The Honeybee You See on a Rooibos Flower Is Not Working

A reader who watched a rooibos field in flower would see honeybees on it, plenty of them, working the yellow blooms the way a honeybee works anything yellow and open. They are not pollinating it. An entomologist who spent years photographing these flowers found that the honeybee takes the nectar and leaves the plant no better off, and that the insects that actually set its seed are a shorter, stranger list: a handful of solitary bees, and, in the Cederberg's own corner of the range, two small wasps built for exactly this flower and little else.

A lock that only the right visitor can spring

Aspalathus linearis is a legume, and a legume flower is not the open dish a daisy offers. Like a pea or a lupin, it hides its nectar behind a hinged lower petal called the keel, and the keel does not simply sit open for anyone. It has to be tripped: an insect has to land in the right place and push in the right way before the mechanism releases, and only then does the flower dust that visitor with pollen and let it at the nectar it came for, entomologist Sarah Gess documented in a 2000 survey of the plant's flower visitors. That nectar is the flower's payment for pollination, and it is not meant to be reachable any other way.

A honeybee has found the other way. On some days, Gess watched honeybees turn up on the flowers in large numbers. A honeybee reaches the nectar from the side, bypassing the keel's trigger point, and comes away with a meal but no pollen to carry onward. It does not pollinate the flowers, she writes, but robs them, taking the nectar illicitly and leaving the keel unsprung.

Who is actually doing it

The insects that do trip the flower are a specific, mostly solitary cast, and which ones show up depends on where in the plant's range you are standing. From the Western Cape across to the Eastern Cape, the reliable visitors are solitary bees: several species each of Chalicodoma and Spinanthidium (relatives of the mason bees, which carry pollen packed on the underside of the abdomen) and of Xylocopa, the carpenter bees, which pack it onto their hind legs instead, a difference plain enough that Gess notes even a casual observer can tell the two groups apart on sight.

Only in the western part of that range, from the Cape east to about Willowmore, does a second, smaller group join the list: pollen wasps, family Vespidae, subfamily Masarinae, a genuinely unusual branch of the wasp family. Most wasps hunt live prey to feed their young. Pollen wasps do not. They provision their nests with pollen and nectar, gathered and carried much the way a bee does, which makes them one of the few wasp lineages in the world that live like bees instead of like wasps. Two, Ceramius clypeatus and C. richardsi, are large enough to trip the flower the same way the solitary bees do, landing on the wing petals and taking a dollop of pollen far forward on the underside. Two more, in the smaller genus Masarina, land on the flower's upper petal instead and come away marked on the front of the head. Gess's own scanning electron micrographs, taken through Rhodes University's microscopy unit, show the pollen sitting in that exact spot, an impact zone about half a millimetre across (the enlarged detail box in her micrographs measures 0.57 millimetres), proof that each species is picking it up in precisely the place its particular approach puts it.

The list runs from specialist to generalist. The two Ceramius appear to visit nothing but Aspalathus. The two Masarina are restricted to the Cape Crotalarieae, the small group of related legumes rooibos belongs to, but not to rooibos alone. The solitary bees keep no such narrow range: none is confined even to that legume group, and they work other flowers when rooibos is not in bloom. "Where present," Gess writes, the Ceramius pair are "the most dependable visitors and therefore the most dependable pollinators," precisely because they have nowhere else to be.

Why the wasps are not everywhere

Dependability has a catch. The pollen wasps need mud, not sand, to build their nest turrets, which means they need standing or seeping water nearby to work it, and they do not nest in loose sandy ground or far from a water source, the same survey notes. The carpenter bees have their own constraint: they need dead plant stems to excavate, pithy ones for the smaller species and woody ones for the larger, so their numbers track the availability of old stems rather than the rooibos itself. A cultivated field supplies neither on its own. Gess's closing line to growers is a plain one: siting a seed-production block without regard for where these insects can actually nest is not a detail, it decides whether the flowers set seed at all. A companion paper she published with Fred Gess in 1994, on the wider Cape group of legumes rooibos belongs to, spells out the same point at length: seed yield in cultivated rooibos runs through a wild insect workforce the field itself does nothing to house.

Rooibos is not alone in this. Honeybush, the Cape's other homegrown legume tisane, runs on the same rule: a 2023 study of Cyclopia found that carpenter bees alone pollinate it, and that honeybees cannot trip the flower open to reach the reward at all. Two indigenous Cape teas, two crops built from flowers that evolved alongside their own particular Cape insects, and neither one sets its seed on the honeybee the rest of agriculture leans on.

None of this is settled science with every gap closed. An Afrikaans-language entomological survey commissioned by the rooibos industry itself, run at Clanwilliam in 1997, found thrips on the flowers during a flowering-season count and could not determine what part, if any, they play in pollination, the report to the trade body records. No one appears to have gone back to settle it in the years since. A cup of rooibos exists because a short list of insects, most of which are not the one a reader would have guessed, keep turning up in the right place at the right angle, and there is still at least one visitor to the flower whose part in the story has never been worked out.