There are roughly 130 coffee species. That is the number botanists currently accept in the genus Coffea, and it is best treated as an approximate, moving figure rather than a settled one: published totals sit anywhere from the low 120s to the mid 130s depending on which checklist you consult and when it was compiled. Two of those species supply almost all the coffee in world trade. A third and a fourth turn up in commerce in tiny volumes. The rest — well over a hundred plants — have never been crops at all, and most of them have never been tasted by anyone outside a research station.
That gap is the real story of this page. A genus with well over a hundred members gave the world a drink made from essentially two of them. Understanding why says more about what farming demands of a plant than any list of Latin names ever could.
How many coffee species are there, and why no source will give you one number
Ask three references and you will get three answers. Global plant checklists have carried figures in the 120s for years; more recent tallies run to the low 130s; some databases quote a higher number still because they count named subspecies and varieties alongside species, or include plants held in living collections that are clearly distinct but not yet formally published under a name. None of these sources is simply wrong. They are counting slightly different things, and the underlying total genuinely grows.
The honest way to state it is this: the number of described coffee species is around 130 and rising. Anyone who gives you a precise figure without a date attached is quoting a snapshot.
Two forces push the count up. The first is survey work. Large parts of the tropical forest where wild Coffea grows — northern Madagascar especially, but also patches of West and Central Africa — have only been surveyed properly in recent decades, and expeditions keep returning with plants that do not match anything already named. As recently as 2021, six species new to science were described from northern Madagascar alone.
The second force is taxonomy itself, and it is the one most people miss.
What “described species” actually means
A described species is not a fact discovered in the forest. It is a judgement, published in the scientific literature, that a group of plants is different enough from its neighbours to deserve its own name. Judgements get revised. When a revision decides that two named things are really one, the count falls. When it decides that one named thing is really two or three, the count rises — without a single new plant being found.
Coffee has recent examples of both, and they are worth knowing because they explain most of the disagreement between sources:
- A merge that added about twenty species. For most of the twentieth century, a closely related group of plants sat in a separate genus called Psilanthus — sometimes labelled “false coffees.” A reassessment published in 2011 concluded the split was not supportable and folded that genus into Coffea. Roughly twenty species joined the genus overnight, and coffee’s native range officially stretched east into tropical Asia and northern Australia, where those plants live. Nothing changed in any forest; a boundary on paper moved.
- A split that turned one species into three. More recently, genomic work on the liberica group concluded that what had long been handled as a single variable species is better understood as three: liberica proper, the plant traded as excelsa (reinstated under the name Coffea dewevrei), and a third, Coffea klainei. Reporting on that work described the known species total moving up by two as a direct result. Again, no new plant — a re-reading of plants already in cultivation and in herbaria.
This is why a page that states the count as a hard number is quietly misleading. The count is a running total of published opinion, and opinion is still being published.
At a glance: the genus versus the crop
| Species | Role in the cup | Rough share of what the world drinks |
|---|---|---|
| Coffea arabica | Almost all specialty coffee and the majority of everyday coffee | Commonly reported at somewhere between 55 and 60 percent of world production, with the share drifting down in recent years |
| Coffea canephora (robusta) | Instant coffee, espresso blends, much of the commodity trade | Effectively all of the remainder |
| Coffea liberica | A real but small trade, with a strong regional following in parts of Southeast Asia | A small fraction of one percent; researchers who have looked closely argue the widely repeated “about one percent” claim is a misquotation and far too high |
| Coffea dewevrei (excelsa) | Small quantities, often blended; recently separated from liberica | Similarly tiny |
| C. racemosa, C. stenophylla, C. eugenioides and a handful of others | Occasional micro-lots, research plots, breeding parents | Negligible — measured in trees, not containers |
| The remaining hundred-plus species | Never grown as crops; many never assessed for flavor at all | None |
If you want the botany of the plant itself — how a coffee tree grows, flowers and fruits — that belongs to our guide to what the Coffea plant is, which also covers the arabica, robusta and liberica trio in detail. This page is about the count and the filter.
The filter: what a coffee species has to survive to become a crop
Farming is a brutally narrow sieve. A wild plant has to pass every one of the following tests, not most of them, before anyone can build an industry on it. Fail one and the species stays in the forest no matter how good it tastes.
It has to yield enough, per tree and per year
This is the killer. Most Coffea species are understorey shrubs that set a modest crop of small fruit, which is a perfectly good strategy for a plant and a hopeless one for a farm. A grower is paying for land, pruning, weeding and harvest labour whether the tree carries a couple of kilos of cherry or a couple of handfuls. Species that were cultivated in the past and then abandoned — including some genuinely admired ones — were generally dropped for exactly this reason: the cup was fine, the harvest was not.
The seed has to be big enough to be worth processing
Coffee is sold by weight of seed, and everything between the tree and the roaster — pulping, drying, hulling, sorting, roasting — is tuned to a seed of a particular size. Several wild species produce seeds so small that the fruit-to-seed ratio makes processing uneconomic in labour terms alone, and the roasting behaviour of a very small, dense seed is different enough that existing equipment and habits do not transfer without rework.
It has to ripen on a schedule a picker can work with
Commercial coffee depends on a reasonably concentrated harvest window. Species that flower opportunistically after rain, or that hold ripe and unripe fruit on the same branch for months, force pickers back over the same tree again and again. Some wild species also take considerably longer from flower to ripe fruit than arabica or robusta, which stretches the growing cycle past what a farm can comfortably absorb.
The architecture has to be harvestable
A coffee tree that grows tall and open, with fruit at the ends of long whippy branches, costs more to pick than one that fruits densely on short laterals within arm’s reach. Liberica is the standing lesson here: a genuinely vigorous species whose sheer size has always complicated its cultivation. We cover that trade-off in the guide to liberica coffee.
It has to carry caffeine — or accept that it will not be sold as coffee
Caffeine content varies enormously across the genus. Some species are reported to carry more than robusta; others carry very little; at least one described from Cameroon is reported to make essentially none at all. A naturally caffeine-free plant is fascinating to breeders and awkward as a straightforward commercial crop, because the drink most people buy is expected to do something. A species that tastes pleasant and does nothing occupies an odd commercial space.
The cup has to be tolerable to a market that already exists
This is the softest test and in practice the most decisive. Coffee drinkers arrived at their expectations through arabica and robusta over centuries. A species that brews to something genuinely unfamiliar — more herbal, more astringent, closer to a tisane than to coffee — has to create its own market rather than slot into one, and that has almost never happened.
There is one more constraint sitting underneath all of these: coffee seed is short-lived. It does not store like grain, losing viability within months rather than years under ordinary conditions. That makes moving a new species around the world, building nurseries and holding genetic material in reserve far more laborious than it is for most crops — a quiet tax on any attempt to domesticate something new.
Where the wild coffee species live
The distribution of the genus is lopsided in a way that surprises most drinkers. Coffee is not a pantropical plant that happens to be farmed in Latin America. Every wild Coffea species is native to the Old World, and the great majority to Africa and the islands off its east coast:
- Continental Africa carries a large share of the genus — published counts for the mainland commonly run to around fifty species or more — spread from West Africa across the Congo basin to the Ethiopian highlands and south towards Mozambique. South-west Cameroon is repeatedly described as the richest single pocket on the mainland.
- Madagascar is the outstanding centre of wild coffee diversity, with a species count that rivals or exceeds the entire African mainland depending on which checklist you use — the great majority of them found nowhere else, many confined to small patches of forest. That story has its own page: Madagascar coffee.
- The Comoros and the Mascarenes (Mauritius and Réunion) contribute a small handful of island endemics with very restricted ranges, several of them of serious conservation concern.
- Tropical Asia and northern Australia hold the roughly twenty species that arrived in the genus through the 2011 merge — the eastern edge of coffee’s natural range, and the part of it least connected to anything in your cup.
Everything grown in Latin America, Asia and Oceania today descends from plants carried out of that Old World range within the last few centuries. The crop travelled; the genus never did.
What the other coffee species are actually like
“Never farmed” does not mean “worthless.” A few of the non-commercial species are studied well enough to sketch, and they show how much variation the filter has been discarding:
- Coffea stenophylla, from Upper West Africa, was grown in the nineteenth century, dropped out of commerce and was relocated growing wild in Sierra Leone in the late 2010s; it has its own guide: stenophylla coffee.
- Coffea racemosa, native to the coast of Mozambique and neighbouring areas, is strikingly drought-hardy, drops its leaves in the dry season and is reported to carry notably little caffeine. Any production is local and very small.
- Coffea eugenioides is widely described as one of arabica’s two ancestral parents. It is reported to be low in caffeine and low yielding, and appears occasionally as a curiosity lot rather than as a crop.
- Many Madagascan species have never been evaluated for beverage use at all. Several are known from a single forest fragment and a few herbarium sheets, which is all the record there is.
Where the record is genuinely thin — and for most of the genus it is very thin indeed — the responsible answer is that nobody knows how these plants taste, yield or behave under cultivation, because nobody has systematically tried. That is not a gap in this article; it is a gap in the literature.
Why the size of the genus matters now
For a century the answer to “how many coffee species are there?” was a trivia question. It stopped being one as the climate envelope for arabica narrowed and leaf rust, borer and drought pressure intensified on the two species that carry the trade. A hundred-plus untested relatives is not an academic inventory; it is the pool from which heat tolerance, disease resistance and drought hardiness would have to come, either as breeding parents or, in a few cases, as crops in their own right. The liberica split matters for the same reason: it did not create a new plant, but it clarified that a heat- and drought-tolerant option had been hiding inside a name.
That pool is also under pressure. Published assessments of the genus have concluded that a large share of wild coffee species — a majority, on the most widely cited work — face a meaningful risk of extinction, driven mainly by forest clearance, climate pressure and the very small ranges many of them occupy. Those figures vary by assessment and by how many species were included, and should be read as indicative rather than exact, but the direction is not seriously disputed, and a striking number of species are not held in any living collection outside their native forest. The count of coffee species is rising on paper at the same time as the plants behind it are getting harder to find.
The bottom line
Roughly 130 coffee species exist as described names in the genus Coffea, with the total drifting upward as forests are surveyed and as taxonomists split and merge what is already known. Two of them — arabica and robusta — account for essentially the entire trade, with liberica and excelsa adding a rounding error and a handful of others appearing as curiosities. The other species were not rejected on taste. They were filtered out by yield, seed size, ripening behaviour, tree shape and the sheer inertia of an industry built around two plants. That filter is why a large genus produced a narrow crop — and why the species it left behind are suddenly worth paying attention to.