A young French naval officer crosses the Atlantic with a coffee seedling in his care. Water runs short. He gives the plant half of his own ration, week after week, until the ship reaches Martinique. From that one surviving seedling grows the coffee of the Americas. The problem with the story is not that it is unlikely. Plants did cross oceans in the hands of officers, and water did run short. The problem is that the account has exactly one witness, and the witness is the hero.
Gabriel de Clieu was a real naval officer and coffee did reach Martinique in the 1720s, but the heroic voyage comes from de Clieu's own published account of himself, written many years after the fact, and the planting record points elsewhere. Jonathan Morris, in Coffee: A Global History (2019), dates the first planting to 1724 and traces the material not to Paris but to two colonial crops already in the ground, one on Bourbon and one in Suriname. The right way to read the water ration is as memoir with something at stake, not as record.
One source, and it is the man in the story
Morris introduces the whole episode with the word supposedly, which is a historian marking a claim he cannot corroborate. The claim runs like this. In 1723 de Clieu took living cuttings out of the Paris botanic gardens and landed them on Martinique. Everything downstream of it, the rationing and the rival passenger who tries to destroy the plant, arrives from the same pen.
Set that against the coffee stories in the same period that do have paperwork. The Mecca trial of 1511 has a named city pasha, Kh'air Beg, three physicians called to testify that the drink caused intoxication, a partial reversal in Cairo, and Beg's dismissal the following year. The 1591 fatwa settling whether a roasted bean still counted as vegetable matter under religious law carries the name of the man who issued it, Sheik ul-Islam Bostanzade Mehmed Effendi. Those events survive because people other than their protagonists wrote them down while they were happening.
De Clieu has de Clieu. That is not proof of invention. It is the reason the story cannot be promoted past anecdote, and the reason it should never have been printed as the founding fact of coffee in the Americas.
The legend says cuttings, the record says seed
This is the discrepancy that gets skipped, and it is not cosmetic.
The story requires living plants. A cutting is a piece of green tissue that has to stay alive in a pot for the length of an ocean crossing, which is precisely what generates the drama about water. The planting record Morris describes involves seed from two other colonies. Seed is cargo. It sits in a sack and needs nothing from anybody's canteen.
Look at where that seed is said to have come from and the dull version writes itself. French planters had coffee growing on Bourbon, the island now called Réunion, from 1715. The stock had come out of Yemen while the Ottoman authorities were banning exactly that trade, and it took so well that the company running the island turned planting into an obligation: by 1724 the French East India Company would take a land concession back from any settler who left it unplanted, and it had gone as far as discussing execution for damaging a tree on purpose. Suriname is the second name in the record. The Dutch put coffee there in 1712 and the colony was shipping by 1721. Two established crops, both already in production, both sitting inside European shipping networks that ran to the Caribbean.
So the boring explanation for coffee arriving on Martinique is that it moved sideways through a colonial system that was by then treating the crop as state property. The Baba Budan story has the same shape. A pilgrim comes home from Mecca with coffee seed hidden on him and plants it in Malabar, and Morris's flatter reading is that Malabar had coffee because the Banyan merchant network was already carrying it there. Coffee's transfers keep turning out to be logistics wearing a hero costume.
Even a careful book carries the date forward
Worth naming, because it shows how hard a memoir is to get back out of the record once it is in.
Morris's own chronology table lists 1723 as the date of the first French plantation on Martinique. His body text, a few dozen pages away, says planting appears to have begun in 1724 with seed from two other origins. Both are in the same book. The table is carrying de Clieu's date and the prose is correcting it.
That is not a reason to distrust the book. It is a demonstration of the actual mechanism by which these stories persist. A date that entered the literature through a self-published heroic account gets copied into tables, and tables get copied into everything else, and by the time a historian is careful about it in the text the number has already been reprinted for two hundred years.
What he had to gain
Here the article is reasoning rather than reporting, and it should say so.
I cannot show you de Clieu's motive. No document in our sources states it. What can be shown is that the incentive existed and was large. By the mid-1720s coffee had become a matter of colonial policy severe enough that a company would take back your land for not growing it. In that setting, a documented claim to have personally originated a colony's crop is not a charming anecdote. It is standing, and standing converts into position and property.
Treat that as the frame, not the finding. The finding is narrower: the story is unwitnessed, it conflicts with the planting record on both the date and the material, and the man who told it was its beneficiary.
What this article will not do
It will not tell you that de Clieu never carried a coffee plant across the Atlantic. He may well have. Naval officers moved botanical material routinely, botanic gardens existed to supply it, and multiple such transfers almost certainly happened without anyone writing a memoir. What collapses under examination is the specific heroic version: the shared water ration and the single ancestral tree presented as the origin point of an entire hemisphere's agriculture.
It also will not tell you which coffee trees in Martinique descended from what. Nobody in these sources can. The genetic narrowness of cultivated arabica is real and the founder populations were tiny, but tracing a living Caribbean tree back to a named 1720s plant is not something the historical record supports, and any article that offers you that lineage is guessing.
The part of the Martinique story that is documented
Martinique became a real origin with a real market. Ukers, writing in 1922, records that Balzac blended Mocha, Bourbon and Martinique for his own coffee, buying each from a different shop in a different Paris street, an errand that took half a day. That is a checkable fact about a named man's habit, and it tells you Martinique coffee had a distinct commercial identity a century after the planting.
And then there is the part the water-ration story sits on top of. Four fifths of the world's coffee at the end of the eighteenth century came off Caribbean plantations, most of it from the French colony of Saint Domingue, and the cost structure that made that possible was enslaved people. The 1789 count for the colony runs to 465,000 enslaved against 30,000 whites and 28,000 free people of colour. P. J. Laborie's 1798 growing manual is the document that hands down the water-channel pulping method the trade called the West Indian process. The same manual tells a planter to wash a whip before it is used on the next person, so that infection does not travel between them.
Notice what got remembered. Out of the entire French Caribbean coffee system, the founding story that survived into general circulation is about one officer being thirsty on a boat. That is not an accident of retelling. A story about a man's personal sacrifice is a comfortable thing to put at the start of a history whose actual first chapter is a slave colony.
How to hold it
Keep the story. Downgrade the claim. Say "de Clieu wrote that he shared his water," which is true and is the more interesting sentence anyway, and stop saying "de Clieu shared his water," which our sources cannot support.
The same test does useful work on things you are being sold today. A farm story on the back of a bag is a single-source account published by the party it flatters, which is structurally identical to a memoir written by a naval officer with a land claim. That does not make it false. It makes it a claim rather than a fact, and the difference shows up in what a farm name is actually allowed to signal.
Coffee history rewards this kind of reading. The goatherd legend turns out to have a firm first-print date and no Ethiopian source, the papal blessing turns out to have interested parties instead of documents, and the earliest person you can actually attach to coffee is a mufti with a death date rather than anyone with a good anecdote. The romance does not evaporate when you check. It relocates.