Cacao sets fruit from a tiny fraction of its flowers. These studies ask whether pollination, rather than fertiliser or pest control, is what holds yields down.
11 studies1990 to 2025Each checked against its published abstract
What these studies add up to
Where it has been tested, pollinating flowers by hand raised yields substantially, and by more than fertiliser or insecticide did.
But more pollination does not always mean more pods. In two studies the trees set more young pods and then shed them, and mature pods did not increase. What a tree can feed limits what it keeps.
The tiny midges long credited with pollinating cacao are not the only candidates. On one set of Indonesian farms none was caught at the flowers at all, and which insects can physically reach the pollen is still being worked out.
Hand pollination works far less well where trees are not compatible with their own pollen, which is common in native cacao.
Giving pollinators somewhere to breed, as rotting husks or stems on the ground, increased fruit set in two studies. Neither counted the insects.
More pollination made more young pods, and the tree shed them
What limits the number of pods a cacao tree carries to maturity: pollination, or the tree's own resources?
How
Groups of 14-year-old trees were hand pollinated, left to natural pollination, or had their fruit removed continuously, and were followed from 1983 to 1986 with counts of flowers, wilted young pods and mature pods.
Finding
Hand-pollinated trees set significantly more young pods than naturally pollinated ones, but the number of mature pods did not differ, because more of the young pods wilted. The authors conclude that fruit set is regulated by the tree's supply of assimilate, with wilting as the adjustment.
Does not establish: That hand pollination never raises yield. Later field experiments found large gains; the two results differ in trees, site and method and have not been reconciled in a single study.
Limits: One planting over four years. The abstract does not give the number of trees or the pollination intensity.
How does the number of pollen grains landing on a stigma relate to whether the fruit survives and how many seeds it holds?
How
Flowers of one upper-Amazon clone (UPA 409) in Côte d'Ivoire were pollinated by hand; the style was cut a day later so the grains on each stigma could be counted without harming the fruit.
Finding
Forty-three grains gave half the maximum fruit set at 28 days. Above 115 grains fruit set levelled off, at 88% of ovaries developing at 28 days and 75% at maturity. Seed number rose with pollen up to about 238 grains and levelled at an average of 58 seeds a pod.
Does not establish: How many grains a visiting insect actually delivers, or that the same numbers hold for other clones.
Limits: One clone at one site, pollinated by hand with compatible pollen. The thresholds are that clone's.
What is known about cacao pollination, and could better pollination raise yields?
How
A review of the published literature.
Finding
Insects other than the ceratopogonid midges usually credited also visit cacao flowers, but their role is unknown. Pollination is described as a neglected limit on yield.
Does not establish: Who the effective pollinators are.
Does giving pollinating midges somewhere to breed increase pollination, and does it also help the predators that eat them?
How
Cacao fruit husks were added to plots of a cacao monoculture, crossed with hand-pollination treatments, and fruit numbers, yield and predator numbers were recorded.
Finding
Adding husks increased the number of fruits per tree, and with hand pollination increased final yield. Spiders and skinks also became more numerous, without reducing pollination.
Does not establish: That the gain came from midges specifically, or that it holds on shaded farms where litter is already plentiful.
Limits: One monoculture farm. The abstract gives no effect sizes and does not report catching or counting the midges themselves.
Do cacao flowers attract their pollinating midges by scent?
How
Flower volatiles were collected and analysed; three midge species were reared in the laboratory through several generations and offered natural and synthetic flower odours in choice tests.
Finding
The flower's scent was made up entirely of straight-chain hydrocarbons, which is unusual among flowers. Midges were weakly attracted to the whole natural odour, with no clear difference between species, and not at all to a blend of its four main components.
Does not establish: That scent is what brings midges to flowers on a farm, or which compounds would be needed to attract them.
Limits: Laboratory choice tests with reared insects. The attraction was weak.
Is cacao yield limited by pollination, or by nutrients and pests?
How
A field experiment in Indonesia comparing hand pollination at different intensities with fertiliser and insecticide treatments.
Finding
Pollinating a small share of flowers by hand increased yield per tree by 51%, and pollinating all of them by 161%. Fertiliser and insecticide gave no comparable gain.
Does not establish: That the gain persists over years, or that it pays everywhere.
Limits: One country and a limited period. The labour was costed for the study farms, not for smallholders in general.
Does making a farm friendlier to pollinators raise cocoa production?
How
On three farms in Côte d'Ivoire, rotting banana stems were placed in the plots, with and without insecticide, for the six months before the 2018–19 main harvest, and young and mature pods were counted.
Finding
Young pods increased by 75% on trees within a metre of the rotting material and by 50% on trees further away, but only in plots sprayed against mirids. Mature pods did not increase significantly over the 25 weeks. Soil nutrients on all three farms were judged inadequate.
Does not establish: That closing the pollination gap raises yield by itself. The authors conclude it will do so only when whatever stops young pods maturing is also dealt with.
Limits: Three farms and one season. Pollinators were not counted; the effect on them is inferred from fruit set.
Which insects visit cacao flowers, and what about a farm and its surroundings brings more of them?
How
Flower visitors were recorded on 18 farms in Central Sulawesi against distance to forest, surrounding habitat, canopy cover and leaf litter; leaf litter was then manipulated experimentally on 24 farms.
Finding
Ants and flies were the commonest visitors. No ceratopogonid midge, the group usually named as cacao's pollinator, was captured. Forest and agroforest around the farm and more canopy cover increased visitors; distance to forest did not matter, and adding leaf litter mattered less than the surroundings.
Does not establish: That ants or these flies pollinate cacao, or that midges are absent from Sulawesi as opposed to missed by the sampling.
Limits: One region. A visitor is not necessarily a pollinator, and the authors call them potential pollinators throughout.
Which insects are physically able to reach the pollen and stigma of a cacao flower?
How
Flowers collected in the field and museum specimens of candidate pollinators were scanned in three dimensions by micro-computed tomography, and the insects' body dimensions were compared with the flower's openings.
Finding
The narrowest point is a side opening between the petal hood and the filament. Candidate pollinators were sorted into three levels of likelihood by how many of their body dimensions fit through it. The scans also located reward structures in the flower whose presence had been unclear.
Does not establish: That an insect small enough to enter does pollinate. The authors describe cacao's pollination biology as unresolved.
Limits: A test of fit, using museum specimens. The abstract does not say how many flowers or which insects fell into each level.
What determines where on a cacao branch flower cushions form, and when they flower?
How
Branch architecture was measured and flowering monitored weekly, node by node, on two genotypes grown under medium and heavy shade.
Finding
A cushion formed once a growth unit reached a certain age, with its likelihood set by the node's position and diameter. Heavy shade, a 90% reduction in light, greatly limited cushion formation whatever the branch's architecture. When cushions flowered was not related to architecture.
Does not establish: How much shade is too much on a farm, or the effect on yield. The study measured cushions and flowering, not pods.
Limits: Two genotypes, in the first reproductive phase of young trees.
Which insects actually pollinate cacao, and in what proportions, in different regions.
Why hand pollination multiplied yield in some experiments and only multiplied wilted young pods in others. Soil, tree age and how many flowers were pollinated all differ between them.
Whether trees pollinated to a higher yield can sustain it for years without being depleted.
Whether hand pollination pays once the labour is costed for a typical smallholding.
How heat and drought act on flowering, pollinators and fruit set together. No study read for this page measured all three.
These are questions the studies read for this page leave unanswered. They are not a claim that nobody has answered them.
We would like to set two Google Analytics cookies to count page views, and to count them in our own cookie-free measurement as well. They do not identify you and nothing is shared with advertisers. Decline and the site works exactly the same — see cookies for what each one does.