Charles I and Nicholas to Boswell (1643)

Summary. Two ciphered letters of the same day, transcribed provisionally by S. Tomokiyo in 2021 and listed on his page of unsolved ciphers. Robert Pitt published the alphabet on 14 September 2026: a 24-letter row repeated four times over the figures 20–115, supplementary homophones in alphabetical blocks at 116–159, nulls below 20, and four word codes. That solution is not claimed here. What this page adds is verification and reading. A permutation control puts Pitt’s row 9.6 standard deviations above 20,000 random rows on the same periodic structure, with none as good. The four graphic signs that Pitt’s notes left unexplained turn out to be word-signs introduced inline — the clerk wrote the sign beside the word the first time he spelled it (good, Cousin, Master, us) and used it alone thereafter — which closes the three framed passages as “our good Cousin your Master”. And the King’s letter is not to Boswell at all: it is his answer and re-credentials to the Duke of Courland’s envoy, stuck at The Hague, sent through Boswell in Boswell’s own cipher, as Nicholas’s covering letter says in so many words. Simpson’s 1893 print of the King’s parallel letter to the Duke, same date, confirms the frame and gives 291 = Duke. Status: read in substance. A dozen single-occurrence word codes and some transcription slips wait on the folios, which are not online.

01 The documents

TNA Discovery has the items at folio level: SP 84/157/96, f. 217, “Charles I to Boswell — in cipher” and SP 84/157/97, f. 219, “Nicholas to Boswell — in cipher”, both 2 November 1643. No images exist online. Tomokiyo’s transcription (blog, 17 September 2021), with asterisks where he was unsure, is the only text; Pitt’s inputs are line for line the same. Sir William Boswell was the King’s resident at The Hague; Sir Edward Nicholas was Secretary of State at Oxford.

Across both letters: 455 numeric groups, 171 distinct, 69 seen once. The bands separate cleanly by usage:

band	occurrences	distinct	seen once	what it is
0–19	45	17	4	nulls (no reading anywhere needs one)
20–115	276	80	17	the alphabet: 24 letters × 4 homophones (96 predicted, 80 seen; q z unused)
116–159	61	31	14	supplementary homophones for a d e i m n o r s t u y, in alphabetical blocks
160–399	25	22	19	word codes, plus a null block 361–386
400–999	48	21	15	word codes continued: 800 to ×14, 588 of ×7, 873 you ×6

02 The key (Pitt’s alphabet)

plain   a  b  c  d  e  f  g  h  i  k  l  m  n  o  p  q  r  s  t  u  w  x  y  z
1st     20 32 21 33 22 34 23 35 24 36 25 37 26 38 27 39 28 40 29 41 30 42 31 43
2nd     44 56 45 57 46 58 47 59 48 60 49 61 50 62 51 63 52 64 53 65 54 66 55 67
3rd     68 80 69 81 70 82 71 83 72 84 73 85 74 86 75 87 76 88 77 89 78 90 79 91
4th     92 104 93 105 94 106 95 107 96 108 97 109 98 110 99 111 100 112 101 113 102 114 103 115

That is letter = row[(n − 20) mod 24] with the row a c e g i l n p r t w y b d f h k m o q s u x z — the odd positions of the 24-letter alphabet (no j, no v) and then the even ones. Each letter’s four values are 24 apart. It is the same type as the regular Stuart keys Tomokiyo describes for Scudamore (1636), Cave (1638) and Ormonde–Clanricarde (1643), and as the Maltravers–Ormonde key. The supplementary block 116–159 is five alphabetical runs (a d e · m n o r s t y, a d e · n o · t u, …), so even its unseen values are constrained; 142 = s is added here from Nicholas’s opening “[His Majesty] comands”.

03 Verification

The 21 maximal runs of four or more alphabet groups (106 letters) were scored with a 4-gram model built from 2.7 million characters of seventeenth- to nineteenth-century English. The null hypothesis permutes the 24 letter identities over the same periodic structure.

	score
Pitt’s row	−217.5
20,000 random rows: mean / sd / best	−333.7 / 12.1 / −271.1
z; rows scoring at least as well	9.6; 0 of 20,000

A free period scan (annealing a 24-letter row for periods 20–32) is not decisive and is recorded as such: at 106 letters the annealer over-fits, and periods 27–32 reach higher scores than the true key. The argument for period 24 is structural — the odd/even construction, the 24-apart homophones, the alphabetical supplementary blocks — and the readings against the clear text (“A C · · knowledge”, “[835] D E R stand”, “Y O U R E [C] R E D E N T I A L S” beside Simpson’s “Our Re-credentials”) are the confirmation.

04 The four signs

Tomokiyo renders four graphic signs: △, 中, □ and +. Each occurs once inside a spelled word and afterwards on its own, and the stand-alone occurrences come in exactly the combination the embedded ones predict.

sign	embedded occurrence	stand-alone	value
△	`47 86 △ 110 33` = GO△OD	“back to OUR △ 中 873 □”	good
中	`93 122 中 41 112` = CO中US-in	“Our 中 873 □ S affeccons”, “our said 中 873 □”, “our △ 中 873 □”	Cousin
□	`61 116 88 □ 77 70 28` = MAS□TER	the same three places	Master
+	`158 + 150` = U+S (“inuiting us”)	“acceptable to +”, “preiudiced +”, “to + to the port of”, “addres by 213 to +”	us

So △ 中 873 □ is “good Cousin your Master”. The clerk wrote the sign beside the word the first time he spelled it — compare “for whom 270 thus hereafter” in the Ormonde letter, where a code is introduced inline — and the transcription placed the sign in the middle of the word. This is the explanation Pitt’s notes said was missing, and with it 873 = you / your fits all six occurrences.

05 Reading

Codes in brackets, proposed transcription corrections in braces, spelling as the cipher gives it.

Charles I to the Duke of Courland’s envoy, Oxford, 2 November 1643 (f. 217):

Sir, Wee have received your ciuill {&} welcom letter of the {6th} of 8ber 1643. But never heard of any [514] letter of Our good Cousin the Duke of Curland’s, your Master, saue those inuiting us to send to the Funerals of the Late [303] Dukes his Father & Unkle [376] deceased of happy memory; to [which] we made an answer, but know not whether the same arriued [289] safe. The Difficulty of passage {we} feare {hath} preiudiced us. Our Cousin your Master’s affeccons & offers are soe reall & acceptable to us, specially in this coniuncture of our Affaires, as we acknowledge [our] [selves] there[by] highly obliged. We [pray] you therefore to present our princely thanks [177] on this behalf, to [assure] our said Cousin your Master, & to lett him understand that we shall most [wel]comly receiue his [assistance] with in [755] [188] [musket]s, match & [powder], in {what} [pro]por[tion] it shall please him to transmit the same to us, to the [port] of Weymouth, Dartmouth, Exeter, Falmouth or [228]. Not doubting but God will shortly enable Us to recompence soe free & soe [gener]ous a c[591] & es[126]. Doe assure your self also of our fauour particular, & doe take your addres by [213] to us in very good part, as also that you haue taken care to avoyde the dangerousnes of the passage hither. And soe wishing you a safe returne back to our good Cousin your Master, to whom w[e] herewith [send] your re-credentials, Wee commit you to the Protection of the Almighty. Given at our court at Oxford the second day of Nouember in the Nineteenth yeare of Our Reigne 1643.

Nicholas to Boswell, 2 November 1643 (f. 219):

Sir, [His Majesty] comands [me] to signify his pleasure to you, that you use all possible diligence & industry by your meanes of the [170] & otherwise [secret]ly to hinder the coming ouer of the [181]s from the [726], for it is here apprehended that they [will] doe more hurt then good, for that they [are] understood here not to affect Monarchy. His Majesty’s two letters herewith sent with flying Seales are put into your character, because most safe & ready; you are to decifer & interpret them. Monsieur F I R X S S appeares not here. …

Word codes with a reading: 800 to, 588 of, 873 you/your, 291 Duke (from Simpson’s printed cipher line 291 588 C U R L A N D on the letter to the Duke), 516 letter, 835 un-, 749 selves, 750 send, 205 assure, 223 by, 629 pray, 640 port, 636 pro-, 406 gener-, 484 His Majesty, 746 secret-. Sorted, the readable codes run alphabetically (a 170–228 … d 291 … o 588–591 … s 726–755, t 800, u 835, y 873), so the nomenclator was probably an alphabetical list; a second letter in the key would test that.

06 What the letters are

The King’s letter is not to Boswell. It answers the envoy’s letter of 6 October, calls the Duke “your Master” three times, sends “your re-credentials”, and wishes him a safe return to his master. Simpson’s Civil War Papers (Scottish History Society Miscellany I, 1893) prints the King’s letter to Duke James of Courland of the same day, calling itself “these Our Re-credentials”, and the envoy’s later memorial: the passage was unsafe and “the ministers of their Majesties at the Hague” dissuaded him from crossing and took his letters. Nicholas tells Boswell that the King’s two letters are “put into your character” and that Boswell is to “decifer & interpret them”. Hence the catalogue’s “Charles I to Boswell”.
“Monsieur F I R X S S appeares not here.” Simpson names Major Georg von Firks as the Duke’s agent. A man expected at Oxford who has not come is the envoy’s own account of himself.
The arms. Muskets, match and powder to Weymouth, Dartmouth, Exeter, Falmouth or a fifth port: the Duke’s memorial says he was told their Majesties needed munitions rather than men, and Nicholas wrote to Roe in December 1643 that “arms and powder are arrived at Dartmouth for his Majesty” (CSPD).
The funerals of Duke Friedrich (d. 1642) and his brother Wilhelm (d. 1640), held together in February 1643: “the Late Dukes his Father & Unkle”.
Nicholas’s instruction to hinder “the coming ouer of the [181]s from the [726]”, people “understood here not to affect Monarchy”, fits the States General’s embassy of mediation resolved on in late 1643 and received at Oxford early in 1644, which the King did not want: 181+S = Ambassadors, 726 = States. That is a historical reading, not a cryptographic one.

07 Method, credit and files

Pitt’s repository robertpitt/charles-i-boswell-cipher (MIT; commit 609ebf4, 14 September 2026) had the alphabet, the supplementary homophones, most nulls and four word codes before this session. It should be cited as the solution of the alphabet.
Band statistics and the decoder (boswell/key.py); 4-gram model (lm4.py); permutation control (permtest.py); period scan (periodscan.py).
Simpson 1893 and CSPD 1641–43 from Internet Archive OCR; TNA references from the Discovery API.

Checked: Pitt’s inputs against the blog text; the alphabet against 20,000 permuted rows; every reading against the tokens; Simpson’s letter and memorial; the TNA and CSPD references. Not checked: the manuscripts (no images online); the Latvian archive descriptions Pitt cites. The identity of the envoy as Firks, the alphabetical-order hypothesis and the proposed transcription slips are inferences to verify before repeating.

08 Sources

S. Tomokiyo, “Charles I–Boswell Cipher (1643)”, blog post of 17 September 2021, and Unsolved Historical Ciphers (cryptiana).
R. Pitt, charles-i-boswell-cipher (GitHub, September 2026).
TNA Discovery, SP 84/157/96–97.
H. F. Morland Simpson (ed.), “Civil War Papers”, Miscellany of the Scottish History Society I (1893), pp. 143–170 (Internet Archive miscellanyofscot01scot).
Calendar of State Papers Domestic 1641–43 (1887), Nicholas to Roe, December 1643.

Analysis scripts and the ciphertext are in the repository (boswell/).