Polyalphabetic ciphers were discussed by classic authors such as Alberti, Trithemius, Bellaso, Porta, and Vigenère, but it may not have been much known until the mid-sixteenth century. The present article presents some polyalphabetic ciphers before 1600.
It should be noted that the following includes not only typical polyalphabetic ciphers that regularly switch alphabets but also schemes that simply provides for more than one alphabet to be used concurrently. (We do not deal with machine ciphers, some of which are very complex polyalphabetic ciphers.)
Leon Battista Alberti's treatise, De componendis cifris or De cifris, written in 1466 or 1467, describes a cipher disk and advises changing the position after enciphering "three or four words".
The treatise, translated by Cosimo Bartoli into Italian, was printed as one of the fifteen included in Opuscoli Morali (1568) (Google, pp.200), with the running title "La Cifra di Leonbattista Alberti" (Charles J. Mendelsohn (1940), "Bibliographical Note on the 'De Cifris' of Leone Battista Alberti", Isis, vol.32, no.1, pp.48-51 (JSTOR); Christie's), but its significance was not recognized for a long time.
Johannes Trithemius' Polygraphia, posthumously published in 1518, described a tabula recta (commonly known as the Vigenere table today), in which the substitution alphabet (corresponding to a row in the table) is switched progressively in a fixed sequence.
Giovan Battista Bellaso's first book, La Cifra Del Sig. Giovan Battista (1553), proposed a scheme in which substitution tables are switched according to a keyword, where each substitution table is for swapping letters in pairs (see another article in Japanese).
Giovanni Battista Porta's De furtivis literarum notis (1563) described various schemes including a cipher disk like Alberti's, switching substitution tables according to a keyword as in Bellaso's pairwise-swapping. A mixed alphabet is also described. (See another article.)
Blaise de Vigenère's Traité de chiffres (1586) proposed the autokey cipher, though the Vigenere cipher as commonly known today refers to a simpler version, combining a square table (in the regular alphabet) with switching according to a keyword (see another article (Academia.edu)).
It is known that cipher collections from the sixteenth century contain many polyalphabetic tables (Wikipedia in Italian), but materials before 1560 are scanty.
An early specimen of a polyalphabetic cipher was recently discovered.
The instructions for use attached to a seemingly simple substitution table in the State Archives of Florence describes switching of substitution tables letter by letter, as described in Marco Vito, "A Florentine 'polyalphabetic' cipher in the 15th century" (HistoCrypt 2025).
The cipher is in Archivio di Stato di Firenze (ASFI), Miscellanea Repubblicana 11, fasc. 292, I, and is endorsed "Cifera di Livorno Conposta per la buona [aromoria] di Piero Capp[o]ni". Piero Capponi (Wikipedia) worked for Lorenzo the Magnificent. Vito considers the cipher is related to one of Capponi's two missions to Livorno (1484-1485, 1486-1487) and gives 1487 as the most likely date.
The substitution alphabet is like the following (my transcription uses approximate characters; the left half is missing because of torn paper).
|
... f g h i k l m n o p q r s t u ... q r h p s l i b 7 t 4 y e z 8 |
Its usage is illustrated with an example of enciphering "papa." The first letter "p" is enciphered as "t" according to the above table. After enciphering one letter, the substitution alphabet is shifted so that the cipher symbol ("4") next to the one just used ("t") aligns with "a" in the top row. Thus,
|
....l m n o p q r s t u a b c d e ... q r h p s l i b 7 t 4 y e z 8 |
From this, the second letter "a" is enciphered as "4". Then, the next symbol ("y") to the one just used ("4") is aligned with "a" in the top row:
|
....k l m n o p q r s t u a b c d ... q r h p s l i b 7 t 4 y e z 8 |
This enciphers the third letter "p" into "l." Again, the next symbol ("i") to the one just used ("l") is aligned with "a" in the top row.
|
....p q r s t u a b c d e f g h i ... q r h p s l i b 7 t 4 y e z 8 |
This enciphers the last letter "a" as "i." Thus, the plaintext "papa" is enciphered into "t4li".
This scheme, determining the relative position (shift) for enciphering the next letter according to the current symbol may be considered as a variant of the autokey cipher invented by Vigenere a century later. It is also remarkable that the mixed alphabet is already used in this cipher.
Last but not least, this case warns us that when we search the archives only for tables with an apparent polyalphabetic nature, we may miss polyalphabetic encryption described in text attached to a seemingly monoalphabetic substitution table.
A scheme similar to the Gronsfeld cipher, a variant of the Vigenere cipher named after a seventeenth-century German count, was employed for Silvio Calandra, ambassador of the Duke of Mantua (see another article).
A letter from Mary, Queen of Scots, to Guzman de Silva, Spanish ambassador in England, dated 16 September 1568, regularly switches among three substitution tables after enciphering every letter (see another article).
A table attached to this letter also seems to be a polyalphabetic cipher table.
My earlier description of this table needs correction. It is noted that every two rows forms a complete alphabet, and differences in style in "d" and "l" seem to be irrelevant. Every even row is a permutation of each other (similarly for odd rows). All in all, this table specifies six substitution alphabets, each consisting of two rows, similar to those described by Bellaso and Porta.
Aloys Meister (1906), Die geheimschrift im dienste der Päpstlichen kurie von ihren anfängen bis zum ende des XVI. Jahrhunderts (Internet Archive) notes a few cases of possible use of polyalphabetic ciphers, based on papers left by Matteo Argenti (born in 1561 (Meister p.59)). Matteo Argenti was a cipher secretary in the curia from 1591 to 1605 after his uncle, Gianbattista Argenti, and, after his dismissal in 1605, compiled papers of his uncle and his own (Meister (1906) p.51, 56, 60).
In 1581, Gianbattista Argenti was challenged by his master, Giacomo Boncompagni, Duke of Sora (Wikipedia), to solve a cipher in a polyalphabetic system and he boasts that he readily solved it (Meister (1906), p.294-295, 58, 86). (Gregory XIII could not appoint Gianbattista as cipher secretary and found a position for him under his "nephew". It was only in 1585 that Gianbattista became cipher secretary (Meister (1906), p.55, 51).)
Solving this would have been difficult even with knowledge from later centuries. The only possibility I can think of is that he knew the system per se and he only had to identify the exact substitution tables and the key, given a hint of a probable word.
A cipher dated 28 September 1589 for Cardinal Caetani [Caietan, Cajetan], the papal legate sent to France, (Wikipedia) includes four substitution tables numbered 1-5 (Meister (1906), p.420). This may have been broken by a contemporary codebreaker, Chorrin (see another article) (Marshal Chorin's visit to London is recorded in "The Elizabethan Court Day by Day" (Folgerpedia) (1595).). As with the case with Panicarola below, the switching may have occurred only once in a few lines.
A polyalphabetic cipher, dated 3 October 1589, was prepared by Gianbattista Argenti for Bishop Panicarola (Francesco Panigarola, sent to France as assistant to the papal legate Caetani (Wikipedia)). But it was considered too complicated and a simpler cipher below was actually used. The instruction provides for switching the alphabet according to a key sentence after enciphering a word. The cipher alphabet consists of letters and Arabic figures (0-9, c-q). (Meister (1906), p.303, 59)
Another polyalphabetic cipher from the same date for Panicarola is based on pairwise swapping.
According to the description of Meister (1906), p.82, in this simplified version, the switching of tables may have occurred only once in a few lines, which was the case when Matteo Argenti practiced such switching with the Cardinal of Florence (Alessandro de' Medici, later Leo XI) (EBSCO, Wikipedia), papal legate in France in the time of Clement VIII (Meister (1906), p.82-83).
Matteo Argenti observed a "key cipher" ("cifra per chiave": a polyalphabetic cipher with table switching according to a keyword/key phrase) was often used in personal communication between princes of the Church (Matteo mentions Cardinal Aldobrandino corresponding with Francesco Albobrandino, his lieutenant in Spain. The latter's mission in Spain is mentioned in Davila (1678), The history of the civil wars of France (EEBO) p.673) and he used it himself with some princes, though usually only a few lines were enciphered with this scheme because of its complexity. In two instances, Matteo Argenti inserted a piece in the polyalphabetic cipher in a ciphertext in another cipher (Meister (1906), p.83-84).
Instead of using the "Vigenere square", the keyword was written vertically, and the whole alphabet was completed to the right, continuing from each letter of the keyword. (This is similar to the description in John Falconer (1685), Cryptomenysis Patefacta (Internet Archive) p.19)
A Vigenere table of one Father Marino Furlani is also recorded. The paper is filed with documents attributed to 1585-1590 and 1555. I believe this is after the publication of Vigenere (1586).
A Vigenere-like table is in the Milanese Archives (Carteggio Sforzesco 1598, p.151; DECODE R5602).
It has 22 rows, each of which has a substitution alphabet in reverse alphabetical order. Somehow, two rows share the same starting position of the reverse alphabet, and a compact table in which two rows sharing the same position are represented by one row is attached below. Although there is nothing that shows the date, I think this belongs to the late sixteenth century.
Héder, M ; Megyesi, B. The DECODE Database of Historical Ciphers and Keys: Version 2. In: Dahlke, C; Megyesi, B (eds.) Proceedings of the 5th International Conference on Historical Cryptology HistoCrypt 2022. Linkoping, Sweden : LiU E-Press (2022) pp. 111-114. , 4 p. [pdf]
Megyesi Beáta, Esslinger Bernhard, Fornés Alicia, Kopal Nils, Láng Benedek, Lasry George, Leeuw Karl de, Pettersson Eva, Wacker Arno, Waldispühl Michelle. Decryption of historical manuscripts: the DECRYPT project. CRYPTOLOGIA 44 : 6 pp. 545-559. , 15 p. (2020) [link]
Megyesi, B., Blomqvist, N., and Pettersson, E. (2019) The DECODE Database: Collection of Historical Ciphers and Keys. In Proceedings of the 2nd International Conference on Historical Cryptology. HistoCrypt 2019, June 23-25, 2019, Mons, Belgium. NEALT Proceedings Series 37, Linköping Electronic Press. [pdf]
A letter from Edward Wilton to the Earl of Essex (1595) used fourteen different ciphers (Daybell p.157, citing CP35/112: 30/10-9/11/1595).
Henry Neville and Robert Cecil used a cipher comprising more than one substitution tables, which are occasionally switched in one letter (see another article). The switching was indicated by a letter indicating the table.
Venetians adopted a cipher with superencryption (the caselle cipher) in 1578 and used it for two decades. Superencryption resulted in polyalphabeticity. (See Paolo Bonavoglia's works I quoted in my blog.)
From about the late sixteenth century, polyalphabetic ciphers were no longer a rarity, though they were never widely used in official correspondence with a possible exception of the Confederates during the US Civil War and possibly many (not very successful) trials during the American Revolutionary War.
In 1636, Thomas Roe devised a polyalphabetic cipher for Elizabeth, Queen of Bohemia (in which a digit indicates the position of the substitution table as in the Gronsfeld cipher), but the cipher was soon abandoned (see another article).
Lord Feilding used a cipher in which three tables are switched every letter (see another article).
A letter from Thurloe to Henry Cromwell (1656) employs a simple but unique polyalphabetic cipher, in which a letter is represented by a pair of figures, the difference of which indicates a letter (see another article).
A Vigenere table as well as a sheet experimenting with it are included in papers related to Edward Nicholas (another article).
Add MS 11252 contains a polyalphabetic table in which columns with mixed alphabets can be switched according to a keyword (see another article).
A Danish diplomatic cipher probably from the later seventeenth century had five substitution tables (see another article).
A cipher disk more elaborate than Alberti's is preserved among papers of Brûlart de Léon (see another article in Japanese or another article at Academia.edu).
Instead of switching a substitution table, some French ciphers switched codes (list of nomenclature) (see another article (1696-1697) and another article (1782)).
It is notorious that James Lovell devised polyalphabetic ciphers (see another article) during the American Revolution for Franklin, John Jay, John Adams.
Thomas Jefferson used a polyalphabetic cipher with his secretary, William Short (see another article).
In one letter between Edmund Randolph and James Madison, a passage particularly sensitive was enciphered with a polyalphabetic cipher in ciphertext encoded with the official code (see another article).
A Vigenere table sent by Rev. William Gordon to George Washington is preserved (see another article, yet another).
Marie-Antoinette used a polyalphabetic cipher in her correspondence with Axel von Fersen (see another article) as well as with the court in Vienna (see another article).
A polyalphabetic cipher is used in a document which is purportedly Napoleon's proclamation to the army (see another article).
The Confederates used polyalphabetic ciphers extensively during the Civil War (see another article).
The Japanese government used cipher disks at an early stage of westernization, but they were only used to change the monoalphabetic substitution table for a particular message (see another article in Japanese).
Lewis Caroll used the Vigenere cipher with his "child-friend". He devised other systems. (See another article in Japanese.)
A challenge cryptogram (Kulp's Cryptogram) sent to Edgar Allan Poe turned out to be in the Vigenere cipher in 1975 (see another article).
The Vigenere cipher and its variant are popular in fictions, e.g., Jules Vernes, La Jangada, Johannes Mario Simmel, The Caesar Code.