John Dee and the Magic Tables in the Book of Soyga

Jim Reeds


We are now in a position to diagnose the copying errors. The mistakes in cells 21 through 30 of line 3 are easily seen to be due to elision of an a from one of the repeating baxi groups; the pattern ends on the right foot again in cell 31 by the insertion of an extra i. We might term this a horizontal eye skip error. The errors in cells 19 through 36 in lines 6, 7, and 8 are seen to result from a vertical eye skip error, as follows. The rightmost 18 cells of lines 6, 7, 8, and 9 of the original are


and of the received copy are


so we see that the copyist deleted the right half of line 6 and duplicated the right half of line 9. There is a transposition error in row 5, cells 4 and 5: the original has lg and the copy has gl. The remaining errors are simple replacements of one letter by another.

5.2. Summary of Actual Errors

In fact all of the types of errors illustrated above occur in both the Sloane 8 and Bodley 908 versions of the tables. There seems to have been one set of original tables which I call A. Our extant versions, Bodley 908 and Sloane 8, seem to have been derived independently from a awed intermediate version which I call C.

The originals A were constructed with the code words as listed in my Table I, by application of the N +f(W) formula; the errors in applying the formula are listed in my Table III. Since errors in applying the N + f(W) formula propagate and spoil everything below and to the right of the error locus, we can be sure that this is the complete list of errors in A. Out of the 46,656 cells in the complete set of tables, only 13 errors were made in applying the formula.
The alternative, that Bodley 908 and Sloane 8 did not share a common original, would require us to believe that exactly these same particular errors (and no others) were committed in working out the originals for both Bodley 908 and for Sloane 8. This is so unlikely under any reasonable model for errors that I reject this alternative in favour of a single shared common original A.

A number of gross eye skip errors were committed in the descent of Bodley 908 and Sloane 8 from A. In Bodley 908's version of T2, row 3, cells 20{35 read axibaxibaxibaxib instead of xibaxibaxibaxiba; that is, an a was inserted at cell 20. In both Bodley 908's and Sloane 8's versions of T24, the right hand half of row 35 was elided and the right hand half of row 34 was duplicated. In both Bodley 908's and Sloane 8's versions of T36, row 3, cells 30 - 36 read baxibax instead of
A's ibaxiba, and row 12 is elided.
I detected seven transposition errors, some unique to Bodley 908 and to Sloane 8, and some shared, as listed in my Table IV.
A tabulation was made of all corresponding places where Sloane 8, Bodley 908, or A were all legible but failed to give unanimous readings of cell entries, except for those involved in the gross eye skips noted above. The tabulation was made again, where all dierences explainable by mere confusion of i/l, u/n, f/s, e/c, or t/r were censored, in an attempt to compensate for possible transcription errors on my part (especially in reading Sloane 8).
The results, in my Table V, again show Bodley 908 and Sloane 8 each have a large number of unique errors in addition to a larger number of shared errors. If either of Sloane 8 or Bodley 908 were copied from the other, the errors unique to the ancestor would have had to have been corrected in the child. Because the text is incoherent, there is no natural 'self repair' mechanism analogous to a scribe's knowledge of orthography or grammar allowing emendation of errors, at least in the large areas of the tables lacking repeating motifs. If both Bodley 908 and Sloane 8 were independently derived from the original A, the 394 (or 223) shared errors would all be the result of accidental occurrence of precisely the same mistakes, independently committed in copying Bodley 908 from A and in copying Sloane 8 from A. This is very unlikely under any reasonable model of copying errors. So we conclude instead that both Bodley 908 and Sloane 8 were derived from a common awed copy, which I call C, of the originals. Because Bodley 908 seems to have fewer disagreements with A than Sloane 8 does, we conclude that Bodley 908 is a more accurate copy of C than Sloane 8 is. Overall, there seems to be a 3/4% copying error rate in going from A to C, a 1/2% error rate in going from C to Sloane 8, and a 1/3% error rate in going from C to Bodley 908.
The same techniques can be used to see what relation Dee's copy of the eight Soyga tables appearing in Sloane 3189, the Book of Enoch, has to Bodley 908 and Sloane 8.
In the rst place, the T13 of Sloane 3189 shows the same mistake in applying the N + f(W) formula (in row 18, column 2) present in the T13 of A. Hence even if not copied directly from Bodley 908 or Sloane 8, the Soyga tables in Sloane 3189 are, like those of Bodley 908 and Sloane 8, ultimately derived from A. A fortiori, they are copies of the Soyga tables, rather than simply creations inspired by, or in the same style as, the Soyga tables.
Secondly, the T2 of Sloane 3189 lacks the gross eye skip error found in row 3 of T3 of Bodley 908. This suggests Sloane 3189 was not copied from Bodley 908, but not strongly so: the eye skip error occurs in the repeating baxibaxi area and could have been corrected by a naive but alert copyist.
Third, looking only at locations where all four of A, Bodley 908, Sloane 8 and Sloane 3189 supply legible values, I found the results in my Table VI. The agree-disagree counts seem to make Sloane 3189 slightly but insignicantly closer to Bodley 908 than to Sloane 8.
Fourth, and more tellingly, the transposition error in T3 of Bodley 908, where there is a kq instead of the correct qk in row 22, is not present in the T3 of Sloane 3189. Unlike the T2 eye skip error, this error is well outside the area of repeating motifs, and so uncorrectable by a naive copyist.
On balance, then, it seems that the Soyga tables in Dee's Book of Enoch, Sloane 3189, are closer in manuscript transmission to Sloane 8 than to Bodley 908.

Assuming that the Sloane 3189 Soyga tables were copied from Sloane 8, the most common copying error was replacing z by x: out of the 477 occurrences of the letter z in the Sloane 8 tables which have corresponding Sloane 3189 versions, it was rendered correctly 441 times, rendered as an x 34 times, and as a q and an r each once. There are 9 instances
where an i was written instead of a y. Overall, there is a 1.5% copying error rate from Sloane 8.
Regardless of which particular manuscript the Book of Enoch got its Soyga table copies from, the questions of why they were copied and what relation they have to the Enochian system are central to furthering our understanding of Dee's relation to the Book of Soyga.
On the one hand it is possible that Dee deliberately copied them (or had them copied) into his notebook (in rearranged sequence: T1, T13, T2, T14, and so on, so both 'Aries' tables were visible on an opening, both 'Taurus' tables visible on the next, etc.) for ready reference, possibly with motives similar to mine in section 3 of this paper, or possibly in order to use them in magical operations. This might have happened some time before 1582, that is, before his 'Enochian' period, in which case their appearance with the Enochian material in Sloane 3189 would be the accidental result of reuse of a largely blank notebook. On the other hand, they might have a more direct connection with the Sloane 3189 Enochian material: they might have been revealed the same way the rest of the Book of Enoch material was (in which case the copying errors could be attributed either laudably to angelic emendation or deplorably to mundane data-entry-clerk error), or they might have been accorded a semi-privileged status, not themselves revealed but worthy of inclusion as an appendix to the Book of Enoch by a principle of virtue-by-association. Even though I see no way to use the methods of this paper to distinguish between these possibilities, I do not hesitate to speculate in the next section about one possible stylistic connection between the Soyga tables and the rest of the Book of Enoch.

6. Comparison with other tables

Large square tabular arrays of letters are quite common in early modern magic works, exhibiting a variety of forms as yet unsurveyed in the scholarly literature. Here I present a brief taxonomy of magic tables according to their internal structure.
The more usual point of view, represented by Yates, pays primary attention to the authors' theories of magic and scant attention to the actual form of the tables: ... in Agrippa's Third Book [on Occult Philosophy] there are elaborate numerical and alphabetical tables for angel-summoning of the type [my emphasis] which Dee and Kelley used in their operations...
These can be seen in Dee's manuscript 'Book of Enoch', British Museum, Sloane MSS. 3189. Cf. the 'Ziruph Tables' in Agrippa's De occult. phil., III, 24. Agrippa was not Dee's and Kelley's only source for practical Cabala, but their minds run on these things within the Agrippan framework. 24
In fact Dee's tables and Agrippa's have completely dierent forms (as can be seen by glancing at my Plates IV and V), so Yates must be using 'of the type' to refer to the authors' intentions and not to their tables' actual appearance or formation.25
My tentative taxonomy begins by crudely dividing all square magic tables into two classes, the small and the large, according to whether they have, say, fewer than fteen rows and columns or more. Among the small tables are those with letters forming words when read either vertically or horizontally, as in the famous square found at Herculaneum,

which are nowadays known as 'word squares'. Word square charms have been in continuous use from Roman times to the present. Many such squares appear in Abraham ben Simeon's Cabala Mystica, which Patai concludes | based in part on an analysis of the text in the squares themselves was written around 1400.26
Small numerical tables like

nowadays known as 'magic squares', have also been used since the late middle ages in Europe and in Asia for far longer as charms or arithmetical amusements. 27 (The numbers in each of the rows, columns, and two main diagonals all add up to the same sum, in this case 65.)
Such a small numerical square appears in the 1514 print Melencolia I of Albrecht Durer (1471{1528); many others are to be found in Agrippa's Book II, where each planet is assigned its own magic square, each square being presented in both Arabic and equivalent Hebrew numerals. 28
As far as I know, all large magic tables in mediaeval or early modern sources are alphabetic. We may divide them into unpatterned and patterned; the latter are subdivided into those in which the form of the pattern is obvious and those in which the pattern is hidden.
Most of Dee's tables in the Book of Enoch are unpatterned: squares and lozenge shaped arrays with 49 rows and columns lled with text in the 'Enochian' language described by Laycock and Whitby. 29 One of these is illustrated in my Plate IV. The text is inscribed in the tables line by line, left to right, one letter per cell, with no space between words. The eight Soyga tables appearing in the same book are of course patterned, but with a hidden pattern; it is tempting to believe that Dee's favourite table size, 49, was inspired by the size of the Soyga tables, 36, since 49 = 7 7 is the next perfect square after 36 = 6 6.
Similarly there are 36 Soyga tables and, as Kelley informed Dee on 24 March 1582/1583, there are to be 49 Enochian tables. 30
There are many large patterned tables in one of Agrippa's Cabalistic chapters.31 They include: an angel chart of no interest to us, a 'right table of commutation', an 'averse table of commutation', an 'irrational averse table of commutation', a 'table of Ziruph', and a 'rational table of Ziruph'.
The three tables of commutation are examples of what are nowadays known as 'Latin squares', N by N tabular arrays of symbols from an N symbol alphabet | in this case the N = 22 letter Hebrew alphabet arranged in such a way that each letter appears just once in each row and in each column. 32
It is possible that Agrippa received the idea of the 'tabula commutationum recta' from Trithemius. Book 5 of Trithemius's Polygraphia (written in 1508 but printed in 1518) contains a 'recta transpositionis tabula' and a 'tabula transpositionis aversa' of exactly the same form as Agrippa's but based on a hybrid 24 letter alphabet formed by adjoining 'w' to the end of the standard 23 letter Latin alphabet.33 These Latin squares are of a particularly simple type, where each row is a shift of its predecessor, giving the table an overall barber-pole pattern of diagonal stripes.
Agrippa's third table of commutation, the 'tabula aversa dicat irrationalis' is a more complex Latin square. The top row and right-hand vertical margin contain the alphabet in its usual order; the bottom row and the left-hand vertical margin contain the alphabet in reversed order. The interior of the table is partially patterned. Most rows contain blocks of letters in consecutive alphabetical order. Because most of these blocks are shifted by one square from corresponding blocks in neighbouring rows, much of the area of the table has a diagonally striped pattern. But there does not seem to be a simple rule specifying the overall conformation of the table. It seems to be the result of an attempt to construct a Latin square as diagonally striped as possible, consistent with the given normal and reversed alphabets appearing in
the margins.
Agrippa's table of Ziruph, illustrated in my Plate V, is possibly copied from Johann Reuchlin (1455{1522), who in turn owes much to the thirteenth century Kabbalist Abraham Abulaa (1240{1292).34
It consists of 22 rows, each with 11 cells per row. In each cell is a pair of Hebrew letters, placed in such a way that each letter appears exactly once in each row. Each row represents a reciprocal substitution alphabet: the letters in each of the 11 pairs are to be substituted for each other. One of these rows gives the 'Atbash' alphabet according to which the rst and last letters of the Hebrew alphabet (aleph and taw) are interchanged, the second and second from last (beth and shin), and so on.35 Successive rows are obtained by alternately shifting all the left hand elements of the pairs to the pair to the left or all the right hand elements to the pair to the right (with a provision for reversing direction when the end is reached) in a kind of contredanse.36
Such substitution alphabets are used in the branch of the practical Cabala known as temurah (permutation) in connection with the operation of tseruf (combination). The intent is to enlarge the scope of Cabalistic correspondences between words and phrases: two words are related not only if they have the same numerical sum, as in usual gematria, but also if the one is equal to the Atbash-transformed version of the other, and so on. The 'rational table of Ziruph' is possibly Agrippa's invention. The size, shape, and general appearance of this table is the same as the Ziruph table, but the pattern by which the letters shift from row to row is slightly dierent.
Not all large patterned tables appearing in the early modern period are magical, however. For instance, a manuscript of Thomas Harriot (1560{1621) contains letter squares intended to illustrate a combinatorial calculation. 37 These tables, like the Soyga tables, are derived from a key word or phrase, but unlike the Soyga tables, the pattern is completely
obvious. Harriot used the key phrases HENRICVS PRINCEPS FECIT and SILO PRINCEPS FECIT to form squares of 21 and 17 rows respectively. The following articial example based on the key word VERITAS illustrates the pattern. (The key phrase starts at the centre and emanates in concentric lozenges towards the corners.)

Each of these tables is accompanied by a numerical calculation, which turns out to give the number of ways the given key phrase can be spelled out in the square, following a path of vertical and horizontal moves to adjacent cells, starting in the centre and nishing in a corner. (The present VERITAS specimen has 80 such paths; the general formula is 4 times the binomial coecient            , when the key phrase has 2n + 1 letters.)

And finally we have the tables in the Book of Soyga as our sole examples of large patterned tables whose pattern is hidden. None of the other tables, intricate as they are, have so complex an underlying pattern as that given by the N + f(W) formula used in the Book of Soyga. It is no wonder that Dee found them perplexing.


1 British Library, MS Sloane 3188, fol. 9r.
2 All of these examples: Christopher Whitby, John Dee's Actions With Spirits, 2 vols (New York: Garland, 1988), i, pp. 146{147.
3 Scrying, a cooperative magical operation during which privileged visual and aural information  in this case from angels |is conveyed to the participants, was much used by Dee. Three diering views of what 'really went on' are presented in Meric Casaubon, A True and Faithful Relation (London, 1659) (which I have not seen), in Whitby, Actions with Spirits, i, and in D. E. Harkness, 'Shows in the Showstone: A Theater of Alchemy and Apocalypse in the Angel Conversations of
John Dee (1527{1608/9)', Renaissance Quarterly, 49 (1996), 707{737.
4 John Dee, Spiritual Diaries, Sloane 3188, fol. 9r, transcribed in Whitby, Actions with Spirits, ii, pp. 17{18 and translated in Whitby, Actions with Spirits, i, pp. 211{ 212.
5 A marginal note on Sloane 3188, fol. 9r, transcribed in Whitby, Actions with Spirits, ii, p. 18, seems to say Kelley and Dee had met for the rst time two days previous to this: 'Note: he had two dayes before made the like demaunde and request unto me: but he went away unsatisfied. For, his coming was to entrap me, if I had any dealing with Wicked spirits as he confessed often tymes after...' See my note 16 for evidence of Kelley's continued ignorance of basic facts about the Book of Soyga a month later.
6 Whitby, Actions with Spirits, i, pp. 146{147; Deborah Elizabeth Harkness, 'The Scientic Reformation: John Dee and the Restitution of Nature' (unpublished Ph.D. dissertation, University of California at Davis, 1994), pp. 317{318, 415. Both guess the Book of Soyga might well have inuenced Dee or Kelley. Harkness, p. 415, suggests that the Book of Soyga's Adamic association | in particular its use of an Adamic language, discussed by Uriel and Il, in Sloane 3188, fols. 9r and 89v| would have especially appealed to Dee. Whitby, Actions with Spirits, i, p. 147, cites I. R. F. Calder as conjecturing that the Book of Soyga is the Voynich manuscript (Yale University, Beinecke Rare Book & Manuscript Library, MS 408), the notorious cipher manuscript described by J. M. Manly, 'Roger Bacon and the Voynich MS', Speculum, 6 (1931), 345{391; if true, this would be a case of solving one mystery by replacing it with a greater. I see no connection between the two books, other than their probable ownership by Dee. The Book of Enoch, also called Liber Logaeth and Liber mysteriorum sextus et sanctus, British Library, MS Sloane 3189, was in effect Dee's lab notebook, written concurrently with the Spiritual Diaries, Sloane 3188. Whitby, Actions with Spirits, i, p. 143, gives a description of its contents.
7 Deborah Harkness, personal communication, 1996, and 'The Nexus of Angelology, Eschatology, and Natural Philosophy in John Dee's Angel Conversations and Library', in this volume.
8 Oxford, Bodleian Library, MS Bodley 908; British Library, MS Sloane 8.
9 This description is based on examination of microlm copies, not on the manuscripts themselves.
10 'Zadzaczadlin': Bodley 908, fol. 69v and Sloane 8, fol. 70v.
11 Robert Turner, Henry Cornelius Agrippa His Fourth Book of Occult Philosophy (London, 1655; reprinted London: Askin, 1978).
12 Bodley 908, fol. 51v.
13 Both in Bodley 908, fol. 42r.
14 As described by, say, Richard Kieckhefer, Magic in the Middle Ages (Cambridge: Cambridge University Press, 1989).
15 'Geber': Bodley 908, fols. 116v and 126r; 'Steganographia': Bodley 908 fol. 123v.
16 Bodley 908, fol. 4r; Sloane 8, fol. 6r. But this directly contradicts what the spirit 'Il' said during a scrying session with Edward Kelley and John Dee on Thursday 18 April 1583, as recorded in Dee's Spiritual Diaries, Sloane 3188, fol. 89v, transcribed in Whitby, Actions with Spirits, ii, p. 332: 'Soyga signieth not Agyos. Soyga alca miketh.' (Dee's or Il's emphasis.) One might take this as evidence of Kelley's unfamiliarity with the Book of Soyga at this early stage in his residence in Dee's household.
17 Wayne Shumaker, The Occult Sciences in the Renaissance: A Study in Intellectual Patterns (Berkeley, California: University of California Press, 1972).
18 D. P. Walker, Spiritual and Demonic Magic from Ficino to Campanella, (London: Warburg Institute, 1958) and Frances A. Yates, Giordano Bruno and the Hermetic Tradition (London: Routledge and Kegan Paul, 1964). It is certain that Hermeticism and Cabalism were important formative inuences on early modern magic, even if Yates's claims about their inuence on early modern science are rejected.
19 Karen de Leon-Jones, personal communication, 1998. I have not found a single table or chart or discussion of such anywhere in the works I have seen of the two great modern historians of the Kabbalah, Gershom Scholem and Moshe Idel.
20 In Bodley 908, at fols. 180 - 197; in Sloane 8, at fols. 102 - 138; see my Table I.
21 Book of Enoch, Sloane 3189, in four openings of the book, between fols. 58 - 65, as shown in my Table I.
22 In Bodley 908, fols. 167r{168v; in Sloane 8, fols. 138v{140v. The Bodley 908
version seems to contain many mistakes.
23 Charles Babbage, 'Notice respecting some Errors common to many Tables of Logarithms', Memoirs of the Astronomical Society, 3 (1827), 65{67, which I have only seen reprinted in Charles Babbage, The works of Charles Babbage, edited by Martin Campbell-Kelly, 11 vols (London: W. Pickering, 1987), ii, pp. 67 - 71. Summarized in Dr Dionysius Lardner, 'Babbage's Calculating Engine', Edinburgh Review, July 1834, no. 120; which I have only seen as reprinted in Charles Babbage and his Calculating Engines, Selected Writings by Charles Babbage and Others, edited by Philip Morrison and Emily Morrison (New York: Dover, 1961), pp. 163 - 224; the discussion of errors in logarithm tables appears on pp. 177 - 183.
24 Yates, Giordano Bruno, p. 149 and note. The tables are in Agrippa, De Occulta Philosophia, iii, 25, not iii, 24.
25 That is, Yates did not care to pay attention to the differences between the tables, possibly because she did not know how to. It is also possible that for Yates, magic tables,  unlike texts or images,  are not subject to the processes of copying, emulation, improvement, and confusion; that is, they are neither vehicles for ideas nor potential sources of evidence in intellectual or cultural history.
26 Raphael Patai, The Jewish Alchemists (Princeton: Princeton University Press, 1994), pp. 277{288.
27 Menso Folkerts, 'Zur Fruhgeschichte der magischen Quadrate in Westeuropa', Sudhos Archiv, 65 (1981), 313{338 gives a detailed survey of the genre. Vladimr Karpenko, 'Between Magic and Science: Numerical Magic Squares', Ambix, 40 (1993), 121{128, surveys alchemical magic squares; in this connection, also see Patai, The Jewish Alchemists, chapter 26.
28 Heinrich Cornelius Agrippa of Nettesheim, De Occulta Philosophia libri tres (Cologne, 1533). I rely on the edition of V. Perrone Compagni (Leiden: Brill, 1992). In a supercilious scholium, Shumaker, The Occult Sciences in the Renaissance, p.139, takes Agrippa to task for a mistake in one of his magic squares. However the mistake is clearly a typographic error present only in the Arabic numeral form of the square, and only in the particular edition Shumaker looked at. (Shumaker, p. 158, note 70, seems to say he relies on 'Henricus Agrippa ab Nettesheym, Opera (Lugduni,
c. 1650?)', which he understands to be printed in London instead of Lyons!) For a discussion of Agrippa's magic squares, see K. A. Nowotny, 'The construction of certain seals and characters in the work of Agrippa of Nettesheim', Journal of the Warburg and Courtauld Institutes, 11 (1949), 46{57 and I. R. F. Calder, 'A note on magic squares in the philosophy of Agrippa of Nettesheim', Journal of the Warburg and Courtauld Institutes, 11 (1949), 196{199.
29 Donald C. Laycock, The Complete Enochian Dictionary: A Dictionary of the Angelic Language as Revealed to Dr. John Dee and Edward Kelley, revised edition (York Beach, Maine: Samuel Weiser, 1994); Whitby, Actions with Spirits, i, pp. 144 - 146.
30 Sloane 3189, fol. 62v, transcribed in Whitby, Actions with Spirits, ii, p. 227. Apparently one of the tables is not to be written, leaving only 48 to be put in the Book of Enoch. It is tempting to compare this with the Book of Soyga's T36
'Magistri', which has a missing row.
31 Agrippa, De Occulta Philosophia, iii, 25, sigs. yiir{yiiiir. These tables are surveyed in a modern reissue of the J[ohn] F[rench] translation (London, 1651) of Agrippa: Three Books of Occult Philosophy, edited by Donald Tyson (St. Paul, Minnesota: Llewellyn Publications, 1993), appendix VII, pp. 762 - 767.
32 J. Denes and A. D. Keedwell, Latin Squares and their Applications (New York: Academic Press, 1974).
33 Johannes Trithemius, Polygraphiae libri sex (Oppenheim, 1518), v, sigs. oijr and oijv. It is barely possible that Trithemius received the idea of such tables from Agrippa, possibly when they met in the winter of 1509/1510. In 1510 Agrippa sent Trithemius a draft of his De occulta philosophia which, according to Compagni (p. 58), lacked the chapter containing the Ziruph tables and tables of commutation.
Trithemius's use of the tabula recta is purely cryptographic, and most printed works on cryptography ever since include such diagrams, often under the name of 'Vigenere table'. Since Agrippa's text does not discuss his tables of commutation it seems more likely, absent any further direct evidence, that Agrippa copied from Trithemius.
34 Johann Reuchlin, De Arte Cabalistica (Hagenau, 1517), Book III, sig. Nvir; I rely on the parallel-text translation of M. Goodman and S. Goodman, of 1983, reissued with introduction by Moshe Idel (Lincoln, Nebraska: University of Nebraska Press, 1993). For gematria and the Reuchlin-Abulaa connection: Gershom Scholem, 'Gematria' in Encyclopaedia Judaica (Jerusalem: Macmillan, 1971) and Gershom Scholem, Major Trends in Jewish Mysticism, second edition, reissued (New York: Schocken, 1995), p. 127.
35 A somewhat similar table of reciprocal substitution alphabets occurs in Giovanni-Battista della Porta, De Occultis Literarum Notis (Naples, 1563) ii, 16; I rely on a facsimile (Zaragoza: Catedra de Criptografa del Centro Politecnico Superior de la Universidad de Zaragoza, 1996) of the 1593 Montbeliard edition. This is a cryptographic work, and Porta tables are almost as much a xture in cryptographic literature as Vigenere tables. Porta's table is based on a 22 letter Latin alphabet with the letter K omitted. In Porta's table but not in the Ziruph table letters from the first half of the alphabet are paired only with letters from the last half.
36 There appears to be one deviation from this pattern. In the seventeenth row of both Reuchlin's and Agrippa's tables the letters sade and resh are paired, as are tet and taw. The rule used to produce the rest of the table would pair taw with sade and pair tet with resh.
37 British Library, Additional MS 6782, fols. 27, 28, with associated calculation on fol. 57. These are briefly described in John W. Shirley, Thomas Harriot: a Biography (Oxford: Oxford University Press, 1983), pp. 419{420, who apparently did not understand the calculation on fol. 57. I intend to address these Harriot tables in a subsequent paper.

John Dee and the Magic Tables in the Book of Soyga

John Dee Section

Main Library