Mary Celeste: Why Did Ten People Leave a Ship That Could Still Sail?

The strangest fact about Mary Celeste is not that a ship disappeared. The ship came back. Men from another vessel put her under control, sailed her to Gibraltar, and delivered her into a system of inspections, correspondence, and court proceedings. Her cargo eventually reached its destination. The people who had begun the voyage did not.

Captain Benjamin Spooner Briggs, his wife Sarah, their little daughter Sophia, and seven crew members had left New York for Genoa in November 1872. In early December, the brigantine Dei Gratia encountered their vessel in the Atlantic. Nobody answered. A boarding party found an empty ship, damaged sails and rigging, water below, and equipment that invited questions. The ship’s boat was missing. So were the people whose decisions might have explained the scene.

A later story supplied an answer in the form of a survivor’s statement. It had a doctor, a conspiracy, murders, and a voyage diverted toward Africa. Arthur Conan Doyle had invented it. Other accounts supplied hot meals, perfect order, sinister crewmen, and increasingly confident declarations that the mystery had been solved. Some of those details tell us more about the making of a legend than about the crossing.

This investigation starts closer to the event: a family letter, contemporary Gibraltar newspaper reports, a diver’s affidavit, an analyst’s report, and a consular account of the delivered cargo. Most are available here through published transcriptions, not a fresh examination of the original manuscripts. That distinction matters. We can examine what the surviving words say without pretending that every original page, physical object, or archival file has been inspected.

Where the evidence stands: Mary Celeste was found deserted, and a salvage crew sailed her to Gibraltar. No authenticated account establishes why Briggs and the others left or what happened to them afterward. The material examined during the inquiry did not substantiate the suspected bloodstains. Emergency abandonment remains a serious explanation; the initiating emergency and the subsequent loss of the people remain unproved. Later experiments address physical possibilities, not a witnessed historical sequence. Royal Museums Greenwich’s overview provides a useful introduction to the unresolved case.

Explore the investigation

1. The first report already contained a problem
2. A working vessel with a history before 1872
3. Ten people, not an anonymous vanished crew
4. The captain's letter opens a different world
5. What the family expected to happen next
6. The cargo: 1,701 barrels and nine empty ones
7. Reconstructing the calendar without inventing a last hour
8. Boarding a vessel that did not answer
9. The position on the slate and the route we cannot draw
10. A diver examined the underside
11. The stains that did not become blood evidence
12. The marks on the bow and two incompatible readings
13. What the salvage court was deciding
14. Weather records from land, decades later
15. Water, the pump, and the fear of a worsening problem
16. Alcohol vapor: a real hazard and an unproved event
17. What the newer experiments actually change
18. The missing boat is the second half of the mystery
19. Mutiny and the difference between a suspicion and a case
20. Pirates, collusion, and the money that came later
21. Conan Doyle supplied the survivor history lacked
22. The ship lived on, and even her wreck became disputed
23. Put the explanations beside the evidence
24. The answer the record permits
25. Sources and further investigation

1. The first report already contained a problem

On December 13, 1872, the Gibraltar Chronicle reported the arrival of an American vessel without her original company. Dei Gratia had arrived the previous day; Mary Celeste followed with men from the other ship aboard. The report was close to the event, but closeness did not make every printed detail correct. Its position for the discovery differed markedly from the position printed in the more developed January account.

The early item gave a latitude of 28 degrees 20 minutes north and a longitude of 17 degrees 30 minutes west. The January 31 report gave 38 degrees 20 minutes north and 17 degrees 15 minutes west. Ten degrees of latitude is an enormous discrepancy. It cannot be dismissed as the slight wandering of a sailing ship between two observations. Nor should a researcher average the two numbers. The sensible next step is to compare the reports with the navigational testimony and identify which event each purports to describe. Transcriptions of the Gibraltar Chronicle reports.

That small exercise establishes the character of the whole investigation. A date or coordinate can be old, precise, and wrong. An impressive number may be a printing error. A modern account may quietly correct it, or may copy it into a new map. The appearance of precision does not establish the quality of the underlying observation.

There is a second distinction in the first report. The ship had not materialized at Gibraltar. People had encountered her offshore, boarded her, and navigated her there. The interval between discovery and examination created a chain of custody, but it also created opportunities for ordinary changes: pumping, sail handling, moving equipment, and the practical work of getting a short-handed vessel into port. An inspector at Gibraltar saw the ship after that passage.

None of this makes the salvors unreliable by definition. It makes their accounts essential. A responsible reconstruction needs to ask who first saw an object, whether it was moved, and whether later descriptions repeat the initial observation or describe a subsequent condition. The vessel was a workplace and a salvage operation before it became a famous puzzle.

The first headline, then, should not be a declaration of a supernatural event or a solved crime. It should be a set of questions about a recoverable ship and an unrecovered company. The strongest evidence begins with those questions still open.

2. A working vessel with a history before 1872

Mary Celeste began as Amazon, built at Spencer’s Island in Nova Scotia in 1861. Her history included a damaging grounding in 1867, changes of ownership, American registration under the new name in 1868, and a substantial refit before the famous voyage. A compilation of her early career describes Joshua Dewis’s shipyard and the sequence of salvage, purchase, and rebuilding. These events make a more useful starting point than the idea of a vessel born under a curse. The Spencer’s Island history.

A ship can survive an accident, receive repairs, and continue trading. That is a material and commercial history, not evidence that later misfortunes belong to one supernatural chain. To decide whether an old grounding contributed to a later emergency, an investigator would need to connect damage, repairs, and a particular failure. Merely placing the two dates in the same paragraph does not supply that connection.

The rig also matters. Mary Celeste was a brigantine, with square sails on the forward mast and fore-and-aft canvas on the after mast. The surviving painting of Amazon offers a useful general silhouette, although the ship underwent later alterations. It should not be treated as a complete plan of her 1872 condition. The watercolor opening this article follows the broad two-masted form while remaining an interpretation.

Published dimensions and tonnage figures require the same care. Figures from before and after a refit describe different states of a vessel. Gross register tonnage is a measure related to enclosed volume, not simply the ship’s weight. Calling it weight can distort a discussion of capacity or stability before the argument has properly begun. This article does not calculate Mary Celeste’s stability from a few general dimensions; the necessary plans and loading details have not been established here.

The refit is relevant to later pump and debris theories, but relevance is not proof. A recently altered ship might contain construction debris. It might also have been thoroughly cleaned. A theory needs evidence of the particular obstruction it proposes, not just a plausible opportunity for an obstruction to exist.

By November 1872, Briggs was commanding a commercial voyage in a vessel in which he had an interest. He had reasons to care about her performance, the delivery of the cargo, and the safety of the people aboard. Understanding those ordinary responsibilities gives the later abandonment its force. A decision to leave would have meant giving up more than a convenient platform on the sea.

3. Ten people, not an anonymous vanished crew

The familiar phrase missing crew can obscure the composition of the company. There were eight working men when the captain is included, plus Sarah Briggs and Sophia. The total was ten. A roster is therefore more than a list of possible suspects; it restores the distinction between the people operating the ship and the family traveling with them.

The names generally given are Benjamin Spooner Briggs, first mate Albert Richardson, second mate Andrew Gilling, steward and cook Edward William Head, and seamen Volkert Lorenzen, Boy or Boz Lorenzen, Arian Martens, and Gottlieb Goodschaal. Published records and transcriptions contain spelling variants, especially for the European sailors. Sarah and Sophia complete the company. This article uses recognizable forms without pretending that every variant denotes a different person. Published crew list; Fay’s discussion of names in the records.

A variant spelling is a practical research problem. It can cause a person to disappear from a database search or make two records seem unrelated. Before inferring that a sailor used an alias, a researcher should compare the handwriting, the clerk’s language, the port, and the other identifying information. A transcribed name is not always a name the person wrote himself.

The men’s jobs also identify different kinds of responsibility. The first and second mates belonged to the command structure. The steward and cook sustained the domestic life of the voyage. The seamen handled the repeated physical work that kept a sailing vessel moving. These roles do not tell us how a dispute unfolded or who supported a particular decision. They tell us why the ship depended on cooperation across distinct tasks.

Briggs’s son Arthur was not aboard. The later family account places him with his grandmother so he could attend school. His absence from the voyage explains why the family’s story continued through a child who survived on land. It should not be converted into an ominous last-minute rescue from a predicted disaster.

When the inquiry considered violence, these people became subjects of speculation without being available to answer it. That imbalance persists. We can easily invent an angry mate or a frightened sailor because neither can contradict the scene. The more responsible task is to retain names, roles, and documented relationships while refusing to supply private thoughts as evidence.

For a comparison with a different missing company, see the Carroll A. Deering investigation. Both ships left physical evidence behind. Neither left a verified final account from the people whose absence created the mystery.

4. The captain’s letter opens a different world

On Sunday, November 3, Briggs wrote to his mother from New York. The letter survives in a family narrative attributed to Oliver Cobb and is reproduced online. It is one of the most valuable pieces of the case because it was written before the disappearance, for an ordinary personal purpose. It was not composed to explain a ghost ship. Briggs’s letter and the surrounding family account.

The letter describes loading completed on Saturday and an expectation of leaving on Monday night or Tuesday if circumstances allowed. It comments favorably on the mate and steward. Briggs thought the vessel was in good trim, while acknowledging that this was his first experience of sailing in her. He could assess preparation without yet knowing exactly how the ship would behave underway.

Those two statements are compatible. They do not reveal a captain secretly terrified of an unseaworthy vessel, nor do they certify that every component would perform perfectly. They show a working mariner distinguishing what he had inspected from what he had yet to experience. It is precisely the sort of distinction later retellings tend to erase.

The domestic details are even more striking. Sarah and Sophia had joined him. A trip involving Central Park became more difficult because the horse epidemic had disrupted streetcar service. A private carriage was too expensive for the outing Briggs had in mind. Sophia had recovered from a cold, enjoyed food, looked at a family album, and asked about people at home. The cabin’s musical life and the plans for letters from abroad belonged to a household temporarily at sea.

These details do not solve the disappearance. Their value is that they give the missing family a present tense before disaster: a child’s appetite, an outing, a budget, a message to relatives. They also resist a common narrative habit, the search for a fatal premonition in every ordinary sentence. A hope to return in spring is a plan, not a coded farewell.

The document has its own limits. The online text is a transcription embedded in a later family account. The letter’s wording and the surrounding recollections should not be treated as if they were all written on November 3. A detail supplied by Cobb may be useful, but its date and basis differ from Briggs’s own report.

Read on its proper terms, the letter makes the later absence more troubling without making it more occult. This was a voyage being prepared by people who expected to continue their lives.

5. What the family expected to happen next

Briggs’s plans for correspondence reveal how he imagined the voyage unfolding. Mail might reach him at Genoa after a period of travel, and later correspondence might find him at Messina. He hoped to be home in spring. The route in his mind extended beyond the single destination usually printed in a summary of the mystery.

The letter also looks backward toward Arthur’s life at home. Briggs discusses practical matters involving the boy’s clothing and skates, and asks for attention to ordinary family arrangements. The details matter because they show continued obligations on shore. They do not read like preparations to erase an identity or abandon the family left behind. That is not a logical proof against a secret plan, but it is actual evidence that any such theory must address.

A fraud theory sometimes assumes that a person could simply step out of one life and into another. The letter makes that abstraction concrete. Briggs had a mother to write to, a son whose needs he discussed, a wife and child aboard, business responsibilities, and destinations where correspondence was expected. A deliberate disappearance would need to account for those relationships as well as for the movement of the ship.

The same reasoning must not be pushed too far. Affectionate correspondence cannot demonstrate that an accident was impossible, that a disagreement never occurred, or that a captain never made a mistaken judgment. People who love their families can still encounter emergencies. A personal letter is evidence about expressed intentions at a particular time, not a guarantee about later conduct.

The later family narrative says Briggs had considered a different livelihood before committing to the voyage and describes improvements intended to make the cabin suitable for Sarah and Sophia. Those recollections help explain the decision to travel together. They do not establish what happened during the final hours. Family memory and voyage evidence overlap, but they are not interchangeable.

There is a poignant parallel with Theodosia Burr’s last voyage aboard the Patriot. In both cases, correspondence records expectations that the sea did not fulfill. The historian can recover the plans and the waiting without inventing the missing person’s final thoughts.

For the reader, the practical question is this: does a proposed explanation fit the documented life, or does it require replacing that life with an invented personality? A persuasive account must do the harder work of explaining what the records actually show.

6. The cargo: 1,701 barrels and nine empty ones

Mary Celeste carried alcohol in barrels. The most useful numerical evidence is a consular report about what was landed at Genoa after the salvage. As quoted from Charles Fay’s book, Consul O. M. Spencer wrote to Consul H. J. Sprague on April 26, 1873, that 1,701 barrels had been delivered, agreeing with the bills of lading, and that nine were empty. The report described that number as not unusually large and the cargo’s condition as excellent. The reproduced consular passage.

That is a more precise and more limited statement than cargo untouched. The barrels were accounted for, but a small number lacked their contents. It is also different from the claim that investigators discovered nine leaking barrels at the moment the vessel was first boarded. The report concerns the outturn at destination, after additional time and handling.

Nine divided by 1,701 is approximately 0.53 percent. That is a fraction of barrel count. It is not a measured fraction of the original liquid volume, a measure of vapor in the hold, or a reconstruction of how much alcohol had escaped before the people left. A barrel can empty over time. The timing and location of any loss are central to the vapor theory and are not supplied by the final count alone.

The quoted report also includes a proof figure and a specific gravity. Historical proof conventions require care; a number labeled proof should not automatically be treated as a modern percentage of alcohol by volume. This article does not infer an exact original formulation from that line, and it does not label the cargo methanol or assume a denaturing recipe without supporting shipping documentation.

The consular judgment that the loss was not unusual is useful context, but it is not a physical experiment. An ordinary amount of leakage in trade could still matter under a particular combination of ventilation and ignition conditions. Conversely, the existence of flammable cargo does not show that those conditions occurred aboard this vessel.

The most valuable next records would be the full bills of lading, stowage information, cask descriptions, and dated inspections of their condition. A hypothesis about a concentrated release needs more than the total cargo count. It needs a chronology of the release. That is the point where an attractive explanation encounters the actual limits of the evidence.

7. Reconstructing the calendar without inventing a last hour

The main chronology begins with Briggs’s November 3 letter and the departure from New York generally dated November 7. It then narrows to the surviving navigational information near the Azores. The January 31 Chronicle report distinguishes a formal log entry dated November 24, sea time, from a slate entry for 8 a.m. on November 25. The slate placed the eastern point of St. Mary’s, or Santa Maria, six miles to the south-southwest. Contemporary report of the entries.

Those are two records, not one sentence with an uncertain date. A slate could hold working observations before transfer into a log. The fact that the final surviving entries differ in form and date is not itself evidence that somebody tampered with them. It is a reason to preserve their wording and function.

Discovery is commonly dated December 4, although many accounts say December 5. Fay’s discussion attributes the discrepancy to civil and nautical dating in the relevant records. The article uses December 4 for the civil date and retains the alternative where it explains a source. It does not assume that every maritime record in 1872 used one universal convention. Fay’s date discussion.

There is a temptation to subtract the last entry from the discovery and call the result the period during which the ship sailed herself. That assumes the people left immediately after the entry. It may be a reasonable scenario to test, but it is not established by the absence of a later entry. They could have remained aboard for an interval that produced no surviving record.

Nor does a stopped record prove that its writer stopped living at that instant. A log can stop because of interruption, loss of pages, a change in routine, or the decision to leave it behind. The task is to establish which possibility the evidence supports, not to choose the most dramatic one.

The later dates are firmer anchors for particular actions: the two ships’ arrivals at Gibraltar, the diver’s December 23 inspection, the stain examinations in early 1873, and the March judgment reported by the Chronicle. These events tell us when the evidence was examined. They cannot be moved backward and presented as observations made beside Santa Maria.

A useful timeline therefore contains both events and records. The empty space between them is part of the result, not a defect to be concealed by a confident narration.

8. Boarding a vessel that did not answer

Royal Museums Greenwich identifies Oliver Deveau and John Wright as the initial boarders. They found nobody aboard, a missing boat, disturbed or damaged sailing equipment, water ingress, a partly dismantled pump, and a sounding rod. The museum’s account also distinguishes the subsequent salvage party: Deveau and two sailors took Mary Celeste to Gibraltar. The museum’s reconstruction.

The first boarding and the salvage passage are often compressed into one scene. Keeping them apart clarifies who could have observed the ship before work began and who later handled it over a longer period. It also prevents an illustration of three salvors from being mistaken for a literal record of the initial boarding party.

What did the empty deck establish? It established that the boarders found no one. It did not establish that everyone had left at the same time, in the same way, or for the same reason. The missing boat and portable navigational equipment make an organized departure a serious inference, but the inference still needs a sequence.

The ship’s condition was neither pristine nor equivalent to a sinking wreck. Both exaggerated descriptions create false choices. If it was perfect, leaving seems inexplicable. If it was obviously sinking, the later successful passage seems impossible. A vessel can have damaged sails, a troublesome pump, and water below while remaining capable of being worked to port.

The boarders’ practical response is therefore important evidence. They did not merely look from a distance and pronounce the ship sound. A small party actually navigated it. That demonstrates a capacity under the conditions of the salvage passage. It does not show what Briggs believed during an earlier emergency, or what condition he thought might develop next.

A reader assessing a theory should ask which condition it explains. Does it explain the first boarders’ observations? Does it explain the state after pumping and sail work? Does it explain the inspector’s later view in harbor? An account that moves freely among those stages can seem to solve a contradiction that it created itself.

The mystery survives this careful reading. It simply becomes a more specific one: how did a vessel that could later be managed by three men lose the ten people who had set out aboard her?

9. The position on the slate and the route we cannot draw

The January Chronicle account gives the last formal log position as 36 degrees 56 minutes north, 27 degrees 20 minutes west. It then describes the later slate observation of Santa Maria. The distinction between a calculated coordinate and a bearing on land is important. They are related navigational evidence, but they are not identical kinds of measurement.

A bearing reports the direction of an identified feature from the observer. Its value depends on the correct identification of the feature, the instrument or method used, and the meaning of the recorded distance. Turning that brief entry into an exact modern map pin requires assumptions. Fay later supplied a proposed coordinate for the slate position, explicitly treating it as a reconstruction rather than a number written in the testimony. The published record and later reconstruction.

The ship’s eventual discovery position lies farther east. Connecting the two points with a straight line would make an attractive route map, but it would not reconstruct the track. Winds, currents, sail configuration, and the time of abandonment could all alter the path. A line between endpoints answers where the endpoints are; it does not answer how the ship traveled between them.

The January report itself tried to reason about drift. It questioned whether the vessel could have covered the implied interval under the conditions it assumed. That is valuable evidence of contemporary reasoning, not a final oceanographic result. An argument can be historically important and still depend on incomplete inputs.

The missing navigation equipment creates another difficulty. Instruments useful to people leaving in a boat may also be the instruments a modern researcher would want to inspect for error. If a chronometer was inaccurate, a later calculation might show that certain recorded positions were inconsistent with other evidence. It would not allow us to examine the missing instrument or prove when an error began.

This is why the article’s timeline avoids a decorative line purporting to be the final voyage. The honest visual is a sequence of recorded observations and later actions. The route between them remains a problem to be modeled under stated assumptions.

Readers of the Beale Ciphers investigation will recognize the same distinction in another form. A set of exact-looking numbers is only as useful as its transcription, conventions, and reference system. Before a coordinate can become a clue, we need to know what its writer meant it to represent.

10. A diver examined the underside

Ricardo Portunato’s affidavit records an inspection of Mary Celeste below the waterline at Gibraltar on December 23. The vessel was then held under a warrant in the Vice-Admiralty proceedings. The language of arrest in this setting concerns the ship’s legal detention; it should not be read as a statement that every person associated with her had been arrested for a crime. Portunato’s affidavit.

The diver described examining the underwater hull, including the keel, sternpost, rudder, and protective metal sheathing. He reported good condition and no signs of a collision or grounding of the kind he was looking for. That is direct inspection evidence about particular surfaces at a particular time.

It is especially relevant to theories that require a major underwater impact. If an explanation says the ship struck an obstacle hard enough to produce a conspicuous structural injury, the diver’s report is a difficulty that must be addressed. The response cannot simply be that anything might have happened. The theory should specify what injury it proposes and why the examination would not have detected it.

At the same time, the affidavit is not a universal certificate that no frightening event could have occurred. It does not measure the crew’s earlier perception of flooding. It does not establish the condition of every pump. It does not identify a sound heard in the hold or a vapor concentration. An examination designed to assess hull condition cannot answer every question about a disappearance.

The date matters as much as the observation. December 23 was after the discovery and the passage to Gibraltar. The report’s strongest use is to establish what Portunato found then. To use it as proof of a condition a month earlier requires an additional argument about what could or could not have changed in between.

This is not an excuse to dismiss inconvenient evidence. It is a way to use evidence at its proper strength. A sound underwater hull weighs against certain damage scenarios. It does not weigh equally against every emergency-abandonment scenario. If the danger was believed to be imminent rather than already catastrophic, the ship’s later survival is part of what the theory is trying to explain.

The affidavit also reminds us that the inquiry involved work beyond speculation in a courtroom. Someone entered the water and inspected the vessel. Recovering that labor makes the historical investigation more substantial than the familiar image of baffled officials confronting an inexplicable ship.

11. The stains that did not become blood evidence

One of the most consequential developments was the examination of material suspected to be blood. Dr. J. Patron’s report describes samples from the vessel and a sword associated with the captain’s cabin. The transcription contains a date inconsistency between its heading and the sequence described in the text, so it is safest to refer to the examinations as occurring in the January–February 1873 stage of the inquiry. The laboratory sequence is clearer than the heading. Patron’s analytical report.

Patron received several paper-wrapped samples. Some contained wood bearing suspicious marks; others involved a bag or related material. He examined what happened when fragments were placed in distilled water, observed the liquid over time, and used heat and microscopy. The report gives intervals of observation rather than a single instant judgment.

That detail matters. The analyst did not simply look at a red mark and accept the investigators’ interpretation. He compared the appearance with what his tests revealed. The water remained clear in the described trials, and the expected signs of blood did not appear. Microscopic examination identified material consistent with wood and iron compounds rather than the blood structures he sought.

The sword received separate attention. Its appearance initially seemed suspicious, but the subsequent examination did not substantiate that impression. The report describes chemical behavior consistent with iron and notes the absence of the expected blood findings. The sheath was also considered. The result undermined the use of the sword and stains as affirmative proof of murder.

A modern reader should resist two opposite exaggerations. The first is to continue calling the stains blood after the analysis failed to establish that. The second is to claim that the tests proved no violence had occurred anywhere aboard the ship. Patron did not conduct a modern DNA examination of every surface. His report addresses the submitted material using the methods he described.

The important historical result is narrower and still powerful: a central apparent clue to bloodshed did not survive examination as blood evidence. A theory of violence could no longer rely on that material as if its nature had been confirmed. It would need other evidence.

This is a rare point in the mystery where uncertainty can be reduced without solving the whole case. We can say that a particular argument lost support. We do not have to preserve every early suspicion merely because the final disappearance remains unexplained.

It is also a lesson in the difference between suspicion and observation. The same reddish appearance can be recorded accurately while its cause is interpreted incorrectly. The analyst’s task was to test the interpretation. In this instance, the investigation produced a correction that later storytelling has too often ignored.

12. The marks on the bow and two incompatible readings

The January Chronicle report described marks on the bow as potentially deliberate damage. In that reading, the physical scene might fit a crime followed by an attempt to make the ship appear damaged or to sink her. The report’s reasoning linked observations of the vessel with a larger suspicion about the missing people.

A later inspection described in the March 4 Chronicle offered a different interpretation. Captain R. W. Shufeldt of the United States naval vessel Plymouth regarded the marks as effects of ordinary strain and the behavior of the planking rather than intentional cutting. He also considered a scenario involving leakage, fear, and departure from the ship. The January and March reports.

These are not two independent discoveries of two different sets of marks. They are competing readings of material associated with the same vessel. The difference is central. If the marks were intentionally cut, they might support an inference of deliberate action. If they arose from working timber or ordinary strain, that inference weakens.

The reader should ask what each observer actually examined and what their explanation required. Did the proposed cutting tool leave a characteristic trace? Did the arrangement of the planks support a structural explanation? Were the marks photographed, measured, or preserved? A modern article cannot answer those questions by drawing an imagined slash in a watercolor.

Nor should Shufeldt’s professional standing be used to turn his entire scenario into a witnessed fact. His inspection may provide a better explanation of particular marks while his account of the crew’s decision remains hypothetical. Expertise can strengthen an interpretation of material without supplying direct knowledge of an unwitnessed event.

This separation is particularly useful in assessing emergency theories. A leak or the fear of one does not require deliberate damage. Conversely, the absence of a proven deliberate cut does not establish which ordinary defect, if any, frightened the company. The physical question and the behavioral question remain connected but distinct.

The contemporary disagreement also changes the way we should narrate the inquiry. It was not a single official conclusion that everyone accepted and later researchers overturned. Different observers offered different assessments while the case was still active. A professional investigation should preserve those disagreements at the point where they occurred.

The bow marks therefore belong in the evidence table with their interpretive history attached. They should not appear as unqualified proof of sabotage, and they should not disappear merely because a preferred modern theory has no use for them.

13. What the salvage court was deciding

The court proceedings arose because Dei Gratia’s people had recovered and brought in a vessel and cargo belonging to others. Their work raised a claim for salvage. The disappearance of the original company made the circumstances unusual and invited a broader inquiry, but the existence of a salvage claim is not itself evidence that the salvors caused the emergency.

The March 15 Chronicle reported the judgment delivered the previous day. It gave an award of £1,700 and stated separate values in dollars for the vessel and cargo: $5,700 and $36,943, totaling $42,643. The report described the award in relation to the property’s value. These figures should remain in their stated currencies; a pound amount cannot be divided directly by a dollar total to produce a meaningful percentage without a conversion basis. The reported judgment.

Modern summaries sometimes give different totals or fractions. That does not automatically mean the court’s arithmetic was incoherent. They may be discussing insured value rather than the values used in the report, converting currencies, rounding, or repeating an earlier summary. This article preserves the contemporary figures and does not manufacture a modern equivalent.

The distinction between compensation and a criminal verdict matters. An award can reflect the court’s assessment of the service and the circumstances of the claim. It does not establish the fate of the missing people. Equally, an unresolved suspicion voiced during proceedings is not a conviction of the men who brought the ship in.

Frederick Solly-Flood’s concern about foul play shaped the inquiry. But the analytical report on the stains and the alternative reading of the bow damage show why an investigator’s conviction must be separated from the evidence supporting it. A determined official can ask useful questions and still favor an explanation that the material does not prove.

The surviving newspaper transcription is not a substitute for the full court file. One awkward passage near the end of the judgment report appears textually problematic. Rather than build an accusation about a witness’s conduct on that wording, this article leaves that point for comparison with the original printed page or court record. Transcription errors should not become moral judgments.

The court’s historical importance lies partly in the documents it generated. Without the salvage proceedings, many observations might never have been recorded in a form available to later readers. The inquiry failed to settle the disappearance, but it preserved evidence that lets us reject some of the most persistent embellishments.

14. Weather records from land, decades later

A letter dated May 27, 1940, by meteorologist José Agostinho revisited weather around the Azores in November 1872. It drew on older observations and correspondence rather than a recovered shipboard weather instrument. The letter identifies Angra and Ponta Delgada as the available observing stations for the period and reports a change in conditions around November 24–25. Agostinho’s letter.

Its details include a front passing Angra during the afternoon and evening of November 25, a change from southwesterly to northwesterly wind, and a wind observation of 62 kilometers per hour at Ponta Delgada the previous evening. It also gives rainfall totals of 14 millimeters at Angra and 29 at Ponta Delgada for the noon-to-noon interval spanning November 24–25. The letter says no relevant earthquake was found in the records and newspapers consulted.

Those figures are regional evidence, not measurements made on Mary Celeste. A rain gauge on land cannot tell us how much water entered the ship. A station wind observation cannot be inserted into a final-hour narrative without considering the vessel’s position and the movement of the weather system. The chart displays the reported station rainfall precisely because its limited meaning can be stated clearly.

The absence of a recorded earthquake is also a bounded finding. It weakens an explanation that relies on a conspicuous local event that those records should have captured. It is not proof that every possible underwater disturbance anywhere in the Atlantic was detected. A proposed earthquake explanation would still need independent geological or observational support, not merely the claim that old records were incomplete.

The 1940 letter belongs to a later stage of investigation. Its author was reconstructing conditions from surviving information, just as modern researchers do. The age of the observations and the date of the reconstruction should both remain visible.

Weather is not a single rival to human decision. A vessel might encounter rough conditions, develop a problem, and then be abandoned because of how the problem was understood. The relevant question is not simply whether a storm existed. It is what sequence connected the weather to the condition of the ship and the choices of the people aboard.

A convincing reconstruction would therefore compare station records, nearby ship observations, the log entries, and the physical evidence under one consistent chronology. It would also show the range of outcomes that remain possible when the timing or position changes. A dramatic storm illustration cannot perform that work.

15. Water, the pump, and the fear of a worsening problem

The partly dismantled pump and the sounding rod are among the most suggestive reported objects. They point toward attention to water in the vessel, but they do not tell us why a component was taken apart or who last handled it. It might have been inspected, repaired, or left in the middle of work. The object is evidence of a condition; the action that produced the condition has to be inferred.

A sounding rod can help establish a depth at a particular place. That is not the same as knowing the total volume of water in an irregular, loaded hull, or the rate at which water is entering. A single depth also does not tell a captain whether the next observation will be higher. The practical concern in a developing emergency is often the trend and the ability to control it.

This distinction offers a plausible route through the central paradox. The ship’s later survival does not prove that Briggs knew she would survive. If he believed water was entering faster than it could be managed, and if he thought a boat offered access to land or temporary safety, departure might have seemed the less dangerous option at the time. That is a conditional reconstruction, not a recovered order from the captain.

Anne MacGregor’s investigation, described by Smithsonian in 2007, combined a drift analysis with a proposed chain involving navigational error, rough weather, and a pump obstructed by debris from earlier cargo or refitting. Oceanographer Phil Richardson helped assess whether the ship could travel between the relevant positions without a crew. The published account presented a plausible scenario, not an eyewitness discovery. Smithsonian’s report on the reconstruction.

The useful test is to separate its links. Was a pump obstructed? Was the obstruction caused by the suggested material? Did it prevent a reliable assessment of the water? Did Briggs believe the ship was in immediate danger? Did the company leave near land? Each question needs its own support. Proving that an unmanned drift was possible does not prove the pump’s condition or the captain’s reasoning.

There is also no need to portray the captain as foolish for a decision whose inputs are lost. Later observers had information he may not have possessed: the fact that the ship remained afloat and could be sailed to port. Judging an earlier decision solely by that later outcome substitutes hindsight for reconstruction.

The pump theory remains compelling because it connects an observed object with an ordinary maritime concern. Its weakness is the missing evidence that would establish the complete sequence. A careful reader can recognize both qualities at once.

16. Alcohol vapor: a real hazard and an unproved event

Alcohol cargo creates another plausible reason for alarm. Ethanol is flammable, and its vapor can form an ignitable mixture with air under suitable conditions. The National Institute for Occupational Safety and Health classifies ethyl alcohol as a flammable liquid. That basic physical fact justifies examining vapor scenarios. It does not establish the composition of the historical cargo or the atmosphere in Mary Celeste’s hold. NIOSH’s ethyl alcohol entry.

Three versions of the theory should be kept separate. In one, an actual brief ignition frightened the company. In another, an unusual sound or pressure event was interpreted as a warning of an impending explosion. In a third, concern about fumes led to a precautionary departure without any ignition. They do not require identical evidence, and success in demonstrating one mechanism does not validate the other two.

The nine empty barrels are relevant but insufficient. The outturn report does not tell us when their contents were lost, how quickly they escaped, or how the hold was ventilated at the time. Nor does a final empty cask establish a concentration of vapor around the crew. A credible model needs those conditions stated as assumptions or supported by records.

The word explosion itself can mislead. Readers may imagine a sustained fire that consumes a ship, while a proposed experiment may concern a brief combustion event with a different pattern of damage. The absence of a burned-out hull can challenge one version without automatically excluding every possible short event. But the reverse is equally important: showing that limited visible damage is possible does not show that such an event occurred here.

A historical explanation also needs a human sequence. What did the people notice? Why did they all leave? Was the departure intended to be temporary? What prevented return? A chemical demonstration addresses only part of that chain. It cannot supply the missing order, the boat’s movements, or the fate of the passengers and crew.

There is no need to invent a specific spark, smoking sailor, or careless act. Those details may make a scene cinematic, but they would turn a proposed mechanism into an accusation without evidence. The investigation can assess physical plausibility while leaving the initiating source unknown.

The vapor theory is therefore neither nonsense nor a demonstrated solution. It is a family of hypotheses whose strongest and weakest points become clearer when the cargo evidence, the ship’s condition, and the missing boat are considered separately before they are joined.

17. What the newer experiments actually change

In April 2026, Chemistry World reporter Mason Wakley described renewed work on the explosion hypothesis involving Jack Rowbotham and Frank Mair at the University of Manchester, with comment from Andrea Sella about earlier experimental work. The reporting accompanied a television reconstruction. Wakley’s public account confirms the participants and the renewed testing, but it is not a complete experimental paper. The reporter’s account of the 2026 investigation.

The development is worth including because it tests a physical possibility that can otherwise be dismissed too casually. It does not justify changing the historical status to solved. A successful demonstration can show that a specified arrangement produces a specified result. To apply that result to Mary Celeste, the arrangement must be compared with the conditions that plausibly existed aboard her.

The most useful questions concern the experiment’s inputs. What material represented the cargo? What temperature and ventilation were assumed? How was the model scaled? What opening or hatch arrangement was used? What was measured afterward, and how often was the result repeated? These are questions for evaluating the reported demonstration, not instructions for reproducing a hazardous experiment.

The distinction between a model and a historical reconstruction is especially important. A small model can reveal a mechanism, but changing scale can affect how heat, pressure, and gas movement interact. A researcher must explain which features of the full-sized vessel the model represents and which it simplifies. The audience should not have to infer that equivalence from a dramatic camera angle.

A second set of questions concerns the historical record. Would the proposed event leave traces that the boarders or inspectors described? Would it disturb the same hatches? Would it explain the absence of the people and the presence of the cargo? Does it require an additional unrecorded event to prevent their return? A theory becomes stronger when it accounts for several independent observations without continually adding new assumptions.

The article therefore treats the 2026 work as renewed support for investigating a mechanism, not as a recovered account of the final minutes. The distinction resembles the recent Kryptos developments: obtaining one part of an answer does not automatically settle a different question. Physical feasibility, historical occurrence, and the fate of the people are separate claims.

The experiment’s greatest value may be to improve the questions. It can tell investigators which conditions matter most and which historical records would be most useful. That is a meaningful advance even when the final verdict remains open.

18. The missing boat is the second half of the mystery

An explanation for leaving the ship is incomplete unless it also explains why nobody returned or reached a place where their survival was recorded. The missing boat is therefore not a minor item on an inventory. It is the likely bridge between the abandonment and the disappearance of the people.

Several scenarios can be stated without pretending that one has been witnessed. The company may have intended to remain near the vessel while a perceived danger passed. They may have intended to head for land. They may have expected help from another ship. Each possibility creates a different set of questions about distance, weather, equipment, and subsequent movement.

The often-repeated tether theory proposes that a line connected the boat to the brigantine and later failed or came loose. It offers an explanation for losing access to a vessel that remained afloat. But no authenticated final account establishes that such a line was used. It belongs in the hypothesis column, not in the opening scene as a fact.

The basic arithmetic of separation helps explain why the idea is plausible without proving it. If two objects move apart at a relative speed of one knot, they separate by half a nautical mile in thirty minutes. At two knots, the same interval gives one nautical mile. These are deliberately chosen examples, not estimates of Mary Celeste’s speed, the boat’s performance, or the conditions that day.

A small initial separation could become a serious problem if the larger vessel continued moving and the people in the boat could not match its course or speed. That reasoning does not require the ship to race away at an implausible rate. It does require evidence about the actual arrangement before it can become a historical reconstruction.

The boat’s absence also does not prove that all ten people were alive when it left. That is one of the difficult limits of the evidence. A peaceful emergency departure, a departure after violence, and a later removal of the boat could all produce absence. The surrounding observations help compare those sequences, but the missing object cannot testify by itself.

The most decisive discovery would be a securely identified contemporary record of the boat or its occupants after separation. An unproven sighting, an anonymous confession, or a story published long afterward would need authentication before it could fill that role. Until then, the boat remains both the strongest practical clue to departure and the largest missing piece of the human outcome.

19. Mutiny and the difference between a suspicion and a case

Mutiny attracted attention because it could explain a missing captain and an abandoned vessel through human conflict. The difficulty is that the theory must explain more than one absence. Briggs, Sarah, Sophia, and every crew member disappeared. A successful mutiny followed by escape still needs an account of where the mutineers went and why no reliable trace of them emerged.

The inquiry’s suspected blood evidence initially appeared capable of supporting violence. Patron’s analysis undermined that use of the material. The alternative explanation of the bow marks further weakened an argument that depended on deliberate damage. A modern mutiny theory cannot simply reuse those early impressions without acknowledging the subsequent examinations.

Briggs’s favorable remarks about the mate and steward are also relevant. They do not prove that relations remained good throughout the crossing, but they are contemporary evidence against inventing a predeparture atmosphere of obvious hostility. If a theory proposes a later rupture, it needs evidence of that change rather than a fictional quarrel inserted into the voyage.

Nationality is not a motive. The presence of European sailors does not explain violence, and spelling differences in their names do not establish concealment. A responsible investigation would look for documented disputes, prior relationships, changes in employment, or reliable witness statements. It would not treat an unfamiliar surname as a sinister clue.

The alcohol cargo sometimes becomes a shortcut: sailors supposedly drank, became violent, and seized the vessel. That is a story about behavior, not a deduction from the existence of barrels. It requires evidence of access, consumption, conflict, and the resulting acts. The delivered cargo count cannot establish those steps.

There is an additional problem of explanatory economy. If a theory requires an unobserved mutiny, an unobserved killing, a deliberate abandonment, and then a separate unobserved disaster to eliminate all survivors, it has accumulated several events. That does not make it impossible. It does mean that its ability to fit the final scene is not enough; independent support becomes especially important.

The fair conclusion is that violence was considered and not demonstrated by the evidence discussed here. It should remain available for testing if new material appears. It should not be elevated to the default answer merely because a crime offers identifiable villains and a sharper ending than an accident.

20. Pirates, collusion, and the money that came later

A theory of external attack faces a similar demand for sequence. It needs an attacking vessel or party, an encounter, a reason to remove or kill the people, and an explanation for leaving a valuable ship and cargo available for salvage. The lack of a conventional theft pattern does not logically exclude every attack, but it removes an easy motive that a piracy story often assumes.

Collusion between the two captains is another proposal. In that version, the disappearance becomes part of a plan to produce a salvage payment. The existence of a payment supplies a possible financial subject to investigate, but it does not demonstrate an agreement. People can receive compensation for responding to a danger they did not create.

The practical burden of the proposal is substantial. It must explain the timing of the ships’ movements, the participation or exclusion of the other men, the presence of Briggs’s wife and child, and the continued disappearance of everyone from Mary Celeste. It must also fit the letter showing Briggs’s expectations of correspondence and return. A conspiracy that makes all inconvenient evidence disappear with its participants is difficult to test and correspondingly weak as a historical conclusion.

The salvors’ conduct should be assessed through their statements, the condition of the recovered vessel, and the proceedings. It should not be inferred from the mere fact that they had an incentive to recover property. An incentive to perform salvage is the reason a salvage system exists; it is not a confession to manufacturing the need for it.

The later history of Mary Celeste includes a separate insurance-fraud episode under a different captain in 1885. That event often casts a backward shadow over 1872. The shadow is not evidence. A ship can be involved in wrongdoing at one point in its career without every earlier misfortune sharing the same cause. The participants, documents, and acts must be kept separate.

A useful financial inquiry would compare actual ownership interests, policies, claims, and payments. It would ask who stood to receive what under a specific outcome. General statements that someone must have profited do not substitute for that accounting. Nor should the absence of a complete modern reconstruction be treated as proof that the records conceal a plot.

The available case supports investigation of financial motives where documents justify it. It does not establish that the 1872 abandonment was arranged for money. The distinction protects both the historical record and the reputations of people too often turned into characters in somebody else’s solution.

21. Conan Doyle supplied the survivor history lacked

Arthur Conan Doyle’s J. Habakuk Jephson’s Statement, published in 1884, is one of the most consequential documents in the story’s afterlife. Its importance is literary and historical, but it is not testimony from an actual survivor of Mary Celeste. The text can be read directly in the public-domain collection The Captain of the Polestar, and Others. Doyle’s story.

The opening adopts the appearance of a documentary account. It places the discovery in December 1873, changes the ship’s name to Marie Celeste, describes her as being towed into Gibraltar, and invokes a newspaper report. The fictional voyage runs from Boston toward Lisbon, with a cargo including tallow and clocks. These are not minor slips in an otherwise authentic witness statement. They belong to an invented narrative built around a recognizable mystery.

The narrator is a doctor who joins the voyage for his health. The story supplies named passengers, private conversations, an apparent journal, and an increasingly sinister fellow traveler, Septimius Goring. Navigation is manipulated, violence removes people, and the voyage reaches the African coast. A strange stone carried by the narrator becomes the device that saves him and enables the story to have a survivor.

The mechanism is elegant as fiction: the real case lacked an account of the missing interval, so Doyle invented someone able to tell it. The narrative can explain a deserted vessel because its author controls every unseen action. That is exactly the power a historical investigator does not possess.

The story’s racialized characterization also matters. It frames revenge and violence through the prejudices and sensational conventions of its period. Those invented motives cannot be imported into the biographies of the real crew. A fictional villain’s speech is evidence about Doyle’s narrative construction, not about a sailor who left New York in 1872.

The comparison chart highlights a few readily testable differences. A reader does not need to decide whether every scene is plausible to recognize that the text is not an authentic deposition. Its dates, route, cargo, names, and boat details already diverge from the documented case.

The lesson is not that fiction should never use history. It is that a convincing documentary voice can travel beyond its original setting. A detail may be repeated by someone who no longer remembers where it came from. Eventually the invented detail becomes part of what everybody thinks was found on the ship.

The Voynich Manuscript investigation raises a related problem: a fluent explanation can feel like a solution before the method supporting it has been demonstrated. In Mary Celeste’s case, narrative coherence is particularly seductive because the strongest historical evidence leaves the most dramatic scene unwritten.

22. The ship lived on, and even her wreck became disputed

Mary Celeste continued in service after the Gibraltar proceedings. Her physical survival is essential to the case: the 1872 disappearance was not the ship’s final loss. In January 1885, under Captain Gilman C. Parker, she was deliberately run onto a reef off Haiti in an attempted insurance fraud. This was a later event involving a different company and a different set of evidence. Overview of the vessel’s later career.

The distinction prevents two errors. The first is to depict Mary Celeste sinking in 1872. The second is to assume that the later fraudulent wrecking solves the earlier disappearance. A vessel’s name can connect two stories without making their causes identical.

In August 2001, the National Underwater and Marine Agency announced that an expedition had identified remains on Rochelais Reef as Mary Celeste. The announcement cited dimensions, construction materials, fastenings, and wood samples. It expressed strong confidence in the identification. NUMA’s original announcement.

A 2024 article on NUMA’s own website acknowledged a serious complication: wood analysis indicated that a sampled timber was still growing years after Mary Celeste’s 1885 loss. The organization thus preserves both the original confident announcement and a later account of evidence that challenges it. NUMA’s later discussion.

The appropriate conclusion here is that the identification is disputed, not that the 2001 announcement remains unqualified fact. A full reassessment would need the sample’s provenance, the dating method and result, and its relationship to the wreck structure. A piece of wood found at a site is not automatically a piece of the vessel being sought. Conversely, a problematic sample needs to be evaluated within the whole archaeological context.

This is a different question from what happened to Briggs’s company. Even a securely identified 1885 wreck would be the remains of a ship that had undergone another dozen years of service after the disappearance. It would not be an untouched time capsule of the scene encountered by Dei Gratia.

The wreck dispute is therefore useful for its own evidence lesson. Historical identification is an argument built from context and material, not a permanent status conferred by an announcement. New analysis can require a confident claim to be revised. That willingness to revise is a strength of investigation, even when it deprives a famous story of its apparent final chapter.

23. Put the explanations beside the evidence

The competing explanations become easier to assess when each is required to answer the same questions. Why did the people leave? Why was the boat missing? Why could the ship later be sailed to port? What physical traces should the proposed event have left? What happened after departure? An explanation that answers only the first question is not yet a complete solution.

An emergency involving water or a pump connects reported equipment with a familiar practical concern. It still needs evidence of the initiating problem, the captain’s assessment, and the boat’s subsequent history. A vapor scenario connects the cargo with a real hazard, but it needs the release conditions and a historically supported chain from alarm to loss. Either could be combined with weather without making weather alone the entire explanation.

Mutiny and external violence offer human causes but have no demonstrated final sequence in the material examined here. The suspected blood evidence cannot be counted as confirmed support. A financial conspiracy needs actual coordination and transactions, not merely the later existence of a salvage award. Extraordinary explanations add mechanisms for which the case provides no affirmative evidence.

These are not equally supported possibilities simply because none has a complete solution. Some begin with observed objects and ordinary processes. Others depend on additional people, acts, or forces that have not been established. The reader can prefer the explanation with the strongest support while retaining uncertainty about the missing links.

No numerical probabilities are assigned. The record does not supply reliable inputs for precise odds. A percentage would risk making editorial judgment look like a statistical measurement. The comparison instead identifies what supports a theory and what evidence would distinguish it from its rivals.

The most useful future discoveries would be specific: an authenticated continuation of a voyage record, a reliable contemporary sighting of the boat, fuller original testimony about equipment and handling, or a demonstrable physical result tied to the actual vessel’s arrangement. A new retelling that rearranges familiar clues is not the same kind of advance.

The reader’s verdict can therefore be provisional and still rigorous. The strongest broad account is that people left a ship because of an emergency or perceived emergency and were then unable to return or reach safety. That account fits important features without requiring a proven crime. The evidence discussed here does not establish the exact emergency, the moment of departure, or the final fate of every person.

24. The answer the record permits

There are clear answers within the mystery. The vessel was Mary Celeste. She was found without her original company. Men from Dei Gratia sailed her to Gibraltar. The suspected stains examined by Patron did not become confirmed blood evidence. Doyle’s survivor narrative was fiction. The ship’s later wrecking was a separate event. The central disappearance remains unresolved.

That combination is more informative than an undifferentiated label of unsolved. It allows the investigation to close some doors while leaving others open. A hot meal or a fictional murderer does not deserve to remain in the evidence merely because the real ending is missing. Correcting an error is a result, even when it does not reveal the boat’s course.

The surviving documents also change the emotional center of the story. Briggs’s letter returns us to a father thinking about his son’s skates, a husband traveling with his family, and a captain planning where letters might find him. The crew list restores names to the men whose work made the voyage possible. The inquiry records restore the people who boarded, sailed, dived, examined, and argued over the evidence.

None of those people should be reduced to an atmosphere. The company did not exist to furnish a ghost story. The salvors did not become criminals merely because their discovery was strange. The investigators did not fail to think of ordinary explanations; several of those explanations were already being debated in 1873.

What remains is an incomplete chain of events with a human loss at its center. The ship’s survival gives us something to examine, but it also creates the enduring question: what danger seemed greater than leaving her? The missing boat carries the second question: what happened after that choice?

Readers can continue through the Flannan Isles investigation, where an empty workplace and a later inspection produce another problem of timing, or the Franklin expedition investigation, where material discoveries and Indigenous knowledge have progressively changed the account. The wider collection sits alongside our Lake Norman history and the Lake Norman Brief article library.

Mary Celeste’s most defensible ending is a boundary: much can be established about the vessel and the investigation, while the final decisions and the fate of the ten people remain beyond the surviving evidence. The reader is entitled to a judgment about the best explanation. The record is not yet entitled to a claim of certainty.

Sources and further investigation

The principal historical texts below were read in published transcriptions. This article does not claim a new examination of the original court file, the original family letter, or the surviving wreck samples. The watercolors are interpretations; the charts distinguish recorded counts, reported observations, and explicitly hypothetical arithmetic.

1. Gibraltar Chronicle reports, December 1872–March 1873. Contemporary reporting on arrival, inspections, competing interpretations, and the salvage judgment; the early position error and an awkward final transcription passage require caution.
2. Benjamin Briggs’s November 3 letter, with Oliver Cobb’s family narrative. The predeparture letter is distinguished from the later narrative surrounding it.
3. Ricardo Portunato’s diver affidavit. Inspection of the underwater hull at Gibraltar on December 23.
4. Dr. J. Patron’s analytical report. Examination of suspected stains and the sword; the transcription’s heading and internal dates are inconsistent.
5. Published crew list. Names and roles, with spelling variants considered alongside other records.
6. Charles Fay’s corrections to common errors. A later research compilation with references to testimony, registers, and consular correspondence.
7. Cargo discussion and the quoted April 26, 1873 consular report. The historical outturn passage is distinguished from modern commentary about chemistry and tax history.
8. Spencer’s Island and the vessel’s early history. A secondary compilation used for the broad construction and ownership sequence.
9. José Agostinho’s 1940 weather letter. A retrospective account of regional observations; its figures are not shipboard measurements.
10. Royal Museums Greenwich: The mystery of the Mary Celeste. Curatorial overview distinguishing the initial boarding from the salvage passage.
11. Smithsonian: Abandoned Ship. Jess Blumberg’s 2007 account of MacGregor and Richardson’s reconstruction, treated as a proposed explanation.
12. NIOSH: Ethyl alcohol. Primary technical reference for the general flammability of ethanol, not an assay of the historical cargo.
13. Mason Wakley’s public account of the 2026 reporting. Confirmation of renewed experimental work; not a complete methods paper.
14. Arthur Conan Doyle: J. Habakuk Jephson’s Statement. Public-domain fiction examined as a source of later narrative confusion.
15. NUMA’s 2001 wreck announcement and its 2024 discussion of the dating problem. These are compared rather than presenting the initial identification as settled.
16. Mary Celeste’s later career and bibliography. A secondary guide to the separate 1885 loss; the original trial record would be needed for a detailed legal reconstruction.