Hostname: page-component-586b7cd67f-vdxz6 Total loading time: 0 Render date: 2024-11-30T20:33:05.608Z Has data issue: false hasContentIssue false
  • English
  • Français

What Poe Knew About Cryptography

Published online by Cambridge University Press:  02 December 2020

W. K. Wimsatt Jr.
W. K. Wimsatt Jr.*
Affiliation:
Yale University
Get access Rights & Permissions [Opens in a new window]

Extract

No one really believed in the authenticity of the answers,” complained Poe. “People will not believe I really decipher the puzzles.” But Poe did decipher the puzzles—that is, he really read a good many cryptograms of a certain simple type. Even the most recent biographers of Poe might be surprised to learn how many cryptograms were sent to him from time to time, and how many he solved or nearly solved. Some of the evidence has been published or cited but not much of it ever in one place, and much of it has been left hidden, especially that part which reveals the limitations of Poe as cryptographer. A large part was not available until September 1941, when the American Antiquarian Society acquired a file of the long-sought Alexander's Weekly Messenger for 1840.

Type
Research Article
Information
PMLA , Volume 58 , Issue 3 , September 1943 , pp. 754 - 779
Copyright
Copyright © Modern Language Association of America, 1943

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Graham's Magazine, xix, 35 (July, 1841).

2 Poe to F. W. Thomas, 4 July, 1841, in Works, ed. R. H. Stoddard (New York, 1884), i, 105.

3 The ensuing study would not have been possible without the cooperation of Professor T. O. Mabbott, who once more has allowed me to take unlimited advantage of his wide familiarity with the materials of Poe research. I have further relied heavily upon Professor John W. Ostrom, who has always been willing to supplement the information in his valuable Check List of Letters to and from Poe (Alderman Library, Charlottesville, 1941); upon Mr. C. S. Brigham, who in advance of republication sent me photostatic copies of the Poe items in the American Antiquarian Society Alexander's for 1840; and upon Colonel William F. Friedman, who has given me the benefit of an expert's advice in several cryptographic questions. See his articles, “Edgar Allan Poe, Cryptographer,” Signal Corps Bulletin, No. 97, pp. 41–53 (July–Sept., 1937); No. 98, pp. 54–72 (Oct.–Dec., 1937); the first of these reprinted from American Literature, viii, 266–280 (Nov., 1936).

I am grateful to the Boston Public Library for permission to quote various documents in the Griswold Collection and to reproduce ciphers in facsimile. Miss Honor McCusker has been of great assistance in this connection.

4 Alexander's Weekly Messenger, Philadelphia, 18 Dec., 1839, p. 2, cols. 1–2. The only recorded copy is at the Ohio State Archaeological and Historical Society, Columbus, Ohio.

5 Graham's Magazine, xix, 34 (July, 1841).

6 Alexander's Weekly Messenger, 15 Jan., 1840, p. 2, col. 4; 22 Jan., p. 2, col. 2; 29 Jan., p. 2, col. 4; 5 Feb., p. 2, col. 3; 12 Feb., p. 2, col. 5; 19 Feb., p. 2, cols. 2–3; 26 Feb., p. 2, col. 4; p. 4, cols. 3–5; 4 March, p. 2, col. 3; p. 4, col. 1; 11 March, p. 2, col. 3; 25 March, p. 2, col. 6; 8 April, p. 2, col. 2; 22 April, p. 2, col. 3; 29 April, p. 2, col. 4.

All the cipher articles as well as some twenty-seven other short articles on miscellaneous literary, scientific, psychological and humorous subjects, of which most if not all were contributed by Poe to Alexander's during the same period, may be conveniently consulted in the reprint by Clarence S. Brigham, “Edgar Allan Poe's Contributions to Alexander's Weekly Mesenger,” Proceedings of the American Antiquarian Society, lii, Part 1 (April, 1942, 45–125). Besides the file of Alexander's for 1840 in the American Antiquarian Society library, a second has apparently come to light (New York Herald Tribune Book Review, May 30, 1943, p. 18, col. 1).

The numbers of Alexander's for 15 and 22 January may also be found at the western Reserve Historical Society, Cleveland, Ohio.

7 C. S. Brigham, loc. cit., pp. 50–51.

8 Graham's Magazine, xviii, 203 (April, 1841).

9 Graham's Magazine, xix, 33–38 (July, 1841).

10 J. A. Harrison, Life and Letters of Poe (New York, 1903), ii, 92–93. The manuscript is in the Griswold Collection at the Boston Public Library. Where Harrison prints “cryptograph” Thomas wrote “cryptography,” and there are in fact two messages written in the same cipher. Poe chose the second and longer, which, however, may be the harder because of its outrageous diction. Cf. notes 57 and 60.

11 Works, ed. R. H. Stoddard (New York, 1884), i, 105.

12 Graham's, xix, 96 (August, 1841).

13 Cf. note 21.

14 Thomas to Poe, 19 July, 1841, Griswold Collection. A. H. Quinn, Poe (New York, 1941), p. 323, refers briefly to this letter.

15 The letter which Thomas dated 3 August (Harrison, op. cit., ii, 95–100) would seem to have been written actually on 3 Sept. (Poe to Thomas, 1 Sept., 1841, facsimile, American Art Association Anderson Galleries, Sale No. 3827, Valuable … Manuscripts … of Mrs. Lucius L. Button et al., 11–12 March, 1930, pp. 89–90, No. 284).

16 For the letter see Harrison, op. cit., ii, 102–103; the cipher, which Harrison omits, is preserved in the Griswold Collection.

17 Thomas to Poe 22 Sept., 1841 implies, however, that on 20 Sept. Poe wrote to Thomas on the subject of cryptography (Griswold Collection).

Colonel W. F. Friedman, Principal Cryptanalyst, Office of the Chief Signal Officer, U. S. War Department, has been kind enough to examine both the Ewing and the Young ciphers and is inclined to think that both will remain unread because of their brevity.

18 Graham's, xix, 192 (October, 1841). He now published the whole of Dr. Frailey's acknowledgment (published partly in August) and also the solution of a short cipher sent on 10 August by a person signing himself Timotheus Whackemwell. By a mistaken guess at the handwriting (“autography!”) Poe had sent the answer to J. N. McJilton, the Baltimore litterateur (cf. Harrison, op. cit., ii, 100).

19 The date of this letter rests on Bolton's statement in his second letter of 4 Nov. Bolton's mistaken impression that the solution had appeared in the September, rather than the October, number does not argue that he was mistaken about the date of his own certificated solution. Bolton's letter and Poe's reply (the latter in facsimile) first came to light in the Memphis Commercial Appeal, 15 Nov., 1925, Section iv, p. 7. They are quoted by W. F. Friedman, Signal Corps Bulletin, No. 97, pp. 48–49; No. 98, pp. 59–60.

20 Memphis Commercial Appeal, 15 Nov., 1925, Section iv, p. 7.

21 The gist of Poe's answer is as follows: “Yours of the 4th is this moment received; and I hasten to exonerate myself. … Our last ‘form’ necessarily goes to press a full month in advance of the day of issue. It often happens, moreover, that the last form in order is not the last in press. … Upon this hint you will easily see the possibility of your letter not having come to hand in season for acknowledgment in the November number. … In our December number (which has been quite ready for 10 days) you will find an unqualified acknowledgment of your claims” (note 20).

Of this explanation it is to be noted: (1) that Poe's acknowledgment of Bolton's solution for the December number may really have been prepared before the arrival of Bolton's 4 November letter and certainly was if the letter did not arrive before 18 November; but (2) that even if it took letters fifteen days to go from Pontotoc to Philadelphia, Bolton's letter sent on 9 September must have arrived in time for the November number. Thomas's acknowledgment of the Frailey solution was sent from Washington on 6 July and was “just received” in time for inclusion in the August number (Graham's, xix, 96)—in the “last form.” Certainly if Bolton's letter of 9 September had been of the same nature, Poe would have found a way to include it in the November number (sent to press presumably about 8 October). In 1844 the September Graham's was not to be made up before 10 August (A. H. Quinn, op. cit., p. 430, Poe to Lowell, 2 July, 1844).

22 Graham's, xix, 308 (Dec., 1841).

23 Ernest D. North, Catalogue of a Choice Collection of Autograph Letters and MSS. (New York, Oct., 1905), pp. 58–59, No. 358. I am indebted to Mr. North for a transcript from his Catalogue, where the letter is printed in full. The original seems not to be further traceable.

24 Signal Corps Bulletin, No. 97, p. 50; No. 98, pp. 59–60.

25 I have not yet been able to determine whether he was.

26 Harrison, op. cit., ii, 102.

27 Tomlin to Poe, 9 August, 1843, Griswold Collection.

28 G. E. Woodberry, Poe (Boston, 1909), ii, 39–41. In a letter to the Literary Review, 9 Sept., 1922, p. 18, Professor J. M. Manly, asking for information about the letter and cipher, said that the letter had been in the possession of William Nelson, of Paterson, N. J., but was no longer traceable. The letter is not mentioned in Anderson Auction Company, Sale No. 1025, Autograph letters … of William Nelson, 16–17 April, 1914.

29 Griswold Collection.

30 Facsimile, American Art Association Anderson Galleries, Sale No. 3800, From the Libraries … of Eustace Conway et al., 16–17 Dec., 1929, pp. 68–69, No. 285.

31 Bangs and Company, Catalogue of—the Collection of a Boston Litterateur, 11 April, 1896, No. 113½.

32 Richmond Times Dispatch, 21 July, 1935, Section v, p. 15. Cf. G. E. Woodberry, Poe (Boston, 1909), ii, 374. The original is in the Valentine Museum.

33 Lowell Weekly Journal, 19 April, 1850, p. 2, col. 4. The article is unsigned, but the authorship is established by a letter from Mrs. Richmond to J. H. Ingram, 4 Jan., 1877 (Ingram Collection, Alderman Library, University of Virginia). I am indebted in this connection to Mr. John Cook Wyllie, Mr. John W. Ostrom, and Mr. Robert H. Haynes.

Cudworth, referring apparently to the ciphers in Alexander's, gives a sample sentence, both cipher and solution, of “one of the easiest.” But this is actually not from any of the Alexander's ciphers or from any other Poe cipher of which I know. Cudworth does not refer by name to Alexander's, but to “one of the weekly papers” of “N. York,” where he says Poe was resident in 1839. This mistake is apparently the basis of J. H. Ingram's like statement (Poe, London, 1880, i, 190–191, where he quotes from Cudworth without giving his name or the reference).

For Poe's visit to Lowell in May, 1849, see A. H. Quinn, op. cit., pp. 613, 622 n.

34 G. E. Woodberry, Poe (Boston, 1909), ii, 40.

35 Alexander's Weekly Messenger, 18 Dec., 1839, p. 2, col. 2.

36 Alexander's Weekly Messenger, 26 Feb., 1840, p. 4, col. 4. Poe further suspected Kulp's cipher because the handwriting was too even.

37 Alexander's Weekly Messenger, 25 March, 1840, p. 2, col. 6.

38 One puzzle and solution which Poe published (26 Feb., p. 4, cols. 3–4) is scarcely a cipher (though included in my count of eleven ciphers with solution, ante p. 755) but a message concealed by being broken into phrases and mixed with other phrases, so that the total makes nonsense. Poe italicized the phrases which make sense. This type of secret writing is similar to the stencil type described later by Poe in Graham's (note 70).

39 Alexander's Weekly Messenger, 25 March, 1840, p. 2, col. 6.

40 Graham's, xix, 34 (July, 1841).

41 W. F. Friedman, Signal Corps Bulletin, No. 97, pp. 50–51 (July–Sept., 1937); No. 98, pp. 65–67 (Oct.–Dec., 1937).

42 Alexander's Weekly Messenger, 19 Feb., 1840, p. 2, col. 2. Cf. John Bartlett, Familiar Quotations (New York, 1914), p. 876, n. 1.

43 Poe's correspondents by choosing a well-known jingle such as this, an enigma in verse, the Lord's Prayer, a passage from the Scriptures, the opening lines of a Shakespearian play, or a few sentences about cryptography, often greatly facilitated his solutions. See especially Alexander's Weekly Messenger, 15 Jan., 1840, p. 2, col. 4; 22 Jan., p. 2, col. 5; 26 Feb., p. 2, col. 4; 11 March, p. 2, col. 3; Graham's, xix, 36 (July, 1841); and note 32.

44 Graham's, xix, 308 (Dec., 1841).

45 Alexander's Weekly Messenger, 26 Feb., 1840, p. 2, col. 4; 11 March, p. 2, col. 3; 29 April, p. 2, col. 4.

46 Notes 36 and 37.

47 Graham's, xix, 96 (August, 1841).

48 Note 18.

49 Note 27. Poe also jotted down some words and incomplete words with points taking the place of letters, just as Legrand does in “The Gold Bug.” Thus: “Ala.ara,” “can you decyp,” “sir,” “thin,” “bear,” “statue,” “.hips,” “fou,” “.title,” “t.s.aloosa,” “July 24,” “twenty.” It will be recalled that the cipher was from the Hon. Alexander B. Meek, of Tuscaloosa, Alabama, and that Tomlin had forwarded it to Poe on 9 August.

50 Note 32.

51 Note 33.

52 Graham's, xviii, 203 (April, 1841).

53 Graham's, xix, 37 (July, 1841).

54 Graham's, xviii, 203 (April, 1841).

55 Graham's, xix, 96 (August, 1841).

56 Graham's, xix, 307 (Dec., 1841).

57 “Without dubiety incipient pretension is apt to terminate in final vulgarity, as parturient mountains have been fabulated to produce muscupular [sic] abortions” (Graham's, xix, 192, October, 1841). With this sentence of the Frailey plain text compare the following from the speech of Johnson's Ghost by Horace and James Smith: “A swelling opening is too often succeeded by an insignificant conclusion. Parturient mountains have ere now produced muscipular abortions” (Rejected Addresses [Boston, 1860], pp. 107–108; the book was originally published in 1812).

58 Graham's, xix, 96 (August, 1841). Cf. note 65.

59 Philadelphia Saturday Museum, 4 March, 1843, p. 2, col. 4 (copy in the University of North Carolina Library). This little article, entitled “Quick Perception,” appeared in the same number of the Saturday Museum as the Hirst biography of Poe (cf. note 109). The cipher was a short one of the simple substitution sort. “‘Why did Adam bite the apple,’ said a schoolmaster to a country boy. ‘Because he had no knife to cut it,’ answered the boy.”

60 Griswold Collection, Thomas to Poe, 12 May, 1845. The Register of All Officers and Agents … of the United States (Washington, 1845), p. 113, lists Charles S. Frailey as a clerk of the Treasury Department, General Land Office.

61 Note 30.

62 This example is added in the margin.

63 I leave the dates untranslated because I do not believe they are determinate. Poe translated “25 September 1843,” and Professor Mabbott has pointed out to me that Poe seems to be working on the very plausible assumption that the signs here employed for numbers are to be taken as fragments of two patterns of crossed lines: + and ∗, the angles standing for numbers 0 to 9 in some symmetrical order, as is common in business codes. If we assume, as Poe did, that ┘ is surely 1, ∠ almost surely 8, and probably 2, then we can fill in the other numbers in a partially symmetrical pattern such as gives Poe's date (25 September, 1843) or in another pattern which gives 24 September, 1839. But both these dates must be rejected because Richard Douglas became a Captain in the Marine Corps on 7 March, 1839; see T. H. S. Hamersly, General Register of the United States Navy and Marine Corps … 1782–1882 (Washington, D. C., 1882), p. 877. It seems likely that Douglas used the angles in some asymmetric order, so that the pattern had really no significance.

64 Poe was particularly troubled by the symbol ⊃, which stands for the recurrent y of the closed vowel sounds in the phonetic spelling of Lieutenant Douglas. Poe translated ⊃ sometimes as s, as he also translated c, which is really s. But he translated ⊃ as r in “our” for “muy”; as i in “neseserri” for “neseseriy”; and as d in “reside” for “resiyv.” In the phrase “yuwnuyted steyts mariyns,” which he suggests may mean “United States Marine,” he ignores the troublesome ⊃ altogether.

65 It will be remembered that Thomas in his testimonial letter of 6 July, 1841 (note 58) recalled how in the previous year at his own lodgings in Philadelphia he and Dow and Poe had talked about Aaron Burr's correspondence in cipher. Poe “laughed” at what he “termed Burr's shallow artifice” and said he could “decipher any such cryptography easily.” This conversation seems to have taken place sometime in May, 1840, or shortly after (Letter of Thomas to Poe, 3 [Sept.] 1841, J. A. Harrison, Life and Letters of Poe, ii, 98–99). Moreover I think that even if Poe knew of Burr's ciphers earlier in the year, when he was solving the ciphers in Alexander's, he had not derived his ideas from Burr. The celebrated letter of Burr to Wilkinson 29 July, 1806, was written in three ciphers, one a simple substitution cipher of arbitrary symbols, one a cipher of special symbols for important words, and one a cipher based on the use of a certain edition of Entick's pocket dictionary. These facts could have been known to Poe accurately from the published account of Burr's trial (T. Carpenter, The Trial of Col. Aaron Burr, Washington City, 1808, Vol. iii, Appendix l, where also are described some cyclic ciphers used by Burr's associates), but vague ideas of the subject must have been widely spread through conversation. See Nathan Schachner, Aaron Burr (New York, 1937), opp. p. 325, a facsimile of the letter and of the key to the arbitrary alphabet and special symbols. Charles Burdett, Margaret Moncrieffe (New York, 1860) gives as frontispiece a facsimile of a key to special symbols used by Burr in cipher correspondence with his daughter.

66 The term “secret writing” (Geheimschrift or occultae scripturae) which from now on he uses at the head of his articles does not appear to be common in the cipher literature of that time. Cf. note 96.

67 For the possibility that “A Descent into the Maelström” was among the Tales of the Folio Club submitted by Poe for the Baltimore Saturday Visiter prize in 1833, see A. H. Quinn, Poe (New York, 1941), pp. 202, 213, 745–746.

68 G. E. Woodberry and E. C. Stedman, Works of Poe (New York, 1894–95), iv, 291; Adolph B. Benson, “Scandinavian References in the Works of Poe,” Journal of English and Germanic Philology, XL, 83–85 (Jan., 1941).

69 Graham's, xviii, 235–241 (May, 1841).

70 One of the puzzles solved in Alexander's (note 38) was similar to this.

71 In all these cases, he revised the wretched phrasing of the Britannica completely. The cipher with two copies of a book he might have found in the Encyclopaedia Americana (Philadelphia, 1836), iv, 63–64, “Cryptography.” Here also he might have found the term “secret writing.” Cf. note 96.

72 Graham's, XIX, 33 (July, 1841).

73 Encyclopaedia … First American Edition (Philadelphia, 1798), v, 16. The whole article on ciphers is contained on pp. 16–18. With Britannica removed from the title, this is the third edition of the Encyclopaedia Britannica (Edinburgh, 1797). Poe might have used as late an edition as the sixth (Edinburgh, 1823), vi, 158–160 (John Crerar Library). The chief source of the Britannica article on ciphers would seem to be William Hooper, Rational Recreations (London, 1774), i, 143–159.

74 Graham's, xix, 38 (July, 1841). “The works of the two latter,” continues Poe, “may be found, we believe, in the library of the Harvard University.” And accordingly Howard Paul, in an article entitled “Recollections of Edgar Allen Poe” (Munsey's Magazine, vii, 557, Aug., 1892) but derived actually in large part from Poe's writings in Graham's, created the myth that Poe had made a pilgrimage to “Harvard University to consult treatises by Trithemius, Vignere, and Niceron on his favorite theme.” The Catalogue of the Library of Harvard University (Cambridge, 1830–31) includes Vigènere's Traicté des Chiffres (Paris, 1586), but none of the cryptographic writings of Niceron, Porta, or Trithemius.

75 Cortell Holsapple, “Poe and Conradus,” American Literature, iv, 62–65 (March, 1932), thinks the cryptograph in “The Gold Bug” may owe something to David Arnold Conradus, “Cryptographia Denudata,” Gentleman's Magazine, xii, 133–134, 185–186, 241–242, 473–475 (March, April, May, Sept., 1742). But there is no reason to think that Poe went to any place so obscure as the Gentleman's Magazine for 1742, either before July, 1841, up to which time he seems to have relied on what he possessed in his “own intellect,” or afterwards, when he had consulted the Britannica and Rees. Cf. note 81.

Mary E. Phillips, Poe the Man (Chicago, 1926), ii, 1230–31, believes Poe the author of an article in Blackwood's, lxii, 422–431 (October, 1847) and, on the basis of an allusion to Champollion in this article, suggests that Poe was indebted to Champollion's work on Egyptian hieroglyphics. Although there seems no sufficient reason for attributing the Blackwood's article to Poe (The attribution rests on a question which Eveleth put to Poe, Letters from George W. Eveleth to Edgar Allan Poe, ed. T. O. Mabbott [New York, 1922], p. 21), it is possible to demonstrate Poe's awareness of Champollion from a passage in Eureka (Works, ed. J. A. Harrison [New York, 1902], xvi, 196). Yet from this it certainly does not follow that Poe owed any cryptographic knowledge to Champollion, who was about as well known as Einstein in our day, and whose method of reading hieroglyphics with the aid of the Rosetta Stone bore no great resemblance to Poe's method of reading English ciphers. See J. G. H. Greppo, Essay on the Hieroglyphic System of M. Champollion, trans. Isaac Stuart (Boston, 1830), pp. 19–25.

The mysterious chasms and rock carvings which Poe introduced into Chapter xxiii of A. Gordon Pym and the explanation of them which he appended in the “Note” at the end of the story may show that by 1838 he was interested in hieroglyphics, but they have no connection with cryptography. The explanation of the hieroglyphs supposes a knowledge of Ethiopic, Arabic, and Egyptian alphabets (Works, ed. J. A. Harrison, New York, 1902, ii, 222–225, 243–244, 329).

This is probably the place to say that the two articles on Poe's cryptography in the Bookman, xii, 4–7 (March, 1903) and xxviii, 450–451 (Jan., 1909) are worthless, the second, the work of Firmin Dredd, being taken, comically enough, in great part verbatim from the earlier work of the editor.

76 “The invention of a perfect cypher was a point to which Lord Chancellor Bacon devoted many months;—devoted them in vain, for the cryptograph which he has thought worthy a place in his De Augmentis, is one which can be solved” (Graham's, xix, 96, August, 1841).

77 “He studied the principles of the art of deciphering with great interest, and invented one cipher so ingenious, that, many years later, he thought it deserving of a place in the De Augmentis” (T. B. Macaulay, Critical and Historical Essays, ii [“Everyman's Library,” No. 226], 303). Macaulay seems to echo Bacon's “aliud inventum subjiciemus; … nec etiam adhuc visa nobis res digna est, quae pereat” (Works, London, 1826, vii, 262; De Augmentis, vi, i). Macaulay's essay on Bacon first appeared in the Edinburgh Review for July, 1837. A serial digest of it appeared in the Southern Literary Messenger for January, February, and March, 1838, and on the first page of the January installment the passage in question was quoted (iv, 9), while in a conspicuous position at the end of an editorial note appeared a line by Pope on Bacon which Poe echoed in the December 1841 Graham's. (Cf. note 89.) Poe was living in New York in the winter of 1838 but may be supposed to have kept an eye on the magazine of which but a year before he had been editor. (For the story of how he later borrowed a volume of the Messenger, see A. H. Quinn, Poe [New York, 1941], 407–409.)

The first American edition of Macaulay's Critical and Miscellaneous Essays appeared in Boston in 1840, two volumes, the last essay of the second volume being that on Bacon (copy in the Harvard College Library; and cf. North American Review, li, 502, Oct., 1840). In 1841 Carey and Hart brought out a second edition in Philadelphia, two volumes containing the same essays as the Boston Edition and in the same order, and a third volume containing other miscellaneous writings (copy in the collection of Professor Carl F. Schreiber, Yale University). The third volume was reviewed by Poe for the June, 1841, number of Graham's (xviii, 294–295), and the concluding words of his review suggest that he was familiar with the first two volumes or at least had them at hand.

78 In Macaulay's essay Poe could have found the title De Augmentis but not the statement that the cipher could be solved (though it was not taking a great risk to say that it could be). In the encyclopedia which we are about to discuss he could have found several statements that the cipher could be solved but not the title De Augmentis.

79 Graham's, xix, 35 (July, 1841).

80 J. O. Bailey, “Poe's ‘Palaestine’,” American Literature, xiii, 44–58 (March, 1941); “Poe's ‘Stonehenge’,” Studies in Philology, xxxviii, 645–651 (October, 1941); “Sources for Poe's Arthur Gordon Pym, ‘Hans Pfaal,‘ and Other Pieces,” PMLA, lvii (June, 1942), 513–535; M. N. Posey, “Notes on Poe's Hans Pfaall,” MLN, xlv, 501–507 (Dec., 1930). I am particularly indebted to Professor Bailey for general advice on Poe's use of encyclopedias and for the opportunity of studying his manuscripts in advance of publication.

81 Envelope of letter, Thomas to Poe, 19 July, 1841, Griswold Collection. Poe used the space on one side of the address for the group of notes which I have placed first, and the space on the other side of the address for the other three groups in the order in which I present them. The four groups are to be found in this order in Rees. Cf. note 97.

82 Abraham Rees, The Cyclopaedia (London, 1819), viii, “Cipher” [p. 24].

83 Ibid. [p. 19]. Both Rees and Poe omit j and v.

84 Graham's, xix, 307 (Dec., 1841).

85 Abraham Rees, op. cit., vin, “Cipher” [pp. 5, 13]. P. 13 quotes Shaw's translation verbatim: “There are three properties required in cyphers, viz.; (1.) that they be easy to write and read; (2.) that they be trusty and undecypherable; and, (3.) if possible, clear of suspicion.” Cf. Peter Shaw, Philosophical Works of Francis Bacon (London, 1733), i, 141.

86 Graham's, xix, 308 (Dec., 1841).

87 Abraham Rees, op. cit., viii, “Cipher” [pp. 14, 29].

88 Essay on Man, iv, 281.

89 Abraham Rees, op. cit., iii, “Francis Bacon” [p. 5]. But cf. note 77. For Poe's acquaintance with Bacon see Killis Campbell, “Poe's Reading,” University of Texas Studies in English, v, 185 (1925).

90 Graham's, xix, 308 (Dec., 1841).

91 (A) John Davys, An Essay on the Art of Decypherìng (London, 1737), includes, pp. 9–23, a “Discourse of Dr. Wallis. Now first publish'd from his Original Manuscript in the Publick Library at Oxford.” Wallis speaks (pp. 11–13) of solving one cipher which contained nearly 300 characters, but he does not explain how the characters were employed. A letter from Wallis' collection (pp. 23–27) shows the use of more than one symbol for a few letters of the plain alphabet, but this letter is read not by Wallis but by Davys. Again, a cipher in which each letter of the alphabet is denoted by more than one symbol (p. 44) is described not by Wallis but by Davys. (B) The Monthly Magazine, xiii, 446, 560 (June, July, 1802); xiv, 252, 521 (Oct., 1802; Jan., 1803) prints a number of letters about ciphers solved by Wallis, but very little is said about the nature of these ciphers. (C) Thomas Hearne, Works (London, 1810), iii, clix–clx, contains a letter in which Wallis tells of solving a cipher “in Numeral Figures, extending in number to above seaven hundred, with many other Characters intermixed.” Doubtless many of these symbols were used as variants for letters of the alphabet, but many may have been used as special symbols for words and syllables—as is true of the first cipher mentioned under (A) above. (D) John Wallis, Operum Mathematicorum (Oxford, 1699), iii, 659–672, presents two ciphers solved by Wallis, neither of which is the sort of which Poe spoke. Tbe numbers used in each of these ciphers run as high as 471, but by no means all the numbers from 1 to 471 are used. (E) Biographia Britannica, vol. vi, part ii (London, 1766), “John Wallis,” pp. 4117, 4121, 4126, 4130, 4136, notes G, P, Z, LL, XX, quotes generously from all but (B) above and from other sources.

92 Abraham Rees, op. cit., viii, “Cipher” [p. 11].

93 Graham's, xix, 308 (Dec., 1841).

94 The treatment of this is one of the weak points of the Rees article, which does not use the term chiffre quarré, and which considers only cycles of alphabets by words or lines (as perhaps in the Alexander's seven-alphabet cipher). Poe's inaccurate statement that “no letter is ever represented twice by the same character” refers obviously to the more difficult cyclic change of alphabets from letter to letter.

95 Lucille King, “Notes on Poe's Sources,” University of Texas Studies in English, x, 131–134 (1930).

96 Encyclopaedia Americana (Philadelphia, 1836), IV, 63–64, “Cryptography.” This article employs the term “secret writing,” a translation of Geheimschrift in the article Geheimschrift (Kryptographie) in Allgemeine Deutsche Real-Encyklopädie, Siebente Originalauflage (Leipzig, F. A. Brockhaus, 1830), iv, 549–550 (University of Chicago Library). The Americana is based on the seventh edition of the Real-Encycklopädie. Poe's use of the term “secret writing” in the July Graham's suggests he may have consulted the Americana at the same time as the Britannica. Cf. note 66.

97 Poe wrote, in the space below the address, and at right angles to the notes which he had taken on cryptography: “Et dit, moi je m'en / Et dit, moi je m'en / Ma foi; moi je men ris / Béranger's ‘petit homme gris’.” Cf. P. J. De Béranger, Oeuvres Complètes (Paris, 1834), i, 38–40, “Le Petit Homme Gris.”

98 The Gold Bug, ed. Hervey Allen and T. O. Mabbott (New York, 1928), p. 68. In revising a copy of Tales, 1845 (the Lorimer Graham copy, now at the Century Club, New York) Poe added to “The Gold Bug” the passage on the chemistry of invisible ink (T. O. Mabbott, op. cit., pp. 58–59, 86–89). The passage bears a fairly close resemblance to the last paragraph of the article “Ink” in Abraham Rees's Cyclopaedia (London, 1819), vol. xix. A cross-reference to “Ink” is given in vol. viii, “Cipher” [p. 17].

99 The Gold Bug, ed. cit., p. 66.

100 Graham's, xix, 33 (July, 1841).

101 W. F. Friedman, Signal Corps Bulletin, No. 97, pp. 52–53 (July-Sept., 1937).

102 Alexander's Weekly, 25 March, 1840, p. 2, col. 6; 22 April, p. 2, col. 3. The italics are mine.

103 Works, ed. R. H. Stoddard (New York, 1884), i, 105.

104 Graham's, xix, 96 (August, 1841).

105 Graham's, xviii, 203 (April, 1841).

106 Graham's, xix, 34–36 (July, 1841).

107 Works, ed. R. H. Stoddard (New York, 1884), i, 105.

108 G. E. Woodberry, Poe (Boston, 1909), ii, 40.

109 Philadelphia Saturday Museum, 4 March, 1843, p. 1, col. 9.

110 The Gold Bug, ed. cit., p. 65.

111 Philadelphia Saturday Museum, 4 March, 1843, p. 1, col. 8.

112 Lowell Weekly Journal, 19 April, 1850, p. 2, col. 4.

113 Clarence S. Brigham, loc. cit., 105–106, 116–117, 123.

114 Alexander's Weekly Messenger, 22 Jan., 1840, p. 2, col. 5.

115 Idem, 26 Feb., 1840, p. 4, col. 5.