SOLVING CIPHER SECRETS

ROM sundry and variously -situated localities on this fair planet has issued forth the request:

"Give us a method of solving the simple substitution cipher, when the normal divisions between words are not present."

Our correspondents say that their efforts to apply the monoliteral frequency tables in the August 15 and September 12 issues of FLYNN'S meet with little or no success.

And further, since there are no spaces between words in a continuously written, or arbitrarily spaced, cipher, they do not find an opportunity to use the word frequency tables in the May 16 issue, nor the method of " collation and analysis " mentioned also in that issue, and further in that of October 31.

So they ask, therefore, if there is not some more effective method of attacking a cipher of this kind.

As a matter of fact, the simple expedient of eliminating the spaces of the normally divided simple substitution cipher actually creates a new and different kind of cipher, resolvable to a certain extent, as will be shown toward the end of this article, by processes applicable to the spaced variety, but really demanding more effectual methods of analysis.

Many such methods have been devised. but of these several, Solving Cipher Secrets will first deal with that one involving the use of digraphs, or two-letter sequences.

In ordinary usage digraph is taken to mean two vowels, or two consonants having but a single sound. But in cipher parlance the term is allowed a broader significance, being used to designate any sequence of any two letters whatsoever.

Accompanying this article is a table showing the frequency per 10,000 digraphs of every possible combination of two letters. This tabulation, prepared especially for readers of this department, was made from a count of 10,000 digraphs in straight English text, and may be used to calculate the approximate number of times any given digraph may be expected to occur in a specimen message, and also in other ways that will soon be apparent.

Since this table represents sequences of letters, rather than pairs, every letter in the text used was made the beginning of a digraph.

Thus, Solving Cipher Secrets would give the digraphs—SO, OL, LV, VI, IN, NG, GC, CI, IP, PH, et cetera and not merely the pairs—SO, LV, IN, GC, IP, HE, RC, and so on.

In the Table of Digraphs herewith, the frequency per 10,000 of any digraph will be found in the square located in the column designated by the first letter of the digraph, and in the row indicated by its second letter. The absence of a number in any square is not to be taken as evidence that such digraph is impossible of occurrence, but rather that it is extremely rare.

By making 10,000 the basis of the count, percentage of any digraph may be found by merely pointing off two places. Thus TH, with a frequency of 377 in 10,000, has a percentage of 3.8; HE occurring 305 times, a percentage of 3.1 ; and so on.

As a further aid, a list of the most frequently used digraphs, arranged in the order of their descending frequencies, will be found under the complete table. Just as E tops the list of single letter frequencies, so here TH towers head and shoulders over all other digraphs, leading its nearest competitor HE by a safe margin.

The numbers in the last row of the complete table represent the totals of digraphs in each column. In effect this is a monoliteral frequency table of 10,000 letters prepared from the same text used for the digraphs.

Finally, before demonstrating the solution of a simple substitution cipher by means of these tables, it should be mentioned that digraphs are not limited to solving this kind of cipher only. This department will show, in future issues, how digraphs can be variously applied in resolving widely different kinds of ciphers.

Now to our cipher, one of the simple substitution type in which, as you may remember, any given letter of the alphabet is always represented in cipher by one fixed substitute, each cipher substitute, conversely, invariably signifies the same letter of the alphabet.

Here is such a cipher. You will observe that there are no divisions between words. In the original of this cipher the writing was continuous. Here the arbitrary grouping by fives is purely for convenience, as this practice in all kinds of ciphers decreases the liability of error in transcription.

53‡‡†  305))  6*;48  26)4‡  .)4‡)  ;8o6*
;48†8  ¶60))  85;;]  8*;:‡  *8†83  (88)5
*†;46  (;88*  96*?;  8)*‡(  ;485)  ;5*†2
:*‡(;  4956*  2(5*-  4)8¶8  *;406  9285)
;)6†8  )4‡‡;  1(‡9;  48081  ;8:8‡  1;48†
85;4)  485†5  28806  *81(‡  9;48;  (88;4
(‡?34  ;48)4  ‡;161  ;:188  ;‡?;

A usual procedure in working with a cipher of arbitrary signs, is first to rewrite it, using letters in place of the signs or characters. Letters can be employed in every phase of cipher work more conveniently than signs, and the use of the latter ordinarily results in making a system more cumbersome in use, without any advantage of being less certain of solution.

The decipherer should, however, use his judgment in such cases. It is often possible to solve many ciphers of this kind in less time than would be required to transcribe them into literal form. In the present instance it is enough to have mentioned this step.

The table of digraphs can most readily be prepared from the cipher by using a ruled form similar to that illustrated in the Table of Digraphs, recording each digraph in its proper square as it occurs in the cipher.

Since every character is the beginning of a digraph, the number of digraphs in a given cipher will always be one less than the total number of characters. Following is a list of all digraphs occurring in the present cryptogram more than once, arranged in the order of their descending frequencies:

;4 12     5* 3    5; 2    ‡‡ 2
48  8     5) 3    28 2    ‡( 2
85  5     8† 3    81 2    ‡; 2
88  5     80 3    8¶ 2    ‡? 2
6*  5     8* 3    1( 2    (8 2
)4  5     8; 3    4) 2    )8 2
†8  4     1; 3    9; 2    )6 2
8)  4     06 3    *† 2    )) 2
4‡  4     (‡ 3    *8 2    ;1 2
*;  4     (; 3    *‡ 2    ;: 2
;8  4     ); 3    ‡9 2    ?; 2

And next you will find a frequency table of all single characters in the cipher, also arranged in that order:

8 34    5 12    9 5     - 1
; 27    6 11    2 5     . 1
4 19    (  9    : 4     ] 1
) 16    †  8    3 4     ———
‡ 15    1  7    ? 3     204
* 14    0  6    ¶ 2

Now that the tables have been explained, and the frequency lists duly prepared, the next question is how to turn this data to the solution of the cipher.

Table of Digraphs and List of Those Most Frequently Used

TH-377   ND-116   IT-90   IS-71   FT-61   DI-55
HE-305   ES-115   EA-84   LE-67   RI-60   WA-55
ER-184   HA-114   NG-83   SE-65   IO-59   DT-54
IN-173   OF-113   OU-80   DE-64   SO-59   EC-54
AN-168   EN-101   AR-78   RA-64   EM-58   CH-53
OU-162   OR- 99   ET-77   TA-64   FO-58   CO-52
RE-139   AS- 98   HI-7S   AL-63   EE-57   LI-52
AT-136   NT- 97   TE-74   VE-63   SI-57   BE-51
ED-126   TI- 93   RO-73   ME-62   SS-57   EO-50
ST-119   TO- 92   NE-72   RT-62   TT-57   IC-50

It has previously been shown that when depending on the frequencies of single letters it would hardly be possible to definitely determine more than one letter, namely E, the substitute for which would ordinarily be easily recognized in that it usually predominates over all others,

By the same token, in a frequency table of the digraphs of any cipher, it would similarly seem impossible to definitely identify more than one digraph, TH, the substitute for which should outnumber all other pairs of cipher characters.

And, indeed, by frequency alone this might well be true. But here is where another principle, that of sequence, comes into play. By frequencies alone results might often be very limited. But also be given sequences, and the trick will be turned very neatly, as you presently shall see.

To gain a point of entry, since TH=377 is the predominating digraph, the most frequently recurring digraph in the cipher may at least tentatively be assigned this value. Accordingly, it may be assumed with some degree of certainty that ;4, which occurs 12 times, is TH.

Now a digraph has within itself, besides the relative frequencies of its individual letters, four additional aids to discovering its identity, namely: (1) its own frequency; (2) that of the digraph reversed; (3) the frequency of its first letter doubled; and (4) that of its second letter doubled.

To illustrate this, consider the following data concerning the digraph TH, its reverse HT, and its doubled letter TT and HH. This information is found in the table of digraphs, and in the cipher, being based on the assumption that ;4 is the equivalent of TH.

A comparison of these values seems to substantiate the assumption that ;4 is TH. In addition, the characters ; and 4 occur respectively 27 and 19 times in the cipher, thus agreeing approximately in frequency with what may be expected of the substitutes for T and H.

Before proceeding to the identification of further digraphs, it might be well to mention that an analysis of digraph frequencies, alone or together with single letter frequencies or other data, can often be made to reveal many facts about ciphers.

For instance, it would be easy to detect if a message, in addition to having been enciphered in simple substitution cipher, had also been written backward. To illustrate take almost any digraph, say TH. In a reversed message the predominating digraph would be HT, with the result that the second letter instead of the first would be the most frequent, either considered as a single letter T, or as a doubled letter TT in a digraph.

In another case, suppose that a message had also been enciphered in a more complex method of transposition than the above. Here monoliteral frequencies would be normal. But digraph frequencies would not, since these latter depend on sequence, and the original sequence would have been destroyed by the act of transposition.

To continue with the digraphs, having tentatively assumed the values of certain characters, as T and H in this instance, examine the table of digraphs in an effort to find if there are any predominating combinations of these known letters with other letters as yet undetermined, such as will permit themselves to be isolated by their frequencies, or by comparison with other known letters or digraphs.

In the present case we have our choice of proceeding with digraphs beginning or ending with T or H. The table will show that there are fewer predominating digraphs beginning with H than there are ending with that letter, or beginning or ending with T. Since it is easier to choose from among few than many, a consideration of digraphs beginning with H is now the logical step.

The most frequently used of these are: HE=305; HA=114; HI=75; HO=42; et cetera, the first of which, HE, should be easy to recognize because of its high frequency.

In the cipher the following digraphs begin with 4, the substitute assumed to be H :

48   8 times;
4‡   4 times;
4)   2 times.

Thus 48 is most probably the equivalent of HE. That 8 equals E can further be tested by trying it in its combinations with the known letter T. This character also occurs 34 times on the frequency table of single characters, thus being without doubt the predominant E.

If digraphs ending in H be now examined, the substitute for S should not be far to seek. For, next to TH, the digraph SH, although not topping the list, is readily told from the others by the fact that S, now that E and T have been eliminated, occurs more frequently doubled than does any other first letter of digraphs ending in H.

Here are the facts from the table. Letters in parentheses have already been identified:

(TH 377)      (TT 57)
 CH S3         CC 6
 SH 40         SS 57
 WH 32         WW 2
(EH 28)       (EE 57)
 GH 25         GG 3

As a further test for S, its substitute must combine freely with the already discovered substitute for T, E, and H, as shown by the digraph frequencies—ST =119, SE=65, and SH =40.

To apply these facts, exclusive of the already discovered ;4 for TH, the following combinations ending in 4 or H occur in the cipher. The doubles of the first characters of these pairs are also given.

)4 5     )) 2
-4 1     —  0
34 1     33 0

Thus )4 would seem to be SH, since its first characters doubled )) occurs twice in the cipher. The character ) as S also tests positively in its combinations with the assumed substitutes for E, H, and T, since S E, S H and S T occur respectively, 2, 5, and 3 times, in the cipher. Besides, the character ) has a monoliteral frequency of 16, in further proof of its claim to the role of S.

Having now discovered E, H, S, and T, the next attempt will be to find O by an examination of digraphs beginning with H. The Table lists the following combinations:

(HE 305)        (EE 57)
 HA 114          AA  3
 HI  75          II  3
 HO  42          OO 29

Of these, HE is known. The frequencies of HA and HI, not to mention HO, are too close to be of any aid. The reversed digraphs, AH, IH, and OH are also of about an even frequency as are also the monoliteral frequencies of A, I, and O. But in the doubled vowels AA=3, II=3, and 00=29, the predominant OO provides the means of differentiation.

In the cipher, the following combinations beginning with 4 or H occur most frequently.

48  8     (=HE)
4‡  4
4)  2     (=HS)

Here 4‡ must be HO, since if ‡ were I or A, on account of the low frequency of II or AA, the doubled character ‡‡ would hardly be found twice in a cipher of this brevity. Besides ‡ checks well in its combinations with other discovered characters, and in its individual frequency.

This assay has already resulted in the recovery of five of the unknown components of the cipher, E, H, O, S, and T, which in their combined frequencies total over 54 % of the whole message. But the important point is that the alphabet has been this far developed without any reference to the text of the enciphered message.

In a longer cipher it might be possible safely to carry this process much further. Different ciphers will also require different procedures. The TH-HE-SH-HO route need not be followed necessarily.

If now the five known letters be substituted throughout, using dashes at the same time for unknown characters, the partially deciphered message will look like this:

--OO ----SS -- THE -- SHO - SHOSTE
- -- THE - E -.-SSE - TT - E - T - O -
E - E --EES--- TH -- TEE ----- TES -
O - THE - ST ------ O - TH ---------
HSE - E - TH ---- E - STS -- ESHOOT --
O - THE - E - TE - EO - THE - E -
THSHE ---- EE --- E -- O - THET -
EETH - O -- HTHESHOT --- T -- EETO
-T. 

Some words are here already deciphered in full. And many others, partly deciphered, may be completed by the method described in FLYNN'S for May 16, 1925, in connection with the normally spaced cipher.

Thus O-T at the very end of the message is obviously OUT, making ? the equivalent of U. TH-O--H near the close now becomes TH-OU-H, plainly THROUGH. Other words, IN, DEGREES, THIRTEEN, NORTHEAST, NORTH, EAST, FIFTY, now follow in rapid succession, in their efforts to get on the band wagon while there is yet time.

Any who care to follow the solution to the finish will find that they have deciphered the classic example of this kind of cipher, propounded by Poe in his " Gold Bug. " But the solution has here been accomplished by an altogether different method.

Poe began by assuming that the most frequently used character in the cipher was the substitute for E. Then, knowing that T H E was a very common word, he searched for the most frequently recurring pair of characters preceding his supposed substitute for E, to which he assigned the value TH, completing the word THE.

Substituting letters for cipher characters as soon as found, and confining himself to the determination of such partly deciphered words as were capable of but a single reasonable interpretation, Poe arrived at still other words, the order of discovery in this particular message, after E and THE, being — TREE, THROUGH, DEGREE, THIRTEEN, and lastly A.

Now, what with methods here explained, and the copious tables in this and previous articles, you should be pretty well set up to handle the business of the continuously written or arbitrarily spaced simple substitution cipher.

"FUEL FOR THE FIRE"

To provide fuel for the fire, here are a couple of ciphers out of the usual run. What they are is a dead secret. You'll never know their mystery until you solve them. And then you'll agree that they were worth your efforts.

The first of these was actually sent to Poe by a correspondent, W. B. Tyler, who believed his cipher impossible of solution, The solution to this cipher has never been published, to our knowledge. Poe does not say that he solved it. But if he did not, it could only have been that he did not try.

About the second one it is best not to say much, for it would probably not then be necessary to print the solution in the next article. As it is, you will have some thinking to do. And if you do not get it, you will be surprised when you see the explanation.

CIPHER No. 1. (W. B. Tyler to E. A. Poe.)

CIPHER No. 2.

9-5765-1   4-1-7   1-54-5-4   5-59-52   8-9-1
41-57   8-1-4   5-4-63   2-2   5-4-2   8-5-761
1   9-2-9-39-2   7   66-1   634   9-57-1-3-9-
78-5-92-7   2   8-5-7-634-2-3   7-45-2   8-64-1
1-7   9-57-1   3   7-7-5-1   2   9-4   5-37   2
8-1-8   562   6537   4-9-4-765-1   634   5-8-9-1
3-9-78-5-92-3   2   2   8-1-2   8-5-8-1-4-41-
4-5-9-2   2   661-51-51   5   5-92   62   8-5-1
5-9-1   2   8-5-35-7-9-6566-6-37   551   3-9-
78-5-96-1-51   5   8-63-1-51-2-37   2   5-i
2   9-6-7   1-1   2   62   8-5-7-5-53   9-55-
55-1   1   66-765-1   45-2   8-64-9.

WHO FIRST SOLVES THIS WINS TEN DOLLARS

Fans, here is a good chance to replenish your bank account. Read how!

DEAR SIR:

I am taking the liberty of sending you a cryptogram which I venture to doubt, if I ever doubted anything, can be solved.

At all events I am willing to give my word of honor that I will send ten dollars for you to forward to the first one sending in a correct solution.

Personally, I think there is a good deal of " bunk " in anybody's claiming to solve any cipher sent in.

HOBART HOLLIS.

Hazleton, Indiana.

Of the several ciphers so far submitted by fans who back up their challenges to solve with a promise of reward, we believe that the present cipher, though difficult, offers the best chance of success.

CIPHER No. 3 (Hobart Hollis).

5-2-20-130-13-10-80-30-140-120-2S-78-78-120-
39-38-33-130-156-160-21-6-72-40-18-20-9-50-
175-52-105-66-48-18-110-30-30-75-10-50-45-
76-8-36-5-100-95-35-26-27-26-75-2-39-88-21-
4-20-40-102-26-78-24-44-72-9-27-68-50-78-28-
22-2-75-40.

Priority of solution will be determined by postmark; and in case of a tie, the best explanation of the method used will determine the winner. Any solution to be considered must be mailed on or before January 23, one month from the date of this issue.

A complete explanation of this cipher will appear in an early issue of Solving Cipher Secrets.

Speaking of challenge ciphers, do you remember the one by Fridolf Holmberg in FLYNN'S for October 31? Well, no one solved it, and so Fridolf still has his money. But if he made his cipher difficult to spare his pocketbook, by so doing he only endangered his life.

People all over the country who tried to cope with his problem are on the warpath with blood in their eye. We advise our correspondent to wear a false mustache or something, and to tread the public highways with caution.

Just read the following letter and message in cipher. They plainly show how the wind is blowing.

DEAR SIR:

I have lately been following your department very closely, and have found some very interesting specimens of ciphers. They afford recreation, study with success, and real concentration, in each issue.

By the way, you may send the inclosed cipher (No. 3, below) to Fridolf Holmberg, who has that pippin in the October 31 issue.

This one has me hooked yet, though the rest were comparatively simple. M y wife is almost crazy, too. She can't seem to get started with it. I remain yours till the game wardens put a closed season on ciphers, then I'll quit my job.

D. WASHBUR N HALL,

Manager W. U. Telegraph Co.

Chehalis, Washington.

CIPHER No. 4 (D. Washburn Hall).

EYGK KXKC ZYKX KNZ HXGJ C YGN
KXAY KN QIKXC YAUBXKT G ZYUSRG
SG O EXXUY KH RROC KN XU YORG-
NKNI LU XGKRJ EGZY UZ MXKHSUN
LRUJOXL TXGC O.

Whoever believes the fair sex is not strong on ciphers must certainly revise his opinion when he tries his hand at this clever creation in the way of a mathematical cipher by Mrs. E. J. Hyatt, East Elmhurst, New York.

Here in its simplest form, it could readily be disguised as a memorandum of expenses, or otherwise. But in spite of the innocence of its appearance, you will find it a wicked, tricky thing to work with. What can you make of this cash account?

CIPHER No. 5 (Mrs. E. J. Hyatt).

 $5.29      $4.78      $5.95      $3.50
  5.80       3.15       5.63       3.74
   .60       i.55       2.80       3.16
  3.35       2.25       6.61       5.23
  3.30       2.19       2.32       3.60
  2.10       3.00       1.89        .70
  3.50       5.29       3.56       6.40
  6.68       5.61       5.71       5.62
  3.14       3.42       3.49       6.60
  6.53       4.25       1.75       1.50
   .27       3.17       4.68       1.81
  4.50        .50       2.17        .33
  3.44       3.60       2.00       2.11
  5.72       5.68       3.20       3.75
  5.00       3.44       5.18       5.54
  4.70       1.77       3.00       5.59
  1.39       6.70       3.22       2.76
  5.66       3.29       5.75     ——————  
  2.00       5.75       1.98     $64.18
  3.51       1.80       4.90
  5.50       4.75       6.59
  3.70       5.68       1.89
  6.55       6.00       4.62
  3.45       4.50       1.77
   .56       2.75       2.25
——————       4.77       3.10
$96.14       5.70       6.26
             2.30       2.15
          ———————    ———————
          $107.64    $104.43

Now try this one from William H. B. Woodbury, Finger-Print Expert, Northvale, New Jersey. Mr. Woodbury did not intend his message to be indecipherable, but only as he expresses it, " to give FLYNN'S cipher fans a bad half hour."

CIPHER No. 6 (William H. B. Woodbury, F. P. E.).

QLBY XLBHH DY RPFODY B
JBK DO LC EFKXYPCEY EDR-
LCPX MCXY YF LBZC XFJC FKC
XFHZC YLCJ ACOFPC YLC DKN
DX IPT LBZC KCZCP JDXXCI B
KCJACP FO OHTKKX XDKEC
ODPXY RCAHDXLCI BHXF YLDKN
TFGP ICRBPYJCKY XFHZDKU EDR-
LCP XCEPCYX YLC JFXY DKY-
CPCXYDKU OCBYGPC RCAHDXLCI
DK BKT JBUBVDKC KFQ CSDX-
YCKY.

The next cipher is a real treat, from a correspondent who has requested us to subscribe his letter with his initials only.

DEAR SIR :

The writer, whose life has been spent in the far places from civilization, is sending you a cipher he picked up in certain South Sea Islands.

This cipher is known as the D. A. It is simplicity itself, and any child can work it — if he gets the key.

However, without the key you will have double the fun trying it out the first time. Figure it out and add your own variations for a second, or more complicated cipher.

J. W. B.

Grand Rapids, Michigan.

CIPHER No. 7 (J. W. B.).

84160 96700 3I980 08217 40097 S0030
16028 05127 74402 10040 01003 39201
36637 46683 22122 07169 02360 5001S
41200 10420 77597 67761 35311 24769
02522 19922 42133 00982 59840 06412
38574 13339 14001 54977 72012 59000
17020 00104 12668 08058 83210 36362
20992 71115 05847 76161 33388 56110
21904 47149 92160 SS52I 36884 05742
25389 92322 10016 08024 01020 58400
17699 20821 11346 20663 30001 88739
00126 68161 30061 77491 76330 41079
35821 44714 66840 05770 21969 24480
12566 22567 77405 56210 18413 04.

The key to this cipher most ingeniously combines complexity in effect with simplicity in use. And it is one that would readily lend itself to the structure of ciphers of greater intricacy. The original of the above cipher was continuously written. Here it is grouped by fives to insure accuracy.

With the one exception of the solution to No. 3, which will be held over till a later date, the solutions to all of the ciphers in this issue will be found in next Solving Cipher Secrets.

The fans can therefore whet up their curiosity as to any unsolved ciphers to the very limit; and this with the previous assurance that their curiosity must soon be gratified.

SOLUTIONS TO PREVIOUS CIPHERS

One criticism of the Fleissner grille, described in the December 5 issue and with method of solving in the issue of January 2, might be the difficulty in safely transmitting the key to a correspondent. To obviate this, all the grilles used in the above articles were constructed by using a simple key word.

Thus, to construct a grille for the key word FATAL, the first letter, F, the sixth in the alphabet, indicates a 6 x 6 grille, which will have nine openings.

If the key word has nine or more letters, use only the first nine. If less than nine, as in this instance, repeat it as required to make up this number. Then number these letters from 1 to 9 in accordance with their alphabetic order, taking any repeated letters from left to right. Finally divide the series into four equal groups, including any remainder with the last group, thus:

F A T A L F A T A
5 1—8 2—7 6—3 9 4

This key word is now transformed into the numerical key: I:1-5; II:2-8; III:6-7; IV: 3-4-9. Which formula can be used as in the December 5 issue in making the grille.

Here are the translations and key words to the December 5 Fleissner ciphers.

CIPHER No. 1. (Key: FLEISSNER.) Our food, water, and ammunition are exhausted. Send supplies and reinforcements or we must surrender within twelve hours. YZ.

CIPHER No. 2. (Key: HEXAGON.) Enemy plans attack from Northwest at daybreak. Prisoners estimate his strength at one hundred and fifty thousand. We are sending fifty thousand infantry reinforcements and one hundred cannon, which will reach you before midnight.

The next two are those of the January 2 issue:

CIPHER No. 1. (Key: FATAL.) The Fleissner grille becomes more difficult to decipher as its size increases and as the length of the message decreases. QXVCRST.

CIPHER No. 2. (Key: FATAL, grille turned over.) This message was written with the same grille as was used in enciphering the first with the difference that the grille was turned over, the under side becoming the upper. XAVTU.

Beginning with this issue, Solving Cipher Secrets is inaugurating a new policy, that of printing in each issue of the department the solutions to all the ciphers in the previous article. Inasmuch as this will considerably broaden the scope of the department, and add to its interest, we believe that it will meet with universal approval. Here are the solutions to the rest of the January 2 ciphers:

CIPHER No. 3. (Captain W. O. Cooper.) " Try and do it."

Captain Cooper's cipher used the numerical key, 1-2-3-4-5. In other words, take the first letter of the first word, the second letter of the second word, and so on up to the fifth letter of the fifth word; repeating the formula as the length of the message requires, thus:

TAKEN WRITING KEY HAWAII
FLYNN'S DECIPHER IDEA BOOK
COMING CRYPTOGRAM

The numerical key here can be any number whatsoever, and it could be conveyed in numerous ways. For example, the date, July 4, 1776, would give the key, 7-4-1-7- 7-6. And, further, as Captain Cooper suggests, before concealing the letters in the words, they can first be enciphered in another system, as by taking the next letter in the alphabet, or by using the Gronsfeld cipher with the same key.

In a longer message, a consideration of the factors of intervals between shorter words, which must represent small key figures would result in the discovery of the number of figures in the key. I f the words be then transcribed into rows of this length, significant letters in any column will be similarly located.

CIPHER No. 4. (William I. Rawsden.) " If FLYNN'S Magazine keeps up the present high quality of their stories, they will deserve the added business which it will bring them."

This message was enciphered by substituting for each letter the twelfth in advance of it in the alphabet. Ciphers of this kind may be solved by the special method suggested in the last article, which is also easily done on the typewriter in this fashion:

UVWXYZABCDEFGHIJK L etc.
RSTUVWXYZABCDEFGHI —
RSTUVWXYZABCDEFGHI —
XYZABCDEFGHIJKLMNO —
KLMNOPQRSTUVWXYZAB —
ZABCDEFGHIJKLMNOPQ —
etc.

CIPHER No. 5. (Gilbert Hagedorn.) " Pacini Kenosha unsolved murder mystery awaits an avenging hand."

This interesting, simple numerical substitution cipher is somewhat troublesome to solve, but only because of its brevity. Dividing 51, the number of cipher groups, by 6, the number of columns, gives a quotient of 5.6, which agrees very closely with the average five letter length of English words, and suggests the columnar arrangement of words.

Here is the alphabet, insofar as the message reveals it:

 A  B  C  D  E  F  G  H  I  J  K  L  M
98 —— 94 92 86 —— 82 80 72 —— 63 62 61
 N  O  P  Q  R  S  T  U  V  W  X  Y  Z
60 59 56 —— 50 44 43 42 40 38 —— —— ——

The substitutes taken alphabetically are in descending numerical order, which fact should have been of some assistance in solving.

CIPHER No. 6. (Dr. G. M. Weitz.) " To my knowledge this cipher has never been solved. Several army cipher experts have given the solution to good luck cipher solvers."

Dr. Weitz's system consists in first reversing the message, and then enciphering it with the key, 123321 by a method exactly similar to the Gronsfeld cipher explained in FLYNN'S for June 6 with the exception that the counting is done backward in the alphabet instead of forward.

Here is a portion of his cryptogram deciphered:

Cipher:     RPBSJNR PBENHB IZRJ etc.
Key:        1233211 233211 1551 etc.
Deciphered: SREVLOS REHPIC KCUL etc.
Transposed: LUCK CIPHER SOLVERS.

Foundations for solving this type of cipher, as well as the Gronsfeld, the Vigenère mentioned below, and other innumerable variations, will be laid in next Solving Cipher Secrets.

CIPHER No. 7. (Fletcher Pratt.) "This cipher is based on the roulette wheel and will defy decipherment unless such an instrument is used in the process. Its advantage is that it is easy to write and read, while the difficulty of solution increases with the length of the message."

Following is the arrangement of Mr. Pratt's thirty-eight-letter alphabet, beneath which is the series of numbers on the roulette wheel at the initial position of the wheel, for enciphering the first word of the message.

  A  B  C  D  E  F  G  M  I  J  K  L  M
—00——1-13—36—24——3—15—34—22——5—17—32—20—
 N  O  P  Q  R  S  T  U  V  W  X  Y  Z
—7—11—30—26——9—29——0——2—14—35—23——4—16—
  E  T  A  N  O  R  I  S  H  D  L  U
—33—21——6—18—31—19——8—12—29—25—10—27—

At the beginning of each word the wheel bearing the numbers is turned one place in a counterclockwise direction. The letters of the second word would thus be represented by numbers one place to the right of those in the initial position; the third word by those two places to the right; and so on, until at the thirty-ninth word the wheel would again be in its initial position.

Cipher: 21——29——22——28   36——5——26——22——3——28    17——2   etc.
Message: T   H   I   S    C  I   P   H  E   R    I   S   etc.

The translation of these few words is enough to demonstrate the method of this ingenious system.

As a fitting conclusion we have listed the solvers of ciphers in FLYNN'S for October 31.

Mr. Bellamy heads the roll of honor in number of ciphers solved, with Mr. Schmutz a close second, and the only one submitting a solution to No. 5, a "Vigenère cipher by William E. Bowms, Corps Area Photographer with the Army Signal Corps at Fort McPherson, Georgia. This cipher will again be referred to when methods of deciphering this type have been explained.

• Arthur Bellamy, Boston, Massachusetts. (1-2-4-6.)
• A. P. Schmutz, Philadelphia, Pennsylvania. (1-5-6.)
• A. "W. Sweeten, M. E., Sharon Hill, Pennsylvania, (1.)
• Charles C. Fulton, Internal Revenue Chemist, Omaha, Nebraska, (1.)
• Thomas J. 'Williams, New York, N. Y. (6.)
• Charles F. Graf, Brooklyn, New York. (6.)
• James Riley, Norwich, Connecticut. (1.)

You will notice that the roll of honor is much shorter than it used to be when Solving Cipher Secrets' was in its primary stages. But strangely the enthusiasm of fans is not waning. Though fewer of them succeed with the more complicated codes, not a single voice has been raised for easier problems.