Kaliningrad-bar

Two years ago an encrypted bottle post was found in Kaliningrad, Russia. Can a reader break this cryptogram?

Click here for the complete top 50 list

Kalinigrad, formerly known as Königsberg, used to be the most Eastern German city, before it became a part of the Soviet Union after World War 2. There is a famous mystery connected to Kaliningrad/Königsberg: the Amber Room.

Last year I found an online article (here’s the Google Translate version of it) mentioning a second mystery related to Kaliningrad. This one is about an encrypted bottle post.

 

Kalinigrad’s second mystery

According to this article, in 2015 roadworkers found a brown glass bottle containing a sheet of paper. On the sheet an encrypted message is written. Here’s the front side:

Kaliningrad-Cryptogram

And this is the rear side:

Kaliningrad-Cryptogram-2

The bottle that contained the message appears to be a mass product from the Soviet era (i.e., after WW2). The paper and the writing style look post-war to me, too.

The encryption method used might be a simple letter substitution (MASC), though the many apostrophes are a little confusing. On the first page, one letter is underlined, another one is even double underlined.

It is completely unclear what the purpose of this bottle post was. Maybe, somebody looking for a place to hide important information thought putting it into a small bottle and digging it somewhere might be a good idea. I can also imagine two lovers who exchanged encrypted love letters this way. Or was the bottle used by a spy to place a message at a dead drop?

 

What my readers found out

I blogged about the Kaliningrad bottle post for the first time in September 2016. Some of my readers published a number of interesting comments.

As it turned out, the cryptogram was discussed in a Russian blog post, too. A reader of this post created the following frequency count:

Kalinigrad-Frequencies

Considering that the message was probably written in the Soviet era, Russian seems to be the most likely cleartext language. According to blog reader Thomas, the index of coincidence of the cryptogram is about 0,054, which is consitstent with Russian.

However, Russian is usually written in the cyrillic alphabet, while we deal here with the Latin alphabet. In addition, blog reader Thomas wrote: “The cryptogram has 37 characters: abcdefghiklmnorstuvwz, d’ f’ h’ l’ m’ n’ n” ö r’ s’ t’ ü z’, e with circonflex, e underlined and n underlined. That’s a bit too much for a simple substitution of the Russian alphabet. Furthermore there are too many double characters (uu, ii, ss, ff, ll, ee, t’t’, n’n’, dd).”

Here’s another interesting comment by Thomas: “If we assume that the apostrophe wasn’t used for creating different cipher characters (t -t’ etc.) we get the frequency sequence e, n, r, i, s. Since this is characteristic of the german language, this might be an indicator for a transposition of a german plaintext.”

I’m quite optimistic that this mystery can be solved. If you have any clues, please leave a comment.

Further reading: Who can decipher these encrypted consular messages?

Kommentare (54)

  1. #1 Jerry McCarthy
    England, Europa
    17. Oktober 2017

    The letters ď and ť are suggestive of Czech/Slovakian

  2. #2 Norbert
    17. Oktober 2017

    Abo

  3. #3 Hans
    17. Oktober 2017

    There are at least two letters underlined one time on the first page (“e”, l. 5 and “n”, l. 17), maybe even a third if one includes the p in line 9 that is underlined curly. The letter in line 13 that is underlined two times is an “e”, too.
    On the second page, there are also two underlined letters (“n”, l. 2 and 5).

  4. #4 Thomas
    17. Oktober 2017

    The frequency bar diagram of this text (e.g. here : http://practicalcryptography.com/cryptanalysis/text-characterisation/monogram-bigram-and-trigram-frequency-counts/) shows a stunning similarity to the frequency distribution of German. The only major difference: To few g and too much w.

  5. #5 Thomas Ernst
    Latrobe
    17. Oktober 2017

    If we include modified characters on their own merit, we are looking at a cipher alphabet of approximately 56 characters (appr. because I may have overlooked some):

    a | b | b. | c | c. | d | d. | d’ |e | e | e. | e. | ê | f | f’ | f. | f | f@ | g | h | h’ | i | k | l | l’ | l. | m | m’ | n | n’ | n”| n | n. | n. | o | ö | p | r | r’ | s | s. | s’ | t | t’ | t* | t. | t’. | u | ü | v | w | z | z. | z’ | ъ | ъ’

    • the “@” after “f” – “f@” – is a placeholder for the first sign on page 1. line 13. It occurs at the height of an apostrophe, but did not turn out that way in this post.
    • The wavy line occurs above, not below the “t”. Since it did not turn out in this post, I have replaced it with a *.
    • Russian “ъ” I used as a placeholder for the sign that is similar to a “z”, but looks more like a “2” turned upside down.

    Occurring doubles are: dd | ee | ff | hh | ii | ll | n’n’ | ss | t’t’ | uu | ww | ъъ

    The precise endstrokes of “v” and “w”, the bottom loop of the “z”, as well as the careful placement of the “apostrophe” at the proper height of “n” and “t” page 1, line 11 indicate deliberate care in writing these characters. This care goes for all the other letters as well, and also for the spaces. This note appears not to have been written in haste, and whatever the underlying language, we can expect a high degree of accuracy.

    The variants of the letters may indicate:
    • syllables, or
    • endings of a highly inflected language, such as Czech or Russian, or
    • diacritical signs, or
    • be partially phonetic, such as the “hard” and the “soft” sign after Russian consonants, or
    • perform all of the above duties at the same time.

    Russian has 12 hard and soft consonants, which in turn yield two alternate sets of “hard” and “soft” vowels à 5 each. In this case, the “apostrophe”, which occurs after 11 letters, and inside words (assuming the divisions between letter groups indeed mean word separations) may indicate the hard or the soft sign, or alternate vowels. It is also possible – if we stick with Russian for the moment – that there exist homophones for either the hard or soft letters. And then are twelve double letters. Perhaps they are not doubles, but hard and soft consonants, while the real doubles are indicated by other signs. Of equal importance is the fact that certain modifiers, such as the circonflexe, the double apostrophe, the double underline, and the wavy line, occur only with some characters, but not with others.

    That the writer uses diacritics from French (circonflexe, tréma), or from both French (circonflexe) and German (Umlaut) implies his familiarity with either one of both languages. The three prominent languages taught at public schools in the USSR after WWII were English, French, German. (http://www.doiserbia.nb.rs/img/doi/0579-6431/2015/0579-64311502305I.pdf). The curriculum underwent some changes, especially during the Cold War; Spanish was added later as a fourth foreign language. I do not think that German is the plaintext language. However, my suggestion of Russian is just a point of departure. In order to solve this, one will have to match a language to the “special” signs as well as the doubled letters.

    According to the Russian article in the Kaliningrad paper that (first?) reported on this, the bottle was scooped up during street excavation work for a gas line in Baltiysk (Pillau till 1945) on “Lenin Street”, houses 64-66. Apparently, the original bottle is no longer; the text itself may be incomplete. So we only have the workers’ impression that the bottle was Soviet vintage, lemonade or beer (did they really look that similar? Did children often confuse one with the other??). The article made special mention of the fact that real cork, and wax was used to seal the bottle. To draw conclusions from this about the social status of the young person who wrote this – and I am sure it was a young person – must be left to the imagination.

    Returning briefly to the cipher alphabet: and what about those poor waifs a, g, k, p, w (I may have overlooked a sign) that receive no modifiers at all? This is of equal importance.

  6. #6 Thomas Ernst
    Latrobe
    17. Oktober 2017

    Regarding the Cyrillic/Latin question: the extreme care with which the writer wrote the Latin characters – see my note above – may indicate that he transposed Cyrillic characters into Latin characters, which he had learned via a second language. Any first time student of Russian or Greek will remember having attempted, at some point, to do the same thing in reverse: write his name, or a longer German, French, English, Latin passage in Cyrillic letters. Not by translation – just by transposition.

  7. #7 Hans
    18. Oktober 2017

    Just a small correction: the second page is in fact a separate sheet, not the rear side of the front page. The Russian article speaks of two sheets of paper having been found and if one looks closely at the left, irregular edge of the torn out sheets, one sees that they don’t match each other. The blue writing imprint seems to result from the sheets being closely stuck together for decades, not from the backside of the paper.

  8. #8 Thomas Ernst
    Latrobe
    18. Oktober 2017

    @ Hans: I don’t think anyone suggested that the two pages constitute one leaf. It is obvious they don’t.

  9. #9 Hans
    18. Oktober 2017

    @ Thomas Ernst: I didn’t mean that you suggested this, but if you look at the article above, it speaks of one sheet with a front and rear side. Like you said, this is obviously not the case. I was only confused by this at first.

  10. #10 Thomas Ernst
    Pittsburgh
    18. Oktober 2017

    Addendum: Hans’ remark about the two sheets of paper points out encore une certaine largesse: cork, wax, one-sided writing. I remember letters from relatives in the GDR who wrote on every last line on both sides of one sheet of paper. Those were the days of the unforgettable Mitropa coffee …

  11. #11 Thomas
    18. Oktober 2017

    The very small lower case “k” might indicate that the writer had learnt Cyrillic handwriting first. Moreover, an unusual “r”. But I doubt the message is a transcription of a text written in Cyrillic.

  12. #12 Norbert
    18. Oktober 2017

    The underlined letters mentioned by Hans occur (more or less) periodically, which suggests that they divide the text into sections of (more or less) even size. Assuming that an underlined letters marks the end of such a section, I get the following sizes:
    166, 169, 162, 169, 169, 144, 5
    (not counting spaces, punctuation and apostrophes/diacritics).

    This is quite regular, apart from the 144-letter section and the last five letters “eimat” (or “eimut”). But if we add the latter plus the ten dots (maybe padding characters?) to the 144-letter section, we get 159 letters which again is a size quite similar to the preceding ones.

    Maybe it’s a block-wise transposition, i. e. a transposition applied to each section separately?

  13. #13 Thomas
    18. Oktober 2017

    Maybe 13×13 complete and incomplete columnar transpositions? Don’t know how to test that.

  14. #14 Thomas Ernst
    Latrobe
    18. Oktober 2017

    Norbert’s “Rosetta” approach is intriguing, especially with regard to the final “………” after “eimat”. The dots appear to imply “and so on”. Which, in turn, would identify “eimat” not as belonging to the previous block, but as the beginning of the next one, which the writer left unfinished – apparently on purpose.

  15. #15 Thomas Ernst
    Latrobe
    18. Oktober 2017

    I can’t help but being curious about the print-through on the second page. It appears to be slanted to the right, which would imply that there was at least one other page between the two survivors. This missing page appears to have had one and a half blank lines in the middle. The second section starts in the right top corner – where you might expect date. So – perhaps a diary?

  16. #16 Hans
    18. Oktober 2017

    I think that the pattern of the print-throughs on the very right of the second sheet suits the first sheet very well, especially the length of the line 11 in comparision with the following lines. For example at the ending of line 13 or the last line, one can even make out the same spaces between the words as one sheet one. Maybe the missing imprint in the middle could have another reason, probably resulting from a special way the sheets were stuck together?

  17. #17 Dampier
    18. Oktober 2017

    abo :]

  18. #18 Thomas Ernst
    Latrobe
    18. Oktober 2017

    My list of characters is incomplete: aside from the placeholder-only (!) “ъ”, there also is a true reversed “s” in the text, the Cyrillic “ƨ” (first line, third word, fourth letter, and the penultimate letter of the first line, and more). “ƨ” also shows with an apostrophe, and is one of the few singletons in the text, alongside “e” and “i”.

  19. #19 Thomas Ernst
    Latrobe
    18. Oktober 2017

    @ Hans: indeed, I missed the matching line-endings! Especially with lines 11 – 14, the second sheet convincingly reflects the line endings of the first sheet. Thus away with middle page and date and diary … However, I am guessing there was text before the first page. But since that can’t be proven one way or another, it is an irrelevant hypothesis.

  20. #20 Thomas Ernst
    Latrobe
    18. Oktober 2017

    And now for some wiki-wisdom about Ukrainian: “an apostrophe is used to indicate the hardness of the sound in the cases when normally the vowel would change the consonant to soft; in other words, it functions like the yer in the Russian alphabet.” (The “yer” – pure coincidence – is my place-holder “ъ”.)

  21. #21 Thomas Ernst
    Latrobe
    19. Oktober 2017

    More Russian: there are 12 cipher values with an apostrophe: d’, f’, h’, l’, m’, n’, r’, s’, t’, z’ ъ’, ƨ’. Hopefully I didn’t overlook one. And one “n” has two apostrophes, I am aware of that. There are 12 Russian hard (and soft) consonants: б’, в’, д’, з’, л’, м’, н’, п’, р’, с’, т’, ф’. – I think a native Russian speaker should give this cipher a shot. The letters will not necessarily be in order – they could be intertwined, backwards, who knows what. But trying out these 12 consonants should go a long way in telling us sth. about this text. Either: yes, it is transcribed Cyrillic, or: no, it can’t be possible, because of … . Only because the Russians who originally published the article couldn’t do anything with the cipher – and I think they actually did NOT exclude Russian in their quick and long list of impossible languages – doesn’t mean they looked at it properly.

  22. #22 Alex Ulyanenkov
    19. Oktober 2017

    The language is Deutch.
    n’ = ne, s’=se, t’=te, ê = ue
    all capital letters – you should exclude
    all underlined leters ( like n. eu. on the second page are not redable) sign “.” after the undelined letters means that you should continue with next word….

    So I am not a big specialist in Deutch, but last page of the message contains the following words:
    “s’e” = see
    “n’eêain. (“n” is undelined) Albot’e” = “neue ailbote”
    Last 3 words of the page:
    “iwêrw wurteheu. (eu – undelined) eimat” = weurde (I think so, but not sure)
    wurde heimat…. (I 100% sure in that last 2 words)

  23. #23 Alex Ulyanenkov
    19. Oktober 2017

    More…
    I am talking about last page …

    ” öfenuinrwe wen wertss cidi kereelrn”. sn’ehd’en i
    aind’ n’aêain. Albot’e”
    ~=~
    “öffnung(?) wer(der(?)) wenn weg die (?) ss cd (?) kriegen senden … ein neue ailbote”…

  24. #24 Thomas
    19. Oktober 2017

    There may be an underlying German (I suppose you mean German and not Dutch) plaintext, the letter frequency seems to point in that direction. But this message is definitely no German plaintext, not even a misspelled one. Yes, “eimat” may be a part of “Heimat”, but that’s all.
    As to the handwriting: Do you think it is typical of someone who had learnt the Cyrillic script first?

  25. #25 Rallinger
    19. Oktober 2017

    Interestingly enough, the 2 sheets seem to origin from a ripped-apart larger sheet. This is the first one mirrored, matching the second one perfectly: https://imgur.com/y3g3Wld

  26. #26 Thomas
    19. Oktober 2017

    Thus the two images show only the right half of the front and the right half of the back of the sheet?

  27. #27 Hans
    19. Oktober 2017

    I think the two sheets are torn out from something like an exercise book or notebook, as only the left edges are irregular. Furthermore, the sheets are lined horizontally as you would expect from an exercise book.

  28. #28 Thomas Ernst
    Pittsburgh
    22. Oktober 2017

    Polyalphabeta uacantibus nullis claueque incognita. Some kid learning Bellaso or Vigenère. Ain’t easy. Will share either solution or surrender. The latter more likely.

  29. #29 Thomas Ernst
    Latrobe
    22. Oktober 2017

    Should like to add: I’ve tried to post my proof’d transcript of the ciphered text several times, but – as Klaus explained to me – a filter of sorts that surmises “Nazi symbols”, such as “ss”, wouldn’t let me. Amusing, I know. – Thus I couldn’t post my transcript. Suffice it to say that the “kid” enciphers the same passage 7 times. The first two attempts, counting bells and whistles, such as apostrophes and periods, ring in at 194 each. Then it was a few less, then a few more. In his final and seventh attempt, the “kid” became bored, and fizzled out after five characters. Lemonade and beer: he must have had the latter! – The all-important apostrophe – and DO note that only line I/3 doesn’t have any at all – is not a phonetic sign of articulation, but indicates a letter: either doubling, or following the previous one. Right now, it looks like the apostrophe is a double-up. My Russian “ƨ” does not occur in this text, it is just a variant of “ъ”. However, the writer of this cipher grew up with the Cyrillic alphabet, as can easily be derived from his writing of “f”, “g”, “v”, “w”, and “z”. Which does not mean that the plaintext is Cyrillic. As a matter of fact – based on my symbol count – it can’t be. The writer is enciphering the same Latin, English, or even German passage 7 times. French certainly is out because of its 13 accented letters. German can be reduced to 26 by writing the “additional” characters as “ae, “oe”, “ue’, “ss”. However, I am not so sure it is German. Russian, at any rate, because of its 33 characters, is out as the plaintext language. – The language is the same for all seven sections, including the fiver at the end. The Latin Pater Noster, up to “sicut et nos dimittimus debitoribus nostris”, is 196 characters without ligatures. But perhaps it’s the incipit Cæsarii De Bello Gallico, who knows. It could be Shakespeare – anything that struck this youngster at the time, and struck him or her so deeply that he/she attempted to embottle it for eternity. – Anyone quick with the proper algorithm on his/her computer should be able to line up sections 1 and 2 in a twiddle, and thus derive the polyalphabetic consequences.

  30. #30 Jerry McCarthy
    England, Europa
    22. Oktober 2017

    #29. @Herr Ernst. I have had similar problems in the past. Our solution was for me to mail the text to Klaus, who could then post the mail’s contents.

  31. #31 Thomas Ernst
    Latrobe
    22. Oktober 2017

    Here are sections 1 and 2 of “Time in a Russian Bottle”. I have replaced the Russian hard sign – that evil little Nazi sympathizer – with an “x”. I leave it to the reader to transpose all apostrophes into the preceding letter. Don’t yet know what to do with the periods. They may function like the apostrophes, and indicate the preceding or following letter, i. e. “m.” = “mn”, or “mm”. Thankfully, there aren’t to many. Notice that with the use of several polyalphabets you are as likely to get “mirrour substitutions” as you are not. Here are the endings of sections 1 and 2:

    1) […] dlu dnh xoaul lel nene ttse.
    2) […] fwx dih szêus xsx ffua ffnp.

    The groupings are arbitrary, and only serve to illustrate above argument.

    Sequuntur sections 1 and 2:

    en’ifvn d’t’öhn’fdê elhxikixacel etê deluwxs
    uxed ganunt’ lnenên’lf x.s.f.d. fxuustx d’e astf’
    uesx cfefdxx helcnunde uwsesd danunhhs c.f.
    d’enac s’eaxa efn’ xuhncfaiw m’am’mt. f.t.’f. wmet’t’cef
    dludn hxo aul lel nenet’se. [194] emnh mwie eiie iael’es
    esmdesxnt xnoxunt’ ex xelfd’ x.l.b. scihss ne in’nw htouk
    sans eighcin möwidgn lwixe öxbns’n hxan’ê uwesgnen’n
    ei sêk dloex’n’ua zn’el’et’ff s.n.c. lnee exadn’ wue
    êfht’iesn’n’f wxdihs zêusxs xf’uaf’np. [194]

    Assuming that the writer stopped with a complete word in section 7, here are all seven variants of it: 1) ennif, 2) emnhm, 3) ihxdö, 4) lelhu, 5) lbawo, 6) A[???]llot, 7) eimat. That word alone is worthy of consideration: three versions beginning with “e”, two variants with “h” as fourth letter, “i” and “h” removed just one step from another in variants 1 and 2.

    It occurred to me that you wouldn’t necessarily consider the Pater noster or De bello gallico ripe for eternity. Perhaps there does hover a whiff of long withered rose petals over these encipherings. Perhaps it was a letter, which might have begun with “Liebe […]”, or a name, “Maria”, or perhaps “Shall [I compare thee to a summer’s day?”].” To end on Will: “There are more things in heaven and earth […] / than are dreamt of in your philosophy.”

  32. #32 Thomas Ernst
    Pittsburgh
    22. Oktober 2017

    Since my previous post took, here is a full transcript. The added character counts in “[…]” (done, of course, without my added line numberings!) set off all seven sections from eachother. If we consider this to be the same text seven times over, the count may be indicative of final word length: a word of 5 letters omitted in section 3, a word of seven letters added in sections 5 and 6. And the first word containing 5 characters – and an “interior” word ending on character 166, for whatever that’s worth … The raised little hickey on the first “f”, line I/13, looks like an encercled “2” to me, and may indicate two apostrophes, which in turn would raise the character count of that section to 188. – The first character of section 6 looks like a capital “A”. Both characters/modifiers occur only once. Any mistakes in my transcript should be easy to correct.

    [I/1] en’ifvn d’t’öhn’fdê elhxikixacel etê deluwxs
    [2] uxed ganunt’ lnenên’lf x.s.f.d. fxuustx d’e astf’
    [3] uesx cfefdxx helcnunde uwsesd danunhhs c.f.
    [4] d’enac s’eaxa efn’ xuhncfaiw m’am’mt. f.t.’f. wmet’t’cef
    [5] dludn hxo aul lel nenet’se. [194] emnh mwie eiie iael’es
    [6] esmdesxnt xnoxunt’ ex xelfd’ x.l.b. scihss ne in’nw htouk
    [7] sans eighcin möwidgn lwixe öxbns’n hxan’ê uwesgnen’n
    [8] ei sêk dloex’n’ua zn’el’et’ff s.n.c. lnee exadn’ wue
    [9] êfht’iesn’n’f wxdihs zêusxs xf’uaf’np. [194] ihx död’e sxunxca
    [10] e wehsnhdn’ mne fu oenin’f iftt’ kt’t’set’ wx’sfue ede ê
    [11] ue ex’xn dde eshse i an’ex’e nunw tmnisen’t’ miabdhhncs
    [12] x.l.n. fneezt’ xande hehxt’wx he ehfs mid’a ue inlg
    [13] f2t’eaveidn’ t’oa iuegnnz’w ngwsaetme. [187] lel hul xetn’h’
    [14] dien’ i nelct’ eweas hlen’sx’ wiê ê eisx’e eomn’f
    [15] glniftha usxualhseen ed’ixt’n’l’ t’nifn’tn’ smwallssn d.s.z.-
    [16] d’xepc x’l’iwt’ sxehui i iufiee ihnxl n’ecnn’ hiuen ênfi
    [17] sengnet’snx’ hgênww hn’ein’ üdna lgs’ eamdemn. [199] lba wof’u
    [18] i mud’n x’nuaht’ sêas’gnn bzexeenn f’exx di eg tnein
    [19] exdsz’i isx’sn’n n’hee xêt iöe dhi ewgwt’mk n.t.d. mze
    [20] i eix’ fnö x’eseeon’ fexisen’lcx’t’ hxhu l’nhu ewt’
    [II/1] öfenuinxwe wen wextss cidi kexeelxn.” sn’ehd’en i
    [2] aind’ n’eêain. [199] Allot’l f’laefdln’ m’n’eev s’e eed-‘a
    [3] tekezwn nie eenn’mn ndew giunn’ xfxess i sx’ee ide
    [4] aif–zd’ln umfdxf n’.xn. x’en’ iuxent wehdmnnu esl’es
    [5] –‘fef nd’e wn’êd dxesmxde gon’xt’ iwêxw wxixeheu. [166]
    [6] eimat . . . . . . . . . . [5]

  33. #33 Thomas Ernst
    Pittsburgh
    22. Oktober 2017

    Continuing my sentimental journey for a moment: sections 1, 2 and 4, 5 are five letters apart. Let’s assume that the unique double underling at the end of section 3 stands for a doubled letter, and the “f2” means two apostrophes: then we have 189 characters for section 3, i. e. the “complete” sections 1 – 5 all vary by one or two five letter words – the same amount of letters as the first word seems to contain. Syntactically, how can you break off on five letters less (sections 1 and 2), or, respectively, add five letters (sections 4, 5) unless they contain the invocation of a name – the same name as at the beginning of the text, let’s say “Maria […] Maria, Maria” (punctuation not enciphered). A copied text (perhaps a song, or a poem), or a text by the writer’s invention. – Cryptologically, this speculation does not get us anywhere. However, it is a pleasant thought.

  34. #34 Thomas Ernst
    Latrobe
    22. Oktober 2017

    Continuing my sentimental journey for a moment: if we consider the unique final double underlining in section 3 as indicative of another letter, and consider “f2″ = f”, then section 3 is 189 characters long. That means that all five “complete” sections are one or two five-letter words apart. Assuming the passage is the same in sections 1 through 5, you could syntactically augment or diminish the text only if the last word is a refrain, or a name – such as “Daisy”, for example. If we take the only five characters of section 7 as one word, the text could start and end on “Daisy […] Daisy, Daisy”, punctuation not enciphered. You pick your own name: Fiona, Gerda, Hazel, Irene, Maria …. One of my cats I called Daisy. She is an orange tabby. Now she has been immortalized in this blog! – This does not help cryptologically, but it is a pleasant thought.

  35. #35 Thomas Ernst
    Latrobe
    22. Oktober 2017

    My screen froze at comment 33; thought it hadn’t made it. Rewrote from memory what is now comment 34. Pick the one you like better.

  36. #36 Thomas Ernst
    Latrobe
    23. Oktober 2017

    My gut instinct tells me this is an autokey cipher, à la Cardano, the autokey shifting in each variant. It is easy to learn and apply, but quite difficult to crack if you have nothing to go by than five letters that you don’t even know. Bonne chance à tous!

  37. #37 Alex Ulyanenkov
    Moscow
    24. Oktober 2017

    As a specialist in image analysis I am collecting as much as possible info re investigated sample:

    – Where, when and how it was found
    – What is the matter of the sample
    – Artifacts of the sample and so on

    In a case if sample is historical artifact – we need to know or reconstruct the history around it.

    If it contains handwriting – we should analyze the handwriting features…

    So, let’s call that process as “sample pre-understanding”.

    Now let’s look to the sample and to the message we are talking about and have a look how method works.

    1. The brown glass bottle with message was found in Baltiysk (Kaliningrad area of Russia) in the region of Lenin street 64-66.
    2. WWII time it was Pillau – Kriegsmarine (as well as U-Boots) base. It was very well prepared for defense. The region of modern Lenin street 64-66 (where bottle was found) corresponds with former Pillau living area called “Alt Pillau” (http://forum.axishistory.com/viewtopic.php?t=164250&start=15).
    3. The brown glass bottle was sealed by cork and wax.
    4. The message was rolled into a tube.
    5. The tube in the middle was covered by band of Al-foil and the foil was fixed by Cu-wire.
    6. The brown glass bottle was broken during excavation works – so, we lost some important evidence which can give us more info re timing.

    Let’s now look to the paper and the message.
    1. Paper is typical for notebooks for wide period of time (1930 – 1960)
    2. Message is written by blue inks – the fountain pen was used (most likely like on the image here – https://dl.backbook.me/full/740c4887bd.jpg) .
    3. Handwriting is very stable, smooth, clear, calm – means the message was written in comfortable and good illumination conditions (enough light), probably at home. The accurate sealing by cork and wax – can also confirm that.
    4. The handwriting is typical for young people 10 – 13 years old, most probably for boy.
    5. The message text contains a lot of specific signs like letters with apostrophes and dots. Tipical for non-Slavonic languages.

    Just a remark – the table with results of frequency analysis in this article contains mistakes because of incorrectly identified symbols…

    What we know about Pillau in WWII and transaction period?
    1. It was hardly destroyed in 1945.
    2. On 20th of January 1945 in accordance the order the evacuation of civil Germans from East Prussia was started.
    3. Before Soviet army cleaned East Prussia from Nazis that strategic territory were very well patrolled by German police and special troops. There are lots of reports in open sources re Soviet reconnaissances which were lost on the territory.

    That means – no unknown strangers (especially Russians, Polish, French people etc) were available on the territory at least for free moving.

    Especially for highly guarded Pillau, which was a last point of defense and which fell on 25th of April 1945 only.

    Starting from May 1945 it was a big immigration flow to East Prussia – from Belorussia, Pskov area, Kalinin (Tver) area , Yaroslavl area and Moscow area to restore the industry of East Prussia. Mostly workers and agriculture people.

    In the period from 22nd of October 1947 till 21st of October 1948 all German East Prussia citizens was moved to the territory of former GDR.

    SO. First conclusion:
    The message was left in the period from 20th of January 1945 till 21st of October 1948.

    Next…

    The Al-foil is one of important signs which tells us that the author was not Russian one because Al-foil was very unusual item for Russian Immigrants .

    The knowledge (especially on the good level) German (enemy language) for immigrants was limited by some common words…

    My opinion – the message was left by German boy 10 – 13 years old.

    Most probably there was a group of friends from Alt Pillau, classmates or/and neighbors with similar age.

    The last group of 10 dots in the message – most probably a symbol of the group or (less probability) signature of the author.

    So – message written in German (Deutch) by using more/less simple coding.

    The message should consist some key words, corresponding to the person it wrote, place and time it was written:
    – Sea(See)
    – Me (Ich)
    – War (Krieg)
    – etc…

    The message probably (as we are talking about child) may contain some grammar errors.

  38. #38 Alex Ulyanenkov
    Moscow
    24. Oktober 2017

    The whole message contains only one capital letter “A” on the last page.

    As I wrote earlier at least some letters can be identified as: n’ = ne, s’=se, t’=te, ê = ue

    “s’e” = see

    We can see probably rule – if the letter is underlined and has a dot in the end – we should skip that letter/letters and skip the space with next word). So, again,

    “n’eêain. (“n” is undelined and with dot in the end ) Albot’l” = “neue ailbote” (mean -a new courier.) The word “Albot’l” has a part “botl” (sounds as bottle?).
    Looks like the bottle called as “a new courier” used because of some reasons) to deliver the message.

    If we will apply the rule above to the last 2 words on the last page:
    “ wurteheu. (eu – undelined and with dot in the end) eimat” we have “wurte heimat”. May be we see a mistake in word “wurde” . So we shoild receive “wurde Heimat”……….

  39. #39 Alex Ulyanenkov
    Moscow
    26. Oktober 2017

    As an idea – what about Plautdietsch? That German dialect was language of East Prussian area.

  40. #40 Alex Ulyanenkov
    Moscow
    26. Oktober 2017

    Here is a link to online Woerterbuch – http://www.plaut-dietsch.de/pages/plathoch.html

  41. #41 Thomas Ernst
    Pittsburgh
    28. Oktober 2017

    From Russia and roundabouts with love: to all those who still search – cease, I found the answer. The text is in Cyrillic (or a transliteration thereof), and it does speak of passion. Not for Daisy, Frida, Gerda, Maria. It speaks of a forbidden passion, to be entombed in a bottle, to be resucistated at a proper time. This passion extended to a prophecy – to come true later on. – The text dates back to the 19th Century. It is – originally – in Cyrillic. However, no Pushkin, Lermontov, Gogol, Dostojevsy, Turgenjev, Tolstoi need apply. If I tell you the solution, there will just be a general “oh, that’s what it is.” Both Norbert and Thomas have helped, indirectly. Norbert with his letter count, Thomas with his instinct that the apostrophes are not absolute, but relative. Indeed, they are relative: “n'” = “np”. That’s the only obvious one; the writer used “np” twice: once to signal the end of section 2 more distinctively: “np” instead of “n’.”, or “n..”: notice how the underlining continues underneath the period. A period underlined means a grammatical period, not a cryptological one. Since “p” can be Latin or Cyrillic (=r), it is difficult to ascertain the alphabetical distance of the other apostrophes from their precedents. However, within the cipher alphabet (!), apostrophied f’ or t’ etc. stand for the same symbols. Not, however, in the plaintext, which – as it appears so far – changes on every letter. – Since I am juggling between Latin, Cyrillic, and transliterated Cyrillic, the workings of this cipher will take me a while to demonstrate in a convincing fashion. – It reminds me of Heidel, 25 years ago: his plaintext was clear: but how he had enciphered it, occupied me for days. – As with Ferdy III, I shall publish full solution of the enciphering process & enciphered passage & background – once I’ve understood it myself – here first.

  42. #42 Hans
    28. Oktober 2017

    @Thomas Ernst: Congratulations, a very impressive achievement! I was confident that you would eventually find the solution. Now I’m really looking forward to getting to know the plaintext and to comprehend how you could figure it out!

  43. #43 Thomas
    28. Oktober 2017

    @Thomas Ernst
    I am champing at the bit – is it German?

  44. #44 Thomas Ernst
    Latrobe
    28. Oktober 2017

    @ Hans, Thomas: the text is in Cyrillic. Why the writer used one Cyrillic character at all – the one alike to the Russian hard sign (which I had to transcribe as an x) I can only surmise. I don’t think it was in order to confuse. This one non-Latin sign was either dictated by substitution need, or – if it indeed represent the Russian hard sign, for which I take it, closed loop or not – the writer left it in plain sight to counter the Ukrainian hard sign, the apostrophe. Which, given the plaintext, would make sense on more than one level. The fact that the writer moves so adeptly and effortlessly between parallel phonetic and graphic universes in enciphering a political text makes such cryptological playfulness quite possible. – Regarding deciphering: the Cyrillic alphabets are difficult for non-natives in that they changed historically (pre-/post-revolutionary Russian, for example), and are different geographically (Ukrainian/Russian) and phonetically (articulation, iotation). Given this to be a “mid-20th Century text”, quite a few things are possible … – I chanced upon the plaintext by one wild guess, after my thinking had turned from sweet romance to phonetics and politics. The text I found yielded almost – differing by only one character in section 1, depending on which version you use – the same character counts for sections 1/2 and 4/5. Call it serendipity. Add the fact that the text contains two different, self-contained sections, of which one is 192 characters long, the other193. Thus sections 1 and 2 may be identical, or not. Should I not get the cryptological details done, I’ll post the plaintext here, and let someone else do the clean-up. – The concept of having the plaintext-value of the apostrophe – or should we rather call it the Ukrainian hard sign? – be determined by its predecessor is admirable! Which, in turn, let me guess at an autokey. Of which I am not sure any longer. However, whether you take Bellaso or Cardano or Vigenère as inspiration, the cipher alphabet changes on every letter. Now comes the time-consuming part: writing the text out, extending each cipher letter throughout the various Cyrillic alphabets, and coming up with – first – a numerical key-sequence, then, with some luck, a text-sequence that makes sense. I remember from the Bellaso thread that someone – I forgot who, apologies! – had solved yet another one of Bellaso’s “exercises” perfectly, plus substitution alphabet-halvings, but couldn’t come up with the key. That’s usually the difficult part (see Heidel). – The plaintext of the Baltiysk-cipher, as mentioned, is political in nature, and I can’t help but thinking that the person who wrote it enjoyed certain societal privileges, and may, at some point, have worked for the Soviet “black chamber”. Why else would you deal with cryptology and several alphabets and probably languages, and be so good at it? This guy – or gal – was no Ferdinand who couldn’t write in cipher when he had a headache. He/she was a true “pro”, perhaps at the beginning of his/her career. With some luck – but that may be too much to hope for – he used his/her name as key. By now, I would like to know that name.

  45. #45 Thomas
    28. Oktober 2017

    I think the character that you take for the Cyrillic hard sign (ъ) and transliterated as x is the (Latin) “r”, otherwise there were no r’s in this text. Furthermore the frequency of this character matches that of the r in a (German) plaintext.

  46. #46 Thomas Ernst
    Latrobe
    29. Oktober 2017

    The Latin “r” is one of the hidden letters, probably one of the apostrophes following either “h” or “t”. Regarding hidden letters: that the “p” makes two appearances at all has two reasons. The first time the writer used it to conclude a section of his text (section 2). Though otherwise he/she did not encipher punctuation marks, he/she set off longer sections of his text with a period. The reason for that is inherent in the plaintext. These grammatical, not cryptologic periods the writer underlined. At the end of section 2 he could have written n.’, with period and apostrophe underlined. For reasons of neatness – that is my best guess – he avoided the ugly clash of a cryptologic with a syntactic punctuation mark, and closed on the intended letter itself – “p”. The second “p”, line 16, I consider a slip on the writer’s part. Besides “r”, I expect to find the remainder of the Latin/Greek alphabet in the apostrophes, such as “j”, “q”, “x”, “y”, and then Cyrillic signs. – Regarding frequency count: with polyalphabets, they are impossible. And triplets like “eee”, or sequential doubles, like “llss”, are indicative – albeit in themselves no proof – of polyalphabets. I should also like to add that the writer used false word separations. – Regarding my initial reliance on character count, which may appear simplistic: there are pre- and post-revolutionary versions of the same text (f. e. more soft signs in the earlier Russian), as well as textual variants, such as switches between Russian “and” and “neither”, which makes for either two or four characters. In other words: I did not just look for exactly matching numbers. However, once I found the proper proportions with the proper text, everything made sense. If there are variances of, let’s say, 3 or 4 characters, I will try to explain. As I will why section 6 only has 165 characters (or maybe 162 or 166), and section 7 only five (no room for error there). The first plaintext letter of which is Д.

  47. #47 Thomas Ernst
    Latrobe
    29. Oktober 2017

    quick correction of above: “[…] at the end of section 2 he could have written n’., with apostrophe and period underlined.”

  48. #48 Thomas Ernst
    Pittsburgh
    30. Oktober 2017

    Am working on matches. Not the ones you strike to light fire. Though they would be helpful, too. – Matches between older and newer Cyrillic alphabets, between romanized forms of Cyrillic alphabets, between letter distances of any of the alphabets. Meanwhile, am slightly assured that the key sequence to the polyalphabets will yield personal information. Which I had been hoping for. – This will either be a big cryptological fire-cracker – or a low-burning torch, dying on its embers, to bury me alongside Hensel and Kortum, chalked mass grave, wherever that may be …

  49. #49 Alex Ulyanenkov
    Moscow
    30. Oktober 2017

    @Thomas Ernst: In general it is possible to use the similarity of writing of some letters in Russian and Latin for mesage coding. Especially in hanwriting. For instance: Latin “n” can be identified as Russian “п” (sounds like Latin “p”), Latin “r” ~ Russian “г” (sounds like Latin “g” in the word “get”), Latin “m” ~ to Russian handwritten “т” (“t”), handwritten “b” ~ “в” (sounds like “v” in the word “victory”), “y” ~ “у” (Latin “u”, but sounds like double “o” in the word “book”), “u” ~ “и” (“i” or “e” with prononciation like “ea” in the word “read”). The apostrophe ” ‘ ” was used in some cases to replace the Russian “hard sign” (“ъ”) – unpronounceable letter, used to show the hardness of previous consonant letter. May be it helps for your therory. But the handwriting of the message is identified as young hand, definetely done not by Russian or other Slavic language.

  50. #50 Thomas Ernst
    Latrobe
    31. Oktober 2017

    @ all who are still following: forget the mass grave. It’ll be a firecracker. Ferdy’s spirit is with me … I have the cipher text. I have the plaintext. Although the writer used false word separations, he always finished a line on a complete word. That explains the short and long line endings. And makes my job a little easier as to variant readings. Now I have to find the polyalphabetic substitution sequence, then the key that determined their sequence. The latter, as mentioned before, can be quite difficult.

  51. #51 Thomas Ernst
    Latrobe
    31. Oktober 2017

    Since it’s nice to associate dates with decipherings, let me just add: as of this Halloween, October 31, 2017, consider the contents of the cipher solved. And that ain’t no trick!

  52. #52 Thomas Ernst
    Latrobe
    7. November 2017

    Am feeling quite sympathetic towards anyone awaiting a quick answer on this one. As mentioned before, my solution sticks like butter to a teflon pan. HOWEVER: the encipherer did not only use false word separations. He did the same with the separation of text segments. Yes, the underlined “endings” are endings of words: yet not of a complete text section. They are run-ons. And the writer made a mistake in line 7. – I could just post the whole cipher-/plaintext without explanation. – However, that’s uncouth! – Am trying to get into the polyalphabetic swing behind it. And that is soooo d…d difficult given that this person enciphered across languages, i. e romanizations of Cyrillic. – It was – by luck – an easy cipher to solve: but becoming close to impossible to explain. – Once I am done, I should consider this cipher the most unique I’ve ever come across. And it shall deserve unique placement in the history of cryptography! Compared to this, Ferdy III was someone living in a bus down by the river …

  53. #53 Harald
    vienna
    13. November 2017

    @thomas great work! i am very excited to read the cleartext and the solution! thx for your work!!

  54. #54 Alex Ulyanenkov
    Moscow
    16. November 2017

    @Thomas Ernst – could you provide the final text?