In his instructions for the expedition to the Pacific, Jefferson anticipated that Meriwether Lewis might find opportunity to send dispatches back by way of western army posts, “putting into cypher whatever might do injury if betrayed” (see Document IV of the group of documents on drafting instructions for Lewis, at 13 Apr. above). Two undated manuscripts, in Jefferson’s handwriting and among his papers at the Library of Congress, record the cipher he worked up for Lewis. Jefferson put no heading on either manuscript, but in one of them he enciphered a hypothetical message that he could speculate might someday come from Lewis, beginning “I am at the head of the Missouri” (see Document I below). Jefferson endorsed the other paper as a “Cypher establd. with Captain Lewis” (Document II). In that endorsement he also identified “Artichokes” as a key for the cipher. That word forms a direct connection between the manuscripts, for in Document I, “artichoke” was the key (he used “antipodes” in Document II). Document I records the first stage of his work on the cipher, shown by unused columns on the right side of the table, his labored effort to explain the method, and his first practice exercise using no key word. Document II is a neatly written manuscript, made after he had developed the system to his liking.

He noted in the endorsement to Document II that he established (“establd.”) the cipher with Lewis. They must have discussed it in person, for the subject of the cipher did not come up in the letters they exchanged during the spring of 1803 while Lewis was in Harpers Ferry, Lancaster, and Philadelphia making preparations for the expedition. Perhaps Jefferson drew up the method in April, when he drafted the instructions for the journey and included the reference to a cipher. If he “established” it with Lewis in person, however, they could not have done so until Lewis returned to Washington in June, and it is possible that Jefferson prepared the cipher as late as that month.

The tableau-style cipher system that Jefferson drew up for Lewis dated from the sixteenth century (Weber, United States Diplomatic Codes description begins Ralph E. Weber, United States Diplomatic Codes and Ciphers, 1775–1938, Chicago, 1979 description ends , 10–12; David Kahn, The Codebreakers: The Story of Secret Writing, rev. ed. [New York, 1996], 133–7, 145–51). The letters of the alphabet appear in order in each row, but the position of the letters shifts from row to row. (Jefferson included the ampersand as, in effect, a 27th letter of the alphabet.) In his first trial of the cipher, at the top of the grid he numbered the columns, leaving the one on the far left unnumbered (see Document I). The user would work through the intended message letter by letter, finding the first letter in the column on the left and reading across to the column numbered “1,” then finding the second letter in the left column and reading across to column 2, and so on. On reaching the 27th letter of the message, the user would begin again with column 1. Jefferson experimentally encrypted a passage that was 27 letters long, “the man whose mind on virtue bent,” to get the result “ujh qft epxbp yvas dd maknpa zcmu.” He followed that exercise with a brief set of directions.

He revised the instructions to incorporate the usage of a key word, a device that was, like the form of the cipher itself, well established (Kahn, Codebreakers, 148). Now instead of finding the numbered column to match the letter’s position in the message, he decided on a key word—“artichoke”—and aligned the letters of the message to be encoded with the letters of the key, repeating the key word as many times as necessary to complete the message. Going then to the cipher table, he found each letter of the message in the left-hand column as before, located the corresponding letter of the key word along the bottom row of the table, and at the intersection of the appropriate row and column found the letter needed for the encryption. As a demonstration, he aligned the first two words of his test passage (“the man”) with the first six letters of “artichoke,” then from the table found the encryption “u z y v d v.” To try the method with a longer passage, which would require repetition of the key word, he experimented at the foot of Document I with the imagined message from Lewis: “I am at the head of the Missouri. all well, and the Indians so far, friendly.” Lewis echoed that demonstration passage when, on his arrival at Pittsburgh on 15 July, he cheerfully wrote the president to announce—not in cipher—“so far and all is well.”

For the fair copy of the cipher, Jefferson, now assuming that a key word would always be employed, did not number the columns (Document II). Instead he added an alphabetical row at the top of the table, to be used to find the letters of the key word. With “antipodes” as his key this time, he enciphered the phrase he had used in the draft (“the man whose mind …”). He also added a new feature, digits alongside the first ten rows of the chart. This element allowed the user to code numerals. To encipher “18,” as shown by an example he gave in the document, one would align the two digits with the first two letters of the key word—a and n in the case of “antipodes”—and then find the intersections of the numeral 1 row with the “a” column and the numeral 8 row with the “n” column. The result was the letters b and v, which—as Jefferson demonstrated but did not explain—should be underlined to indicate that they stood for numbers, not letters of the alphabet.

In his instructions for the cipher, Jefferson did not discuss his criteria for the selection of key words. “Artichoke” and “antipodes,” however, are each nine letters long and no letter appears more than once in either word. Employing either word as the key therefore utilizes nine columns of the cipher table, whereas a word such as “attention,” also nine letters long, would use only six columns. Making “artichoke” plural, as he did in the endorsement to Document II, expands to ten the number of columns put into service. Ease of recollection may be one reason he chose to experiment with key words beginning with the letter a, or perhaps he consulted an alphabetically ordered word list, such as Samuel Johnson’s Dictionary of the English Language, and early in his quest found words that met his criteria.

The distinctive phrase that Jefferson enciphered as a practice exercise in each document—“the man whose mind on virtue bent”—was the first line of an interpretation by the poet Thomas Blacklock of an ode by Horace. Although Jefferson was probably not familiar with Blacklock’s poetry, he knew very well the work in which the ode first appeared: David Hume’s history of Britain. In a volume first published in 1757, Hume recounted the story of Cornelis de Witt of the Netherlands, who in 1672 was falsely accused of plotting against the stadtholder and, under torture, maintained his innocence by repeatedly reciting the ode in Latin. In a footnote, Hume, who at the time was Blacklock’s friend and benefactor, included an English translation that he had solicited from the poet. Blacklock’s three verses begin:

“The man, whose mind on virtue bent,

 Pursues some greatly good intent,

   With undiverted aim,

 Serene, beholds the angry crowd;

 Nor can their clamors, fierce and loud,

   His stubborn honor tame.”

The verse was quoted a few times (including praise by Oliver Goldsmith) but was likely not very widely known. Jefferson had long acquaintance with Hume’s History, though, purchasing his first set of the volumes in 1764. Although he came to denounce Hume as an “Apologist” of first the Stuarts, and then of their many predecessors on the English throne going back to Saxon and Norman times, he acknowledged the powerful “charms of stile” that made the philosopher’s prose dangerously beguiling to unsuspecting readers. Jefferson had once been vulnerable himself: “I remember well the enthusiasm with which I devoured it when young,” he recalled of the History in 1810. Perhaps he found the ode’s opening line to be a practical exercise for his cipher in 1803 because it uses 15 different letters, more than half the alphabet. It is 27 letters in length, which conveniently required him to utilize every column of the table in his first practice and to begin again with column 1. But what really led him to choose that passage, rather than some other phrase that could serve the purpose, is not known. About a year earlier, he used the Lord’s Prayer and an address from the first Continental Congress as exercises using the cipher invented by Robert Patterson (David Hume, The History of Great Britain. Vol. II. Containing the Commonwealth, and the Reigns of Charles II. and James II. [London, 1757], 224–5; Hume, The History of Great Britain. Vol. II, 2d ed. [London, 1759], 224; Ernest Campbell Mossner, The Forgotten Hume: Le bon David [New York, 1943], 18–37; DNB description begins H. C. G. Matthew and Brian Harrison, eds., Oxford Dictionary of National Biography, In Association with The British Academy, From the Earliest Times to the Year 2000, Oxford, 2004, 60 vols. description ends ; H. Trevor Colbourn, The Lamp of Experience: Whig History and the Intellectual Origins of the American Revolution [Chapel Hill, 1965], 158–9, 177–81, 217; John M. Werner, “David Hume and America,” Journal of the History of Ideas, 33 [1972], 454–6; London Magazine, 40 [1771], 538; The Miscellaneous Works of Oliver Goldsmith, M.B. A New Edition, 4 vols. [London, 1801], 4:406–7; Maarten Prak, The Dutch Republic in the Seventeenth Century: The Golden Age, trans. Diane Webb [Cambridge, 2005], 53; LCB description begins Douglas L. Wilson, ed., Jefferson’s Literary Commonplace Book, Princeton, 1989, The Papers of Thomas Jefferson, Second Series description ends , 167–9; Sowerby, description begins E. Millicent Sowerby, comp., Catalogue of the Library of Thomas Jefferson, Washington, D.C., 1952–59, 5 vols. description ends No. 370; RS description begins J. Jefferson Looney and others, eds., The Papers of Thomas Jefferson: Retirement Series, Princeton, 2004- , 9 vols. description ends , 3:5; Vol. 37:275–81).

If Lewis ever implemented the cipher, no documents in which he did so have been found. Jefferson may have expected him to encrypt communications only in the event of extraordinary circumstances. The contingency foreseen in the instructions, that Lewis might recruit Native Americans as couriers to take dispatches to Cahokia or Kaskaskia, did not arise. The last documents that Lewis and William Clark sent from Fort Mandan in April 1805, before they continued up the Missouri River, they left in plain text and put in the care of the crew that returned downriver to St. Louis in the expedition’s keelboat. Lewis did not write to the president again until September 1806, after he reached St. Louis on his return journey (Jackson, Lewis and Clark description begins Donald Jackson, ed., The Letters of the Lewis and Clark Expedition, with Related Documents, 1783–1854, 2d ed., Urbana, Ill., 1978 description ends , 1:231–42, 317–25).