An encrypted message from 1939, sent from the German battleship Schleswig-Holstein to an address in Kiel, Germany, waits to be solved. Can a reader help?

The Schleswig-Holstein, a German battleship, fired the first shots of World War II when she bombarded Danzig in the early morning hours of September 1st, 1939.

 

The Schleswig-Holstein

The Schleswig-Holstein, named for the German federal state of Schleswig-Holstein, was built from 1905-1908. She fought in both World Wars. During World War I, she saw front-line service in the Battle of Jutland. As one of the few battleships permitted for Germany by the terms of the Treaty of Versailles, the Schleswig-Holstein was again pressed into fleet service in the 1920s. In 1935 she was converted into a training ship. The Schleswig-Holstein was used as a training vessel for the majority of WW2, before she was sunk by British bombers in December 1944.

Schleswig-Holstein

 

An unsolved cryptogram

My friend Tobias Schrödel, who is known, among other things, as a comedy hacker, book author, and crypto book expert, owns a very interesting document: an encrypted teleprinter message that was sent from the Schleswig-Holstein to a household in Kiel, Germany, in 1939. Here it is:

Schleswig-Holstein-Cryptogram

This message, which I will call “Schleswig-Holstein cryptogram”, was contained in the following envelope:

Schleswig-Holstein-1939-Vorderseite-2

The recipient of this mail was a person or family named Dabis in Kiel, Germany. The surname Dabis is quite rare in Germany, but it exists in the Kiel area. The letter was stamped on January 6th, 1939, in Kiel. On the rear side of the envelope a shopping list is noted (I’m quite sure it was written by the receiver):

Schleswig-Holstein-1939-Kuvert-Rueckseite

When Tobias purchased this envelope, it contained a second document (a telegraph):

Schleswig-Holstein-1939-Telegram-Weihnachten-1938-Vorderseite

This sheet, which is not encrypted, is addressed to the same person (Dabis) and dated 24.12.1938. The text is a Christmas greeting from a “lonesome seaman”. It is quite likely that the envelope originally only contained one of the two sheets and that the other one was put there for storage.

 

How to solve it?

Here’s a transcription of the Schleswig-Holstein cryptogram:

ZASUG RNGSE LNHET NLLUI TDSEE OCBAU AMTXU DBNAN NEUFX SPSRR UDLTH+

According to George Lasry, who is known to many readers of this blog, it is probably not a teleprinter cipher and not an Enigma.

If neither a teletype cipher machine (like the Siemens & Halske Geheimschreiber or the Lorenz machine) nor an Enigma was used, a codebook is certainly an option. In addition, many manual ciphers were in use in the German navy. Having some background knowledge about the cipher technology used on the Schleswig-Holstein would be very helpful.

Can a reader say more about this cryptogram and its background? Tobias and I would be very interested.


Further reading: An encrypted letter from World War 2

Linkedin: https://www.linkedin.com/groups/13501820
Facebook: https://www.facebook.com/groups/763282653806483/

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Kommentare (35)

  1. #1 Jim Gillogly
    United States
    2. September 2017

    I suggest that it’s likely a transposition cipher: the IC is much too high for a good machine system, and it uses all letters except jkqvwy.

  2. #2 Michaela Ellguth
    Berlin
    2. September 2017

    Die MS Schleswig Holstein war ein Flaggen- und Schulschiff. Sie befand sich am 6. Januar 1939 wahrscheinlich in Kiel zur Wartung, bevor sie dann im August 1939 nach Danzig aufbrach.
    Kommandant: 8. Juni 1938 bis 25. April 1939 Kapitän zur See Gustav Kieseritzky.
    Es gibt die Marine-Kameradschaft Schleswig-Holstein. Da das ein Kadattenausbildungsschiff war, besteht de Möglichkeit, das heute noch jemand lebt, der auf dem Schiff zur Ausbildung war und dort könnte man ggf. mal fragen, ob das jemand lesen kann oder vielleicht sogar kennt https://www.sophie-x.de/historie/linienschiff/linienschiff-1.htm
    Alle Verschlüsselungen bestehen aus 5 Zeichen, sind also keine Worte. Ich tippe auf teilverschlüsselte Abkürzungen.

  3. #3 Tobias Schrödel
    München
    2. September 2017

    Hello Klaus,
    thank you for this article!
    The envelope is for sure from the encrypted message and *not* from the christmas greeting. It is hard to see, but the date 6-1-39 is written in the post stamp.
    One can see it quite clear in a b/w version:
    https://www.sichere.it/download/05_Stempel_Kuvert_bw.jpg
    Many greetings!
    Tobias

  4. #4 Jim Gillogly
    2. September 2017

    Normal ADFGVX is not possible: it only uses 6 letters in the ciphertext. ÜBCHI is not ruled out by the statistics I noted above, but perhaps the cipher is too short for unambiguous solution unless the keys are short.

    Perhaps the two X’s appear together at the end for padding, or in the middle together for punctuation.

  5. #5 TWO
    2. September 2017

    I agree with Gillogly.

    Most likely a transposition system

    In the beginning the German word ZU is easily visible.

    ZU is the German word for TO

    regards,

    TWO

  6. #6 Thomas
    3. September 2017

    Da der verschlüsselte Text auf dem Formular des Marinenachrichtendienstes und der Text auf dem Formular der Reichspost dieselbe Anzahl von Buchstaben haben und auch zusammen in dem Umschlag waren, würde ich Folgendes vermuten: Auf dem ersten Formular befindet sich der von der Marinefunkstelle in Kiel aufgenommene verschlüsselte Telegrammtext, der dann entschlüsselt auf ein Formular der Reichspost zur Zustellung an den Empfänger übertragen wurde, das vermutlich aber bei der empfangenen Nachricht verblieben ist. Die Weihnachtsgrüße scheinen mir der (zugegebenermaßen allerdings nicht besonders geheimhaltungsbedürftige) Klartext der verschlüsselten Nachricht zu sein. Eine Transpositionschiffre wäre es dann nicht.

  7. #7 Jim Gillogly
    3. September 2017

    Ah. Interesting, but by my count the text in the Reichspost message has 56 letters, while the enciphered text has 55. Perhaps they would match if (z.b.) “wuenscht” were spelled “wunscht” in the cipher.

    I don’t think a commercial code is likely, since it’s hard to see how the assumed plaintext would be broken up into 11 code groups; and in any case, commercial codes were typically more pronounceable.

    In favor of transposition: the missing letters are all low frequency in German, the index of coincidence is high, and the number of vowels (31%) is similar to the number of vowels in typical German (36%)… I may be a little off on the vowels because of umlauted letters being represented with e as in “wuenscht”, but in any case the ratio is close. The relatively low frequency of E would be unusual, but having an H to potentially follow the C is reassuring.

    I’ve tried it as ÜBCHI (double columnar transposition with a single key) with no success. One source said the plaintext messages were padded to even periods, which would mean key length 5 or 11. I’m confident that there is no double transposition solution with a single key for either length. Other sources don’t mention the padding, but I’m fairly confident that there’s no solution up to around length 13 with or without padding.

  8. #8 anderer Michael
    3. September 2017

    Die Schleswig Holstein war vom 29.10.38 bis 10.03 .1939 als Schulschiff im Atlantik unterwegs.
    Es gab einen Offizier mit Namen Martin Dabis bei der Reichsmarine. Mir ist bekannt, dass er ein U-Boot – Jagdschiff befehligte mit dem prosaischen Namen UJ 1703 und dem Zusatz “Mob FD63” bei der 17.UJ Flottille, stationiert in Norwegen.Dieses wurde am 5.3.1944 durch Flugzeuge der Westallierten versenkt.Er muss den Angriff überlebt haben , ihm wurde später ein hoher Orden verliehen.Über sein weiteres Schicksal oder sonstige biographische Angaben habe ich nichts gefunden.Ob dieser Martin Dabis identisch mit einem Dabis als Absender ist , weiß ich nicht, weil ich die Crewlisten nur sehr unvollständig einsehen konnte
    Ich ging davon aus, dass der Inhalt privat sein muss und dem Klartext entsprechen könnte.Ich kann mir nicht vorstellen , dass die Reichsmarine banale private Texte mit der Enigma versendet ( möglich ist allerdings alles, schließlich wurde die Enigma bis 1944 für nicht entschlüsselbar gehalten, von den Deutschen).Solche kurzen einfachen vorhersehbaren Texte bieten eine guten Ansatzpunkt für eine Entschlüsselung.Also spekulierte ich , dass eines der vielen bereits im 1.Weltkrieg verwendeten Verfahren zur Funkübermittlung mittels Morsecode verwendet wurde. Warum so eine Nachricht überhaupt verschlüsselt wurde, kann ich nur spekulieren.Vielleicht zu Trainingszwecken.Die Schleswig-Holstein war ein Schulschiff und später war nicht immer eine Enigma vorhanden.
    Also, ich bin bei den dreisprachigen Texten , die ich so mühevoll las ,etwas mit den Begriffen durcheinander gekommen.Tut mir leid Herr Schmeh.
    Sie haben meine nicht sehr fundierten kurzen Kommentare ( dumm könnte man auch sagen) gelöscht , ist nicht schlimm. Ich bin deswegen nicht beleidigt.

    Könnte es eine Vigenere Verschlüsselung mit einem im Vergleich zum Text längeren Schlüssel sein, das soll dann faktisch nicht lösbar sei, sofern der Klartext nicht bekannt ist ( und ich nicht wieder etwas verwechselt habe oder es doch eine Enigma Nachricht ist.)An einer Kasiki Analyse habe ich mich erfolglos versucht, ich bin auf ein Codeword mit zwei Buchstaben gekommen.Aber das kann es wohl nicht sein.

  9. #9 Jim Gillogly
    3. September 2017

    Either Enigma or Vigenere would have much more even letter frequencies. Tobias Schrödel’s observation strongly suggests that the Christmas greeting and the enciphered message are not directly related. In any case the cipher is not an encryption of the Christmas greeting with Enigma, since two of the first 55 letters would have the same plaintext and ciphertext, which cannot happen with Enigma. If one shortened “wuenscht” by one letter, three of the 55 would match.

  10. #10 Thomas
    3. September 2017

    The German navy made use of the ´Reservehandverfahren´ (Manual Backup Procedure) when no Enigma was at hand. Michael Hörenberg provides a very detailed description of this method (including a Reservehandverfahren program in part 5):
    https://enigma.hoerenberg.com/index.php?cat=Reservehandverfahren
    The ´Marinedienstvorschrift` shown on Hörtenberg´s Website dates from 1940, but I could not find out when this procedure was introduced the first time.
    BTW: I wonder why this message contains five letter groups, according to many sources the navy, unlike the army, used to divide their ciphertexts into four letter groups.

  11. #11 Thomas
    3. September 2017

    Another manual cipher of the navy was the `Werftschlüssel´. This cipher was based upon bigram substitution and yielded five letter groups. See Klaus Schmeh, Codeknacker gegen Codemacher, 3rd ed., p. 66, https://de.wikipedia.org/wiki/Werftschl%C3%BCssel

  12. #12 Tobias Schrödel
    München
    3. September 2017

    Thank you all for your contribution!!!

    THE ENVELOPE AND WHAT IT CONTAINED:
    As Jim Gillogly already noted, I strongly think, that the envelope originally contained the cipher message only.
    Here is why:
    Both (envelope & cipher text) show the date Jan 6th, 1939 -> see this b/w pic:
    https://www.sichere.it/download/05_Stempel_Kuvert_bw.jpg
    And secondly: Klaus did not use all the pics I provided. On the backside of the christmas greeting there is a handwritten “arrival notice” stating: Arrived Dec 24th, 1938.
    https://www.sichere.it/download/06_Telegram_Weihnachten_1938_Rueckseite.jpg
    Both messages were simply kept by the receiver in one envelope over the years. Thats all, I think.

    THE IDEA OF A TRANSPOSITION:
    I very much like the idea of a more or less simple transposition. During the last years (I bought that envelope around 2008) I was so stuck on a teleprinter cipher, that I never had other possibilities in mind. But your thoughts make sense! And as the message is possibly (maybe even likely?) a private message, a paper & pen cipher seems more likely to me, than the official “Reservehandverfahren” or “Werftschlüssel”.

    Again – thank you all for your help!
    Tobias

  13. #13 werner67
    3. September 2017

    Unabhängig vom Verfahren: ich halte es nicht für unwahrscheinlich, dass auch triviale Nachrichten gezielt verschlüsselt wurden. Das bedeutet nämlich einen bedeutenden Mehraufwand für eventuelle Abhörstationen und der Entscheid, was nun wichtig ist und was nicht, dauert viel länger.

    Regardless of the procedure: I do not think it is unlikely that trivial messages also specifically have been encrypted. This means a significant additional effort for possible listening stations and the decision as to what is important and what is not will take much longer.

  14. #14 Jim Gillogly
    3. September 2017

    I agree that trivial messages might well be sent: one way cryptanalysts would find an entry to new cipher systems was trying the standard training messages operators would send (e.g. well-known drinking songs), matching the length of previous such messages to ones in the new system. A ship on a training assignment might be more likely to do this than one on active duty.

    Regarding the system: I think we can eliminate both the Reservehandverfahren and the Werftschlüssel system. Both used digraph tables, which would cause much greater mixing.

    The link Thomas gave for RHV shows a ciphertext of comparable length to the Schleswig-Holstein message: JIVGECURUOKLWDSQSPLCLFXRMMMIWSGMJLHJFQTJSWNFOKLSCDJGRVZVHVGWVMDVZYLGEVLYLVRUPYVBXWNL. Unlike the S-H message, this has a low index of coincidence and good coverage of the alphabet, with only one letter (A) unused. The vowel percentage is about 11%, much lower than S-H or German. Nearly all of the most common letters in German have low frequency in the RHV sample.

    The link Thomas gave to the Wikipedia article on Werftschlüssel implies an even number of 5-letter groups, since they are enciphered in vertical pairs — if there was no procedure for a single 5-letter group at the end, this would rule this system out for the 11-group S-H message. In any case, Werftschlüssel would also give a much smoother frequency count because of the digraph substitutions.

  15. #15 Max Baertl
    3. September 2017

    Maybe it’s some kind of route transposition.

  16. #16 Jim Gillogly
    4. September 2017

    Yes, or perhaps a columnar with blocked-out cells.

  17. #17 Thomas
    4. September 2017

    On the other hand the index of coincidence seems to be to low for a pure transposition cipher. What catches the eye are the German words SEE, BAU and AMT. Maybe an imperfect/incomplete transposition enhanced with substitution or nulls? That could explain the medium IC.

  18. #18 Thomas
    4. September 2017

    Sorry, I overlooked the German word NEU. Hence, this short ciphertext contains four legible words. Mere coincidence?

  19. #19 Jim Gillogly
    4. September 2017

    If Max’s suggestion is correct those words could be where the route goes left to right.

  20. #20 Thomas
    5. September 2017

    Ein paar Kleinigkeiten noch zum Hintergrund:

    Das Adressbuch von Kiel für 1940 führt unter der Anschrift Hohenrade 42 als Haushaltungsvorstand (weitere Familienmitglieder sind nicht angegeben) auf: Günther Dabis, Marineoffizier (https://wiki-de.genealogy.net/w/index.php?title=Datei%3AKiel-AB-1940.djvu&page=86).
    Ein Herr Dabis (Martin oder Günther oder…) lässt sich der Kadettenliste des Oktober 1938 eingetretenen Crewjahrgangs (“Crew 38”) für die “Schleswig Holstein” nicht entnehmen, weshalb es sich wohl um ein Mitglied der Stammbesatzung gehandelt hat.

    Im Januar 1939 (seit Oktober 1938) war die “Schleswig Holstein” auf ihrer letzten Ausbildungsfahrt um Südafrika und nach Westindien. (enthält das Telegramm vielleicht Angaben dazu, wo sich das Schiff gerade befand?)

  21. #21 Norbert
    6. September 2017

    If it’s a route transposition, we should maybe consider the possibility of a triangle, because …

    f
    if
    tyf
    ivel
    etter
    scanea
    silybew
    rittenin
    atriangle
    likethisxx

    The plaintext might have been written in one specific route into the triangle (or whatever shape), and read out in another route.

  22. #22 Thomas
    6. September 2017

    Good idea, as 55 equals the sum of the numbers up to ten. Up to now I’ve tested only 5×11. However, the two X probably are nulls at the end, preceded by the name (MARTIN, GUENTHER or whoever). GUENTHER seems unlikely because – if the cipher is a pure transposition – in German the C should be followed by the H.

  23. #23 Thomas
    6. September 2017

    BTW: In a forum I found that “Maprifu’ stands for ‘Marineprisenfunk’. Can anyone confirm that? Still musing about the meaning. ‘Prise’ is the German word for ‘prize’ in the sense of maritime law (captured ships). Marine Prize Radio???

  24. #24 Jim Gillogly
    6. September 2017

    A Google search on “mapriful” called up two pointers to postage stamp auctions, including the note “War navy 1937, marconigram “MaPriFu” from liner “Silesia” over coastal radio station Kielradio to the fiance of a crew members. Extremely rare form with… “. Going to the website showed a number of auctions by a Dr. Reinhard Fischer, but when I visited the site I didn’t see the referenced message. it’s a very long URL, shortened here: https://goo.gl/btQW4W . Could the “Pri” stand for something else, like Privat?

    Regarding the triangular array – I noticed that also, and tried a number of patterns with no success. I’m dubious, because that method does not generalize directly to other lengths of messages. It would be fine for a puzzle, but that’s not the context here.

  25. #25 Jim Gillogly
    6. September 2017

    Groups 3 and 11 each have an H, so there are enough to handle both CH and Gunther if necessary.

  26. #26 Thomas
    6. September 2017

    @Jim
    If you haven’t found the Silesia message from 1937: https://www.delcampe.net/fr/collections/timbres/allemagne-non-classes/kriegsmarine-1937-funktelegramm-maprifu-ab-linienschiff-schlesien-ueber-kuestenfunkstelle-kielradio-an-395875656.html
    There is another maprifu message from 1931: https://picclick.de/Telegramm-1931-Marine-Nachrichtendienst-Kielradio-Kiel-Radio-K%C3%B6ln-132270594000.html
    ‘pri’ = privat was also my first guess, but in a German army mail forum ‘Marineprisenfunktelegramme’ – a term I never heard – are mentioned, unfortunately without further explanation.

  27. #27 Dan Girard
    6. September 2017

    Re #24:

    There’s a French site that has an image of the message in that auction:
    https://www.delcampe.net/fr/prestige/catalogs/show-4847?category_id=1121936
    It’s in the second row on the page, first item on the left.

    As to the “Schleswig-Holstein” message, there’s one other thing I was wondering about, in addition to the meaning of “MaPriFu.” In what circumstances would a navy permit an encrypted message to be sent from a warship to a private residence?

  28. #28 Jim Gillogly
    6. September 2017

    The second of those two messages is from a “dtsch krgsch koeln” which I assume is this: https://en.wikipedia.org/wiki/German_cruiser_K%C3%B6ln . If it were Privat, would “PrivatFunk” mean “Commercial radio”? Regarding Dan’s question – perhaps one could explain the system to the security officer and let him verify the contents. But I agree – it does seem odd. Standard practice for censors on the Allied side was to paraphrase any message that looked at all “off” and send that instead – Kahn has a cute example at the top of page 516. This wouldn’t be as easy with a cipher.

  29. #29 Gerd
    7. September 2017

    Vielleicht ist mit “Prisen” “Privat-Sendung” gemeint? Der Marineprisenfunk wäre dann ein Dienst mit dem private Mitteilungen der Besatzung übermittelt werden?

    Gerd

  30. #30 Thomas
    7. September 2017

    @Gerd
    Hört sich gut an. Das würde jedenfalls mehr Sinn machen als die “Prise” im Seerecht.

  31. #31 anderer Michael
    7. September 2017

    Maybe maprifu =MArine Personen RIchtFUnk
    I read a list of german naval ( or is marine the correct word?)abbrevations.
    P = Person
    RiFu = RIchtFUnk
    But I didn’t find the abbrevation maprifu

    The German prize law was effective on 3. september 1939.

    Dan
    “In what circumstances would a navy permit an encrypted message to be sent from a warship to a private residence?”
    Very good question. The cruiser Schleswig -Holstein was a school-ship ( training ship).
    Maybe the message was an exercise to train the cadets. They used an old cyper method and maybe the keyword is MAP (Ma Pri Fu).
    But I don’t have the experience and the skill to rate my theory.

  32. #32 anderer Michael
    7. September 2017

    Sorry a mistake
    The possible keywords , not MAP. I thought MPR or MPRF.

  33. #33 Alex
    8. September 2017

    Ist es auch möglich das auf solchen Wegen, gezielte sinnlose Wörterreihenfolgen per Post verschickt wurden, um die Alliierten zu verunsichern ?

  34. #34 Klaus Schmeh
    28. März 2019

    Bernd Kunz via email:
    Gern möchte ich Ihnen den Hinweis geben, dass es sich bei der gesuchten Person sicherlich um FKpt a.D. Gunther Dabis, * 30.11.1911 in Schorbus, + 6.6.1990 in Weinheim handelt.

  35. #35 Gustaaf Plancke
    Brügge, Belgien
    24. Juni 2022

    == bzgl Nachricht #23 :
    “RiFu” bedeutet hier warscheinlich “RichtFunk”, diese Abkürzung wurde auch von Siemens nach dem Kriege weiter verwendet
    (ich habe 21 Jahre bei Siemens gearbeitet im Bereich Wehrtechnik – Richtfunkgeräte und Verschlüsselungstechnik)
    = = bzgl Nachricht #28 :
    “dtsch krgsch koeln” bedeutet vielleicht deutsche Kriegsschule Köln” ?
    ihr werd es nicht glauben aber ich hatte gestern Nacht einen Traum und in diesem Traum redete man über die deutsche Kriegsschule Köln. Warum ? ? ? Das kann ja kein Zufall sein.