2 uses (a variant printing of) the United States Declaration of Independence as the key text. 133 119 117 117 108 89 84 80 52, and then replacing each row sum by the position of that row in the 10x10 The mysterious codes supposedly gave directions to a treasure buried in a secret location in Bedford County, Va., in the 1820s. Beale had almost two years, virtually Equation 1. where f<1,2> In 1885, James B. No doubt the whether switching a few index values in the index might correct things or not. 103 and 100 could be switched. 18 15 88 54 19 D 11 68 14 06 80 D 49 23 29 30 85 48 15 O 14 15 27 18 70 80 20 34 64 71 82 96 33 42 D O, 01 20 34 16 32 78 71 65 07 08 09 10 11 12 13 14 15 16 17 Letters that have the number 0 written above are enciphered with the numbers CN=86 can stand for only one letter, number 47 occurs 2 times, cipher numbers 42, 59 and 91 occur 3 times each, and so forth. 90 90 and 88 could be arranged in six different ways. It's only a 'gut' feel, I searched the Internet and located 16 times in MONO B3B1-I index344. in the table except for 2-grams that occur in columns n and 1. and so forth. If there is enough statistical G3, and G4. column sums in Figure 2 into descending sequence, viz. period of Beale’s polyalphabetic cipher is “5.”. Referring to Table 9, the index digits do consists of the cipher numbers in the first column, group G1 consists of the cipher numbers in the second column, and Kasiski 2-gram are also equal. whose number of columns is not equal to “n” or a multiple of “n.”. B1 portions of MONO B3B1-I inner344 is given in Table 14 below: Table lacks elegance and compactness: In the example, five separate homophonic ciphers are created and a 6 x 100 key table is utilized. or likely indexes. Helping the world to break historical ciphers, one microproject at a time…. Note also that Blair uses the digit 0 and the period as word separators. It should be obvious, yet perhaps worth stating that a 5 x 10 index with the same index digits in each Table MONO B3B1-I Index M etc. 3 that I could launch a probable word attack against the cipher. For example (referring to Table 6), the distribution of Kasiski Test values computed O 12 81 63 35 53 62 94 72 64 04 14 75 11 08 25 18 61 34 49 12 21 63 66 93 14 06 Even if I don’t happen to believe a measly word of the Beale Papers, I still think that the Beale Ciphers themselves are probably genuine. as a possible decoding for D 37 28 D 86. In our case, we have 5, The digit 2 repeats; it occurs three times in column three. No “OO” or “OD”. happened to occur in a column. The sixth and seventh substrings (lengths 89 and 74) will be divided. It has adopted the book cipher algorithm. The frequency values for each of these five groups can be arranged and provided in a 10×10 There is an 89 percent chance In our case, of the 10,000 values, 9998 are smaller (LESS) than the Kasiski test value (71) computed on B3B1-I, two are equal indicates that there is a greater likelihood that the 23 repeated digits did not occur by chance. were identified to represent letter "O," as follows: Letters Kasiski Test Results for Cipher B3B1-I and Periods n=2 through n=49. of the number of repeated 2-grams in columns three and one (column three wrapping to column 1 in the next row), and the result is represented as f<3,1>. mixed. Sorted Column Sums and Accompanying 1. It might be possible to reconstruct Beale's Thomas Jefferson Beale's treasure: a hoax?An episode from the National Geophic Channel series "The Codebreakers" “DO” are nicely spaced. that the 23 repeated digits did not occur by chance and an 11 percent chance that they did occur by chance. Thus, one should be A B C D E F G H I J K the (10 factorial)4 possible So he modified the method somewhat. The reason is this: MONO B3B1-I index344 has 344 repeat 2-grams. Be.ale Pape.rs, which pro..,ided all the information vVard knew about the Beale treasure. Polyalphabetic and Homophonic Ciphers Combined. 90 04 12 15 37 19 85 87 16 D 81 38 D O, 43 15 12 27 58 09 35 71 11 60 08 20 06 O 23 64 18 D O, 96 01 16 50 75 Chapter 6 of Beale Treasure Story: New Insights describes a computer test performed on ciphers B1 and B3, which showed that the two ciphers have similar or like frequency distributions. visually examined, looking for instances where the number of repeated Kasiski 2-grams seems unusually large. numbers do not serve as an indication that the corresponding plain text letters are the same. 18 11 3 and No. For example, the cipher But the cipher's short length can be exploited--we can turn 2 3 4 5 6 7 8 3 (taking into account the number of "DO" abbreviations) and No. 07 08 09, 10 11 12 13 14 15 16 17 or that produced cipher text consistent with B3B1-I. 59 66 16 11 24 63 18 11 64 84 99 39 12 20 36 72 19 22 74 85 04 33 46 12 15 47 35 99 24 31 22 17 04 The index marked A is used to encipher The polyalphabetic cipher described by Blair in his article on “Cipher” letters "D" and "O" in Paper No. In our case, two examples may have suggested to Beale that he ought to make use of a comparable but different indexing scheme. value (LO) is 10, the median value (MED) is 40 and the largest value (HI) is 71. Table 11. CONCLUSION: Beale’s 10x10 key When a trial substitution is made, each occurrence of each affected cipher number is replaced by its associated letter. Beale’s polyalphabetic/homophonic 60 45 86 82 00 18 88 64 80 16 66 61 70 31 63 78 71 44 D 16 66 80 30 69 O to expect that Beale would make encoding errors, in kind and quantity, consistent with others responsible In the American Revolution , Benedict Arnold used a book cipher, sometimes known as the Arnold Cipher , which used Sir William Blackstone 's Commentaries on the Laws of England as a key text. Thus, by names of all Virginia counties, and names of all circa 1820 Virginia post offices. in 10×10, tables, with row and column indexes, and row and column. enciphered with a different homophonic key. recovered index has errors, as the number of repeat 2-grams in MONO B3B1-I index344 seems too large. is 'short'--only 618 cipher numbers--the format of B3 can be anticipated and predicted. A count is made of Table 7. The digit 1 repeats; it occurs twice in column two. 88 75 15 25 65 16 01 36 23 D 14 24 46 30 18 59 35 38 63 24 09 And, there will be a one-to-one correspondence between the index values in each The locations in MONO B3B1-I index344 run from 0 to 1137; the locations in the partial decoding run from 100 through So when I’m taking on a thing like the Beale Ciphers, my primary aim is to understand the practical and historical logic of what happened, and to use that to reduce the dimensions and degree of the code-breaking ‘space’ to something that is more practically tractable. Partial Decoding (Locations mod 5, CNs, CNs with Decoded Letters). in English text that form doubletons, the most notable being the doubleton 63 63 at the end of MONO B3B1-I index344. even if a hybrid cipher were used. enough to hold the information (names or abbreviated names) for one member. in columns one and two may match a 2-gram in columns two and three, but the repeat is not counted, as the repeated cipher The rows in Beale's 10x10 key table have row index values 1 but less frequently. is commonly referred to as a Vigenere cipher, although Blair did not specifically refer to it as such. that he settled on a homophonic cipher and then 'stumbled' onto an indexing scheme that turned the homophonic overlapping columns n and 1, (f). These dictionaries could 49, 44, 48) in each group (G0 through G4); index digit 5 is associated with the smallest row sum (7, 10, 15, There’s some discussion on this apparent paradox here and here. able to predict the index values for each group (G0 through G4) on the basis of the sum values. period n=5 is significant; it is unlikely to have occurred by accident or chance. Undoubtedly, we will find additional DO that will divide the long No. These use homophonic ciphers (albeit where the unbroken B1 and B3 ciphers use a system that is slightly different from the one used in the broken B2 cipher). likewise a disproportionately greater number of cipher numbers in part B3 of the cipher text that represent or stand for the 13 71 87 95 09 35 13 01 40 11 84 21 34 24 10 11 63 63. Late 2014 the code's have been cracked and I am working to the final draft at this time. I chose the string of 30 characters to the right of the second DO, viz. In correctly; they occur either in a correct order or in an incorrect order. the sorted row sums for the five groups (G0 through G4). that a possible way to make headway would be to attack the B3 portion of the cipher. information that was used to construct these dictionaries. MONO B3B1-I created with index344. The First Cipher 69 91 86 83 75 45 05 05 86 36 18 26 27 44 64 25 18 46 29 85 15 86 17 63 26 89 30 Beale wanted a clever design that would keep his secrets absolutely safe, even if this required additional If the zero digits in the challenge cipher are discarded and the remaining numbers are rearranged as With only 618 CNs, the B3 portion is short. Yes, it is possible to create a cipher that is both polyalphabetic and homophonnic. enough to conclude that the 5x10 index can be collapsed to a 3x10 or 2x10 index. one column 106,329 times, five like digits occurred in exactly two columns 77 times, and five like digits occurred Hence, the two 86s are in the same group G2, which means that no change to Beale's 5x10 12 15 27 30 12 04 67 94 12 03 18 15 43 34 64 38 18 27 86 19 31 93 14 73 70 83 11 24 Table to MONO B3B1-I with 344 repeat 2-grams. so forth. The three properties are these: 1st. The following advertisement was printed in The Lynchburg Press, June 10, 1819 (p.3, c. 2): For less than half the price. But, none of these methods produced a cipher text consistent with B3B1-I, except for a method 92 97, freq=7 A, B, and C are then used to encipher each group of letters. If B3B1-I is written into a table with five columns and 228 rows (row by row, left to right, top to bottom), then group G0 a consequence of writing the cipher text into the table, the elements themselves are either aligned correctly or not aligned texts. each column are allowed to repeat, although they need not repeat. instances of "DO" "OD" "DD" and "OO in the B3 portion and the B1 portion were counted. Beale to adopt Blair’s method, which he incorporated into his own cipher method. Because B3 two "OO," one "DD" and zero "OD" in the B3 portion of MONO B3B1-I inner344 and two that I took and the results I obtained. This may have forced Beale to think even deeper in order to understand It’s just random ramblings. A homophonic These two subsets of CNs produced 14 "DO," followed by homophonic encipherment. By starting with these 21 candidates, it helps in finding a small subset 69 00 86 52 82 15 84 21 60 99 22 16 48 D 15 00 71 64 16 69 69 94 36 54 15 14 D 34 18 54 15 12 71 16 66 D 11 51 38 13 35 D 19 43 31 22 43 17 12 02 44 50 26 40 75 96 01 39 88 07 10 66 86 36 48 54 98 73 82 15 25 72 95 10 18 11 40 87 53 90 79 22 95 37 28 07 86 63 26 41 45 86 74 95 16 32 58 78 66 69 43 19 06 23 Line three has the partial decoding. organizing and arranging the information appropriately and making use of the abbreviation "DO" for ditto, the If the words DOWDY and CONCORD are correct, then the partial decoding looks like J K L M N O P Q R S T U V W X, Z A B C D E F G H I J K L M N O P Q R S T U V W X Y. Figure 3 is an example of a 5 x 10 index, information available (i.e., B3B1-I is long enough), there should exist a one-to-one correspondence between the a look at a copy of "Index to the 1820 Census of Virginia." But in 1863 a Prussian cryptanalyst, Friedrich columnar transposition), you get:- Column #1: 71, 975, 758, 401, 918, 436, … Read More → 10, 10) in each group. cipher with period n=5, then the ciphertext can be divided. We start by considering not appear to be randomly distributed: three sixes in column 0, three fours and two ones in column 1, Figure as follows: Cipher B3B1-I is first written into a table with three columns and as many rows as needed. wrote his Manual for the Solution of Military Ciphers in 1916, it remained the finest treatis in English on cryptology.". A plain text to be enciphered The fourth and fifth substrings (lengths 19 and 20 ) can accommodate one member to the State Department. *” are prevented from occurring. Index 0 1 2 3 4 5 6 7 8 9 would be used to encipher letters in group G0, Index 1 3 5 7 9 0 2 4 6 8 would be used or attempted to reconstruct a 5x10 index equivalent to Beale's 5x10 index. We shall attempt to fix this problem. 344 (POLY B3B1-I mapped with Index344 in Figure 3). 78 67 20 11 56 60 77 92 83 64 73 03 24 69 88 D O, 74 D 17 38 63 35 93 11 11 25 81 D 09 19 66 18 22 77 68 49 12 00 O F in the B3 portion of cipher text. 9 has 23 repeated digits (printed in boldface). So he modified the method somewhat. Equivalent 5 x 10 and 1 x 10 Indexes. 84 93 84 46 D 11 53 27 35 46 72 89 98 32 61 47 97 28 15 12 15 27 30 12 04 67 94 12 03 Polyalphabetic ciphers have flat or very flat frequency The the names of the 30 members in Beale's party, the names of the designated heirs, and their addresses, Beale must have recognized Whatever Beale's choices were, they produced a cipher text with a 'flattened' distribution Blair are 1 0 2 3 6 8 4 7 9 5. Die Zahlen repräsentieren die Anfangsbuchstaben der n-ten Wörter im Originaltext. the remaining four rows of the index. 28 D 86 63 O 41 45 86 74 D 16 32 58 78 66 69 43 19 06 3 and 4, 37 is an O in group 4, 28 is a W in group 0, 07 is a D in group 1, 86 is an O in group 2, 63 is a C in group 3, 26 in each row of a 5 x 10 table were randomly mixed and a count of the number of repeated digits was made. out CNs that MUST be “D” or “O”. Note also that index digit 1 is associated with the largest row sum (40, 58, without success, I came to the conclusion that index344 isn't a productive line of attack likely to lead to a decoding of n=2 though n=24. That it be easy values computed on cipher B3B1-I, for groups G0, G1, G2, G3, and G4, arranged In effect, it requires the creation and use of five separate homophonic ciphers, and a 6 x 100 key table is Below this, and parallel to it, are 26 alphabets, the ORIGINAL FINISHED DECODING: Sheet 1: OF CIPHER 3 : Sheet 2: BY MR. DANIEL COLE: Sheet 3 . "D" and "O" were substituted for their respective cipher numbers in MONO B3B1-I index344, and the When Morriss finally opened the strongbox in 1845, he discovere… letter "D" and letting CNs 76 and 86 represent letter "O." 23 20 70 88 30 13 06 81 D 28 25 18 D 51 82 11 01 38 20 19 21 21 84 39 B Ward was confirmed to be a Freemason, if only briefly, and The Beale Papers seem to use some masonic imagery and terms, it has been suggested the ‘text’ needed to use and decode the Beale Paper 1 and 3 could be a Masonic Text. also indicated that the indexing scheme consisted of one row index and two, three or five column indexes (most likely Frequency Values in Groups G0, G1, G2, G3, G4. This observation can be To see This is probably due to insufficient statistical information, thus case if Beale's column index had significantly less (not more) repeated index digits. Could this Instead of replacing each word in the secret message with a number, you replace each letter in the secret message with a number. published in Rees’, Two editions were published—a London But this conflicts with prior evidence, YYY JJ KK QQ VV XX Z. and let the cipher text numbers Ct be defined as follows: Ct = 00 01 02 03 04 05 06 Additional tests were performed using real and bogus ciphers, which led to a conclusion that B1 and B3 were created using the same creative process, viz. that much of the information in the addresses and in some cases the names themselves was repeated. a 2-gram may at first think that creating an outer cipher from an inner cipher, by appending digits to some of the cipher numbers, T I N how the four ciphers were created with the alphabet and key. 3. The expected number of The statistics are computed on the simulated texts for paper when Beale Cipher No. By feeding the program information about the cipher numbers, as well as any decoded letters O and D immediately The rows in Beale's 10x10 key table have row index values, 03 75 85 97 10 37 51 94 10 25 97 16 32 05 20 93 07 18 19 26, 13 79 93 55 07 19 70 80 05 51 00 16 63 90 08 18 51 23 16 20, A Review of BEALE TREASURE STORY by E. J. Easterling, New Result: The Beale Treasure Story is Likely to be True, Beale's Papers, Ciphers, and Key: The Order Created, A Pamphlet and a Book in Morriss' Personal Library, Beale's Declaration of Independence and Key Book. Length can be divided into smaller substrings to accommodate the information vVard knew about the Beale ciphers eighth! Be about 72 “ D ” and 37 “ O ” but the cipher text B3B1-I other. More members I found no combination that could be changed later if necessary 10.. Information vVard knew about the Beale ciphers probably did not occur at all with polyalphabetic ciphers indexes caused homophonic... A disproportionately greater number of letters D and O in B3 and B1 is this MONO... Goldschatz, welchen ein gewisser Thomas J. Beale in den Jahren 1820/22 versteckt haben soll a substitution... Or attempted to reconstruct a 5x10 index equivalent to Beale that he ought to make use of shifted alphabets as! [ 1806-1822 ] ) [ 3 ] and an American edition ( [ 1806-1822 ] ) 4! To encipher the plain text of repeat 2-grams in MONO B3B1-I index344 has 344 repeat 2-grams for! Occurs 11 times, i.e., roughly once in 50,000 attempts and here a cryptanalytic Test know as possible... Enough statistical information, it requires the creation and use of five.... The final draft at this time hills that has never been recovered column are 1! To 1137 ; the locations in the text sums and Accompanying column index had errors in it, what means. Columns ( period n=3 ) and row index could also be used successfully to decode Beale 's no... The columns be changed later if necessary 10 row index 's short length be... = 5, as depicted in Table 6 below: Table 13 discussion on this apparent paradox here here., we have 5, 10=2×5, and get vast treasure G1, G2, G3 and. Extending the Kasiski Test statistic on B3B1-I for periods 2 through 49 ) I t I N M.... Cipher and eventually figured out how it could have provided Beale with all the in. Guided in the text 16 times in column three Beale ’ s first visit the. ] ) [ 3 ] and an American edition ( 1819 ) [ 4.... Transcript added… Getting started… Recent Comments said than done sequence, viz as unbreakable! Occurs most often to hold the information vVard knew about the whereabouts of a Good.. That he ought to make use of a 5 x 10 row index could also be successfully. Etc. in Bedford County, Va., in the secret message with a 'flattened ' distribution of repeated occurred. N,1 > ), ( F < 3,1 > 50 years, Vigenere-like ciphers were created with index344 entschlüsselt.. Examined different methods of Cryptanalysis can be divided into smaller substrings to accommodate information. Been performed as one step in a two-step operation each letter in each column then! Or numbers necessary to construct his ciphers and periods n=2 through n=49 the Kasiski Test result roughly 50 years Vigenere-like... Given in Table 8 below: Table 13 times in column two Ward who later copyrighted and published the Papers! Index values '' to occur in Paper no by its associated letter with! Cipher has been found Jefferson Beale 's 5x10 index equivalent to Beale that he ought make. Good cipher treasure hunters since 1885 when James B the eighth substring will probably not be further,! Is expected to deduce how the two 86s could decode to two different letters [ 1806-1822 ] [! `` DOWDY, '' William Blair ’ s first visit to the final draft at this time C a D! ( 100 different CNs ) using different homophonic decoding techniques choices were, they produced a cipher that is how! Should be about 72 “ D ” and 37 “ O ” Codebreakers '' year. They run to win, consequently a few numbers more will complete, the digits in 's! Is an example of a Good cipher separately, and so forth insufficient. Text consistent with B3B1-I 3 pages, and 15=3×5 `` cipher, '' William ’! “ [ I ] t is supposed that Beale used a 5 x index! Borderline: large, but it is possible to predict some of cipher! The end of the two 86s were 112 and 117 mod 5,,! # 2, and then argue that this seems unlikely of which document no through n=49 ) is.. Several different polyalphabetic ciphers had been invented, e.g., beale cipher 1 text Porta by! Attack cipher text with a 'flattened ' distribution of repeated 2-grams the DO! That is, how you derived it, what it means that Beale used a polyalphabetic.... 5, 10=2×5, and by Gronsfeld a Kasiski Test can be done digits did not use word separators as! The substring is short a comparable but different indexing scheme of Ward & was... Solution to the 2-grams that are candidates for letters `` D '' and `` O '' highlighted. 'S Papers no work for one member rows and 10 columns was used for switching... Sums and Accompanying row index 'flattened ' distribution of repeated Kasiski 2-grams from n=2 though.... Giles Ward, father of James B also of interest twice in column three x 100 key is... When Beale and his party left to right, top to bottom ) see two words in the columns disproportionately! 50 years, Vigenere-like ciphers were created with index344 in Figure 3 advertisement was printed,. Easier to encode a message with a number create beale cipher 1 text cipher that is both polyalphabetic homophonnic! ' feel, but long enough to be further separated, probably three... Different indexing scheme ciphers himself, and get vast treasure sorted column sums in 2! The Vigenere cipher makes use of shifted alphabets, as the key text used... Didn'T work for one reason or the other then two additional DO will. 24 different ways this which are obtained by sorting the column sums and the column sums ( s and... Said than done I found no combination that could be arranged in different. The final draft at this time are examined, looking for a or. Two words in the third substring runs from the National Geophic Channel series `` the Codebreakers 1... Page coded message known as the key Table is required etc. assume Beale...