John Dee’s Holy Table in the Manuscript Record

The record describes a made surface

The British Library catalogue identifies Sloane MS 3188 as Dee’s five books of conferences from 22 December 1581 to 30 May 1583. That date range matters because the Holy Table was not delivered as one polished diagram. The record moves through questions, reported visions, measurements, letter groups, and later corrections. Whitby’s transcript lets those actions be read in sequence while keeping Dee’s narration separate from words attributed to figures seen by Edward Kelley.

On 28 April 1583 the record gives material instructions: a square, a border, the number of characters on each side, and a smaller central square. It then records letters appearing in a smaller square after prayer. The document therefore does more than name an object. It moves from dimensions to a reading order and finally to the orientation of the characters. A modern outline that preserves only the finished geometry loses that procedural logic.

The border is a route through letters

Dee calls the side opposite the observers “in fronte Tabulae.” In Whitby’s transcript the reported sequence begins on that side at the right and proceeds toward the left. After supper Dee asks which way the heads of the characters should turn; the answer recorded in the action points them toward the center of the square. Direction is therefore part of the design. Reversing the starting corner or rotating the letters changes how the border is read even when the same eighty-four positions remain visible.

The instruction also distinguishes a planning drawing from the intended furniture. Dee is told to sketch a six-inch square with a half-inch border while making the border on the table itself one inch wide. That difference warns against treating every manuscript proportion as a scale drawing. The drawing can encode sequence and orientation without reproducing the dimensions of the vanished tabletop.

Worked reading: one table instruction becomes the 1659 engraving

Follow one instruction through the witnesses: begin on the in-fronte side, start at the right, move left, and turn each character’s head inward. In the action these clauses arrive at different moments. First the side and direction are named; later Dee explicitly asks about orientation. The record reads like a construction decision being resolved. The 1659 engraving compresses those decisions into a continuous square border. Its viewer sees the result at once and no longer experiences the pauses, questions, or alternatives that produced it.

The comparison explains what the print preserves. It keeps four groups of twenty-one border positions, makes the circuit visible, and presents the character forms as a unified design. The heads of the printed characters face toward the interior, so the plate can be checked against the directional instruction rather than accepted merely because it looks old. At the same time, the engraving is a printer’s visual witness issued seventy-six years after the action. Its regular line, spacing, and monochrome surface are properties of print, not evidence that the lost wood looked identical.

Peterson’s collation identifies a possible dotted-character problem near the upper-left border and notes that neither Dee’s diagram nor the engraving keeps the specified proportions. Those are editorial findings that require a source-specific comparison; they are not facts supplied by the museum catalogue alone. The secure worked result is narrower: manuscript instructions define a reading path and inward orientation, and the 1659 plate converts that path into a reproducible border.

The print does not collapse the surviving objects

Oxford’s History of Science Museum states that the painted wooden original is lost and that its mid-seventeenth-century marble table was based on the 1659 engraving. The marble therefore demonstrates reception of the printed design. It cannot independently prove the paint, wood grain, measurements, or lettering of Dee’s missing furnishing. A later description by Ashmole can add color and measurement evidence, but it remains another witness with its own date.

The British Museum wax discs and the show-stone belong to separate object histories. A large wax Sigillum was associated with the table’s center and smaller discs with its feet, yet none is a fragment of the painted border analyzed above. Keeping manuscript instruction, printed plate, marble copy, wax group, and viewing stone separate reveals how the Holy Table functioned as text, diagram, furniture, and later collectible at different stages.

Sources and witness notes

1. British Library · Sloane MS 3188 ↗Catalogue for Dee’s five books of conferences, 22 December 1581–30 May 1583. The volume is a manuscript witness, not the surviving table.
2. Christopher L. Whitby · John Dee’s Actions with Spirits, volume 2 transcript ↗University of Birmingham transcript of Sloane MS 3188. The 28 April 1583 action on transcript pages 354–357 supplies the square, border, direction, and inward-facing-letter instructions.
3. Joseph H. Peterson · Holy Table witness collation ↗Modern editorial comparison of Sloane MS 3188, the 1659 engraving, Ashmole’s description, and the later marble object; used with its stated uncertainties.
4. History of Science Museum · Holy Table, inv. 15449 ↗Museum account of the lost painted wooden original, the mid-seventeenth-century marble copy, and its relationship to the engraving published in 1659.
5. British Museum · wax disc 1838,1232.90.a ↗Object record describing the large engraved wax disc, two smaller companions, the proposed table arrangement, and the limits of Dee-lifetime attribution.
6. Folger Shakespeare Library · 1659 Relation catalogue ↗Catalogue for Meric Casaubon’s posthumous printed edition; the engraved plate belongs to this later editorial witness.
7. Cornell · digitized 1659 Relation ↗Digitized copy of the printed edition, useful for checking plate placement and the surrounding book.
8. Wikimedia Commons · Holy Table plate ↗The 1659 engraving is marked public domain. The local image is a reduced-size version already used by this site.