Fish-Stunning, Oral Exposure and the Dose Gap
The single most common argument for jewel vine's safety is not a study — it is an inference from tradition: people have caught and eaten tuba-stunned fish for centuries across Southeast Asia, without recorded mass poisoning, so the plant "must" be safe. The parent page names this argument and states its conclusion is wrong. This article does the arithmetic behind that conclusion, using real, published concentration and clearance data rather than assertion, and then extends the same honest treatment to two pieces of newer literature — a genuine 2023 ethnobotanical report and a 2024 rat study — that a search engine might surface as apparent evidence of internal use, and that deserve exactly the same non-dispositive, non-alarmist scrutiny.
Table of Contents
- The Short Answer
- The Practice, Briefly
- Where the "The Fish Were Eaten" Argument Breaks
- How Fast Rotenone Actually Disappears From Treated Water
- Where It Goes: Sediment, Tissue and No Bioaccumulation
- Oral Pharmacokinetics: Why Route Changes Everything
- Printing the Gap: Root Concentration Versus Water Concentration
- A Genuine 2023 Ethnobotanical Report, and What It Does and Doesn't Establish
- A 2024 Rat Study, and Why It Doesn't Change the Verdict
- What This Page Is Not Claiming
- Numbered Findings: What Is Not Known
- Key Research Papers
- External Resources
- Connections
- Featured Videos
The Short Answer
Eating a fish stunned with dilute rotenone in a stream or pool is not evidence that eating the root, or a concentrated rotenone product, is safe. The water concentrations used to stun fish are deliberately low, they fall further before a fish is caught and cooked, rotenone does not concentrate in fish flesh above the surrounding water, and mammalian gut absorption and liver metabolism handle a swallowed dose very differently from a fish's gill-absorbed one. Two newer pieces of published research — a 2023 report of internal use by one Indigenous community and a 2024 rat study of a leaf extract — are real, and neither of them changes that conclusion, for reasons explained in full below.
The Practice, Briefly
The parent page describes tuba fishing in detail: pounded Derris elliptica root is swirled through a dammed section of slow-moving stream or a tidal pool, the fish lose coordination within minutes and rise to the surface, and they are collected by hand or basket — then eaten. This is a real, historically and culturally significant practice across Malaysia, Indonesia, the Philippines, Vietnam, Papua New Guinea and the Pacific, and it is not in dispute. What is in dispute is the inference some readers draw from it: that centuries of apparently uneventful fish-eating proves the plant is safe to consume more directly. This article is about why that specific inference fails, not about whether the practice happened or whether the fish were genuinely eaten — they were.
Where the "The Fish Were Eaten" Argument Breaks
Laid out as a syllogism, the argument runs: (1) rotenone-stunned fish were eaten for centuries; (2) no mass poisoning from this practice is recorded; (3) therefore rotenone, and by extension the root that produces it, is safe for humans to consume. The conclusion does not follow, because premise (1) smuggles in an assumption the rest of this article tests directly: that the dose a person receives from eating a stunned fish is comparable, or even in the same broad range, to the dose in a mouthful of root or a rotenone-based product. It is not. The remainder of this article establishes, with real published numbers, three separate reasons the fish-eating dose is dramatically smaller than the root or product dose: concentration (how much rotenone was ever in the water to begin with), clearance (how much of that remains by the time a fish is caught and cooked), and route (how the body handles a swallowed dose differently from a gill-absorbed one). Each is real, measured science, not a rhetorical move to dismiss the tradition.
How Fast Rotenone Actually Disappears From Treated Water
Modern fisheries-management applications of rotenone are the best-documented, most precisely measured analogue available for what happens in a tuba-fishing pool, because regulators require exactly this kind of monitoring before and after a treatment. Three real field and laboratory studies give concrete numbers:
- Diamond Lake, Oregon (2006 treatment). A 2014 Environmental Toxicology and Chemistry study found that rotenone dissipated quickly from lake water — approximately 75% was gone within 2 days, with an average water half-life of 4.5 days under first-order kinetics. Neither rotenone nor its major metabolite, rotenolone, was detected in groundwater, sediment, or aquatic plants.
- Lake Davis, California (2007 treatment). A 2012 study in the same journal measured a rotenone water half-life of 5.6 days, a sediment half-life of 31.1 days, and — the figure that matters most for the fish-eating question — a tissue half-life of 6.1 days in brown bullhead catfish, with none of the formulation's constituents found to bioaccumulate as a result of treatment. The study describes brown bullhead as a rotenone-resistant species, so these tissue figures come from fish that tolerate the poison rather than from fish stunned by it, and rotenone and rotenolone lasted longer in tissue (below detection only within 212 days) than in water.
- Photochemical degradation under simulated sunlight. A 2021 Environmental Science & Technology study measured rotenone half-lives of just 4.2 to 20.1 hours under sunlight exposure typical of the high-latitude lakes it studied — an order of magnitude faster than the multi-day water half-lives measured in the lake field studies above, because direct sunlight is a far more aggressive degradation pathway than dilution and microbial breakdown alone.
A tropical tuba-fishing pool — shallow, sun-exposed, warm — sits closer to the fast, sunlight-driven end of this range than to a deep temperate lake. The parent page's statement that "sunlight and warm water break it down within days" is consistent with, and now backed by, these real measurements.
Where It Goes: Sediment, Tissue and No Bioaccumulation
The Lake Davis study's most important finding for this article is not the half-life number itself but what it says about fish tissue specifically: rotenone and its metabolite rotenolone persisted somewhat longer in sediment and tissue than in water — not detectable in sediment after 157 days, and below detection in tissue within 212 days — but critically, "none of the constituents was found to bioaccumulate in tissues as a result of treatment." Bioaccumulation is the process by which a chemical concentrates in an organism's tissue to levels higher than in its surrounding environment, the way mercury concentrates up a food chain. Rotenone did not do this in Lake Davis. From that finding, this page infers that a fish's tissue concentration does not build up above the water concentration it was exposed to; the study itself did not report tissue levels as matching water levels, and the fish it tested were a rotenone-resistant species.
That single finding does most of the work in this article. By this page's inference (not a measurement the study reports), a fish caught during or shortly after a tuba treatment carries a body burden in roughly the same low concentration range as the water it was swimming in — not a concentrated dose built up over the fish's tissues, which is the scenario that would actually be dangerous.
Oral Pharmacokinetics: Why Route Changes Everything
A 2021 Chemical Research in Toxicology review by Innos and Hickey, specifically examining the pharmacokinetics behind rotenone's use as a Parkinson's disease research tool, states plainly that "rotenone is metabolized extensively in vivo, and choice of route of exposure influences greatly the dose used." The parent page's account of gill absorption versus mammalian first-pass metabolism is the mechanistic explanation for exactly this review's conclusion: a fish absorbs a dissolved, lipophilic molecule directly across its gills into arterial blood, while a mammal swallowing the same molecule sends it through the gut wall and then the liver's cytochrome P450 system before much of it reaches systemic circulation at all. The same review also confirms, from the toxicological literature it surveys, that "rotenone degrades when exposed to light or water" — the mechanism behind the half-life figures above — and notes that even the vehicle a dose is dissolved in measurably changes the outcome in laboratory rotenone-poisoning models, which is a further reason a stunned fish's low-concentration, water-borne exposure is not a stand-in for a concentrated oral dose of root or product.
Long-term exposure is a separate question from a single swallowed dose. In a US case-control study nested in the Agricultural Health Study (110 people with Parkinson's disease and 358 controls), lifetime use of rotenone as a pesticide was associated with Parkinson's disease (odds ratio 2.5; 95% confidence interval 1.3–4.7).
Printing the Gap: Root Concentration Versus Water Concentration
This site's evidence doctrine calls for printing the arithmetic behind a safety or danger claim rather than asserting a conclusion, so that a reader can check it. Here is the comparison, using only figures already established on this site or sourced above, and comparing like with like — oral-route concentrations against other oral-route concentrations, not water concentrations against body-weight doses, which are not directly comparable units.
- Dry derris root, as historically traded: the parent page states this was commonly a few percent of dry weight, reported considerably higher in selected cultivated strains. Call the low end of "a few percent" about 1–5% rotenone by weight — that is 10,000 to 50,000 milligrams of rotenone per kilogram of root (10,000–50,000 parts per million).
- Water in a treated lake or pool: fisheries agencies treat at concentrations described on the parent page as "well under one part per million" — that is, under roughly 1 milligram per kilogram of water, and per the field studies above, that already-low figure is itself falling on a multi-day (or, under strong sun, multi-hour) half-life from the moment of application.
- Fish tissue at the time of catching: by this page's inference from the Lake Davis no-bioaccumulation finding (the study did not compare tissue levels with water levels, and its fish were a rotenone-resistant species), tissue concentration does not exceed water concentration — so a caught fish's flesh would sit in roughly the same sub-one-part-per-million range as the water, not the tens-of-thousands-of-parts-per-million range of the root itself.
Comparing the root's own rotenone concentration (10,000–50,000 ppm) to a treated pool's water concentration and the fish tissue inferred to follow it (well under 1 ppm, and falling) gives a gap on the order of four to five orders of magnitude — a ten-thousand-fold-or-greater difference in concentration, before oral bioavailability and first-pass metabolism reduce the fish-eating exposure even further. That gap, not any special property of fish digestion, is the real reason centuries of tuba-fishing did not produce mass poisoning while the case reports on the parent page describe real deaths from concentrated root or product. This arithmetic describes a gap; it is not a calculation of any "safe" amount of root or rotenone product to consume — no such amount has been established, and the parent page is explicit that none should be sought.
A Genuine 2023 Ethnobotanical Report, and What It Does and Doesn't Establish
A 2023 Journal of Ethnopharmacology paper documented the medicinal plant knowledge of the Kenyah community of Sarawak, Borneo, based on interviews with 24 respondents across four longhouses. Of 61 plant species recorded, only seven were cited by more than 20% of respondents, and the paper specifically flagged four species used for indications "scarcely reported" in the existing ethnobotanical literature. Derris elliptica was one of them — recorded by the Kenyah as a treatment for fever and influenza.
This is real, recent, peer-reviewed field data, and it deserves to be taken seriously as a documented practice rather than dismissed. It is worth being precise about what it does and does not establish:
- What it establishes: at least one specific Indigenous community genuinely uses Derris elliptica internally for fever and influenza. The paper does not appear to be in error, and there is no reason to doubt the ethnographic record.
- What it does not specify: the paper's own abstract does not describe the exact preparation, part used, dose, or frequency — and this site's policy is to refuse a number or a detail it cannot verify rather than guess at one. Whether this use involves the root specifically, what concentration, and how often, are not stated in what this site could verify.
- What it does not establish: safety. An ethnobotanical citation records that a practice exists; it is not a toxicology study, a case-series follow-up, or a pharmacovigilance system. The same paper notes that traditional knowledge in this community is concentrated in respondents over 40 and described by its own authors as being lost as younger community members turn to modern medicine — consistent with a shrinking, largely undocumented-by-modern-surveillance practice, which is a reason for caution about undetected harm rather than reassurance about its absence. "No case reports" from a practice with essentially no modern medical surveillance reflects absent monitoring, not a demonstrated safety record.
- What it does not contradict: the parent page's statement that jewel vine "has never been an internal remedy in any major traditional system" refers specifically to Ayurveda, traditional Chinese medicine, Thai traditional medicine, Jamu and Vietnamese thuốc nam — formal, codified traditional medical systems with texts, training and pharmacopoeias. A single Indigenous community's folk practice, however real and worth documenting, is a different kind of evidence from those systems, and this finding does not contradict the parent page's specific claim. It does add a genuine nuance the parent page does not currently mention: in this survey, Derris elliptica was one of seven plants named by five or more of the 24 respondents, and over 95% of the respondents were 40 years or older.
A 2024 Rat Study, and Why It Doesn't Change the Verdict
A 2024 Saudi Pharmaceutical Journal paper investigated the "therapeutic potential" of a Derris elliptica methanolic leaf extract in streptozotocin-induced diabetic Sprague Dawley rats. Rats received 200 or 400 mg/kg of the extract orally for 14 days, alongside acute oral toxicity testing in normal rats. The study reported no signs of toxicity or mortality, a significant reduction in blood glucose and body weight, reduced total cholesterol, increased insulin secretion at the higher dose, and improved liver-enzyme markers — concluding that the extract "may have therapeutic potential for the treatment of diabetes mellitus."
This is a real, peer-reviewed paper, not a fabrication, and it deserves the same direct treatment as the ethnobotanical finding above: taken seriously, and precisely contextualised, rather than either hidden or credited more than it earns.
- The plant part is different, and that difference plausibly matters. The parent page is explicit that rotenone concentrates in the root of Derris elliptica, which is the part traded, studied in poisoning cases, and discussed throughout this site. This 2024 study used a leaf extract. A leaf plausibly carries a substantially different, likely much lower, rotenoid content than root material — this site has not located a direct assay comparing the two in this specific study, so the magnitude of the difference is not asserted here, only the direction the plant's own established chemistry would predict.
- The paper's own historical framing is not corroborated elsewhere. Its introduction states that D. elliptica "has been used to treat diabetes for centuries," without a verifiable citation this site could trace. No traditional-use record located anywhere in this Benefits leg's research — including the genuine Kenyah ethnobotanical report above, which documents fever and influenza, a different indication, in a different community — corroborates a centuries-old diabetes tradition. Per this site's policy of refusing an unverifiable number or claim, that specific sentence in the paper's own introduction is flagged here as unsupported rather than repeated as background fact.
- It is one small, short, unreplicated study. Fourteen days of dosing, a single research group, no independent replication located. The parent page's own "Why It Turns Up on Herb Lists" section already names the general pattern this instance fits: laboratory bioactivity findings in a plant with a dangerous reputation are periodically published and then read by others as more supportive than the underlying study can bear. A single rodent study reporting no observed toxicity over two weeks, in a plant part not otherwise implicated in the poisoning literature, does not establish long-term safety, does not establish human relevance, and does not establish that root material would behave the same way.
- Nothing here changes the verdict on root material or human use. This study cannot be read as evidence that any part of Derris elliptica, taken by mouth, is safe or beneficial for a person with diabetes or any other condition. The parent page's warning stands, in full, regardless of this finding.
What This Page Is Not Claiming
- This is not a claim that tuba-fishing is dangerous to the people who practise it. The dose-gap arithmetic above is consistent with, and helps explain, the historical absence of mass poisoning from eating stunned fish. That is different from any claim about the root or concentrated products.
- This is not a dosing calculation. No figure on this page describes a safe amount of root, extract or rotenone product for a person to consume. The arithmetic above measures a gap between two concentrations; it does not define a floor beneath which consumption becomes acceptable.
- This is not a dismissal of either the Kenyah ethnobotanical record or the 2024 rat study as fake or worthless. Both are real, peer-reviewed contributions to the literature, reported here in full rather than omitted, and neither is presented as proof of anything beyond what it actually shows.
- This is not the final word on either finding. Further research, including species-and-part-specific rotenoid assays and any follow-up safety surveillance on the Kenyah practice, could add real information this article does not currently have.
Numbered Findings: What Is Not Known
- The rotenoid content of the leaf extract used in the 2024 diabetic-rat study, compared directly to root material from the same plant population, was not located — the part-substitution argument above is a plausible inference from established plant chemistry, not a directly measured comparison.
- The exact preparation, dose, frequency and part used in the Kenyah fever/influenza practice are not specified in the source this site could verify.
- No pharmacovigilance or case-series follow-up on the Kenyah practice was located, so its real-world safety record, positive or negative, is unmonitored rather than established.
- A direct, controlled comparison of fish-tissue rotenone concentration immediately after a tuba-fishing treatment, using modern analytical methods, was not located — the figures used above come from regulated modern fisheries-management treatments, which is the closest available analogue but not an identical practice.
- No human oral pharmacokinetic study of rotenone at any dose was located. The absorption and metabolism discussion above draws on animal pharmacokinetic data and general toxicological principles, not a human trial.
What is known about swallowed rotenone in people comes mainly from poisoning reports. A 2024 scoping review of acute rotenone poisoning screened 2,631 publications and identified 11 case reports describing 18 victims, 15 of whom were hospitalised and five of whom died. Most cases occurred in private homes, where the victims unknowingly ate rotenone-containing plants. Vomiting and metabolic acidosis (a build-up of acid in the blood) were the most common signs. There is no antidote, and treatment is supportive. The reviewers add that cases may be under-diagnosed because the symptoms are non-specific, and under-reported in English-language journals because most cases occurred in non-English-speaking countries.
Key Research Papers
- Innos J, Hickey MA. Using rotenone to model Parkinson's disease in mice: a review of the role of pharmacokinetics. Chemical Research in Toxicology. 2021;34(5):1223-1239. — PubMed PMID: 33961406
- Finlayson BJ, Eilers JM, Huchko HA. Fate and behavior of rotenone in Diamond Lake, Oregon, USA following invasive tui chub eradication. Environmental Toxicology and Chemistry. 2014;33(7):1650-1655. — 4.5-day water half-life. — PubMed PMID: 24733691
- Vasquez ME, Rinderneck J, Newman J, McMillin S, Finlayson B, Mekebri A, Crane D, Tjeerdema RS. Rotenone formulation fate in Lake Davis following the 2007 treatment. Environmental Toxicology and Chemistry. 2012;31(5):1032-1041. — 5.6-day water half-life, no tissue bioaccumulation (brown bullhead catfish, a rotenone-resistant species). — PubMed PMID: 22407799
- Redman ZC, Wesolowski J, Tomco PL. Photochemical pathways of rotenone and deguelin degradation: implications for rotenoid attenuation and persistence in high-latitude lakes. Environmental Science & Technology. 2021;55(8):4974-4983. — hours-scale half-lives under sunlight. — PubMed PMID: 33724011
- Cavoski I, Caboni P, Sarais G, Cabras P, Miano T. Photodegradation of rotenone in soils under environmental conditions. Journal of Agricultural and Food Chemistry. 2007;55(17):7069-7074. — 5–7 hour half-life under direct sunlight. — PubMed PMID: 17658825
- Melo KM, Oliveira R, Grisolia CK, Domingues I, Pieczarka JC, de Souza Filho J, Nagamachi CY. Short-term exposure to low doses of rotenone induces developmental, biochemical, behavioral, and histological changes in fish. Environmental Science and Pollution Research. 2015;22(18):13926-13938. — 96-hour zebrafish LC50 of 12.2 µg/L. — PubMed PMID: 25948382
- Sundara Rajoo K, Lepun P, Alan R, Singh Karam D, Abdu A, Rosli Z, Izani N, James Gerusu G. Ethnobotanical study of medicinal plants used by the Kenyah community of Borneo. Journal of Ethnopharmacology. 2023;301:115780. — records Derris elliptica use for fever and influenza. — PubMed PMID: 36202163
- Abd Rahman R, Jayasingh Chellammal HS, Ali Shah SA, Mohd Zohdi R, Ramachandran D, Mohsin HF. Exploring the therapeutic potential of Derris elliptica (Wall.) Benth in streptozotocin-induced diabetic rats: phytochemical characterization and antidiabetic evaluation. Saudi Pharmaceutical Journal. 2024;32(4):102016. — PubMed PMID: 38463183
- Wiwattanapatapee R, Sae-Yun A, Petcharat J, Ovatlarnporn C, Itharat A. Development and evaluation of granule and emulsifiable concentrate formulations containing Derris elliptica extract for crop pest control. Journal of Agricultural and Food Chemistry. 2009;57(23):11234-11241. — PubMed PMID: 19904939
- Zhou Y, Wang K, Yan C, Li WS, Li H, Zhang N, Zhang ZX. Effect of two formulations on the decline curves and residue levels of rotenone in cabbage and soil under field conditions. Ecotoxicology and Environmental Safety. 2014;104:23-27. — PubMed PMID: 24632119
- Tat J, Heskett K, Boss GR. Acute rotenone poisoning: a scoping review. Heliyon. 2024;10(7):e28334. — PubMed PMID: 38633629
- Tanner CM, Kamel F, Ross GW, et al.. Rotenone, paraquat, and Parkinson's disease. Environmental Health Perspectives. 2011;119(6):866-872. — PubMed PMID: 21269927
External Resources
- PubMed — the index behind every citation on this page.
- US EPA — Pesticide Ingredients — current rotenone registration and use data.
- National Poison Control Center (US) — for any suspected pesticide or plant-poisoning exposure.
- US Fish and Wildlife Service — background on modern rotenone use in fisheries management.