Somebody Fed Rooibos to Nine Hundred Tilapia

Rooibos is a beverage crop. It grows in the Cederberg, in the Western Cape, and for as long as anyone has grown it commercially the question of what to do with the harvest has had one answer: dry it, pack it, pour hot water over it, sit down. The Ministry has found that answer entirely sufficient since 1931 and has never felt the need to improve on it.

So it took a moment to absorb the news that a freshwater laboratory in Mthatha, in the Eastern Cape, several hundred kilometres from the nearest rooibos field and with no grower anywhere near it, spent eight weeks feeding the stuff to fish. Not as a curiosity. As feed.

Nine hundred tilapia and three diets

A 2025 study in the journal Biology¹ by Grace Okuthe, Bongile Bhomela and Noluyolo Vundisa, at the Freshwater Laboratory of Walter Sisulu University, took 900 Mozambique tilapia larvae, very small things averaging 0.54 g and 2.22 cm, and divided them across nine tanks. Three diets, three tanks each, so every result has a triplicate behind it.

One group got a standard commercial basal diet, with no rooibos added. The second got that diet with a fermented rooibos extract worked in at 30% (w/v). The third got the same, with green rooibos, the unfermented kind. Then everybody waited eight weeks, which is the part of science this Ministry approves of most.

The number a fish farmer actually reads

Feed conversion ratio is grams of feed per gram of fish gained, and lower is better. Feed is the largest single cost in commercial aquaculture, so this one figure is very nearly the whole business.

The control fish came in at 2.32, with a standard deviation of 0.57, which is a wide spread and worth noticing. The fermented-rooibos group managed 1.50 ± 0.25. The green-rooibos group managed 1.41 ± 0.07, the best of the three by a clear margin and the steadiest of the three besides. On the same feed, the rooibos fish were simply getting more out of it.

Prof. Okuthe put the commercial case plainly to Food For Mzansi³: "These may sound like small numbers, but in commercial aquaculture, even fractional improvements in feed conversion translate into substantial cost savings, faster production cycles and reduced environmental impact." Her account of the mechanism is modest and rather nicely put. "These compounds work in harmony with the fish's digestive system, helping them get more from the same amount of feed."

One caution about the numbers you may meet elsewhere. South African agricultural media, both Food For Mzansi and, in Afrikaans, Maroela Media⁴, reported headline figures of roughly 2.83% faster growth and 0.9% improved feed conversion for the green-rooibos group. Those percentages are the press framing. They are not the paper's own reported statistics, and the Ministry could not reconcile them against the ratios above. Take the ratios as the record. The story was also picked up by the aquaculture trade press, which ran its own account of the work².

The number that looks like bad news

Here is where an honest report has to slow down.

Condition factor is a rough plumpness-for-length measure, and on that measure the rooibos fish lost. The control group scored 13.14 ± 4.87. The green-rooibos group scored 6.70 ± 2.53 and the fermented group 6.29 ± 2.45. Read quickly, that is the fish on plain feed coming out ahead.

The paper's own discussion offers a kinder reading: "the control fish might have accumulated more adipose tissue relative to muscle or, conversely, the rooibos extracts encouraged leaner muscle mass gain," and notes that "efficient linear growth and lean muscle deposition are often more desirable than a high condition factor if this implies excessive fat accumulation." That is plausible. A fish that is longer and leaner for its feed is a better fish to sell than a fish that is merely fatter.

But it is an interpretation, not a finding. The study did not rule out the less flattering possibility, and it does not claim to have. Two readings sit on the same number, and the trial as run cannot separate them. The Ministry declines to pick one on the reader's behalf.

Not the molecule anyone expected

Readers who have spent time in this Ministry's filing room will assume they know which compound did the work in the green extract, because green rooibos and aspalathin are usually spoken of in the same breath.

They did not, in this case. The dominant phenolic compound the team identified in the green-rooibos extract was protocatechuic acid, alongside ferulic, vanillic, syringic, caffeic and p-coumaric acids. Aspalathin was not the standout. Whether that reflects the extraction, the batch, or something about the plant material is not a question this trial set out to answer, and the Ministry will not answer it on the authors' behalf either.

What showed up in the liver and spleen

The team ran a micronucleus assay, a standard test for DNA damage, and the rooibos-supplemented diets "did not enhance the frequencies of MN" compared with the control, meaning no detectable increase in DNA damage from the diet. That is a clean result, and it is a narrower one than it sounds: it says the extract did no measurable harm on that test, not that it protected anything.

The tissue work is less comfortable and should not be tucked away. Both rooibos groups showed "fat vacuoles and large hepatocytes with displaced nuclei" in the liver compared with the control fish. In the spleen, the green-rooibos group showed "fewer macrophages within the diffuse white pulp" than the control. The authors call these findings preliminary and say the liver changes "necessitate further investigation into optimal inclusion levels and long-term physiological impacts." A 30% inclusion rate is a great deal of anything, and the paper is asking the obvious question about it rather than waving it past.

They list the rest of the unfinished work themselves: the molecular mechanisms are unknown, gut health and immune response need proper study, and nobody has yet checked what any of this does to the eating quality of the fish. Next they intend to try the extract on other freshwater species, and on shrimp.

What the Cederberg might get out of it

Most commercial fish feed leans on imported or synthetic additives. A locally grown substitute would cut an input cost that otherwise leaves the country, and would open a market for rooibos that has nothing to do with cups, with work attached in harvesting and processing. That is the case Okuthe makes: "This research shows how an indigenous crop can open new economic pathways... it also supports South Africa's commitment under the Nagoya Protocol."

That last clause carries more weight than it looks. Rooibos is one of very few crops in the world where the industry pays a share of its earnings back to the Khoi and San peoples for their traditional knowledge of the plant, an arrangement this Ministry has already catalogued in full. Every new use found for the bush passes through that agreement. A fish-feed market would not be an exception to it. "South Africa has an opportunity to lead here," Okuthe said. "We're combining heritage, science and sustainability in a way that resonates with global markets."

The Ministry's position

Nine tanks. Eight weeks. Fish the length of a paperclip at the start. One promising ratio, one awkward ratio with two possible explanations and no way yet to tell them apart, and a set of livers the authors want a longer look at.

This is not a policy. It is a first result, honestly reported, including the parts that do not flatter it, which is more than can be said for most first results. The Ministry files it under promising and unfinished, notes that a plant which asks nothing of the soil it grows in may turn out to ask very little of a fish either, and closes, as ever, at four.