book cover
From FLYNN's January 22, 1927


Edited by M. E. Ohaver

HILE Edgar Allan Poe is commonly associated with cryptography through his masterful tale, "The GoldBug," a certain obscure review, published in Graham's Magazine for April, 1841, and antedating the story by more than two years, must also receive its full share of attention.

This review concerns a collection of short biographical studies, "Sketches of Conspicuous Living Characters of France," that had originally been published at Paris in weekly numbers by some one who styled himself "un homme de rien," but which had now been translated into English by Robert M. Walsh, and just offered—in 1841—in book form by Lea & Blanchard, Philadelphia publishers.

A glance at the array of celebrities listed in this volume is alone enough to mark it as of exceptional interest:


To a man of Poe's bent several of these sketches must have held an extraordinary appeal. Thus—aside from the notices of poets, writers, and novelists—the author of "The Gold-Bug" must have been keenly interested in the colorful life of Jean Lafitte, French corsair of the Gulf of Mexico.

And incidentally, could the name of Poe's celebrated detective, the Chevalier C. Au - guste Dupin, to whom is ascribed the distinction of being the original detective of fiction, have been suggested by the Dupin of the sketches? Poe's Dupin, at any rate, made his first appearance in "The Murders in the Rue Morgue," also in the April, 1841, issue of Graham's Magazine.

We would like to believe, however, that Poe was interested most of all in the Ber - ryer sketch, where it is told how a letter addressed by the Duchess of Berry to the Legitimists of Paris was accompanied by a long note in cipher, the key to which she had forgotten to give.

"The penetrating mind of Berryer," writes the unknown biographer 'of this key, "soon discovered it. It was this phrase substituted for the twenty-four letters of the (French) alphabet—Le gouvernement provisoire."

To this statement Poe took exception in no uncertain terms. "All this may be very well in anecdote," he wrote, "but we cannot understand the extraordinary penetration required in the matter."

Also, in the essay mentioned below, he added: "The assertion that Berryer 'soon discovered the key phrase,' merely proves that the writer of these memoirs is entirely innocent of cryptographical knowledge.

"Monsieur B. no doubt ascertained the key phrase, but it was merely to satisfy his curiosity, after the riddle had been read. He made no use of the key in deciphering. The lock was picked."

To prove his contention, Poe issued the following remarkable challenge:

The phrase "Le gouvernement provisoire" is French, and the note in cipher was addressed to Frenchmen. The difficulty of deciphering may well he supposed much greater had the key been in a foreign tongue; yet any one who will take the trouble may address as a note in the same manner as here proposed, and the key phrase may be either in French, Italian, Spanish, German, Latin, or Greek—or in any of the dialects of these languages—and we pledge ourselves for the solution of the riddle. The experiment may afford our readers some amusement—let them try it.

Whereupon there ensued a lively and exciting controversy between Poe and his readers, which contributed largely to what is probably Poe's most important work on ciphers, his essay, "A Few Words on Secret Writing," published in the July, 1841, issue of Graham's Magazine, and, curiously enough, retitled "Cryptography" in some editions of his works.

In response to the challenge Poe received two cryptograms of the type designated, both from an unknown correspondent at Stonington, Connecticut, who signed himself S. D. L.

In one of these an English key phrase was employed, and in the other a Latin. Poe proved his claim by promptly solving them both. The first of these is used for illustrative purposes in the present article.

The key to the key phrase cipher, a description of which is by this time necessary, consists of a phrase or sentence known only to the corresponding parties.

This key phrase may be in any language. but when intended for use with the natural English alphabet it must consist of twenty-six letters.

To prepare the cipher alphabet, merely write the key phrase, letter for letter, below the letters of the alphabet, taking both in regular order. The following enciphering key is formed on the key phrase to the example sent to Poe, "A word to the wise is sufficient."

        Enciphering Key
Message:  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 Z
Cipher:   a w o r d t o t h e w i s e i s s u f f i c i e n t 

Any desired letter can now be represented in cipher by the key phrase letter directly below it. Thus, in the short message at (a), Y=, 0=/, and so on, as shown in the cryptogram at (b).

(a) Message: Y O U R  L I F E  I S  I N  D A N G E R.
(b) Cipher:  N i i u  i h t d  h f  h e  r a e o d u.

When enciphering, the key phrase alphabet is thus used exactly as a simple substitution alphabet. Rut the process of deciphering is not always so easy. For it is possible here for a given cipher letter to have more than one fixed alphabetical equivalent.

Perhaps the easiest way to decipher such a cryptogram is to prepare a deciphering key, like that subjoined for the present key phrase, showing the several equivalents for each symbol:

Cipher:  a c d e f h i n o r s t u w
Message: A V E J S I L Y C D M F R B
               N T   O   G   P H   K
               X     U       Q Z

To decipher the cryptogram (b), write beneath each symbol, as shown at (c), the one or more letters found in the deciphering key. Then try to form words from these letters, using, of course, only one letter for each symbol, as illustrated at (d) :

(b) N i i u  i h t d  h f  h e  r a e o d u.
(c) Y L L R  L I F E  I S  I J  D A J C E R.
      O O    O   H      T    N      N G
      W W    W   Z           X      X
(d) Y O U R  L I F E  I S  I N  D A N C E R.
             W I F E  I T       D A N G E R. 

In this type of cipher it is thus sometimes possible for a given group to signify any one of several different meanings. For example, ihtd can be translated either as LIFE or WIFE. Similarly, hf could be either IS or IT, and raeodu might be DANCER or DANGER.

In such cases the words actually intended must be determined by context, or by knowledge of conditions referred to in the message. The receiver of the above cryptogram might thus be certain that it should read: YOUR LIFE IS IN DANGER; but otherwise the message might remain ambiguous.

One of Poe's correspondents, W. B. Tyler, with this peculiarity of the cipher in mind, remarked that this was "not a very desirable system, as it would scarcely be more useful than a lock without its key, or with one that did not fit its wards."

Poe was in full accord with this criticism, stating that a model cipher should not only comply with Sir Francis Bacon's three essentials—in being (1) easy to write and read, (2) trusty and undecipherable, and (3) clear of suspicion—but should possess the additional virtue of also being, with the key, "promptly and certainly decipherable."

And he had already, in his essay, outlined the following unusual method of attaining this fourth property in the key phrase cipher:

To obviate, therefore, the exceeding difficulty of deciphering this species of cryptograph, on the part of the possessors of the key phrase, and to confine the deep intricacy of the puzzle to those for whom the cipher was not designed, it becomes necessary that some order should be agreed upon by the parties corresponding—some order in reference to which those characters are to be read which represent more than one letter—and this order must be held in view by the writer of the cryptograph. It may be agreed, for example, that the first time an i occurs in the cipher, it is to be understood as representing that character which stands against the first i in the key phrase, that the second time an i occurs it must be supposed to represent that letter which stands opposed to the second i in the key phrase, et cetera, et cetera. Thus the location of each cipherical letter must be considered in connection with the character itself in order to determine its exact signification.

It would seem, however, that this device merely shifts the burden from the shoulders of the translator to those of the inditer of the cryptogram. For upon him would devolve the even greater, if not impossible, task of so wording his message that the letters in each of These series would always be repeated in alphabetical order.

Further, Poe's proposal would seem to violate what might be termed a fifth essential of the model cipher, that it be capable of transmitting any message whatsoever, regardless of its wording.


Now that the cipher has been fully discussed, how to resolve it without the key is next in order. But first it must be said that key phrase cipher can usually be identified by certain easily recognizable characteristics: less than twenty-six letters will be employed; frequently used letters will ordinarily predominate as cipher symbols; word divisions will be normal; and there will occur many combinations impossible in simple substitution ciphers, such as iiiir (go) in the following example.

In explaining the analysis, the first of the two key phrase ciphers sent to Poe will be used. The numbers interspersed in the following transcription of this cipher are added here to facilitate reference.

"Cauhiif aud ftd sdftirf ithot (5) tacd
wdde rdchfdr tiu fuaefsbffheo (10) fdoudf
hetiusafhie tuis ied herhchriai (15) fi aeiftdu
wn sdaef it (20) iuhfheo, hiidohwid wn aen
deodsf (25) ths tiu itis hf iaf (30) iuhoheaiin
rdffhedr; aer ftd auf (33) it ftif fdoudfin
oissiehoafheo hefdiihodeod (40 ) taf wdde
odeduaiin fdusdr ounsfiouastn. (45) Saen
fsdohdf it fdoudf iuhfheo (50) idud weiie fi
ftd aeohdeff; (55) fisdfhsdf a fiacdf tdar iaf
(60) ftacdr aer ftd ouiie iuhffde (65) isie
ihft fisd herdihwid oiiiiuheo (70) tiihr, atfdu
ithot ftd tahu {75) wdheo sdushffdr fi ouii
aoahe, (80) hetiusafhie oiiir wd fuaefshffdr
ihft (83) ihffid raeodu ftaf rhfoicdun iiiir (90)
defid iefhi ftd aswiiafiun dshffid (95) fatdin
udaotdr hff rdffheafhie. Ounsfiouastn, (100)
tiidcdu, siud, suisduin dswuaodf ftifd (105)
sirdf it iuhfheo ithot aud (no ) uderdudr
idohwid iein wn sdaef (115) it fisd desiaeafiun
wdn ithot (120) sawdf weiie ftd udai
fhochthoafhie (123) it ftd ohstduf dssiindr fi
(130) hff siffdffiu."

Poe did not divulge his method of solving the key phrase cipher. Nevertheless, it will now be shown how the cipher can readily be resolved by means of well known principles, some of which, at least, Poe himself used on simple substitution ciphers. The principles here illustrated—this article by no mean exhausts the subject—are those of (1) frequency, (2) comparison, and (3) context, as applied to the determination of letters, letter combinations, affixes, and words.

Suppose that we begin with simple frequencies. In "The Gold-Bug" Poe assigns the value E to the most used symbol. But here the most used symbol might represent several letters, and E might not be one of them. However, since E averages 12.5% (see table in FLYNN'S WEEKLY for January 23, 1926) any symbol in the present cryptogram with a frequency of about 12.5% or more, at once becomes probable material for E.

i  115  15.9%     o   41  5.6%
f  103  14.2      s   40  5.5
d   94  13.1      r   26  3.6
h   61   8.4      n   20  2.7
e   53   7.3      w   16  2.2
t   50   6.9      c    8  1.1  
u   48   6.6      ————————————
a   47   6.5         722

According to the above frequency table, i, f, and d, stand capable of assuming the role of E in the present cryptogram.

Next, Poe tells us that the most used word in the language is THE. In the word table in FLYNN'S W'EEKLY for May 16, 1925, the average of THE is placed at 6.3%. The most used three-letter group in the present cryptogram is ftd, occurring eight times, or six per cent of the one hundred and thirty-two words in the message. The third letter, d, of this group is already probable material for E. Hence, it may be assumed that one of the meanings, at any rate, of ftd is THE.

It is well to check the frequency of each letter as it is discovered with that of the symbol representing it. Thus, the small difference between d (13.1%) and E (12.5%) indicates that if d also stands for any other letter or letters, they must be of low frequency.

The next principle to be illustrated is that of "collation and analysis of shorter words," mentioned by Poe in "The GoldBug," but also applicable to this type of cipher. In the following table all groups of from one to five letters are listed, to facilitate comparison:

  (1)                      (3)                 (4)                  (5)

a   57               aud  2-110            tacd  6           ithot  5-73-109-
                     ftd  3-34-54-63-      wdde  7-42                 120
                            74-93-123-     tuis  13          sdaef  19-115
  (2)                       127            itis  28          weiie  52-122
                     tiu  9-27             ftif  37          ouiie  64
fi  16-53-78-        ied  14               Saen  46          tiihr  7 1
      130            aen  24               idud  51          atfdu  72
wn  15-23-114        ths  26               tdar  59          wdheo  76
it  20-36-48-        iaf  30-60            isie  66          aoahe  80
      107-116-       acr  33-62            ihft  67-85       oiiir  82
      126            auf  35               fisd  68-117      iiiir  90
hf  29               taf  41               tahu  75          dcliil 91
wd  83               hff  98-131           ouii  70          iefhi  92
                     wdn  119              ftaf  88          ftifd  105
                                           siud  102         sirdf  106
                                           iein  113         sawdf  121
                                           udai  124

There is only one single letter word in the whole cryptogram, a (57); from which, however, the symbol a can be assigned one or both the values A and I.

Were there now any two letter groups beginning with a, such words as AM, AN, AS, AT, IF, IN, IS, IT, could be considered. But passing on to the three letter groups we find aud (2-110), aen (24), aer (33-62), and auf (35), beginning with a, which might thus signify such words as ALL, AND, ANY, ARE, ITS, and so on. Since a=A and d=E, it is probable that aud=ARE, giving the value, u=R.

Another valuable means of comparison applicable to this kind of cipher is afforded by short words beginning or ending in CH, SH, TH, and WH; especially by those beginning with TH, which—on account of the wide use of such words as THE, THAN, THAT, THExM, THEN, THEY, THIS THUS, THEIR, THERE, THESE, THOSE, and so on—is ordinarily the predominating initial digraph.

The prefix ft- predominates here among the short words. Accordingly, all the groups are listed that begin and end with ft, since this presumably is TH. Also—having the other H combinations in mind—all groups having t as the second or last symbol are similarly set down.

  ftd      THE
  ftaf    THAT
  ftif    THUS
  ftifd   THOSE
ihft    WITH
  itis   WHOM
  ithot  WHICH

It is easy enough to determine the meanings of these groups by comparing their symbols, as in simple substitution cipher. The only difference is that a given symbol can here be the substitute for several different letters. The English equivalents are all fairly obvious with the exception, perhaps, of ftif, which might he taken for THIS were it not that the symbol for I is h, as found in ihft =WITH, and ithot=WHICH.

Newly discovered meanings for symbols should be checked by their frequencies. Thus, it is compatible here for f (14.2% ) to represent both T (9.5% ) and S (6.2%). Another important asset is the analysis of the beginnings and endings of the longer words, most of which are formed by means of prefixes and suffixes.

To assist in determining these affixes, the accompanying table has been prepared, showing the number of times per thousand words—counting only those of six or more ietters—the most used combinations can be expected, on an average, to occur. The one thousand words from which the table was made constituted approximately thirty per cent of the whole number of words of all lengths in straight English text.


First Letter: S- 111; A- 88; C- 85; P- 82; M- 70; B- 69; R- 68; D- 65; E- 63; F- 55; I- 43; T- 33; L- 29; W- 28; O- 27; G- 22; H- 17; N- 15; U- 12; J- 9; Q- 5; V- 4; Y- 3; K- 1.

First two letters: RE- 49; CO- 46; PR- 43; BE- 32; IN- 31; DE- 30; PE- 23; BU- 21; MA- 21; EX- 19; PO- 19; ST- 19; SE- 17; LE- 15; SA- 15; EN- 14; MO- 14; FI- 13; FR- 13; WI- 13; AL- 11; GE- 11; PA- 11; SH- 11; SI- 11; SU- 11; AD- 10; CH- 10; DI- 10; DR- 10; RA- 10; TH- 10.

First three letters: PRO- 33; CON- 22; COM- 12; IND- 11; BEC- 10; BUS- 10; PER- 10.

First four Letters: BECA- 8; CONS- 7; INTE- 7; SHOU- 7; CONT- 6; WITH- 6; PROB- 5; CONG- 5.


Last Letter: -S 183, -E 141; -N 126; -D 116; -Y 101; -T 86; -G 82; -R 80; -L 40; -H 13; -C 9; -M 8; -F 6; -K 3; -P 2; -W 2; -A 1; -O 1.

Last two Letters: -ED 89; -ON 83; - NG 81; -ER 67; -NT 43; -ES 34; -RS 34; AL 33; -LY 33; -SS 27; -LE 24; -TS 24; -RY 22; -TY 19; -CE 18; -RE 18; -AN 18; -NS 18; -SE 17; -VE 17; -EN 13; -ND 11; -GE 11; -ST 11; -TE 10.

Last three letters: -ING 81; -ION 67; -ERS 33; -ENT 32; -TED 28; -ESS 27; -RED 13; -NCE 13; -SED 12; -TER 12; -VER 12; -ONS 12; -NTS 12; -HER 11; -URE 10; -DER 10.

Last four Ltters: -TION 60; -NESS 16; -DING 15; -MENT 15; -THER 10; -ATED 9; -TURE 9; -ALLY 9; -KING 8; -TING 8; -RESS 8; -ANCE 7; -LING 7; VING 7; -ERAL 7; -SION 7; -ENTS 7; -ARED 6; -SSED 6; -TIVE 6; -HING 6; -WING 6; -ICAL 6; -TTER 6; -EVER 6; -TORS 6; -IONS 6; -CTED 5; -ITED 5; -FORE 5; NING 5; -ONAL 3; -ICAN 5; -RDER 5; -DENT 5.

The table can be used in many ways. Thus by frequency alone, -ED (89), -ING (81), and -TION (60), would more than likely be found among the most used endings of two, three, and four letters, respectively.

Affixes can also be identified by comparison with words, or with each other. For example, various letters of -TION can occur as commonly used single words: I, IT, TO, ON, NO, NOT. And again, the prefix IN-, and the suffixes -TION and -ING, have certain letters in common.

To illustrate how to use the table with the present type of cipher, take the most used four letter endings in the cryptogram, fhie (12-81-99-125), and fheo (10-21-39-50-108). A comparison of the symbols in common to these endings indicates that they are -TION and -TING, respectively.

Checking with two letter groups, we find hf (29), and fi (16-53-78-130), which would thus he IT and TO. Also he-, which by this supposition would he IN-, is used five times (12-15-40-69-81) as an initial combination.

It is probably unnecessary to mention that as each new value is discovered it must be set down in the keys, which can thus be gradually reconstructed. For example, the deciphering key now stands thus:

Cipher:  a c d e f h i n o r s t u w
Message: A - E N S I O - C - M H R -
                 T   U   G

This key is now sufficiently developed to attempt the translation of some of the longer words, or even of parts of the message. But in substituting in this way from an incomplete alphabet it must he remembered that a given symbol can possibly signify other letters besides those already discovered. At this stage of the game, context is also, a valuable asset in deciding letters and words. The first few words of the cryptogram, deciphered herewith, afford an excellent illustration of these points.

C a u h i i f  a u d  f t d  s d f t i r f   i t h o t   t a c d   w d d e   r d c h f d r  ...
- A R I O O S  A R E  S H E  M E S H O - S   O H I C H   H A - E   - E E N   - E - I S E -  ...
        U U T         T          T   U   T   U     G                                  T
        W W                          W       W
V A R I O U S  A R E  T H E  M E T H O D S   W H I C H   H A V E   B E E N   D E V I S E D

By a continuation of the same processes the reader, should he be curious enough, can decipher the entire cryptogram.

It is important to observe, however, that all the deciphering already done has been without knowledge of, or reference to, the key phrase. And the results would have been the same had the key phrase been in a foreign language, or even a mere jumble of letters.

But while it is thus possible to resolve the cipher and subsequently reconstruct the key phrase, the latter can also, notwithstanding Poe's statements, prove a valuable aid in decipherment.

For, as a general rule, symbolic values that would produce unpronounceable or impossible combinations in the key phrase can be rejected. Also, if the language of the key phrase is known, it should often be possible to reconstruct the latter—since for the most part it will consist of such symbols as occur in the cryptogram—even when only a few scattered words of the message have been deciphered.

For example, the new values just found, c=V, r=D, and w=B, fit in well with those already determined in the subjoined partially reconstructed key phrase, where it should require no great stretch of the imagination to supply the missing letters.

a w o r d - o t h - - - s

e i - - u f f i c i - - -

Such is the cipher that Poe juggled so dexterously to the mystification and confusion of his readers; a cipher that might well be likened to a chameleon, since its symbols and groups are capable of various meanings, taking on shades most suitable to the complexion of the communicated intelligence.

Unlike the simple substitution cipher, nearly every group here, even if the key phrase is at hand, is a miniature problem in itself, that must be resolved also with the major problem of the whole cryptogram in mind.

Now it is set down in the annals of cryptography that this chameleon of ciphers has been resolved by one Pierre Antoine Berryer, French political orator and chief of the Legitimists, and one Edgar Allan Poe, American poet, writer of tales, essayist and critic.

All of which should he an incentive for the reader to try his own skill at the two key phrase ciphers appended hereto.

The first of these tells how to locate a certain secret place of concealment. The key phrase is one of twenty-four letters (not "Le gouvernement provisoire" ) in a foreign language, adapted to our alphabet by combining I-J and U-V.

In the hiding place disclosed by the first cipher are to he found two objects, together with a note of instructions as to their disposal. This note, enciphered by means of a twenty-six letter English key phrase, constitutes the text of the second example.

Remember, now, Poe and Berryer solved the key phrase cipher!

Can you?


Ye ufll eadyfau Igdlyedvyfe fe vol eddalu
uffuufde difal vol tyidddh dv Eddreluu, y oyv
deeyulevdllh aofe d uledlv uodyer, uyuvdev
uyg yeeolu ye d dyrov-tyel edfl vol uolddoldu
fe vol eeyrov ye olddtuye ulayel voldl
itdnfelu. Uyvoye vol edhov, uyuetfulu ih
vol Ifalllev fe voyu uodyer, y efaeu vol uledlv
uoyeo, odayer udyale Ih edvold vf oyu rddal,
vole vadelu idee vf cduvle aofe 11, deu uytt,
du y dl eldvdye, elald dltldul 11 aevyt y uyl
iluyul oyl.


Oo rcn ro em orco oytss loolrent ayn onltna
estln nyntn oo yollnc ayoo eooars tcl ayn
Irconooorc ro yoo otaynto orssono tcl Itoeno,
o Irnoons yoe ar stm ayn staant nerc ayn
ootn, tcl ar loolytttn ayn ootoa ocar yoo rnc
yntl. Or enoa oytss yn oyonsl ayn enertm
ro yoo tclnoarto, tcl oettn yoeonso aynot


In response to popular request that Solving Cipher Secrets appear more often, it has been decided by the powers that he to make the department a weekly event, at least, for a probational period.

Many items of interest will be found in the weekly department, including fuller discussions of "Ciphers From Our Readers," than have heretofore been possible. These weekly installments are not expected to interfere with the feature length articles, which will continue, as now, to be published at intervals of several weeks. We would be glad to hear what you think of the weekly plan.

Coming now to the December 18 ciphers, No. 1 used an alphabet constructed on the key word SPHINX. Note how the letters are written by columns into the numerical key block, where 11=S, 12=Q, and so on.

    S P H I N X
    A B C D E F
    G J K L M O
    Q R T U V W
    Y Z -

  1 2 3 4 5 6 7 8 9
1 S Q B Z K I U M F
2 A Y J H T D N V O
3 G P R C - L E X W

The message in No. 1, "True skill at deciphering consists not only in seeing what is before our eyes, but also in foreseeing what is yet to come," was enciphered with the key word VEIL (28-37-16-36).

No. 2 used the straight Nihilist key—see FLYNN'S WEEKLY for March 28, 1925. The key word was ONE (34-33-15), and the message, "It was Benjamin Franklin who said: 'Three may keep a secret if two of them are dead.' "

No. 3 (Robert A. VanderPyl) was a modified Gronsfeld—see FLYNN'S WEEKLY for June 6, 1925—using subtraction as well as addition. Each cipher group comprised two words of the message, which was: "It shouldn't be beyond the ingenuity of man to compose an insoluble cipher, hut if this can he solved I should admit that it cannot he done." Each period of twelve letters was enciphered as follows:

            (plus)      (minus)
           ┌───────┐   ┌───────┐  
Key:     0 1 2 3 4 5 0 1 2 3 4 5
Text:    I T S H O U L D N T B E
Cipher:  i u u k s z l c l q x z

The recurrent group tg (67-19=48) suggested 12 as a possible period, tgc being easily recognizable as THE with the key, minus 0-1-2.

No. 4 (T. P. Spence) used the following simple numerical alphabet, constructed from the first five symbols preceding the cryptogram proper, as follows: (B) read numbers 1 to 26 Backwards; (23) start at 23; (1) begin with 1st letter of alphabet as 23; (5) skip 5 letters; (6) take 6th letter. and so on through the alphabet.

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

The last two symbols of the key (1-3) meant that only the first and third numbers of each cipher group were to he read, all other numbers being nulls. The message: "Geo. I will be off at seven thirty. T."

No. 5 (M. Walker) was an excellent example of the concealment type, every third letter being significant, thus HAMAN DVILE..., the message being: "Have two suit cases in the alley at twelve thirty."

Observe that, through the proper selection of nulls, the cryptogram took on the form of a meaningless jumble of words.

No. 6 (Wilburn N. Milhauser) requested, in simple numerical cipher: "Please publish some easy ciphers with instructions how to solve them." We believe Wilburn will find something to his taste in the weekly issues of the department.

The order of transposition in No. 7 (Arthur Bellamy) is based on the order of letters in the message, as here shown by the shortest (7b) "of his four cryptograms, which conveyed the message: "Prepare to attack Richmond Thursday."

To encipher, take the letters in alphabetical order, -writing as the cryptogram whatever letters follow them in the message. Thus, the letters after the A's are RTCY, which thus become the first four letters of the cryptogram. Next, take the letters after the B's, if any; and so on. The first letter of the message is taken as the next following the final letter.

To decipher, arrange the letters of the cryptogram in alphabetical order (1), writing the cryptogram below, as at (2), whereby each letter in line (1) has below it the letter next following it in the message.

(1) A A A A C C D D E E H H I K M
(2) R T C Y K H T A P T M U C R O
  N O O P P R R R R S T T T T U Y
  D A N R A E E I S D O T A H R P

Below Y—the last letter of the present message as can be indicated by a special key if desired—is found P, the first letter of the message. Below the first P is R, the second letter of the message; and so on.

An autokey transposition cipher is something out of the usual run. Mr. Bellamy's other messages were: (7a), "Destroy all British commerce. Sink without trace." (7c), "This cipher has no key and gives its own solution." (7d), "At zero hour July 6 make quick drive for West Pond bridge."

The solutions to the ciphers in this issue, including the translation of the Poe cipher, will be printed in the next article.


Space permitting, we had intended to publish all the entries in the "Radio Cipher Contest" recently promulgated in the October 2 issue of FLYNN'S WEEKLY. AS some of these are quite lengthy, however, we regret that we must limit ourselves to the winning entry.

It was unnecessary to judge the contestants by the dates of their entries, for, while all gave the correct translation of the cryptogram, only one solution was received that otherwise fully met all the requirements.

Mr. Castle himself acted as final judge, awarding the radio set to Mr. Charles W. Clapp, whose entry, in full, was as follows:


I have solved the Radio Cipher given in the October 2 issue, and give my translation of the message as follows: (Note Mr. Clapp here appends the correct translation).

The letters of the message were first changed into numbers according to the alphabet given below (10 being used as a word spacer). The key used was likewise changed into numbers and written continuously below the numbers of the message in the manner of the Nihilist Cipher. The key used in this case was, "Castle," followed by a word spacer. The separate numbers of the message were then added to the key numbers immediately below them to form the cipher.

                        29 B
                20 C    30 F
I A     11 D    21 G    31 J
2 E     12 H    22 K    32 N
3 I     13 L    23 O    33 R
4 M     14 P    24 S    34 V
5 Q     15 T    25 W    35 Z
6 U     16 X            
7 Y    
        10=word, space

To solve a cipher of this type, it is first necessary to determine the length of the key. To do this, tabulate the recurrent groups and the intervals between them, according to the Kasiski principle. On examining the intervals, with their factors, one factor usually predominates, which can usually be taken to the length of the key. In this case it was seven. Copy out the cipher, arranging it in rows of seven each, one under the other. Now all of the figures in each column would have been enciphered by the same key number. In such a long example as this it was more than likely that in each of the seven columns, the figure I as part of the message would occur af least once. Following this assumption I determined the lowest number in each of the columns and took as the key number for each column one less than the lowest number.

In this case the numerical key 20-1-24-15- 13-2-10 was arrived at. The key was then subtracted from the numbers of the cipher in the usual way, giving a simple substitution cipher of the numeral type. A frequency table was then prepared for this version of the cipher and the alphabet used was determined by the use of this, and by trial and error.

Since the frequency of "10" was too high to represent "e" it was allotted the ro1e of word spacer. The cipher then became a simple substitution cipher with normal word divisions and was easily deciphered.


114 Brook Street, Sandwich, Ontario, Canada.

October 4, 1926.

Besides the above entry, three others were received, all of which, however, fell short of the requirements of Rule 2, which stated that an entry should consist of (a) a translation of the cryptogram, (b) a description of the cipher system, and (c) an explanation of the method used in deciphering it.

Mr. F. M. Walters, 902 Kennebec Stregt, Pittsburgh, Pennsylvania, gave the correct translation, but failed to include a description of the cipher system. Further, while his analysis revealed the several alphabets, it did not find that they were all derived from a primary alphabet by means of a key word. (Entry dated: October 2, 1926.)

Dr. G. A. Ferrell, 404 Realty Building, Bessemer, Alabama, also gave the correct translation, but he described the cipher as consisting of seven unrelated mixed alphabets. Dr. Ferrell, too, did not find the primary alphabet. (Entry dated: October 8, 1926.)

Mr. Fredrik Pilstrand, 684 Warren Street, Brooklyn, New York, submitted the correct translation, but without any description of the system. Mr. Pilstrand discovered the primary alphabet and the numerical key, but neglected to mention that the latter was "CASTLE-". (Entry dated: October 17, 1926.)

To this we will now briefly append our regular lists of solvers. The following answers were received to the September 4 ciphers:

Twenty-six answers, all told, were received to the October 2 ciphers as follows: