Text: Richard Bolton to Edgar Allan Poe — January 10, 1842

Pontotoc, January 10, 1842

Dear Sir,

A press of business at the close of the year has delayed my reply to your very complimentary letter of Nov. 18. I suspected too, that your succeeding number would contain a correct print of the two cryptographs of Mr. Tyler, and I intended attempting their solution which I thought possible.

Mr. Reneau has just loaned me the January number of your magazine my copy not having yet arrived and I find it unnecessary to delay any longer.

As to the process by which I effected the solution of Dr. Frailey’s cyper it is useless to repeat all my abortive efforts and guesses. Suffice it to say that from a comparison of various words reduced to syllables I ascertained the vowels and particularly e from its frequent use and u from its unfrequent occurrence. I and a from their being used per se. If representing “ed” a common terminal syllable in long words next became known. Having advanced thus far the cipher fi-fvti and nia-fvti representing reduce and produce gave me a thread to the labyrinth after which my progress was comparatively easy. ITAGI did not as you suppose materially aid me. Having thus many links of the chain the links still wanting were supplied in making sense of the article. Some of the words, I must admit gave me trouble. The grammatical sequence of long and short words incident to language was of much assistance to me. Candour compels me to add that had the letters been continuous and not divided into words I should probably have failed. I therefore pay homage to you as King of Secret Readers.

From the insight I have thus obtained into the principles of secret writing I am convinced that a single cypher where one or more characters are used invariably for the same letter is capable of solution. The key phrase cypher may however be rendered more difficult by a judicious interpollation of the remaining letters. This would puzzle one not having the key they would be stricken out as meaning nothing.

A plan has suggested itself to me which I flatter myself may be tried by Lord Bacon’s rule and not be found wanting. It is easily written and certainly read with the key and insoluble without it and how little the chance of guessing at the key it may be calculated by arithmetic. By this plan each letter of the alphabet represents each other letter according to a variable rule, which rule may be varied with different correspondents.

Suppose a cypher in which the third preceding letter is used instead of the proper letter as dila for good. This is a simple cypher whose key is No. 3. Again suppose the key to be 5.7.3.1; so that for each division of four letters, the fifth preceding shall be used. The cypher of good will then be bhlc. This is a much more difficult cypher, inasmuch as each letter of the alphabet may be represented by four other letters. Again suppose, after every twentieth letter, the key to be continually changed — for instance after the first twenty letters the key to be 13 14, 10, 8; after the key to be 8, 10, 6, 4, and so on successively. This will render it impossible to analyse the cypher.

My plan is of this character. Let the letters of the alphabet in their natural order represent 1, 2, 3, etc., to 26. Let word be the key in the present instance, that is 23, 15, 18, 4. Let the letters be written continuously and the key be changed after every twenty-fourth letter, the key of the first division to be the above key increased by four — 17, 19, 22, 3, and the key of every other division to be the above key increased by the corresponding number of the last letter of the last letter of the previous division.

For example, take the following:

The enemy has invested our works and is vigorously pushing the siege. We can hold out but three days longer when if not relieved we must surrender. The cryptograph will be: Here word, or 23, 15, 18, , 4 is the key. The movable to be applied to is d or 4 in the first division of 24 letters —

To Edgar A. Poe, Esq., editor Graham’s Gentleman’s Magazine, Philadelphia, Pa.