The Mystery of Marie Rogêt: 1845 Rewrites the Serial’s Clues

Which texts are compared here?

Four printings are compared. They are quoted from the transcription made by the Edgar Allan Poe Society of Baltimore, with every quotation on this page also read on a scan of the printed page. Each printing is what this site calls a one-evening text, a single fixed witness, so the serial and the book are kept apart and are never merged into one “standard” version of the story.

• Part I, The Ladies’ Companion (New York), vol. XVIII, no. 1, November 1842, pp. 15–20, 6,862 words.
• Part II, vol. XVIII, no. 2, December 1842, pp. 93–99, 6,598 words.
• Part III, vol. XVIII, no. 4, February 1843, pp. 162–167, 6,211 words.
• Tales (New York: Wiley and Putnam, 1845), pp. 151–199, 19,589 words of text plus 24 footnotes.

The three parts appeared in nos. 1, 2 and 4 of the magazine’s volume XVIII. The serial has 103 paragraphs and the 1845 text has 102, because one paragraph of Part II disappears. According to the Society’s list of texts, no manuscript of the tale survives, and a copy of Tales that carries later handwritten changes by Poe, known as the “J. L. Graham” copy, is a further version that is not used here.

Two lines on the serial’s printed pages differ from the transcription. In both places this page follows the printed pages. On page 165 of the February 1843 number the deed is “known only to one living human being,” in the singular, and on page 167 the editors write “brought to pass; that an individual assassin,” without the extra “and” that appears in the transcription.

What happens in each installment?

The story moves the Rue Morgue pair, Dupin and his unnamed friend, onto a case that the narrator openly ties to New York. Part I sets out the case: Marie, a young woman who once worked in a perfume shop, leaves her mother’s house on a Sunday morning saying she will visit her aunt in the Rue des Drômes, and on the fourth day her body is found “floating in the Seine” (p. 16).

The rest of the installment gathers what the Paris papers print about the body, the suitor St. Eustache, and the thicket near the Barrière du Roule where some boys later find her belongings.

Part II gives most of its pages to Dupin as he takes the newspapers apart, continuing his answer to one paper’s argument that the body cannot be Marie’s at all, and ends with the popular view that a gang killed her. Part III argues against the gang, turns to the evidence of Madame Deluc, who keeps a roadside inn near the thicket, and follows a man of dark complexion and a missing boat before the editors break in with their bracketed note.

The four printings, with pages, paragraphs and words, as CSV (where the data comes from) Data: The Mystery of Marie Roget, The Ladies’ Companion (1842–43) and Tales (1845) texts, E. A. Poe Society of Baltimore transcriptions, retrieved 2026-11-03

The narrator is the unnamed friend who told “The Murders in the Rue Morgue,” and he opens by saying he wrote that earlier tale “about a year ago.” The Paris police who come to Dupin here are the same force he mocked in 1841, when he named a real Paris officer; the memoirs signed by Vidocq, the officer Dupin named are another Paris crime story with a real name on the title page.

What did Poe change about the ending?

Sixteen changes of wording are followed on this page, chosen because each one alters what the tale points to, and each is quoted with its page numbers from every printing that carries it: 12 from both, and the 4 passages cut or added in 1845 from the one printing that has them. Nine of the 16 fall in the text of Part III, six in Part II and one in Part I, so the changes gather toward the end of the argument rather than spreading evenly through it.

The serial’s promise

The first change comes in Part I, before the crime is even told. The November 1842 text says the police tried many times to hire Dupin and calls this case “The only instance, nevertheless, in which such attempt proved successful, was the instance to which I have already alluded” (p. 15). In 1845 the same place reads “One of the most remarkable instances” (p. 153), so the 1845 text no longer says that this was the only case the police persuaded him to take.

The serial’s last page makes a firmer claim. After Dupin’s last speech, a bracketed note signed “Eds.” says that the part of the manuscript which follows up his clue has been left out, and then it reports the outcome of the case:

Whatever the bracket leaves out stays in what this site calls the narrator’s pocket, which sorts held-back information into three kinds: not-yet-known (the teller does not know it either), delayed (the teller knows it and saves it for its place) and concealed (the teller knows it and deliberately keeps it back). The bracket describes a deliberate omission, which makes it the concealed kind in both printings. The difference is that the editors, not the narrator, say they are keeping it back.

One, or two

The largest group of changes sits in Dupin’s summing-up in Part III. The serial counts a single person who knows the truth and then asks “And who that one?” (p. 165). The 1845 text cuts the question, along with the promise after it that the answer will not be difficult to find. The count changes as well:

The next sentence changes too. The serial reads “We have attained the idea of a murder perpetrated” (p. 165), while the 1845 text reads “We have attained the idea either of a fatal accident under the roof of Madame Deluc, or of a murder perpetrated” (p. 194). A few pages earlier the 1845 text had already added “There might have been a wrong here, or, more possibly, an accident at Madame Deluc’s” (p. 189), in a passage where the serial moves straight on.

The thicket loses its certainty in the same way. In the serial Dupin says “That it was the scene, I believe” (p. 162) and later “You will here bear in mind that I admit the thicket as the scene of the outrage” (p. 164). In 1845 the first sentence becomes “I may or I may not believe” (p. 185) and the second loses its admission. The serial’s line “We are not forced to suppose a premeditated design of murder or of violation” (p. 165) is cut entirely.

Marie’s purpose and the naval officer

Part II changes the reasons given for Marie leaving home. Where the serial has her think “I am to meet a certain person for the purpose of elopement” (p. 98), the 1845 text adds “or for certain other purposes known only to myself” (p. 183), and a later clause adds “or not until certain concealments are effected” (p. 183). Neither printing says what those purposes or concealments were, and this page does not guess.

The naval officer of her first disappearance moves a little too. In the serial the gap between Marie’s two disappearances “is precisely the general period of the cruises of our men-of-war” (p. 98), while in 1845 it “is a few months more than” that period (p. 182). The serial also has Dupin attack “the culpable remissness of the police” in a sentence of Part II that the 1845 text drops.

The 16 changes, quoted with pages from the printings that carry them, as CSV (where the data comes from) Data: The Mystery of Marie Roget, The Ladies’ Companion (1842–43) and Tales (1845) texts, E. A. Poe Society of Baltimore transcriptions, retrieved 2026-11-03

n = 16 changes: 12 reworded, 3 cut and 1 added in 1845. The CSV gives both quotations for the 12 reworded rows and the one quotation for each cut or added passage, with the edition and page of each.

Which footnotes did Poe add in 1845?

Apart from the magazine’s notes on where each installment continued, the serial printed one footnote of its own, “See murders in the Rue Morgue.” (Part I, p. 19). The 1845 text has 24 footnotes, so 23 are new. Eleven of the new notes name a New York or Philadelphia newspaper standing behind a Paris paper in the story, and seven give a place or a person, such as “Nassau Street.” on page 153.

The longest note, the one on page 151, runs to 248 words over two pages. It tells the reader that the tale followed “the essential” facts of the real murder while only paralleling the rest, and that “the investigation of the truth was the object.” Its last sentence claims that two confessions, one of them by “the Madame Deluc of the narrative,” confirmed the conclusion that the tale had reached.

All 24 footnotes of 1845, with pages and kinds, as CSV (where the data comes from) Data: The Mystery of Marie Roget, The Ladies’ Companion (1842–43) and Tales (1845) texts, E. A. Poe Society of Baltimore transcriptions, retrieved 2026-11-03

Poe’s claim has been disputed. Richard Fusco, in a 1978 study published by the Poe Society, argues that Poe invented the support he claims from the confession of the real woman behind Madame Deluc, and that her reported statement came between the second and third installments rather than long after publication. Fusco also reads the 1845 changes as Poe softening the certainty of 1842. This page reports both accounts side by side and does not test Fusco’s dating, which rests on sources it has not read.

Motif grid: what the two texts point to

The grid compares the two texts on seven features of the case as each printing tells it. Each feature has a written rule, given in the definitions file, for what counts as present and which forms count as transformed.

The grid, cell by cell with pages and evidence, as CSV (where the data comes from) Data: The Mystery of Marie Roget, The Ladies’ Companion (1842–43) and Tales (1845) texts, E. A. Poe Society of Baltimore transcriptions, retrieved 2026-11-03

n = 2 texts checked: the three-part serial in The Ladies’ Companion (1842–43) and Tales (1845). Of the 14 cells, 10 are present, 2 transformed and 2 absent in the checked text, and none is undetermined. “Transformed” here means that the 1845 text allows “one, or two” people, and that it moves the confessions from the editors’ note to Poe’s own footnote.

The seven feature definitions, as CSV (where the data comes from) Data: The Mystery of Marie Roget, The Ladies’ Companion (1842–43) and Tales (1845) texts, E. A. Poe Society of Baltimore transcriptions, retrieved 2026-11-03

What do the two texts not share?

Three features hold in both printings. Both report the newspapers’ view that a gang killed Marie and have Dupin argue against it, both follow a seaman of dark complexion who is linked to the naval officer of her first disappearance, and both print the editors’ note that leaves out the follow-up of the clue.

Four features differ. Only the 1845 text raises an accident at Madame Deluc’s inn and gives New York names in its footnotes, while only the serial counts a single person who knows the deed and reports a conviction on confession. The wording outside these 16 changes differs too, but mostly in small ways: 187 of the 306 marked differences are matters of spelling, punctuation or typeface, such as the serial’s “segar-girl” becoming the “cigar-girl” of 1845.

What the revisions do not settle

The comparison shows what each printing says, not what happened to Mary Rogers, a young woman whose death this page does not try to explain. Neither text names a killer, and the 1845 text adds possibilities without choosing among them. Whether the revisions answered news of the real case is a question about letters and newspapers that this page has not read, so the order of events in 1842–43 is left to the studies listed below.

The last word on the outcome belongs to the editors’ bracket in each printing. In February 1843 the note signed “Eds.” told readers of The Ladies’ Companion that an individual assassin had confessed; in the 1845 Tales the same bracket says only that “the result desired was brought to pass,” and Poe’s own footnote, at the head of the tale, carries the claim about confessions instead.