Cacao Pollination by Midges: Why Yields Are So Low

The cacao tree is one of the most economically important crops on the planet, and almost nobody knows that the entire global chocolate industry depends on a roughly 2-millimeter biting fly that breeds in damp leaf litter. Cacao isn’t pollinated by bees. It can’t be. The flowers are too small, too oddly structured, and they offer no nectar reward. Instead, Theobroma cacao relies almost entirely on tiny midges in the family Ceratopogonidae — the same group of insects most people know as no-see-ums or punkies — to move pollen from flower to flower.

The result is a botanical inefficiency that has shaped the entire economic structure of the chocolate industry. Only about 1 to 5 percent of a cacao tree’s flowers ever become pods. Mechanization can’t help — there’s no machine that replaces a midge. And because the midges depend on damp, shaded leaf litter to breed, plantation design has to keep biodiversity, shade, and decomposing organic matter in the picture, which makes intensive monoculture farming counterproductive. This is one of the reasons that 90 percent of the world’s cacao is still grown by smallholder farmers on plots of three hectares or less, and why the industry has 6.5 million producers averaging 56 years old rather than a few thousand industrial operators.

Understanding why cacao is the way it is starts with the flower itself.

Cacao Flowers Are Built for Midges, Not Bees

Cacao flowers are structurally unusual. They grow directly out of the trunk and main branches in a botanical pattern called cauliflory, rather than from leaf axils on small branches the way most flowering plants work. They’re tiny — about 1 to 1.5 cm across — and they hang on long, slender stalks. The structure has five sepals, five petals, and a complicated central arrangement of stamens, staminodes, and a five-lobed pistil, all packed into a space barely large enough to see clearly without a magnifier.

The petals form small hooded cups around the anthers, which means pollen isn’t sitting out in the open the way it is on a daisy or a clover. To reach the pollen, an insect has to physically force its way into the petal hood — and then somehow exit and find another flower’s stigma. Bees can’t do this. Honeybees and most native bees are too large for the petal hoods, their hairy bodies don’t fit the geometry, and the flowers offer essentially no nectar reward. Cacao flowers don’t recruit bee foragers in the way nectar-rich flowers do.

Flowering is also stupendously prolific in a way that doesn’t match the actual reproductive output. A mature cacao tree can produce 10,000 to 100,000 flowers in a year. Of those, only about 1 to 5 percent — and on commercial plantations often as few as 0.3 to 3 percent — become pods that mature to harvest. Most flowers fall off the tree within 24 to 48 hours of opening, unfertilized. This isn’t a defect; it’s how the system is built. The tree puts up enormous numbers of flowers betting that midges will manage to find a tiny fraction of them.

Ceratopogonidae Midges Are the Real Workers

The pollinators are biting midges in the family Ceratopogonidae — a large family of more than 6,200 named species of small flies, mostly under 3 mm long, that breed in damp, decomposing organic matter. The genus most associated with cacao pollination is Forcipomyia, with much of the documented activity coming from species in the subgenus Forcipomyia (Euprojoannisia). Forcipomyia squamipennis, formally described by Ingram and Macfie in 1924, is the species most often cited as a primary pollinator in West Africa. Forcipomyia hardyi fills the equivalent role in Hawaii. Different regions have different dominant species, but the family-level pattern is consistent worldwide.

Biting midges have been proposed as the main cacao pollinators since the 1940s, beginning with Billes (1941) and Posnette (1944), whose Gold Coast (now Ghana) studies established the framework that subsequent research has refined rather than replaced. The midges are small enough to enter the petal hood, fuzzy enough to carry pollen on their bodies, and active during the dawn and dusk hours when cacao flowers are receptive. Adult midges are weak fliers — most individuals don’t disperse more than a few meters from their daytime shelters in any given flight — so a cacao plot needs midge breeding habitat right inside or immediately adjacent to it. Suitable habitat means moist, shaded leaf litter, decomposing plant matter, rotting cacao pod husks left on the orchard floor, and standing water in tree hollows or banana-pseudostem cavities — exactly the kind of environment a traditional shaded cacao plot provides as a side effect of how it’s structured.

The mechanism of pollen transfer is mechanical and almost accidental. A midge enters a flower, contacts the anthers while moving inside the petal hood, picks up pollen on its body hairs, and then visits another flower. If the second flower is on a different tree (cacao is largely self-incompatible — most varieties can’t pollinate themselves), and if the midge happens to make contact with the stigma with enough viable pollen grains, fertilization can occur. Each flower needs roughly 100 to 250 pollen grains to fertilize all its ovules, and a single midge typically carries only a few to 30 grains at a time. The transfer is messy, low-volume, and unreliable, which is why low single-digit flower-to-pod success rates are the norm even in well-managed plots.

The Low Conversion Rate Shapes the Whole Industry

The economic implication of midge pollination is hard to overstate. A cacao tree producing 50,000 flowers in a year and converting 2 percent to pods yields about 1,000 pods. Each pod contains 20 to 50 seeds. After fermentation and drying, a hectare of mature trees produces, on average, 400 to 600 kg of dry beans per year — sometimes much less in unimproved smallholder conditions, occasionally 2,000 kg or more on intensively managed CCN-51 plantations. By comparison, a hectare of coffee can yield 1,000 to 3,000 kg of green beans, and a hectare of corn yields 8,000 to 12,000 kg.

This low yield per tree, combined with the labor-intensive nature of harvesting (pods don’t ripen all at once, so each tree requires multiple visits per season), is what keeps cacao economically rational only at smallholder scale. A 3-hectare smallholder plot producing 1,500 kg of dry beans per year — about average for a Dominican Republic fine-flavor farmer — generates roughly $2,500 in net annual income. There’s no farm-scale efficiency play that turns this into industrial agriculture, because the bottleneck is biology.

The 6.5 million cacao farmers globally, with an average age of 56 years, exist as a structural consequence of midge biology. If cacao could be wind-pollinated, like maize or wheat, it would have been industrialized a century ago. If it could be bee-pollinated, like apples or almonds, it would have followed those crops into managed-pollinator monoculture. Instead, it’s stuck with a wild insect that requires biodiversity, shade, leaf litter, and intact forest edges to thrive — which means the farms have to be small, mixed, and embedded in tropical ecosystems. The structural cocoa price volatility of the last few years has pushed this point hard: when supply contracts, there is no spare industrial capacity to bring online, because the industrial form doesn’t exist.

Plantation Design Either Helps the Midges or Hurts Them

The traditional shaded cacao plot is, almost by accident, a midge habitat. The classic agroforestry design grows cacao under a canopy of taller shade trees — bananas, plantains, fruit trees, and native forest species — with leaf litter accumulating on the ground and decomposing throughout the year. The shade keeps the orchard floor cool and moist; the leaf litter and rotting husks provide breeding substrate for midge larvae; the diversity of vegetation supports adult midge feeding; and the canopy provides cover from wind that would otherwise scatter the small flies before they could find flowers.

When cacao plantations move toward intensive monoculture — full-sun planting, herbicide-cleared orchard floors, no shade canopy, no leaf litter — pollination tends to collapse. Studies in West Africa and Latin America have repeatedly documented pollinator populations and pod set falling sharply in cleared, full-sun plots compared to traditional shaded plots within the same region. The midge populations don’t show up. The flowers fall off. Yields plummet despite the added sunlight that, in theory, should drive more photosynthesis.

This puts cacao agriculture in an unusual position relative to the rest of tropical commodity farming. For most crops — palm oil, soy, cattle — clearing forest and converting to monoculture increases yield. For cacao, it can decrease yield, because the pollinator infrastructure depends on the kind of biodiversity that monoculture destroys. The economic incentive, oddly, aligns with conservation: if you want a productive cacao farm, you need to keep some forest around it, leave the leaf litter on the ground, and avoid the kind of clean-floor management that defines industrial agriculture. The most extreme version of this — wild Bolivian chocolate silvestre, harvested from forest understory — is also the version where the pollinator system needs no help at all.

Hand Pollination Works, But Doesn’t Scale

Researchers have known about midge dependence for more than 80 years, and there have been recurring attempts to bypass the bottleneck through hand pollination. The technique works: a worker takes a fine brush or toothpick, opens a flower, transfers pollen from another tree’s flower onto the stigma, and moves on. Hand-pollinated flowers convert to pods at rates of 50 to 70 percent — roughly 25 to 50 times the natural rate. Some research stations have used this method to double or triple yields on small experimental plots.

The problem is labor. A skilled hand-pollinator can process maybe 200 to 400 flowers per hour. A mature tree has 10,000 to 100,000 flowers per year, which means hand-pollinating a single tree to its full theoretical capacity would take dozens of person-hours per season. Hand-pollinating a 3-hectare smallholder plot with 1,000 to 1,500 trees would require an entirely impractical labor commitment. Some Indonesian and Sri Lankan operations have done partial hand pollination on premium varieties — particularly Criollo, where each pod is worth significantly more — but no one has ever made it work as a primary pollination strategy at commercial scale.

Pheromone manipulation and breeding-substrate enrichment have also been tested. The idea is to attract more midges into a plantation using synthetic versions of the natural pheromones that draw mates, or to deliberately leave pod husks and damp organic debris around the plot to seed midge breeding sites. Trials in Ghana and Latin America have shown some success in increasing local midge populations and pod set, but the results haven’t been consistent enough to translate into a deployable commercial product. As of the mid-2020s, the leading sustainability strategies are still about preserving and restoring midge habitat rather than replacing the midges.

Self-Incompatibility Adds Another Layer

Most cacao varieties are self-incompatible — meaning a flower can’t be successfully pollinated by pollen from the same tree, or even from a closely related clone. Pound first documented this in 1932; the Theobroma form of incompatibility is unusual in that the rejection reaction abscises the entire flower rather than just blocking pollen-tube growth. The midge has to move pollen between genetically distinct trees, which means the geography of a plantation matters: a block of identical clones won’t reliably set pods even with healthy midge populations, because the pollen the midges are moving is functionally the same. CCN-51, the high-yield Ecuadorian hybrid, is partially self-compatible, which is one of the reasons it produces the yields it does — but most fine-flavor varieties, including most Criollo and Trinitario lines, require cross-pollination.

This is why mixed-genetic plantings often outperform monoclonal ones, even before accounting for midge habitat. A plot with three or four genetic lines interspersed gets better pollination success than a plot with a single clone, simply because the pollen the midges are moving is more likely to fertilize successfully. The traditional smallholder farm — with a mix of inherited seedlings, grafted improvements, and unintentional volunteer trees — is often inadvertently optimized for cross-pollination in a way that uniform plantation plantings are not.

For more on the genetic structure of cacao varieties and how the traditional Criollo, Forastero, and Trinitario classification works alongside the modern 10-cluster genetic framework, see the companion piece on cacao genetics.

What This Means for the Chocolate You Buy

Every bar of chocolate, from a 95-cent Hershey’s to a $25 craft single-origin, traces back to midges that nobody pays. The pollinator runs on free labor; no one breeds them in commercial operations; their populations rise and fall with how well farmers happen to maintain the conditions they need. A farmer who clears too aggressively, sprays the wrong insecticide, or moves to a full-sun system loses pollination support whether they realize it or not. A farmer who keeps shade trees, leaves pod husks rotting on the ground, and tolerates some weediness around the plot edges supports the system without doing anything specific.

This is why sustainability stories about cacao tend to focus on agroforestry and biodiversity rather than on direct yield optimization. The yield is downstream of the ecosystem. The ecosystem is what supports the midges. And the midges are the bottleneck on which the entire industry — 6.5 million farmers, 200-plus US craft bean-to-bar makers, a multi-billion-dollar global market — actually depends.

Bees get the press. Midges do the work.