book cover
From FLYNN's October 2, 1926


Edited by M. E. Ohaver

HE cipher system treated in this article should be familiar to most readers of detective literature, and therefore to perusers of this magazine, through the fact that it was used by Arthur B. Reeve in his collection of tales, "The Treasure Train."

Reeve's book was published in 1917. But the cipher to which we refer has been in the limelight also on at least two previous occasions.

It was described in the third installment of a valuable and interesting article on cryptography by George Wilkes, that ran in The Cosmopolitan Magazine for February, March, and April, 1904.

And again it appeared in the chapter on ciphers in Cecil Bullivant's book, "Home Fun," published in 1910.

The present article, however, differs from all the above mentioned works in its presentation of the cipher in that it shows how to determine if an unknown cryptogram is in this system, and offers a ready method of solution.

In explaining this we shall employ a cryptogram, which, for the sake of a realistic setting, we will suppose to figure in a big pay roll holdup.

Following the commission of our supposed crime, let us further assume that a round-up of suspicious characters nets at least one man who is unable or unwilling to give a satisfactory account of himself.

On the person of our suspect is found a piece of paper, bearing the following characters, evidently a message in cipher:


Given a severe grilling, our prisoner refuses to explain the piece of writing. However, we still have a card up our sleeve. We can grille the cryptogram itself. Perhaps it will be more communicative.

A visual inspection of the writing, with it? paucity of vowels, and predominance of X's and &''s, is sufficient if i itself to show that the cipher is not one of the transposition class.

This conclusion may be verified, if desired, by fully performing the Hitt transposition test, given in FLYNN'S WEEKLY for September 4. But first we must append a frequency table of the seventy-eight characters of the cryptogram.


In a transposition cipher the total frequencies of the groups AEIOU, LNRST, and JKQXZ, will not vary more than 5% one way or the other from the normal averages of 40%, 30%, and 2% respectively.

Here the group LNRST tests normally. But AEIOU falls short of its 35% minimum, and JKQXZ is far in excess of its 7% limit.

   A  4   L  2    J  1
   E      N  5    K  2
   I  6   R  3    Q  1
   O      S 10    X  9
   U  3   T  2    Z
   ————   ————    ————
     13     22      13
(16.6%) (28.2%) (16.6%)
(Percentages of 78 characters)

Already our cipher is beginning to talk. Even so soon as this has it yielded the information that it is not of the transposition class. If we keep right after it, maybe we will secure further admissions, or even a full confession.

Now that a transposition cipher seems out of the question, we can consider the possibility of a substitution cipher. And here again the character & engages our attention.

Assuming tentatively that our cipher is one in which each character represents a single letter or character in the original text, this & may mean that a 27-character alphabet is being used, so that an additional character will be available for some additional sign, as, for instance, a space between words.

That the cipher is of the simple substitution type would seem to be indicated by the 48% total of the five most used characters, S-X-&-I-N, in the cryptogram. In average English text the five most used characters, E-T-A-O-N, or (when a spacer is used) space-E-T-A-O, comprise approximately 45% and 50% respectively of the whole number of characters.

However, we must not therefore rush headlong into the conclusion that our cipher is of the simple substitution type, either with or without a spacer. For there are many obstacles in the way of such an assumption.

For instance, if we suppose that S=space, the twenty-one consecutive characters from the 46th to the 67th of the cryptogram would represent a single word. On the other hand, if S=E, these same characters constitute a series in which E is not used.

Similar observations can be made upon X, &, I, and other frequently used characters, one of which must, in all probability, represent the space or the E in a simple substitution cipher.

Words of extreme length, and long series of letters without E, are, of course, entirely possible. The above conditions, however, are not strongly indicative of a simple substitution cipher.

Additional proof to this effect, however, is somewhat to be desired. And it is readily afforded by testing the cryptogram with the Kasiski principle, fully described in FLYNN'S WEEKLY for August 7.

According to Kasiski, if a cipher is of the simple substitution type, recurrent groups will be rather numerous, but at irregular intervals. While in a multiple alphabet cipher using a fixed series of alphabets, the recurrent groups will be comparatively few, but at regular intervals in most cases.

Here is a list of the recurrent groups in the present cryptogram:

(Group)   (Location)
CX        12, 18. S&        21, 45, 67, 77. S&&BSNA   21, 67.
SN        25, 35, 71. AF        32, 62.

In this list the majority of intervals— these are the differences between any two numbers locating any given group—are multiples of two, which, according to Kasiski, would seem to indicate the use of two cipher alphabets.

To the experienced eye, however, this table speaks volumes, and, incidentally, illustrates that the famed Kasiski principle can be put to other uses than to distinguish between single and multiple alphabet ciphers, or to determine the number of alphabets in a multiple alphabet cipher.

First, all the groups in the table, with but one exception, consist of an even number of characters. In a multiple alphabet cipher recurrent groups of an odd number of characters are common. Even in the present short cryptogram, if it were in a two-alphabet system, there could easily be a couple of three-letter recurrent groups.

Again, the majority of groups in the present cryptogram occur on the odd numbered characters. In a multiple alphabet cipher groups can recur indifferently on both even and odd numbers.

Finally, the intervals between the recurrent groups are here multiples of two. This is, of course, also possible in multiple alphabet cipher, especially if it uses two alphabets.

But, to sum up these peculiarities, when the majority of the recurrent groups in any cryptogram consist of an even number of characters (as 2, 4, 6 ...). occurring on the odd numbered characters of the cryptogram, with intervals that are multiples of two, the indication is a cipher based on digraphs in which each pair of text letters is represented by a group of two characters in cipher.

In the straight alphabetic type of digraph cipher there will be (26×26=) 676 different combinations. The most used digraphs, in the order of their descending frequency (see FLYNN'S WEEKLY for January 23, 1926), are TH, HE, ER, IN, AN, ON, and so on.

In a cipher using an additional character, as a spacer, there will be (27×27=) 729 different combinations. If a space (-) is used, on account of its high frequency, the most used digraphs in their descending order will run somewhat as follows: E- ; TH; -T; HE; S-; T-; ER; and so on.

In a simple substitution cipher the most used, or one of the most used, characters will ordinarily be the substitute for E, the most used letter, or for the space, when the latter is used.

Similarily, in a fixed digraph cipher one of the most used combinations of characters will ordinarily represent TH in the straight alphabetic type, or E- where the space is employed.

A tabulation of the characters of the present cryptogram by twos, GV, XW, SG, KR ..., reveals that the most used pairs are: S&, which occur four times; SN, used three times; and &B, twice. All other two-character groups are used once only.

In a digraph cipher employing a word-space, as this one presumably does, it may be assumed with some degree of certainty that one or more of the most used cipher groups (here S&, SN, and &B) represent one or more of the most frequently occurring digraphs, E-, TH-, T, HE, and so on.

Owing to the peculiar construction of the present system it is unnecessary to use a general method for digraph ciphers. As will soon be evident, to correctly assume the identity of but one cipher group (for example, that SN=E- ) is sufficient to determine the adjustment of the two alphabets on which the cipher is based.

But while this cipher is thus capable of solution by identifying only one digraph, due to an amazing oversight of its constructor, it can be solved by a still more direct method.

To explain this let us consider for the moment that, as the Kasiski test seemed to indicate, the cipher is of the multiple alphabet type, using two alphabets alternately. In this case, if the straight alphabetic order has been used, both alphabets can easily be determined by identifying only one character in each.

In the following frequency table, alphabet No. 1 comprises the odd numbered characters, G, X, S, K ..., of the cryptogram; and alphabet No. 2 includes the even numbered characters, V, W, G, R, and so on.

Alphabet No. 1: 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  &
   Frequencies: 2        1     4  2        1  1  2  1                 10 2  1  1  1  6  1     3

Alphabet No. 2: 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  &
   Frequencies: 2  2  2  1        1     6     1     2  5     1  1  3        2  1  1  3        5

In alphabet No. 1 the most used character is S, employed ten times; and in alphabet No. 2 it is I, used six times. These two characters have a total frequency of 16, or 20.5% of 78, the whole number of characters in the cryptogram,

Now if S and I happen to have the same significance in their respective alphabets, the above percentage would seem to indicate that our cipher uses a spacer. For the word-space has an average frequency of 18% in average English text; while E, When a spacer is not used, averages only about 12%.

Assuming, therefore, that spacers are employed, the most logical place for them, in a straight arrangement of the alphabet, is after Z. Here are the two cipher alphabets adjusted on the supposition that S and I are space substitutes:

 Text Alphabet: 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  —

Alphabet No. 1: T  U  V  W  X  Y  Z  &  A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S 
   Frequencies: 2  1  1  1  6  1     3  2        1     4  2        1  1  2  1                 10

Alphabet No. 2: J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z  &  A  B  C  D  E  F  G  H  I  
   Frequencies:    1     2  5     1  1  3        2  1  1  3        5  2  2  2  1        1     6

Observe that, by this arrangement, large numbers in both cipher alphabets are placed in conjunction with the letters of highest frequency, E-T-A-O-N, in the text alphabet. Also that small numbers come opposite those of lowest frequency, J-K-Q-X-Z. These are very favorable indications that we are proceeding upon correct assumptions.

The tentative alphabets can now be tried on a short section of the cryptogram. For convenience this is taken by twos, as shown at (a):

(b) OM EN -Y SI B- RU EI -D ...
(c) MO NE Y- IS -B UR IE D- ...
(d) Money    is buried ...

In the first pair, GV. the first letter, G, signifies O in alphabet No. 1; and the second letter, V, stands for M in alphabet No. 2. GV thus equals OM . Similarly, the second pair of characters, XW, signifies EN; and so on.

At first sight the translation in line (b), OMEN-YSIB-RUEI- D ..., does not look very promising. But observe, at (c), the miracle that results by merely reversing the pairs in line (b). The translation at (d) is now clearly evident. Here is the whole message: Money is buried under the big stone in cellar of the old house near the river.

Here, then, is a cipher that—by using two alphabets and reversing the pairs — seems to combine the principles of substitution and transposition. As a matter of fact, however, these results are identical with those attained by the altogether different system referred to at the beginning of this article.

The cipher in question is based on digraphs employing the elaborate table illustrated on the following page.

To use the digraph cipher, first select any two desired key letters, as, for instance, the key J-T of the present example. The alphabet beginning with the first key letter is now written across the top of the table, and that beginning with the second key letter is written down the side. The character & is in both cases written after Z.

To encipher any message, as, for instance, that one already used in connection with the previously explained method, first arrange the letters of the text (d) in pairs as shown at (c), and then find the substitute for each digraph, given at (a), as follows:

Thus MO, the first digraph, is found in the G row of the table, as indicated by the key alphabet at the side, and in the V column, as indicated by the key alphabet across the top. Accordingly MO is represented by GV in cipher. In the same way, the substitute for the second digraph, NE, is XW; and so on.

Here, seemingly, we have a cipher of digraphs. But actually, when resolved to its simplest elements, it consists of two simple alphabets. Why is this?

Simply because all the digraphs in any column of the table have the same first letter, and the cipher substitutes for these digraphs will have the same second character, taken from the key alphabet across the top of the table. Thus AA, AB, AC, AD ... will, in the J-T key, be represented in cipher by the groups TJ, UJ, VJ , W J ... respectively.

Likewise, all the digraphs in any horizontal row of the table have the same second letter, and the cipher substitutes for these digraphs will have the same first character, taken from the key alphabet down the side of the table. For instance, AA, BA, C A, DA ... will, in the present key, be represented by the cipher groups TJ, TK, TL, TM ... respectively.

The present cipher, which, incidentally, is reputed to have been used by "the War Office of a well known Continental Power," is thus merely another example of a complexity that is more apparent than real.

J K L M N O P Q R S T U V W X Y Z & A B C D E F G H I
T aa ba ca da ea fa ga ha ia ja ka la ma na oa pa qa ra sa ta ua va wa xa ya za -a
U ab bb cb db eb fb gb hb ib jb kb lb mb nb ob pb qb rb sb tb ub vb wb xb yb zb -b
V ac bc cc dc ec fc gc hc ic jc kc lc mc nc oc pc qc rc sc tc uc vc wc xc yc zc -c
W ad bd cd dd ed fd gd hd id jd kd ld md nd od pd qd rd sd td ud vd wd xd yd zd -d
X ae be ce de ee fe ge he ie je ke le me ne oe pe qe re se te ue ve we xe ye ze -e
Y af bf cf df ef ff gf hf if jf kf lf mf nf of pf qf rf sf tf uf vf wf xf yf zf -f
Z ag bg cg dg eg fg gg hg ig jg kg lg mg ng og pg qg rg sg tg ug vg wg xg yg zg -g
& ah bh ch dh eh fh gh hh ih jh kh lh mh nh oh ph qh rh sh th uh vh wh xh yh zh -h
A ai bi ci di ei fi gi hi ii ji ki li mi ni oi pi qi ri si ti ui vi wi xi yi zi -i
B aj bj cj dj ej fj gj hj ij jj kj lj mj nj oj pj qj rj sj tj uj vj wj xj yj zj -j
C ak bk ck dk ek fk gk hk ik jk kk lk mk nk ok pk qk rk sk tk uk vk wk xk yk zk -k
D al bl cl dl el fl gl hl il jl kl ll ml nl ol pl ql rl sl tl ul vl wl xl yl zl -l
E am bm cm dm em fm gm hm im jm km lm mm nm om pm qm rm sm tm um vm wm xm ym zm -m
F an bn cn dn en fn gn hn in jn kn ln mn nn on pn qn rn sn tn un vn wn xn yn zn -n
G ao bo co do eo fo go ho io jo ko lo mo no oo po qo ro so to uo vo wo xo yo zo -o
H ap bp cp dp ep fp gp hp ip jp kp lp mp np op pp qp rp sp tp up vp wp xp yp zp -p
I aq bq cq dq eq fq gq hq iq jq kq lq mq nq oq pq qq rq sq tq uq vq wq xq yq zq -q
J ar br cr dr er fr gr hr ir jr kr lr mr nr or pr qr rr sr tr ur vr wr xr yr zr -r
K as bs cs ds es fs gs hs is js ks ls ms ns os ps qs rs ss ts us vs ws xs ys zs -s
L at bt ct dt et ft gt ht it jt kt lt mt nt ot pt qt rt st tt ut vt wt xt yt zt -t
M au bu cu du eu fu gu hu iu ju ku lu mu nu ou pu qu ru su tu uu vu wu xu yu zu -u
N av bv cv dv ev fv gv hv iv jv kv lv mv nv ov pv qv rv sv tv uv vv wv xv yv zv -v
O aw bw cw dw ew fw gw hw iw jw kw lw mw nw ow pw qw rw sw tw uw vw ww xw yw zw -w
P ax bx cx dx ex fx gx hx ix jx kx lx mx nx ox px qx rx sx tx ux vx wx xx yx zx -x
Q ay by cy dy ey fy gy hy iy jy ky ly my ny oy py qy ry sy ty uy vy wy xy yy zy -y
R az bz cz dz ez fz gz hz iz jz kz lz mz nz oz pz qz rz sz tz uz vz wz xz yz zz -z
S 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- --

Intended originally as an intricate system based on digraphs, it turns out to be, in effect, one that employs two simple alphabets, combined with a reversal of pairs of letters. The table of digraphs is an altogether unnecessary accessory. The cipher can just as well be operated from the two simple alphabets, both of which can be readily reconstructed from memory.

But it is time now to return to our "prisoner," who has almost been forgotten while we were busy unraveling the mysteries of his cipher.

Perhaps, at this very moment, he is pondering on the efficacy of the cipher, and the advisability of admitting his complicity in the holdup.

Too late!

Even now the police are searching for the old house along the river. And perhaps within the hour they will have recovered the stolen money.

Besides, in the meantime, our prisoner has been identified. And a search of his room has uncovered two more messages, apparently in the same cipher.

Perhaps these cryptograms will contain further information about the robbery. Maybe they will afford a clew as to the whereabouts of the other criminals.

Do you want to aid in their capture?

Then decipher one or both of the following two cryptograms, and send your solutions, posthaste, to this department!






Attention, fans!

Here is an opportunity to profit by your knowledge of cryptography.

The first solver of cipher No. 3 in this issue, who complies with the rules given below, will be given a genuine Superiorflex three tube radio set, free of all cost.

This unprecedented offer is made possible through the generosity and interest of Mr, C. A. Castle, of the Electrical Research and Manufacturing Company, Waterloo, Iowa, makers of the set

The set offered is a regular one hundred dollar model, that will retail the coming season at seventy-five dollars. It is finished in an attractive walnut cabinet, is wonderfully efficient, and of the highest class material and workmanship.

Win this set and you won't lack entertainment. It will bring in practically every large broadcasting station in the country with perfect clearness and plenty of volume to operate a loud speaker.

The contest will be governed by the following simple rules:

(1) Any one is eligible to compete for the prize. It is not necessary to be a subscriber to FLYNN'S WEEKLY.

(2) An entry must consist of a translation of the cryptogram, a description of the cipher system, and an explanation of the method used in deciphering It. These descriptions can be as long or as short as you please.

(3) Any entry, to he accepted, must he postmarked not later than one month from the date of this issue of FLYNN'S WEEKLY.

(4) The winner of the radio set will be that person submitting the first correct entry, priority being determined by postmark.

(5) In case two or more correct entries mailed on the same date are tied for first place, the one accompanied by the best explanation of the method of solution will be declared the winner.

(6) Address all entries: Radio Cipher Contest, FLYNN'S WEEKLY, 280 Broadway, New York, N. Y.

A complete exposition of Mr. Castle's cipher, together with a list of solvers, if any, will be printed in an early issue of this department.

Here is the Radio Cipher, fans! Now get busy!

CIPHER No. 3 (C. A. Castle).



Not so long ago we received a cryptogram that turned out to be in a modification of the digraph cipher described in this issue. This cipher is included somewhere in the following collection. Find it, and see if you can solve it. It will further demonstrate the practicability of the scheme of analysis outlined in the present article.

The following letter came with No. 4:


I submit the following cipher. Like all the other amateurs, I think my cipher is undecipherable to those who do not know the secret of its formation.

R. M. PACKARD, Boston, Mass.

CIPHER No. 4 (R. M. Packard).


The next one hails from Louisville, Kentucky. See what you can do with it.

CIPHER No. 5 (Paul Napier).


No. 6 is by Mr. J. Levine, Los Angeles, California, who has sent us a large number of his original and ingenious constructions, some of which have been published in previous issues.

The present cipher is so devised that it is capable of conveying two entirely different messages at the same time, all of the numbers being used for both messages.

Sir Francis Bacon's biliteral cipher is one of this general nature, as is another and simpler system also described by this author in his " Advancement of Learning. " A double writing of this kind has the advantage that the supposed message could be explained, to divert suspicion, leaving the real message still secure.

Can you unravel either or both the messages conveyed in Mr. Levine's cipher?

CIPHER No. 6 (J. Levine).

24   5   30½   10½   20   19   28
3   7l½   16½   27   4   22½   2l½
28   3. 

To round out the list for this time we are appending an example of the kind of cipher, or crypt, employed as a form of puzzle by the National Puzzler's League.

Ciphers, or, more exactly, cipher-puzzles, of this type are regularly printed every Sunday in the Red Magic section of the New York World and other newspapers, as well as in the Enigma, the official organ of the Puzzler's League, for any information about which our readers can address the league's secretary, Mr. Lewis Trent, 1391 Jesup Avenue, New York, N. Y.

To cipher fans the crypt is better known as the simple substitution cipher. A literal alphabet is always used in crypts, any given letter of the message always being represented by the same substitute in cipher, and any given cipher substitute fixedly signifying but one certain letter in the message. Normal word divisions are observed.

To make these crypts more difficult of solution it is customary to employ unusual wordings of the messages. Thus, in our own example, every word consists of six letters, and no word contains the same letter more than once. And yet the message is a grammatically correct sentence.

Come on now, all you puzzlers! Solve this crypt, and send us your solution. If you solve it, you can read the following question: What did the HGSDCI find in the AWGUSO? But you may not be able to answer it.



Explanations of all the ciphers in this issue, excepting No. 3, will be printed in next Solving Cipher Secrets. In the meantime, fans, do your best, and let us know how you make out.


The first of the September 4 ciphers was of the transposition class, Nihilist type, described in detail in FLYNN'S WEEKLY for March 20, 1926.

The key word was VOLSK, which gives the numerical key 5-3-2-4-1. The incident referred to in the message, below, is one of the many attempts made by the Nihilists to assassinate Alexander II:

The mine laid on the railway embankment near Alcxandrovsk failed to explode and the Tsar passed over the spot immjured.

No. 2 was a Confederacy cipher, being a copy of a cryptogram actually sent by Jefferson Davis to General E. K. Smith during the Civil War.

The cipher system used was the Vigenère chiffre carré, described with a method of solution in FLYNN'S WEEKLY for February 20, 1926. In the following translation the words of the message that were in cipher are capitalized.

Montgomery, 30th.


Shreveport, La., vi a Wi.:

What are you doing to execute the instructions sent you to FORWARD TROOPS TO EAST SIDE OF TH E MISSISSIPPI? if success will be more certain you can substitute WHARTON'S CAVALRY COMMAND For WALLER'S INFANTRY DIVISION, by which you may effect A CROSSING above that part OF THE RIVER patrolled by the LARGER CLASS OF GUNBOATS.


The assumptions that JN VAS and HJ OPG were OF THE would give the key letters VICTO and TEVIC respectively, which combine into the larger key fragment TEVICTO.

The interval between the groups providing the identical key letters VIC is 75, suggesting a key of 5, 15, 25, or 75 letters. A five letter key is impossible, since already seven key letters are discovered. And it is unnecessary to look further than a 15 letter supposition, because TEVICTO works satisfactorily at intervals of 15 letters. Needless to add, the whole key, COMPLETE VICTORY, is readily found.

Cipher No. 3 (Rudolph O. Casperson) was also a Vigenère, with the key NEVERTHELESS, the message being: "Believe me, Solving Cipher Secrets is some department."

There were a number of recurrent groups here, but, as it chanced, they were all accidentals—see FLYNN'S WEEKLY for August 7, 1926—and were thus of no volue in determining the period by the Kasiski method.

About the best way to solve this would be to look for THE, which would give VEM with OLQ. Trying both THE and VEM as parts of the key, the latter gives IPH (suggesting CIPHER ) ISS, and MEN at intervals of 12 letters from OLQ.

The solution to No. 4 (W. R. Bell) was: "This is the International Morse code which is known the world over." Capital letters were used for dashes (—) and small letters for dots (.), all the letters in any group forming the letter in the Morse alphabet, thus:

Cipher:   A   g i r l   t o   s e e (etc.)
Morse:    -   . . . .   . .   . . . (etc.)
Message:  T   H         I     S     (etc.) 

No. 5 (Frank Dodd) was another of the International Morse ciphers. In this one the dot (.) was represented by a figure 5, 7, or 9; and a dash (—) by a figure 6, 8, or 0. The figures 1, 2, 3, 4 were used as follows: any one of them represented a space (*) between letters or words; any two of them, a comma; and any three of them, a period.

A capital letter was indicated by preceding the letter to be capitalized with K. Mr. Dodd's message: " We need fifty (K) Ford cars at once."

Message:   W      *  E  *  N    *  E
Morse:     . — —     .     — .     .
Figures:   9 0 8  4  5  2  6 7  1  5 (etc.)
Regrouped:  90845       26715    

Still another of the telegraph alphabet ciphers was No. 6 (Carlton Bell). In this one the letters that rested upon the line of writing—as, for instance, a, b, c, d, et cetera —stood for the dots of the Morse code. While the letters that crossed through the line of writing—g, j, p, q, y—stood for the dashes. In the script alphabet, the letters f and z also cross the line (as also does the italic f), and therefore represent dashes.

Mr. Bail's system reserved two letters to indicate endings of words; t being used for a dot in that location, and g for a dash. Mr. Bell's message: " Solving a cipher is like picking a lock; one uses no key. " A short example:

Cipher:   a n d   g y p   o p e n   d r i p
Morse:    . . .   — — —   . — . .   . . . —
Message:  S       O       L         V

The last of the September 4 ciphers, No. 7 (John Q. Boyer), was based on the following alphabet, in which each number stands for either of two different letters:

A  B  C  D    I  J  K  L  M  N    U  V  W
1  2  3  4    5  6  7  8  9  10   11 12 l3
E  F  G  H    O  P  Q  R  S  T    Y  X  Z

This alphabet is so arranged, and the words of the message so chosen, that each group of the cryptogram could represent two words. One series of these words (a) formed the message: " Do not send for Jim as I can get some bid on land near Lima"; but the other series (b), while all-dictionary words, did not form a sentence.

(a)  D O    N O T    S E N  D ...
     4—5   10—5-10   9—1-10—4 ...
(b)  h i    t i n    m a t  h ...

What Mr. Boyer has here attempted is a cipher with more than one translation. Mr. Boyer has also submitted the following, based on the same alphabet: 1-8-5-9  3-1-10  2-5-4  13-1-10-4  2-8-5-10-4  10-5-10-1.  8-5-9-1   9-1-10   8-5-9-10  4-5-1-8. Here only the letters above the numbers in the alphabet are used for the sentence; and those below the numbers for the word series. It's easier than it looks.

So much for the September 4 solutions. Now for the list of July 3 solvers. The list this time is short, but it means a lot. You will note that all three solvers conquered No. 1, which affords ample proof that it is possible to solve a transposition cipher by anagramraing, especially if some idea is had as to the nature of the message.

The order here is that in which the solutions were submitted:

As you will observe, no one climbed aboard No. 3, Mr. Winsor's free subscription offering. Nevertheless, we think all three members of the above triumvirate are to be congratulated.

And now for the list of May 22 solvers held over from the last issue of Solving Cipher Secrets.

In the list herewith, you will note that not a single cipher of that date escaped without casualties. The book cipher No. 1, went down to an ignominious defeat before the onslaught of the Holmeses and Lecoqs of FLYNN'S WEEKLY. Nos 2 and 4 seem to have afforded the most difficulty. All solutions are listed in the order submitted.

Keep your answers coming, fans! These ciphers are not always so hard as they look. Send in your solutions, and look for your name in the next list.