Coffea stenophylla is a wild West African coffee species that tastes like high-quality arabica but grows at mean annual temperatures 6.2 to 6.8 °C higher than arabica’s. In a blind tasting, 81% of the panel said it was arabica, against 98% and 44% for two real arabicas and 7% for a robusta. It is not yet a commercial crop.
The heat figure is from Davis, Mieulet, Moat, Sarmu and Haggar, “Arabica-like flavour in a heat-tolerant wild coffee species”, Nature Plants 7, 413–418 (2021). The tasting figures are in the authors’ own write-up and in Daily Coffee News‘ report, not in the paper’s abstract. The two arabica controls were a Sidamo from Ethiopia (98%) and a Sul de Minas from Brazil (44%), and the robusta came from Flores, Indonesia (7%); 47% of the judges also called the stenophylla sample “something new”. Almost nothing you can buy today is stenophylla.
Table of Contents
What the research found
Coffea stenophylla is native to Upper West Africa and had been effectively lost to cultivation for most of a century. Researchers relocated it growing wild in Sierra Leone, and also worked with material originally from Ivory Coast that had been kept alive in a collection on Réunion. Both sources went into the 2021 study.
Heat tolerance
The paper’s abstract states that stenophylla grows and crops under the same rainfall conditions as robusta and Liberica, but at mean annual temperatures 6.2 to 6.8 °C higher than arabica. Writing up their own work, the authors put the absolute figure at a mean annual temperature requirement of 24.9 °C, “which is 1.9 °C higher than robusta, and a staggering 6.2–6.8 °C higher than Arabica.”
Note carefully what that figure is and is not. It is a mean annual temperature for the climate the plant grows in — not a peak temperature it survives, and not a heat-shock threshold. It says stenophylla can be farmed in places arabica cannot be farmed at all. It does not say a stenophylla tree will shrug off a heatwave that kills an arabica tree next to it.
| Species | Mean annual temperature it grows at |
|---|---|
| C. stenophylla | 24.9 °C |
| Robusta | 1.9 °C below stenophylla |
| Arabica | 6.2–6.8 °C below stenophylla |
The blind tasting
The Sierra Leone sample was first tasted by Union Hand-Roasted Coffee in London. A full evaluation of the Ivory Coast sample followed in December 2020 and early January 2021, run with CIRAD at its sensory laboratory in Montpellier: judges from three independent panels (Nespresso, Supremo and CIRAD) and four other companies, 15 in all, tasted it blind alongside arabica and robusta. Asked whether the stenophylla sample was arabica, 81% said yes, a share of all their yes/no answers. The authors’ write-up counts 18 judges in total. The authors describe the profile as complex, with natural sweetness, medium to high acidity, fruitiness and good body.
Daily Coffee News also reports an SCA cupping score of 80.25, just over the 80-point specialty line, but that came from a separate cupping: three professional evaluators in London scored the wild Sierra Leone sample in August 2020, from about 10 g of coffee. The caveat that matters most for that score is that it was wild coffee, “wild plants selected at random and crudely processed”, not a cultivated and carefully prepared lot. It is a floor for what the species might do, not a ceiling.
This site has not tasted stenophylla. Nobody reading this page is likely to have the chance to soon. Everything above is what a published panel reported, and it is presented as their result rather than ours.
How much caffeine does stenophylla have?
Similar to arabica’s, according to a 2025 study that compared the two with robusta. It analyzed unroasted beans, including seven wild stenophylla samples from Sierra Leone, found robusta highest, and its authors write: “The caffeine contents of Arabica and stenophylla were found to be similar (TOST P = 0.077).” That equivalence test allowed a difference of up to 25% of the mean, and 0.077 is above the usual 0.05 cutoff. The study compared relative amounts (peak areas) and gives no percentage, so this page prints no caffeine figure. The same study found theacrine, a related alkaloid, in stenophylla but in no arabica or robusta sample, “the first record of this alkaloid in coffee beans” (Jan-Smith et al., npj Science of Food 2025).
Where it grows
A 2025 genetics study of wild Sierra Leonean stenophylla — Lahai and colleagues, “Genetic basis of phenotypic diversity in C. stenophylla“, Frontiers in Genetics — records its collection sites at 375 m, 416 m and 466 m above sea level. That is lowland forest. Arabica’s quality reputation is built on high-altitude farming, where cooler temperatures slow cherry development; stenophylla appears to deliver an arabica-like cup without needing the altitude. If it holds up under cultivation, that is the commercially interesting part.
A taxonomy note, because this gets muddled
Coffea stenophylla is its own species. It is not a variety of arabica, not a variety of robusta, and — despite sharing a West African range — not part of the Liberica group. That last point matters because the Liberica group has recently been reclassified, with C. liberica split into three species including C. dewevrei, the coffee sold as excelsa. Our page on excelsa coffee beans covers that reclassification and its sources; it has nothing to do with stenophylla, and the two questions should not be run together.
For the species you can actually buy, see arabica coffee beans, robusta coffee beans and liberica coffee beans.
Why anyone cares
Arabica supplies most of the world’s speciality coffee and is fussy about temperature. As growing regions warm, the land that suits arabica shrinks and moves uphill, and there is only so much uphill. A species that produces an arabica-like cup at 24.9 °C mean annual temperature, at low elevation, widens the map rather than shifting it. Our page on the climate change impact on coffee farming covers the pressure this is a response to.
What stands between that idea and a bag on a shelf is ordinary and considerable: stenophylla has essentially no agronomic package behind it — no established yields, no disease-resistance data at scale, no processing conventions, no supply chain, and no buyers used to asking for it. The 2025 genetics work above exists precisely because breeders need a genetic map before they can select cultivars. Expect years, not months.
Can you buy stenophylla coffee?
Rarely, and check what you are buying. Because stenophylla is famous and scarce, the name is worth more than the coffee is available. If a retailer offers it, ask which population or collection it came from and expect a specific answer.
If you do get some, treat it as you would an arabica of unknown origin. Start at 55 g of coffee per 1 litre of water — about 1:18 by weight, the standard brew ratio in SCA Standard 310-2021, Home Coffee Brewers, section 7.1.5: “The standard brew ratio of 55 g of coffee per 1.000 kg of 17±3°C water”. Remember that a coffee “cup” is 6 fl oz rather than 8, and that stronger means more coffee, not less. For water temperature see our page on the best temperature for French press coffee. A filter method suits a coffee described as sweet, fruity and medium-to-high in acidity better than espresso does — but this is reasoning from the panel’s descriptors, not from a cup anyone here has drunk.
Corrections to earlier versions of this page
An earlier version of this page published three tables of numbers that were not real. They have been deleted rather than corrected, because correcting an invented table into a different invented table is not an improvement. What was on this page, and why each one had to go:
- A “Flavor Profile Ratings” table scoring stenophylla and arabica out of 10 on fruitiness, sweetness, acidity and body, described as “based on expert tasting panels”. No such panel produced those scores. There was a real sensory panel, described above. Its judges scored attributes such as fruity, acidity and body out of 10, but the paper reports those scores as radar graphs, and the earlier table’s numbers did not come from it.
- A “Nutritional Profile” table giving caffeine content as “1.2 mg/100g” for stenophylla and “1.5” for arabica. Those units are impossible. Arabica caffeine is usually reported as a percentage of dry weight — around 1.2 per cent, which is 1,200 mg per 100 g, a thousand times the figure printed here. The chlorogenic acid, lipid and protein figures in the same table had no source either. The 2025 study cited above compares its caffeine with arabica’s but gives no percentage, so this page prints no caffeine figure at all rather than a plausible-looking one.
- A “Climate Tolerance” table giving stenophylla a maximum temperature tolerance of 30 °C against arabica’s 29 °C. A one-degree advantage is not the finding, and it badly understates the point of the species. The published figure is a gap of 6.2 to 6.8 °C in mean annual temperature — a difference that changes which countries can grow premium coffee, not a rounding.
- It also said stenophylla “prospers in well-drained soils often found in high-altitude areas like Sierra Leone and Liberia”. Stenophylla is a lowland species; that is much of why it matters. The 2025 genetics survey above collected its wild Sierra Leonean material at 375 m, 416 m and 466 m above sea level. For comparison, our page on Kenyan coffee puts that country’s arabica at roughly 1,400–2,100 m.
Also corrected: an earlier version said it had not found a caffeine measurement for stenophylla in any source it had read; the 2025 study above compares its caffeine with arabica’s. It said stenophylla grows at temperatures “hotter than arabica can tolerate”; the paper reports a mean annual temperature 6.2 to 6.8 °C higher than arabica’s. It credited 18 judges to the CIRAD tasting, where the paper’s supplement gives 15, and it reported the London cupping score as part of that tasting. And its list of removed tables said the real panel did not score attributes out of 10, which it did.
