IT isn't always possible to do that which you intended. For example, it was our plan to present in this issue a cipher that at one time threatened to play a great part in the World War. But further investigation indicated certain weaknesses which unfitted it for our use.
So we are offering instead one that Poe dealt with. Edgar Allan Poe, by the way, was one of the greatest students of ciphers and cryptograms. There were few codes that could baffle him for long.
After you have read what Mr. Ohaver has to say about Poe's method, study the cipher that follows. Then see if you're as good a man as Poe. Be sure to study and apply the frequency tables in this issue. They are vital.
ENTION of the name Edgar Allan Poe usually conjures up memories of hauntingly melodious verse, of trenchant criticism, or of delicate pastels in prose. His grotesque tales of horror, imagination, and mystery are familiar to all, as are also those tales of ratiocination in which he exploits his character, the Chevalier Auguste Dupin, who, after Voltaire's Zadig, is the original detective of fiction.
However, it is not Poe the romancer, poet, or essayist that forms the subject of this article, but Poe in another and less familiar field, namely that of cryptography, where his work, even if not of a comprehensive scope, still displays the distinctive touch that characterizes him as a literary artist. It is not at all hard to understand why Poe should be attracted by the science of ciphers. It is just that sort of subject in which we should expect a peculiar genius such as his to revel with delight.
And we may therefore with tolerable certainty surmise that he felt its fascination even as a very young man, notwithstanding the fact that his earliest connection with this obscure branch of knowledge as revealed in his own writings is not until 1839, when he had reached the age of twenty-eight years.
In 1838 Poe settled in Philadelphia. And during the following year there appeared in Alexander's Messenger, a weekly newspaper of that city, a section conducted by Poe devoted to the solution of ciphers. This venture of Poe's is in all probability the original cipher department, but in any case it is the first section of the kind of importance in any periodical.
The newspaper in question was for a period of several months considerably taken up with hieroglyphic and cabalistic looking cryptograms that poured in on Poe from all parts of the country.
Many regarded the matter in the light of a humbug, inserted in the paper to give it a queer air for the purpose of attracting attention.
Others, not familar with the methods involved in the solutions of such problems, accused Poe not only of solving the ciphers, but of preparing them besides. It is hardly necessary to say that Poe, of course, solved all the ciphers in good faith.
Most of these cryptograms were written in what is ordinarily termed the simple substitution alphabet. In a cipher of this class a given letter of the alphabet is always represented in cipher by a certain letter, figure, or character. And conversely, a given cipher character is invariably the substitute for one certain letter of the alphabet and no other.
In 1840 Poe became editor of Graham's Magazine, to which in July of the following year he contributed an essay on cryptography. This is concerned primarily with an account of the key-phrase cipher, the most difficult of all the ciphers that he is known to have solved.
Poe prefers to leave us mystified as to his method of treating this cipher. A complete discussion of the system, however, and of what was probably Poe's method of solving it will form the basis of a future article of this series, in which you will be given a key-phrase cipher to solve for yourself.
We have now arrived at what is without doubt the most important of Poe's work along this line, namely, his immortal tale, "The Gold Bug." This was published as a prize story in the June, 1843, number of the Philadelphia Dollar Magazine.
It concerns itself with the recovery of pirate's treasure by means of a message in cipher, and may be cited as the classic example of the cipher in fiction.
The appearance of this story marked the popularization of the cipher in this field. Stories containing cryptograms immediately became all the rage. Many of these were palpable imitations of Poe. At the present time the cipher is an almost indispensable accessory to the writer of mystery and detective tales.
Having completed this general survey of Poe's work in cryptography, let us now turn to his method of solving the simple substitution cipher, and to the resolution of that variety in particular where divisions occur between the words.
Poe described his method of attack on this kind of cipher as a process of "collation and analysis." By this he meant that, beginning preferably with the short words of the cipher, he compared those groups which had characters in common.
Here it must be said that Poe was not the originator of the method he used. It had already been in common use long before his time as the usual method of solving such ciphers.
To further assist in the solution of ciphers by this process the editor of this department has prepared the accompanying tables of the fifty most frequently occurring words in the English language, compiled from tabulations of one hundred thousand words as used in average written English.
In Table I these words are arranged in the order of their frequencies, the numbers after each word indicating the average frequency per one thousand words. The, for example, occurs an average of sixty-three times per one thousand words. This means that about six words out of every hundred, or one out of every sixteen, will be the word the.
It will be observed that the first ten words in this table make up over one-fourth —or exactly 26.4%—of all the words used in average reading matter. And the entire fifty constitute nearly one-half—exactly 48.9%—of all such words. When it is considered that our language comprises some half million words this fact seems astounding. But these fifty words are for the greater part connective particles whose use, even in the tersest sentences, it is impossible to escape. Some of them may therefore be expected to be present even in the shortest of ciphers.
| TABLE I | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| the | 63 | it | 12 | at | 7 | from | 5 | been | 4 |
| and | 34 | was | 10 | we | 7 | are | 5 | would | 4 |
| of | 34 | is | 9 | on | 7 | all | 4 | she | 4 |
| to | 32 | will | 9 | he | 6 | me | 4 | or | 3 |
| I | 24 | as | 9 | by | 6 | so | 4 | there | 3 |
| a | 19 | have | 8 | but | 6 | if | 4 | her | 3 |
| in | 19 | not | 8 | my | 6 | they | 4 | an | 3 |
| that | 14 | with | 8 | this | 6 | had | 4 | when | 3 |
| you | 13 | be | 8 | his | 5 | has | 4 | some | 3 |
| for | 12 | your | 8 | which | 5 | were | 4 | any | 3 |
| TABLE II | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| a | 19 | an | 3 | all | 4 | been | 4 | there | 8 |
| I | 24 | as | 9 | and | 34 | from | B | which | 5 |
| at | 7 | any | 3 | have | 8 | would | 4 | ||
| be | 8 | are | 5 | some | 3 | ||||
| by | 6 | but | 6 | that | 14 | ||||
| he | 6 | for | 12 | they | 4 | ||||
| if | 4 | had | 4 | this | 6 | ||||
| in | 19 | has | 4 | were | 4 | ||||
| is | 9 | his | 5 | when | 8 | ||||
| it | 12 | not | 8 | will | 9 | ||||
| me | 4 | she | 4 | with | 8 | ||||
| my | 6 | the | 63 | your | 8 | ||||
| of | 84 | was | 10 | ||||||
| on | 7 | you | 13 | ||||||
| or | 3 | ||||||||
| so | 4 | ||||||||
| to | 82 | ||||||||
| we | 7 | ||||||||
In Table II the words are grouped according to their lengths, with alphabetical arrangement in each group to facilitate comparison.
To illustrate the application of these tables as well as the method used by Poe, suppose it is desired to decipher the short cryptogram subjoined in line (a).
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) | |
| (a) | [:. | $?]. | °)' | :)] | &†'; | -)☞] | )'* | ) | ]:†☞[ | [†';£. |
| (b) | --- | --- | -an | -a- | --n- | -a-- | and | a | ----- | --n--- |
| (c) | The | --se | -an | has | --n- | ea-s | and | a | sh--t | t-n--e |
| (d) | The | --se | -an | has | --n- | ears | and | a | sh-rt | t-n--e |
| (e) | The | --se | -an | has | -on- | ears | and | a | short | ton--e |
First of all it is possible to solve such a cipher by one or more of several methods. These various processes will from time to time be elucidated in these columns, but for the present we will limit ourselves strictly to the method under discussion.
In the present instance we may begin operations on the assumption that the eighth cipher group ) is probably a or I. If the latter, then the seventh group )'* , which has the same character as its initial, would more than likely be its. This, however would give the improbable sequence its I.
On the other hand, if we assume that ), equals a, then group (7) might be and, any, or are, giving the sequences and a, any a, and are a. The first of these is the most likely, providing a, d, and n as tentative values for three of the cipher characters. Substitute these throughout and the partially deciphered message will stand as in line (b).
Now group (4) -a- might be can, had, has, may, or was, the middle letter of all these being a. Has is the most likely, since the first character of the group : is identical with the second character of group (1) , which can thus be assumed to be the. Substitute the values e, h, s, and t, and the cipher appears as at (c).
Substitute all the letters of the alphabet successively in group (6), ea-s, and we have ears and eats as two probabilities. The latter is impossible, since in our alphabet t is already represented by [. Our cipher now stands as at (d).
Group (9), sh-rt, might be shirt or short. The latter seems more probable because the results by substitution in groups (5) and (10) seem more likely, as shown in line (e).
It is unnecessary to continue the solution further. The reader may, if he chooses, carry it out to a successful finish. The partially discovered alphabet now stands as follows:
| Normal 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 |
| Cipher Alphabet: | ) | * | . | : | ' | † | ☞ | ] | [ |
Armed with the above exposition of Poe's principle of collation and analysis, and further equipped with the modern frequency tables, you are now prepared to meet a real test in the cipher given below.
This is enciphered in one of Poe's original keys, as is also the shorter specimen above. Cipher alphabets of this sort are termed arbitrary as distinguished from the methodized alphabet used in the Augustus cipher in FLYNN'S for February 21.
In the methodized alphabet, you will remember, the discovery of a single letter will usually lead to the discovery of others, or even of the entire alphabet.
Not so with the arbitrary alphabet. In it each discovery is independent, and does not in itself lead to the identification of any other cipher characters.
Application of the frequency tables is up to the ingenuity of the decipherer. The different comparisons possible are practically unlimited. But the solution of a cipher should not be solely by means of words found in the tables.
This list of words is only intended as an aid in searching out the vulnerable points of the cipher. Incidentally, it would be wise for you to preserve these tables. They will be of constant value in solving other and different ciphers in subsequent issues.
Here is your test! It hides one of Poe's much discussed statements about ciphers.
What is it?
Decipher it. and then tell us if you agree with Poe.
18] .8()‡*) -5* 28 95†8 ;‡ 28o68¶8 ;45; 6; 6) *‡; Q?6;8 5* 85): ;46*3 ;‡ 6*¶8*; 5 98;4‡† ‡1 )8-(8; ](6;6*3 ]46-4 )4500 251108 6*¶8); 635;6‡* :8; 6; 95: 28 (‡?*†O: 5))8(;8† ;4S; 4?9S* 6*38*?6;: -5**‡; -‡*-‡-; 5 -6.48( ]46-4 4?95* 6*38*?6;: -5**‡; (8)‡O¶8
Here is a message enciphered in an ingenious key submitted by Mr. Hyman Wacks, Brooklyn, New York.
Owing to the fact that it is here given without any suggestions whatever as to its treatment, it is expected to offer more than the usual resistance against solution.
ECYN RGVA GI GH FUHQAN. UTCGF MA. XCIDXCHA UDDAMXD CH VGNID HUDGCHUR.
After you have deciphered the above it should be interesting to try to restore the complete cipher alphabet.
Apparently this alphabet has been constructed in an arbitrary fashion, but if you will look closely, you will discover that there is method in it after all.
It would be profitable for you to try to discover the principle on which the arrangement of letters in this cipher alphabet is based. Look for the answer in the next article of this series.
How did you make out with the biliteral cipher in our last article? If you were one of those who heeded the advice to consider everything with an eye of suspicion, you probably discovered the cipher lurking in the innocent looking cartridge belt.
You will remember that in the instructions how to decipher the biliteral cipher you were told that the a form of any given character or letter outnumbered the b form approximately two to one.
There were thirteen loops in the belt filled with cartridges, as compared with the seven unfilled loops. The correct deduction is, therefore, that the filled loops represent a characters, and the unfilled loops b characters in Bacon's biliteral alphabet.
Assigning these values, dividing into groups of five, and substituting the corresponding letters of the alphabet as per details of this cipher in FLYNN'S for April 25, gives the series: aaaab (B) ; aaaaa (A) ; aaaba (C) ; abbab (0) ; abbaa (N) ; revealing the concealed word BACON.
Perhaps you never before thought it possible for any one by a few deft and unsuspicious moves to form a cipher obscure to the uninitiated, but as plain as day to those in on the secret.
The Bacon biliteral cipher shows you just how this can be done.
Look in our next article for the solution of the Poe cipher and also for a complete exposition, including method of solution, of the famous Gronsfeld cipher.
Magazines must work a long time in advance. This paragraph was written on April 6, and we are closing Solving Cipher Secrets for May 16. Down to date the following readers have sent in correct answers to the message in the Nihilist code, which appeared in our issue of March 28. Many of them answered the challenge with messages of their own, which will be dealt with later.
