Habsburg Codes and Ciphers

The present article presents some German ciphers related to the Habsburgs found in publications.

Further sources will be found in Leopold Auer, "Die Verwendung von Chiffren in der diplomatischen Korrespondenz des Kaiserhofes im 17. und 18. Jahrhundert" in Geheime Post (2015) as well as other articles cited below. Moreover, Jakub Mírka's series of articles in Crypto-World (in issues 11-12/2012, 1-2/2013, 3-4/2013, 11-12/2013) beginning with "Raně novověká šifrovaná korespondence ve fondech šlechtických rodinných archivů Státního oblastního archivu v Plzni, část I." provide many images (and probably much more in the text) about cipher materials in the archives in Pilsen (now in the Czech Republic). (Bohemia was part of the Habsburg estates since 1526.)

Another article of mine: Some German Ciphers: 1540-1815 describes some German ciphers not described herein.

Before Emperor Chares V

Vowel Substitution during the Middle Ages

In the Middle Ages, rudimentary secret writing was used, in which typically only vowels were replaced with (i) other letters or symbols, (ii) dots, (iii) crosses (X, XX, XXX, ...), (iv) Roman numerals, or (iv) (in the 15th century) figures (Meister (1902), Die Anfange der Modernen Diplomatischen Geheimschrift, p.5-11, p.17-18). (Archival sources of other forms of secret writing are given in Bischoff, Bernhard (1954), "Übersicht über die nichtdilpomatischen Geheimschriften des Mittelalters", MIOG, vol.62, summarized in Philip Neal, "A list of mediaeval cipher manuscripts" at Voynich Sources.)

Replacing vowels (a, e, i, o, u) with other letters (b, f, k, p, x) (Meister (1902) p.6) might cause a trouble in deciphering because of bivalance of the substituted letters but such a potential confusion can be avoided by reciprocal substitution. An example of such reciprocal substitution (replacing a, e, i, o, u with m, k, d, t, h and replacing m, k, d, t, h with a, e, i, o, u) is found in 1433.

Rudolph IV (1339-1365)

Use of a substitution cipher alphabet is recorded in several documents (Meister (1902) p.11; Bischoff (1954) no.62) related to Rudolph IV of Austria (1339-1365), including his epitaph (Wikipedia). This substantially corresponds to what is known as the "Alphabetum Kaldeorum" recorded in 1428 (Wikipedia). But this isolated example, as well as similar ciphers such as one of Hildegard von Bingen (1098-1179) (Meister (1902) p.10), is not necessarily the origin of later development of ciphers.

Friedrich III (1415-1493)

Frederick III, Holy Roman Emperor from 1452 (the first emperor from the House of Habsburg), is known for a reciprocal substitution cipher in which "a" and "e" were interchanged and similarly for the letters in pairs "i" and "o", "b" and "c", "d" and "f", etc. (Bischoff (1954), no.12). The Emperor claimed he devised this himself ("hab ich selbs gedacht").


The Emperor's boast might be justified in view of mere vowel substitution systems used at the time (replacing a, e, i, o, u with 1, 2, 3, 4, 5 (Bischoff (1954) no.133) or with strokes (Bischoff (1954) no.136)), though a complete substitution cipher alphabet with Arabic numerals might also have been known sometime in his reign (Bischoff (1954) no.141). Not only the letters in the alphabet but also "et", "est", "con", "tur" were covered by the substitution.

Hans Keller, an envoy of the Emperor to Italy in the 1480s, appears to have been given a cipher based on reciprocal substitution (Mitis, Oskar Freih von, "Blanquette und Chiffren zur italienischen Reise eines kaiserlichen Gesandten 1480", MIOG vol.26; cited in Auer p.154).

Maximilian I (1459-1519)

During the time of Maximilian I, son of Frederick III and Holy Roman Emperor from 1508, envoys used cipher (Auer p.154). It was to Maximilian that, in 1508, Abbot Trithemius dedicated Polygraphia (Noël L. Brann, Trithemius and Magical Theology (1999) p.133; Noël L. Brann, The Abbot Trithemius (1462-1516) (1981) p.93; Latin dedication), which would be posthumously printed in 1518 to become the first printed book on cryptology.

One cipher used in a letter from Rome to the Emperor in 1513 is reconstructed in Anton Walder, "Der letzte Ritter als erster Verschlüssler im Reich" in Geheime Post (2015). It is a simple cipher that employs arbitrary symbols and some Arabic figures to represent letters in the alphabet as well as "et", "con", and some words. It provides for homophones for vowels and "t".

Margaret of Austria (1519)

A similar cipher (with less figures) used in a letter from a Nassau (probably Count Henry of Nassau-Breda) to Margaret of Austria (daughter of Maximilian I and Governor of the Habsburg Netherlands), dated Achembourg, 13 April 1519, is known (see another article).

A 15th-century Cipher?

Below is another simple cipher between the Emperor and the Duke of Ferrara found in the archives of Modena (Meister (1902) p.38). (The family of Este was given the title of Duke of Modena and Reggio in 1452 and then Duke of Ferrara in 1471.) Meister says it belongs to the 15th century (p.33), though this uses more figures and less arbitrary symbols compared with those of Charles V and Margaret of Austria above.

The vocabulary in the nomenclature might help in dating this cipher. In 1509 during the War of the League of Cambrai, the Duke of Ferrara, Francesco Gonzaga, took Verona for Emperor Maximilian I and proceeded to the Venetian stronghold at Legnano but, before the battle began, the Duke was surprised and was taken prisoner (Emma Lucas (2014), Lucrezia Borgia (Google); Wikipedia). Legnano was taken by the French in 1510 (William Gilmore Simms (1847), The Life of the Chevalier Bayard (Google)). The context seems to match but is not enough to deny Meister's statement that this cipher belongs to the 15th century.

Meister says there is another cipher from the time of Charles V with this (p.33).

German cryptology was still in its infancy. In 1521, when Giovanni Soro, a Venetian cipher expert, submitted his plan to write a treatise about his experience with foreign ciphers, it was to cover "Latin, Italian, Spanish, and French ciphers." (Meister (1902) p.23)

Ciphers during Personal Union with Spain under Emperor Charles V

The Holy Roman Emperor Charles V, grandson of Maximilian I, inherited the Spanish crown in 1516 because untimely deaths had left his mother, Juana the Mad, the sole heir to the legacy of the Catholic Monarchs, King Ferdinand of Aragon and Queen Isabella of Castile.

Born and bred in the Netherlands, Charles was more versed in French than German. He even did not spend much time in Germany until the 1540s (Wikipedia).

At the time, Spanish ciphers were more advanced than German ciphers but apparently did not leave a lasting mark on German ciphers.

Flemish Ciphers?

Charles is known to have used cipher with his tutor Adrian of Utrecht (later Pope Adrian VI).

His correspondence with his ambassador in England, Louis de Praët (who came from Flanders) around 1524-1525 also used a cipher, of which the key was shared with the court of Margaret of Austria, Governor of the Netherlands. The cipher was simple enough to be learnt by heart. (See another article.) It might have been similar to the cipher between Nassau and Margaret above.

Spanish Cipher

On the other hand, the Emperor's letter to his ambassador in Genoa, Lope de Soria, used a complicated Spanish-style cipher.

There were other similar ciphers, including one used between Charles V and his brother Ferdinand, King of Hungary and later Emperor Ferdinand I (see another article).

Iñigo de Mendoza, successor of Louis de Praët as ambassador in London, appears to have possessed a Spanish cipher. Charles V personally pointed out its superiority over a Flemish cipher. (See another article)

... as the cipher which you possess is safer than hers [Margaret's], we will use yours, for the purpose of transmitting our orders and wishes thereupon. It would not do that such important matters as these should fall into the hands of our enemies, so that by deciphering the contents of the despatches they might gain any advantage.
Charles V to Iñigo de Mendoza, 30 May 1527

Vowel Indicator System

Spanish ciphers were developing themselves. Several ciphers used in the 1540s and possibly also in the 1530s employed a vowel indicator system, in which diacritic signs indicating vowels are combined with consonant symbols to systematically form syllables. (For discussion of the origin of such a scheme, see another article). Such a vowel indicator system was used in many ciphers during the reign of Philip II (see another article).

On the other hand, although Ferdinand I knew of such ciphers, it appears they did not take root in Germany.

Imperial Ciphers After Charles V

Ambassador to Rome (1568)

While many ciphers use graphic signs or (later) Arabic figures, one cipher used by an Austrian ambassador to Rome employ a two-letter combination to represent each letter in the alphabet and nomenclature elements (in the latter case, the first letter is often capitalized). More than 100 nulls ("errantes") are provided, which look similar to the other cipher letters (that is, nulls cannot be distinguished by other elements by their appearance). (Láng (2020), quoted below, p.88; Kt 13, Fasc 19, f.9-15)

Such a bigram cipher is also seen in a Venetian cipher of Michiel Surian (1557-1558) (in superscript form) (see another article), a cipher of Augustus of Brunswick-Lüneburg (see another article), and a French cipher at the time of the War of the Spanish Succession (see another article). The similarity may be coincidence.

Maximilian II (1575)

A cipher used by Maximilian II was identified in Nils Kopal & Michelle Waldispühl (2021), "Deciphering three diplomatic letters sent by Maximillian II in 1575" (Cryptologia, DOI). The authors succeeded in a ciphertext-only cryptanalysis by using a homophonic solver algorithm and thereafter, found a key in a database of manuscript ciphers in HHStA (see below). At first, the language model of French or Latin led nowhere. When a German language model was used, after a few restarts of the algorithm, meaningful fragments of German appeared.

Since space between words is kept in the ciphertext, frequently used words such as der, die, das could be easily identified. Umlauts are used in a naive manner. That is, ä is enciphered as a symbol for "a" with two dots above.

The cipher ("Cyffra nova ad Poloniam" from 1572) (identified as "Staatskanzlei Interiora, Chiffrenschlüssel, Kt. 13. Fasc. 20. f 28" in HHStA and named "ÖStA_HHStA_Stk_Int_Chiffrenschlüssel_fasc_20_28" in the DECODE database) assigns one or two graphic signs to each letter of the alphabet as well as double letters. Latin words (as well as Arabic figures 1-24) are used as nomenclature elements. For example, "Benignus" means Archduke Ernest of Austria and "1" stands for "Germania". Since there are some symbols that cannot be read with the original key, there may have been a variant in the cipher. Interestingly, Latin code words are inflected according to the corresponding plaintext word. For example, the genitive form "benigni" is read as "son's" (i.e., Archduke Ernest). Similarly, "ater" (meaning "unsure") appears in the nominative case in one place, but in the dative ("Atro") in another instance occurring after a preposition.

Nulls ("errantes") employ Latin words as well as graphic signs. This means that nulls cannot be distinguished from symbols for letters or nomenclature elements by their appearance. Although not many nulls are used, the name of the sender ("Maximilian") at the beginning was broken with multiple nulls in two letters (e.g., "###MAX#IMILI#AN##").

The decihpered letters turned out to be addressed to ambassadors in Poland and concern the election of King of Poland in 1575-1576. King Henry of Valois secretly left Poland in June 1574 to claim the French throne (to become Henry III) after the death of his brother Charles IX. In 1575, interregnum was declared and the election process began. In December, Maximilian II was declared King of Poland, but the decision was opposed by the Polish nobility. In the end, the election of the candidate supported by the nobility was confirmed. (Wikipedia)

The deciphered letter of 7 July 1575 (the date revealed by deciphering) includes arguments in favor of his son, Archduke Ernest, in the royal election. The second letter dated 24 December 1575 shows concern about the divided opinion in Poland and expresses a wish for a definite decision of his election. It also mentions a possibility of war. The third ciphertext (postscript to a letter of 23 December 1575) is a short summary of the second letter.

The authors further proceeded to achieve a preliminary decipherment of two letters written by a Lithuanian nobleman Jan Chodkiewicz to Maximillian II in 1574 and 1575 in another cipher (Nils Kopal, Michelle Waldispühl, "Two Encrypted Diplomatic Letters Sent by Jan Chodkiewicz to Emperor Maximilian II in 1574-1575" in HistoCrypt 2021 (pdf)). This time, the task was harder than before because there is no space between words.

The authors also noticed that a cipher of a Hungarian nobleman Andreas Dudith printed in an edition by Kotońska (1998) has nomenclature elements that fit for Chodkiewicz's letter, which suggests a possibility that the same nomenclature elements were shared among different ciphers.

Imperial Ciphers in the 17th to 18th Century

Ferdinand II

In 1621, Ferdinand II used a monoalphabetic cipher in his correspondence with Jacobus Curtis (Jakob Kurtz), his Polish trustee (Katherine Ellison, Susan Kim ed., A Material History of Medieval and Early Modern Ciphers (2017)).

Archduke Ferdinand and Archduke Leopold (1609-1610)

A cipher used between Archduke Ferdinand (later, Holy Roman Emperor Ferdinand III) and Archduke Leopold is described in Carolin Pecho, "Der Habsburger-Code. Chiffrierte Briefe von Erzherzog Ferdinand an Erzherzog Leopold während des Erbfolgekrieges um Jülich-Kleve als Versuche der Gemeinschaftsstiftung (1609-1610)" in Geheime Post (2015).

It was a simple substitution cipher based on pairing, as reconstructed by Pecho. (The following is the version used by Archduke Ferdinand. Archduke Leopold used a slightly modified version in which X and W were exchanged.)


The key can be derived by some manipulation using a keyword "Carolus." An idea of creating a substitution table with such a memorable keyword was made widely known by Giovan Battista Bellaso in the 1550s (See the links in another article (in Japanese)).

The brother archdukes' political positions were diverging at the time, with Archduke Leopold closer to the Emperor. Pecho considers the use of cipher in this period was more as a bond between the brothers rather than for strict secrecy and points out that their later corresopndence without cipher was more formal and in a reserved tone.

Ferdinand III and Archduke Leopold (1640-1645)

The brothers Ferdinand III and Archduke Leopold used cipher in their private correspondence in 1640-1643 and 1645 (Auer p.158). The ciphertext (in Latin) was broken only in 2017 by Thomas Ernst, a German philologist and historian (Klausis Krypto Kolumne).

It was basically a numerical substitution cipher but digits may alternatively be written in geometrical figures having the number of strokes equal to the digit. For example, a square, "M", "W", "Σ" all have four strokes and thus mean "4". Similarly, a triangle means "3", and so on. This discovery was the breakthrough in Thomas Ernst's codebreaking. He must have noticed that these glyphs as well as Arabic figures tend to occur in pairs, which suggests that the whole cipher essentially consists of nothing more than two-digit pairs.

Once geometrical figures were rendered in Arabic figures, the next breakthrough was an idea that A is 01 or 11, E is 02 or 12, I is 03 or 13, and O is 04 or 14, motivated by a special position of AEIOU for the Habsburgs. This turned out to be successful. (A similar success of a seemingly too-good-to-be-true hypothesis is described in another article with regard to papal ciphers.) This hypothesis led to a partial reading of a marginal note "Zifra 21(?) 13(I) 42(?) 04(O) 23(?) 04(O) 33(?) 03(I) 43(?) 02(E) 01(A)" at the beginning of the letter. It is natural to assume this specifies the cipher used in the letter and Thomas Ernst had tried hypotheses such as "Zifra FRATELLO", "Zifra LEOPOLDO", etc. After such trials, his knowledge in history immediately reminded him of the name Piccolomini (it is common to suppress double letters in enciphering). "Zifra Piccolominea" meant "Piccolomini's cipher."

The development of the codebreaking process can be traced in many comments made by Thomas Ernst in Klausis Krypto Kolumne. (The first report of his solution is on 15 September 2017.) It is remarkable Thomas Ernst guessed 12=E even before the initial breakthrough (his comments on 15 July 2017).

(The above cipher follows the latest reading of 9 October 2017 but there remain uncertainties about symbols for B, H, P. His official publication should be consulted when it is made available.)

In the 17th century, ciphers tended to use less and less graphic symbols in favor of Arabic figures in Germany as well as in England, France, and Spain. This Ferdinand III-Leopold cipher is interesting in being essentially a figure cipher with a deceptive appearance of mixed use of figures and symbols.

Ferdinand III

In 1632, the Emperor's secret agents used a cipher of graphic symbols in Italian letters (A Material History of Medieval and Early Modern Ciphers (2017)).

Athanasius Kircher (1602-1680) was occasionally involved with decoding intercepted Swedish letters for Ferdinand III (Karl de Leeuw, Cryptology and Statecraft in the Dutch Republic, p.8). (Ferdinand's reign was 1637-1657. Sweden was involved in the Thirty Years' War on the Protestant side.) (Kircher claimed to have deciphered antient Egyptian writing in hieroglyphs, which later turned out to be incorrect. In 1666, he was sent the Voynich Manuscript in the hope that he may be able to decipher it. (Wikipedia).)

Maximilian von und zu Trauttmansdorff

Fig.15 of Mírka (3-4/2013) is a letter of 20 June 1637 partly in cipher from Johann Karl von Schönburg (Wikipedia) to Maximilian von und zu Trauttmansdorff (Wikipedia), who was Prime Minister of Emperor Ferdinand III and headed the delegation of the Empire in negotations for the Peace of Westphalia from 1645 to 1647. The reconstructed cipher is no.8 of Mírka (11-12/2013).

Fig. 1 (p.18, monoalphabetic substitution cipher, February 1644) and Fig. 2 (p.19, homophonic substitution cipher, syllables, nulls called "Errantes", February 1644) of Mírka (11-12/2012) seem to be related to Maximilian of Trauttmansdorff. The latter looks similar to the cipher for Marchese de Grana from the 1660s below.

Fig. 4 (p.23) of Mírka (11-12/2012) is a letter mostly in cipher, with some interlined decipherment, from Ferdinand Sigismund Kurtz von Senftenau (Wikipedia) to Maximilian of Trauttmansdorff on 17 January 1639. The reconstructed cipher, given as no.9 of Mírka (11-12/2013), looks similar to the cipher for Marchese de Grana from the 1660s below. A reconstruction of another cipher used by the same correspondents in 1646 is no.10 of Mírka (11-12/2013).

Fig. 5 (p.26) of Mírka (11-12/2012) is a letter partly in cipher, with interlined decipherment, to Maximilian of Trauttmansdorff on 7 April 1625 from Constantinople.

Fig. 6 (p.28) of Mírka (11-12/2012) is a letter mostly in numerical cipher, with decipherment in the margin, to Maximilian of Trauttmansdorff on 14 July 1646 from Joachim Friedrich von Blumenthal (Wikipedia), who represented both Brandenburg and the Holy Roman Empire at the negotiations for the Peace of Westphalia.

No.11 and no.12 of Mírka (11-12/2013) are ciphers reconstructed from Latin correspondence in 1645 between Franz Paul de Lisol (?), Johann Tasselt (?), and Maxmilian of Trauttmansdorff.

Two Ciphers of Imperial Ministers from the Thirty Years' War Period

Several letters in cipher written by Imperial ministers to the Holy Roman Emperor in the period of the Thirty Years' War, preserved in the Vienna Court Archives and photographed by Peter Nüchterlein, were solved in 2018. Two ciphers were identified.

Cipher of Figures/Symbols (ca.1637-1641?)

One is a homophonic substitution cipher with Arabic figures and graphic symbols. It was solved by Thomas Bosbach and Nagra independently within the day (Klausis Krypto Kolumne).

One deciphered passage of one letter reads: "Die Landgravin zu Hessen sol an ihren ragt(?) abhie Doct. Teichman geschrieben haben, daß sie von Herzog George beredet worden." This indicates the letter may have been written between 1637, when Landgravine of Hesse, Amalie Elisabeth, became regent of Hesse-Kassel (Wikipedia), and 1641, when Duke George of Brunswick-Lüneburg died (Wikipedia).

While Nüchterlein considered this pertains to Trautmannsdorf, Thomas Bosbach considers a Vienese resident in Brunswick-Lüneburg (such as Count of Tattenbach-Reinstein, Viennese ambassador to Brunswick) is more likely because the Duke of Brunswick-Lüneburg is mentioned simply as "Herzog George."

Numerical Cipher (ca.1637-1648?)

The other is a homophonic substitution cipher in figures. It was solved by Thomas Bosbach in one night (Klausis Krypto Kolumne).

A deciphered passage in one letter reads: "Die gefassete Luneburg und Brunsweigische Resolution alle dinge ohne vorbewusst test Hern Ertzbischofs zu Bremen nicht Erfolgt." Thomas Bosbach considers the Count of Tattenbach-Reinstein was the sender.

(The figures in brackets are conjectures. The placement of W after X, Y, and Z is also found in Gaspar Schott's book. See another article.)

General Cipher for Imperial Ministers (ca.1680-1690)

Fig. 7 (p.3) of Mírka (1-2/2013), the same as no.2 of Mírka (11-12/2013), is a general cipher for imperial ministers (ca.1680-1690). Use of graphic symbols and lack of syllable representation make it look an older style than the 1691 cipher for Conte Marsigli below.

Leopold I

Emperor Leopold I (reigned 1658-1705) personally used cipher (Auer p.158).

Marchese di Grana

A cipher for Marchese di Grana is given by Auer (Fig. 1) as a typical example with increased use of figures rather than graphical symbols. Grana was on various diplomatic missions in the Holy Roman Empire and Spain in 1664.

This cipher is similar to much older Spanish cipher Cg.13 used around 1587-1592 and other similar ciphers (which I call "syllabic numerical ciphers") used around 1592-1593 (see another article) as well as some contemporary Spanish ciphers (see another article, Cg.51 and "A Cipher from Devos (1967)").

There was another cipher in an instruction for the same Marchese di Grana in (1668), which was also used in correspondence between Leopold I and Pötting.

According to Auer, these two, both characteristic in the syllable representation, were characteristic of the Reichskanzlei in the 17th century. It is possible that the "syllabic numerical ciphers" were imported from the Habsburgs in Spain but evidence for such a link is yet to be found.

Again, according to Auer (n.51), there was already two-part code of a few hundred words. (The size increases to 5000 in the time of Maria Theresia.)

Emperor Charles VI

Fig. 8 (p.6) of Mírka (1-2/2013) is a letter of 3 February 1720 of Emperor Charles VI partly in numerical cipher. Not deciphered but can be read with a key found in the archives.

Emperor Charles VI (1685-1740) used cipher in part of his private diary (Auer p.158).

Ciphers with Imperial Ambassadors to Constantinople

In 1628-1629, Johann Ludwig Kuefstein, ambassador in Constantinople, used a cipher consisting of graphic symbols in writing to Ferdinand III. It was almost a monoalphabetic cipher, with a few homophones. Use of such graphic symbols had been common around 1571, when Carolus Rym wrote from Constantinople (A Material History of Medieval and Early Modern Ciphers (2017)).

Eight ciphers given to ambassadors and residents in Constantinople from 1643 to 1741 are described in some detail in Gerhard Kay Birkner, "Briefe durch Feindesland. Die chiffrierte Post Wien-Istanbul um 1700" in Geheime Post (2015).

All are figure ciphers, representing single letters, syllables, words and names (and also punctuations, months, currencies, and numbers from 1740B). Nulls ("Errantes") are provided except for the 1691 and 1740B ciphers.

1643: Cipher for Johann Rudolf Schmidt Baron von Schwarzenhorn, resident in Constantinople.

1645: Copy of the 1643 cipher for Hermann Graf Czernin von und zu Chudenitz, ambassador in Constantinople.

1691: Cipher for Luigi Ferdinando, Conte di Marsigli. The Emperor was then at war with the Ottomans (1683-1699) (Wikipedia) but the Count was sent on secret peacekeeping missions from 1690 to 1692 in Constantinople and Wallachia.

This cipher is reproduced as Fig. 1. Each letter of the alphabet is assigned three figures; each syllable is given two figures. The nomenclature includes figures 216-347 for names and words (in Italian). Marginal instructions are in Latin. Although the arrangement of figures is largely alphabetical, it is somewhat improved when compared with the Cipher for Grana above in the 1660s.

1699: Cipher for Wolfgang von Oettingen-Wallerstein, who was imperial ambassador for negotiations for the Peace of Karlowitz (Wikipedia).

1734: Cipher for Leopold Freiherr von Talmann, resident in Constantinople.

1740A, 1740B: Ciphers for Anton Corfitz, Graf von Uhlefeld, ambassador, then resident in Constantinople in 1740-1741. Two-part codes, necessitated by the increased size. 1740A has about 900 figures. Although 1740B had figures from 1 to 2312, the net number of entries was about 850 because of many blanks. 1740B, in part reproduced in Fig.2, has fewer homophones for letters than the 1691 cipher, but it is a significant improvement in respect to irregularity in assignment and increased size.

1741: Cipher for Heinrich Christoph Edler von Penckler, secretary under Uhlefeld, then resident. It had a range of figures 1-800, with few blanks (extended to 900 by a different hand).

Cipher with Ambassador to Rome

Fig.3 (p.20) of Mírka (11-12/2012), the same as no.6 of Mírka (11-12/2013), is a cipher for Juan Alvaro Cienfuegos Villazon (Wikipedia), cardinal and the Holy Roman Emperor's ambassador to Rome. This looks like the 1691 cipher for Conte di Marsigli above.

Family Windisch-Graetz

Fig. 14 (p.8) of Mírka (3-4/2013) is about Gottlieb von Windisch-Graetz (Wikipedia) (1673). The reconstructed cipher is no.7 of Mírka (11-12/2013).

Figs. 12-13 (p.5-6) of Mírka (3-4/2013) is a cipher for Leopold Victorin of Windisch-Graetz, an imperial envoy to the Low Countries (ca. 1720). Not only syllables "ba", "be" ,..., bigrams "ab", "ac", "ad", "af", ... are provided.

Nos. 1-6 of Mírka (11-12/2012) are original ciphers preserved in the family archives of Windisch-Graetz. No.1 and no.2 (mentioned above) appear to be in German. No.2 and no.4 are in French. No.5 and no.6 (mentioned above) are in Italian.


Marie-Antoinette used a polyalphabetic cipher in her correspondence with the court in Vienna. It is the same cipher as the one she used with Axel von Fersen. (See another article.) It is yet to be found out whether the cipher was introduced by the Queen, Fersen, or Vienna.

Four Hundred Years of Habsburg Cryptography

(Section added in March 2021)

Nearly 500 manuscript ciphers from the 16th to 19th centuries are preserved in the archives of the State Chancellery (Staatskanzlei) of the House, Court, and State Archives (Haus-, Hof- und Staatsarchiv, HHStA) in the National Archives of Austria (Österreichisches Staatsarchiv, ÖStA) in nine large boxes (ÖStA HHStA Staatskanzlei Interiora Chiffrenschlüssel Kt. 13-21), of which an overview is given in Benedek Láng (2020), "Was it a Sudden Shift in Professionalization? Austrian Cryptology and a Description of the Staatskanzlei Key Collection in the Haus-, Hof- und Staatsarchiv of Vienna", Proceedings of the 3rd International Conference on Historical Cryptology, HistoCrypt 2020 (LiU Electronic Press, DOI).

The following outlines some aspects of the development of Habsburg ciphers relying on Láng (2020).

From Sheets to Codebooks

Typical ciphers among the 16th to 17th century ciphers in Kt.13 and Kt.14 consist of one page, with homophonic substitution with three or four symbols for each letter of the alphabet and a nomenclature with about one hundred entries.

During the reign of Emperor Charles VI (1711-1740), ciphers were provided on a sheet. Later, ciphers in the form of a booklet were common. Many ciphers (in French) for ambassadors from the second half of the 18th century (Kt.15, Fasc.21 and 22) had more than a thousand (but less than 10000) entries. In the second part of Maria Theresa's time (1740-1780), two-part code of about 10000 entries with 2-4 page instructions was the norm (Kt.17, Fasc.24).

Láng considers the turning point was the creation in 1742 of the State Chancellery for foreign affairs separate from the Court Chancellery for domestic affairs (Wikipedia), though, of course, "evolution of the methods is not uniform".


Symbols constituting the cipher alphabet of the 16th to 17th century ciphers in Kt.13 and Kt.14 are usually numbers. Graphic signs are also used. (I guess graphic signs gave way to Arabic figures in the first half of the 17th century, but this needs to be checked with the database records.)

Code Words

Before a nomenclature with Arabic figures became common, foreign words were used to represent names in many ciphers.

One cipher used in relation to a Polish delegate (Kt.13, Fasc.19, f.28) has elements such as: Dux (primus), Princeps (secundus), Pontifex maximus (bonus), Imperator (gravis), Imperatrix (mens), Palatinus Cracoviensis (meaning "Palatine of Krakow"?) (species).

A cipher of Ogier Ghislain de Busbecq, a Habsburg delegate to the Turks between 1554-1562, and those of Castaldo and Caraffa have "sophisticated metaphors in the nomenclator table": Papa (Andromedes), Cardinalis (Antistes), Petrus Aldombrandinus (Amorius), Imperator (Benignus), Rex (Bruno).

One cipher (Kt.14, f.311-313) has code names such as Papa (pater), Imperator Carolus (dominus), Rex Francorum (patronus), Rex Angliae (theologus), Rex Poloniae (amicus), Eques (vacca). ("Rex Angliae" suggests this does not belong to the Elizabethan period. Maybe Charles V's time?)

Two-Part Code

Two-part code provides an enciphering table sorted alphabetically ("chiffre chiffrant") and a deciphering table sorted numerically ("chiffre déchiffrant").

An undated cipher used in relation to Berlin (Kt.13, Fasc.19, f.32-33, 34-36) is one example of two-part code.

A cipher of Prince Eugene of Savoy from the 1690s is an early example of large two-part code, consisting of two- to four-digit figures up to 2400. (None of the contemporary French ciphers as described in another article has numbers above 1000.) (Prince Eugene distinguished himself in the war with the Ottomans in the 1680s. When war broke out with France, he fought on the Rihne and in Italy. After the Duke of Savoy, his second cousin, allied with France, Prince Eugene was given command of the imperial forces in the east and won a decisive victory against the Ottomans at the Battle of Zenta in 1697.)

A cipher for delegate Hoffman in London (1721) (Kt.13, f.152-157) also appears to be two-part code. It has one thousand elements, assigning three codes to each letter and each bigram (ab, ac, ad, ...). For example, "a" can be represented by 51, 41, or 40; "ab" by 122, 44, 41. (The same number occurs in different entries. Rather than homophones, the three columns may be three separate codes.)

It appears that two-part code consisting of separate "chiffrant" and "déchiffrant" parts was common in the reign of Emperor Charles VI (1711-1740). But the two-part principle is not always adhered to, even with a code having a large vocabulary. A code (Kt 15, f.3-14) is a four-digit code with nearly 10000 entries, but the arrangement is alphabetical.

Even in 1750, a similar one-part code (Kt 15, f.19) was used (in relation to France). It is a large one-page code with nearly 2000 entries.

A "French speaking system from Milano" (1824) (Kt 15, f.38-47, 48-53) is a two-part code in a booklet form. It consists of four-digit numbers.


Láng observes that nulls in Habsburg ciphers tend to include all the types of symbols. It may prevent a codebreaker from recognizing nulls from their appearance.

A cipher for Carolus Rym (1570) (Kt. 14, f.291-302) employs conjunctions and other common words in Latin as nulls (e.g., quapropter, deinde, simulatque, quoniam, mandavimus, dedimus, renunciatum). It may mislead a codebreaker.

A code from 1750 (Kt 15, f.19) consists of odd numbers, and all even numbers served as nulls.

Printed Template

There are printed templates (Kt. 14, f.132-135, 136-141), on which an alphabet and a large list for a nomenclature are printed. This allows a new code to be prepared by simply filling numbers in the entries. (Similar printed templates were used in England around 1690 (see another article). Later, it was common and was also used by Americans (see another article etc.).)


From the 18th century, not only short remarks, but two-page long instructions were provided to guide the proper use of the code.

©2018 S.Tomokiyo
First posted on 10 September 2018. Last modified on 3 October 2021.
Cryptiana: Articles on Historical Cryptography
inserted by FC2 system