mirror could have acquired a famous name after Dee’s death. The period between his lifetime and the secure collecting history is precisely where a mistaken identification could enter. Chemical analysis of Mexican stone would produce the same result whether Dee had owned it or someone else had.²

The early Hawes testimony makes that alternative less simple than it first appears. It requires an explanation for the Dee-associated flat stone in Pontois’s house, even if that stone was a different object. The testimony deserves weight without being made to prove more than its description permits. A familiar supernatural claim can contain useful evidence about a possession while remaining unreliable about the possession’s supposed powers.²

Whitby’s objection concerns identification from Dee’s own records. The surviving references he examined did not uniquely pick out the museum disc. The best-known optical glass was concave; other stones were crystals. An image of Kelley consulting a flat black mirror throughout the conferences therefore goes further than the identification available in those passages. The research does not make ownership impossible, but it denies us that easy documentary match.⁶

A European-source copy is now a weaker explanation. The material tested matched Mexico, so a substitute made from European obsidian would not account for the result. One can imagine Mexican raw material worked later in Europe, but the paper supplies no evidence for such a scenario. The known Mexican forms and geological result already offer an account without introducing an otherwise undocumented workshop.²

The dating uncertainty survives alongside this stronger geographical conclusion. An early colonial mirror could be just as Mexican as a pre-conquest one. Its manufacture might reflect continuity of Indigenous skill under changed rule. Calling the object Aztec is a useful museum and archaeological classification, but cannot establish the individual maker’s identity or the circumstances in which that person worked.²

There are also limits to identifying its first function. It resembles mirrors associated with religion and elite display, yet no excavated setting ties it to a particular ritual. Circular mirrors could be attached to bodies or sculptures; some may have been incorporated into colonial Christian crosses. The possibility of reuse widens the object’s history without allowing a precise sequence to be written for this example.²

The rectangular slabs offer a warning against assuming that every polished black surface meant the same thing. The research discusses their use as portable Christian altars and, in later European hands, as supports for paintings. A shared material could serve purposes that new owners understood very differently. The stone could remain physically recognisable while the explanation given for keeping it changed.²

The Dee attribution stands in that space between a good historical case and a complete chain of custody. Dismissing it as an empty legend would discard the early testimony and the object’s collecting history. Treating it as certain would conceal the missing identification and transfers. The open question is narrow enough to investigate: whether this particular surviving Mexican disc is the flat stone associated with Dee soon after his death.²

Angels, demons and a polished surface

The British Library’s catalogue describes Sloane MS 3188 as Dee’s conferences with angels, beginning in December 1581 and continuing into 1583. It identifies the use of a crystal ball, and records a later copy made by Elias Ashmole. Some leaves are missing.⁴

The arrangement could produce an imposing record. Tables, letters, names and instructions gave the communications a form that could be copied and studied. Dee’s activities belonged to a search for hidden order, connected with his other attempts to understand nature and divine knowledge. Their persistence shows his commitment to the undertaking. It does not independently authenticate the speakers he believed were addressing him.⁴⁵

Rampling’s account of the continental years also shows how spiritual and practical expectations could diverge. Patrons had an interest in what alchemy promised to accomplish, and Kelley developed a reputation in that field. By the later 1580s, his efforts to leave the scryer’s role coincided with his growing success as an alchemist. The partnership had material pressures and changing opportunities as well as its recorded spiritual ambitions.⁵

Later interpreters brought their own religious purposes. Meric Casaubon published part of Dee’s spiritual record in 1659 as A True and Faithful Relation. As the Jesus College library’s account explains, his preface defended the existence of spirits while attacking claims to religious inspiration that he considered dangerous. He accepted Dee’s sincerity, but understood the episode through the danger of false revelation.⁹

Walpole found a different use for the relic. His Strawberry Hill description presented it as evidence of ancestral credulity and public deception. The case label kept the Devil’s Looking-glass phrase alive, borrowing the authority and entertainment of satire. Both the believer and the amused collector could value the same object, even while disagreeing about what its past demonstrated.¹⁷⁹

The 2021 investigation recovered a part of the mirror’s history invisible in its English labels: the source of the stone itself.²

What the mirror can now tell us

As of 7 November 2026, the accessed research establishes Pachuca obsidian for the mirror associated with Dee. Its exact date of manufacture, original setting and route into European possession remain unresolved. The Dee connection has historical support, including the early Pontois evidence, but no continuous documented chain.¹²

Research in Mexico continues to make the wider material history more precise. In May 2025, INAH reported analysis of 788 obsidian objects from the Templo Mayor, dating from about 1375 to 1520. They drew on eight source areas. The work found a more varied pattern of acquisition over time, involving materials from territories outside Mexica control as well as Pachuca’s valued green obsidian.¹⁰

That study concerned other artefacts, not Dee’s mirror. Its relevance is the complexity it restores to the world behind the object. A geological source is a place in a system of production and exchange, not the name of a single people or a guarantee of one ritual purpose. Recovering that system requires excavated contexts of the kind this mirror has lost.²¹⁰

The mirror remains paired with a case bearing Walpole’s description of Dee’s spirits. The label preserves the English name that made it famous; the polished disc preserves the skill of the unknown Mexican craftsperson who made it.¹²