Meriwether Lewis · Natchez Trace · 11 October 1809

The Death at Grinder's Stand

Meriwether Lewis died in the night of 10 to 11 October 1809 at a log inn on the Natchez Trace, shot twice. This page weighs suicide, murder and accident against every document that survives.

Every quotation below comes from a saved copy of the document named, and each copy carries its citation, its web address and the date it was fetched. The copies, the working notes, the data and the code are in the files that accompany this page.

What happened

On the evening of 10 October 1809 a rider in a loose white gown striped with blue came up alone to two log cabins in the woods of what is now Lewis County, Tennessee, about 72 miles south-west of Nashville, and asked for a night's lodging. The place was Grinder's Stand, one of the settlers' cabins that took in travellers along the Natchez Trace, the old road that ran north-east from the Mississippi through the Chickasaw country to Nashville. The rider was Meriwether Lewis, who had led the Lewis and Clark expedition. As an army captain he had commanded, with William Clark, the expedition that crossed the continent to the Pacific and back in 1804 to 1806. Before that he had been private secretary to President Thomas Jefferson, who planned the expedition and remained his patron. In 1807 Jefferson had made him governor of the Louisiana Territory, the northern part of the land bought from France in 1803, governed from St. Louis. He was 35 years old and on his way to Washington.

His affairs in St. Louis had gone badly by the summer of 1809. Jefferson had left office in March 1809, and in July the War Department under the new president, James Madison, refused to pay some of the bills Lewis had drawn on the government for public spending as governor. The refusal, once known in St. Louis, brought his private creditors down on him. He was at open odds with his deputy, the territorial secretary Frederick Bates. The book of the expedition, which he had promised to write, was three years overdue. He set out on 4 September to defend his accounts in person, carrying the expedition's journals and duplicates of his vouchers, and went down the Mississippi by boat, ill with fever. At Fort Pickering, the army post on the Chickasaw Bluffs where Memphis now stands, the commander, Captain Gilbert Russell, found him deranged and kept him there a fortnight, until he seemed recovered. Lewis then took to the road with James Neelly, the United States government's agent to the Chickasaw nation, through whose land the Trace ran; with his own servant, John Pernier, a man Jefferson had known for six years; and with Neelly's servant. Two days before Grinder's Stand, two of the horses strayed in the night, and Neelly stayed behind to find them while Lewis rode on with the two servants.

The man of the house, Robert Grinder, was away. His wife, Priscilla Grinder, gave Lewis the main cabin and took her children to the kitchen cabin. The servants slept in the stable, some distance off. As Mrs Grinder later told it, Lewis ate little, paced the yard talking to himself, flushed with anger and then turned courteous ("Madam this is a very pleasant evening"), asked his servant for his gunpowder, refused a bed and lay down on the floor on bearskins and a buffalo robe. In the night she heard a pistol shot, something heavy falling, the words "O Lord!", and then a second shot. Lewis came to her door and asked for water, and she was too frightened to answer. When the servants reached him, he told them he had shot himself and asked again for water. He had been shot in the head and in the body, and by some accounts cut with a razor. He was alive at sunrise and dead soon after. Neelly came up later that day and buried him near the stand.

Neelly wrote to Jefferson from Nashville a week later that the governor had killed himself in a fit of derangement. Jefferson accepted the report, and so did William Clark, Lewis's partner in command on the expedition and now the government's Indian agent at St. Louis, whose first words on reading a newspaper report were "I fear this report has too much truth". The Nashville, St. Louis and Washington newspapers printed the death as suicide. No doctor saw the body and no inquest was held.

Why it is a mystery, and what each side says

None of the people who wrote about the night was in the room. Everything comes from Mrs Grinder, who told her story at least four times over 36 years and told it differently each time, and from two servants whose words survive only as others reported them. No one recorded the things that would settle a shooting now: powder marks, the position of the pistols, the state of his hands. The head wound is described in five incompatible ways by five writers who had it at second hand.

The case for suicide was made by the people who knew him. Jefferson wrote in 1813 that Lewis had been "subject to hypocondriac affections" from early life; the word then meant depression. Russell said the boat crew had reported two attempts on his own life on the river, and a letter written at Nashville thirteen days before the death says the same. Neelly found him "at times deranged in mind" on the road. The pressures described above, the protested bills, the quarrel with Bates and the overdue book, were known to all of them. Most historians since have held that view: Dawson Phelps, the National Park Service historian of the Trace, in 1956; Donald Jackson, editor of the expedition's letters, in 1962; and most of the writers who have followed them.

The case for murder began in Tennessee. In 1848 the state legislature appointed a committee to put a monument over the grave, and the committee's report to the legislature of 1849 to 1850 said that death "by the hands of an assassin" seemed "more probable", giving as its grounds the loneliness of the place and a story that the servant had fled with the money and horses. Local tradition accused the innkeeper. The Lewis family accused "his Spanish servant". In 1962 the novelist Vardis Fisher, in Suicide or Murder?, argued that every document was second-hand and every witness interested. Since then the surgeon E. G. Chuinard (1991) has held that Lewis could not have survived and moved as described, and named Neelly as the killer; the historian John Guice (1998 to 2006) has argued that the story never fitted, from the moonless night to the two hours Mrs Grinder waited; the forensic scientist James Starrs and the independent researcher Kira Gale (2009, 2016) have argued for a plot by General James Wilkinson, the army's senior officer and Lewis's predecessor as governor, and have called Russell's statement of 1811 a forgery; and Tony Turnbow, a historian of the Trace, has argued for a robber and a cover story arranged by Neelly. In 1996 a coroner's jury in the county where the grave lies recommended that the body be exhumed; a federal court in 1998 and the Department of the Interior, which manages the grave as part of the Natchez Trace Parkway, in 2010 were asked to allow it, and it has not been opened. Thomas Danisi, Lewis's biographer, holds a third position: Lewis fired the shots in a malarial delirium, and "suicide" is the wrong word for it. At a panel of historians in 2009, two hundred years after the death, the historians put suicide at 96 percent; the audience voted 350 to 20 for murder.

How I judged it

I read every document from 1809 to 1813 that bears on the death, in the editions listed in section 13, and the modern debate from Fisher to 2025, in full where it is published in articles and in fragments where it is in books I could not obtain. I put the events in date order, laid Mrs Grinder's tellings side by side, checked the one claim that she saw something against the phase of the moon, and compared the reported wounds with modern forensic case series of suicides and homicides involving more than one shot. Then I scored sixteen pieces of evidence, each as a range of how much more likely it would be if Lewis shot himself than if someone else did, and combined the scores under several starting assumptions, so that a reader who disagrees with any number can change it and see what follows. The combination puts self-infliction at 92 percent in its central run and at 63 to 76 percent when every item is discounted; a reader who starts from a three-to-one presumption of murder and discounts every item to a quarter ends at 32 percent.

My own figure is about 85 percent for self-inflicted, about 15 for murder and under 1 for accident. It sits below the central run because the model cannot allow for how badly the body was examined: nobody who saw it knew what to look for. The case rests mainly on the month before the death rather than on the night itself, for which there is no reliable witness. Attempts on his own life were reported thirteen days before he died, through chains that do not depend on the one disputed document. The two men who knew him best accepted the news at once. The two men who could have robbed him went straight to his patrons, and his pistols, rifle, watch, horses and two trunks of papers were all accounted for. The reported wound pattern, a head shot that missed the brain, then a shot to the trunk, then hours of survival, is one that suicides with a pistol are documented to produce, and every forensic pathologist who has examined the surgical claim that it is impossible has rejected it. Every murder theory requires at least one of the three people at the stand to have lied about a man the servants found alive and talking, and the conspiracy version requires the record to have been false from a fortnight before he died. Murder keeps about 15 percent for two reasons: the body was never examined for the signs that would decide a shooting, and the model cannot allow for the possibility that the three people at the stand agreed on a false story.

What this page covers

The verdict panel below gives the figures, a table showing how far they move when the starting assumptions change, and a list of what is new here. Section 1 introduces the people and the documents and defines the three possibilities weighed. Section 2 sets out the dated record from Lewis's departure from St. Louis to the last statement by a contemporary, each entry naming its document. Section 3 asks who was in a position to know what. Section 4 lays Mrs Grinder's tellings side by side and tests the one visual claim against the moon. Section 5 covers the wounds, the pistols and what modern forensic series say about such injuries. Section 6 covers Lewis's state of mind: his debts, his illness, the attempts on the boat and the five diagnoses that have been argued. Section 7 takes the homicide theories one by one, each stated in its proponents' words. Section 8 explains the scoring behind the figures. Section 9 answers the strongest opposing writers by name and counts where historians stand. Sections 10 to 13 say what would overturn the verdict, what is new and what agrees with earlier work, what I could not read, and where every document can be found.

Contents

1. Who is who, and the three possibilities
2. What the record says, in order
3. Who saw what
4. Mrs Grinder's story, version by version
5. The wounds and the pistols
6. The state of his mind
7. The homicide theories, one by one
8. How I weighed it
9. The strongest case against, answered by name
10. What would overturn this verdict
11. What is new, and what agrees with earlier work
12. What I could not read
13. Sources

1. Who is who, and the three possibilities

The people who appear in what follows, in roughly the order they enter the story.

• Meriwether Lewis (1774 to 1809): army officer; private secretary to President Jefferson from 1801 to 1803; leader, with William Clark, of the expedition to the Pacific of 1804 to 1806; governor of the Louisiana Territory from 1807.
• Thomas Jefferson: president from 1801 to 1809, who planned the expedition and remained Lewis's patron. His letters of 1809 to 1810 and the memoir of Lewis he wrote in 1813 for the published journals are quoted below.
• William Clark: Lewis's partner in command on the expedition; in 1809 the government's Indian agent and the militia general at St. Louis. His letters to his brother Jonathan Clark in Kentucky are the most private record of the death.
• James Madison: president from March 1809, to whom Jefferson forwarded the news and the servant.
• Frederick Bates: secretary of the Louisiana Territory, Lewis's deputy and his enemy.
• Amos Stoddard: an army major, the first American commandant at St. Louis in 1804 and a friend of Lewis, who wrote to him from Fort Pickering.
• Gilbert Russell: the army captain commanding Fort Pickering, who held Lewis there from 15 to 29 September 1809. He wrote to Jefferson twice in January 1810 and gave a statement in November 1811 whose authenticity is disputed.
• James House: a captain at Nashville who wrote to Bates on 28 September 1809 with news of Lewis's condition, thirteen days before the death.
• James Neelly: the United States agent to the Chickasaw nation, who travelled with Lewis from Fort Pickering, buried him, kept his pistols, and wrote the first report of the death to Jefferson on 18 October 1809.
• John Pernier: Lewis's servant, whom Jefferson had known for six years. The Lewis family later called him "his Spanish servant" and "a Frenchman". He died in Washington on 29 April 1810.
• Robert and Priscilla Grinder (the name is also spelt Griner): the settlers who kept the stand. He was away that night; she is the only adult witness.
• John Brahan: a captain at Nashville who wrote out Neelly's letter of 18 October 1809 and wrote three of his own the same day.
• Alexander Wilson: the ornithologist, who stopped at the stand on 6 May 1810 and wrote down what Mrs Grinder told him. His letter was printed in the Port Folio, a Philadelphia magazine, in 1812, and in the editions of his works by George Ord (1828) and Alexander Grosart (1876).
• Isaac Coles: the President's private secretary, who opened Lewis's trunks in Washington with Clark on 5 January 1810.
• John Marks and William D. Meriwether: members of Lewis's family who settled his estate; Marks went to Neelly's in the winter of 1811 to 1812 to recover the effects.
• John Christoph Süverman: a blind former servant of Jefferson's, in whose house in Washington Pernier died.
• General James Wilkinson: the army's senior officer and Lewis's predecessor as governor, long suspected and later proved to have been a paid agent of Spain. The conspiracy theories make him the instigator of a murder.

The documents most often cited are the letters of Neelly and Brahan to Jefferson of 18 October 1809; the Nashville Democratic Clarion of 20 October, the first newspaper report; the Missouri Gazette of St. Louis and the National Intelligencer of Washington, which printed the news in November; Clark's letters to his brother of 28 October and 26 November; Russell's letters to Jefferson of 4 and 31 January 1810 and his statement of 26 November 1811; Pernier's petition to Jefferson of 10 February 1810; Wilson's letter of 18 May 1810; Jefferson's memoir of 1813; an account of the night that Mrs Grinder gave to a passing stranger, a schoolteacher, about 1839, printed in 1845; and the report, presented in 1849 to 1850, of the committee appointed by the Tennessee legislature in 1848 to put a monument over the grave. Most of the letters are printed in Founders Online, the National Archives' free online edition of the founders' papers, and in The Papers of Thomas Jefferson: Retirement Series, or in Donald Jackson's Letters of the Lewis and Clark Expedition (1962; second edition 1978). Section 13 lists them with their identifiers.

The modern writers who appear most often: Dawson Phelps, the National Park Service historian of the Natchez Trace, whose article of 1956 made the documentary case for suicide; Vardis Fisher, a novelist, whose Suicide or Murder? (1962) opened the case for murder; Paul Russell Cutright, a historian of the expedition (1986); E. G. Chuinard, a surgeon and expedition historian, who argued murder in 1991; James Holmberg of the Filson Historical Society in Louisville, who edited Clark's letters to his brother (1992, 2002); John Guice, a historian at the University of Southern Mississippi, who argued murder between 1998 and 2006; Thomas Danisi, Lewis's biographer, who argues for a self-inflicted death that was not suicide (2002 to 2014); Tom Chalkley, who wrote the fullest study of Pernier (2004); Ann Rogers, on the order of the shots (2005); Michael Carrick, a firearms historian, on the pistols (2006, 2009); James Starrs, a professor of law and forensic science who led the campaign to exhume the body, and Kira Gale, an independent researcher, who argued murder together in 2009, and Gale alone for a Wilkinson conspiracy in 2015 and 2016; David Peck, a physician, on the medical questions (2009, 2022); Clay Jenkinson, on Lewis's character (2011); and Tony Turnbow, a historian of the Trace, on Neelly (2012, 2015). Their works are listed in section 13 and the strongest of them are answered in section 9. We Proceeded On, cited throughout, is the journal of the Lewis and Clark Trail Heritage Foundation, in which most of this debate has been conducted.

Three hypotheses are weighed throughout. Self-inflicted (S): Lewis fired both shots himself, whether by decision or in delirium. Homicide (H): someone else fired at least the fatal shot, whoever it was and whatever the motive. Accident (A): nobody intended the wounds. Homicide is treated as one hypothesis, though its advocates disagree about the killer; section 7 takes their versions separately, and section 8 says how a reader who holds one of them can re-enter the scoring for that case alone. The documents of 1809 to 1813 are quoted in their own words and kept apart from what later writers made of them, and the case for murder is stated in its authors' words before it is answered.

2. What the record says, in order

The table gives the dated record from Lewis's departure from St. Louis to the last statement by a contemporary. Every entry names its document. Section 13 lists the documents with their identifiers in Founders Online or their printed source.

Three entries in the table carry most of the weight in what follows. The derangement and the attempts on his life were reported before he died: House's letter of 28 September, Russell's detention of Lewis and his papers, and Neelly's provision of a man and a horse all belong to the fortnight before Grinder's Stand. The two people closest to him, Jefferson and Clark, accepted the report at once, and Clark's first reaction on reading a newspaper was fear that it was true. And the two men with the most to gain from a robbery, Neelly and Pernier, both went straight to Lewis's patrons: Neelly wrote to Jefferson within a week, and Pernier rode 600 miles to Monticello, Jefferson's house in Virginia, and then to the President's house in Washington to ask for his wages.

3. Who saw what

The table separates what each source could have known at first hand from what it reports.

Everything known about the night comes from Mrs Grinder and the two servants, and it reached the record through Neelly, who arrived after the death. Each homicide theory therefore has to hold that one or more of those three lied, and the theories differ about which. Because the suicide reading takes all three at their word, the weight in what follows falls on two things that rest on other witnesses or on outside knowledge: the state of Lewis's mind in the weeks before, and the pattern of the wounds.

4. Mrs Grinder's story, version by version

Priscilla Grinder told her story at least four times to people who wrote it down, over 36 years. The first three tellings are within seven months of the death. The table sets them side by side, with the later traditions after them.

Later traditions, none earlier than 1876: Marcus Wright's Tennessee informants (a shot, a request for water, a second shot, "a bullet-hole under the chin"); Polly Spencer via Christina Anthony (one shot after supper, Grinder "at once suspected", tried and released, "only 25 cents" found); Pease 1894 (a rifle shot, "Joshua Grinder", the house burned); the Griner slaves Pete and Malinda, interviewed about 1879 (two shots, all went together to the room, death at noon); the Lewis family (murdered by "his Spanish servant" who "was never afterwards heard from"). They are collected with their dates in the accompanying notes on the timeline and the versions (notes/C_timeline_setting_versions.md, section 3).

The stable core

The three tellings of 1809 to 1810 agree on what a listener in a nearby cabin could know: two pistol shots in the night, a wound to the head and one to the trunk, a request for water, survival until sunrise, and words to the servants that he had done it himself. They disagree on what a listener could not know or would be tempted to improve: the hour ("before midnight" against "about three"), whether she roused the servants at once or at daybreak, and whether she saw anything. Wilson's version is the only one with movement in the yard. It is also the only one in which she admits doing nothing for two hours, and that is not a detail a witness invents to her own credit. Clay Jenkinson made that point in his study of Lewis's character (2011).

The details that appear only in later tellings are these. The razor first appears in a Nashville newspaper two days after Neelly arrived; Neelly's own letter of the same week does not mention it, and Madison's letter of 30 October has a dirk instead. The path of the ball through the body appears only in 1811. Three shots, riders at dark and a servant in the dead man's clothes appear only in an account told thirty years on to a passing stranger and printed six years after that, in a newspaper 1,200 miles away. That account is the sole documentary basis for the theory that Pernier killed him. I could not get its full text (section 12); the fragments quoted by Fisher, Cutright, Chalkley and Gale are all I have read.

The moon

The historian John Guice showed in 2002 that the night was moonless and asked how Mrs Grinder could have seen Lewis stagger across the yard. I repeated the calculation for the site (35.52 N, 87.46 W): new moon on 9 October 1809; on 10 October the moon set at 18:35 local mean time, an hour after sunset, four percent lit; at three in the morning it was 62 degrees below the horizon; sunrise was at 06:02. Guice is right that there was no moonlight. What follows from it is narrower than the murder case has made it. Wilson's letter names no light at all (no moon, fire, candle or lantern), and the one telling with a visual claim is also the one that says the cabin logs were "open, and unplastered", which describes light escaping from inside the room as much as a line of sight. The visual claim is a weak part of one second-hand version. It does not touch the version Neelly carried to Nashville within the week, in which she heard the shots and woke the servants. The calculation is in code/astronomy_1809.py and data/astronomy_1809.txt.

5. The wounds and the pistols

What was described

No doctor saw the wounds; the descriptions come from Mrs Grinder, the servants and Neelly, through the writers named here. The head wound is one that "grazed the top of his head" (Clarion), in which "the ball did not take effect" (Brahan), which "glancing, was ineffectual" (Madison), which was "only making a furrow over it" (Russell), or in which "a piece of the forehead was blown off, and had exposed the brains, without having bled much" (Wilson). The trunk wound was "a little below the Breast" (Neelly), "through his intestines" (Clarion), or entered the breast and passed "downward thro' his body" to exit "low down near his back bone" (Russell). The cuts were made with a razor on "neck, arm and ham" (Clarion; the ham is the back of the thigh) or "from head to foot" (Russell), or with a dirk (Madison), or there were none (Neelly, Wilson, Jefferson). No one recorded powder marks, singeing, the position of the pistols, blood on the hands or the distribution of blood in the room. The Lewis researcher Ann Rogers showed in 2005 that the order of the shots, head first and then chest, is a presumption: Brahan wrote "first it was thought in the head" and "it is said". Every later writer on both sides has used that order.

The pistols were a "Brace of Pistols" (Neelly), which Neelly kept and never surrendered, so their calibre is unknowable. The firearms historian Michael Carrick argued in 2006 and 2009 that they were the cased pair Lewis bought from Booth in Philadelphia on 20 May 1807 for $87, gentlemen's travelling pistols with barrels of four to seven inches, and that the .69 calibre North & Cheney army pistols shown on the covers of Guice (2006) and Starrs and Gale (2009), and asserted by Gale (2016), were not with him. Two shots "immediately" one after the other from single-shot flintlocks require two loaded pistols. Wilson's Mrs Grinder has Lewis asking his servant for his powder canister on arrival; Russell has him loading the pistols in the night.

What is known about such wounds

The forensic literature does not say what happened at Grinder's Stand; it says how common each element of the reported sequence is in modern cases, which matters because the medical argument for murder since Chuinard has been that the sequence is impossible for a suicide. The series below were gathered from PubMed, the index of medical journals, from abstracts only; the identifiers are in the accompanying medical notes (notes/B_medical_forensic.md, section 3c).

Put together: two self-inflicted pistol shots are rare but well documented, and within that rare group a first shot that misses the brain, followed by a shot to the trunk, is common (five of the eleven cases in the largest series described). A tangential forehead wound is survivable with the power to act. Hours of survival with walking and speech after a trunk wound that spares the heart is ordinary. Cutting after a failed shooting is a named category. The Clarion's "through his intestines" and Russell's low exit near the backbone both point away from the heart. The surgeon E. G. Chuinard's opinion of 1991 that Lewis "could not have gone through all the perambulations described" is contradicted by every forensic pathologist who has addressed it since, including the one who testified at the 1996 inquest.

The same table must be read from the other side. More than one shot is the rule in homicides. Nobody looked for soot or described the hands. The razor cuts on "neck, arm and ham" read as a man seeking his own arteries, but they could also be read as an attack, and no one recorded whether they were shallow and parallel, as hesitation cuts are, or on the hands and forearms, as defensive cuts are. The physical evidence therefore excludes neither reading, but it does establish two things: that the sequence the witnesses described is one that suicides produce, and that the point on which the murder case has rested medically since 1991 is wrong.

An exhumation could still answer part of this. The frontal bone would show a groove or a hole; ribs or vertebrae could fix the trunk trajectory; a retained ball would give the calibre and a wad the range. The grave was opened in 1848 by the monument committee, which reported only that "the upper part of the skeleton" was examined and "such evidences found as to leave no doubt of the place of interment". Since the landscaping of 1935 and the re-founding of the monument on concrete in 2000, it probably lies under about four feet of fill and a concrete base. The Department of the Interior's final refusal to open it is dated 25 August 2010.

6. The state of his mind

The pressures

By the summer of 1809 Lewis was in an open quarrel with his territorial secretary, Frederick Bates, who wrote to his brother on 14 July that he lamented "the unpopularity of the Governor". On 15 July the War Department refused to honour drafts Lewis had drawn for public purposes. The refusal reached St. Louis in August. Lewis told Major Amos Stoddard, his friend and the first American commandant at St. Louis, on 22 September what it had done: "the protest of a draught however just has drawn upon me at one moment all my private debts which have excessively embarrassed me." He was carrying the vouchers to Washington to "put all matters right". The journals of the expedition, three years after his return, were unwritten, and Jefferson had asked when they would be ready. Clark, who saw him in August, wrote to his brother that Lewis was distressed, and on reading the newspaper in October his first thought was "the weight of his mind has overcome him".

The illness

The Missouri Gazette, the St. Louis newspaper, reported that Lewis left St. Louis "afflicted with fever" (2 Nov 1809, as quoted by Danisi). He wrote from the Bluffs that he was "very much exhausted from the heat of the climate, but having taken medicine feel much better". Russell, who had him for a fortnight, said he arrived "in a state of mental derangement", was "stoped until he would recover", and "in about Six days" was "perfectly restored". Russell's 1811 statement adds that the derangement "appeared to have been produced as much by indisposition as by other causes", that "By much severe depletion during his illness he had been considerably reduced and debilitated", and that "in three or four days" after leaving "he was again affected with the same mental disease". Stoddard's letter of 28 September, printed in St. Louis, has him "under the hands of the surgeons mate Dr. Smith", the only physician known to have attended him. The Missouri Gazette, reporting the death in St. Louis, where his illness had been seen, added: "The governor has been of late very much afflicted with fever, which never failed of depriving him of his reason." Neelly found him "at times deranged in mind" in the Chickasaw nation. The Nashville Clarion spoke of "a deranging malady for about six weeks". Mrs Grinder's Lewis paced, talked to himself, flushed "as if it had come on him in a fit", and was courteous between the fits.

The attempts on the boat

The murder case treats the suicide attempts on the river as an invention of Russell's 1811 statement, which a document examiner has called a forgery (section 9), but the attempts are attested four times, and three of the four do not depend on the 1811 statement. Only House's chain is certainly independent of Russell as a source: Clark's summary is of Russell's own letters, and the Intelligencer's informant is not named.

1. James House to Frederick Bates, Nashville, 28 September 1809, thirteen days before the death: Lewis "had made several attempts to put an end to his own existence" and Russell was "keeping a strict watch over him to prevent his committing violence on himself". House had it from Stoddard, who had it from a man "immediately from the Bluffs". It is hearsay, but it was written before the death and so cannot have been shaped by it.
2. Russell's lost letters to Clark, as Clark summarised them to his brother on 26 November 1809: Lewis "had attempted to kill himself before he got there", and "made his will at the Bluffs and left Wm. Merriwether & myself Execiters".
3. The National Intelligencer, 27 November 1809, from information reaching Washington: "he became deranged in his mind, and made some attempts on his life; but the people in the boat secured him".
4. Russell's statement, 26 November 1811: the crew told him "that he had made two attempts to Kill himself, in one of which he had nearly succeeded".

To make these a cover story, the conspiracy must include House, Stoddard, an unnamed traveller, Russell, whoever informed the Intelligencer and Clark's correspondent, and it must have begun before Lewis died. Gale (2016) accepts the consequence and names House and Bates as conspirators.

What was the matter with him

Five diagnoses have been argued. The table gives each in its strongest form and its best objection, both from the sources I read. The detail, with quotations and page references, is in the accompanying medical notes (notes/B_medical_forensic.md, section 2).

Jefferson's sentence about the family is the only statement of heredity: "a constitutional disposition in all the nearer branches of the family of his name, and was more immediately inherited by him from his father." Flick (2021) found in the daybook of Dr George Gilmer, the Virginia physician who treated the family, that William Lewis, the father, was repeatedly prescribed purgatives and asafoetida in the 1770s, which were then treatments for melancholia among much else. Lewis's half-brother Dr John Marks was confined for insanity in 1819 (Chuinard, citing Luttig's journal). These records do not support a diagnosis, but they show what family history Jefferson had in mind when he wrote the sentence.

The verdict does not depend on which diagnosis is right. Whether the derangement was depressive, febrile, toxic or a mixture, the man described by House, Russell, Neelly, Clark and Mrs Grinder was one in whom attempts on his own life had been observed and guarded against for a month before the night at Grinder's. That observation is what the weighing in section 8 uses, and it concerns the weeks before the death rather than the night.

7. The homicide theories, one by one

Each theory is stated with the evidence its proponents cite, then tested against the record. Where a theory's own case is quoted, the quotation is marked at the left.

The Grinders

"Old Grinder, who was of Indian blood, was at once suspected of the murder, ran away, was captured on Cane creek, brought back, and tried; but the proof not being positive, he was released. Only 25 cents was found on the person of Captain Lewis after he was shot. Old Grinder soon afterward removed to the western part of the State, and it was reported in his old neighborhood had bought a number of slaves and a farm."

James D. Park, Nashville American, 6 September 1891, from Christina Anthony's memory of the servant girl Polly Spencer; reprinted in Coues 1893. The earliest print of the Grinder accusation I found.

Every 1809 account that mentions the household has Robert Grinder away from home that night. No 1809 source suspects the household. The tradition of a trial rests on Park's 1891 informant and Pease's 1894 magazine piece, which Pease himself conceded represented "the local stories and traditions"; no inquest or court record has ever been found, and the county's own historians in 1894, 1900 and 1904 recorded a counter-tradition, from Colonel Robert M. Cooper, who knew the district from the 1850s, that "Grinder was not suspected ... Grinder did not flee from the county". The prosperity is documented and modest: Spence's county history (1900) has Griner back in the Duck River bend where he had first settled, buying 100 acres for $250 in 1814 and another 100 in 1824, with sons who became militia officers. Pease's "plantations and slaves" are not in the deeds. Against the household stand two facts: the servants who found Lewis reported his words that he had done it himself, and the woman who told Wilson the fullest version told it to her own discredit. As noted in section 4, a household that had killed him could have reported that he was found dead, and what she reported instead was that she left a wounded man for two hours because she was afraid.

John Pernier

"Have you ever heard of the report, that Governor Lewis did not destroy his own life, but was murdered by his servant, a Frenchman, who stole his money and horses, returned to Natchez, and was never afterwards heard of?"

Meriwether Lewis Clark, William Clark's son, to the Rev. Mr Cressey, quoted in the Tennessee monument committee's report of 1849–50 as the report's stated ground, beside the loneliness of the place, for finding assassination "more probable".

This is the theory the monument committee endorsed in its report of 1849 to 1850, the Lewis family repeated in 1904, and the 1845 account (Pernier in Lewis's clothes with a gold watch) supports. Every clause of the sentence is false on the documents. Pernier did not return to Natchez; he rode with Neelly to Nashville, was sent on by Neelly with fifteen dollars, reached Monticello by 21 November, carried Jefferson's letter to the President, petitioned Jefferson on 10 February 1810 for $271.50 in unpaid wages, "resqued from prison", and died in Washington on 29 April 1810 in the house of a blind ex-servant of Jefferson's. He did not steal the horses: Neelly reported one horse lost in the wilderness and sent Pernier on the other, and Jefferson told him to sell it and credit the estate. He did not vanish with a watch. Neelly listed "his Rifle, Silver watch, Brace of Pistols, dirk & tomahawk" among the effects in his own care on 18 October 1809, and Brahan wrote the same day that Neelly "has got his Silver watch". In January 1810 Coles listed "His Watch" among the articles of value in Washington, without saying which of the three trunks he opened it came from. In January 1812 an unsigned note in the Lewis-Marks papers said that Neelly, who still carried the pistols and dirk, had Lewis's "gold watch" (Gale 2016, n. 30; I could not get the note). Whether these letters describe one watch or two cannot be settled from the copies I read. Gale reads the 1812 watch as the one of the 1845 story, recovered from Pernier; no document of 1809 puts any watch with Pernier, and the one that names it puts it with Neelly a week after the death. The one contemporary sentence against him, Russell's of 31 January 1810, sits in a paragraph about Neelly giving Lewis "every chance to seek an opportunity to destroy himself"; the Retirement Series editors read it as "blaming the suicide on Pernier's negligence", and Russell's 1811 statement dropped it. Pernier's own death by laudanum six months on can be read as guilt or as the end of a destitute man whose master had "followed" the same road first; Jefferson, who had known Pernier for six years, read it the second way. Chalkley (2004) reached the same conclusion from the documents then in print, and Pernier's petition, read in full, adds his own voice to it. Gale (2016) read the same petition, and the letter of John Christoph Süverman in whose house Pernier died, as the record of a witness silenced by General James Wilkinson, the army's senior officer, whom the conspiracy theories make the instigator; neither document names a threat, and the account they give is of wages, debt and sickness.

James Neelly

"Neelly, who had first sought and loaded Lewis's pistols, fired in the dark toward Lewis's raising body, and this shot barely grooved Lewis's scalp, but momentarily stunned him ... The second shot entered the chest and went downward to exit in the low back, a course explained by the fact that Lewis was in the semi-upright position of coming erect."

E. G. Chuinard, "How Did Meriwether Lewis Die? It Was Murder", We Proceeded On, 1991–92.

Neelly is the least reliable of the witnesses, and Fisher, Chuinard and Turnbow have each cast him as the killer or as the man who covered the killing up. He was, by his own account, absent; he kept the pistols, dirk, rifle and a horse and never gave them up; he claimed to have lent Lewis money, which Russell denied ("he had none himself"); he was in debt; Tony Turnbow, a historian of the Trace, found in 2012 that he was named as defendant in a suit at Franklin, Tennessee, with a hearing on 11 October 1809, and a second letter of 18 October datelined from the Chickasaw Agency, about 200 miles from Nashville by the Trace. Thomas Danisi, Lewis's biographer, answered in 2013 and 2014 that the agent was a man of 68 whom the Chickasaw chiefs had asked for as "an old Gentleman", that he drew $100 in cash on 24 September from David Hogg, the government's trader (the factor) at the Bluffs, so Russell's "none himself" is wrong, that he had retained the attorney Jenkin Whiteside, that the court minute's formula does not show who was present, that rheumatism explains letters in other hands, and that the folklore route by the Agency house is unproven. That dispute is unresolved, and I leave it so, because it does not bear on the death. On any reading Neelly was not in the room, and his account of the night is Mrs Grinder's and the servants'. For Neelly to be the killer, the woman and both servants must have lied or been deceived about a man they found alive and talking, and Neelly must have ridden back in the night to shoot a companion who was riding to Washington in his company anyway, then written to Jefferson within the week. Keeping a dead man's pistols is evidence of dishonesty about property, and nothing in the record connects it to the shooting.

Robbers on the Trace

The Trace had a reputation, and Guice (2006) argued that it was less safe in 1809 than Dawson Phelps had said in 1956. The candidates named are "land pirates", the Harpes (active in the 1790s), John Murrell (a child in 1809), and "Tom Runions", who first appears in Richard Dillon's biography of Lewis of 1965 with moccasin tracks and a rifle-butt print at the cabin; no earlier source names him, and Gale (2016) makes him Grinder's nephew. The post-rider Robert O. Smith, who in the same tradition found the body by the road, was born about 1806 (Holmberg 2006). Against all of them: no 1809 source mentions strangers at the stand; the servants, 200 yards off, came when called or at daylight and reported no one else; and the robbers left the brace of pistols, the rifle, the dirk, the tomahawk, the silver watch, the horses and two trunks of papers. What was taken, if anything was taken, is an unknown part of "more than One hund $ in Notes & Specia" that Russell had heard "it is said could be found" nowhere; the check he worried about reached Washington. I found no documented robbery with murder near Grinder's in 1808 to 1810 in the sources listed in section 13, and I could not get Phelps's article.

Political assassination

"There are two clues in this letter indicating House and Bates were involved in the conspiracy. First, the identity of the man who saved Lewis from 'several attempts to put an end to his own existence' is not revealed. The real purpose of the letter ... is to tell Bates where the key to the locked room could be found."

Kira Gale, "A Cold Case Mystery", We Proceeded On 42:4 (November 2016), on House's letter of 28 September 1809.

The conspiracy theories (David Leon Chandler's The Jefferson Conspiracies, 1994; Gale 2015 and 2016; Turnbow's later work on General Wilkinson) hold that Lewis was killed for what his papers contained, and that the suicide story was arranged. Their evidence is the absence of an inquiry, the interest of Wilkinson and Bates in Lewis's removal, Jefferson's readiness to accept suicide, and the forgery claim against Russell's statement. Their difficulty is the one that defeats the story of Pernier's flight to Natchez: the documents contradict them. To hold the theory one must treat House's letter, Russell's two letters and statement, Neelly's and Brahan's letters, the Clarion's report, Mrs Grinder's four tellings, Pernier's conduct, Clark's private letters to his brother and Jefferson's private letters to Madison as coordinated falsehoods, and must explain why the coordination began a fortnight before the death; no document supports any part of this. Wilkinson himself was in New Orleans from March 1809 to February 1810 (Danisi 2009), and the two Russell letters that Clark received in November 1809, which the theory calls forgeries, have not been seen by anyone since Clark read them. The absence of a federal inquiry has a plain explanation in the sources: Neelly, who buried him, reported the death as suicide; Russell confirmed the derangement; and Jefferson and Clark, who knew him best, accepted it.

Accident

A single accidental discharge while loading or handling a pistol in the night could account for one wound, but not for two shots from two pistols, for the razor cuts, or for the words "I have done the business". The model's central run gives a simple accident under one tenth of one percent, and the sceptical runs lift it to 2 percent under the even prior and to at most 2.5 percent under the murder-leaning prior; only the flat prior, which starts it at a third and then shrinks every item to a quarter, lifts it to 17 percent. The verdict's "under 1 percent" is my judgment on the two shots, the razor and the words, conditional on the items as entered.

Within self-infliction

Two positions remain that accept the shots as Lewis's own and dispute the word suicide. Danisi holds that the acts were "self-surgery" in a malarial delirium: the man did not mean to die. Peck (2009) holds that they were the end of a depressive and alcoholic decline and that Lewis meant to die. The record cannot separate a decision from a delirium, and the dying words go both ways: "I have done the business" and "so hard to die" sound like a man who meant it; the request to "heal my wounds" sounds like one who did not. The verdict as stated covers both positions.

8. How I weighed it

This section explains the scoring that produced the figures in the verdict panel. It is a Bayesian tally, which means that each piece of evidence is scored by how likely it would be under each hypothesis, the scores are multiplied together, and a starting assumption (the prior) is updated by the product. The scores are my judgments. Sixteen evidence items were drawn from the record above. For each, I entered a range for the probability of seeing that evidence if Lewis shot himself (S), if he was killed (H), and if it was an accident (A). Only the ratios matter. The model picks each score at random within its range (evenly on a logarithmic scale), multiplies across items, applies a prior and rescales the three results to sum to one, 100,000 times, and reports the middle value and the central 90 percent of the results. Items that rest on one underlying fact or one chain of transmission are grouped, and in the conservative rule reported here only one item per group is used in each draw, chosen at random. Two sceptical settings shrink every ratio's distance from one by a half and by three quarters. Leaving each item out in turn shows what it carries. The code is code/evidence_model.py; the inputs, with a "basis" line naming the files behind each item, are data/evidence_items.json; the outputs are data/evidence_model_results_one.json and ..._product.json.

Groups: E1 to E5 (state of mind) count once per draw; E7 and E8 (the Grinder chain) count once per draw. The last column is the median posterior for S under the even prior when that item is removed; the full-set figure is 92 percent.

Several features of the table deserve comment. The state-of-mind block is the heaviest, and it is heavy because it comes through separate chains from the month before the death: the man from the Bluffs whose report House wrote down on 28 September, Russell at Fort Pickering, Neelly on the road, Pernier as Clark reported him, and Mrs Grinder at the stand. Only House's was written before the death, and only Russell's is first-hand. The two items that favour homicide most, the money and the later tradition, are weak on inspection: the check that "could not be found" reached Washington, and one of the monument committee's two grounds was a sentence about Pernier that the documents refute. The wound pattern is entered as nearly neutral, with the widest range in the table, because the base rates cut both ways and the facts that would decide it were never recorded. The valuables left with the body carry more than any other single item against robbery. No single item carries the verdict, as the leave-one-out figures in the panel show.

The model has four limits. It cannot allow for collusion: if Mrs Grinder, both servants and Neelly lied together, most of the items in the table fall away, and no number in it measures how likely that is. It cannot allow for the absence of examination: no one looked for soot or powder marks, and the model gives an unrecorded finding no weight either way. It treats the sub-hypotheses of homicide as one, so a reader who believes only in a Grinder robbery, or only in a political killing, should re-enter the items for that case alone; the item notes say which way each cuts. And the ranges are wide by design, which is why the 90 percent bands are wide. The central figure is the middle value of 100,000 combinations of judged scores, and it should be read as a summary of those judgments.

9. The strongest case against, answered by name

I read the works below in full unless I say otherwise. Each is given its strongest point and an answer from the record.

Where the field stands. Among authors who have stated a position in print since 2006, I count 13 for suicide (Holmberg, Jenkinson, West, Nicandri, Peck and Peck, Flick, Rogers, Chalkley, Merritt, Kirkpatrick, Crosby, Bartley, Morris), 6 for murder in three versions that exclude one another (Guice's unknown outlaw; Turnbow's robber, with the suicide story as a cover for Neelly; the Wilkinson conspiracy of Gale, Stroud, Mandrell, and Schrag and Haze), 3 for a self-inflicted death that was not suicide (Danisi, Jackson, J. W. Fisher) and 6 or 7 undecided (Starrs, Buckley, Bartsch, Strickland, the Lewis and Clark Trail Heritage Foundation's board in 2008, Fausz and Gavin). Of the six who have worked most with new documents (Holmberg, Jenkinson, Danisi, Guice, Turnbow, Gale), two say suicide, one says accident and three say murder. At the 2009 bicentennial panel the historians put suicide at 96 percent and the audience voted 350 to 20 for murder (Holmberg and Jenkinson, WPO 50:3). The table, with each author's basis, is in the accompanying notes on the scholarship (notes/D_scholarship_theories.md, section 7c). "Agrees with the majority" on this page means that majority and the documentary case behind it, in a field that is not settled.

Vardis Fisher, a novelist: Suicide or Murder? (1962)

I read this only in snippets (section 12). Fisher's method, as reported by Chuinard, Cutright and Guice, is to show that every document is second-hand and every witness interested, and that the physical story is incredible; he named Pernier, Neelly, the Grinders and Russell as possible killers and reached no verdict. The first point is right, and this page accordingly gives most weight to the evidence from before the death and to base rates. The second is answered in section 5.

E. G. Chuinard, a surgeon: "It Was Murder" (1991–92)

I read this in full. His chief evidence is surgical: that Lewis could not have survived the second shot for two hours and moved as described, and that the absence of recorded powder marks and blood trails shows a cover-up. The first is answered by the forensic pathologists who have addressed it (Francisco at the 1996 inquest; Wagner, Carrison, Leo and Mortenson via Peck and Peck 2022; Peck 2009; Danisi's trauma surgeons) and by the case series in section 5. The second treats the absence of any record of powder marks or blood as if someone had recorded that there were none; in fact Brahan has the servants find him "weltering in his blood". Chuinard's malaria is discussed in section 6; his Neelly reconstruction in section 7.

John D. W. Guice, a historian at the University of Southern Mississippi: "A Fatal Rendezvous" (1998), "Moonlight and Meriwether Lewis" (2002), By His Own Hand? (2006, ed.)

I read the articles in full and the book in snippets. Guice's case is cumulative: no autopsy or inquest; Neelly unreliable and interested; Mrs Grinder's changing story and the moonless night; the two-hour survival; Jefferson's readiness to believe the worst of a man he thought "hypochondriac"; a Trace less safe than Phelps allowed; a coroner's jury that found "little evidence to support the suicide theory". Each is engaged above: the moon (section 4), the survival (section 5), Neelly (section 7), the Trace (section 7). Jefferson's readiness is real, and it is evidence in the other direction too: he had watched Lewis's depressions at close quarters for two years. The coroner's jury heard a case coordinated by the exhumation's proponents, and its stated reason for exhumation was the importance of the person "to the history of Lewis County". Guice's own summary of the odds, from a psychologist, was "60-40" against suicide; the sceptical runs in section 8 show which discounts produce a figure like that.

James Starrs, a professor of law and forensic science, and Kira Gale, an independent researcher: The Death of Meriwether Lewis: A Historic Crime Scene Investigation (2009); Gale, "A Cold Case Mystery" (2016)

The book in snippets; the article and Gale's letters in full. The specific claims are: that the pistols were .69 calibre army pistols (Carrick shows they were not); that Russell's statement of 1811 is a forgery, on the testimony of the former FBI examiner Gerald Richards that it "was neither written nor signed by Captain Gilbert Russell or Major Jonathan Williams" (Donald Jackson attributed the hand to Jonathan Williams, the army colonel who took the statement, and the statement's substance is attested three times elsewhere, section 6); that Russell's two accounts of the length of the derangement, "about Six days" and "15 Days", cannot come from one man (Clark's summary of Russell's lost letters gives "15 Days in a State of Derangement most of the time", so both figures are Russell's, one for the acute phase and one for the detention); and that House, Bates, Wilkinson and Jefferson formed a conspiracy. The last is answered in section 7: it requires the record to be false from a fortnight before the death. On the forgery, Richards's testimony as quoted by Guice (2006) and Gale is that neither the text nor the two signatures on the document, which is at Indiana University's Lilly Library, are in Russell's or Williams's hand, which is what a clerk's copy of a statement and its endorsement looks like; Fisher had already called it "copied from a copy", and Richards is nowhere quoted as saying the content was invented. Williams and Russell were both at Frederick, Maryland, in November 1811 for Wilkinson's court-martial, which is also the plain explanation of how Williams came to take a statement from Russell there.

Tony Turnbow, "The Man Who Abandoned Lewis" (2012) and reply (2015); Thomas Danisi, "The Real James Neelly" (2013–14)

I read all of these in full. Turnbow's documentary finds (the Franklin suit, the Agency dateline, Neelly's debts and later charges) are real, and so are Danisi's answers (the agent's age and commission, the $100 draft, the attorney, the court formula); whether Neelly reached Grinder's on 11 October or later is unresolved. For the verdict, Neelly's whereabouts on the 11th do not bear on the night of the 10th, for which he was never a witness.

The Tennessee monument committee (appointed 1848, report 1849–50) and the Lewis family

The committee's sentence, "It seems to be more probable that he died by the hands of an assassin", has been quoted by a federal court and every murder advocate since. Its stated grounds are the loneliness of the place and a letter from Clark's son repeating the story that the servant "stole his money and horses, returned to Natchez, and was never afterwards heard of". Section 7 shows that story to be false in each particular. The committee opened the grave and recorded nothing about the skull beyond identification. The family tradition (Wheeler 1904) adds a gold watch, seen years later with a man taken for "the Spanish servant", and papers that vanished; the 1809 letters put the watch in Neelly's care at Nashville a week after the death (section 7), and the papers went to Philadelphia for the 1814 edition.

Thomas Danisi, "self-surgery, not suicide" (2002, 2009, 2012)

Danisi accepts that Lewis fired the shots and argues that malaria, and no wish to die, moved his hand. Loge's, Peck's and Crosby's objections are in section 6. Whether or not the diagnosis holds, Danisi's position is one of the two that the verdict's "self-inflicted" covers.

10. What would overturn this verdict

• The skeleton. A frontal bone showing a penetrating entry from behind or above, rather than a tangential groove; rib or vertebral damage showing a trunk trajectory that a man cannot give himself; a retained ball of a calibre no gentleman's pistol fired; a wad or soot pattern showing a shot from beyond arm's reach. Any of these would end the argument. The Department of the Interior refused exhumation on 25 August 2010 as a "final decision"; no later application, suit or bill was found, and the family's site has since lapsed.
• A record of an 1809 inquest in Hickman or Maury County naming Grinder or anyone else, or the "Whiteside" jury of local tradition. None has been found in 130 years of looking, and the county sources of 1894 to 1904 say the papers are lost or never existed.
• The full text of the 1845 account showing that Mrs Grinder's third version contains checkable detail, rather than the accretions its fragments suggest.
• House's letter proving misdated, or Stoddard's informant proving to be Neelly or Russell coaching the story, which would remove the only pre-death attestation of the attempts. The original is in the Missouri History Museum and has been printed by Fisher and Danisi.
• A document from Pernier's last months showing money or goods of Lewis's in his hands beyond the horse he was told to sell.
• Evidence that Neelly was at Grinder's on the night of the 10th, which would make him a witness and a suspect at once; no document places him there.

11. What is new, and what agrees with earlier work

New, as far as I can tell from what I read

• Pernier's petition of 10 February 1810, read in full from the text in the Retirement Series of Jefferson's papers. Chalkley (2004), the fullest study of Pernier, knew of the wage claim through Jefferson's letter to Madison and of the death through Jefferson's letter to Meriwether; the petition itself, with its account, its "refused" and its "resqued from prison", was not used by Chalkley. Gale (2016) used its $271.50 and its fear of prison in the opposite reading, so what is new here is the full reading and the sequence, and the document itself is not.
• The three attestations of the boat attempts that do not depend on Russell's 1811 statement, and the observation that one of them predates the death, as an answer to the forgery thesis. House's letter has been quoted since Fisher; its bearing on the forgery claim is drawn here.
• The base rates for multiple-shot suicides, unplanned complex suicides and capability to act, applied to the reported wound pattern (gathered from the forensic literature, section 5).
• The evidence model with its sensitivity, leave-one-out and grouping printed beside the headline.
• Small points about dates and documents. Wilson's visit was Sunday 6 May 1810; the Port Folio's "May 28th, 1811" led Jenkinson (2011) to place it "sixteen months later"; Ord (1828) and Grosart (1876) had already dated the letter 18 May 1810, and the day is fixed by the letter's own "Friday morning, May 4th". Russell's check "none of which it is said could be found" reached Washington (Retirement Series editors, citing Jackson). The monument committee's "assassin" (report of 1849 to 1850) rests on two stated grounds, the loneliness of the place and a story about the servant that the documents refute. The Missouri Gazette's "fever, which never failed of depriving him of his reason" and Stoddard's "surgeons mate Dr. Smith" are contemporary St. Louis testimony that none of the books or articles listed in section 13 uses.

Agrees with earlier work, with credit

• Dawson Phelps, the National Park Service historian of the Trace (1956), and Donald Jackson, editor of the expedition's letters (1962, 1978): the documentary case for suicide, and the texts on which later writers have relied.
• James Holmberg of the Filson Historical Society (1992, 2002): Clark's letters to his brother, which supply the "too much truth" and the summary of Russell's lost letters.
• John Guice (2002): the moonless night. My calculation agrees with his figures for the site.
• Ann Rogers (2005): the shot sequence is a presumption.
• Michael Carrick (2006, 2009): the pistols were gentlemen's travelling pistols, calibre unknowable, and no army .69s.
• Thomas Chalkley (2004): the documentary case against Pernier as killer.
• David Peck, a physician (2009, 2022): the medical answer to Chuinard; the first to put probabilities on the medical questions.
• Thomas Danisi (2013–14): the agent's age, commission and finances; and Tony Turnbow (2012): the Franklin suit, which keeps the Neelly question open.
• Clay Jenkinson (2011): the reading of Wilson's Mrs Grinder as a witness against her own interest.
• The four sets of working notes that accompany this page, on the documents, the medicine, the timeline and the versions, and the scholarship (notes/A_documents_report.md, notes/B_medical_forensic.md, notes/C_timeline_setting_versions.md, notes/D_scholarship_theories.md), on which sections 2 to 7 and 9 draw throughout.

12. What I could not read

• Not read in full: Fisher 1962; Guice ed. 2006; Starrs and Gale 2009; Danisi and Jackson 2009; Danisi 2012; Gale 2015; Dillon 1965; Ambrose 1996; Kushner 1981; Phelps 1956; Donald Jackson's Letters of the Lewis and Clark Expedition (the inventory of Lewis's effects taken at Nashville by Thomas Freeman on 23 November 1809, and the full Russell statement, were read only in excerpts); Chandler 1994; Peck and Peck 2021; Stroud 2018; Morris 2004; Jenkinson 2000 and 2011 beyond chapter VI; Turnbow's two books. I read everything on the death in We Proceeded On, the journal of the Lewis and Clark Trail Heritage Foundation, from 1974 to 2025 (index in data/wpo_death_index.tsv), and the books only as complete sets of search-hit fragments, so I am sure of the position each takes and not of its page-level argument. The digital libraries JSTOR, HathiTrust and Google Books were out of reach, and the lending libraries' copies gave only search snippets. So the principal opposing works were read in full where they are articles (Chuinard, Guice's articles, Gale 2016, Turnbow, Danisi's articles, Peck, Holmberg, Rogers, Carrick, Chalkley) and only in fragments where they are books.
• Not obtained at all: the 1845 account's full text (a manuscript version is in the Draper manuscripts at the Wisconsin Historical Society, 29CC33, per Morris 2004 and Starrs and Gale 2009); the 1996 inquest transcript of 364 pages, with the testimony of Fackler, Haag and Bass, known only through Starrs and Gale's excerpts and the WPO reports; the Avery thesis of 1978; the print-only WPO issues of November 2025 and February 2026; House's letter beyond the quoted phrases; the Clarion's own issue (read in the Kentucky Gazette and Enquirer reprints); the Tennessee House Journal appendix of 1849–50 (read in the transcription that Elliott Coues verified for his 1893 edition of the expedition's history); Richards's testimony except as Gale reports it; the Marks and Neelly correspondence of 1811–12 except as Fisher and Gale quote it; the will Lewis wrote at New Madrid on 11 September 1809.
• Forensic literature read in abstracts only; the base rates are from modern populations with modern weapons and are used as orders of magnitude rather than as measurements of 1809.
• No archival work. Everything here is from digitised or printed sources. I did not consult the Bates Papers at the Missouri History Museum, the Draper manuscripts in Wisconsin, the Grace Lewis Miller papers (a twentieth-century researcher's collection) or the Tennessee State Library's microfilm.
• The model rests on judged inputs. They can be argued item by item, and the wide bands show how far that argument can go.
• Transcription. The Kentucky Gazette and Missouri Gazette texts were read from page images and are not verified by machine text recognition (OCR); the Clarion phrases "he had now done for himself" and "about 7 o'clock" were checked against the Enquirer reprint as well.
• Discrepancies. The accompanying notes on the documents (notes/A_documents_report.md) list 24 points on which the record of 1809 to 1813 disagrees with itself and 20 documents I could not get. The points that bear on the verdict (the hour of the shots, the wounds and the razor, when the servants came, the earlier attempts, the effects and the watch, the burial date, the date of Wilson's letter) are treated in sections 2 to 7. The rest (the distance from Nashville, the state of the journals, Neelly's dates of office) do not change a conclusion here.

13. Sources

Primary documents (saved copies in the sources/ folder that accompanies this page, each with citation, URL and the date fetched)

• Frederick Bates to Richard Bates, 14 July and 9 Nov 1809. Marshall, ed., Life and Papers of Frederick Bates (1926), vol. 2. archive.org
• Meriwether Lewis to James Madison, Chickasaw Bluffs, 16 Sept 1809. Founders Online, Madison/03-01-02-0420
• Lewis to Amos Stoddard, 22 Sept 1809. Thwaites, Original Journals 7:368. archive.org
• James House to Frederick Bates, 28 Sept 1809. Bates Papers, Missouri History Museum; quoted in Fisher 1962 p. 87, Danisi and Danisi, WPO 38:4 (2012), Gale, WPO 42:4 (2016). Fragments only.
• Amos Stoddard, Nashville, 28 Sept 1809, extract in the Missouri Gazette, 19 Oct 1809, p. 3; and the Gazette's notices of 4 Oct and 2 Nov 1809. Read from microfilm frames, archive.org item st-louis-la-mo-2-222.
• James Neelly to Thomas Jefferson, Nashville, 18 Oct 1809. Jefferson/03-01-02-0478
• John Brahan to Jefferson, 18 Oct 1809. Jefferson/03-01-02-0476; Brahan to Stoddard, same day, Thwaites 7:389.
• Nashville Democratic Clarion, 20 Oct 1809, as reprinted in the Kentucky Gazette, 31 Oct 1809, p. 3 (Chronicling America), and the Richmond Enquirer, 17 Nov 1809, p. 2.
• James Madison to Jefferson, 30 Oct 1809. Jefferson/03-01-02-0497
• Jefferson to C. & A. Conrad, 23 Nov and 11 Dec 1809. 03-02-02-0011, 0045
• Jefferson to Madison, 26 Nov 1809. Jefferson/03-02-02-0015; Madison to Jefferson, 27 Nov and 11 Dec 1809.
• National Intelligencer, 27 Nov 1809, p. 3. Chronicling America
• National Intelligencer, 13 Nov 1809, p. 3 (first notice) and 15 Nov 1809, p. 1 (the Russellville letter of 20 Oct, from the Lexington paper of 28 Oct). Chronicling America, read from the page images.
• William Clark to Jonathan Clark, 26 Aug, 28 Oct, c. 8 Nov and 26 Nov 1809, as printed by Holmberg, WPO 18:4 (Nov 1992). PDF
• Gilbert C. Russell to Jefferson, 4 Jan 1810 (03-02-02-0088) and 31 Jan 1810 (0153); Jefferson to Russell, 18 Apr 1810 (0290).
• Isaac A. Coles to Jefferson, 5 Jan 1810. 03-02-02-0090
• John Pernier to Jefferson, 10 Feb 1810. 03-02-02-0171
• William Dickson to Jefferson, 20 Feb 1810 (0199); Jefferson to Dickson, 20 Apr 1810 (0295); W. D. Meriwether to Jefferson, 14 Mar 1810 (0240).
• John Christoph Süverman to Jefferson, 5 May 1810. 03-02-02-0321; Jefferson to W. D. Meriwether, 21 Aug 1810, 03-03-02-0020.
• Alexander Wilson to Alexander Lawson, Natchez, 18 May 1810: The Port Folio, Jan 1812, pp. 34–47 (archive.org); Ord, Sketch of the Life of Alexander Wilson (1828); Grosart, Poems and Literary Prose of Alexander Wilson (1876).
• Russell's statement, 26 Nov 1811, Jackson 2:573–75, as excerpted in WPO 28:1 (Feb 2002). PDF
• Jefferson to Paul Allen, 18 Aug 1813 (the memoir of Lewis). 03-06-02-0341-0002
• Report of the Lewis Monument Committee, Tennessee, 1849–50, in Coues, History of the Expedition (1893), 1:lvii–lxii. archive.org
• Marks to Reuben Lewis, 22 Jan 1812, and an unsigned note on the pistols and gold watch, Lewis-Marks Collection, University of Virginia Library. Neither obtained; quoted by Carrick, WPO 32:1 (2006), from Fisher 1962 p. 136, and by Gale, WPO 42:4 (2016), n. 30.

Scholarship read (saved under sources/scholarship/ and sources/medical/; WPO = We Proceeded On, lewisandclark.org/wpo)

• Bakeless, Partners in Discovery (1947), death chapter. Bartsch, "Not Manic, Not an Alcoholic", WPO 51:3 (2025). Carrick, letter, WPO 32:1 (2006); "Meriwether Lewis's Pistols", WPO 35:4 (2009). Chalkley, "The Valet Did It?", WPO 30:4 (2004); reviews, WPO 35:3 (2009). Chuinard, "How Did Meriwether Lewis Die? It Was Murder", WPO 17:3, 17:4, 18:1 (1991–92). Coues 1893, memoir notes. Crosby, letter, WPO 39:3 (2013). Cutright, "Rest, Rest, Perturbed Spirit", WPO 12:1 (1986). Danisi, "Meriwether Lewis's Death: Malaria", WPO 28:1 (2002); letters 2009–13; "The Chickasaw Agency Letter", WPO 39:1 (2013); "The Real James Neelly", WPO 40:4 (2014). Danisi and Danisi, "Uncovering Jefferson's Account", WPO 38:4 (2012). Flick, "The Gilmer Daybook", WPO 47:2 (2021). Gale, letters 2008–13; "A Cold Case Mystery", WPO 42:4 (2016). Giesecke, "1809 Newspapers", WPO 35:3 (2009). Guice, "A Fatal Rendezvous", WPO 24:2 (1998); "Moonlight and Meriwether Lewis", WPO 28:1 (2002); review of Gale, WPO 42:1 (2016). Holmberg, "I Wish You to See & Know All", WPO 18:4 (1992); "Writing What Deserves to Be Read", WPO 50:1 (2024). Jenkinson, The Character of Meriwether Lewis (2011), ch. 6. Loge, "Meriwether Lewis and Malaria", WPO 28:2 (2002). Moore, "The Death of Meriwether Lewis", American Historical Magazine (1904). Peck, "Death Myths and Probabilities", WPO 35:4 (2009); letters 2010–13; Peck and Peck, "The Search for Truth", WPO 48:1 (2022). Ravenholt, "Triumph then Despair", Epidemiology 5 (1994). Rogers, "The Gunshots at Grinder's Stand", WPO 31:4 (2005); "Hypocondriac Affections", WPO 36:1 (2010). Spence, History of Hickman County (1900). Turnbow, "The Man Who Abandoned Lewis", WPO 38:2 (2012); reply, WPO 41:2 (2015). Westefeld and Less, Suicide and Life-Threatening Behavior 34 (2004), abstract. Wheeler, The Trail of Lewis and Clark (1904), vol. 1. Wright, Pease and Horsley, Southern Magazine (1894). In re Exhumation of Meriwether Lewis, 999 F. Supp. 1066 (M.D. Tenn. 1998). National Park Service pages and the Cultural Landscapes Inventory for the Meriwether Lewis site. The forty forensic papers are listed with PMIDs and DOIs in notes/B_medical_forensic.md, section 3c.

Files

• README.md: the verdict, its confidence, what is new, what it agrees with, and what was not read, in summary.
• report.html: this page. report-src/DESIGN.md: the design plan.
• notes/A_documents_report.md (the record of 1809 to 1813 file by file, 24 discrepancies, 20 documents not obtained), notes/B_medical_forensic.md, notes/C_timeline_setting_versions.md, notes/D_scholarship_theories.md (the debate work by work 1809 to 2026, seven homicide theories in strongest form, the suicide case as its proponents put it, provenance disputes, the 2010 decision, a table of positions), notes/ACCESS.md (how each source was reached).
• data/evidence_items.json, data/evidence_model_results_one.json, data/evidence_model_results_product.json, data/astronomy_1809.txt, the Founders Online title indexes data/fo_index_*.tsv, data/wpo_death_index.tsv (every We Proceeded On item on the death, 1974 to 2025).
• code/evidence_model.py, code/astronomy_1809.py, code/wbfetch.py, code/fo_index.py, code/fo_source.py.